task e2e-llm-inference-service has failed: "step-fail-if-needed" exited with code 1: Error [get-kubeconfig] Found kubeconfig secret: cluster-gszfc-admin-kubeconfig [get-kubeconfig] Wrote kubeconfig to /credentials/cluster-gszfc-kubeconfig [get-kubeconfig] Found admin password secret: cluster-gszfc-admin-password [get-kubeconfig] Retrieved username [get-kubeconfig] Wrote password to /credentials/cluster-gszfc-password [get-kubeconfig] API Server URL: https://a0e16ce1ce6064cbb9315ffe26d645d0-c337f2f095db2de9.elb.us-east-1.amazonaws.com:6443 [get-kubeconfig] Console URL: https://console-openshift-console.apps.9f3251de-dc08-4a17-bab2-d85f341a8f2d.prod.konfluxeaas.com [clone-repo] autofix/rhoaieng-71245 [clone-repo] https://github.com/opendatahub-io/kserve [clone-repo] Cloning into '/workspace/source'... [clone-repo] Updating files: 90% (2794/3093) Updating files: 91% (2815/3093) Updating files: 92% (2846/3093) Updating files: 93% (2877/3093) Updating files: 94% (2908/3093) Updating files: 95% (2939/3093) Updating files: 96% (2970/3093) Updating files: 97% (3001/3093) Updating files: 98% (3032/3093) Updating files: 99% (3063/3093) Updating files: 100% (3093/3093) Updating files: 100% (3093/3093), done. [e2e-llm-inference-service] + bash [e2e-llm-inference-service] + STATUS_FILE=/test-status/deploy-and-e2e-status [e2e-llm-inference-service] + echo failed [e2e-llm-inference-service] + COMPONENT_NAME=kserve-agent-ci [e2e-llm-inference-service] ++ jq -r --arg component_name kserve-agent-ci '.[$component_name].image' [e2e-llm-inference-service] + export KSERVE_AGENT_IMAGE=quay.io/opendatahub/kserve-agent@sha256:0b6de7012f6344e0be5c95ac8deb96d56b96edd624ee25e9ad8ecbad40286f86 [e2e-llm-inference-service] + KSERVE_AGENT_IMAGE=quay.io/opendatahub/kserve-agent@sha256:0b6de7012f6344e0be5c95ac8deb96d56b96edd624ee25e9ad8ecbad40286f86 [e2e-llm-inference-service] + COMPONENT_NAME=kserve-controller-ci [e2e-llm-inference-service] ++ jq -r --arg component_name kserve-controller-ci '.[$component_name].image' [e2e-llm-inference-service] + export KSERVE_CONTROLLER_IMAGE=quay.io/opendatahub/kserve-controller@sha256:d717f59e7fc6132421a696b06ade56c2521428323575a477d0441e0491bdb66c [e2e-llm-inference-service] + KSERVE_CONTROLLER_IMAGE=quay.io/opendatahub/kserve-controller@sha256:d717f59e7fc6132421a696b06ade56c2521428323575a477d0441e0491bdb66c [e2e-llm-inference-service] + COMPONENT_NAME=kserve-router-ci [e2e-llm-inference-service] ++ jq -r --arg component_name kserve-router-ci '.[$component_name].image' [e2e-llm-inference-service] + export KSERVE_ROUTER_IMAGE=quay.io/opendatahub/kserve-router@sha256:3bfaa94d16d20ed0bc6642496b738f1c154c88684ea41dc8df471bf4e0a2804c [e2e-llm-inference-service] + KSERVE_ROUTER_IMAGE=quay.io/opendatahub/kserve-router@sha256:3bfaa94d16d20ed0bc6642496b738f1c154c88684ea41dc8df471bf4e0a2804c [e2e-llm-inference-service] + COMPONENT_NAME=kserve-storage-initializer-ci [e2e-llm-inference-service] ++ jq -r --arg component_name kserve-storage-initializer-ci '.[$component_name].image' [e2e-llm-inference-service] + export STORAGE_INITIALIZER_IMAGE=quay.io/opendatahub/kserve-storage-initializer@sha256:13ad765c0f323200e1db7b45c99661287e65c5158d6b54b16139829636d20be4 [e2e-llm-inference-service] + STORAGE_INITIALIZER_IMAGE=quay.io/opendatahub/kserve-storage-initializer@sha256:13ad765c0f323200e1db7b45c99661287e65c5158d6b54b16139829636d20be4 [e2e-llm-inference-service] + COMPONENT_NAME=odh-kserve-llmisvc-controller-ci [e2e-llm-inference-service] ++ jq -r --arg component_name odh-kserve-llmisvc-controller-ci '.[$component_name].image' [e2e-llm-inference-service] + export LLMISVC_CONTROLLER_IMAGE=quay.io/opendatahub/odh-kserve-llmisvc-controller@sha256:611712697e2f5e409e4ec5b04fe487b6af8ff2eadd54cbb3a7ef8d15d81f4ae5 [e2e-llm-inference-service] + LLMISVC_CONTROLLER_IMAGE=quay.io/opendatahub/odh-kserve-llmisvc-controller@sha256:611712697e2f5e409e4ec5b04fe487b6af8ff2eadd54cbb3a7ef8d15d81f4ae5 [e2e-llm-inference-service] + ./test/scripts/openshift-ci/run-e2e-tests.sh 'llminferenceservice and cluster_cpu and not autoscaling and not tracing' 2 llm-d [e2e-llm-inference-service] Installing on cluster [e2e-llm-inference-service] Using namespace: kserve for KServe components [e2e-llm-inference-service] SKLEARN_IMAGE=quay.io/opendatahub/sklearn-serving-runtime:odh-pr-1690 [e2e-llm-inference-service] OPT_125M_MODEL_URI=s3://example-models/facebook/opt-125m [e2e-llm-inference-service] ERROR_404_ISVC_IMAGE=quay.io/opendatahub/error-404-isvc:odh-pr-1690 [e2e-llm-inference-service] SUCCESS_200_ISVC_IMAGE=quay.io/opendatahub/success-200-isvc:odh-pr-1690 [e2e-llm-inference-service] [INFO] Installing Kustomize v5.8.1 for linux/amd64... [e2e-llm-inference-service] [SUCCESS] Successfully installed Kustomize v5.8.1 to /workspace/source/bin/kustomize [e2e-llm-inference-service] v5.8.1 [e2e-llm-inference-service] make: Entering directory '/workspace/source' [e2e-llm-inference-service] [INFO] Installing yq v4.52.1 for linux/amd64... [e2e-llm-inference-service] [SUCCESS] Successfully installed yq v4.52.1 to /workspace/source/bin/yq [e2e-llm-inference-service] yq (https://github.com/mikefarah/yq/) version v4.52.1 [e2e-llm-inference-service] make: Leaving directory '/workspace/source' [e2e-llm-inference-service] Installing KServe Python SDK ... [e2e-llm-inference-service] [INFO] Installing uv 0.7.8 for linux/amd64... [e2e-llm-inference-service] [SUCCESS] Successfully installed uv 0.7.8 to /workspace/source/bin/uv [e2e-llm-inference-service] warning: Failed to read project metadata (No `pyproject.toml` found in current directory or any parent directory). Running `uv self version` for compatibility. This fallback will be removed in the future; pass `--preview` to force an error. [e2e-llm-inference-service] uv 0.7.8 [e2e-llm-inference-service] Creating virtual environment... [e2e-llm-inference-service] warning: virtualenv's `--clear` has no effect (uv always clears the virtual environment) [e2e-llm-inference-service] Using CPython 3.9.25 interpreter at: /usr/bin/python3 [e2e-llm-inference-service] Creating virtual environment at: .venv [e2e-llm-inference-service] /workspace/source [e2e-llm-inference-service] Using CPython 3.11.13 interpreter at: /usr/bin/python3.11 [e2e-llm-inference-service] Creating virtual environment at: .venv [e2e-llm-inference-service] Resolved 266 packages in 1ms [e2e-llm-inference-service] Building kserve @ file:///workspace/source/python/kserve [e2e-llm-inference-service] Downloading pandas (12.5MiB) [e2e-llm-inference-service] Downloading kubernetes (1.9MiB) [e2e-llm-inference-service] Downloading cryptography (4.3MiB) [e2e-llm-inference-service] Downloading aiohttp (1.7MiB) [e2e-llm-inference-service] Downloading setuptools (1.2MiB) [e2e-llm-inference-service] Downloading uvloop (3.8MiB) [e2e-llm-inference-service] Downloading pydantic-core (2.0MiB) [e2e-llm-inference-service] Downloading portforward (3.9MiB) [e2e-llm-inference-service] Downloading black (1.6MiB) [e2e-llm-inference-service] Downloading grpcio (6.4MiB) [e2e-llm-inference-service] Downloading grpcio-tools (2.5MiB) [e2e-llm-inference-service] Downloading pyarrow (40.1MiB) [e2e-llm-inference-service] Downloading mypy (17.2MiB) [e2e-llm-inference-service] Downloading botocore (12.9MiB) [e2e-llm-inference-service] Downloading numpy (15.7MiB) [e2e-llm-inference-service] Building timeout-sampler==1.0.3 [e2e-llm-inference-service] Building python-simple-logger==2.0.19 [e2e-llm-inference-service] Downloading aiohttp [e2e-llm-inference-service] Downloading black [e2e-llm-inference-service] Downloading pydantic-core [e2e-llm-inference-service] Downloading grpcio-tools [e2e-llm-inference-service] Downloading setuptools [e2e-llm-inference-service] Built python-simple-logger==2.0.19 [e2e-llm-inference-service] Downloading portforward [e2e-llm-inference-service] Downloading uvloop [e2e-llm-inference-service] Downloading cryptography [e2e-llm-inference-service] Downloading grpcio [e2e-llm-inference-service] Downloading kubernetes [e2e-llm-inference-service] Built timeout-sampler==1.0.3 [e2e-llm-inference-service] Built kserve @ file:///workspace/source/python/kserve [e2e-llm-inference-service] Downloading numpy [e2e-llm-inference-service] Downloading pandas [e2e-llm-inference-service] Downloading botocore [e2e-llm-inference-service] Downloading pyarrow [e2e-llm-inference-service] Downloading mypy [e2e-llm-inference-service] Prepared 101 packages in 1.87s [e2e-llm-inference-service] warning: Failed to hardlink files; falling back to full copy. This may lead to degraded performance. [e2e-llm-inference-service] If the cache and target directories are on different filesystems, hardlinking may not be supported. [e2e-llm-inference-service] If this is intentional, set `export UV_LINK_MODE=copy` or use `--link-mode=copy` to suppress this warning. [e2e-llm-inference-service] Installed 101 packages in 606ms [e2e-llm-inference-service] + aiohappyeyeballs==2.6.1 [e2e-llm-inference-service] + aiohttp==3.13.3 [e2e-llm-inference-service] + aiosignal==1.4.0 [e2e-llm-inference-service] + annotated-doc==0.0.4 [e2e-llm-inference-service] + annotated-types==0.7.0 [e2e-llm-inference-service] + anyio==4.9.0 [e2e-llm-inference-service] + attrs==25.3.0 [e2e-llm-inference-service] + avro==1.12.0 [e2e-llm-inference-service] + black==24.3.0 [e2e-llm-inference-service] + boto3==1.37.35 [e2e-llm-inference-service] + botocore==1.37.35 [e2e-llm-inference-service] + cachetools==5.5.2 [e2e-llm-inference-service] + certifi==2025.1.31 [e2e-llm-inference-service] + cffi==2.0.0 [e2e-llm-inference-service] + charset-normalizer==3.4.1 [e2e-llm-inference-service] + click==8.1.8 [e2e-llm-inference-service] + cloudevents==1.11.0 [e2e-llm-inference-service] + colorama==0.4.6 [e2e-llm-inference-service] + colorlog==6.10.1 [e2e-llm-inference-service] + coverage==7.8.0 [e2e-llm-inference-service] + cryptography==46.0.5 [e2e-llm-inference-service] + deprecation==2.1.0 [e2e-llm-inference-service] + durationpy==0.9 [e2e-llm-inference-service] + execnet==2.1.1 [e2e-llm-inference-service] + fastapi==0.136.3 [e2e-llm-inference-service] + frozenlist==1.5.0 [e2e-llm-inference-service] + google-auth==2.39.0 [e2e-llm-inference-service] + grpc-interceptor==0.15.4 [e2e-llm-inference-service] + grpcio==1.78.1 [e2e-llm-inference-service] + grpcio-testing==1.78.1 [e2e-llm-inference-service] + grpcio-tools==1.78.1 [e2e-llm-inference-service] + h11==0.16.0 [e2e-llm-inference-service] + httpcore==1.0.9 [e2e-llm-inference-service] + httptools==0.6.4 [e2e-llm-inference-service] + httpx==0.27.2 [e2e-llm-inference-service] + httpx-retries==0.4.5 [e2e-llm-inference-service] + idna==3.10 [e2e-llm-inference-service] + iniconfig==2.1.0 [e2e-llm-inference-service] + jinja2==3.1.6 [e2e-llm-inference-service] + jmespath==1.0.1 [e2e-llm-inference-service] + kserve==0.19.0 (from file:///workspace/source/python/kserve) [e2e-llm-inference-service] + kubernetes==32.0.1 [e2e-llm-inference-service] + markupsafe==3.0.2 [e2e-llm-inference-service] + multidict==6.4.3 [e2e-llm-inference-service] + mypy==0.991 [e2e-llm-inference-service] + mypy-extensions==1.0.0 [e2e-llm-inference-service] + numpy==2.2.4 [e2e-llm-inference-service] + oauthlib==3.2.2 [e2e-llm-inference-service] + orjson==3.10.16 [e2e-llm-inference-service] + packaging==24.2 [e2e-llm-inference-service] + pandas==2.2.3 [e2e-llm-inference-service] + pathspec==0.12.1 [e2e-llm-inference-service] + platformdirs==4.3.7 [e2e-llm-inference-service] + pluggy==1.5.0 [e2e-llm-inference-service] + portforward==0.7.1 [e2e-llm-inference-service] + prometheus-client==0.21.1 [e2e-llm-inference-service] + propcache==0.3.1 [e2e-llm-inference-service] + protobuf==6.33.5 [e2e-llm-inference-service] + psutil==5.9.8 [e2e-llm-inference-service] + pyarrow==19.0.1 [e2e-llm-inference-service] + pyasn1==0.6.3 [e2e-llm-inference-service] + pyasn1-modules==0.4.2 [e2e-llm-inference-service] + pycparser==2.22 [e2e-llm-inference-service] + pydantic==2.12.4 [e2e-llm-inference-service] + pydantic-core==2.41.5 [e2e-llm-inference-service] + pyjwt==2.12.1 [e2e-llm-inference-service] + pytest==7.4.4 [e2e-llm-inference-service] + pytest-asyncio==0.23.8 [e2e-llm-inference-service] + pytest-cov==5.0.0 [e2e-llm-inference-service] + pytest-httpx==0.30.0 [e2e-llm-inference-service] + pytest-json-report==1.5.0 [e2e-llm-inference-service] + pytest-metadata==3.1.1 [e2e-llm-inference-service] + pytest-xdist==3.6.1 [e2e-llm-inference-service] + python-dateutil==2.9.0.post0 [e2e-llm-inference-service] + python-dotenv==1.1.0 [e2e-llm-inference-service] + python-multipart==0.0.22 [e2e-llm-inference-service] + python-simple-logger==2.0.19 [e2e-llm-inference-service] + pytz==2025.2 [e2e-llm-inference-service] + pyyaml==6.0.2 [e2e-llm-inference-service] + requests==2.32.3 [e2e-llm-inference-service] + requests-oauthlib==2.0.0 [e2e-llm-inference-service] + rsa==4.9.1 [e2e-llm-inference-service] + s3transfer==0.11.4 [e2e-llm-inference-service] + setuptools==78.1.0 [e2e-llm-inference-service] + six==1.17.0 [e2e-llm-inference-service] + sniffio==1.3.1 [e2e-llm-inference-service] + starlette==1.2.1 [e2e-llm-inference-service] + tabulate==0.9.0 [e2e-llm-inference-service] + timeout-sampler==1.0.3 [e2e-llm-inference-service] + timing-asgi==0.3.1 [e2e-llm-inference-service] + tomlkit==0.13.2 [e2e-llm-inference-service] + typing-extensions==4.15.0 [e2e-llm-inference-service] + typing-inspection==0.4.2 [e2e-llm-inference-service] + tzdata==2025.2 [e2e-llm-inference-service] + urllib3==2.6.2 [e2e-llm-inference-service] + uvicorn==0.34.1 [e2e-llm-inference-service] + uvloop==0.21.0 [e2e-llm-inference-service] + watchfiles==1.0.5 [e2e-llm-inference-service] + websocket-client==1.8.0 [e2e-llm-inference-service] + websockets==15.0.1 [e2e-llm-inference-service] + yarl==1.20.0 [e2e-llm-inference-service] Audited 1 package in 53ms [e2e-llm-inference-service] /workspace/source [e2e-llm-inference-service] [INFO] Installing Kustomize v5.8.1 for linux/amd64... [e2e-llm-inference-service] [INFO] Kustomize v5.8.1 is already installed in /workspace/source/bin (>= v5.8.1) [e2e-llm-inference-service] make: Entering directory '/workspace/source' [e2e-llm-inference-service] make: Leaving directory '/workspace/source' [e2e-llm-inference-service] Now using project "kserve" on server "https://a0e16ce1ce6064cbb9315ffe26d645d0-c337f2f095db2de9.elb.us-east-1.amazonaws.com:6443". [e2e-llm-inference-service] [e2e-llm-inference-service] You can add applications to this project with the 'new-app' command. For example, try: [e2e-llm-inference-service] [e2e-llm-inference-service] oc new-app rails-postgresql-example [e2e-llm-inference-service] [e2e-llm-inference-service] to build a new example application in Ruby. Or use kubectl to deploy a simple Kubernetes application: [e2e-llm-inference-service] [e2e-llm-inference-service] kubectl create deployment hello-node --image=registry.k8s.io/e2e-test-images/agnhost:2.43 -- /agnhost serve-hostname [e2e-llm-inference-service] [e2e-llm-inference-service] [INFO] Installing Kustomize v5.8.1 for linux/amd64... [e2e-llm-inference-service] [INFO] Kustomize v5.8.1 is already installed in /workspace/source/bin (>= v5.8.1) [e2e-llm-inference-service] make: Entering directory '/workspace/source' [e2e-llm-inference-service] make: Leaving directory '/workspace/source' [e2e-llm-inference-service] Creating namespace openshift-keda... [e2e-llm-inference-service] namespace/openshift-keda created [e2e-llm-inference-service] Namespace openshift-keda created/ensured. [e2e-llm-inference-service] --- [e2e-llm-inference-service] Creating OperatorGroup openshift-keda... [e2e-llm-inference-service] operatorgroup.operators.coreos.com/openshift-keda created [e2e-llm-inference-service] OperatorGroup openshift-keda created/ensured. [e2e-llm-inference-service] --- [e2e-llm-inference-service] Creating Subscription for openshift-custom-metrics-autoscaler-operator... [e2e-llm-inference-service] subscription.operators.coreos.com/openshift-custom-metrics-autoscaler-operator created [e2e-llm-inference-service] Subscription openshift-custom-metrics-autoscaler-operator created/ensured. [e2e-llm-inference-service] --- [e2e-llm-inference-service] Waiting for openshift-custom-metrics-autoscaler-operator CSV to become ready... [e2e-llm-inference-service] Waiting for CSV to be installed for subscription openshift-custom-metrics-autoscaler-operator... (0/600) [e2e-llm-inference-service] Waiting for CSV to be installed for subscription openshift-custom-metrics-autoscaler-operator... (5/600) [e2e-llm-inference-service] Waiting for CSV to be installed for subscription openshift-custom-metrics-autoscaler-operator... (10/600) [e2e-llm-inference-service] Waiting for CSV to be installed for subscription openshift-custom-metrics-autoscaler-operator... (15/600) [e2e-llm-inference-service] Waiting for CSV to be installed for subscription openshift-custom-metrics-autoscaler-operator... (20/600) [e2e-llm-inference-service] Waiting for CSV to be installed for subscription openshift-custom-metrics-autoscaler-operator... (25/600) [e2e-llm-inference-service] CSV custom-metrics-autoscaler.v2.19.0-1 found, but not yet Succeeded (Phase: Installing). Waiting... (30/600) [e2e-llm-inference-service] CSV custom-metrics-autoscaler.v2.19.0-1 found, but not yet Succeeded (Phase: Installing). Waiting... (35/600) [e2e-llm-inference-service] CSV custom-metrics-autoscaler.v2.19.0-1 found, but not yet Succeeded (Phase: Installing). Waiting... (40/600) [e2e-llm-inference-service] CSV custom-metrics-autoscaler.v2.19.0-1 found, but not yet Succeeded (Phase: Installing). Waiting... (45/600) [e2e-llm-inference-service] CSV custom-metrics-autoscaler.v2.19.0-1 found, but not yet Succeeded (Phase: Installing). Waiting... (50/600) [e2e-llm-inference-service] CSV custom-metrics-autoscaler.v2.19.0-1 is ready (Phase: Succeeded). [e2e-llm-inference-service] --- [e2e-llm-inference-service] Applying KedaController custom resource... [e2e-llm-inference-service] Warning: resource kedacontrollers/keda is missing the kubectl.kubernetes.io/last-applied-configuration annotation which is required by oc apply. oc apply should only be used on resources created declaratively by either oc create --save-config or oc apply. The missing annotation will be patched automatically. [e2e-llm-inference-service] kedacontroller.keda.sh/keda configured [e2e-llm-inference-service] KedaController custom resource applied. [e2e-llm-inference-service] --- [e2e-llm-inference-service] Allowing time for KEDA components to be provisioned by the operator ... [e2e-llm-inference-service] Waiting for KEDA Operator pod (selector: "app=keda-operator") to be ready in namespace openshift-keda... [e2e-llm-inference-service] Waiting for pod -l "app=keda-operator" in namespace "openshift-keda" to be created... [e2e-llm-inference-service] Pod -l "app=keda-operator" in namespace "openshift-keda" found. [e2e-llm-inference-service] Current pods for -l "app=keda-operator" in namespace "openshift-keda": [e2e-llm-inference-service] NAME READY STATUS RESTARTS AGE [e2e-llm-inference-service] keda-operator-68bff59c-pnbm5 1/1 Running 0 42s [e2e-llm-inference-service] Waiting up to 120s for pod(s) -l "app=keda-operator" in namespace "openshift-keda" to become ready... [e2e-llm-inference-service] pod/keda-operator-68bff59c-pnbm5 condition met [e2e-llm-inference-service] Pod(s) -l "app=keda-operator" in namespace "openshift-keda" are ready. [e2e-llm-inference-service] KEDA Operator pod is ready. [e2e-llm-inference-service] Waiting for KEDA Metrics API Server pod (selector: "app=keda-metrics-apiserver") to be ready in namespace openshift-keda... [e2e-llm-inference-service] Waiting for pod -l "app=keda-metrics-apiserver" in namespace "openshift-keda" to be created... [e2e-llm-inference-service] Pod -l "app=keda-metrics-apiserver" in namespace "openshift-keda" found. [e2e-llm-inference-service] Current pods for -l "app=keda-metrics-apiserver" in namespace "openshift-keda": [e2e-llm-inference-service] NAME READY STATUS RESTARTS AGE [e2e-llm-inference-service] keda-metrics-apiserver-559f75f947-rm46z 1/1 Running 0 48s [e2e-llm-inference-service] Waiting up to 120s for pod(s) -l "app=keda-metrics-apiserver" in namespace "openshift-keda" to become ready... [e2e-llm-inference-service] pod/keda-metrics-apiserver-559f75f947-rm46z condition met [e2e-llm-inference-service] Pod(s) -l "app=keda-metrics-apiserver" in namespace "openshift-keda" are ready. [e2e-llm-inference-service] KEDA Metrics API Server pod is ready. [e2e-llm-inference-service] Waiting for KEDA Webhook pod (selector: "app=keda-admission-webhooks") to be ready in namespace openshift-keda... [e2e-llm-inference-service] Waiting for pod -l "app=keda-admission-webhooks" in namespace "openshift-keda" to be created... [e2e-llm-inference-service] Pod -l "app=keda-admission-webhooks" in namespace "openshift-keda" found. [e2e-llm-inference-service] Current pods for -l "app=keda-admission-webhooks" in namespace "openshift-keda": [e2e-llm-inference-service] NAME READY STATUS RESTARTS AGE [e2e-llm-inference-service] keda-admission-5fbd5c4644-kqr2b 1/1 Running 0 52s [e2e-llm-inference-service] Waiting up to 120s for pod(s) -l "app=keda-admission-webhooks" in namespace "openshift-keda" to become ready... [e2e-llm-inference-service] pod/keda-admission-5fbd5c4644-kqr2b condition met [e2e-llm-inference-service] Pod(s) -l "app=keda-admission-webhooks" in namespace "openshift-keda" are ready. [e2e-llm-inference-service] KEDA Webhook pod is ready. [e2e-llm-inference-service] --- [e2e-llm-inference-service] ✅ KEDA deployment script finished successfully. [e2e-llm-inference-service] KSERVE_CONTROLLER_IMAGE=quay.io/opendatahub/kserve-controller@sha256:d717f59e7fc6132421a696b06ade56c2521428323575a477d0441e0491bdb66c [e2e-llm-inference-service] LLMISVC_CONTROLLER_IMAGE=quay.io/opendatahub/odh-kserve-llmisvc-controller@sha256:611712697e2f5e409e4ec5b04fe487b6af8ff2eadd54cbb3a7ef8d15d81f4ae5 [e2e-llm-inference-service] KSERVE_AGENT_IMAGE=quay.io/opendatahub/kserve-agent@sha256:0b6de7012f6344e0be5c95ac8deb96d56b96edd624ee25e9ad8ecbad40286f86 [e2e-llm-inference-service] KSERVE_ROUTER_IMAGE=quay.io/opendatahub/kserve-router@sha256:3bfaa94d16d20ed0bc6642496b738f1c154c88684ea41dc8df471bf4e0a2804c [e2e-llm-inference-service] STORAGE_INITIALIZER_IMAGE=quay.io/opendatahub/kserve-storage-initializer@sha256:13ad765c0f323200e1db7b45c99661287e65c5158d6b54b16139829636d20be4 [e2e-llm-inference-service] Installing KServe via kustomize... [e2e-llm-inference-service] # Warning: 'commonLabels' is deprecated. Please use 'labels' instead. Run 'kustomize edit fix' to update your Kustomization automatically. [e2e-llm-inference-service] # Warning: 'commonLabels' is deprecated. Please use 'labels' instead. Run 'kustomize edit fix' to update your Kustomization automatically. [e2e-llm-inference-service] # Warning: 'commonLabels' is deprecated. Please use 'labels' instead. Run 'kustomize edit fix' to update your Kustomization automatically. [e2e-llm-inference-service] # Warning: 'commonLabels' is deprecated. Please use 'labels' instead. Run 'kustomize edit fix' to update your Kustomization automatically. [e2e-llm-inference-service] # Warning: 'commonLabels' is deprecated. Please use 'labels' instead. Run 'kustomize edit fix' to update your Kustomization automatically. [e2e-llm-inference-service] customresourcedefinition.apiextensions.k8s.io/clusterstoragecontainers.serving.kserve.io serverside-applied [e2e-llm-inference-service] customresourcedefinition.apiextensions.k8s.io/datascienceclusters.datasciencecluster.opendatahub.io serverside-applied [e2e-llm-inference-service] customresourcedefinition.apiextensions.k8s.io/dscinitializations.dscinitialization.opendatahub.io serverside-applied [e2e-llm-inference-service] customresourcedefinition.apiextensions.k8s.io/inferencegraphs.serving.kserve.io serverside-applied [e2e-llm-inference-service] customresourcedefinition.apiextensions.k8s.io/inferencemodelrewrites.inference.networking.x-k8s.io serverside-applied [e2e-llm-inference-service] customresourcedefinition.apiextensions.k8s.io/inferenceobjectives.inference.networking.x-k8s.io serverside-applied [e2e-llm-inference-service] customresourcedefinition.apiextensions.k8s.io/inferencepoolimports.inference.networking.x-k8s.io serverside-applied [e2e-llm-inference-service] customresourcedefinition.apiextensions.k8s.io/inferencepools.inference.networking.k8s.io serverside-applied [e2e-llm-inference-service] customresourcedefinition.apiextensions.k8s.io/inferencepools.inference.networking.x-k8s.io serverside-applied [e2e-llm-inference-service] customresourcedefinition.apiextensions.k8s.io/inferenceservices.serving.kserve.io serverside-applied [e2e-llm-inference-service] customresourcedefinition.apiextensions.k8s.io/llminferenceserviceconfigs.serving.kserve.io serverside-applied [e2e-llm-inference-service] customresourcedefinition.apiextensions.k8s.io/llminferenceservices.serving.kserve.io serverside-applied [e2e-llm-inference-service] customresourcedefinition.apiextensions.k8s.io/servingruntimes.serving.kserve.io serverside-applied [e2e-llm-inference-service] customresourcedefinition.apiextensions.k8s.io/trainedmodels.serving.kserve.io serverside-applied [e2e-llm-inference-service] Waiting for CRDs to be established... [e2e-llm-inference-service] Waiting for CRD inferenceservices.serving.kserve.io to appear (timeout: 90s)… [e2e-llm-inference-service] CRD inferenceservices.serving.kserve.io detected — waiting for it to become Established (timeout: 90s)… [e2e-llm-inference-service] customresourcedefinition.apiextensions.k8s.io/inferenceservices.serving.kserve.io condition met [e2e-llm-inference-service] Waiting for CRD llminferenceserviceconfigs.serving.kserve.io to appear (timeout: 90s)… [e2e-llm-inference-service] CRD llminferenceserviceconfigs.serving.kserve.io detected — waiting for it to become Established (timeout: 90s)… [e2e-llm-inference-service] customresourcedefinition.apiextensions.k8s.io/llminferenceserviceconfigs.serving.kserve.io condition met [e2e-llm-inference-service] Waiting for CRD clusterstoragecontainers.serving.kserve.io to appear (timeout: 90s)… [e2e-llm-inference-service] CRD clusterstoragecontainers.serving.kserve.io detected — waiting for it to become Established (timeout: 90s)… [e2e-llm-inference-service] customresourcedefinition.apiextensions.k8s.io/clusterstoragecontainers.serving.kserve.io condition met [e2e-llm-inference-service] Waiting for CRD datascienceclusters.datasciencecluster.opendatahub.io to appear (timeout: 90s)… [e2e-llm-inference-service] CRD datascienceclusters.datasciencecluster.opendatahub.io detected — waiting for it to become Established (timeout: 90s)… [e2e-llm-inference-service] customresourcedefinition.apiextensions.k8s.io/datascienceclusters.datasciencecluster.opendatahub.io condition met [e2e-llm-inference-service] Applying all resources... [e2e-llm-inference-service] customresourcedefinition.apiextensions.k8s.io/clusterstoragecontainers.serving.kserve.io serverside-applied [e2e-llm-inference-service] customresourcedefinition.apiextensions.k8s.io/datascienceclusters.datasciencecluster.opendatahub.io serverside-applied [e2e-llm-inference-service] customresourcedefinition.apiextensions.k8s.io/dscinitializations.dscinitialization.opendatahub.io serverside-applied [e2e-llm-inference-service] customresourcedefinition.apiextensions.k8s.io/inferencegraphs.serving.kserve.io serverside-applied [e2e-llm-inference-service] customresourcedefinition.apiextensions.k8s.io/inferencemodelrewrites.inference.networking.x-k8s.io serverside-applied [e2e-llm-inference-service] customresourcedefinition.apiextensions.k8s.io/inferenceobjectives.inference.networking.x-k8s.io serverside-applied [e2e-llm-inference-service] customresourcedefinition.apiextensions.k8s.io/inferencepoolimports.inference.networking.x-k8s.io serverside-applied [e2e-llm-inference-service] customresourcedefinition.apiextensions.k8s.io/inferencepools.inference.networking.k8s.io serverside-applied [e2e-llm-inference-service] customresourcedefinition.apiextensions.k8s.io/inferencepools.inference.networking.x-k8s.io serverside-applied [e2e-llm-inference-service] customresourcedefinition.apiextensions.k8s.io/inferenceservices.serving.kserve.io serverside-applied [e2e-llm-inference-service] customresourcedefinition.apiextensions.k8s.io/llminferenceserviceconfigs.serving.kserve.io serverside-applied [e2e-llm-inference-service] customresourcedefinition.apiextensions.k8s.io/llminferenceservices.serving.kserve.io serverside-applied [e2e-llm-inference-service] customresourcedefinition.apiextensions.k8s.io/servingruntimes.serving.kserve.io serverside-applied [e2e-llm-inference-service] customresourcedefinition.apiextensions.k8s.io/trainedmodels.serving.kserve.io serverside-applied [e2e-llm-inference-service] serviceaccount/kserve-controller-manager serverside-applied [e2e-llm-inference-service] serviceaccount/llmisvc-controller-manager serverside-applied [e2e-llm-inference-service] role.rbac.authorization.k8s.io/kserve-leader-election-role serverside-applied [e2e-llm-inference-service] role.rbac.authorization.k8s.io/kserve-llmisvcconfig-read-access serverside-applied [e2e-llm-inference-service] role.rbac.authorization.k8s.io/llmisvc-leader-election-role serverside-applied [e2e-llm-inference-service] clusterrole.rbac.authorization.k8s.io/kserve-admin serverside-applied [e2e-llm-inference-service] clusterrole.rbac.authorization.k8s.io/kserve-edit serverside-applied [e2e-llm-inference-service] clusterrole.rbac.authorization.k8s.io/kserve-inferenceservice-distro-role serverside-applied [e2e-llm-inference-service] clusterrole.rbac.authorization.k8s.io/kserve-llmisvc-distro-role serverside-applied [e2e-llm-inference-service] clusterrole.rbac.authorization.k8s.io/kserve-llmisvc-manager-role serverside-applied [e2e-llm-inference-service] clusterrole.rbac.authorization.k8s.io/kserve-manager-role serverside-applied [e2e-llm-inference-service] clusterrole.rbac.authorization.k8s.io/kserve-metrics-reader-cluster-role serverside-applied [e2e-llm-inference-service] clusterrole.rbac.authorization.k8s.io/kserve-proxy-role serverside-applied [e2e-llm-inference-service] clusterrole.rbac.authorization.k8s.io/kserve-view serverside-applied [e2e-llm-inference-service] clusterrole.rbac.authorization.k8s.io/openshift-ai-inferenceservice-image-volume-scc serverside-applied [e2e-llm-inference-service] clusterrole.rbac.authorization.k8s.io/openshift-ai-llminferenceservice-scc serverside-applied [e2e-llm-inference-service] rolebinding.rbac.authorization.k8s.io/kserve-leader-election-rolebinding serverside-applied [e2e-llm-inference-service] rolebinding.rbac.authorization.k8s.io/kserve-llmisvcconfig-read-access serverside-applied [e2e-llm-inference-service] rolebinding.rbac.authorization.k8s.io/llmisvc-leader-election-rolebinding serverside-applied [e2e-llm-inference-service] clusterrolebinding.rbac.authorization.k8s.io/kserve-inferenceservice-distro-rolebinding serverside-applied [e2e-llm-inference-service] clusterrolebinding.rbac.authorization.k8s.io/kserve-llmisvc-distro-rolebinding serverside-applied [e2e-llm-inference-service] clusterrolebinding.rbac.authorization.k8s.io/kserve-manager-rolebinding serverside-applied [e2e-llm-inference-service] clusterrolebinding.rbac.authorization.k8s.io/kserve-proxy-rolebinding serverside-applied [e2e-llm-inference-service] clusterrolebinding.rbac.authorization.k8s.io/llmisvc-manager-rolebinding serverside-applied [e2e-llm-inference-service] configmap/inferenceservice-config serverside-applied [e2e-llm-inference-service] configmap/kserve-parameters serverside-applied [e2e-llm-inference-service] secret/kserve-webhook-server-secret serverside-applied [e2e-llm-inference-service] secret/mlpipeline-s3-artifact serverside-applied [e2e-llm-inference-service] service/kserve-controller-manager-metrics-service serverside-applied [e2e-llm-inference-service] service/kserve-controller-manager-service serverside-applied [e2e-llm-inference-service] service/kserve-webhook-server-service serverside-applied [e2e-llm-inference-service] service/llmisvc-controller-manager-service serverside-applied [e2e-llm-inference-service] service/llmisvc-webhook-server-service serverside-applied [e2e-llm-inference-service] service/s3-service serverside-applied [e2e-llm-inference-service] deployment.apps/kserve-controller-manager serverside-applied [e2e-llm-inference-service] deployment.apps/llmisvc-controller-manager serverside-applied [e2e-llm-inference-service] deployment.apps/seaweedfs serverside-applied [e2e-llm-inference-service] networkpolicy.networking.k8s.io/kserve-controller-manager serverside-applied [e2e-llm-inference-service] securitycontextconstraints.security.openshift.io/openshift-ai-inferenceservice-image-volume-scc serverside-applied [e2e-llm-inference-service] securitycontextconstraints.security.openshift.io/openshift-ai-llminferenceservice-scc serverside-applied [e2e-llm-inference-service] clusterstoragecontainer.serving.kserve.io/default serverside-applied [e2e-llm-inference-service] llminferenceserviceconfig.serving.kserve.io/kserve-config-llm-decode-template serverside-applied [e2e-llm-inference-service] llminferenceserviceconfig.serving.kserve.io/kserve-config-llm-decode-worker-data-parallel serverside-applied [e2e-llm-inference-service] llminferenceserviceconfig.serving.kserve.io/kserve-config-llm-prefill-template serverside-applied [e2e-llm-inference-service] llminferenceserviceconfig.serving.kserve.io/kserve-config-llm-prefill-template-nvidia-cuda serverside-applied [e2e-llm-inference-service] llminferenceserviceconfig.serving.kserve.io/kserve-config-llm-prefill-template-nvidia-cuda-fast-1 serverside-applied [e2e-llm-inference-service] llminferenceserviceconfig.serving.kserve.io/kserve-config-llm-prefill-template-nvidia-cuda-fast-2 serverside-applied [e2e-llm-inference-service] llminferenceserviceconfig.serving.kserve.io/kserve-config-llm-prefill-worker-data-parallel serverside-applied [e2e-llm-inference-service] llminferenceserviceconfig.serving.kserve.io/kserve-config-llm-router-route serverside-applied [e2e-llm-inference-service] llminferenceserviceconfig.serving.kserve.io/kserve-config-llm-scheduler serverside-applied [e2e-llm-inference-service] llminferenceserviceconfig.serving.kserve.io/kserve-config-llm-template serverside-applied [e2e-llm-inference-service] llminferenceserviceconfig.serving.kserve.io/kserve-config-llm-template-amd-rocm serverside-applied [e2e-llm-inference-service] llminferenceserviceconfig.serving.kserve.io/kserve-config-llm-template-amd-rocm-fast-1 serverside-applied [e2e-llm-inference-service] llminferenceserviceconfig.serving.kserve.io/kserve-config-llm-template-amd-rocm-fast-2 serverside-applied [e2e-llm-inference-service] llminferenceserviceconfig.serving.kserve.io/kserve-config-llm-template-ibm-spyre-ppc64le serverside-applied [e2e-llm-inference-service] llminferenceserviceconfig.serving.kserve.io/kserve-config-llm-template-ibm-spyre-ppc64le-fast-1 serverside-applied [e2e-llm-inference-service] llminferenceserviceconfig.serving.kserve.io/kserve-config-llm-template-ibm-spyre-ppc64le-fast-2 serverside-applied [e2e-llm-inference-service] llminferenceserviceconfig.serving.kserve.io/kserve-config-llm-template-ibm-spyre-s390x serverside-applied [e2e-llm-inference-service] llminferenceserviceconfig.serving.kserve.io/kserve-config-llm-template-ibm-spyre-s390x-fast-1 serverside-applied [e2e-llm-inference-service] llminferenceserviceconfig.serving.kserve.io/kserve-config-llm-template-ibm-spyre-s390x-fast-2 serverside-applied [e2e-llm-inference-service] llminferenceserviceconfig.serving.kserve.io/kserve-config-llm-template-ibm-spyre-x86 serverside-applied [e2e-llm-inference-service] llminferenceserviceconfig.serving.kserve.io/kserve-config-llm-template-ibm-spyre-x86-fast-1 serverside-applied [e2e-llm-inference-service] llminferenceserviceconfig.serving.kserve.io/kserve-config-llm-template-ibm-spyre-x86-fast-2 serverside-applied [e2e-llm-inference-service] llminferenceserviceconfig.serving.kserve.io/kserve-config-llm-template-intel-gaudi serverside-applied [e2e-llm-inference-service] llminferenceserviceconfig.serving.kserve.io/kserve-config-llm-template-intel-gaudi-fast-1 serverside-applied [e2e-llm-inference-service] llminferenceserviceconfig.serving.kserve.io/kserve-config-llm-template-intel-gaudi-fast-2 serverside-applied [e2e-llm-inference-service] llminferenceserviceconfig.serving.kserve.io/kserve-config-llm-template-nvidia-cuda serverside-applied [e2e-llm-inference-service] llminferenceserviceconfig.serving.kserve.io/kserve-config-llm-template-nvidia-cuda-fast-1 serverside-applied [e2e-llm-inference-service] llminferenceserviceconfig.serving.kserve.io/kserve-config-llm-template-nvidia-cuda-fast-2 serverside-applied [e2e-llm-inference-service] llminferenceserviceconfig.serving.kserve.io/kserve-config-llm-tracing serverside-applied [e2e-llm-inference-service] llminferenceserviceconfig.serving.kserve.io/kserve-config-llm-worker-data-parallel serverside-applied [e2e-llm-inference-service] mutatingwebhookconfiguration.admissionregistration.k8s.io/inferenceservice.serving.kserve.io serverside-applied [e2e-llm-inference-service] mutatingwebhookconfiguration.admissionregistration.k8s.io/llminferenceservice.serving.kserve.io serverside-applied [e2e-llm-inference-service] validatingwebhookconfiguration.admissionregistration.k8s.io/inferencegraph.serving.kserve.io serverside-applied [e2e-llm-inference-service] validatingwebhookconfiguration.admissionregistration.k8s.io/inferenceservice.serving.kserve.io serverside-applied [e2e-llm-inference-service] validatingwebhookconfiguration.admissionregistration.k8s.io/llminferenceservice.serving.kserve.io serverside-applied [e2e-llm-inference-service] validatingwebhookconfiguration.admissionregistration.k8s.io/llminferenceserviceconfig.serving.kserve.io serverside-applied [e2e-llm-inference-service] validatingwebhookconfiguration.admissionregistration.k8s.io/servingruntime.serving.kserve.io serverside-applied [e2e-llm-inference-service] validatingwebhookconfiguration.admissionregistration.k8s.io/trainedmodel.serving.kserve.io serverside-applied [e2e-llm-inference-service] Waiting for llmisvc-controller-manager to be ready... [e2e-llm-inference-service] Waiting for pod -l "control-plane=llmisvc-controller-manager" in namespace "kserve" to be created... [e2e-llm-inference-service] Pod -l "control-plane=llmisvc-controller-manager" in namespace "kserve" found. [e2e-llm-inference-service] Current pods for -l "control-plane=llmisvc-controller-manager" in namespace "kserve": [e2e-llm-inference-service] NAME READY STATUS RESTARTS AGE [e2e-llm-inference-service] llmisvc-controller-manager-756677495b-n7q2l 0/1 Running 0 7s [e2e-llm-inference-service] Waiting up to 600s for pod(s) -l "control-plane=llmisvc-controller-manager" in namespace "kserve" to become ready... [e2e-llm-inference-service] pod/llmisvc-controller-manager-756677495b-n7q2l condition met [e2e-llm-inference-service] Pod(s) -l "control-plane=llmisvc-controller-manager" in namespace "kserve" are ready. [e2e-llm-inference-service] Re-applying LLMInferenceServiceConfig resources with webhook validation... [e2e-llm-inference-service] Warning: modifying well-known config kserve/kserve-config-llm-decode-template is not recommended. Consider creating a custom config instead [e2e-llm-inference-service] llminferenceserviceconfig.serving.kserve.io/kserve-config-llm-decode-template serverside-applied [e2e-llm-inference-service] Warning: modifying well-known config kserve/kserve-config-llm-decode-worker-data-parallel is not recommended. Consider creating a custom config instead [e2e-llm-inference-service] llminferenceserviceconfig.serving.kserve.io/kserve-config-llm-decode-worker-data-parallel serverside-applied [e2e-llm-inference-service] Warning: modifying well-known config kserve/kserve-config-llm-prefill-template is not recommended. Consider creating a custom config instead [e2e-llm-inference-service] llminferenceserviceconfig.serving.kserve.io/kserve-config-llm-prefill-template serverside-applied [e2e-llm-inference-service] Warning: modifying well-known config kserve/kserve-config-llm-prefill-worker-data-parallel is not recommended. Consider creating a custom config instead [e2e-llm-inference-service] llminferenceserviceconfig.serving.kserve.io/kserve-config-llm-prefill-worker-data-parallel serverside-applied [e2e-llm-inference-service] llminferenceserviceconfig.serving.kserve.io/kserve-config-llm-router-route serverside-applied [e2e-llm-inference-service] Warning: modifying well-known config kserve/kserve-config-llm-scheduler is not recommended. Consider creating a custom config instead [e2e-llm-inference-service] llminferenceserviceconfig.serving.kserve.io/kserve-config-llm-scheduler serverside-applied [e2e-llm-inference-service] Warning: modifying well-known config kserve/kserve-config-llm-template is not recommended. Consider creating a custom config instead [e2e-llm-inference-service] llminferenceserviceconfig.serving.kserve.io/kserve-config-llm-template serverside-applied [e2e-llm-inference-service] llminferenceserviceconfig.serving.kserve.io/kserve-config-llm-tracing serverside-applied [e2e-llm-inference-service] Warning: modifying well-known config kserve/kserve-config-llm-worker-data-parallel is not recommended. Consider creating a custom config instead [e2e-llm-inference-service] llminferenceserviceconfig.serving.kserve.io/kserve-config-llm-worker-data-parallel serverside-applied [e2e-llm-inference-service] Applying DSC/DSCI resources... [e2e-llm-inference-service] dscinitialization.dscinitialization.opendatahub.io/test-dsci created [e2e-llm-inference-service] datasciencecluster.datasciencecluster.opendatahub.io/test-dsc created [e2e-llm-inference-service] KServe manual installation complete [e2e-llm-inference-service] 🔧 Configuration: [e2e-llm-inference-service] KServe deployment: ❌ disabled [e2e-llm-inference-service] Kuadrant deployment: ✅ enabled [e2e-llm-inference-service] [e2e-llm-inference-service] Checking OpenShift server version...(4.21.23) [e2e-llm-inference-service] 🎯 Server version (4.21.23) is 4.19.9 or higher - continue with the script [e2e-llm-inference-service] ⏳ Installing cert-manager [e2e-llm-inference-service] namespace/cert-manager-operator created [e2e-llm-inference-service] operatorgroup.operators.coreos.com/openshift-cert-manager-operator created [e2e-llm-inference-service] subscription.operators.coreos.com/openshift-cert-manager-operator created [e2e-llm-inference-service] Waiting for openshift-cert-manager-operator CSV to become ready... [e2e-llm-inference-service] Waiting for CSV to be installed for subscription openshift-cert-manager-operator... (0/300) [e2e-llm-inference-service] Waiting for CSV to be installed for subscription openshift-cert-manager-operator... (5/300) [e2e-llm-inference-service] Waiting for CSV to be installed for subscription openshift-cert-manager-operator... (10/300) [e2e-llm-inference-service] CSV cert-manager-operator.v1.20.0 found, but not yet Succeeded (Phase: Installing). Waiting... (15/300) [e2e-llm-inference-service] CSV cert-manager-operator.v1.20.0 is ready (Phase: Succeeded). [e2e-llm-inference-service] Waiting for CRD certificates.cert-manager.io to appear (timeout: 90s)… [e2e-llm-inference-service] CRD certificates.cert-manager.io detected — waiting for it to become Established (timeout: 90s)… [e2e-llm-inference-service] customresourcedefinition.apiextensions.k8s.io/certificates.cert-manager.io condition met [e2e-llm-inference-service] ✅ Cert-manager installed [e2e-llm-inference-service] ⏳ Installing openshift-lws-operator [e2e-llm-inference-service] namespace/openshift-lws-operator created [e2e-llm-inference-service] operatorgroup.operators.coreos.com/leader-worker-set created [e2e-llm-inference-service] subscription.operators.coreos.com/leader-worker-set created [e2e-llm-inference-service] Waiting for leader-worker-set CSV to become ready... [e2e-llm-inference-service] Waiting for CSV to be installed for subscription leader-worker-set... (0/300) [e2e-llm-inference-service] Waiting for CSV to be installed for subscription leader-worker-set... (5/300) [e2e-llm-inference-service] Waiting for CSV to be installed for subscription leader-worker-set... (10/300) [e2e-llm-inference-service] CSV leader-worker-set.v1.0.0 found, but not yet Succeeded (Phase: Installing). Waiting... (15/300) [e2e-llm-inference-service] CSV leader-worker-set.v1.0.0 is ready (Phase: Succeeded). [e2e-llm-inference-service] Waiting for CRD leaderworkersetoperators.operator.openshift.io to appear (timeout: 90s)… [e2e-llm-inference-service] CRD leaderworkersetoperators.operator.openshift.io detected — waiting for it to become Established (timeout: 90s)… [e2e-llm-inference-service] customresourcedefinition.apiextensions.k8s.io/leaderworkersetoperators.operator.openshift.io condition met [e2e-llm-inference-service] leaderworkersetoperator.operator.openshift.io/cluster created [e2e-llm-inference-service] ⏳ waiting for openshift-lws-operator to be ready.… [e2e-llm-inference-service] Waiting for pod -l "name=openshift-lws-operator" in namespace "openshift-lws-operator" to be created... [e2e-llm-inference-service] Pod -l "name=openshift-lws-operator" in namespace "openshift-lws-operator" found. [e2e-llm-inference-service] Current pods for -l "name=openshift-lws-operator" in namespace "openshift-lws-operator": [e2e-llm-inference-service] NAME READY STATUS RESTARTS AGE [e2e-llm-inference-service] openshift-lws-operator-fd8ccff4c-x4j62 1/1 Running 0 13s [e2e-llm-inference-service] Waiting up to 600s for pod(s) -l "name=openshift-lws-operator" in namespace "openshift-lws-operator" to become ready... [e2e-llm-inference-service] pod/openshift-lws-operator-fd8ccff4c-x4j62 condition met [e2e-llm-inference-service] Pod(s) -l "name=openshift-lws-operator" in namespace "openshift-lws-operator" are ready. [e2e-llm-inference-service] ✅ openshift-lws-operator installed [e2e-llm-inference-service] gatewayclass.gateway.networking.k8s.io/openshift-default created [e2e-llm-inference-service] Waiting for pod -l "app=istiod" in namespace "openshift-ingress" to be created... [e2e-llm-inference-service] Pod -l "app=istiod" in namespace "openshift-ingress" found. [e2e-llm-inference-service] Current pods for -l "app=istiod" in namespace "openshift-ingress": [e2e-llm-inference-service] NAME READY STATUS RESTARTS AGE [e2e-llm-inference-service] istiod-openshift-gateway-94bb8fbfd-rffhz 1/1 Running 0 5s [e2e-llm-inference-service] Waiting up to 600s for pod(s) -l "app=istiod" in namespace "openshift-ingress" to become ready... [e2e-llm-inference-service] pod/istiod-openshift-gateway-94bb8fbfd-rffhz condition met [e2e-llm-inference-service] Pod(s) -l "app=istiod" in namespace "openshift-ingress" are ready. [e2e-llm-inference-service] ⏳ Creating a Gateway [e2e-llm-inference-service] Error from server (AlreadyExists): namespaces "openshift-ingress" already exists [e2e-llm-inference-service] gateway.gateway.networking.k8s.io/openshift-ai-inference created [e2e-llm-inference-service] Waiting for pod -l "serving.kserve.io/gateway=kserve-ingress-gateway" in namespace "openshift-ingress" to be created... [e2e-llm-inference-service] Pod -l "serving.kserve.io/gateway=kserve-ingress-gateway" in namespace "openshift-ingress" found. [e2e-llm-inference-service] Current pods for -l "serving.kserve.io/gateway=kserve-ingress-gateway" in namespace "openshift-ingress": [e2e-llm-inference-service] NAME READY STATUS RESTARTS AGE [e2e-llm-inference-service] openshift-ai-inference-openshift-default-9455cf8d5-6wqcr 1/1 Running 0 5s [e2e-llm-inference-service] Waiting up to 600s for pod(s) -l "serving.kserve.io/gateway=kserve-ingress-gateway" in namespace "openshift-ingress" to become ready... [e2e-llm-inference-service] pod/openshift-ai-inference-openshift-default-9455cf8d5-6wqcr condition met [e2e-llm-inference-service] Pod(s) -l "serving.kserve.io/gateway=kserve-ingress-gateway" in namespace "openshift-ingress" are ready. [e2e-llm-inference-service] ⏳ Installing RHCL(Kuadrant) operator [e2e-llm-inference-service] namespace/kuadrant-system created [e2e-llm-inference-service] subscription.operators.coreos.com/rhcl-operator created [e2e-llm-inference-service] operatorgroup.operators.coreos.com/kuadrant created [e2e-llm-inference-service] Waiting for rhcl-operator CSV to become ready... [e2e-llm-inference-service] Waiting for CSV to be installed for subscription rhcl-operator... (0/600) [e2e-llm-inference-service] Waiting for CSV to be installed for subscription rhcl-operator... (5/600) [e2e-llm-inference-service] Waiting for CSV to be installed for subscription rhcl-operator... (10/600) [e2e-llm-inference-service] Waiting for CSV to be installed for subscription rhcl-operator... (15/600) [e2e-llm-inference-service] CSV rhcl-operator.v1.4.1 found, but not yet Succeeded (Phase: Installing). Waiting... (20/600) [e2e-llm-inference-service] CSV rhcl-operator.v1.4.1 found, but not yet Succeeded (Phase: Installing). Waiting... (25/600) [e2e-llm-inference-service] CSV rhcl-operator.v1.4.1 found, but not yet Succeeded (Phase: Installing). Waiting... (30/600) [e2e-llm-inference-service] CSV rhcl-operator.v1.4.1 is ready (Phase: Succeeded). [e2e-llm-inference-service] Waiting for CRD kuadrants.kuadrant.io to appear (timeout: 90s)… [e2e-llm-inference-service] CRD kuadrants.kuadrant.io detected — waiting for it to become Established (timeout: 90s)… [e2e-llm-inference-service] customresourcedefinition.apiextensions.k8s.io/kuadrants.kuadrant.io condition met [e2e-llm-inference-service] Waiting for apiserver discovery /apis/kuadrant.io/v1beta1 to list kuadrants (timeout: 120s)… [e2e-llm-inference-service] Discovery for kuadrant.io/v1beta1 includes kuadrants. [e2e-llm-inference-service] ⏳ sleeping 30s after discovery (RESTMapper can trail discovery)… [e2e-llm-inference-service] kuadrant.kuadrant.io/kuadrant created [e2e-llm-inference-service] ⏳ waiting for Kuadrant Ready (attempt 1/2, timeout 5m)… [e2e-llm-inference-service] kuadrant.kuadrant.io/kuadrant condition met [e2e-llm-inference-service] Waiting for pod -l "control-plane=authorino-operator" in namespace "kuadrant-system" to be created... [e2e-llm-inference-service] Pod -l "control-plane=authorino-operator" in namespace "kuadrant-system" found. [e2e-llm-inference-service] Current pods for -l "control-plane=authorino-operator" in namespace "kuadrant-system": [e2e-llm-inference-service] NAME READY STATUS RESTARTS AGE [e2e-llm-inference-service] authorino-operator-6d85f6564-dt8p7 1/1 Running 0 67s [e2e-llm-inference-service] Waiting up to 600s for pod(s) -l "control-plane=authorino-operator" in namespace "kuadrant-system" to become ready... [e2e-llm-inference-service] pod/authorino-operator-6d85f6564-dt8p7 condition met [e2e-llm-inference-service] Pod(s) -l "control-plane=authorino-operator" in namespace "kuadrant-system" are ready. [e2e-llm-inference-service] ⏳ waiting for authorino service to be created... [e2e-llm-inference-service] service/authorino-authorino-authorization condition met [e2e-llm-inference-service] service/authorino-authorino-authorization annotated [e2e-llm-inference-service] Warning: resource authorinos/authorino is missing the kubectl.kubernetes.io/last-applied-configuration annotation which is required by oc apply. oc apply should only be used on resources created declaratively by either oc create --save-config or oc apply. The missing annotation will be patched automatically. [e2e-llm-inference-service] authorino.operator.authorino.kuadrant.io/authorino configured [e2e-llm-inference-service] Waiting for pod -l "control-plane=authorino-operator" in namespace "kuadrant-system" to be created... [e2e-llm-inference-service] Pod -l "control-plane=authorino-operator" in namespace "kuadrant-system" found. [e2e-llm-inference-service] Current pods for -l "control-plane=authorino-operator" in namespace "kuadrant-system": [e2e-llm-inference-service] NAME READY STATUS RESTARTS AGE [e2e-llm-inference-service] authorino-operator-6d85f6564-dt8p7 1/1 Running 0 77s [e2e-llm-inference-service] Waiting up to 600s for pod(s) -l "control-plane=authorino-operator" in namespace "kuadrant-system" to become ready... [e2e-llm-inference-service] pod/authorino-operator-6d85f6564-dt8p7 condition met [e2e-llm-inference-service] Pod(s) -l "control-plane=authorino-operator" in namespace "kuadrant-system" are ready. [e2e-llm-inference-service] ✅ kuadrant(authorino) installed [e2e-llm-inference-service] Patching ingress domain... [e2e-llm-inference-service] configmap/inferenceservice-config patched [e2e-llm-inference-service] pod "kserve-controller-manager-7598d7bb4-tbnz8" deleted [e2e-llm-inference-service] Waiting for kserve-controller-manager to be ready... [e2e-llm-inference-service] pod/kserve-controller-manager-7598d7bb4-gzlcn condition met [e2e-llm-inference-service] Installing ODH Model Controller manually... [e2e-llm-inference-service] customresourcedefinition.apiextensions.k8s.io/accounts.nim.opendatahub.io created [e2e-llm-inference-service] serviceaccount/model-serving-api created [e2e-llm-inference-service] serviceaccount/odh-model-controller created [e2e-llm-inference-service] role.rbac.authorization.k8s.io/leader-election-role created [e2e-llm-inference-service] clusterrole.rbac.authorization.k8s.io/account-editor-role created [e2e-llm-inference-service] clusterrole.rbac.authorization.k8s.io/account-viewer-role created [e2e-llm-inference-service] clusterrole.rbac.authorization.k8s.io/kserve-prometheus-k8s created [e2e-llm-inference-service] clusterrole.rbac.authorization.k8s.io/metrics-reader created [e2e-llm-inference-service] clusterrole.rbac.authorization.k8s.io/model-serving-api created [e2e-llm-inference-service] clusterrole.rbac.authorization.k8s.io/odh-model-controller-role created [e2e-llm-inference-service] clusterrole.rbac.authorization.k8s.io/proxy-role created [e2e-llm-inference-service] rolebinding.rbac.authorization.k8s.io/leader-election-rolebinding created [e2e-llm-inference-service] clusterrolebinding.rbac.authorization.k8s.io/model-serving-api created [e2e-llm-inference-service] clusterrolebinding.rbac.authorization.k8s.io/odh-model-controller-rolebinding-opendatahub created [e2e-llm-inference-service] clusterrolebinding.rbac.authorization.k8s.io/proxy-rolebinding created [e2e-llm-inference-service] configmap/odh-model-controller-parameters created [e2e-llm-inference-service] service/model-serving-api created [e2e-llm-inference-service] service/odh-model-controller-metrics-service created [e2e-llm-inference-service] service/odh-model-controller-webhook-service created [e2e-llm-inference-service] deployment.apps/model-serving-api created [e2e-llm-inference-service] deployment.apps/odh-model-controller created [e2e-llm-inference-service] servicemonitor.monitoring.coreos.com/model-serving-api-metrics created [e2e-llm-inference-service] servicemonitor.monitoring.coreos.com/odh-model-controller-metrics-monitor created [e2e-llm-inference-service] template.template.openshift.io/autogluon-runtime-template created [e2e-llm-inference-service] template.template.openshift.io/guardrails-detector-huggingface-serving-template created [e2e-llm-inference-service] template.template.openshift.io/kserve-ovms created [e2e-llm-inference-service] template.template.openshift.io/mlserver-runtime-template created [e2e-llm-inference-service] template.template.openshift.io/vllm-cpu-runtime-template created [e2e-llm-inference-service] template.template.openshift.io/vllm-cpu-runtime-template-fast-1 created [e2e-llm-inference-service] template.template.openshift.io/vllm-cpu-runtime-template-fast-2 created [e2e-llm-inference-service] template.template.openshift.io/vllm-cpu-x86-runtime-template created [e2e-llm-inference-service] template.template.openshift.io/vllm-cpu-x86-runtime-template-fast-1 created [e2e-llm-inference-service] template.template.openshift.io/vllm-cpu-x86-runtime-template-fast-2 created [e2e-llm-inference-service] template.template.openshift.io/vllm-cuda-runtime-template created [e2e-llm-inference-service] template.template.openshift.io/vllm-cuda-runtime-template-fast-1 created [e2e-llm-inference-service] template.template.openshift.io/vllm-cuda-runtime-template-fast-2 created [e2e-llm-inference-service] template.template.openshift.io/vllm-gaudi-runtime-template created [e2e-llm-inference-service] template.template.openshift.io/vllm-gaudi-runtime-template-fast-1 created [e2e-llm-inference-service] template.template.openshift.io/vllm-gaudi-runtime-template-fast-2 created [e2e-llm-inference-service] template.template.openshift.io/vllm-multinode-runtime-template created [e2e-llm-inference-service] template.template.openshift.io/vllm-multinode-runtime-template-fast-1 created [e2e-llm-inference-service] template.template.openshift.io/vllm-multinode-runtime-template-fast-2 created [e2e-llm-inference-service] template.template.openshift.io/vllm-rocm-runtime-template created [e2e-llm-inference-service] template.template.openshift.io/vllm-rocm-runtime-template-fast-1 created [e2e-llm-inference-service] template.template.openshift.io/vllm-rocm-runtime-template-fast-2 created [e2e-llm-inference-service] template.template.openshift.io/vllm-spyre-ppc64le-runtime-template created [e2e-llm-inference-service] template.template.openshift.io/vllm-spyre-ppc64le-runtime-template-fast-1 created [e2e-llm-inference-service] template.template.openshift.io/vllm-spyre-ppc64le-runtime-template-fast-2 created [e2e-llm-inference-service] template.template.openshift.io/vllm-spyre-s390x-runtime-template created [e2e-llm-inference-service] template.template.openshift.io/vllm-spyre-s390x-runtime-template-fast-1 created [e2e-llm-inference-service] template.template.openshift.io/vllm-spyre-s390x-runtime-template-fast-2 created [e2e-llm-inference-service] template.template.openshift.io/vllm-spyre-x86-runtime-template created [e2e-llm-inference-service] template.template.openshift.io/vllm-spyre-x86-runtime-template-fast-1 created [e2e-llm-inference-service] template.template.openshift.io/vllm-spyre-x86-runtime-template-fast-2 created [e2e-llm-inference-service] mutatingwebhookconfiguration.admissionregistration.k8s.io/mutating.odh-model-controller.opendatahub.io created [e2e-llm-inference-service] validatingwebhookconfiguration.admissionregistration.k8s.io/validating.odh-model-controller.opendatahub.io created [e2e-llm-inference-service] Waiting for deployment "odh-model-controller" rollout to finish: 0 of 1 updated replicas are available... [e2e-llm-inference-service] deployment "odh-model-controller" successfully rolled out [e2e-llm-inference-service] networkpolicy.networking.k8s.io/allow-all created [e2e-llm-inference-service] KServe setup complete (namespace: kserve) [e2e-llm-inference-service] Add testing models to SeaweedFS S3 storage ... [e2e-llm-inference-service] Waiting for SeaweedFS deployment to be ready... [e2e-llm-inference-service] deployment "seaweedfs" successfully rolled out [e2e-llm-inference-service] S3 init job not completed, re-creating... [e2e-llm-inference-service] job.batch/s3-init replaced [e2e-llm-inference-service] Waiting for S3 init job to complete... [e2e-llm-inference-service] job.batch/s3-init condition met [e2e-llm-inference-service] Prepare CI namespace and install ServingRuntimes [e2e-llm-inference-service] Setting up CI namespace: kserve-ci-e2e-test [e2e-llm-inference-service] Tearing down CI namespace: kserve-ci-e2e-test [e2e-llm-inference-service] Namespace kserve-ci-e2e-test does not exist, skipping deletion [e2e-llm-inference-service] CI namespace teardown complete [e2e-llm-inference-service] Creating namespace kserve-ci-e2e-test [e2e-llm-inference-service] namespace/kserve-ci-e2e-test created [e2e-llm-inference-service] Applying S3 artifact secret [e2e-llm-inference-service] secret/mlpipeline-s3-artifact created [e2e-llm-inference-service] Applying storage-config secret [e2e-llm-inference-service] secret/storage-config created [e2e-llm-inference-service] Applying SeaweedFS S3 credentials secret [e2e-llm-inference-service] secret/seaweedfs-s3-creds created [e2e-llm-inference-service] Linking seaweedfs-s3-creds to default service account [e2e-llm-inference-service] Creating odh-trusted-ca-bundle configmap [e2e-llm-inference-service] configmap/odh-trusted-ca-bundle created [e2e-llm-inference-service] Installing ServingRuntimes [e2e-llm-inference-service] servingruntime.serving.kserve.io/kserve-autogluonserver created [e2e-llm-inference-service] servingruntime.serving.kserve.io/kserve-huggingfaceserver created [e2e-llm-inference-service] servingruntime.serving.kserve.io/kserve-huggingfaceserver-multinode created [e2e-llm-inference-service] servingruntime.serving.kserve.io/kserve-lgbserver created [e2e-llm-inference-service] servingruntime.serving.kserve.io/kserve-mlserver created [e2e-llm-inference-service] servingruntime.serving.kserve.io/kserve-paddleserver created [e2e-llm-inference-service] servingruntime.serving.kserve.io/kserve-pmmlserver created [e2e-llm-inference-service] servingruntime.serving.kserve.io/kserve-predictiveserver created [e2e-llm-inference-service] servingruntime.serving.kserve.io/kserve-sklearnserver created [e2e-llm-inference-service] servingruntime.serving.kserve.io/kserve-tensorflow-serving created [e2e-llm-inference-service] servingruntime.serving.kserve.io/kserve-torchserve created [e2e-llm-inference-service] servingruntime.serving.kserve.io/kserve-tritonserver created [e2e-llm-inference-service] servingruntime.serving.kserve.io/kserve-vllmserver created [e2e-llm-inference-service] servingruntime.serving.kserve.io/kserve-xgbserver created [e2e-llm-inference-service] CI namespace setup complete [e2e-llm-inference-service] Setup complete [e2e-llm-inference-service] === E2E cluster / operator summary === [e2e-llm-inference-service] Client Version: 4.20.11 [e2e-llm-inference-service] Kustomize Version: v5.6.0 [e2e-llm-inference-service] Server Version: 4.21.23 [e2e-llm-inference-service] Kubernetes Version: v1.34.8 [e2e-llm-inference-service] ClusterVersion desired: 4.21.23 [e2e-llm-inference-service] ClusterVersion history (latest): 4.21.23 (Completed) [e2e-llm-inference-service] CSVs in kuadrant-system: [e2e-llm-inference-service] authorino-operator.v1.4.1 Succeeded [e2e-llm-inference-service] cert-manager-operator.v1.20.0 Succeeded [e2e-llm-inference-service] dns-operator.v1.4.0 Succeeded [e2e-llm-inference-service] limitador-operator.v1.4.0 Succeeded [e2e-llm-inference-service] rhcl-operator.v1.4.1 Succeeded [e2e-llm-inference-service] CSVs in openshift-keda: [e2e-llm-inference-service] authorino-operator.v1.4.1 Succeeded [e2e-llm-inference-service] cert-manager-operator.v1.20.0 Succeeded [e2e-llm-inference-service] custom-metrics-autoscaler.v2.19.0-1 Succeeded [e2e-llm-inference-service] dns-operator.v1.4.0 Succeeded [e2e-llm-inference-service] limitador-operator.v1.4.0 Succeeded [e2e-llm-inference-service] rhcl-operator.v1.4.1 Succeeded [e2e-llm-inference-service] CSVs in cert-manager-operator: [e2e-llm-inference-service] authorino-operator.v1.4.1 Succeeded [e2e-llm-inference-service] cert-manager-operator.v1.20.0 Succeeded [e2e-llm-inference-service] dns-operator.v1.4.0 Succeeded [e2e-llm-inference-service] limitador-operator.v1.4.0 Succeeded [e2e-llm-inference-service] rhcl-operator.v1.4.1 Succeeded [e2e-llm-inference-service] CSVs in openshift-lws-operator: [e2e-llm-inference-service] authorino-operator.v1.4.1 Succeeded [e2e-llm-inference-service] cert-manager-operator.v1.20.0 Succeeded [e2e-llm-inference-service] dns-operator.v1.4.0 Succeeded [e2e-llm-inference-service] leader-worker-set.v1.0.0 Succeeded [e2e-llm-inference-service] limitador-operator.v1.4.0 Succeeded [e2e-llm-inference-service] rhcl-operator.v1.4.1 Succeeded [e2e-llm-inference-service] CSVs in openshift-operators (ODH / shared operators, filtered): [e2e-llm-inference-service] authorino-operator.v1.4.1 Succeeded [e2e-llm-inference-service] dns-operator.v1.4.0 Succeeded [e2e-llm-inference-service] limitador-operator.v1.4.0 Succeeded [e2e-llm-inference-service] rhcl-operator.v1.4.1 Succeeded [e2e-llm-inference-service] Kuadrant / Authorino (diagnostics): [e2e-llm-inference-service] CRD kuadrants.kuadrant.io versions: v1beta1 served=true storage=true [e2e-llm-inference-service] Subscriptions in kuadrant-system: [e2e-llm-inference-service] authorino-operator-stable-redhat-operators-openshift-marketplace stable redhat-operators authorino-operator.v1.4.1 [e2e-llm-inference-service] dns-operator-stable-redhat-operators-openshift-marketplace stable redhat-operators dns-operator.v1.4.0 [e2e-llm-inference-service] limitador-operator-stable-redhat-operators-openshift-marketplace stable redhat-operators limitador-operator.v1.4.0 [e2e-llm-inference-service] rhcl-operator stable redhat-operators rhcl-operator.v1.4.1 [e2e-llm-inference-service] Kuadrant CR conditions (kuadrant/kuadrant-system): [e2e-llm-inference-service] Ready=True (Ready) [e2e-llm-inference-service] KServe deployments in kserve: [e2e-llm-inference-service] kserve-controller-manager: ready=1 image=quay.io/opendatahub/kserve-controller@sha256:d717f59e7fc6132421a696b06ade56c2521428323575a477d0441e0491bdb66c [e2e-llm-inference-service] imageID: quay.io/opendatahub/kserve-controller@sha256:8907023b4ab54ba592bbb716ca1db137e376c8cc20b09ec775d6662bae986210 [e2e-llm-inference-service] odh-model-controller: ready=1 image=quay.io/opendatahub/odh-model-controller:fast [e2e-llm-inference-service] imageID: quay.io/opendatahub/odh-model-controller@sha256:ba14297d782dbbb694e4f2a6abe57097a1df94a299c0b74e93d304df4467fd9d [e2e-llm-inference-service] llmisvc-controller-manager: ready=1 image=quay.io/opendatahub/odh-kserve-llmisvc-controller@sha256:611712697e2f5e409e4ec5b04fe487b6af8ff2eadd54cbb3a7ef8d15d81f4ae5 [e2e-llm-inference-service] imageID: quay.io/opendatahub/odh-kserve-llmisvc-controller@sha256:57f052785f919c3711c91da64ea097b1b8d366e1ed531c5a20e31c2bba9b1c89 [e2e-llm-inference-service] === End E2E cluster / operator summary === [e2e-llm-inference-service] /workspace/source [e2e-llm-inference-service] CA certificate extracted [e2e-llm-inference-service] REQUESTS_CA_BUNDLE=/tmp/ca.crt [e2e-llm-inference-service] Run E2E tests: llminferenceservice and cluster_cpu and not autoscaling and not tracing [e2e-llm-inference-service] Starting E2E functional tests ... [e2e-llm-inference-service] Parallelism requested for pytest is 2 [e2e-llm-inference-service] ============================= test session starts ============================== [e2e-llm-inference-service] platform linux -- Python 3.11.13, pytest-7.4.4, pluggy-1.5.0 -- /workspace/source/python/kserve/.venv/bin/python [e2e-llm-inference-service] cachedir: .pytest_cache [e2e-llm-inference-service] metadata: {'Python': '3.11.13', 'Platform': 'Linux-5.14.0-570.122.1.el9_6.x86_64-x86_64-with-glibc2.34', 'Packages': {'pytest': '7.4.4', 'pluggy': '1.5.0'}, 'Plugins': {'anyio': '4.9.0', 'asyncio': '0.23.8', 'metadata': '3.1.1', 'httpx': '0.30.0', 'json-report': '1.5.0', 'xdist': '3.6.1', 'cov': '5.0.0'}, 'PLATFORM': 'el9'} [e2e-llm-inference-service] rootdir: /workspace/source/test/e2e [e2e-llm-inference-service] configfile: pytest.ini [e2e-llm-inference-service] plugins: anyio-4.9.0, asyncio-0.23.8, metadata-3.1.1, httpx-0.30.0, json-report-1.5.0, xdist-3.6.1, cov-5.0.0 [e2e-llm-inference-service] asyncio: mode=Mode.STRICT [e2e-llm-inference-service] created: 2/2 workers [e2e-llm-inference-service] 2 workers [42 items] [e2e-llm-inference-service] [e2e-llm-inference-service] scheduling tests via WorkStealingScheduling [e2e-llm-inference-service] [e2e-llm-inference-service] llmisvc/test_gateway_section_name.py::test_gateway_section_name_propagation[cluster_single_node-cluster_cpu-with-section-name] [e2e-llm-inference-service] llmisvc/test_llm_inference_service.py::test_llm_inference_service[cluster_cpu-cluster_single_node-router-managed-scheduler-with-precise-prefix-cache-inline-config-workload-llmd-simulator-kvcache] 2026-07-07 14:21:35.559 6138 kserve INFO [conftest.py:configure_logger():40] Logger configured [e2e-llm-inference-service] 2026-07-07 14:21:35.559 6141 kserve INFO [conftest.py:configure_logger():40] Logger configured [e2e-llm-inference-service] 2026-07-07 14:21:35.572 6138 kserve.trace Checking Gateway router-gateway-1 in namespace kserve-ci-e2e-test [e2e-llm-inference-service] 2026-07-07 14:21:35.572 6138 kserve.trace INFO [gw_api.py:create_or_update_gateway():34] Checking Gateway router-gateway-1 in namespace kserve-ci-e2e-test [e2e-llm-inference-service] 2026-07-07 14:21:35.599 6138 kserve.trace Resource not found, creating Gateway router-gateway-1 [e2e-llm-inference-service] 2026-07-07 14:21:35.599 6138 kserve.trace INFO [gw_api.py:create_or_update_gateway():62] Resource not found, creating Gateway router-gateway-1 [e2e-llm-inference-service] 2026-07-07 14:21:35.614 6138 kserve.trace ✓ Successfully created Gateway router-gateway-1 [e2e-llm-inference-service] 2026-07-07 14:21:35.614 6138 kserve.trace INFO [gw_api.py:create_or_update_gateway():70] ✓ Successfully created Gateway router-gateway-1 [e2e-llm-inference-service] [e2e-llm-inference-service] [gw0] PASSED llmisvc/test_gateway_section_name.py::test_gateway_section_name_propagation[cluster_single_node-cluster_cpu-with-section-name] [e2e-llm-inference-service] llmisvc/test_gateway_section_name.py::test_gateway_section_name_propagation[cluster_single_node-cluster_cpu-without-section-name] 2026-07-07 14:21:47.679 6138 kserve.trace Checking Gateway router-gateway-1 in namespace kserve-ci-e2e-test [e2e-llm-inference-service] 2026-07-07 14:21:47.679 6138 kserve.trace INFO [gw_api.py:create_or_update_gateway():34] Checking Gateway router-gateway-1 in namespace kserve-ci-e2e-test [e2e-llm-inference-service] 2026-07-07 14:21:47.724 6138 kserve.trace ✓ Successfully updated Gateway router-gateway-1 [e2e-llm-inference-service] 2026-07-07 14:21:47.724 6138 kserve.trace INFO [gw_api.py:create_or_update_gateway():57] ✓ Successfully updated Gateway router-gateway-1 [e2e-llm-inference-service] [e2e-llm-inference-service] [gw0] PASSED llmisvc/test_gateway_section_name.py::test_gateway_section_name_propagation[cluster_single_node-cluster_cpu-without-section-name] [e2e-llm-inference-service] llmisvc/test_llm_auth.py::test_llm_auth_enabled_requires_token[cluster_cpu-cluster_single_node-auth-enabled-default] [e2e-llm-inference-service] [gw1] PASSED llmisvc/test_llm_inference_service.py::test_llm_inference_service[cluster_cpu-cluster_single_node-router-managed-scheduler-with-precise-prefix-cache-inline-config-workload-llmd-simulator-kvcache] [e2e-llm-inference-service] llmisvc/test_llm_inference_service.py::test_llm_inference_service[cluster_cpu-cluster_single_node-router-managed-workload-llmd-simulator0] [e2e-llm-inference-service] [gw1] PASSED llmisvc/test_llm_inference_service.py::test_llm_inference_service[cluster_cpu-cluster_single_node-router-managed-workload-llmd-simulator0] [e2e-llm-inference-service] llmisvc/test_llm_inference_service.py::test_llm_inference_service[cluster_cpu-cluster_single_node-router-managed-workload-llmd-simulator1] [e2e-llm-inference-service] [gw0] PASSED llmisvc/test_llm_auth.py::test_llm_auth_enabled_requires_token[cluster_cpu-cluster_single_node-auth-enabled-default] [e2e-llm-inference-service] llmisvc/test_llm_auth.py::test_llm_auth_invalid_token_rejected[cluster_cpu-cluster_single_node-auth-invalid-token] [e2e-llm-inference-service] [gw1] PASSED llmisvc/test_llm_inference_service.py::test_llm_inference_service[cluster_cpu-cluster_single_node-router-managed-workload-llmd-simulator1] [e2e-llm-inference-service] llmisvc/test_llm_inference_service.py::test_llm_inference_service[cluster_cpu-cluster_single_node-router-managed-workload-llmd-simulator2] [e2e-llm-inference-service] [gw0] PASSED llmisvc/test_llm_auth.py::test_llm_auth_invalid_token_rejected[cluster_cpu-cluster_single_node-auth-invalid-token] [e2e-llm-inference-service] llmisvc/test_llm_auth.py::test_llm_auth_disabled_no_token_required[cluster_cpu-cluster_single_node-auth-disabled] [e2e-llm-inference-service] [gw1] PASSED llmisvc/test_llm_inference_service.py::test_llm_inference_service[cluster_cpu-cluster_single_node-router-managed-workload-llmd-simulator2] [e2e-llm-inference-service] llmisvc/test_llm_inference_service.py::test_llm_inference_service[cluster_cpu-cluster_single_node-router-managed-workload-single-cpu-model-fb-opt-125m-with-lora-hf0] [e2e-llm-inference-service] [gw0] PASSED llmisvc/test_llm_auth.py::test_llm_auth_disabled_no_token_required[cluster_cpu-cluster_single_node-auth-disabled] [e2e-llm-inference-service] llmisvc/test_llm_inference_service.py::test_llm_inference_service[cluster_cpu-cluster_single_node-router-with-gateway-ref-router-with-managed-route-model-fb-opt-125m-workload-llmd-simulator] 2026-07-07 14:30:52.809 6138 kserve.trace Checking Gateway router-gateway-1 in namespace kserve-ci-e2e-test [e2e-llm-inference-service] 2026-07-07 14:30:52.809 6138 kserve.trace INFO [gw_api.py:create_or_update_gateway():34] Checking Gateway router-gateway-1 in namespace kserve-ci-e2e-test [e2e-llm-inference-service] 2026-07-07 14:30:52.866 6138 kserve.trace ✓ Successfully updated Gateway router-gateway-1 [e2e-llm-inference-service] 2026-07-07 14:30:52.866 6138 kserve.trace INFO [gw_api.py:create_or_update_gateway():57] ✓ Successfully updated Gateway router-gateway-1 [e2e-llm-inference-service] [e2e-llm-inference-service] [gw0] PASSED llmisvc/test_llm_inference_service.py::test_llm_inference_service[cluster_cpu-cluster_single_node-router-with-gateway-ref-router-with-managed-route-model-fb-opt-125m-workload-llmd-simulator] [e2e-llm-inference-service] llmisvc/test_llm_inference_service.py::test_llm_inference_service[cluster_cpu-cluster_single_node-router-managed-workload-single-cpu-model-fb-opt-125m] [e2e-llm-inference-service] [gw1] FAILED llmisvc/test_llm_inference_service.py::test_llm_inference_service[cluster_cpu-cluster_single_node-router-managed-workload-single-cpu-model-fb-opt-125m-with-lora-hf0] [e2e-llm-inference-service] llmisvc/test_llm_inference_service.py::test_llm_inference_service[cluster_cpu-cluster_single_node-router-managed-workload-single-cpu-model-fb-opt-125m-with-lora-hf1] [e2e-llm-inference-service] [gw0] FAILED llmisvc/test_llm_inference_service.py::test_llm_inference_service[cluster_cpu-cluster_single_node-router-managed-workload-single-cpu-model-fb-opt-125m] [e2e-llm-inference-service] llmisvc/test_llm_inference_service.py::test_llm_inference_service[cluster_cpu-cluster_single_node-router-custom-route-timeout-scheduler-managed-workload-single-cpu-model-fb-opt-125m] [e2e-llm-inference-service] [gw1] FAILED llmisvc/test_llm_inference_service.py::test_llm_inference_service[cluster_cpu-cluster_single_node-router-managed-workload-single-cpu-model-fb-opt-125m-with-lora-hf1] [e2e-llm-inference-service] llmisvc/test_llm_inference_service.py::test_llm_inference_service[cluster_cpu-cluster_single_node-router-managed-workload-single-cpu-model-pvc] [e2e-llm-inference-service] [gw0] FAILED llmisvc/test_llm_inference_service.py::test_llm_inference_service[cluster_cpu-cluster_single_node-router-custom-route-timeout-scheduler-managed-workload-single-cpu-model-fb-opt-125m] [e2e-llm-inference-service] llmisvc/test_llm_inference_service.py::test_llm_inference_service[cluster_cpu-cluster_single_node-router-with-refs-scheduler-managed-workload-single-cpu-model-fb-opt-125m] 2026-07-07 15:06:05.344 6138 kserve.trace Checking Gateway router-gateway-1 in namespace kserve-ci-e2e-test [e2e-llm-inference-service] 2026-07-07 15:06:05.344 6138 kserve.trace INFO [gw_api.py:create_or_update_gateway():34] Checking Gateway router-gateway-1 in namespace kserve-ci-e2e-test [e2e-llm-inference-service] 2026-07-07 15:06:05.380 6138 kserve.trace ✓ Successfully updated Gateway router-gateway-1 [e2e-llm-inference-service] 2026-07-07 15:06:05.380 6138 kserve.trace INFO [gw_api.py:create_or_update_gateway():57] ✓ Successfully updated Gateway router-gateway-1 [e2e-llm-inference-service] 2026-07-07 15:06:05.380 6138 kserve.trace Checking HttpRoute router-route-1 in namespace kserve-ci-e2e-test [e2e-llm-inference-service] 2026-07-07 15:06:05.380 6138 kserve.trace INFO [gw_api.py:create_or_update_route():121] Checking HttpRoute router-route-1 in namespace kserve-ci-e2e-test [e2e-llm-inference-service] 2026-07-07 15:06:05.387 6138 kserve.trace Resource not found, creating HttpRoute router-route-1 [e2e-llm-inference-service] 2026-07-07 15:06:05.387 6138 kserve.trace INFO [gw_api.py:create_or_update_route():149] Resource not found, creating HttpRoute router-route-1 [e2e-llm-inference-service] 2026-07-07 15:06:05.399 6138 kserve.trace ✓ Successfully created HttpRoute router-route-1 [e2e-llm-inference-service] 2026-07-07 15:06:05.399 6138 kserve.trace INFO [gw_api.py:create_or_update_route():157] ✓ Successfully created HttpRoute router-route-1 [e2e-llm-inference-service] 2026-07-07 15:06:05.399 6138 kserve.trace Checking HttpRoute router-route-2 in namespace kserve-ci-e2e-test [e2e-llm-inference-service] 2026-07-07 15:06:05.399 6138 kserve.trace INFO [gw_api.py:create_or_update_route():121] Checking HttpRoute router-route-2 in namespace kserve-ci-e2e-test [e2e-llm-inference-service] 2026-07-07 15:06:05.402 6138 kserve.trace Resource not found, creating HttpRoute router-route-2 [e2e-llm-inference-service] 2026-07-07 15:06:05.402 6138 kserve.trace INFO [gw_api.py:create_or_update_route():149] Resource not found, creating HttpRoute router-route-2 [e2e-llm-inference-service] 2026-07-07 15:06:05.412 6138 kserve.trace ✓ Successfully created HttpRoute router-route-2 [e2e-llm-inference-service] 2026-07-07 15:06:05.412 6138 kserve.trace INFO [gw_api.py:create_or_update_route():157] ✓ Successfully created HttpRoute router-route-2 [e2e-llm-inference-service] [e2e-llm-inference-service] [gw0] PASSED llmisvc/test_llm_inference_service.py::test_llm_inference_service[cluster_cpu-cluster_single_node-router-with-refs-scheduler-managed-workload-single-cpu-model-fb-opt-125m] [e2e-llm-inference-service] llmisvc/test_llm_inference_service.py::test_llm_inference_service[cluster_cpu-cluster_single_node-router-managed-workload-pd-cpu-model-fb-opt-125m] [e2e-llm-inference-service] [gw1] PASSED llmisvc/test_llm_inference_service.py::test_llm_inference_service[cluster_cpu-cluster_single_node-router-managed-workload-single-cpu-model-pvc] [e2e-llm-inference-service] llmisvc/test_llm_inference_service.py::test_llm_inference_service[cluster_cpu-cluster_single_node-router-managed-workload-pd-cpu-model-pvc] [e2e-llm-inference-service] [gw1] PASSED llmisvc/test_llm_inference_service.py::test_llm_inference_service[cluster_cpu-cluster_single_node-router-managed-workload-pd-cpu-model-pvc] [e2e-llm-inference-service] llmisvc/test_llm_inference_service.py::test_llm_inference_service[cluster_cpu-cluster_multi_node-router-managed-workload-simulated-dp-ep-cpu-model-pvc] [e2e-llm-inference-service] [gw0] PASSED llmisvc/test_llm_inference_service.py::test_llm_inference_service[cluster_cpu-cluster_single_node-router-managed-workload-pd-cpu-model-fb-opt-125m] [e2e-llm-inference-service] llmisvc/test_llm_inference_service.py::test_llm_inference_service[cluster_cpu-cluster_single_node-router-custom-route-timeout-pd-scheduler-managed-workload-pd-cpu-model-fb-opt-125m] [e2e-llm-inference-service] [gw0] PASSED llmisvc/test_llm_inference_service.py::test_llm_inference_service[cluster_cpu-cluster_single_node-router-custom-route-timeout-pd-scheduler-managed-workload-pd-cpu-model-fb-opt-125m] [e2e-llm-inference-service] llmisvc/test_llm_inference_service.py::test_llm_inference_service[cluster_cpu-cluster_single_node-router-with-refs-pd-scheduler-managed-workload-pd-cpu-model-fb-opt-125m] 2026-07-07 15:24:08.611 6138 kserve.trace Checking Gateway router-gateway-2 in namespace kserve-ci-e2e-test [e2e-llm-inference-service] 2026-07-07 15:24:08.611 6138 kserve.trace INFO [gw_api.py:create_or_update_gateway():34] Checking Gateway router-gateway-2 in namespace kserve-ci-e2e-test [e2e-llm-inference-service] 2026-07-07 15:24:08.645 6138 kserve.trace Resource not found, creating Gateway router-gateway-2 [e2e-llm-inference-service] 2026-07-07 15:24:08.645 6138 kserve.trace INFO [gw_api.py:create_or_update_gateway():62] Resource not found, creating Gateway router-gateway-2 [e2e-llm-inference-service] 2026-07-07 15:24:08.652 6138 kserve.trace ✓ Successfully created Gateway router-gateway-2 [e2e-llm-inference-service] 2026-07-07 15:24:08.652 6138 kserve.trace INFO [gw_api.py:create_or_update_gateway():70] ✓ Successfully created Gateway router-gateway-2 [e2e-llm-inference-service] 2026-07-07 15:24:08.652 6138 kserve.trace Checking HttpRoute router-route-3 in namespace kserve-ci-e2e-test [e2e-llm-inference-service] 2026-07-07 15:24:08.652 6138 kserve.trace INFO [gw_api.py:create_or_update_route():121] Checking HttpRoute router-route-3 in namespace kserve-ci-e2e-test [e2e-llm-inference-service] 2026-07-07 15:24:08.658 6138 kserve.trace Resource not found, creating HttpRoute router-route-3 [e2e-llm-inference-service] 2026-07-07 15:24:08.658 6138 kserve.trace INFO [gw_api.py:create_or_update_route():149] Resource not found, creating HttpRoute router-route-3 [e2e-llm-inference-service] 2026-07-07 15:24:08.669 6138 kserve.trace ✓ Successfully created HttpRoute router-route-3 [e2e-llm-inference-service] 2026-07-07 15:24:08.669 6138 kserve.trace INFO [gw_api.py:create_or_update_route():157] ✓ Successfully created HttpRoute router-route-3 [e2e-llm-inference-service] 2026-07-07 15:24:08.669 6138 kserve.trace Checking HttpRoute router-route-4 in namespace kserve-ci-e2e-test [e2e-llm-inference-service] 2026-07-07 15:24:08.669 6138 kserve.trace INFO [gw_api.py:create_or_update_route():121] Checking HttpRoute router-route-4 in namespace kserve-ci-e2e-test [e2e-llm-inference-service] 2026-07-07 15:24:08.672 6138 kserve.trace Resource not found, creating HttpRoute router-route-4 [e2e-llm-inference-service] 2026-07-07 15:24:08.672 6138 kserve.trace INFO [gw_api.py:create_or_update_route():149] Resource not found, creating HttpRoute router-route-4 [e2e-llm-inference-service] 2026-07-07 15:24:08.682 6138 kserve.trace ✓ Successfully created HttpRoute router-route-4 [e2e-llm-inference-service] 2026-07-07 15:24:08.682 6138 kserve.trace INFO [gw_api.py:create_or_update_route():157] ✓ Successfully created HttpRoute router-route-4 [e2e-llm-inference-service] [e2e-llm-inference-service] [gw1] PASSED llmisvc/test_llm_inference_service.py::test_llm_inference_service[cluster_cpu-cluster_multi_node-router-managed-workload-simulated-dp-ep-cpu-model-pvc] [e2e-llm-inference-service] llmisvc/test_llm_inference_service_conversion.py::TestLLMInferenceServiceConversion::test_v1alpha1_to_v1alpha2_conversion [e2e-llm-inference-service] [gw1] PASSED llmisvc/test_llm_inference_service_conversion.py::TestLLMInferenceServiceConversion::test_v1alpha1_to_v1alpha2_conversion [e2e-llm-inference-service] llmisvc/test_llm_inference_service_conversion.py::TestLLMInferenceServiceConversion::test_v1alpha2_to_v1alpha1_conversion [e2e-llm-inference-service] [gw1] PASSED llmisvc/test_llm_inference_service_conversion.py::TestLLMInferenceServiceConversion::test_v1alpha2_to_v1alpha1_conversion [e2e-llm-inference-service] llmisvc/test_llm_inference_service_conversion.py::TestLLMInferenceServiceConversion::test_criticality_preservation_via_annotations [e2e-llm-inference-service] [gw1] PASSED llmisvc/test_llm_inference_service_conversion.py::TestLLMInferenceServiceConversion::test_criticality_preservation_via_annotations [e2e-llm-inference-service] llmisvc/test_llm_inference_service_conversion.py::TestLLMInferenceServiceConversion::test_lora_criticality_preservation [e2e-llm-inference-service] [gw1] PASSED llmisvc/test_llm_inference_service_conversion.py::TestLLMInferenceServiceConversion::test_lora_criticality_preservation [e2e-llm-inference-service] llmisvc/test_llm_inference_service_conversion.py::TestLLMInferenceServiceConversion::test_round_trip_conversion_preserves_fields [e2e-llm-inference-service] [gw1] PASSED llmisvc/test_llm_inference_service_conversion.py::TestLLMInferenceServiceConversion::test_round_trip_conversion_preserves_fields [e2e-llm-inference-service] llmisvc/test_llm_inference_service_stop.py::test_llm_stop_feature[cluster_cpu-cluster_single_node-router-managed-workload-single-cpu-model-fb-opt-125m] [e2e-llm-inference-service] [gw0] PASSED llmisvc/test_llm_inference_service.py::test_llm_inference_service[cluster_cpu-cluster_single_node-router-with-refs-pd-scheduler-managed-workload-pd-cpu-model-fb-opt-125m] [e2e-llm-inference-service] llmisvc/test_llm_inference_service.py::test_llm_inference_service[cluster_cpu-cluster_single_node-router-no-scheduler-workload-single-cpu-model-fb-opt-125m] [e2e-llm-inference-service] [gw0] PASSED llmisvc/test_llm_inference_service.py::test_llm_inference_service[cluster_cpu-cluster_single_node-router-no-scheduler-workload-single-cpu-model-fb-opt-125m] [e2e-llm-inference-service] llmisvc/test_llm_inference_service.py::test_llm_inference_service[cluster_cpu-cluster_multi_node-router-managed-workload-simulated-dp-ep-cpu-model-fb-opt-125m] [e2e-llm-inference-service] [gw1] PASSED llmisvc/test_llm_inference_service_stop.py::test_llm_stop_feature[cluster_cpu-cluster_single_node-router-managed-workload-single-cpu-model-fb-opt-125m] [e2e-llm-inference-service] llmisvc/test_llm_lora_adapters.py::test_llm_with_lora_adapters[cluster_cpu-single-lora-adapter-hf] [e2e-llm-inference-service] [gw1] PASSED llmisvc/test_llm_lora_adapters.py::test_llm_with_lora_adapters[cluster_cpu-single-lora-adapter-hf] [e2e-llm-inference-service] llmisvc/test_llm_lora_adapters.py::test_llm_with_lora_adapters[cluster_cpu-multiple-lora-adapters] [e2e-llm-inference-service] [gw0] PASSED llmisvc/test_llm_inference_service.py::test_llm_inference_service[cluster_cpu-cluster_multi_node-router-managed-workload-simulated-dp-ep-cpu-model-fb-opt-125m] [e2e-llm-inference-service] llmisvc/test_llm_inference_service.py::test_llm_inference_service[cluster_cpu-cluster_single_node-router-managed-scheduler-with-inline-config-workload-llmd-simulator] [e2e-llm-inference-service] [gw0] PASSED llmisvc/test_llm_inference_service.py::test_llm_inference_service[cluster_cpu-cluster_single_node-router-managed-scheduler-with-inline-config-workload-llmd-simulator] [e2e-llm-inference-service] llmisvc/test_llm_inference_service.py::test_llm_inference_service[cluster_cpu-cluster_single_node-router-managed-workload-llmd-simulator-model-qwen2.5-0.5b] [e2e-llm-inference-service] [gw1] PASSED llmisvc/test_llm_lora_adapters.py::test_llm_with_lora_adapters[cluster_cpu-multiple-lora-adapters] [e2e-llm-inference-service] llmisvc/test_llm_tls.py::test_llm_tls_resources[cluster_cpu-cluster_single_node-router-managed-workload-single-cpu-model-fb-opt-125m] [e2e-llm-inference-service] [gw0] PASSED llmisvc/test_llm_inference_service.py::test_llm_inference_service[cluster_cpu-cluster_single_node-router-managed-workload-llmd-simulator-model-qwen2.5-0.5b] [e2e-llm-inference-service] llmisvc/test_llm_inference_service.py::test_llm_inference_service[cluster_cpu-cluster_single_node-router-managed-scheduler-with-configmap-ref-workload-llmd-simulator] [e2e-llm-inference-service] [gw0] PASSED llmisvc/test_llm_inference_service.py::test_llm_inference_service[cluster_cpu-cluster_single_node-router-managed-scheduler-with-configmap-ref-workload-llmd-simulator] [e2e-llm-inference-service] llmisvc/test_llm_inference_service.py::test_llm_inference_service[cluster_cpu-cluster_single_node-router-managed-scheduler-with-replicas-workload-llmd-simulator] [e2e-llm-inference-service] [gw1] PASSED llmisvc/test_llm_tls.py::test_llm_tls_resources[cluster_cpu-cluster_single_node-router-managed-workload-single-cpu-model-fb-opt-125m] [e2e-llm-inference-service] llmisvc/test_prestop_hook.py::test_prestop_hook[cluster_cpu-cluster_single_node-router-managed-workload-single-cpu-model-fb-opt-125m] [e2e-llm-inference-service] [gw0] PASSED llmisvc/test_llm_inference_service.py::test_llm_inference_service[cluster_cpu-cluster_single_node-router-managed-scheduler-with-replicas-workload-llmd-simulator] [e2e-llm-inference-service] llmisvc/test_llm_inference_service.py::test_llm_inference_service[cluster_cpu-cluster_single_node-router-managed-scheduler-with-custom-template-workload-llmd-simulator] [e2e-llm-inference-service] [gw0] PASSED llmisvc/test_llm_inference_service.py::test_llm_inference_service[cluster_cpu-cluster_single_node-router-managed-scheduler-with-custom-template-workload-llmd-simulator] [e2e-llm-inference-service] llmisvc/test_llm_inference_service.py::test_llm_inference_service[cluster_cpu-cluster_single_node-router-managed-scheduler-v06-pd-config-migration-workload-llmd-simulator-pd] [e2e-llm-inference-service] [gw0] PASSED llmisvc/test_llm_inference_service.py::test_llm_inference_service[cluster_cpu-cluster_single_node-router-managed-scheduler-v06-pd-config-migration-workload-llmd-simulator-pd] [e2e-llm-inference-service] llmisvc/test_llm_inference_service.py::test_llm_inference_service[cluster_cpu-cluster_single_node-router-managed-scheduler-v06-nonzero-threshold-migration-workload-llmd-simulator-pd] [e2e-llm-inference-service] [gw1] PASSED llmisvc/test_prestop_hook.py::test_prestop_hook[cluster_cpu-cluster_single_node-router-managed-workload-single-cpu-model-fb-opt-125m] [e2e-llm-inference-service] llmisvc/test_rolling_upgrade.py::test_rolling_upgrade_coordination[cluster_cpu-cluster_single_node-router-managed-workload-llmd-simulator-model-fb-opt-125m] [e2e-llm-inference-service] [gw0] PASSED llmisvc/test_llm_inference_service.py::test_llm_inference_service[cluster_cpu-cluster_single_node-router-managed-scheduler-v06-nonzero-threshold-migration-workload-llmd-simulator-pd] [e2e-llm-inference-service] llmisvc/test_storage_version_migration.py::TestStorageVersionMigration::test_storage_version_migration_after_simulated_upgrade [e2e-llm-inference-service] [gw0] PASSED llmisvc/test_storage_version_migration.py::TestStorageVersionMigration::test_storage_version_migration_after_simulated_upgrade [e2e-llm-inference-service] [gw1] FAILED llmisvc/test_rolling_upgrade.py::test_rolling_upgrade_coordination[cluster_cpu-cluster_single_node-router-managed-workload-llmd-simulator-model-fb-opt-125m] [e2e-llm-inference-service] [e2e-llm-inference-service] =================================== FAILURES =================================== [e2e-llm-inference-service] _ test_llm_inference_service[router-managed-workload-single-cpu-model-fb-opt-125m-with-lora-hf0] _ [e2e-llm-inference-service] [gw1] linux -- Python 3.11.13 /workspace/source/python/kserve/.venv/bin/python [e2e-llm-inference-service] [e2e-llm-inference-service] self = [e2e-llm-inference-service] conn = [e2e-llm-inference-service] method = 'POST', url = '/v1/completions' [e2e-llm-inference-service] body = b'{"model": "publishers/kserve-ci-e2e-test/models/lora-adapter-1", "prompt": "KServe is a", "max_tokens": 20}' [e2e-llm-inference-service] headers = {'User-Agent': 'python-requests/2.32.3', 'Accept-Encoding': 'gzip, deflate', 'Accept': '*/*', 'Connection': 'keep-aliv...lication/json', 'X-Gateway-Model-Name': 'publishers/kserve-ci-e2e-test/models/lora-adapter-1', 'Content-Length': '107'} [e2e-llm-inference-service] retries = Retry(total=0, connect=None, read=None, redirect=None, status=None) [e2e-llm-inference-service] timeout = Timeout(connect=60, read=60, total=None), chunked = False [e2e-llm-inference-service] response_conn = [e2e-llm-inference-service] preload_content = False, decode_content = False, enforce_content_length = True [e2e-llm-inference-service] [e2e-llm-inference-service] def _make_request( [e2e-llm-inference-service] self, [e2e-llm-inference-service] conn: BaseHTTPConnection, [e2e-llm-inference-service] method: str, [e2e-llm-inference-service] url: str, [e2e-llm-inference-service] body: _TYPE_BODY | None = None, [e2e-llm-inference-service] headers: typing.Mapping[str, str] | None = None, [e2e-llm-inference-service] retries: Retry | None = None, [e2e-llm-inference-service] timeout: _TYPE_TIMEOUT = _DEFAULT_TIMEOUT, [e2e-llm-inference-service] chunked: bool = False, [e2e-llm-inference-service] response_conn: BaseHTTPConnection | None = None, [e2e-llm-inference-service] preload_content: bool = True, [e2e-llm-inference-service] decode_content: bool = True, [e2e-llm-inference-service] enforce_content_length: bool = True, [e2e-llm-inference-service] ) -> BaseHTTPResponse: [e2e-llm-inference-service] """ [e2e-llm-inference-service] Perform a request on a given urllib connection object taken from our [e2e-llm-inference-service] pool. [e2e-llm-inference-service] [e2e-llm-inference-service] :param conn: [e2e-llm-inference-service] a connection from one of our connection pools [e2e-llm-inference-service] [e2e-llm-inference-service] :param method: [e2e-llm-inference-service] HTTP request method (such as GET, POST, PUT, etc.) [e2e-llm-inference-service] [e2e-llm-inference-service] :param url: [e2e-llm-inference-service] The URL to perform the request on. [e2e-llm-inference-service] [e2e-llm-inference-service] :param body: [e2e-llm-inference-service] Data to send in the request body, either :class:`str`, :class:`bytes`, [e2e-llm-inference-service] an iterable of :class:`str`/:class:`bytes`, or a file-like object. [e2e-llm-inference-service] [e2e-llm-inference-service] :param headers: [e2e-llm-inference-service] Dictionary of custom headers to send, such as User-Agent, [e2e-llm-inference-service] If-None-Match, etc. If None, pool headers are used. If provided, [e2e-llm-inference-service] these headers completely replace any pool-specific headers. [e2e-llm-inference-service] [e2e-llm-inference-service] :param retries: [e2e-llm-inference-service] Configure the number of retries to allow before raising a [e2e-llm-inference-service] :class:`~urllib3.exceptions.MaxRetryError` exception. [e2e-llm-inference-service] [e2e-llm-inference-service] Pass ``None`` to retry until you receive a response. Pass a [e2e-llm-inference-service] :class:`~urllib3.util.retry.Retry` object for fine-grained control [e2e-llm-inference-service] over different types of retries. [e2e-llm-inference-service] Pass an integer number to retry connection errors that many times, [e2e-llm-inference-service] but no other types of errors. Pass zero to never retry. [e2e-llm-inference-service] [e2e-llm-inference-service] If ``False``, then retries are disabled and any exception is raised [e2e-llm-inference-service] immediately. Also, instead of raising a MaxRetryError on redirects, [e2e-llm-inference-service] the redirect response will be returned. [e2e-llm-inference-service] [e2e-llm-inference-service] :type retries: :class:`~urllib3.util.retry.Retry`, False, or an int. [e2e-llm-inference-service] [e2e-llm-inference-service] :param timeout: [e2e-llm-inference-service] If specified, overrides the default timeout for this one [e2e-llm-inference-service] request. It may be a float (in seconds) or an instance of [e2e-llm-inference-service] :class:`urllib3.util.Timeout`. [e2e-llm-inference-service] [e2e-llm-inference-service] :param chunked: [e2e-llm-inference-service] If True, urllib3 will send the body using chunked transfer [e2e-llm-inference-service] encoding. Otherwise, urllib3 will send the body using the standard [e2e-llm-inference-service] content-length form. Defaults to False. [e2e-llm-inference-service] [e2e-llm-inference-service] :param response_conn: [e2e-llm-inference-service] Set this to ``None`` if you will handle releasing the connection or [e2e-llm-inference-service] set the connection to have the response release it. [e2e-llm-inference-service] [e2e-llm-inference-service] :param preload_content: [e2e-llm-inference-service] If True, the response's body will be preloaded during construction. [e2e-llm-inference-service] [e2e-llm-inference-service] :param decode_content: [e2e-llm-inference-service] If True, will attempt to decode the body based on the [e2e-llm-inference-service] 'content-encoding' header. [e2e-llm-inference-service] [e2e-llm-inference-service] :param enforce_content_length: [e2e-llm-inference-service] Enforce content length checking. Body returned by server must match [e2e-llm-inference-service] value of Content-Length header, if present. Otherwise, raise error. [e2e-llm-inference-service] """ [e2e-llm-inference-service] self.num_requests += 1 [e2e-llm-inference-service] [e2e-llm-inference-service] timeout_obj = self._get_timeout(timeout) [e2e-llm-inference-service] timeout_obj.start_connect() [e2e-llm-inference-service] conn.timeout = Timeout.resolve_default_timeout(timeout_obj.connect_timeout) [e2e-llm-inference-service] [e2e-llm-inference-service] try: [e2e-llm-inference-service] # Trigger any extra validation we need to do. [e2e-llm-inference-service] try: [e2e-llm-inference-service] self._validate_conn(conn) [e2e-llm-inference-service] except (SocketTimeout, BaseSSLError) as e: [e2e-llm-inference-service] self._raise_timeout(err=e, url=url, timeout_value=conn.timeout) [e2e-llm-inference-service] raise [e2e-llm-inference-service] [e2e-llm-inference-service] # _validate_conn() starts the connection to an HTTPS proxy [e2e-llm-inference-service] # so we need to wrap errors with 'ProxyError' here too. [e2e-llm-inference-service] except ( [e2e-llm-inference-service] OSError, [e2e-llm-inference-service] NewConnectionError, [e2e-llm-inference-service] TimeoutError, [e2e-llm-inference-service] BaseSSLError, [e2e-llm-inference-service] CertificateError, [e2e-llm-inference-service] SSLError, [e2e-llm-inference-service] ) as e: [e2e-llm-inference-service] new_e: Exception = e [e2e-llm-inference-service] if isinstance(e, (BaseSSLError, CertificateError)): [e2e-llm-inference-service] new_e = SSLError(e) [e2e-llm-inference-service] # If the connection didn't successfully connect to it's proxy [e2e-llm-inference-service] # then there [e2e-llm-inference-service] if isinstance( [e2e-llm-inference-service] new_e, (OSError, NewConnectionError, TimeoutError, SSLError) [e2e-llm-inference-service] ) and (conn and conn.proxy and not conn.has_connected_to_proxy): [e2e-llm-inference-service] new_e = _wrap_proxy_error(new_e, conn.proxy.scheme) [e2e-llm-inference-service] raise new_e [e2e-llm-inference-service] [e2e-llm-inference-service] # conn.request() calls http.client.*.request, not the method in [e2e-llm-inference-service] # urllib3.request. It also calls makefile (recv) on the socket. [e2e-llm-inference-service] try: [e2e-llm-inference-service] conn.request( [e2e-llm-inference-service] method, [e2e-llm-inference-service] url, [e2e-llm-inference-service] body=body, [e2e-llm-inference-service] headers=headers, [e2e-llm-inference-service] chunked=chunked, [e2e-llm-inference-service] preload_content=preload_content, [e2e-llm-inference-service] decode_content=decode_content, [e2e-llm-inference-service] enforce_content_length=enforce_content_length, [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] # We are swallowing BrokenPipeError (errno.EPIPE) since the server is [e2e-llm-inference-service] # legitimately able to close the connection after sending a valid response. [e2e-llm-inference-service] # With this behaviour, the received response is still readable. [e2e-llm-inference-service] except BrokenPipeError: [e2e-llm-inference-service] pass [e2e-llm-inference-service] except OSError as e: [e2e-llm-inference-service] # MacOS/Linux [e2e-llm-inference-service] # EPROTOTYPE and ECONNRESET are needed on macOS [e2e-llm-inference-service] # https://erickt.github.io/blog/2014/11/19/adventures-in-debugging-a-potential-osx-kernel-bug/ [e2e-llm-inference-service] # Condition changed later to emit ECONNRESET instead of only EPROTOTYPE. [e2e-llm-inference-service] if e.errno != errno.EPROTOTYPE and e.errno != errno.ECONNRESET: [e2e-llm-inference-service] raise [e2e-llm-inference-service] [e2e-llm-inference-service] # Reset the timeout for the recv() on the socket [e2e-llm-inference-service] read_timeout = timeout_obj.read_timeout [e2e-llm-inference-service] [e2e-llm-inference-service] if not conn.is_closed: [e2e-llm-inference-service] # In Python 3 socket.py will catch EAGAIN and return None when you [e2e-llm-inference-service] # try and read into the file pointer created by http.client, which [e2e-llm-inference-service] # instead raises a BadStatusLine exception. Instead of catching [e2e-llm-inference-service] # the exception and assuming all BadStatusLine exceptions are read [e2e-llm-inference-service] # timeouts, check for a zero timeout before making the request. [e2e-llm-inference-service] if read_timeout == 0: [e2e-llm-inference-service] raise ReadTimeoutError( [e2e-llm-inference-service] self, url, f"Read timed out. (read timeout={read_timeout})" [e2e-llm-inference-service] ) [e2e-llm-inference-service] conn.timeout = read_timeout [e2e-llm-inference-service] [e2e-llm-inference-service] # Receive the response from the server [e2e-llm-inference-service] try: [e2e-llm-inference-service] > response = conn.getresponse() [e2e-llm-inference-service] [e2e-llm-inference-service] ../../python/kserve/.venv/lib64/python3.11/site-packages/urllib3/connectionpool.py:534: [e2e-llm-inference-service] _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ [e2e-llm-inference-service] [e2e-llm-inference-service] self = [e2e-llm-inference-service] [e2e-llm-inference-service] def getresponse( # type: ignore[override] [e2e-llm-inference-service] self, [e2e-llm-inference-service] ) -> HTTPResponse: [e2e-llm-inference-service] """ [e2e-llm-inference-service] Get the response from the server. [e2e-llm-inference-service] [e2e-llm-inference-service] If the HTTPConnection is in the correct state, returns an instance of HTTPResponse or of whatever object is returned by the response_class variable. [e2e-llm-inference-service] [e2e-llm-inference-service] If a request has not been sent or if a previous response has not be handled, ResponseNotReady is raised. If the HTTP response indicates that the connection should be closed, then it will be closed before the response is returned. When the connection is closed, the underlying socket is closed. [e2e-llm-inference-service] """ [e2e-llm-inference-service] # Raise the same error as http.client.HTTPConnection [e2e-llm-inference-service] if self._response_options is None: [e2e-llm-inference-service] raise ResponseNotReady() [e2e-llm-inference-service] [e2e-llm-inference-service] # Reset this attribute for being used again. [e2e-llm-inference-service] resp_options = self._response_options [e2e-llm-inference-service] self._response_options = None [e2e-llm-inference-service] [e2e-llm-inference-service] # Since the connection's timeout value may have been updated [e2e-llm-inference-service] # we need to set the timeout on the socket. [e2e-llm-inference-service] self.sock.settimeout(self.timeout) [e2e-llm-inference-service] [e2e-llm-inference-service] # This is needed here to avoid circular import errors [e2e-llm-inference-service] from .response import HTTPResponse [e2e-llm-inference-service] [e2e-llm-inference-service] # Save a reference to the shutdown function before ownership is passed [e2e-llm-inference-service] # to httplib_response [e2e-llm-inference-service] # TODO should we implement it everywhere? [e2e-llm-inference-service] _shutdown = getattr(self.sock, "shutdown", None) [e2e-llm-inference-service] [e2e-llm-inference-service] # Get the response from http.client.HTTPConnection [e2e-llm-inference-service] > httplib_response = super().getresponse() [e2e-llm-inference-service] [e2e-llm-inference-service] ../../python/kserve/.venv/lib64/python3.11/site-packages/urllib3/connection.py:571: [e2e-llm-inference-service] _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ [e2e-llm-inference-service] [e2e-llm-inference-service] self = [e2e-llm-inference-service] [e2e-llm-inference-service] def getresponse(self): [e2e-llm-inference-service] """Get the response from the server. [e2e-llm-inference-service] [e2e-llm-inference-service] If the HTTPConnection is in the correct state, returns an [e2e-llm-inference-service] instance of HTTPResponse or of whatever object is returned by [e2e-llm-inference-service] the response_class variable. [e2e-llm-inference-service] [e2e-llm-inference-service] If a request has not been sent or if a previous response has [e2e-llm-inference-service] not be handled, ResponseNotReady is raised. If the HTTP [e2e-llm-inference-service] response indicates that the connection should be closed, then [e2e-llm-inference-service] it will be closed before the response is returned. When the [e2e-llm-inference-service] connection is closed, the underlying socket is closed. [e2e-llm-inference-service] """ [e2e-llm-inference-service] [e2e-llm-inference-service] # if a prior response has been completed, then forget about it. [e2e-llm-inference-service] if self.__response and self.__response.isclosed(): [e2e-llm-inference-service] self.__response = None [e2e-llm-inference-service] [e2e-llm-inference-service] # if a prior response exists, then it must be completed (otherwise, we [e2e-llm-inference-service] # cannot read this response's header to determine the connection-close [e2e-llm-inference-service] # behavior) [e2e-llm-inference-service] # [e2e-llm-inference-service] # note: if a prior response existed, but was connection-close, then the [e2e-llm-inference-service] # socket and response were made independent of this HTTPConnection [e2e-llm-inference-service] # object since a new request requires that we open a whole new [e2e-llm-inference-service] # connection [e2e-llm-inference-service] # [e2e-llm-inference-service] # this means the prior response had one of two states: [e2e-llm-inference-service] # 1) will_close: this connection was reset and the prior socket and [e2e-llm-inference-service] # response operate independently [e2e-llm-inference-service] # 2) persistent: the response was retained and we await its [e2e-llm-inference-service] # isclosed() status to become true. [e2e-llm-inference-service] # [e2e-llm-inference-service] if self.__state != _CS_REQ_SENT or self.__response: [e2e-llm-inference-service] raise ResponseNotReady(self.__state) [e2e-llm-inference-service] [e2e-llm-inference-service] if self.debuglevel > 0: [e2e-llm-inference-service] response = self.response_class(self.sock, self.debuglevel, [e2e-llm-inference-service] method=self._method) [e2e-llm-inference-service] else: [e2e-llm-inference-service] response = self.response_class(self.sock, method=self._method) [e2e-llm-inference-service] [e2e-llm-inference-service] try: [e2e-llm-inference-service] try: [e2e-llm-inference-service] > response.begin() [e2e-llm-inference-service] [e2e-llm-inference-service] /usr/lib64/python3.11/http/client.py:1395: [e2e-llm-inference-service] _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ [e2e-llm-inference-service] [e2e-llm-inference-service] self = [e2e-llm-inference-service] [e2e-llm-inference-service] def begin(self): [e2e-llm-inference-service] if self.headers is not None: [e2e-llm-inference-service] # we've already started reading the response [e2e-llm-inference-service] return [e2e-llm-inference-service] [e2e-llm-inference-service] # read until we get a non-100 response [e2e-llm-inference-service] while True: [e2e-llm-inference-service] > version, status, reason = self._read_status() [e2e-llm-inference-service] [e2e-llm-inference-service] /usr/lib64/python3.11/http/client.py:325: [e2e-llm-inference-service] _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ [e2e-llm-inference-service] [e2e-llm-inference-service] self = [e2e-llm-inference-service] [e2e-llm-inference-service] def _read_status(self): [e2e-llm-inference-service] > line = str(self.fp.readline(_MAXLINE + 1), "iso-8859-1") [e2e-llm-inference-service] [e2e-llm-inference-service] /usr/lib64/python3.11/http/client.py:286: [e2e-llm-inference-service] _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ [e2e-llm-inference-service] [e2e-llm-inference-service] self = [e2e-llm-inference-service] b = [e2e-llm-inference-service] [e2e-llm-inference-service] def readinto(self, b): [e2e-llm-inference-service] """Read up to len(b) bytes into the writable buffer *b* and return [e2e-llm-inference-service] the number of bytes read. If the socket is non-blocking and no bytes [e2e-llm-inference-service] are available, None is returned. [e2e-llm-inference-service] [e2e-llm-inference-service] If *b* is non-empty, a 0 return value indicates that the connection [e2e-llm-inference-service] was shutdown at the other end. [e2e-llm-inference-service] """ [e2e-llm-inference-service] self._checkClosed() [e2e-llm-inference-service] self._checkReadable() [e2e-llm-inference-service] if self._timeout_occurred: [e2e-llm-inference-service] raise OSError("cannot read from timed out object") [e2e-llm-inference-service] while True: [e2e-llm-inference-service] try: [e2e-llm-inference-service] > return self._sock.recv_into(b) [e2e-llm-inference-service] E TimeoutError: timed out [e2e-llm-inference-service] [e2e-llm-inference-service] /usr/lib64/python3.11/socket.py:718: TimeoutError [e2e-llm-inference-service] [e2e-llm-inference-service] The above exception was the direct cause of the following exception: [e2e-llm-inference-service] [e2e-llm-inference-service] self = [e2e-llm-inference-service] method = 'POST', url = '/v1/completions' [e2e-llm-inference-service] body = b'{"model": "publishers/kserve-ci-e2e-test/models/lora-adapter-1", "prompt": "KServe is a", "max_tokens": 20}' [e2e-llm-inference-service] headers = {'User-Agent': 'python-requests/2.32.3', 'Accept-Encoding': 'gzip, deflate', 'Accept': '*/*', 'Connection': 'keep-aliv...lication/json', 'X-Gateway-Model-Name': 'publishers/kserve-ci-e2e-test/models/lora-adapter-1', 'Content-Length': '107'} [e2e-llm-inference-service] retries = Retry(total=0, connect=None, read=None, redirect=None, status=None) [e2e-llm-inference-service] redirect = False, assert_same_host = False [e2e-llm-inference-service] timeout = Timeout(connect=60, read=60, total=None), pool_timeout = None [e2e-llm-inference-service] release_conn = False, chunked = False, body_pos = None, preload_content = False [e2e-llm-inference-service] decode_content = False, response_kw = {} [e2e-llm-inference-service] parsed_url = Url(scheme=None, auth=None, host=None, port=None, path='/v1/completions', query=None, fragment=None) [e2e-llm-inference-service] destination_scheme = None, conn = None, release_this_conn = True [e2e-llm-inference-service] http_tunnel_required = False, err = None, clean_exit = False [e2e-llm-inference-service] [e2e-llm-inference-service] def urlopen( # type: ignore[override] [e2e-llm-inference-service] self, [e2e-llm-inference-service] method: str, [e2e-llm-inference-service] url: str, [e2e-llm-inference-service] body: _TYPE_BODY | None = None, [e2e-llm-inference-service] headers: typing.Mapping[str, str] | None = None, [e2e-llm-inference-service] retries: Retry | bool | int | None = None, [e2e-llm-inference-service] redirect: bool = True, [e2e-llm-inference-service] assert_same_host: bool = True, [e2e-llm-inference-service] timeout: _TYPE_TIMEOUT = _DEFAULT_TIMEOUT, [e2e-llm-inference-service] pool_timeout: int | None = None, [e2e-llm-inference-service] release_conn: bool | None = None, [e2e-llm-inference-service] chunked: bool = False, [e2e-llm-inference-service] body_pos: _TYPE_BODY_POSITION | None = None, [e2e-llm-inference-service] preload_content: bool = True, [e2e-llm-inference-service] decode_content: bool = True, [e2e-llm-inference-service] **response_kw: typing.Any, [e2e-llm-inference-service] ) -> BaseHTTPResponse: [e2e-llm-inference-service] """ [e2e-llm-inference-service] Get a connection from the pool and perform an HTTP request. This is the [e2e-llm-inference-service] lowest level call for making a request, so you'll need to specify all [e2e-llm-inference-service] the raw details. [e2e-llm-inference-service] [e2e-llm-inference-service] .. note:: [e2e-llm-inference-service] [e2e-llm-inference-service] More commonly, it's appropriate to use a convenience method [e2e-llm-inference-service] such as :meth:`request`. [e2e-llm-inference-service] [e2e-llm-inference-service] .. note:: [e2e-llm-inference-service] [e2e-llm-inference-service] `release_conn` will only behave as expected if [e2e-llm-inference-service] `preload_content=False` because we want to make [e2e-llm-inference-service] `preload_content=False` the default behaviour someday soon without [e2e-llm-inference-service] breaking backwards compatibility. [e2e-llm-inference-service] [e2e-llm-inference-service] :param method: [e2e-llm-inference-service] HTTP request method (such as GET, POST, PUT, etc.) [e2e-llm-inference-service] [e2e-llm-inference-service] :param url: [e2e-llm-inference-service] The URL to perform the request on. [e2e-llm-inference-service] [e2e-llm-inference-service] :param body: [e2e-llm-inference-service] Data to send in the request body, either :class:`str`, :class:`bytes`, [e2e-llm-inference-service] an iterable of :class:`str`/:class:`bytes`, or a file-like object. [e2e-llm-inference-service] [e2e-llm-inference-service] :param headers: [e2e-llm-inference-service] Dictionary of custom headers to send, such as User-Agent, [e2e-llm-inference-service] If-None-Match, etc. If None, pool headers are used. If provided, [e2e-llm-inference-service] these headers completely replace any pool-specific headers. [e2e-llm-inference-service] [e2e-llm-inference-service] :param retries: [e2e-llm-inference-service] Configure the number of retries to allow before raising a [e2e-llm-inference-service] :class:`~urllib3.exceptions.MaxRetryError` exception. [e2e-llm-inference-service] [e2e-llm-inference-service] If ``None`` (default) will retry 3 times, see ``Retry.DEFAULT``. Pass a [e2e-llm-inference-service] :class:`~urllib3.util.retry.Retry` object for fine-grained control [e2e-llm-inference-service] over different types of retries. [e2e-llm-inference-service] Pass an integer number to retry connection errors that many times, [e2e-llm-inference-service] but no other types of errors. Pass zero to never retry. [e2e-llm-inference-service] [e2e-llm-inference-service] If ``False``, then retries are disabled and any exception is raised [e2e-llm-inference-service] immediately. Also, instead of raising a MaxRetryError on redirects, [e2e-llm-inference-service] the redirect response will be returned. [e2e-llm-inference-service] [e2e-llm-inference-service] :type retries: :class:`~urllib3.util.retry.Retry`, False, or an int. [e2e-llm-inference-service] [e2e-llm-inference-service] :param redirect: [e2e-llm-inference-service] If True, automatically handle redirects (status codes 301, 302, [e2e-llm-inference-service] 303, 307, 308). Each redirect counts as a retry. Disabling retries [e2e-llm-inference-service] will disable redirect, too. [e2e-llm-inference-service] [e2e-llm-inference-service] :param assert_same_host: [e2e-llm-inference-service] If ``True``, will make sure that the host of the pool requests is [e2e-llm-inference-service] consistent else will raise HostChangedError. When ``False``, you can [e2e-llm-inference-service] use the pool on an HTTP proxy and request foreign hosts. [e2e-llm-inference-service] [e2e-llm-inference-service] :param timeout: [e2e-llm-inference-service] If specified, overrides the default timeout for this one [e2e-llm-inference-service] request. It may be a float (in seconds) or an instance of [e2e-llm-inference-service] :class:`urllib3.util.Timeout`. [e2e-llm-inference-service] [e2e-llm-inference-service] :param pool_timeout: [e2e-llm-inference-service] If set and the pool is set to block=True, then this method will [e2e-llm-inference-service] block for ``pool_timeout`` seconds and raise EmptyPoolError if no [e2e-llm-inference-service] connection is available within the time period. [e2e-llm-inference-service] [e2e-llm-inference-service] :param bool preload_content: [e2e-llm-inference-service] If True, the response's body will be preloaded into memory. [e2e-llm-inference-service] [e2e-llm-inference-service] :param bool decode_content: [e2e-llm-inference-service] If True, will attempt to decode the body based on the [e2e-llm-inference-service] 'content-encoding' header. [e2e-llm-inference-service] [e2e-llm-inference-service] :param release_conn: [e2e-llm-inference-service] If False, then the urlopen call will not release the connection [e2e-llm-inference-service] back into the pool once a response is received (but will release if [e2e-llm-inference-service] you read the entire contents of the response such as when [e2e-llm-inference-service] `preload_content=True`). This is useful if you're not preloading [e2e-llm-inference-service] the response's content immediately. You will need to call [e2e-llm-inference-service] ``r.release_conn()`` on the response ``r`` to return the connection [e2e-llm-inference-service] back into the pool. If None, it takes the value of ``preload_content`` [e2e-llm-inference-service] which defaults to ``True``. [e2e-llm-inference-service] [e2e-llm-inference-service] :param bool chunked: [e2e-llm-inference-service] If True, urllib3 will send the body using chunked transfer [e2e-llm-inference-service] encoding. Otherwise, urllib3 will send the body using the standard [e2e-llm-inference-service] content-length form. Defaults to False. [e2e-llm-inference-service] [e2e-llm-inference-service] :param int body_pos: [e2e-llm-inference-service] Position to seek to in file-like body in the event of a retry or [e2e-llm-inference-service] redirect. Typically this won't need to be set because urllib3 will [e2e-llm-inference-service] auto-populate the value when needed. [e2e-llm-inference-service] """ [e2e-llm-inference-service] parsed_url = parse_url(url) [e2e-llm-inference-service] destination_scheme = parsed_url.scheme [e2e-llm-inference-service] [e2e-llm-inference-service] if headers is None: [e2e-llm-inference-service] headers = self.headers [e2e-llm-inference-service] [e2e-llm-inference-service] if not isinstance(retries, Retry): [e2e-llm-inference-service] retries = Retry.from_int(retries, redirect=redirect, default=self.retries) [e2e-llm-inference-service] [e2e-llm-inference-service] if release_conn is None: [e2e-llm-inference-service] release_conn = preload_content [e2e-llm-inference-service] [e2e-llm-inference-service] # Check host [e2e-llm-inference-service] if assert_same_host and not self.is_same_host(url): [e2e-llm-inference-service] raise HostChangedError(self, url, retries) [e2e-llm-inference-service] [e2e-llm-inference-service] # Ensure that the URL we're connecting to is properly encoded [e2e-llm-inference-service] if url.startswith("/"): [e2e-llm-inference-service] url = to_str(_encode_target(url)) [e2e-llm-inference-service] else: [e2e-llm-inference-service] url = to_str(parsed_url.url) [e2e-llm-inference-service] [e2e-llm-inference-service] conn = None [e2e-llm-inference-service] [e2e-llm-inference-service] # Track whether `conn` needs to be released before [e2e-llm-inference-service] # returning/raising/recursing. Update this variable if necessary, and [e2e-llm-inference-service] # leave `release_conn` constant throughout the function. That way, if [e2e-llm-inference-service] # the function recurses, the original value of `release_conn` will be [e2e-llm-inference-service] # passed down into the recursive call, and its value will be respected. [e2e-llm-inference-service] # [e2e-llm-inference-service] # See issue #651 [1] for details. [e2e-llm-inference-service] # [e2e-llm-inference-service] # [1] [e2e-llm-inference-service] release_this_conn = release_conn [e2e-llm-inference-service] [e2e-llm-inference-service] http_tunnel_required = connection_requires_http_tunnel( [e2e-llm-inference-service] self.proxy, self.proxy_config, destination_scheme [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] # Merge the proxy headers. Only done when not using HTTP CONNECT. We [e2e-llm-inference-service] # have to copy the headers dict so we can safely change it without those [e2e-llm-inference-service] # changes being reflected in anyone else's copy. [e2e-llm-inference-service] if not http_tunnel_required: [e2e-llm-inference-service] headers = headers.copy() # type: ignore[attr-defined] [e2e-llm-inference-service] headers.update(self.proxy_headers) # type: ignore[union-attr] [e2e-llm-inference-service] [e2e-llm-inference-service] # Must keep the exception bound to a separate variable or else Python 3 [e2e-llm-inference-service] # complains about UnboundLocalError. [e2e-llm-inference-service] err = None [e2e-llm-inference-service] [e2e-llm-inference-service] # Keep track of whether we cleanly exited the except block. This [e2e-llm-inference-service] # ensures we do proper cleanup in finally. [e2e-llm-inference-service] clean_exit = False [e2e-llm-inference-service] [e2e-llm-inference-service] # Rewind body position, if needed. Record current position [e2e-llm-inference-service] # for future rewinds in the event of a redirect/retry. [e2e-llm-inference-service] body_pos = set_file_position(body, body_pos) [e2e-llm-inference-service] [e2e-llm-inference-service] try: [e2e-llm-inference-service] # Request a connection from the queue. [e2e-llm-inference-service] timeout_obj = self._get_timeout(timeout) [e2e-llm-inference-service] conn = self._get_conn(timeout=pool_timeout) [e2e-llm-inference-service] [e2e-llm-inference-service] conn.timeout = timeout_obj.connect_timeout # type: ignore[assignment] [e2e-llm-inference-service] [e2e-llm-inference-service] # Is this a closed/new connection that requires CONNECT tunnelling? [e2e-llm-inference-service] if self.proxy is not None and http_tunnel_required and conn.is_closed: [e2e-llm-inference-service] try: [e2e-llm-inference-service] self._prepare_proxy(conn) [e2e-llm-inference-service] except (BaseSSLError, OSError, SocketTimeout) as e: [e2e-llm-inference-service] self._raise_timeout( [e2e-llm-inference-service] err=e, url=self.proxy.url, timeout_value=conn.timeout [e2e-llm-inference-service] ) [e2e-llm-inference-service] raise [e2e-llm-inference-service] [e2e-llm-inference-service] # If we're going to release the connection in ``finally:``, then [e2e-llm-inference-service] # the response doesn't need to know about the connection. Otherwise [e2e-llm-inference-service] # it will also try to release it and we'll have a double-release [e2e-llm-inference-service] # mess. [e2e-llm-inference-service] response_conn = conn if not release_conn else None [e2e-llm-inference-service] [e2e-llm-inference-service] # Make the request on the HTTPConnection object [e2e-llm-inference-service] > response = self._make_request( [e2e-llm-inference-service] conn, [e2e-llm-inference-service] method, [e2e-llm-inference-service] url, [e2e-llm-inference-service] timeout=timeout_obj, [e2e-llm-inference-service] body=body, [e2e-llm-inference-service] headers=headers, [e2e-llm-inference-service] chunked=chunked, [e2e-llm-inference-service] retries=retries, [e2e-llm-inference-service] response_conn=response_conn, [e2e-llm-inference-service] preload_content=preload_content, [e2e-llm-inference-service] decode_content=decode_content, [e2e-llm-inference-service] **response_kw, [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] ../../python/kserve/.venv/lib64/python3.11/site-packages/urllib3/connectionpool.py:787: [e2e-llm-inference-service] _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ [e2e-llm-inference-service] [e2e-llm-inference-service] self = [e2e-llm-inference-service] conn = [e2e-llm-inference-service] method = 'POST', url = '/v1/completions' [e2e-llm-inference-service] body = b'{"model": "publishers/kserve-ci-e2e-test/models/lora-adapter-1", "prompt": "KServe is a", "max_tokens": 20}' [e2e-llm-inference-service] headers = {'User-Agent': 'python-requests/2.32.3', 'Accept-Encoding': 'gzip, deflate', 'Accept': '*/*', 'Connection': 'keep-aliv...lication/json', 'X-Gateway-Model-Name': 'publishers/kserve-ci-e2e-test/models/lora-adapter-1', 'Content-Length': '107'} [e2e-llm-inference-service] retries = Retry(total=0, connect=None, read=None, redirect=None, status=None) [e2e-llm-inference-service] timeout = Timeout(connect=60, read=60, total=None), chunked = False [e2e-llm-inference-service] response_conn = [e2e-llm-inference-service] preload_content = False, decode_content = False, enforce_content_length = True [e2e-llm-inference-service] [e2e-llm-inference-service] def _make_request( [e2e-llm-inference-service] self, [e2e-llm-inference-service] conn: BaseHTTPConnection, [e2e-llm-inference-service] method: str, [e2e-llm-inference-service] url: str, [e2e-llm-inference-service] body: _TYPE_BODY | None = None, [e2e-llm-inference-service] headers: typing.Mapping[str, str] | None = None, [e2e-llm-inference-service] retries: Retry | None = None, [e2e-llm-inference-service] timeout: _TYPE_TIMEOUT = _DEFAULT_TIMEOUT, [e2e-llm-inference-service] chunked: bool = False, [e2e-llm-inference-service] response_conn: BaseHTTPConnection | None = None, [e2e-llm-inference-service] preload_content: bool = True, [e2e-llm-inference-service] decode_content: bool = True, [e2e-llm-inference-service] enforce_content_length: bool = True, [e2e-llm-inference-service] ) -> BaseHTTPResponse: [e2e-llm-inference-service] """ [e2e-llm-inference-service] Perform a request on a given urllib connection object taken from our [e2e-llm-inference-service] pool. [e2e-llm-inference-service] [e2e-llm-inference-service] :param conn: [e2e-llm-inference-service] a connection from one of our connection pools [e2e-llm-inference-service] [e2e-llm-inference-service] :param method: [e2e-llm-inference-service] HTTP request method (such as GET, POST, PUT, etc.) [e2e-llm-inference-service] [e2e-llm-inference-service] :param url: [e2e-llm-inference-service] The URL to perform the request on. [e2e-llm-inference-service] [e2e-llm-inference-service] :param body: [e2e-llm-inference-service] Data to send in the request body, either :class:`str`, :class:`bytes`, [e2e-llm-inference-service] an iterable of :class:`str`/:class:`bytes`, or a file-like object. [e2e-llm-inference-service] [e2e-llm-inference-service] :param headers: [e2e-llm-inference-service] Dictionary of custom headers to send, such as User-Agent, [e2e-llm-inference-service] If-None-Match, etc. If None, pool headers are used. If provided, [e2e-llm-inference-service] these headers completely replace any pool-specific headers. [e2e-llm-inference-service] [e2e-llm-inference-service] :param retries: [e2e-llm-inference-service] Configure the number of retries to allow before raising a [e2e-llm-inference-service] :class:`~urllib3.exceptions.MaxRetryError` exception. [e2e-llm-inference-service] [e2e-llm-inference-service] Pass ``None`` to retry until you receive a response. Pass a [e2e-llm-inference-service] :class:`~urllib3.util.retry.Retry` object for fine-grained control [e2e-llm-inference-service] over different types of retries. [e2e-llm-inference-service] Pass an integer number to retry connection errors that many times, [e2e-llm-inference-service] but no other types of errors. Pass zero to never retry. [e2e-llm-inference-service] [e2e-llm-inference-service] If ``False``, then retries are disabled and any exception is raised [e2e-llm-inference-service] immediately. Also, instead of raising a MaxRetryError on redirects, [e2e-llm-inference-service] the redirect response will be returned. [e2e-llm-inference-service] [e2e-llm-inference-service] :type retries: :class:`~urllib3.util.retry.Retry`, False, or an int. [e2e-llm-inference-service] [e2e-llm-inference-service] :param timeout: [e2e-llm-inference-service] If specified, overrides the default timeout for this one [e2e-llm-inference-service] request. It may be a float (in seconds) or an instance of [e2e-llm-inference-service] :class:`urllib3.util.Timeout`. [e2e-llm-inference-service] [e2e-llm-inference-service] :param chunked: [e2e-llm-inference-service] If True, urllib3 will send the body using chunked transfer [e2e-llm-inference-service] encoding. Otherwise, urllib3 will send the body using the standard [e2e-llm-inference-service] content-length form. Defaults to False. [e2e-llm-inference-service] [e2e-llm-inference-service] :param response_conn: [e2e-llm-inference-service] Set this to ``None`` if you will handle releasing the connection or [e2e-llm-inference-service] set the connection to have the response release it. [e2e-llm-inference-service] [e2e-llm-inference-service] :param preload_content: [e2e-llm-inference-service] If True, the response's body will be preloaded during construction. [e2e-llm-inference-service] [e2e-llm-inference-service] :param decode_content: [e2e-llm-inference-service] If True, will attempt to decode the body based on the [e2e-llm-inference-service] 'content-encoding' header. [e2e-llm-inference-service] [e2e-llm-inference-service] :param enforce_content_length: [e2e-llm-inference-service] Enforce content length checking. Body returned by server must match [e2e-llm-inference-service] value of Content-Length header, if present. Otherwise, raise error. [e2e-llm-inference-service] """ [e2e-llm-inference-service] self.num_requests += 1 [e2e-llm-inference-service] [e2e-llm-inference-service] timeout_obj = self._get_timeout(timeout) [e2e-llm-inference-service] timeout_obj.start_connect() [e2e-llm-inference-service] conn.timeout = Timeout.resolve_default_timeout(timeout_obj.connect_timeout) [e2e-llm-inference-service] [e2e-llm-inference-service] try: [e2e-llm-inference-service] # Trigger any extra validation we need to do. [e2e-llm-inference-service] try: [e2e-llm-inference-service] self._validate_conn(conn) [e2e-llm-inference-service] except (SocketTimeout, BaseSSLError) as e: [e2e-llm-inference-service] self._raise_timeout(err=e, url=url, timeout_value=conn.timeout) [e2e-llm-inference-service] raise [e2e-llm-inference-service] [e2e-llm-inference-service] # _validate_conn() starts the connection to an HTTPS proxy [e2e-llm-inference-service] # so we need to wrap errors with 'ProxyError' here too. [e2e-llm-inference-service] except ( [e2e-llm-inference-service] OSError, [e2e-llm-inference-service] NewConnectionError, [e2e-llm-inference-service] TimeoutError, [e2e-llm-inference-service] BaseSSLError, [e2e-llm-inference-service] CertificateError, [e2e-llm-inference-service] SSLError, [e2e-llm-inference-service] ) as e: [e2e-llm-inference-service] new_e: Exception = e [e2e-llm-inference-service] if isinstance(e, (BaseSSLError, CertificateError)): [e2e-llm-inference-service] new_e = SSLError(e) [e2e-llm-inference-service] # If the connection didn't successfully connect to it's proxy [e2e-llm-inference-service] # then there [e2e-llm-inference-service] if isinstance( [e2e-llm-inference-service] new_e, (OSError, NewConnectionError, TimeoutError, SSLError) [e2e-llm-inference-service] ) and (conn and conn.proxy and not conn.has_connected_to_proxy): [e2e-llm-inference-service] new_e = _wrap_proxy_error(new_e, conn.proxy.scheme) [e2e-llm-inference-service] raise new_e [e2e-llm-inference-service] [e2e-llm-inference-service] # conn.request() calls http.client.*.request, not the method in [e2e-llm-inference-service] # urllib3.request. It also calls makefile (recv) on the socket. [e2e-llm-inference-service] try: [e2e-llm-inference-service] conn.request( [e2e-llm-inference-service] method, [e2e-llm-inference-service] url, [e2e-llm-inference-service] body=body, [e2e-llm-inference-service] headers=headers, [e2e-llm-inference-service] chunked=chunked, [e2e-llm-inference-service] preload_content=preload_content, [e2e-llm-inference-service] decode_content=decode_content, [e2e-llm-inference-service] enforce_content_length=enforce_content_length, [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] # We are swallowing BrokenPipeError (errno.EPIPE) since the server is [e2e-llm-inference-service] # legitimately able to close the connection after sending a valid response. [e2e-llm-inference-service] # With this behaviour, the received response is still readable. [e2e-llm-inference-service] except BrokenPipeError: [e2e-llm-inference-service] pass [e2e-llm-inference-service] except OSError as e: [e2e-llm-inference-service] # MacOS/Linux [e2e-llm-inference-service] # EPROTOTYPE and ECONNRESET are needed on macOS [e2e-llm-inference-service] # https://erickt.github.io/blog/2014/11/19/adventures-in-debugging-a-potential-osx-kernel-bug/ [e2e-llm-inference-service] # Condition changed later to emit ECONNRESET instead of only EPROTOTYPE. [e2e-llm-inference-service] if e.errno != errno.EPROTOTYPE and e.errno != errno.ECONNRESET: [e2e-llm-inference-service] raise [e2e-llm-inference-service] [e2e-llm-inference-service] # Reset the timeout for the recv() on the socket [e2e-llm-inference-service] read_timeout = timeout_obj.read_timeout [e2e-llm-inference-service] [e2e-llm-inference-service] if not conn.is_closed: [e2e-llm-inference-service] # In Python 3 socket.py will catch EAGAIN and return None when you [e2e-llm-inference-service] # try and read into the file pointer created by http.client, which [e2e-llm-inference-service] # instead raises a BadStatusLine exception. Instead of catching [e2e-llm-inference-service] # the exception and assuming all BadStatusLine exceptions are read [e2e-llm-inference-service] # timeouts, check for a zero timeout before making the request. [e2e-llm-inference-service] if read_timeout == 0: [e2e-llm-inference-service] raise ReadTimeoutError( [e2e-llm-inference-service] self, url, f"Read timed out. (read timeout={read_timeout})" [e2e-llm-inference-service] ) [e2e-llm-inference-service] conn.timeout = read_timeout [e2e-llm-inference-service] [e2e-llm-inference-service] # Receive the response from the server [e2e-llm-inference-service] try: [e2e-llm-inference-service] response = conn.getresponse() [e2e-llm-inference-service] except (BaseSSLError, OSError) as e: [e2e-llm-inference-service] > self._raise_timeout(err=e, url=url, timeout_value=read_timeout) [e2e-llm-inference-service] [e2e-llm-inference-service] ../../python/kserve/.venv/lib64/python3.11/site-packages/urllib3/connectionpool.py:536: [e2e-llm-inference-service] _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ [e2e-llm-inference-service] [e2e-llm-inference-service] self = [e2e-llm-inference-service] err = TimeoutError('timed out'), url = '/v1/completions', timeout_value = 60 [e2e-llm-inference-service] [e2e-llm-inference-service] def _raise_timeout( [e2e-llm-inference-service] self, [e2e-llm-inference-service] err: BaseSSLError | OSError | SocketTimeout, [e2e-llm-inference-service] url: str, [e2e-llm-inference-service] timeout_value: _TYPE_TIMEOUT | None, [e2e-llm-inference-service] ) -> None: [e2e-llm-inference-service] """Is the error actually a timeout? Will raise a ReadTimeout or pass""" [e2e-llm-inference-service] [e2e-llm-inference-service] if isinstance(err, SocketTimeout): [e2e-llm-inference-service] > raise ReadTimeoutError( [e2e-llm-inference-service] self, url, f"Read timed out. (read timeout={timeout_value})" [e2e-llm-inference-service] ) from err [e2e-llm-inference-service] E urllib3.exceptions.ReadTimeoutError: HTTPConnectionPool(host='ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com', port=80): Read timed out. (read timeout=60) [e2e-llm-inference-service] [e2e-llm-inference-service] ../../python/kserve/.venv/lib64/python3.11/site-packages/urllib3/connectionpool.py:367: ReadTimeoutError [e2e-llm-inference-service] [e2e-llm-inference-service] The above exception was the direct cause of the following exception: [e2e-llm-inference-service] [e2e-llm-inference-service] self = [e2e-llm-inference-service] request = , stream = False [e2e-llm-inference-service] timeout = Timeout(connect=60, read=60, total=None), verify = '/tmp/ca.crt' [e2e-llm-inference-service] cert = None, proxies = OrderedDict() [e2e-llm-inference-service] [e2e-llm-inference-service] def send( [e2e-llm-inference-service] self, request, stream=False, timeout=None, verify=True, cert=None, proxies=None [e2e-llm-inference-service] ): [e2e-llm-inference-service] """Sends PreparedRequest object. Returns Response object. [e2e-llm-inference-service] [e2e-llm-inference-service] :param request: The :class:`PreparedRequest ` being sent. [e2e-llm-inference-service] :param stream: (optional) Whether to stream the request content. [e2e-llm-inference-service] :param timeout: (optional) How long to wait for the server to send [e2e-llm-inference-service] data before giving up, as a float, or a :ref:`(connect timeout, [e2e-llm-inference-service] read timeout) ` tuple. [e2e-llm-inference-service] :type timeout: float or tuple or urllib3 Timeout object [e2e-llm-inference-service] :param verify: (optional) Either a boolean, in which case it controls whether [e2e-llm-inference-service] we verify the server's TLS certificate, or a string, in which case it [e2e-llm-inference-service] must be a path to a CA bundle to use [e2e-llm-inference-service] :param cert: (optional) Any user-provided SSL certificate to be trusted. [e2e-llm-inference-service] :param proxies: (optional) The proxies dictionary to apply to the request. [e2e-llm-inference-service] :rtype: requests.Response [e2e-llm-inference-service] """ [e2e-llm-inference-service] [e2e-llm-inference-service] try: [e2e-llm-inference-service] conn = self.get_connection_with_tls_context( [e2e-llm-inference-service] request, verify, proxies=proxies, cert=cert [e2e-llm-inference-service] ) [e2e-llm-inference-service] except LocationValueError as e: [e2e-llm-inference-service] raise InvalidURL(e, request=request) [e2e-llm-inference-service] [e2e-llm-inference-service] self.cert_verify(conn, request.url, verify, cert) [e2e-llm-inference-service] url = self.request_url(request, proxies) [e2e-llm-inference-service] self.add_headers( [e2e-llm-inference-service] request, [e2e-llm-inference-service] stream=stream, [e2e-llm-inference-service] timeout=timeout, [e2e-llm-inference-service] verify=verify, [e2e-llm-inference-service] cert=cert, [e2e-llm-inference-service] proxies=proxies, [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] chunked = not (request.body is None or "Content-Length" in request.headers) [e2e-llm-inference-service] [e2e-llm-inference-service] if isinstance(timeout, tuple): [e2e-llm-inference-service] try: [e2e-llm-inference-service] connect, read = timeout [e2e-llm-inference-service] timeout = TimeoutSauce(connect=connect, read=read) [e2e-llm-inference-service] except ValueError: [e2e-llm-inference-service] raise ValueError( [e2e-llm-inference-service] f"Invalid timeout {timeout}. Pass a (connect, read) timeout tuple, " [e2e-llm-inference-service] f"or a single float to set both timeouts to the same value." [e2e-llm-inference-service] ) [e2e-llm-inference-service] elif isinstance(timeout, TimeoutSauce): [e2e-llm-inference-service] pass [e2e-llm-inference-service] else: [e2e-llm-inference-service] timeout = TimeoutSauce(connect=timeout, read=timeout) [e2e-llm-inference-service] [e2e-llm-inference-service] try: [e2e-llm-inference-service] > resp = conn.urlopen( [e2e-llm-inference-service] method=request.method, [e2e-llm-inference-service] url=url, [e2e-llm-inference-service] body=request.body, [e2e-llm-inference-service] headers=request.headers, [e2e-llm-inference-service] redirect=False, [e2e-llm-inference-service] assert_same_host=False, [e2e-llm-inference-service] preload_content=False, [e2e-llm-inference-service] decode_content=False, [e2e-llm-inference-service] retries=self.max_retries, [e2e-llm-inference-service] timeout=timeout, [e2e-llm-inference-service] chunked=chunked, [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] ../../python/kserve/.venv/lib64/python3.11/site-packages/requests/adapters.py:667: [e2e-llm-inference-service] _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ [e2e-llm-inference-service] [e2e-llm-inference-service] self = [e2e-llm-inference-service] method = 'POST', url = '/v1/completions' [e2e-llm-inference-service] body = b'{"model": "publishers/kserve-ci-e2e-test/models/lora-adapter-1", "prompt": "KServe is a", "max_tokens": 20}' [e2e-llm-inference-service] headers = {'User-Agent': 'python-requests/2.32.3', 'Accept-Encoding': 'gzip, deflate', 'Accept': '*/*', 'Connection': 'keep-aliv...lication/json', 'X-Gateway-Model-Name': 'publishers/kserve-ci-e2e-test/models/lora-adapter-1', 'Content-Length': '107'} [e2e-llm-inference-service] retries = Retry(total=7, connect=None, read=None, redirect=None, status=None) [e2e-llm-inference-service] redirect = False, assert_same_host = False [e2e-llm-inference-service] timeout = Timeout(connect=60, read=60, total=None), pool_timeout = None [e2e-llm-inference-service] release_conn = False, chunked = False, body_pos = None, preload_content = False [e2e-llm-inference-service] decode_content = False, response_kw = {} [e2e-llm-inference-service] parsed_url = Url(scheme=None, auth=None, host=None, port=None, path='/v1/completions', query=None, fragment=None) [e2e-llm-inference-service] destination_scheme = None, conn = None, release_this_conn = True [e2e-llm-inference-service] http_tunnel_required = False [e2e-llm-inference-service] err = ReadTimeoutError("HTTPConnectionPool(host='ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com', port=80): Read timed out. (read timeout=60)") [e2e-llm-inference-service] clean_exit = False [e2e-llm-inference-service] [e2e-llm-inference-service] def urlopen( # type: ignore[override] [e2e-llm-inference-service] self, [e2e-llm-inference-service] method: str, [e2e-llm-inference-service] url: str, [e2e-llm-inference-service] body: _TYPE_BODY | None = None, [e2e-llm-inference-service] headers: typing.Mapping[str, str] | None = None, [e2e-llm-inference-service] retries: Retry | bool | int | None = None, [e2e-llm-inference-service] redirect: bool = True, [e2e-llm-inference-service] assert_same_host: bool = True, [e2e-llm-inference-service] timeout: _TYPE_TIMEOUT = _DEFAULT_TIMEOUT, [e2e-llm-inference-service] pool_timeout: int | None = None, [e2e-llm-inference-service] release_conn: bool | None = None, [e2e-llm-inference-service] chunked: bool = False, [e2e-llm-inference-service] body_pos: _TYPE_BODY_POSITION | None = None, [e2e-llm-inference-service] preload_content: bool = True, [e2e-llm-inference-service] decode_content: bool = True, [e2e-llm-inference-service] **response_kw: typing.Any, [e2e-llm-inference-service] ) -> BaseHTTPResponse: [e2e-llm-inference-service] """ [e2e-llm-inference-service] Get a connection from the pool and perform an HTTP request. This is the [e2e-llm-inference-service] lowest level call for making a request, so you'll need to specify all [e2e-llm-inference-service] the raw details. [e2e-llm-inference-service] [e2e-llm-inference-service] .. note:: [e2e-llm-inference-service] [e2e-llm-inference-service] More commonly, it's appropriate to use a convenience method [e2e-llm-inference-service] such as :meth:`request`. [e2e-llm-inference-service] [e2e-llm-inference-service] .. note:: [e2e-llm-inference-service] [e2e-llm-inference-service] `release_conn` will only behave as expected if [e2e-llm-inference-service] `preload_content=False` because we want to make [e2e-llm-inference-service] `preload_content=False` the default behaviour someday soon without [e2e-llm-inference-service] breaking backwards compatibility. [e2e-llm-inference-service] [e2e-llm-inference-service] :param method: [e2e-llm-inference-service] HTTP request method (such as GET, POST, PUT, etc.) [e2e-llm-inference-service] [e2e-llm-inference-service] :param url: [e2e-llm-inference-service] The URL to perform the request on. [e2e-llm-inference-service] [e2e-llm-inference-service] :param body: [e2e-llm-inference-service] Data to send in the request body, either :class:`str`, :class:`bytes`, [e2e-llm-inference-service] an iterable of :class:`str`/:class:`bytes`, or a file-like object. [e2e-llm-inference-service] [e2e-llm-inference-service] :param headers: [e2e-llm-inference-service] Dictionary of custom headers to send, such as User-Agent, [e2e-llm-inference-service] If-None-Match, etc. If None, pool headers are used. If provided, [e2e-llm-inference-service] these headers completely replace any pool-specific headers. [e2e-llm-inference-service] [e2e-llm-inference-service] :param retries: [e2e-llm-inference-service] Configure the number of retries to allow before raising a [e2e-llm-inference-service] :class:`~urllib3.exceptions.MaxRetryError` exception. [e2e-llm-inference-service] [e2e-llm-inference-service] If ``None`` (default) will retry 3 times, see ``Retry.DEFAULT``. Pass a [e2e-llm-inference-service] :class:`~urllib3.util.retry.Retry` object for fine-grained control [e2e-llm-inference-service] over different types of retries. [e2e-llm-inference-service] Pass an integer number to retry connection errors that many times, [e2e-llm-inference-service] but no other types of errors. Pass zero to never retry. [e2e-llm-inference-service] [e2e-llm-inference-service] If ``False``, then retries are disabled and any exception is raised [e2e-llm-inference-service] immediately. Also, instead of raising a MaxRetryError on redirects, [e2e-llm-inference-service] the redirect response will be returned. [e2e-llm-inference-service] [e2e-llm-inference-service] :type retries: :class:`~urllib3.util.retry.Retry`, False, or an int. [e2e-llm-inference-service] [e2e-llm-inference-service] :param redirect: [e2e-llm-inference-service] If True, automatically handle redirects (status codes 301, 302, [e2e-llm-inference-service] 303, 307, 308). Each redirect counts as a retry. Disabling retries [e2e-llm-inference-service] will disable redirect, too. [e2e-llm-inference-service] [e2e-llm-inference-service] :param assert_same_host: [e2e-llm-inference-service] If ``True``, will make sure that the host of the pool requests is [e2e-llm-inference-service] consistent else will raise HostChangedError. When ``False``, you can [e2e-llm-inference-service] use the pool on an HTTP proxy and request foreign hosts. [e2e-llm-inference-service] [e2e-llm-inference-service] :param timeout: [e2e-llm-inference-service] If specified, overrides the default timeout for this one [e2e-llm-inference-service] request. It may be a float (in seconds) or an instance of [e2e-llm-inference-service] :class:`urllib3.util.Timeout`. [e2e-llm-inference-service] [e2e-llm-inference-service] :param pool_timeout: [e2e-llm-inference-service] If set and the pool is set to block=True, then this method will [e2e-llm-inference-service] block for ``pool_timeout`` seconds and raise EmptyPoolError if no [e2e-llm-inference-service] connection is available within the time period. [e2e-llm-inference-service] [e2e-llm-inference-service] :param bool preload_content: [e2e-llm-inference-service] If True, the response's body will be preloaded into memory. [e2e-llm-inference-service] [e2e-llm-inference-service] :param bool decode_content: [e2e-llm-inference-service] If True, will attempt to decode the body based on the [e2e-llm-inference-service] 'content-encoding' header. [e2e-llm-inference-service] [e2e-llm-inference-service] :param release_conn: [e2e-llm-inference-service] If False, then the urlopen call will not release the connection [e2e-llm-inference-service] back into the pool once a response is received (but will release if [e2e-llm-inference-service] you read the entire contents of the response such as when [e2e-llm-inference-service] `preload_content=True`). This is useful if you're not preloading [e2e-llm-inference-service] the response's content immediately. You will need to call [e2e-llm-inference-service] ``r.release_conn()`` on the response ``r`` to return the connection [e2e-llm-inference-service] back into the pool. If None, it takes the value of ``preload_content`` [e2e-llm-inference-service] which defaults to ``True``. [e2e-llm-inference-service] [e2e-llm-inference-service] :param bool chunked: [e2e-llm-inference-service] If True, urllib3 will send the body using chunked transfer [e2e-llm-inference-service] encoding. Otherwise, urllib3 will send the body using the standard [e2e-llm-inference-service] content-length form. Defaults to False. [e2e-llm-inference-service] [e2e-llm-inference-service] :param int body_pos: [e2e-llm-inference-service] Position to seek to in file-like body in the event of a retry or [e2e-llm-inference-service] redirect. Typically this won't need to be set because urllib3 will [e2e-llm-inference-service] auto-populate the value when needed. [e2e-llm-inference-service] """ [e2e-llm-inference-service] parsed_url = parse_url(url) [e2e-llm-inference-service] destination_scheme = parsed_url.scheme [e2e-llm-inference-service] [e2e-llm-inference-service] if headers is None: [e2e-llm-inference-service] headers = self.headers [e2e-llm-inference-service] [e2e-llm-inference-service] if not isinstance(retries, Retry): [e2e-llm-inference-service] retries = Retry.from_int(retries, redirect=redirect, default=self.retries) [e2e-llm-inference-service] [e2e-llm-inference-service] if release_conn is None: [e2e-llm-inference-service] release_conn = preload_content [e2e-llm-inference-service] [e2e-llm-inference-service] # Check host [e2e-llm-inference-service] if assert_same_host and not self.is_same_host(url): [e2e-llm-inference-service] raise HostChangedError(self, url, retries) [e2e-llm-inference-service] [e2e-llm-inference-service] # Ensure that the URL we're connecting to is properly encoded [e2e-llm-inference-service] if url.startswith("/"): [e2e-llm-inference-service] url = to_str(_encode_target(url)) [e2e-llm-inference-service] else: [e2e-llm-inference-service] url = to_str(parsed_url.url) [e2e-llm-inference-service] [e2e-llm-inference-service] conn = None [e2e-llm-inference-service] [e2e-llm-inference-service] # Track whether `conn` needs to be released before [e2e-llm-inference-service] # returning/raising/recursing. Update this variable if necessary, and [e2e-llm-inference-service] # leave `release_conn` constant throughout the function. That way, if [e2e-llm-inference-service] # the function recurses, the original value of `release_conn` will be [e2e-llm-inference-service] # passed down into the recursive call, and its value will be respected. [e2e-llm-inference-service] # [e2e-llm-inference-service] # See issue #651 [1] for details. [e2e-llm-inference-service] # [e2e-llm-inference-service] # [1] [e2e-llm-inference-service] release_this_conn = release_conn [e2e-llm-inference-service] [e2e-llm-inference-service] http_tunnel_required = connection_requires_http_tunnel( [e2e-llm-inference-service] self.proxy, self.proxy_config, destination_scheme [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] # Merge the proxy headers. Only done when not using HTTP CONNECT. We [e2e-llm-inference-service] # have to copy the headers dict so we can safely change it without those [e2e-llm-inference-service] # changes being reflected in anyone else's copy. [e2e-llm-inference-service] if not http_tunnel_required: [e2e-llm-inference-service] headers = headers.copy() # type: ignore[attr-defined] [e2e-llm-inference-service] headers.update(self.proxy_headers) # type: ignore[union-attr] [e2e-llm-inference-service] [e2e-llm-inference-service] # Must keep the exception bound to a separate variable or else Python 3 [e2e-llm-inference-service] # complains about UnboundLocalError. [e2e-llm-inference-service] err = None [e2e-llm-inference-service] [e2e-llm-inference-service] # Keep track of whether we cleanly exited the except block. This [e2e-llm-inference-service] # ensures we do proper cleanup in finally. [e2e-llm-inference-service] clean_exit = False [e2e-llm-inference-service] [e2e-llm-inference-service] # Rewind body position, if needed. Record current position [e2e-llm-inference-service] # for future rewinds in the event of a redirect/retry. [e2e-llm-inference-service] body_pos = set_file_position(body, body_pos) [e2e-llm-inference-service] [e2e-llm-inference-service] try: [e2e-llm-inference-service] # Request a connection from the queue. [e2e-llm-inference-service] timeout_obj = self._get_timeout(timeout) [e2e-llm-inference-service] conn = self._get_conn(timeout=pool_timeout) [e2e-llm-inference-service] [e2e-llm-inference-service] conn.timeout = timeout_obj.connect_timeout # type: ignore[assignment] [e2e-llm-inference-service] [e2e-llm-inference-service] # Is this a closed/new connection that requires CONNECT tunnelling? [e2e-llm-inference-service] if self.proxy is not None and http_tunnel_required and conn.is_closed: [e2e-llm-inference-service] try: [e2e-llm-inference-service] self._prepare_proxy(conn) [e2e-llm-inference-service] except (BaseSSLError, OSError, SocketTimeout) as e: [e2e-llm-inference-service] self._raise_timeout( [e2e-llm-inference-service] err=e, url=self.proxy.url, timeout_value=conn.timeout [e2e-llm-inference-service] ) [e2e-llm-inference-service] raise [e2e-llm-inference-service] [e2e-llm-inference-service] # If we're going to release the connection in ``finally:``, then [e2e-llm-inference-service] # the response doesn't need to know about the connection. Otherwise [e2e-llm-inference-service] # it will also try to release it and we'll have a double-release [e2e-llm-inference-service] # mess. [e2e-llm-inference-service] response_conn = conn if not release_conn else None [e2e-llm-inference-service] [e2e-llm-inference-service] # Make the request on the HTTPConnection object [e2e-llm-inference-service] response = self._make_request( [e2e-llm-inference-service] conn, [e2e-llm-inference-service] method, [e2e-llm-inference-service] url, [e2e-llm-inference-service] timeout=timeout_obj, [e2e-llm-inference-service] body=body, [e2e-llm-inference-service] headers=headers, [e2e-llm-inference-service] chunked=chunked, [e2e-llm-inference-service] retries=retries, [e2e-llm-inference-service] response_conn=response_conn, [e2e-llm-inference-service] preload_content=preload_content, [e2e-llm-inference-service] decode_content=decode_content, [e2e-llm-inference-service] **response_kw, [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] # Everything went great! [e2e-llm-inference-service] clean_exit = True [e2e-llm-inference-service] [e2e-llm-inference-service] except EmptyPoolError: [e2e-llm-inference-service] # Didn't get a connection from the pool, no need to clean up [e2e-llm-inference-service] clean_exit = True [e2e-llm-inference-service] release_this_conn = False [e2e-llm-inference-service] raise [e2e-llm-inference-service] [e2e-llm-inference-service] except ( [e2e-llm-inference-service] TimeoutError, [e2e-llm-inference-service] HTTPException, [e2e-llm-inference-service] OSError, [e2e-llm-inference-service] ProtocolError, [e2e-llm-inference-service] BaseSSLError, [e2e-llm-inference-service] SSLError, [e2e-llm-inference-service] CertificateError, [e2e-llm-inference-service] ProxyError, [e2e-llm-inference-service] ) as e: [e2e-llm-inference-service] # Discard the connection for these exceptions. It will be [e2e-llm-inference-service] # replaced during the next _get_conn() call. [e2e-llm-inference-service] clean_exit = False [e2e-llm-inference-service] new_e: Exception = e [e2e-llm-inference-service] if isinstance(e, (BaseSSLError, CertificateError)): [e2e-llm-inference-service] new_e = SSLError(e) [e2e-llm-inference-service] if isinstance( [e2e-llm-inference-service] new_e, [e2e-llm-inference-service] ( [e2e-llm-inference-service] OSError, [e2e-llm-inference-service] NewConnectionError, [e2e-llm-inference-service] TimeoutError, [e2e-llm-inference-service] SSLError, [e2e-llm-inference-service] HTTPException, [e2e-llm-inference-service] ), [e2e-llm-inference-service] ) and (conn and conn.proxy and not conn.has_connected_to_proxy): [e2e-llm-inference-service] new_e = _wrap_proxy_error(new_e, conn.proxy.scheme) [e2e-llm-inference-service] elif isinstance(new_e, (OSError, HTTPException)): [e2e-llm-inference-service] new_e = ProtocolError("Connection aborted.", new_e) [e2e-llm-inference-service] [e2e-llm-inference-service] retries = retries.increment( [e2e-llm-inference-service] method, url, error=new_e, _pool=self, _stacktrace=sys.exc_info()[2] [e2e-llm-inference-service] ) [e2e-llm-inference-service] retries.sleep() [e2e-llm-inference-service] [e2e-llm-inference-service] # Keep track of the error for the retry warning. [e2e-llm-inference-service] err = e [e2e-llm-inference-service] [e2e-llm-inference-service] finally: [e2e-llm-inference-service] if not clean_exit: [e2e-llm-inference-service] # We hit some kind of exception, handled or otherwise. We need [e2e-llm-inference-service] # to throw the connection away unless explicitly told not to. [e2e-llm-inference-service] # Close the connection, set the variable to None, and make sure [e2e-llm-inference-service] # we put the None back in the pool to avoid leaking it. [e2e-llm-inference-service] if conn: [e2e-llm-inference-service] conn.close() [e2e-llm-inference-service] conn = None [e2e-llm-inference-service] release_this_conn = True [e2e-llm-inference-service] [e2e-llm-inference-service] if release_this_conn: [e2e-llm-inference-service] # Put the connection back to be reused. If the connection is [e2e-llm-inference-service] # expired then it will be None, which will get replaced with a [e2e-llm-inference-service] # fresh connection during _get_conn. [e2e-llm-inference-service] self._put_conn(conn) [e2e-llm-inference-service] [e2e-llm-inference-service] if not conn: [e2e-llm-inference-service] # Try again [e2e-llm-inference-service] log.warning( [e2e-llm-inference-service] "Retrying (%r) after connection broken by '%r': %s", retries, err, url [e2e-llm-inference-service] ) [e2e-llm-inference-service] > return self.urlopen( [e2e-llm-inference-service] method, [e2e-llm-inference-service] url, [e2e-llm-inference-service] body, [e2e-llm-inference-service] headers, [e2e-llm-inference-service] retries, [e2e-llm-inference-service] redirect, [e2e-llm-inference-service] assert_same_host, [e2e-llm-inference-service] timeout=timeout, [e2e-llm-inference-service] pool_timeout=pool_timeout, [e2e-llm-inference-service] release_conn=release_conn, [e2e-llm-inference-service] chunked=chunked, [e2e-llm-inference-service] body_pos=body_pos, [e2e-llm-inference-service] preload_content=preload_content, [e2e-llm-inference-service] decode_content=decode_content, [e2e-llm-inference-service] **response_kw, [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] ../../python/kserve/.venv/lib64/python3.11/site-packages/urllib3/connectionpool.py:871: [e2e-llm-inference-service] _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ [e2e-llm-inference-service] [e2e-llm-inference-service] self = [e2e-llm-inference-service] method = 'POST', url = '/v1/completions' [e2e-llm-inference-service] body = b'{"model": "publishers/kserve-ci-e2e-test/models/lora-adapter-1", "prompt": "KServe is a", "max_tokens": 20}' [e2e-llm-inference-service] headers = {'User-Agent': 'python-requests/2.32.3', 'Accept-Encoding': 'gzip, deflate', 'Accept': '*/*', 'Connection': 'keep-aliv...lication/json', 'X-Gateway-Model-Name': 'publishers/kserve-ci-e2e-test/models/lora-adapter-1', 'Content-Length': '107'} [e2e-llm-inference-service] retries = Retry(total=6, connect=None, read=None, redirect=None, status=None) [e2e-llm-inference-service] redirect = False, assert_same_host = False [e2e-llm-inference-service] timeout = Timeout(connect=60, read=60, total=None), pool_timeout = None [e2e-llm-inference-service] release_conn = False, chunked = False, body_pos = None, preload_content = False [e2e-llm-inference-service] decode_content = False, response_kw = {} [e2e-llm-inference-service] parsed_url = Url(scheme=None, auth=None, host=None, port=None, path='/v1/completions', query=None, fragment=None) [e2e-llm-inference-service] destination_scheme = None, conn = None, release_this_conn = True [e2e-llm-inference-service] http_tunnel_required = False [e2e-llm-inference-service] err = ReadTimeoutError("HTTPConnectionPool(host='ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com', port=80): Read timed out. (read timeout=60)") [e2e-llm-inference-service] clean_exit = False [e2e-llm-inference-service] [e2e-llm-inference-service] def urlopen( # type: ignore[override] [e2e-llm-inference-service] self, [e2e-llm-inference-service] method: str, [e2e-llm-inference-service] url: str, [e2e-llm-inference-service] body: _TYPE_BODY | None = None, [e2e-llm-inference-service] headers: typing.Mapping[str, str] | None = None, [e2e-llm-inference-service] retries: Retry | bool | int | None = None, [e2e-llm-inference-service] redirect: bool = True, [e2e-llm-inference-service] assert_same_host: bool = True, [e2e-llm-inference-service] timeout: _TYPE_TIMEOUT = _DEFAULT_TIMEOUT, [e2e-llm-inference-service] pool_timeout: int | None = None, [e2e-llm-inference-service] release_conn: bool | None = None, [e2e-llm-inference-service] chunked: bool = False, [e2e-llm-inference-service] body_pos: _TYPE_BODY_POSITION | None = None, [e2e-llm-inference-service] preload_content: bool = True, [e2e-llm-inference-service] decode_content: bool = True, [e2e-llm-inference-service] **response_kw: typing.Any, [e2e-llm-inference-service] ) -> BaseHTTPResponse: [e2e-llm-inference-service] """ [e2e-llm-inference-service] Get a connection from the pool and perform an HTTP request. This is the [e2e-llm-inference-service] lowest level call for making a request, so you'll need to specify all [e2e-llm-inference-service] the raw details. [e2e-llm-inference-service] [e2e-llm-inference-service] .. note:: [e2e-llm-inference-service] [e2e-llm-inference-service] More commonly, it's appropriate to use a convenience method [e2e-llm-inference-service] such as :meth:`request`. [e2e-llm-inference-service] [e2e-llm-inference-service] .. note:: [e2e-llm-inference-service] [e2e-llm-inference-service] `release_conn` will only behave as expected if [e2e-llm-inference-service] `preload_content=False` because we want to make [e2e-llm-inference-service] `preload_content=False` the default behaviour someday soon without [e2e-llm-inference-service] breaking backwards compatibility. [e2e-llm-inference-service] [e2e-llm-inference-service] :param method: [e2e-llm-inference-service] HTTP request method (such as GET, POST, PUT, etc.) [e2e-llm-inference-service] [e2e-llm-inference-service] :param url: [e2e-llm-inference-service] The URL to perform the request on. [e2e-llm-inference-service] [e2e-llm-inference-service] :param body: [e2e-llm-inference-service] Data to send in the request body, either :class:`str`, :class:`bytes`, [e2e-llm-inference-service] an iterable of :class:`str`/:class:`bytes`, or a file-like object. [e2e-llm-inference-service] [e2e-llm-inference-service] :param headers: [e2e-llm-inference-service] Dictionary of custom headers to send, such as User-Agent, [e2e-llm-inference-service] If-None-Match, etc. If None, pool headers are used. If provided, [e2e-llm-inference-service] these headers completely replace any pool-specific headers. [e2e-llm-inference-service] [e2e-llm-inference-service] :param retries: [e2e-llm-inference-service] Configure the number of retries to allow before raising a [e2e-llm-inference-service] :class:`~urllib3.exceptions.MaxRetryError` exception. [e2e-llm-inference-service] [e2e-llm-inference-service] If ``None`` (default) will retry 3 times, see ``Retry.DEFAULT``. Pass a [e2e-llm-inference-service] :class:`~urllib3.util.retry.Retry` object for fine-grained control [e2e-llm-inference-service] over different types of retries. [e2e-llm-inference-service] Pass an integer number to retry connection errors that many times, [e2e-llm-inference-service] but no other types of errors. Pass zero to never retry. [e2e-llm-inference-service] [e2e-llm-inference-service] If ``False``, then retries are disabled and any exception is raised [e2e-llm-inference-service] immediately. Also, instead of raising a MaxRetryError on redirects, [e2e-llm-inference-service] the redirect response will be returned. [e2e-llm-inference-service] [e2e-llm-inference-service] :type retries: :class:`~urllib3.util.retry.Retry`, False, or an int. [e2e-llm-inference-service] [e2e-llm-inference-service] :param redirect: [e2e-llm-inference-service] If True, automatically handle redirects (status codes 301, 302, [e2e-llm-inference-service] 303, 307, 308). Each redirect counts as a retry. Disabling retries [e2e-llm-inference-service] will disable redirect, too. [e2e-llm-inference-service] [e2e-llm-inference-service] :param assert_same_host: [e2e-llm-inference-service] If ``True``, will make sure that the host of the pool requests is [e2e-llm-inference-service] consistent else will raise HostChangedError. When ``False``, you can [e2e-llm-inference-service] use the pool on an HTTP proxy and request foreign hosts. [e2e-llm-inference-service] [e2e-llm-inference-service] :param timeout: [e2e-llm-inference-service] If specified, overrides the default timeout for this one [e2e-llm-inference-service] request. It may be a float (in seconds) or an instance of [e2e-llm-inference-service] :class:`urllib3.util.Timeout`. [e2e-llm-inference-service] [e2e-llm-inference-service] :param pool_timeout: [e2e-llm-inference-service] If set and the pool is set to block=True, then this method will [e2e-llm-inference-service] block for ``pool_timeout`` seconds and raise EmptyPoolError if no [e2e-llm-inference-service] connection is available within the time period. [e2e-llm-inference-service] [e2e-llm-inference-service] :param bool preload_content: [e2e-llm-inference-service] If True, the response's body will be preloaded into memory. [e2e-llm-inference-service] [e2e-llm-inference-service] :param bool decode_content: [e2e-llm-inference-service] If True, will attempt to decode the body based on the [e2e-llm-inference-service] 'content-encoding' header. [e2e-llm-inference-service] [e2e-llm-inference-service] :param release_conn: [e2e-llm-inference-service] If False, then the urlopen call will not release the connection [e2e-llm-inference-service] back into the pool once a response is received (but will release if [e2e-llm-inference-service] you read the entire contents of the response such as when [e2e-llm-inference-service] `preload_content=True`). This is useful if you're not preloading [e2e-llm-inference-service] the response's content immediately. You will need to call [e2e-llm-inference-service] ``r.release_conn()`` on the response ``r`` to return the connection [e2e-llm-inference-service] back into the pool. If None, it takes the value of ``preload_content`` [e2e-llm-inference-service] which defaults to ``True``. [e2e-llm-inference-service] [e2e-llm-inference-service] :param bool chunked: [e2e-llm-inference-service] If True, urllib3 will send the body using chunked transfer [e2e-llm-inference-service] encoding. Otherwise, urllib3 will send the body using the standard [e2e-llm-inference-service] content-length form. Defaults to False. [e2e-llm-inference-service] [e2e-llm-inference-service] :param int body_pos: [e2e-llm-inference-service] Position to seek to in file-like body in the event of a retry or [e2e-llm-inference-service] redirect. Typically this won't need to be set because urllib3 will [e2e-llm-inference-service] auto-populate the value when needed. [e2e-llm-inference-service] """ [e2e-llm-inference-service] parsed_url = parse_url(url) [e2e-llm-inference-service] destination_scheme = parsed_url.scheme [e2e-llm-inference-service] [e2e-llm-inference-service] if headers is None: [e2e-llm-inference-service] headers = self.headers [e2e-llm-inference-service] [e2e-llm-inference-service] if not isinstance(retries, Retry): [e2e-llm-inference-service] retries = Retry.from_int(retries, redirect=redirect, default=self.retries) [e2e-llm-inference-service] [e2e-llm-inference-service] if release_conn is None: [e2e-llm-inference-service] release_conn = preload_content [e2e-llm-inference-service] [e2e-llm-inference-service] # Check host [e2e-llm-inference-service] if assert_same_host and not self.is_same_host(url): [e2e-llm-inference-service] raise HostChangedError(self, url, retries) [e2e-llm-inference-service] [e2e-llm-inference-service] # Ensure that the URL we're connecting to is properly encoded [e2e-llm-inference-service] if url.startswith("/"): [e2e-llm-inference-service] url = to_str(_encode_target(url)) [e2e-llm-inference-service] else: [e2e-llm-inference-service] url = to_str(parsed_url.url) [e2e-llm-inference-service] [e2e-llm-inference-service] conn = None [e2e-llm-inference-service] [e2e-llm-inference-service] # Track whether `conn` needs to be released before [e2e-llm-inference-service] # returning/raising/recursing. Update this variable if necessary, and [e2e-llm-inference-service] # leave `release_conn` constant throughout the function. That way, if [e2e-llm-inference-service] # the function recurses, the original value of `release_conn` will be [e2e-llm-inference-service] # passed down into the recursive call, and its value will be respected. [e2e-llm-inference-service] # [e2e-llm-inference-service] # See issue #651 [1] for details. [e2e-llm-inference-service] # [e2e-llm-inference-service] # [1] [e2e-llm-inference-service] release_this_conn = release_conn [e2e-llm-inference-service] [e2e-llm-inference-service] http_tunnel_required = connection_requires_http_tunnel( [e2e-llm-inference-service] self.proxy, self.proxy_config, destination_scheme [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] # Merge the proxy headers. Only done when not using HTTP CONNECT. We [e2e-llm-inference-service] # have to copy the headers dict so we can safely change it without those [e2e-llm-inference-service] # changes being reflected in anyone else's copy. [e2e-llm-inference-service] if not http_tunnel_required: [e2e-llm-inference-service] headers = headers.copy() # type: ignore[attr-defined] [e2e-llm-inference-service] headers.update(self.proxy_headers) # type: ignore[union-attr] [e2e-llm-inference-service] [e2e-llm-inference-service] # Must keep the exception bound to a separate variable or else Python 3 [e2e-llm-inference-service] # complains about UnboundLocalError. [e2e-llm-inference-service] err = None [e2e-llm-inference-service] [e2e-llm-inference-service] # Keep track of whether we cleanly exited the except block. This [e2e-llm-inference-service] # ensures we do proper cleanup in finally. [e2e-llm-inference-service] clean_exit = False [e2e-llm-inference-service] [e2e-llm-inference-service] # Rewind body position, if needed. Record current position [e2e-llm-inference-service] # for future rewinds in the event of a redirect/retry. [e2e-llm-inference-service] body_pos = set_file_position(body, body_pos) [e2e-llm-inference-service] [e2e-llm-inference-service] try: [e2e-llm-inference-service] # Request a connection from the queue. [e2e-llm-inference-service] timeout_obj = self._get_timeout(timeout) [e2e-llm-inference-service] conn = self._get_conn(timeout=pool_timeout) [e2e-llm-inference-service] [e2e-llm-inference-service] conn.timeout = timeout_obj.connect_timeout # type: ignore[assignment] [e2e-llm-inference-service] [e2e-llm-inference-service] # Is this a closed/new connection that requires CONNECT tunnelling? [e2e-llm-inference-service] if self.proxy is not None and http_tunnel_required and conn.is_closed: [e2e-llm-inference-service] try: [e2e-llm-inference-service] self._prepare_proxy(conn) [e2e-llm-inference-service] except (BaseSSLError, OSError, SocketTimeout) as e: [e2e-llm-inference-service] self._raise_timeout( [e2e-llm-inference-service] err=e, url=self.proxy.url, timeout_value=conn.timeout [e2e-llm-inference-service] ) [e2e-llm-inference-service] raise [e2e-llm-inference-service] [e2e-llm-inference-service] # If we're going to release the connection in ``finally:``, then [e2e-llm-inference-service] # the response doesn't need to know about the connection. Otherwise [e2e-llm-inference-service] # it will also try to release it and we'll have a double-release [e2e-llm-inference-service] # mess. [e2e-llm-inference-service] response_conn = conn if not release_conn else None [e2e-llm-inference-service] [e2e-llm-inference-service] # Make the request on the HTTPConnection object [e2e-llm-inference-service] response = self._make_request( [e2e-llm-inference-service] conn, [e2e-llm-inference-service] method, [e2e-llm-inference-service] url, [e2e-llm-inference-service] timeout=timeout_obj, [e2e-llm-inference-service] body=body, [e2e-llm-inference-service] headers=headers, [e2e-llm-inference-service] chunked=chunked, [e2e-llm-inference-service] retries=retries, [e2e-llm-inference-service] response_conn=response_conn, [e2e-llm-inference-service] preload_content=preload_content, [e2e-llm-inference-service] decode_content=decode_content, [e2e-llm-inference-service] **response_kw, [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] # Everything went great! [e2e-llm-inference-service] clean_exit = True [e2e-llm-inference-service] [e2e-llm-inference-service] except EmptyPoolError: [e2e-llm-inference-service] # Didn't get a connection from the pool, no need to clean up [e2e-llm-inference-service] clean_exit = True [e2e-llm-inference-service] release_this_conn = False [e2e-llm-inference-service] raise [e2e-llm-inference-service] [e2e-llm-inference-service] except ( [e2e-llm-inference-service] TimeoutError, [e2e-llm-inference-service] HTTPException, [e2e-llm-inference-service] OSError, [e2e-llm-inference-service] ProtocolError, [e2e-llm-inference-service] BaseSSLError, [e2e-llm-inference-service] SSLError, [e2e-llm-inference-service] CertificateError, [e2e-llm-inference-service] ProxyError, [e2e-llm-inference-service] ) as e: [e2e-llm-inference-service] # Discard the connection for these exceptions. It will be [e2e-llm-inference-service] # replaced during the next _get_conn() call. [e2e-llm-inference-service] clean_exit = False [e2e-llm-inference-service] new_e: Exception = e [e2e-llm-inference-service] if isinstance(e, (BaseSSLError, CertificateError)): [e2e-llm-inference-service] new_e = SSLError(e) [e2e-llm-inference-service] if isinstance( [e2e-llm-inference-service] new_e, [e2e-llm-inference-service] ( [e2e-llm-inference-service] OSError, [e2e-llm-inference-service] NewConnectionError, [e2e-llm-inference-service] TimeoutError, [e2e-llm-inference-service] SSLError, [e2e-llm-inference-service] HTTPException, [e2e-llm-inference-service] ), [e2e-llm-inference-service] ) and (conn and conn.proxy and not conn.has_connected_to_proxy): [e2e-llm-inference-service] new_e = _wrap_proxy_error(new_e, conn.proxy.scheme) [e2e-llm-inference-service] elif isinstance(new_e, (OSError, HTTPException)): [e2e-llm-inference-service] new_e = ProtocolError("Connection aborted.", new_e) [e2e-llm-inference-service] [e2e-llm-inference-service] retries = retries.increment( [e2e-llm-inference-service] method, url, error=new_e, _pool=self, _stacktrace=sys.exc_info()[2] [e2e-llm-inference-service] ) [e2e-llm-inference-service] retries.sleep() [e2e-llm-inference-service] [e2e-llm-inference-service] # Keep track of the error for the retry warning. [e2e-llm-inference-service] err = e [e2e-llm-inference-service] [e2e-llm-inference-service] finally: [e2e-llm-inference-service] if not clean_exit: [e2e-llm-inference-service] # We hit some kind of exception, handled or otherwise. We need [e2e-llm-inference-service] # to throw the connection away unless explicitly told not to. [e2e-llm-inference-service] # Close the connection, set the variable to None, and make sure [e2e-llm-inference-service] # we put the None back in the pool to avoid leaking it. [e2e-llm-inference-service] if conn: [e2e-llm-inference-service] conn.close() [e2e-llm-inference-service] conn = None [e2e-llm-inference-service] release_this_conn = True [e2e-llm-inference-service] [e2e-llm-inference-service] if release_this_conn: [e2e-llm-inference-service] # Put the connection back to be reused. If the connection is [e2e-llm-inference-service] # expired then it will be None, which will get replaced with a [e2e-llm-inference-service] # fresh connection during _get_conn. [e2e-llm-inference-service] self._put_conn(conn) [e2e-llm-inference-service] [e2e-llm-inference-service] if not conn: [e2e-llm-inference-service] # Try again [e2e-llm-inference-service] log.warning( [e2e-llm-inference-service] "Retrying (%r) after connection broken by '%r': %s", retries, err, url [e2e-llm-inference-service] ) [e2e-llm-inference-service] > return self.urlopen( [e2e-llm-inference-service] method, [e2e-llm-inference-service] url, [e2e-llm-inference-service] body, [e2e-llm-inference-service] headers, [e2e-llm-inference-service] retries, [e2e-llm-inference-service] redirect, [e2e-llm-inference-service] assert_same_host, [e2e-llm-inference-service] timeout=timeout, [e2e-llm-inference-service] pool_timeout=pool_timeout, [e2e-llm-inference-service] release_conn=release_conn, [e2e-llm-inference-service] chunked=chunked, [e2e-llm-inference-service] body_pos=body_pos, [e2e-llm-inference-service] preload_content=preload_content, [e2e-llm-inference-service] decode_content=decode_content, [e2e-llm-inference-service] **response_kw, [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] ../../python/kserve/.venv/lib64/python3.11/site-packages/urllib3/connectionpool.py:871: [e2e-llm-inference-service] _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ [e2e-llm-inference-service] [e2e-llm-inference-service] self = [e2e-llm-inference-service] method = 'POST', url = '/v1/completions' [e2e-llm-inference-service] body = b'{"model": "publishers/kserve-ci-e2e-test/models/lora-adapter-1", "prompt": "KServe is a", "max_tokens": 20}' [e2e-llm-inference-service] headers = {'User-Agent': 'python-requests/2.32.3', 'Accept-Encoding': 'gzip, deflate', 'Accept': '*/*', 'Connection': 'keep-aliv...lication/json', 'X-Gateway-Model-Name': 'publishers/kserve-ci-e2e-test/models/lora-adapter-1', 'Content-Length': '107'} [e2e-llm-inference-service] retries = Retry(total=5, connect=None, read=None, redirect=None, status=None) [e2e-llm-inference-service] redirect = False, assert_same_host = False [e2e-llm-inference-service] timeout = Timeout(connect=60, read=60, total=None), pool_timeout = None [e2e-llm-inference-service] release_conn = False, chunked = False, body_pos = None, preload_content = False [e2e-llm-inference-service] decode_content = False, response_kw = {} [e2e-llm-inference-service] parsed_url = Url(scheme=None, auth=None, host=None, port=None, path='/v1/completions', query=None, fragment=None) [e2e-llm-inference-service] destination_scheme = None, conn = None, release_this_conn = True [e2e-llm-inference-service] http_tunnel_required = False [e2e-llm-inference-service] err = ReadTimeoutError("HTTPConnectionPool(host='ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com', port=80): Read timed out. (read timeout=60)") [e2e-llm-inference-service] clean_exit = False [e2e-llm-inference-service] [e2e-llm-inference-service] def urlopen( # type: ignore[override] [e2e-llm-inference-service] self, [e2e-llm-inference-service] method: str, [e2e-llm-inference-service] url: str, [e2e-llm-inference-service] body: _TYPE_BODY | None = None, [e2e-llm-inference-service] headers: typing.Mapping[str, str] | None = None, [e2e-llm-inference-service] retries: Retry | bool | int | None = None, [e2e-llm-inference-service] redirect: bool = True, [e2e-llm-inference-service] assert_same_host: bool = True, [e2e-llm-inference-service] timeout: _TYPE_TIMEOUT = _DEFAULT_TIMEOUT, [e2e-llm-inference-service] pool_timeout: int | None = None, [e2e-llm-inference-service] release_conn: bool | None = None, [e2e-llm-inference-service] chunked: bool = False, [e2e-llm-inference-service] body_pos: _TYPE_BODY_POSITION | None = None, [e2e-llm-inference-service] preload_content: bool = True, [e2e-llm-inference-service] decode_content: bool = True, [e2e-llm-inference-service] **response_kw: typing.Any, [e2e-llm-inference-service] ) -> BaseHTTPResponse: [e2e-llm-inference-service] """ [e2e-llm-inference-service] Get a connection from the pool and perform an HTTP request. This is the [e2e-llm-inference-service] lowest level call for making a request, so you'll need to specify all [e2e-llm-inference-service] the raw details. [e2e-llm-inference-service] [e2e-llm-inference-service] .. note:: [e2e-llm-inference-service] [e2e-llm-inference-service] More commonly, it's appropriate to use a convenience method [e2e-llm-inference-service] such as :meth:`request`. [e2e-llm-inference-service] [e2e-llm-inference-service] .. note:: [e2e-llm-inference-service] [e2e-llm-inference-service] `release_conn` will only behave as expected if [e2e-llm-inference-service] `preload_content=False` because we want to make [e2e-llm-inference-service] `preload_content=False` the default behaviour someday soon without [e2e-llm-inference-service] breaking backwards compatibility. [e2e-llm-inference-service] [e2e-llm-inference-service] :param method: [e2e-llm-inference-service] HTTP request method (such as GET, POST, PUT, etc.) [e2e-llm-inference-service] [e2e-llm-inference-service] :param url: [e2e-llm-inference-service] The URL to perform the request on. [e2e-llm-inference-service] [e2e-llm-inference-service] :param body: [e2e-llm-inference-service] Data to send in the request body, either :class:`str`, :class:`bytes`, [e2e-llm-inference-service] an iterable of :class:`str`/:class:`bytes`, or a file-like object. [e2e-llm-inference-service] [e2e-llm-inference-service] :param headers: [e2e-llm-inference-service] Dictionary of custom headers to send, such as User-Agent, [e2e-llm-inference-service] If-None-Match, etc. If None, pool headers are used. If provided, [e2e-llm-inference-service] these headers completely replace any pool-specific headers. [e2e-llm-inference-service] [e2e-llm-inference-service] :param retries: [e2e-llm-inference-service] Configure the number of retries to allow before raising a [e2e-llm-inference-service] :class:`~urllib3.exceptions.MaxRetryError` exception. [e2e-llm-inference-service] [e2e-llm-inference-service] If ``None`` (default) will retry 3 times, see ``Retry.DEFAULT``. Pass a [e2e-llm-inference-service] :class:`~urllib3.util.retry.Retry` object for fine-grained control [e2e-llm-inference-service] over different types of retries. [e2e-llm-inference-service] Pass an integer number to retry connection errors that many times, [e2e-llm-inference-service] but no other types of errors. Pass zero to never retry. [e2e-llm-inference-service] [e2e-llm-inference-service] If ``False``, then retries are disabled and any exception is raised [e2e-llm-inference-service] immediately. Also, instead of raising a MaxRetryError on redirects, [e2e-llm-inference-service] the redirect response will be returned. [e2e-llm-inference-service] [e2e-llm-inference-service] :type retries: :class:`~urllib3.util.retry.Retry`, False, or an int. [e2e-llm-inference-service] [e2e-llm-inference-service] :param redirect: [e2e-llm-inference-service] If True, automatically handle redirects (status codes 301, 302, [e2e-llm-inference-service] 303, 307, 308). Each redirect counts as a retry. Disabling retries [e2e-llm-inference-service] will disable redirect, too. [e2e-llm-inference-service] [e2e-llm-inference-service] :param assert_same_host: [e2e-llm-inference-service] If ``True``, will make sure that the host of the pool requests is [e2e-llm-inference-service] consistent else will raise HostChangedError. When ``False``, you can [e2e-llm-inference-service] use the pool on an HTTP proxy and request foreign hosts. [e2e-llm-inference-service] [e2e-llm-inference-service] :param timeout: [e2e-llm-inference-service] If specified, overrides the default timeout for this one [e2e-llm-inference-service] request. It may be a float (in seconds) or an instance of [e2e-llm-inference-service] :class:`urllib3.util.Timeout`. [e2e-llm-inference-service] [e2e-llm-inference-service] :param pool_timeout: [e2e-llm-inference-service] If set and the pool is set to block=True, then this method will [e2e-llm-inference-service] block for ``pool_timeout`` seconds and raise EmptyPoolError if no [e2e-llm-inference-service] connection is available within the time period. [e2e-llm-inference-service] [e2e-llm-inference-service] :param bool preload_content: [e2e-llm-inference-service] If True, the response's body will be preloaded into memory. [e2e-llm-inference-service] [e2e-llm-inference-service] :param bool decode_content: [e2e-llm-inference-service] If True, will attempt to decode the body based on the [e2e-llm-inference-service] 'content-encoding' header. [e2e-llm-inference-service] [e2e-llm-inference-service] :param release_conn: [e2e-llm-inference-service] If False, then the urlopen call will not release the connection [e2e-llm-inference-service] back into the pool once a response is received (but will release if [e2e-llm-inference-service] you read the entire contents of the response such as when [e2e-llm-inference-service] `preload_content=True`). This is useful if you're not preloading [e2e-llm-inference-service] the response's content immediately. You will need to call [e2e-llm-inference-service] ``r.release_conn()`` on the response ``r`` to return the connection [e2e-llm-inference-service] back into the pool. If None, it takes the value of ``preload_content`` [e2e-llm-inference-service] which defaults to ``True``. [e2e-llm-inference-service] [e2e-llm-inference-service] :param bool chunked: [e2e-llm-inference-service] If True, urllib3 will send the body using chunked transfer [e2e-llm-inference-service] encoding. Otherwise, urllib3 will send the body using the standard [e2e-llm-inference-service] content-length form. Defaults to False. [e2e-llm-inference-service] [e2e-llm-inference-service] :param int body_pos: [e2e-llm-inference-service] Position to seek to in file-like body in the event of a retry or [e2e-llm-inference-service] redirect. Typically this won't need to be set because urllib3 will [e2e-llm-inference-service] auto-populate the value when needed. [e2e-llm-inference-service] """ [e2e-llm-inference-service] parsed_url = parse_url(url) [e2e-llm-inference-service] destination_scheme = parsed_url.scheme [e2e-llm-inference-service] [e2e-llm-inference-service] if headers is None: [e2e-llm-inference-service] headers = self.headers [e2e-llm-inference-service] [e2e-llm-inference-service] if not isinstance(retries, Retry): [e2e-llm-inference-service] retries = Retry.from_int(retries, redirect=redirect, default=self.retries) [e2e-llm-inference-service] [e2e-llm-inference-service] if release_conn is None: [e2e-llm-inference-service] release_conn = preload_content [e2e-llm-inference-service] [e2e-llm-inference-service] # Check host [e2e-llm-inference-service] if assert_same_host and not self.is_same_host(url): [e2e-llm-inference-service] raise HostChangedError(self, url, retries) [e2e-llm-inference-service] [e2e-llm-inference-service] # Ensure that the URL we're connecting to is properly encoded [e2e-llm-inference-service] if url.startswith("/"): [e2e-llm-inference-service] url = to_str(_encode_target(url)) [e2e-llm-inference-service] else: [e2e-llm-inference-service] url = to_str(parsed_url.url) [e2e-llm-inference-service] [e2e-llm-inference-service] conn = None [e2e-llm-inference-service] [e2e-llm-inference-service] # Track whether `conn` needs to be released before [e2e-llm-inference-service] # returning/raising/recursing. Update this variable if necessary, and [e2e-llm-inference-service] # leave `release_conn` constant throughout the function. That way, if [e2e-llm-inference-service] # the function recurses, the original value of `release_conn` will be [e2e-llm-inference-service] # passed down into the recursive call, and its value will be respected. [e2e-llm-inference-service] # [e2e-llm-inference-service] # See issue #651 [1] for details. [e2e-llm-inference-service] # [e2e-llm-inference-service] # [1] [e2e-llm-inference-service] release_this_conn = release_conn [e2e-llm-inference-service] [e2e-llm-inference-service] http_tunnel_required = connection_requires_http_tunnel( [e2e-llm-inference-service] self.proxy, self.proxy_config, destination_scheme [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] # Merge the proxy headers. Only done when not using HTTP CONNECT. We [e2e-llm-inference-service] # have to copy the headers dict so we can safely change it without those [e2e-llm-inference-service] # changes being reflected in anyone else's copy. [e2e-llm-inference-service] if not http_tunnel_required: [e2e-llm-inference-service] headers = headers.copy() # type: ignore[attr-defined] [e2e-llm-inference-service] headers.update(self.proxy_headers) # type: ignore[union-attr] [e2e-llm-inference-service] [e2e-llm-inference-service] # Must keep the exception bound to a separate variable or else Python 3 [e2e-llm-inference-service] # complains about UnboundLocalError. [e2e-llm-inference-service] err = None [e2e-llm-inference-service] [e2e-llm-inference-service] # Keep track of whether we cleanly exited the except block. This [e2e-llm-inference-service] # ensures we do proper cleanup in finally. [e2e-llm-inference-service] clean_exit = False [e2e-llm-inference-service] [e2e-llm-inference-service] # Rewind body position, if needed. Record current position [e2e-llm-inference-service] # for future rewinds in the event of a redirect/retry. [e2e-llm-inference-service] body_pos = set_file_position(body, body_pos) [e2e-llm-inference-service] [e2e-llm-inference-service] try: [e2e-llm-inference-service] # Request a connection from the queue. [e2e-llm-inference-service] timeout_obj = self._get_timeout(timeout) [e2e-llm-inference-service] conn = self._get_conn(timeout=pool_timeout) [e2e-llm-inference-service] [e2e-llm-inference-service] conn.timeout = timeout_obj.connect_timeout # type: ignore[assignment] [e2e-llm-inference-service] [e2e-llm-inference-service] # Is this a closed/new connection that requires CONNECT tunnelling? [e2e-llm-inference-service] if self.proxy is not None and http_tunnel_required and conn.is_closed: [e2e-llm-inference-service] try: [e2e-llm-inference-service] self._prepare_proxy(conn) [e2e-llm-inference-service] except (BaseSSLError, OSError, SocketTimeout) as e: [e2e-llm-inference-service] self._raise_timeout( [e2e-llm-inference-service] err=e, url=self.proxy.url, timeout_value=conn.timeout [e2e-llm-inference-service] ) [e2e-llm-inference-service] raise [e2e-llm-inference-service] [e2e-llm-inference-service] # If we're going to release the connection in ``finally:``, then [e2e-llm-inference-service] # the response doesn't need to know about the connection. Otherwise [e2e-llm-inference-service] # it will also try to release it and we'll have a double-release [e2e-llm-inference-service] # mess. [e2e-llm-inference-service] response_conn = conn if not release_conn else None [e2e-llm-inference-service] [e2e-llm-inference-service] # Make the request on the HTTPConnection object [e2e-llm-inference-service] response = self._make_request( [e2e-llm-inference-service] conn, [e2e-llm-inference-service] method, [e2e-llm-inference-service] url, [e2e-llm-inference-service] timeout=timeout_obj, [e2e-llm-inference-service] body=body, [e2e-llm-inference-service] headers=headers, [e2e-llm-inference-service] chunked=chunked, [e2e-llm-inference-service] retries=retries, [e2e-llm-inference-service] response_conn=response_conn, [e2e-llm-inference-service] preload_content=preload_content, [e2e-llm-inference-service] decode_content=decode_content, [e2e-llm-inference-service] **response_kw, [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] # Everything went great! [e2e-llm-inference-service] clean_exit = True [e2e-llm-inference-service] [e2e-llm-inference-service] except EmptyPoolError: [e2e-llm-inference-service] # Didn't get a connection from the pool, no need to clean up [e2e-llm-inference-service] clean_exit = True [e2e-llm-inference-service] release_this_conn = False [e2e-llm-inference-service] raise [e2e-llm-inference-service] [e2e-llm-inference-service] except ( [e2e-llm-inference-service] TimeoutError, [e2e-llm-inference-service] HTTPException, [e2e-llm-inference-service] OSError, [e2e-llm-inference-service] ProtocolError, [e2e-llm-inference-service] BaseSSLError, [e2e-llm-inference-service] SSLError, [e2e-llm-inference-service] CertificateError, [e2e-llm-inference-service] ProxyError, [e2e-llm-inference-service] ) as e: [e2e-llm-inference-service] # Discard the connection for these exceptions. It will be [e2e-llm-inference-service] # replaced during the next _get_conn() call. [e2e-llm-inference-service] clean_exit = False [e2e-llm-inference-service] new_e: Exception = e [e2e-llm-inference-service] if isinstance(e, (BaseSSLError, CertificateError)): [e2e-llm-inference-service] new_e = SSLError(e) [e2e-llm-inference-service] if isinstance( [e2e-llm-inference-service] new_e, [e2e-llm-inference-service] ( [e2e-llm-inference-service] OSError, [e2e-llm-inference-service] NewConnectionError, [e2e-llm-inference-service] TimeoutError, [e2e-llm-inference-service] SSLError, [e2e-llm-inference-service] HTTPException, [e2e-llm-inference-service] ), [e2e-llm-inference-service] ) and (conn and conn.proxy and not conn.has_connected_to_proxy): [e2e-llm-inference-service] new_e = _wrap_proxy_error(new_e, conn.proxy.scheme) [e2e-llm-inference-service] elif isinstance(new_e, (OSError, HTTPException)): [e2e-llm-inference-service] new_e = ProtocolError("Connection aborted.", new_e) [e2e-llm-inference-service] [e2e-llm-inference-service] retries = retries.increment( [e2e-llm-inference-service] method, url, error=new_e, _pool=self, _stacktrace=sys.exc_info()[2] [e2e-llm-inference-service] ) [e2e-llm-inference-service] retries.sleep() [e2e-llm-inference-service] [e2e-llm-inference-service] # Keep track of the error for the retry warning. [e2e-llm-inference-service] err = e [e2e-llm-inference-service] [e2e-llm-inference-service] finally: [e2e-llm-inference-service] if not clean_exit: [e2e-llm-inference-service] # We hit some kind of exception, handled or otherwise. We need [e2e-llm-inference-service] # to throw the connection away unless explicitly told not to. [e2e-llm-inference-service] # Close the connection, set the variable to None, and make sure [e2e-llm-inference-service] # we put the None back in the pool to avoid leaking it. [e2e-llm-inference-service] if conn: [e2e-llm-inference-service] conn.close() [e2e-llm-inference-service] conn = None [e2e-llm-inference-service] release_this_conn = True [e2e-llm-inference-service] [e2e-llm-inference-service] if release_this_conn: [e2e-llm-inference-service] # Put the connection back to be reused. If the connection is [e2e-llm-inference-service] # expired then it will be None, which will get replaced with a [e2e-llm-inference-service] # fresh connection during _get_conn. [e2e-llm-inference-service] self._put_conn(conn) [e2e-llm-inference-service] [e2e-llm-inference-service] if not conn: [e2e-llm-inference-service] # Try again [e2e-llm-inference-service] log.warning( [e2e-llm-inference-service] "Retrying (%r) after connection broken by '%r': %s", retries, err, url [e2e-llm-inference-service] ) [e2e-llm-inference-service] > return self.urlopen( [e2e-llm-inference-service] method, [e2e-llm-inference-service] url, [e2e-llm-inference-service] body, [e2e-llm-inference-service] headers, [e2e-llm-inference-service] retries, [e2e-llm-inference-service] redirect, [e2e-llm-inference-service] assert_same_host, [e2e-llm-inference-service] timeout=timeout, [e2e-llm-inference-service] pool_timeout=pool_timeout, [e2e-llm-inference-service] release_conn=release_conn, [e2e-llm-inference-service] chunked=chunked, [e2e-llm-inference-service] body_pos=body_pos, [e2e-llm-inference-service] preload_content=preload_content, [e2e-llm-inference-service] decode_content=decode_content, [e2e-llm-inference-service] **response_kw, [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] ../../python/kserve/.venv/lib64/python3.11/site-packages/urllib3/connectionpool.py:871: [e2e-llm-inference-service] _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ [e2e-llm-inference-service] [e2e-llm-inference-service] self = [e2e-llm-inference-service] method = 'POST', url = '/v1/completions' [e2e-llm-inference-service] body = b'{"model": "publishers/kserve-ci-e2e-test/models/lora-adapter-1", "prompt": "KServe is a", "max_tokens": 20}' [e2e-llm-inference-service] headers = {'User-Agent': 'python-requests/2.32.3', 'Accept-Encoding': 'gzip, deflate', 'Accept': '*/*', 'Connection': 'keep-aliv...lication/json', 'X-Gateway-Model-Name': 'publishers/kserve-ci-e2e-test/models/lora-adapter-1', 'Content-Length': '107'} [e2e-llm-inference-service] retries = Retry(total=4, connect=None, read=None, redirect=None, status=None) [e2e-llm-inference-service] redirect = False, assert_same_host = False [e2e-llm-inference-service] timeout = Timeout(connect=60, read=60, total=None), pool_timeout = None [e2e-llm-inference-service] release_conn = False, chunked = False, body_pos = None, preload_content = False [e2e-llm-inference-service] decode_content = False, response_kw = {} [e2e-llm-inference-service] parsed_url = Url(scheme=None, auth=None, host=None, port=None, path='/v1/completions', query=None, fragment=None) [e2e-llm-inference-service] destination_scheme = None, conn = None, release_this_conn = True [e2e-llm-inference-service] http_tunnel_required = False [e2e-llm-inference-service] err = ReadTimeoutError("HTTPConnectionPool(host='ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com', port=80): Read timed out. (read timeout=60)") [e2e-llm-inference-service] clean_exit = False [e2e-llm-inference-service] [e2e-llm-inference-service] def urlopen( # type: ignore[override] [e2e-llm-inference-service] self, [e2e-llm-inference-service] method: str, [e2e-llm-inference-service] url: str, [e2e-llm-inference-service] body: _TYPE_BODY | None = None, [e2e-llm-inference-service] headers: typing.Mapping[str, str] | None = None, [e2e-llm-inference-service] retries: Retry | bool | int | None = None, [e2e-llm-inference-service] redirect: bool = True, [e2e-llm-inference-service] assert_same_host: bool = True, [e2e-llm-inference-service] timeout: _TYPE_TIMEOUT = _DEFAULT_TIMEOUT, [e2e-llm-inference-service] pool_timeout: int | None = None, [e2e-llm-inference-service] release_conn: bool | None = None, [e2e-llm-inference-service] chunked: bool = False, [e2e-llm-inference-service] body_pos: _TYPE_BODY_POSITION | None = None, [e2e-llm-inference-service] preload_content: bool = True, [e2e-llm-inference-service] decode_content: bool = True, [e2e-llm-inference-service] **response_kw: typing.Any, [e2e-llm-inference-service] ) -> BaseHTTPResponse: [e2e-llm-inference-service] """ [e2e-llm-inference-service] Get a connection from the pool and perform an HTTP request. This is the [e2e-llm-inference-service] lowest level call for making a request, so you'll need to specify all [e2e-llm-inference-service] the raw details. [e2e-llm-inference-service] [e2e-llm-inference-service] .. note:: [e2e-llm-inference-service] [e2e-llm-inference-service] More commonly, it's appropriate to use a convenience method [e2e-llm-inference-service] such as :meth:`request`. [e2e-llm-inference-service] [e2e-llm-inference-service] .. note:: [e2e-llm-inference-service] [e2e-llm-inference-service] `release_conn` will only behave as expected if [e2e-llm-inference-service] `preload_content=False` because we want to make [e2e-llm-inference-service] `preload_content=False` the default behaviour someday soon without [e2e-llm-inference-service] breaking backwards compatibility. [e2e-llm-inference-service] [e2e-llm-inference-service] :param method: [e2e-llm-inference-service] HTTP request method (such as GET, POST, PUT, etc.) [e2e-llm-inference-service] [e2e-llm-inference-service] :param url: [e2e-llm-inference-service] The URL to perform the request on. [e2e-llm-inference-service] [e2e-llm-inference-service] :param body: [e2e-llm-inference-service] Data to send in the request body, either :class:`str`, :class:`bytes`, [e2e-llm-inference-service] an iterable of :class:`str`/:class:`bytes`, or a file-like object. [e2e-llm-inference-service] [e2e-llm-inference-service] :param headers: [e2e-llm-inference-service] Dictionary of custom headers to send, such as User-Agent, [e2e-llm-inference-service] If-None-Match, etc. If None, pool headers are used. If provided, [e2e-llm-inference-service] these headers completely replace any pool-specific headers. [e2e-llm-inference-service] [e2e-llm-inference-service] :param retries: [e2e-llm-inference-service] Configure the number of retries to allow before raising a [e2e-llm-inference-service] :class:`~urllib3.exceptions.MaxRetryError` exception. [e2e-llm-inference-service] [e2e-llm-inference-service] If ``None`` (default) will retry 3 times, see ``Retry.DEFAULT``. Pass a [e2e-llm-inference-service] :class:`~urllib3.util.retry.Retry` object for fine-grained control [e2e-llm-inference-service] over different types of retries. [e2e-llm-inference-service] Pass an integer number to retry connection errors that many times, [e2e-llm-inference-service] but no other types of errors. Pass zero to never retry. [e2e-llm-inference-service] [e2e-llm-inference-service] If ``False``, then retries are disabled and any exception is raised [e2e-llm-inference-service] immediately. Also, instead of raising a MaxRetryError on redirects, [e2e-llm-inference-service] the redirect response will be returned. [e2e-llm-inference-service] [e2e-llm-inference-service] :type retries: :class:`~urllib3.util.retry.Retry`, False, or an int. [e2e-llm-inference-service] [e2e-llm-inference-service] :param redirect: [e2e-llm-inference-service] If True, automatically handle redirects (status codes 301, 302, [e2e-llm-inference-service] 303, 307, 308). Each redirect counts as a retry. Disabling retries [e2e-llm-inference-service] will disable redirect, too. [e2e-llm-inference-service] [e2e-llm-inference-service] :param assert_same_host: [e2e-llm-inference-service] If ``True``, will make sure that the host of the pool requests is [e2e-llm-inference-service] consistent else will raise HostChangedError. When ``False``, you can [e2e-llm-inference-service] use the pool on an HTTP proxy and request foreign hosts. [e2e-llm-inference-service] [e2e-llm-inference-service] :param timeout: [e2e-llm-inference-service] If specified, overrides the default timeout for this one [e2e-llm-inference-service] request. It may be a float (in seconds) or an instance of [e2e-llm-inference-service] :class:`urllib3.util.Timeout`. [e2e-llm-inference-service] [e2e-llm-inference-service] :param pool_timeout: [e2e-llm-inference-service] If set and the pool is set to block=True, then this method will [e2e-llm-inference-service] block for ``pool_timeout`` seconds and raise EmptyPoolError if no [e2e-llm-inference-service] connection is available within the time period. [e2e-llm-inference-service] [e2e-llm-inference-service] :param bool preload_content: [e2e-llm-inference-service] If True, the response's body will be preloaded into memory. [e2e-llm-inference-service] [e2e-llm-inference-service] :param bool decode_content: [e2e-llm-inference-service] If True, will attempt to decode the body based on the [e2e-llm-inference-service] 'content-encoding' header. [e2e-llm-inference-service] [e2e-llm-inference-service] :param release_conn: [e2e-llm-inference-service] If False, then the urlopen call will not release the connection [e2e-llm-inference-service] back into the pool once a response is received (but will release if [e2e-llm-inference-service] you read the entire contents of the response such as when [e2e-llm-inference-service] `preload_content=True`). This is useful if you're not preloading [e2e-llm-inference-service] the response's content immediately. You will need to call [e2e-llm-inference-service] ``r.release_conn()`` on the response ``r`` to return the connection [e2e-llm-inference-service] back into the pool. If None, it takes the value of ``preload_content`` [e2e-llm-inference-service] which defaults to ``True``. [e2e-llm-inference-service] [e2e-llm-inference-service] :param bool chunked: [e2e-llm-inference-service] If True, urllib3 will send the body using chunked transfer [e2e-llm-inference-service] encoding. Otherwise, urllib3 will send the body using the standard [e2e-llm-inference-service] content-length form. Defaults to False. [e2e-llm-inference-service] [e2e-llm-inference-service] :param int body_pos: [e2e-llm-inference-service] Position to seek to in file-like body in the event of a retry or [e2e-llm-inference-service] redirect. Typically this won't need to be set because urllib3 will [e2e-llm-inference-service] auto-populate the value when needed. [e2e-llm-inference-service] """ [e2e-llm-inference-service] parsed_url = parse_url(url) [e2e-llm-inference-service] destination_scheme = parsed_url.scheme [e2e-llm-inference-service] [e2e-llm-inference-service] if headers is None: [e2e-llm-inference-service] headers = self.headers [e2e-llm-inference-service] [e2e-llm-inference-service] if not isinstance(retries, Retry): [e2e-llm-inference-service] retries = Retry.from_int(retries, redirect=redirect, default=self.retries) [e2e-llm-inference-service] [e2e-llm-inference-service] if release_conn is None: [e2e-llm-inference-service] release_conn = preload_content [e2e-llm-inference-service] [e2e-llm-inference-service] # Check host [e2e-llm-inference-service] if assert_same_host and not self.is_same_host(url): [e2e-llm-inference-service] raise HostChangedError(self, url, retries) [e2e-llm-inference-service] [e2e-llm-inference-service] # Ensure that the URL we're connecting to is properly encoded [e2e-llm-inference-service] if url.startswith("/"): [e2e-llm-inference-service] url = to_str(_encode_target(url)) [e2e-llm-inference-service] else: [e2e-llm-inference-service] url = to_str(parsed_url.url) [e2e-llm-inference-service] [e2e-llm-inference-service] conn = None [e2e-llm-inference-service] [e2e-llm-inference-service] # Track whether `conn` needs to be released before [e2e-llm-inference-service] # returning/raising/recursing. Update this variable if necessary, and [e2e-llm-inference-service] # leave `release_conn` constant throughout the function. That way, if [e2e-llm-inference-service] # the function recurses, the original value of `release_conn` will be [e2e-llm-inference-service] # passed down into the recursive call, and its value will be respected. [e2e-llm-inference-service] # [e2e-llm-inference-service] # See issue #651 [1] for details. [e2e-llm-inference-service] # [e2e-llm-inference-service] # [1] [e2e-llm-inference-service] release_this_conn = release_conn [e2e-llm-inference-service] [e2e-llm-inference-service] http_tunnel_required = connection_requires_http_tunnel( [e2e-llm-inference-service] self.proxy, self.proxy_config, destination_scheme [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] # Merge the proxy headers. Only done when not using HTTP CONNECT. We [e2e-llm-inference-service] # have to copy the headers dict so we can safely change it without those [e2e-llm-inference-service] # changes being reflected in anyone else's copy. [e2e-llm-inference-service] if not http_tunnel_required: [e2e-llm-inference-service] headers = headers.copy() # type: ignore[attr-defined] [e2e-llm-inference-service] headers.update(self.proxy_headers) # type: ignore[union-attr] [e2e-llm-inference-service] [e2e-llm-inference-service] # Must keep the exception bound to a separate variable or else Python 3 [e2e-llm-inference-service] # complains about UnboundLocalError. [e2e-llm-inference-service] err = None [e2e-llm-inference-service] [e2e-llm-inference-service] # Keep track of whether we cleanly exited the except block. This [e2e-llm-inference-service] # ensures we do proper cleanup in finally. [e2e-llm-inference-service] clean_exit = False [e2e-llm-inference-service] [e2e-llm-inference-service] # Rewind body position, if needed. Record current position [e2e-llm-inference-service] # for future rewinds in the event of a redirect/retry. [e2e-llm-inference-service] body_pos = set_file_position(body, body_pos) [e2e-llm-inference-service] [e2e-llm-inference-service] try: [e2e-llm-inference-service] # Request a connection from the queue. [e2e-llm-inference-service] timeout_obj = self._get_timeout(timeout) [e2e-llm-inference-service] conn = self._get_conn(timeout=pool_timeout) [e2e-llm-inference-service] [e2e-llm-inference-service] conn.timeout = timeout_obj.connect_timeout # type: ignore[assignment] [e2e-llm-inference-service] [e2e-llm-inference-service] # Is this a closed/new connection that requires CONNECT tunnelling? [e2e-llm-inference-service] if self.proxy is not None and http_tunnel_required and conn.is_closed: [e2e-llm-inference-service] try: [e2e-llm-inference-service] self._prepare_proxy(conn) [e2e-llm-inference-service] except (BaseSSLError, OSError, SocketTimeout) as e: [e2e-llm-inference-service] self._raise_timeout( [e2e-llm-inference-service] err=e, url=self.proxy.url, timeout_value=conn.timeout [e2e-llm-inference-service] ) [e2e-llm-inference-service] raise [e2e-llm-inference-service] [e2e-llm-inference-service] # If we're going to release the connection in ``finally:``, then [e2e-llm-inference-service] # the response doesn't need to know about the connection. Otherwise [e2e-llm-inference-service] # it will also try to release it and we'll have a double-release [e2e-llm-inference-service] # mess. [e2e-llm-inference-service] response_conn = conn if not release_conn else None [e2e-llm-inference-service] [e2e-llm-inference-service] # Make the request on the HTTPConnection object [e2e-llm-inference-service] response = self._make_request( [e2e-llm-inference-service] conn, [e2e-llm-inference-service] method, [e2e-llm-inference-service] url, [e2e-llm-inference-service] timeout=timeout_obj, [e2e-llm-inference-service] body=body, [e2e-llm-inference-service] headers=headers, [e2e-llm-inference-service] chunked=chunked, [e2e-llm-inference-service] retries=retries, [e2e-llm-inference-service] response_conn=response_conn, [e2e-llm-inference-service] preload_content=preload_content, [e2e-llm-inference-service] decode_content=decode_content, [e2e-llm-inference-service] **response_kw, [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] # Everything went great! [e2e-llm-inference-service] clean_exit = True [e2e-llm-inference-service] [e2e-llm-inference-service] except EmptyPoolError: [e2e-llm-inference-service] # Didn't get a connection from the pool, no need to clean up [e2e-llm-inference-service] clean_exit = True [e2e-llm-inference-service] release_this_conn = False [e2e-llm-inference-service] raise [e2e-llm-inference-service] [e2e-llm-inference-service] except ( [e2e-llm-inference-service] TimeoutError, [e2e-llm-inference-service] HTTPException, [e2e-llm-inference-service] OSError, [e2e-llm-inference-service] ProtocolError, [e2e-llm-inference-service] BaseSSLError, [e2e-llm-inference-service] SSLError, [e2e-llm-inference-service] CertificateError, [e2e-llm-inference-service] ProxyError, [e2e-llm-inference-service] ) as e: [e2e-llm-inference-service] # Discard the connection for these exceptions. It will be [e2e-llm-inference-service] # replaced during the next _get_conn() call. [e2e-llm-inference-service] clean_exit = False [e2e-llm-inference-service] new_e: Exception = e [e2e-llm-inference-service] if isinstance(e, (BaseSSLError, CertificateError)): [e2e-llm-inference-service] new_e = SSLError(e) [e2e-llm-inference-service] if isinstance( [e2e-llm-inference-service] new_e, [e2e-llm-inference-service] ( [e2e-llm-inference-service] OSError, [e2e-llm-inference-service] NewConnectionError, [e2e-llm-inference-service] TimeoutError, [e2e-llm-inference-service] SSLError, [e2e-llm-inference-service] HTTPException, [e2e-llm-inference-service] ), [e2e-llm-inference-service] ) and (conn and conn.proxy and not conn.has_connected_to_proxy): [e2e-llm-inference-service] new_e = _wrap_proxy_error(new_e, conn.proxy.scheme) [e2e-llm-inference-service] elif isinstance(new_e, (OSError, HTTPException)): [e2e-llm-inference-service] new_e = ProtocolError("Connection aborted.", new_e) [e2e-llm-inference-service] [e2e-llm-inference-service] retries = retries.increment( [e2e-llm-inference-service] method, url, error=new_e, _pool=self, _stacktrace=sys.exc_info()[2] [e2e-llm-inference-service] ) [e2e-llm-inference-service] retries.sleep() [e2e-llm-inference-service] [e2e-llm-inference-service] # Keep track of the error for the retry warning. [e2e-llm-inference-service] err = e [e2e-llm-inference-service] [e2e-llm-inference-service] finally: [e2e-llm-inference-service] if not clean_exit: [e2e-llm-inference-service] # We hit some kind of exception, handled or otherwise. We need [e2e-llm-inference-service] # to throw the connection away unless explicitly told not to. [e2e-llm-inference-service] # Close the connection, set the variable to None, and make sure [e2e-llm-inference-service] # we put the None back in the pool to avoid leaking it. [e2e-llm-inference-service] if conn: [e2e-llm-inference-service] conn.close() [e2e-llm-inference-service] conn = None [e2e-llm-inference-service] release_this_conn = True [e2e-llm-inference-service] [e2e-llm-inference-service] if release_this_conn: [e2e-llm-inference-service] # Put the connection back to be reused. If the connection is [e2e-llm-inference-service] # expired then it will be None, which will get replaced with a [e2e-llm-inference-service] # fresh connection during _get_conn. [e2e-llm-inference-service] self._put_conn(conn) [e2e-llm-inference-service] [e2e-llm-inference-service] if not conn: [e2e-llm-inference-service] # Try again [e2e-llm-inference-service] log.warning( [e2e-llm-inference-service] "Retrying (%r) after connection broken by '%r': %s", retries, err, url [e2e-llm-inference-service] ) [e2e-llm-inference-service] > return self.urlopen( [e2e-llm-inference-service] method, [e2e-llm-inference-service] url, [e2e-llm-inference-service] body, [e2e-llm-inference-service] headers, [e2e-llm-inference-service] retries, [e2e-llm-inference-service] redirect, [e2e-llm-inference-service] assert_same_host, [e2e-llm-inference-service] timeout=timeout, [e2e-llm-inference-service] pool_timeout=pool_timeout, [e2e-llm-inference-service] release_conn=release_conn, [e2e-llm-inference-service] chunked=chunked, [e2e-llm-inference-service] body_pos=body_pos, [e2e-llm-inference-service] preload_content=preload_content, [e2e-llm-inference-service] decode_content=decode_content, [e2e-llm-inference-service] **response_kw, [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] ../../python/kserve/.venv/lib64/python3.11/site-packages/urllib3/connectionpool.py:871: [e2e-llm-inference-service] _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ [e2e-llm-inference-service] [e2e-llm-inference-service] self = [e2e-llm-inference-service] method = 'POST', url = '/v1/completions' [e2e-llm-inference-service] body = b'{"model": "publishers/kserve-ci-e2e-test/models/lora-adapter-1", "prompt": "KServe is a", "max_tokens": 20}' [e2e-llm-inference-service] headers = {'User-Agent': 'python-requests/2.32.3', 'Accept-Encoding': 'gzip, deflate', 'Accept': '*/*', 'Connection': 'keep-aliv...lication/json', 'X-Gateway-Model-Name': 'publishers/kserve-ci-e2e-test/models/lora-adapter-1', 'Content-Length': '107'} [e2e-llm-inference-service] retries = Retry(total=3, connect=None, read=None, redirect=None, status=None) [e2e-llm-inference-service] redirect = False, assert_same_host = False [e2e-llm-inference-service] timeout = Timeout(connect=60, read=60, total=None), pool_timeout = None [e2e-llm-inference-service] release_conn = False, chunked = False, body_pos = None, preload_content = False [e2e-llm-inference-service] decode_content = False, response_kw = {} [e2e-llm-inference-service] parsed_url = Url(scheme=None, auth=None, host=None, port=None, path='/v1/completions', query=None, fragment=None) [e2e-llm-inference-service] destination_scheme = None, conn = None, release_this_conn = True [e2e-llm-inference-service] http_tunnel_required = False [e2e-llm-inference-service] err = ReadTimeoutError("HTTPConnectionPool(host='ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com', port=80): Read timed out. (read timeout=60)") [e2e-llm-inference-service] clean_exit = False [e2e-llm-inference-service] [e2e-llm-inference-service] def urlopen( # type: ignore[override] [e2e-llm-inference-service] self, [e2e-llm-inference-service] method: str, [e2e-llm-inference-service] url: str, [e2e-llm-inference-service] body: _TYPE_BODY | None = None, [e2e-llm-inference-service] headers: typing.Mapping[str, str] | None = None, [e2e-llm-inference-service] retries: Retry | bool | int | None = None, [e2e-llm-inference-service] redirect: bool = True, [e2e-llm-inference-service] assert_same_host: bool = True, [e2e-llm-inference-service] timeout: _TYPE_TIMEOUT = _DEFAULT_TIMEOUT, [e2e-llm-inference-service] pool_timeout: int | None = None, [e2e-llm-inference-service] release_conn: bool | None = None, [e2e-llm-inference-service] chunked: bool = False, [e2e-llm-inference-service] body_pos: _TYPE_BODY_POSITION | None = None, [e2e-llm-inference-service] preload_content: bool = True, [e2e-llm-inference-service] decode_content: bool = True, [e2e-llm-inference-service] **response_kw: typing.Any, [e2e-llm-inference-service] ) -> BaseHTTPResponse: [e2e-llm-inference-service] """ [e2e-llm-inference-service] Get a connection from the pool and perform an HTTP request. This is the [e2e-llm-inference-service] lowest level call for making a request, so you'll need to specify all [e2e-llm-inference-service] the raw details. [e2e-llm-inference-service] [e2e-llm-inference-service] .. note:: [e2e-llm-inference-service] [e2e-llm-inference-service] More commonly, it's appropriate to use a convenience method [e2e-llm-inference-service] such as :meth:`request`. [e2e-llm-inference-service] [e2e-llm-inference-service] .. note:: [e2e-llm-inference-service] [e2e-llm-inference-service] `release_conn` will only behave as expected if [e2e-llm-inference-service] `preload_content=False` because we want to make [e2e-llm-inference-service] `preload_content=False` the default behaviour someday soon without [e2e-llm-inference-service] breaking backwards compatibility. [e2e-llm-inference-service] [e2e-llm-inference-service] :param method: [e2e-llm-inference-service] HTTP request method (such as GET, POST, PUT, etc.) [e2e-llm-inference-service] [e2e-llm-inference-service] :param url: [e2e-llm-inference-service] The URL to perform the request on. [e2e-llm-inference-service] [e2e-llm-inference-service] :param body: [e2e-llm-inference-service] Data to send in the request body, either :class:`str`, :class:`bytes`, [e2e-llm-inference-service] an iterable of :class:`str`/:class:`bytes`, or a file-like object. [e2e-llm-inference-service] [e2e-llm-inference-service] :param headers: [e2e-llm-inference-service] Dictionary of custom headers to send, such as User-Agent, [e2e-llm-inference-service] If-None-Match, etc. If None, pool headers are used. If provided, [e2e-llm-inference-service] these headers completely replace any pool-specific headers. [e2e-llm-inference-service] [e2e-llm-inference-service] :param retries: [e2e-llm-inference-service] Configure the number of retries to allow before raising a [e2e-llm-inference-service] :class:`~urllib3.exceptions.MaxRetryError` exception. [e2e-llm-inference-service] [e2e-llm-inference-service] If ``None`` (default) will retry 3 times, see ``Retry.DEFAULT``. Pass a [e2e-llm-inference-service] :class:`~urllib3.util.retry.Retry` object for fine-grained control [e2e-llm-inference-service] over different types of retries. [e2e-llm-inference-service] Pass an integer number to retry connection errors that many times, [e2e-llm-inference-service] but no other types of errors. Pass zero to never retry. [e2e-llm-inference-service] [e2e-llm-inference-service] If ``False``, then retries are disabled and any exception is raised [e2e-llm-inference-service] immediately. Also, instead of raising a MaxRetryError on redirects, [e2e-llm-inference-service] the redirect response will be returned. [e2e-llm-inference-service] [e2e-llm-inference-service] :type retries: :class:`~urllib3.util.retry.Retry`, False, or an int. [e2e-llm-inference-service] [e2e-llm-inference-service] :param redirect: [e2e-llm-inference-service] If True, automatically handle redirects (status codes 301, 302, [e2e-llm-inference-service] 303, 307, 308). Each redirect counts as a retry. Disabling retries [e2e-llm-inference-service] will disable redirect, too. [e2e-llm-inference-service] [e2e-llm-inference-service] :param assert_same_host: [e2e-llm-inference-service] If ``True``, will make sure that the host of the pool requests is [e2e-llm-inference-service] consistent else will raise HostChangedError. When ``False``, you can [e2e-llm-inference-service] use the pool on an HTTP proxy and request foreign hosts. [e2e-llm-inference-service] [e2e-llm-inference-service] :param timeout: [e2e-llm-inference-service] If specified, overrides the default timeout for this one [e2e-llm-inference-service] request. It may be a float (in seconds) or an instance of [e2e-llm-inference-service] :class:`urllib3.util.Timeout`. [e2e-llm-inference-service] [e2e-llm-inference-service] :param pool_timeout: [e2e-llm-inference-service] If set and the pool is set to block=True, then this method will [e2e-llm-inference-service] block for ``pool_timeout`` seconds and raise EmptyPoolError if no [e2e-llm-inference-service] connection is available within the time period. [e2e-llm-inference-service] [e2e-llm-inference-service] :param bool preload_content: [e2e-llm-inference-service] If True, the response's body will be preloaded into memory. [e2e-llm-inference-service] [e2e-llm-inference-service] :param bool decode_content: [e2e-llm-inference-service] If True, will attempt to decode the body based on the [e2e-llm-inference-service] 'content-encoding' header. [e2e-llm-inference-service] [e2e-llm-inference-service] :param release_conn: [e2e-llm-inference-service] If False, then the urlopen call will not release the connection [e2e-llm-inference-service] back into the pool once a response is received (but will release if [e2e-llm-inference-service] you read the entire contents of the response such as when [e2e-llm-inference-service] `preload_content=True`). This is useful if you're not preloading [e2e-llm-inference-service] the response's content immediately. You will need to call [e2e-llm-inference-service] ``r.release_conn()`` on the response ``r`` to return the connection [e2e-llm-inference-service] back into the pool. If None, it takes the value of ``preload_content`` [e2e-llm-inference-service] which defaults to ``True``. [e2e-llm-inference-service] [e2e-llm-inference-service] :param bool chunked: [e2e-llm-inference-service] If True, urllib3 will send the body using chunked transfer [e2e-llm-inference-service] encoding. Otherwise, urllib3 will send the body using the standard [e2e-llm-inference-service] content-length form. Defaults to False. [e2e-llm-inference-service] [e2e-llm-inference-service] :param int body_pos: [e2e-llm-inference-service] Position to seek to in file-like body in the event of a retry or [e2e-llm-inference-service] redirect. Typically this won't need to be set because urllib3 will [e2e-llm-inference-service] auto-populate the value when needed. [e2e-llm-inference-service] """ [e2e-llm-inference-service] parsed_url = parse_url(url) [e2e-llm-inference-service] destination_scheme = parsed_url.scheme [e2e-llm-inference-service] [e2e-llm-inference-service] if headers is None: [e2e-llm-inference-service] headers = self.headers [e2e-llm-inference-service] [e2e-llm-inference-service] if not isinstance(retries, Retry): [e2e-llm-inference-service] retries = Retry.from_int(retries, redirect=redirect, default=self.retries) [e2e-llm-inference-service] [e2e-llm-inference-service] if release_conn is None: [e2e-llm-inference-service] release_conn = preload_content [e2e-llm-inference-service] [e2e-llm-inference-service] # Check host [e2e-llm-inference-service] if assert_same_host and not self.is_same_host(url): [e2e-llm-inference-service] raise HostChangedError(self, url, retries) [e2e-llm-inference-service] [e2e-llm-inference-service] # Ensure that the URL we're connecting to is properly encoded [e2e-llm-inference-service] if url.startswith("/"): [e2e-llm-inference-service] url = to_str(_encode_target(url)) [e2e-llm-inference-service] else: [e2e-llm-inference-service] url = to_str(parsed_url.url) [e2e-llm-inference-service] [e2e-llm-inference-service] conn = None [e2e-llm-inference-service] [e2e-llm-inference-service] # Track whether `conn` needs to be released before [e2e-llm-inference-service] # returning/raising/recursing. Update this variable if necessary, and [e2e-llm-inference-service] # leave `release_conn` constant throughout the function. That way, if [e2e-llm-inference-service] # the function recurses, the original value of `release_conn` will be [e2e-llm-inference-service] # passed down into the recursive call, and its value will be respected. [e2e-llm-inference-service] # [e2e-llm-inference-service] # See issue #651 [1] for details. [e2e-llm-inference-service] # [e2e-llm-inference-service] # [1] [e2e-llm-inference-service] release_this_conn = release_conn [e2e-llm-inference-service] [e2e-llm-inference-service] http_tunnel_required = connection_requires_http_tunnel( [e2e-llm-inference-service] self.proxy, self.proxy_config, destination_scheme [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] # Merge the proxy headers. Only done when not using HTTP CONNECT. We [e2e-llm-inference-service] # have to copy the headers dict so we can safely change it without those [e2e-llm-inference-service] # changes being reflected in anyone else's copy. [e2e-llm-inference-service] if not http_tunnel_required: [e2e-llm-inference-service] headers = headers.copy() # type: ignore[attr-defined] [e2e-llm-inference-service] headers.update(self.proxy_headers) # type: ignore[union-attr] [e2e-llm-inference-service] [e2e-llm-inference-service] # Must keep the exception bound to a separate variable or else Python 3 [e2e-llm-inference-service] # complains about UnboundLocalError. [e2e-llm-inference-service] err = None [e2e-llm-inference-service] [e2e-llm-inference-service] # Keep track of whether we cleanly exited the except block. This [e2e-llm-inference-service] # ensures we do proper cleanup in finally. [e2e-llm-inference-service] clean_exit = False [e2e-llm-inference-service] [e2e-llm-inference-service] # Rewind body position, if needed. Record current position [e2e-llm-inference-service] # for future rewinds in the event of a redirect/retry. [e2e-llm-inference-service] body_pos = set_file_position(body, body_pos) [e2e-llm-inference-service] [e2e-llm-inference-service] try: [e2e-llm-inference-service] # Request a connection from the queue. [e2e-llm-inference-service] timeout_obj = self._get_timeout(timeout) [e2e-llm-inference-service] conn = self._get_conn(timeout=pool_timeout) [e2e-llm-inference-service] [e2e-llm-inference-service] conn.timeout = timeout_obj.connect_timeout # type: ignore[assignment] [e2e-llm-inference-service] [e2e-llm-inference-service] # Is this a closed/new connection that requires CONNECT tunnelling? [e2e-llm-inference-service] if self.proxy is not None and http_tunnel_required and conn.is_closed: [e2e-llm-inference-service] try: [e2e-llm-inference-service] self._prepare_proxy(conn) [e2e-llm-inference-service] except (BaseSSLError, OSError, SocketTimeout) as e: [e2e-llm-inference-service] self._raise_timeout( [e2e-llm-inference-service] err=e, url=self.proxy.url, timeout_value=conn.timeout [e2e-llm-inference-service] ) [e2e-llm-inference-service] raise [e2e-llm-inference-service] [e2e-llm-inference-service] # If we're going to release the connection in ``finally:``, then [e2e-llm-inference-service] # the response doesn't need to know about the connection. Otherwise [e2e-llm-inference-service] # it will also try to release it and we'll have a double-release [e2e-llm-inference-service] # mess. [e2e-llm-inference-service] response_conn = conn if not release_conn else None [e2e-llm-inference-service] [e2e-llm-inference-service] # Make the request on the HTTPConnection object [e2e-llm-inference-service] response = self._make_request( [e2e-llm-inference-service] conn, [e2e-llm-inference-service] method, [e2e-llm-inference-service] url, [e2e-llm-inference-service] timeout=timeout_obj, [e2e-llm-inference-service] body=body, [e2e-llm-inference-service] headers=headers, [e2e-llm-inference-service] chunked=chunked, [e2e-llm-inference-service] retries=retries, [e2e-llm-inference-service] response_conn=response_conn, [e2e-llm-inference-service] preload_content=preload_content, [e2e-llm-inference-service] decode_content=decode_content, [e2e-llm-inference-service] **response_kw, [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] # Everything went great! [e2e-llm-inference-service] clean_exit = True [e2e-llm-inference-service] [e2e-llm-inference-service] except EmptyPoolError: [e2e-llm-inference-service] # Didn't get a connection from the pool, no need to clean up [e2e-llm-inference-service] clean_exit = True [e2e-llm-inference-service] release_this_conn = False [e2e-llm-inference-service] raise [e2e-llm-inference-service] [e2e-llm-inference-service] except ( [e2e-llm-inference-service] TimeoutError, [e2e-llm-inference-service] HTTPException, [e2e-llm-inference-service] OSError, [e2e-llm-inference-service] ProtocolError, [e2e-llm-inference-service] BaseSSLError, [e2e-llm-inference-service] SSLError, [e2e-llm-inference-service] CertificateError, [e2e-llm-inference-service] ProxyError, [e2e-llm-inference-service] ) as e: [e2e-llm-inference-service] # Discard the connection for these exceptions. It will be [e2e-llm-inference-service] # replaced during the next _get_conn() call. [e2e-llm-inference-service] clean_exit = False [e2e-llm-inference-service] new_e: Exception = e [e2e-llm-inference-service] if isinstance(e, (BaseSSLError, CertificateError)): [e2e-llm-inference-service] new_e = SSLError(e) [e2e-llm-inference-service] if isinstance( [e2e-llm-inference-service] new_e, [e2e-llm-inference-service] ( [e2e-llm-inference-service] OSError, [e2e-llm-inference-service] NewConnectionError, [e2e-llm-inference-service] TimeoutError, [e2e-llm-inference-service] SSLError, [e2e-llm-inference-service] HTTPException, [e2e-llm-inference-service] ), [e2e-llm-inference-service] ) and (conn and conn.proxy and not conn.has_connected_to_proxy): [e2e-llm-inference-service] new_e = _wrap_proxy_error(new_e, conn.proxy.scheme) [e2e-llm-inference-service] elif isinstance(new_e, (OSError, HTTPException)): [e2e-llm-inference-service] new_e = ProtocolError("Connection aborted.", new_e) [e2e-llm-inference-service] [e2e-llm-inference-service] retries = retries.increment( [e2e-llm-inference-service] method, url, error=new_e, _pool=self, _stacktrace=sys.exc_info()[2] [e2e-llm-inference-service] ) [e2e-llm-inference-service] retries.sleep() [e2e-llm-inference-service] [e2e-llm-inference-service] # Keep track of the error for the retry warning. [e2e-llm-inference-service] err = e [e2e-llm-inference-service] [e2e-llm-inference-service] finally: [e2e-llm-inference-service] if not clean_exit: [e2e-llm-inference-service] # We hit some kind of exception, handled or otherwise. We need [e2e-llm-inference-service] # to throw the connection away unless explicitly told not to. [e2e-llm-inference-service] # Close the connection, set the variable to None, and make sure [e2e-llm-inference-service] # we put the None back in the pool to avoid leaking it. [e2e-llm-inference-service] if conn: [e2e-llm-inference-service] conn.close() [e2e-llm-inference-service] conn = None [e2e-llm-inference-service] release_this_conn = True [e2e-llm-inference-service] [e2e-llm-inference-service] if release_this_conn: [e2e-llm-inference-service] # Put the connection back to be reused. If the connection is [e2e-llm-inference-service] # expired then it will be None, which will get replaced with a [e2e-llm-inference-service] # fresh connection during _get_conn. [e2e-llm-inference-service] self._put_conn(conn) [e2e-llm-inference-service] [e2e-llm-inference-service] if not conn: [e2e-llm-inference-service] # Try again [e2e-llm-inference-service] log.warning( [e2e-llm-inference-service] "Retrying (%r) after connection broken by '%r': %s", retries, err, url [e2e-llm-inference-service] ) [e2e-llm-inference-service] > return self.urlopen( [e2e-llm-inference-service] method, [e2e-llm-inference-service] url, [e2e-llm-inference-service] body, [e2e-llm-inference-service] headers, [e2e-llm-inference-service] retries, [e2e-llm-inference-service] redirect, [e2e-llm-inference-service] assert_same_host, [e2e-llm-inference-service] timeout=timeout, [e2e-llm-inference-service] pool_timeout=pool_timeout, [e2e-llm-inference-service] release_conn=release_conn, [e2e-llm-inference-service] chunked=chunked, [e2e-llm-inference-service] body_pos=body_pos, [e2e-llm-inference-service] preload_content=preload_content, [e2e-llm-inference-service] decode_content=decode_content, [e2e-llm-inference-service] **response_kw, [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] ../../python/kserve/.venv/lib64/python3.11/site-packages/urllib3/connectionpool.py:871: [e2e-llm-inference-service] _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ [e2e-llm-inference-service] [e2e-llm-inference-service] self = [e2e-llm-inference-service] method = 'POST', url = '/v1/completions' [e2e-llm-inference-service] body = b'{"model": "publishers/kserve-ci-e2e-test/models/lora-adapter-1", "prompt": "KServe is a", "max_tokens": 20}' [e2e-llm-inference-service] headers = {'User-Agent': 'python-requests/2.32.3', 'Accept-Encoding': 'gzip, deflate', 'Accept': '*/*', 'Connection': 'keep-aliv...lication/json', 'X-Gateway-Model-Name': 'publishers/kserve-ci-e2e-test/models/lora-adapter-1', 'Content-Length': '107'} [e2e-llm-inference-service] retries = Retry(total=2, connect=None, read=None, redirect=None, status=None) [e2e-llm-inference-service] redirect = False, assert_same_host = False [e2e-llm-inference-service] timeout = Timeout(connect=60, read=60, total=None), pool_timeout = None [e2e-llm-inference-service] release_conn = False, chunked = False, body_pos = None, preload_content = False [e2e-llm-inference-service] decode_content = False, response_kw = {} [e2e-llm-inference-service] parsed_url = Url(scheme=None, auth=None, host=None, port=None, path='/v1/completions', query=None, fragment=None) [e2e-llm-inference-service] destination_scheme = None, conn = None, release_this_conn = True [e2e-llm-inference-service] http_tunnel_required = False [e2e-llm-inference-service] err = ReadTimeoutError("HTTPConnectionPool(host='ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com', port=80): Read timed out. (read timeout=60)") [e2e-llm-inference-service] clean_exit = False [e2e-llm-inference-service] [e2e-llm-inference-service] def urlopen( # type: ignore[override] [e2e-llm-inference-service] self, [e2e-llm-inference-service] method: str, [e2e-llm-inference-service] url: str, [e2e-llm-inference-service] body: _TYPE_BODY | None = None, [e2e-llm-inference-service] headers: typing.Mapping[str, str] | None = None, [e2e-llm-inference-service] retries: Retry | bool | int | None = None, [e2e-llm-inference-service] redirect: bool = True, [e2e-llm-inference-service] assert_same_host: bool = True, [e2e-llm-inference-service] timeout: _TYPE_TIMEOUT = _DEFAULT_TIMEOUT, [e2e-llm-inference-service] pool_timeout: int | None = None, [e2e-llm-inference-service] release_conn: bool | None = None, [e2e-llm-inference-service] chunked: bool = False, [e2e-llm-inference-service] body_pos: _TYPE_BODY_POSITION | None = None, [e2e-llm-inference-service] preload_content: bool = True, [e2e-llm-inference-service] decode_content: bool = True, [e2e-llm-inference-service] **response_kw: typing.Any, [e2e-llm-inference-service] ) -> BaseHTTPResponse: [e2e-llm-inference-service] """ [e2e-llm-inference-service] Get a connection from the pool and perform an HTTP request. This is the [e2e-llm-inference-service] lowest level call for making a request, so you'll need to specify all [e2e-llm-inference-service] the raw details. [e2e-llm-inference-service] [e2e-llm-inference-service] .. note:: [e2e-llm-inference-service] [e2e-llm-inference-service] More commonly, it's appropriate to use a convenience method [e2e-llm-inference-service] such as :meth:`request`. [e2e-llm-inference-service] [e2e-llm-inference-service] .. note:: [e2e-llm-inference-service] [e2e-llm-inference-service] `release_conn` will only behave as expected if [e2e-llm-inference-service] `preload_content=False` because we want to make [e2e-llm-inference-service] `preload_content=False` the default behaviour someday soon without [e2e-llm-inference-service] breaking backwards compatibility. [e2e-llm-inference-service] [e2e-llm-inference-service] :param method: [e2e-llm-inference-service] HTTP request method (such as GET, POST, PUT, etc.) [e2e-llm-inference-service] [e2e-llm-inference-service] :param url: [e2e-llm-inference-service] The URL to perform the request on. [e2e-llm-inference-service] [e2e-llm-inference-service] :param body: [e2e-llm-inference-service] Data to send in the request body, either :class:`str`, :class:`bytes`, [e2e-llm-inference-service] an iterable of :class:`str`/:class:`bytes`, or a file-like object. [e2e-llm-inference-service] [e2e-llm-inference-service] :param headers: [e2e-llm-inference-service] Dictionary of custom headers to send, such as User-Agent, [e2e-llm-inference-service] If-None-Match, etc. If None, pool headers are used. If provided, [e2e-llm-inference-service] these headers completely replace any pool-specific headers. [e2e-llm-inference-service] [e2e-llm-inference-service] :param retries: [e2e-llm-inference-service] Configure the number of retries to allow before raising a [e2e-llm-inference-service] :class:`~urllib3.exceptions.MaxRetryError` exception. [e2e-llm-inference-service] [e2e-llm-inference-service] If ``None`` (default) will retry 3 times, see ``Retry.DEFAULT``. Pass a [e2e-llm-inference-service] :class:`~urllib3.util.retry.Retry` object for fine-grained control [e2e-llm-inference-service] over different types of retries. [e2e-llm-inference-service] Pass an integer number to retry connection errors that many times, [e2e-llm-inference-service] but no other types of errors. Pass zero to never retry. [e2e-llm-inference-service] [e2e-llm-inference-service] If ``False``, then retries are disabled and any exception is raised [e2e-llm-inference-service] immediately. Also, instead of raising a MaxRetryError on redirects, [e2e-llm-inference-service] the redirect response will be returned. [e2e-llm-inference-service] [e2e-llm-inference-service] :type retries: :class:`~urllib3.util.retry.Retry`, False, or an int. [e2e-llm-inference-service] [e2e-llm-inference-service] :param redirect: [e2e-llm-inference-service] If True, automatically handle redirects (status codes 301, 302, [e2e-llm-inference-service] 303, 307, 308). Each redirect counts as a retry. Disabling retries [e2e-llm-inference-service] will disable redirect, too. [e2e-llm-inference-service] [e2e-llm-inference-service] :param assert_same_host: [e2e-llm-inference-service] If ``True``, will make sure that the host of the pool requests is [e2e-llm-inference-service] consistent else will raise HostChangedError. When ``False``, you can [e2e-llm-inference-service] use the pool on an HTTP proxy and request foreign hosts. [e2e-llm-inference-service] [e2e-llm-inference-service] :param timeout: [e2e-llm-inference-service] If specified, overrides the default timeout for this one [e2e-llm-inference-service] request. It may be a float (in seconds) or an instance of [e2e-llm-inference-service] :class:`urllib3.util.Timeout`. [e2e-llm-inference-service] [e2e-llm-inference-service] :param pool_timeout: [e2e-llm-inference-service] If set and the pool is set to block=True, then this method will [e2e-llm-inference-service] block for ``pool_timeout`` seconds and raise EmptyPoolError if no [e2e-llm-inference-service] connection is available within the time period. [e2e-llm-inference-service] [e2e-llm-inference-service] :param bool preload_content: [e2e-llm-inference-service] If True, the response's body will be preloaded into memory. [e2e-llm-inference-service] [e2e-llm-inference-service] :param bool decode_content: [e2e-llm-inference-service] If True, will attempt to decode the body based on the [e2e-llm-inference-service] 'content-encoding' header. [e2e-llm-inference-service] [e2e-llm-inference-service] :param release_conn: [e2e-llm-inference-service] If False, then the urlopen call will not release the connection [e2e-llm-inference-service] back into the pool once a response is received (but will release if [e2e-llm-inference-service] you read the entire contents of the response such as when [e2e-llm-inference-service] `preload_content=True`). This is useful if you're not preloading [e2e-llm-inference-service] the response's content immediately. You will need to call [e2e-llm-inference-service] ``r.release_conn()`` on the response ``r`` to return the connection [e2e-llm-inference-service] back into the pool. If None, it takes the value of ``preload_content`` [e2e-llm-inference-service] which defaults to ``True``. [e2e-llm-inference-service] [e2e-llm-inference-service] :param bool chunked: [e2e-llm-inference-service] If True, urllib3 will send the body using chunked transfer [e2e-llm-inference-service] encoding. Otherwise, urllib3 will send the body using the standard [e2e-llm-inference-service] content-length form. Defaults to False. [e2e-llm-inference-service] [e2e-llm-inference-service] :param int body_pos: [e2e-llm-inference-service] Position to seek to in file-like body in the event of a retry or [e2e-llm-inference-service] redirect. Typically this won't need to be set because urllib3 will [e2e-llm-inference-service] auto-populate the value when needed. [e2e-llm-inference-service] """ [e2e-llm-inference-service] parsed_url = parse_url(url) [e2e-llm-inference-service] destination_scheme = parsed_url.scheme [e2e-llm-inference-service] [e2e-llm-inference-service] if headers is None: [e2e-llm-inference-service] headers = self.headers [e2e-llm-inference-service] [e2e-llm-inference-service] if not isinstance(retries, Retry): [e2e-llm-inference-service] retries = Retry.from_int(retries, redirect=redirect, default=self.retries) [e2e-llm-inference-service] [e2e-llm-inference-service] if release_conn is None: [e2e-llm-inference-service] release_conn = preload_content [e2e-llm-inference-service] [e2e-llm-inference-service] # Check host [e2e-llm-inference-service] if assert_same_host and not self.is_same_host(url): [e2e-llm-inference-service] raise HostChangedError(self, url, retries) [e2e-llm-inference-service] [e2e-llm-inference-service] # Ensure that the URL we're connecting to is properly encoded [e2e-llm-inference-service] if url.startswith("/"): [e2e-llm-inference-service] url = to_str(_encode_target(url)) [e2e-llm-inference-service] else: [e2e-llm-inference-service] url = to_str(parsed_url.url) [e2e-llm-inference-service] [e2e-llm-inference-service] conn = None [e2e-llm-inference-service] [e2e-llm-inference-service] # Track whether `conn` needs to be released before [e2e-llm-inference-service] # returning/raising/recursing. Update this variable if necessary, and [e2e-llm-inference-service] # leave `release_conn` constant throughout the function. That way, if [e2e-llm-inference-service] # the function recurses, the original value of `release_conn` will be [e2e-llm-inference-service] # passed down into the recursive call, and its value will be respected. [e2e-llm-inference-service] # [e2e-llm-inference-service] # See issue #651 [1] for details. [e2e-llm-inference-service] # [e2e-llm-inference-service] # [1] [e2e-llm-inference-service] release_this_conn = release_conn [e2e-llm-inference-service] [e2e-llm-inference-service] http_tunnel_required = connection_requires_http_tunnel( [e2e-llm-inference-service] self.proxy, self.proxy_config, destination_scheme [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] # Merge the proxy headers. Only done when not using HTTP CONNECT. We [e2e-llm-inference-service] # have to copy the headers dict so we can safely change it without those [e2e-llm-inference-service] # changes being reflected in anyone else's copy. [e2e-llm-inference-service] if not http_tunnel_required: [e2e-llm-inference-service] headers = headers.copy() # type: ignore[attr-defined] [e2e-llm-inference-service] headers.update(self.proxy_headers) # type: ignore[union-attr] [e2e-llm-inference-service] [e2e-llm-inference-service] # Must keep the exception bound to a separate variable or else Python 3 [e2e-llm-inference-service] # complains about UnboundLocalError. [e2e-llm-inference-service] err = None [e2e-llm-inference-service] [e2e-llm-inference-service] # Keep track of whether we cleanly exited the except block. This [e2e-llm-inference-service] # ensures we do proper cleanup in finally. [e2e-llm-inference-service] clean_exit = False [e2e-llm-inference-service] [e2e-llm-inference-service] # Rewind body position, if needed. Record current position [e2e-llm-inference-service] # for future rewinds in the event of a redirect/retry. [e2e-llm-inference-service] body_pos = set_file_position(body, body_pos) [e2e-llm-inference-service] [e2e-llm-inference-service] try: [e2e-llm-inference-service] # Request a connection from the queue. [e2e-llm-inference-service] timeout_obj = self._get_timeout(timeout) [e2e-llm-inference-service] conn = self._get_conn(timeout=pool_timeout) [e2e-llm-inference-service] [e2e-llm-inference-service] conn.timeout = timeout_obj.connect_timeout # type: ignore[assignment] [e2e-llm-inference-service] [e2e-llm-inference-service] # Is this a closed/new connection that requires CONNECT tunnelling? [e2e-llm-inference-service] if self.proxy is not None and http_tunnel_required and conn.is_closed: [e2e-llm-inference-service] try: [e2e-llm-inference-service] self._prepare_proxy(conn) [e2e-llm-inference-service] except (BaseSSLError, OSError, SocketTimeout) as e: [e2e-llm-inference-service] self._raise_timeout( [e2e-llm-inference-service] err=e, url=self.proxy.url, timeout_value=conn.timeout [e2e-llm-inference-service] ) [e2e-llm-inference-service] raise [e2e-llm-inference-service] [e2e-llm-inference-service] # If we're going to release the connection in ``finally:``, then [e2e-llm-inference-service] # the response doesn't need to know about the connection. Otherwise [e2e-llm-inference-service] # it will also try to release it and we'll have a double-release [e2e-llm-inference-service] # mess. [e2e-llm-inference-service] response_conn = conn if not release_conn else None [e2e-llm-inference-service] [e2e-llm-inference-service] # Make the request on the HTTPConnection object [e2e-llm-inference-service] response = self._make_request( [e2e-llm-inference-service] conn, [e2e-llm-inference-service] method, [e2e-llm-inference-service] url, [e2e-llm-inference-service] timeout=timeout_obj, [e2e-llm-inference-service] body=body, [e2e-llm-inference-service] headers=headers, [e2e-llm-inference-service] chunked=chunked, [e2e-llm-inference-service] retries=retries, [e2e-llm-inference-service] response_conn=response_conn, [e2e-llm-inference-service] preload_content=preload_content, [e2e-llm-inference-service] decode_content=decode_content, [e2e-llm-inference-service] **response_kw, [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] # Everything went great! [e2e-llm-inference-service] clean_exit = True [e2e-llm-inference-service] [e2e-llm-inference-service] except EmptyPoolError: [e2e-llm-inference-service] # Didn't get a connection from the pool, no need to clean up [e2e-llm-inference-service] clean_exit = True [e2e-llm-inference-service] release_this_conn = False [e2e-llm-inference-service] raise [e2e-llm-inference-service] [e2e-llm-inference-service] except ( [e2e-llm-inference-service] TimeoutError, [e2e-llm-inference-service] HTTPException, [e2e-llm-inference-service] OSError, [e2e-llm-inference-service] ProtocolError, [e2e-llm-inference-service] BaseSSLError, [e2e-llm-inference-service] SSLError, [e2e-llm-inference-service] CertificateError, [e2e-llm-inference-service] ProxyError, [e2e-llm-inference-service] ) as e: [e2e-llm-inference-service] # Discard the connection for these exceptions. It will be [e2e-llm-inference-service] # replaced during the next _get_conn() call. [e2e-llm-inference-service] clean_exit = False [e2e-llm-inference-service] new_e: Exception = e [e2e-llm-inference-service] if isinstance(e, (BaseSSLError, CertificateError)): [e2e-llm-inference-service] new_e = SSLError(e) [e2e-llm-inference-service] if isinstance( [e2e-llm-inference-service] new_e, [e2e-llm-inference-service] ( [e2e-llm-inference-service] OSError, [e2e-llm-inference-service] NewConnectionError, [e2e-llm-inference-service] TimeoutError, [e2e-llm-inference-service] SSLError, [e2e-llm-inference-service] HTTPException, [e2e-llm-inference-service] ), [e2e-llm-inference-service] ) and (conn and conn.proxy and not conn.has_connected_to_proxy): [e2e-llm-inference-service] new_e = _wrap_proxy_error(new_e, conn.proxy.scheme) [e2e-llm-inference-service] elif isinstance(new_e, (OSError, HTTPException)): [e2e-llm-inference-service] new_e = ProtocolError("Connection aborted.", new_e) [e2e-llm-inference-service] [e2e-llm-inference-service] retries = retries.increment( [e2e-llm-inference-service] method, url, error=new_e, _pool=self, _stacktrace=sys.exc_info()[2] [e2e-llm-inference-service] ) [e2e-llm-inference-service] retries.sleep() [e2e-llm-inference-service] [e2e-llm-inference-service] # Keep track of the error for the retry warning. [e2e-llm-inference-service] err = e [e2e-llm-inference-service] [e2e-llm-inference-service] finally: [e2e-llm-inference-service] if not clean_exit: [e2e-llm-inference-service] # We hit some kind of exception, handled or otherwise. We need [e2e-llm-inference-service] # to throw the connection away unless explicitly told not to. [e2e-llm-inference-service] # Close the connection, set the variable to None, and make sure [e2e-llm-inference-service] # we put the None back in the pool to avoid leaking it. [e2e-llm-inference-service] if conn: [e2e-llm-inference-service] conn.close() [e2e-llm-inference-service] conn = None [e2e-llm-inference-service] release_this_conn = True [e2e-llm-inference-service] [e2e-llm-inference-service] if release_this_conn: [e2e-llm-inference-service] # Put the connection back to be reused. If the connection is [e2e-llm-inference-service] # expired then it will be None, which will get replaced with a [e2e-llm-inference-service] # fresh connection during _get_conn. [e2e-llm-inference-service] self._put_conn(conn) [e2e-llm-inference-service] [e2e-llm-inference-service] if not conn: [e2e-llm-inference-service] # Try again [e2e-llm-inference-service] log.warning( [e2e-llm-inference-service] "Retrying (%r) after connection broken by '%r': %s", retries, err, url [e2e-llm-inference-service] ) [e2e-llm-inference-service] > return self.urlopen( [e2e-llm-inference-service] method, [e2e-llm-inference-service] url, [e2e-llm-inference-service] body, [e2e-llm-inference-service] headers, [e2e-llm-inference-service] retries, [e2e-llm-inference-service] redirect, [e2e-llm-inference-service] assert_same_host, [e2e-llm-inference-service] timeout=timeout, [e2e-llm-inference-service] pool_timeout=pool_timeout, [e2e-llm-inference-service] release_conn=release_conn, [e2e-llm-inference-service] chunked=chunked, [e2e-llm-inference-service] body_pos=body_pos, [e2e-llm-inference-service] preload_content=preload_content, [e2e-llm-inference-service] decode_content=decode_content, [e2e-llm-inference-service] **response_kw, [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] ../../python/kserve/.venv/lib64/python3.11/site-packages/urllib3/connectionpool.py:871: [e2e-llm-inference-service] _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ [e2e-llm-inference-service] [e2e-llm-inference-service] self = [e2e-llm-inference-service] method = 'POST', url = '/v1/completions' [e2e-llm-inference-service] body = b'{"model": "publishers/kserve-ci-e2e-test/models/lora-adapter-1", "prompt": "KServe is a", "max_tokens": 20}' [e2e-llm-inference-service] headers = {'User-Agent': 'python-requests/2.32.3', 'Accept-Encoding': 'gzip, deflate', 'Accept': '*/*', 'Connection': 'keep-aliv...lication/json', 'X-Gateway-Model-Name': 'publishers/kserve-ci-e2e-test/models/lora-adapter-1', 'Content-Length': '107'} [e2e-llm-inference-service] retries = Retry(total=1, connect=None, read=None, redirect=None, status=None) [e2e-llm-inference-service] redirect = False, assert_same_host = False [e2e-llm-inference-service] timeout = Timeout(connect=60, read=60, total=None), pool_timeout = None [e2e-llm-inference-service] release_conn = False, chunked = False, body_pos = None, preload_content = False [e2e-llm-inference-service] decode_content = False, response_kw = {} [e2e-llm-inference-service] parsed_url = Url(scheme=None, auth=None, host=None, port=None, path='/v1/completions', query=None, fragment=None) [e2e-llm-inference-service] destination_scheme = None, conn = None, release_this_conn = True [e2e-llm-inference-service] http_tunnel_required = False [e2e-llm-inference-service] err = ReadTimeoutError("HTTPConnectionPool(host='ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com', port=80): Read timed out. (read timeout=60)") [e2e-llm-inference-service] clean_exit = False [e2e-llm-inference-service] [e2e-llm-inference-service] def urlopen( # type: ignore[override] [e2e-llm-inference-service] self, [e2e-llm-inference-service] method: str, [e2e-llm-inference-service] url: str, [e2e-llm-inference-service] body: _TYPE_BODY | None = None, [e2e-llm-inference-service] headers: typing.Mapping[str, str] | None = None, [e2e-llm-inference-service] retries: Retry | bool | int | None = None, [e2e-llm-inference-service] redirect: bool = True, [e2e-llm-inference-service] assert_same_host: bool = True, [e2e-llm-inference-service] timeout: _TYPE_TIMEOUT = _DEFAULT_TIMEOUT, [e2e-llm-inference-service] pool_timeout: int | None = None, [e2e-llm-inference-service] release_conn: bool | None = None, [e2e-llm-inference-service] chunked: bool = False, [e2e-llm-inference-service] body_pos: _TYPE_BODY_POSITION | None = None, [e2e-llm-inference-service] preload_content: bool = True, [e2e-llm-inference-service] decode_content: bool = True, [e2e-llm-inference-service] **response_kw: typing.Any, [e2e-llm-inference-service] ) -> BaseHTTPResponse: [e2e-llm-inference-service] """ [e2e-llm-inference-service] Get a connection from the pool and perform an HTTP request. This is the [e2e-llm-inference-service] lowest level call for making a request, so you'll need to specify all [e2e-llm-inference-service] the raw details. [e2e-llm-inference-service] [e2e-llm-inference-service] .. note:: [e2e-llm-inference-service] [e2e-llm-inference-service] More commonly, it's appropriate to use a convenience method [e2e-llm-inference-service] such as :meth:`request`. [e2e-llm-inference-service] [e2e-llm-inference-service] .. note:: [e2e-llm-inference-service] [e2e-llm-inference-service] `release_conn` will only behave as expected if [e2e-llm-inference-service] `preload_content=False` because we want to make [e2e-llm-inference-service] `preload_content=False` the default behaviour someday soon without [e2e-llm-inference-service] breaking backwards compatibility. [e2e-llm-inference-service] [e2e-llm-inference-service] :param method: [e2e-llm-inference-service] HTTP request method (such as GET, POST, PUT, etc.) [e2e-llm-inference-service] [e2e-llm-inference-service] :param url: [e2e-llm-inference-service] The URL to perform the request on. [e2e-llm-inference-service] [e2e-llm-inference-service] :param body: [e2e-llm-inference-service] Data to send in the request body, either :class:`str`, :class:`bytes`, [e2e-llm-inference-service] an iterable of :class:`str`/:class:`bytes`, or a file-like object. [e2e-llm-inference-service] [e2e-llm-inference-service] :param headers: [e2e-llm-inference-service] Dictionary of custom headers to send, such as User-Agent, [e2e-llm-inference-service] If-None-Match, etc. If None, pool headers are used. If provided, [e2e-llm-inference-service] these headers completely replace any pool-specific headers. [e2e-llm-inference-service] [e2e-llm-inference-service] :param retries: [e2e-llm-inference-service] Configure the number of retries to allow before raising a [e2e-llm-inference-service] :class:`~urllib3.exceptions.MaxRetryError` exception. [e2e-llm-inference-service] [e2e-llm-inference-service] If ``None`` (default) will retry 3 times, see ``Retry.DEFAULT``. Pass a [e2e-llm-inference-service] :class:`~urllib3.util.retry.Retry` object for fine-grained control [e2e-llm-inference-service] over different types of retries. [e2e-llm-inference-service] Pass an integer number to retry connection errors that many times, [e2e-llm-inference-service] but no other types of errors. Pass zero to never retry. [e2e-llm-inference-service] [e2e-llm-inference-service] If ``False``, then retries are disabled and any exception is raised [e2e-llm-inference-service] immediately. Also, instead of raising a MaxRetryError on redirects, [e2e-llm-inference-service] the redirect response will be returned. [e2e-llm-inference-service] [e2e-llm-inference-service] :type retries: :class:`~urllib3.util.retry.Retry`, False, or an int. [e2e-llm-inference-service] [e2e-llm-inference-service] :param redirect: [e2e-llm-inference-service] If True, automatically handle redirects (status codes 301, 302, [e2e-llm-inference-service] 303, 307, 308). Each redirect counts as a retry. Disabling retries [e2e-llm-inference-service] will disable redirect, too. [e2e-llm-inference-service] [e2e-llm-inference-service] :param assert_same_host: [e2e-llm-inference-service] If ``True``, will make sure that the host of the pool requests is [e2e-llm-inference-service] consistent else will raise HostChangedError. When ``False``, you can [e2e-llm-inference-service] use the pool on an HTTP proxy and request foreign hosts. [e2e-llm-inference-service] [e2e-llm-inference-service] :param timeout: [e2e-llm-inference-service] If specified, overrides the default timeout for this one [e2e-llm-inference-service] request. It may be a float (in seconds) or an instance of [e2e-llm-inference-service] :class:`urllib3.util.Timeout`. [e2e-llm-inference-service] [e2e-llm-inference-service] :param pool_timeout: [e2e-llm-inference-service] If set and the pool is set to block=True, then this method will [e2e-llm-inference-service] block for ``pool_timeout`` seconds and raise EmptyPoolError if no [e2e-llm-inference-service] connection is available within the time period. [e2e-llm-inference-service] [e2e-llm-inference-service] :param bool preload_content: [e2e-llm-inference-service] If True, the response's body will be preloaded into memory. [e2e-llm-inference-service] [e2e-llm-inference-service] :param bool decode_content: [e2e-llm-inference-service] If True, will attempt to decode the body based on the [e2e-llm-inference-service] 'content-encoding' header. [e2e-llm-inference-service] [e2e-llm-inference-service] :param release_conn: [e2e-llm-inference-service] If False, then the urlopen call will not release the connection [e2e-llm-inference-service] back into the pool once a response is received (but will release if [e2e-llm-inference-service] you read the entire contents of the response such as when [e2e-llm-inference-service] `preload_content=True`). This is useful if you're not preloading [e2e-llm-inference-service] the response's content immediately. You will need to call [e2e-llm-inference-service] ``r.release_conn()`` on the response ``r`` to return the connection [e2e-llm-inference-service] back into the pool. If None, it takes the value of ``preload_content`` [e2e-llm-inference-service] which defaults to ``True``. [e2e-llm-inference-service] [e2e-llm-inference-service] :param bool chunked: [e2e-llm-inference-service] If True, urllib3 will send the body using chunked transfer [e2e-llm-inference-service] encoding. Otherwise, urllib3 will send the body using the standard [e2e-llm-inference-service] content-length form. Defaults to False. [e2e-llm-inference-service] [e2e-llm-inference-service] :param int body_pos: [e2e-llm-inference-service] Position to seek to in file-like body in the event of a retry or [e2e-llm-inference-service] redirect. Typically this won't need to be set because urllib3 will [e2e-llm-inference-service] auto-populate the value when needed. [e2e-llm-inference-service] """ [e2e-llm-inference-service] parsed_url = parse_url(url) [e2e-llm-inference-service] destination_scheme = parsed_url.scheme [e2e-llm-inference-service] [e2e-llm-inference-service] if headers is None: [e2e-llm-inference-service] headers = self.headers [e2e-llm-inference-service] [e2e-llm-inference-service] if not isinstance(retries, Retry): [e2e-llm-inference-service] retries = Retry.from_int(retries, redirect=redirect, default=self.retries) [e2e-llm-inference-service] [e2e-llm-inference-service] if release_conn is None: [e2e-llm-inference-service] release_conn = preload_content [e2e-llm-inference-service] [e2e-llm-inference-service] # Check host [e2e-llm-inference-service] if assert_same_host and not self.is_same_host(url): [e2e-llm-inference-service] raise HostChangedError(self, url, retries) [e2e-llm-inference-service] [e2e-llm-inference-service] # Ensure that the URL we're connecting to is properly encoded [e2e-llm-inference-service] if url.startswith("/"): [e2e-llm-inference-service] url = to_str(_encode_target(url)) [e2e-llm-inference-service] else: [e2e-llm-inference-service] url = to_str(parsed_url.url) [e2e-llm-inference-service] [e2e-llm-inference-service] conn = None [e2e-llm-inference-service] [e2e-llm-inference-service] # Track whether `conn` needs to be released before [e2e-llm-inference-service] # returning/raising/recursing. Update this variable if necessary, and [e2e-llm-inference-service] # leave `release_conn` constant throughout the function. That way, if [e2e-llm-inference-service] # the function recurses, the original value of `release_conn` will be [e2e-llm-inference-service] # passed down into the recursive call, and its value will be respected. [e2e-llm-inference-service] # [e2e-llm-inference-service] # See issue #651 [1] for details. [e2e-llm-inference-service] # [e2e-llm-inference-service] # [1] [e2e-llm-inference-service] release_this_conn = release_conn [e2e-llm-inference-service] [e2e-llm-inference-service] http_tunnel_required = connection_requires_http_tunnel( [e2e-llm-inference-service] self.proxy, self.proxy_config, destination_scheme [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] # Merge the proxy headers. Only done when not using HTTP CONNECT. We [e2e-llm-inference-service] # have to copy the headers dict so we can safely change it without those [e2e-llm-inference-service] # changes being reflected in anyone else's copy. [e2e-llm-inference-service] if not http_tunnel_required: [e2e-llm-inference-service] headers = headers.copy() # type: ignore[attr-defined] [e2e-llm-inference-service] headers.update(self.proxy_headers) # type: ignore[union-attr] [e2e-llm-inference-service] [e2e-llm-inference-service] # Must keep the exception bound to a separate variable or else Python 3 [e2e-llm-inference-service] # complains about UnboundLocalError. [e2e-llm-inference-service] err = None [e2e-llm-inference-service] [e2e-llm-inference-service] # Keep track of whether we cleanly exited the except block. This [e2e-llm-inference-service] # ensures we do proper cleanup in finally. [e2e-llm-inference-service] clean_exit = False [e2e-llm-inference-service] [e2e-llm-inference-service] # Rewind body position, if needed. Record current position [e2e-llm-inference-service] # for future rewinds in the event of a redirect/retry. [e2e-llm-inference-service] body_pos = set_file_position(body, body_pos) [e2e-llm-inference-service] [e2e-llm-inference-service] try: [e2e-llm-inference-service] # Request a connection from the queue. [e2e-llm-inference-service] timeout_obj = self._get_timeout(timeout) [e2e-llm-inference-service] conn = self._get_conn(timeout=pool_timeout) [e2e-llm-inference-service] [e2e-llm-inference-service] conn.timeout = timeout_obj.connect_timeout # type: ignore[assignment] [e2e-llm-inference-service] [e2e-llm-inference-service] # Is this a closed/new connection that requires CONNECT tunnelling? [e2e-llm-inference-service] if self.proxy is not None and http_tunnel_required and conn.is_closed: [e2e-llm-inference-service] try: [e2e-llm-inference-service] self._prepare_proxy(conn) [e2e-llm-inference-service] except (BaseSSLError, OSError, SocketTimeout) as e: [e2e-llm-inference-service] self._raise_timeout( [e2e-llm-inference-service] err=e, url=self.proxy.url, timeout_value=conn.timeout [e2e-llm-inference-service] ) [e2e-llm-inference-service] raise [e2e-llm-inference-service] [e2e-llm-inference-service] # If we're going to release the connection in ``finally:``, then [e2e-llm-inference-service] # the response doesn't need to know about the connection. Otherwise [e2e-llm-inference-service] # it will also try to release it and we'll have a double-release [e2e-llm-inference-service] # mess. [e2e-llm-inference-service] response_conn = conn if not release_conn else None [e2e-llm-inference-service] [e2e-llm-inference-service] # Make the request on the HTTPConnection object [e2e-llm-inference-service] response = self._make_request( [e2e-llm-inference-service] conn, [e2e-llm-inference-service] method, [e2e-llm-inference-service] url, [e2e-llm-inference-service] timeout=timeout_obj, [e2e-llm-inference-service] body=body, [e2e-llm-inference-service] headers=headers, [e2e-llm-inference-service] chunked=chunked, [e2e-llm-inference-service] retries=retries, [e2e-llm-inference-service] response_conn=response_conn, [e2e-llm-inference-service] preload_content=preload_content, [e2e-llm-inference-service] decode_content=decode_content, [e2e-llm-inference-service] **response_kw, [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] # Everything went great! [e2e-llm-inference-service] clean_exit = True [e2e-llm-inference-service] [e2e-llm-inference-service] except EmptyPoolError: [e2e-llm-inference-service] # Didn't get a connection from the pool, no need to clean up [e2e-llm-inference-service] clean_exit = True [e2e-llm-inference-service] release_this_conn = False [e2e-llm-inference-service] raise [e2e-llm-inference-service] [e2e-llm-inference-service] except ( [e2e-llm-inference-service] TimeoutError, [e2e-llm-inference-service] HTTPException, [e2e-llm-inference-service] OSError, [e2e-llm-inference-service] ProtocolError, [e2e-llm-inference-service] BaseSSLError, [e2e-llm-inference-service] SSLError, [e2e-llm-inference-service] CertificateError, [e2e-llm-inference-service] ProxyError, [e2e-llm-inference-service] ) as e: [e2e-llm-inference-service] # Discard the connection for these exceptions. It will be [e2e-llm-inference-service] # replaced during the next _get_conn() call. [e2e-llm-inference-service] clean_exit = False [e2e-llm-inference-service] new_e: Exception = e [e2e-llm-inference-service] if isinstance(e, (BaseSSLError, CertificateError)): [e2e-llm-inference-service] new_e = SSLError(e) [e2e-llm-inference-service] if isinstance( [e2e-llm-inference-service] new_e, [e2e-llm-inference-service] ( [e2e-llm-inference-service] OSError, [e2e-llm-inference-service] NewConnectionError, [e2e-llm-inference-service] TimeoutError, [e2e-llm-inference-service] SSLError, [e2e-llm-inference-service] HTTPException, [e2e-llm-inference-service] ), [e2e-llm-inference-service] ) and (conn and conn.proxy and not conn.has_connected_to_proxy): [e2e-llm-inference-service] new_e = _wrap_proxy_error(new_e, conn.proxy.scheme) [e2e-llm-inference-service] elif isinstance(new_e, (OSError, HTTPException)): [e2e-llm-inference-service] new_e = ProtocolError("Connection aborted.", new_e) [e2e-llm-inference-service] [e2e-llm-inference-service] retries = retries.increment( [e2e-llm-inference-service] method, url, error=new_e, _pool=self, _stacktrace=sys.exc_info()[2] [e2e-llm-inference-service] ) [e2e-llm-inference-service] retries.sleep() [e2e-llm-inference-service] [e2e-llm-inference-service] # Keep track of the error for the retry warning. [e2e-llm-inference-service] err = e [e2e-llm-inference-service] [e2e-llm-inference-service] finally: [e2e-llm-inference-service] if not clean_exit: [e2e-llm-inference-service] # We hit some kind of exception, handled or otherwise. We need [e2e-llm-inference-service] # to throw the connection away unless explicitly told not to. [e2e-llm-inference-service] # Close the connection, set the variable to None, and make sure [e2e-llm-inference-service] # we put the None back in the pool to avoid leaking it. [e2e-llm-inference-service] if conn: [e2e-llm-inference-service] conn.close() [e2e-llm-inference-service] conn = None [e2e-llm-inference-service] release_this_conn = True [e2e-llm-inference-service] [e2e-llm-inference-service] if release_this_conn: [e2e-llm-inference-service] # Put the connection back to be reused. If the connection is [e2e-llm-inference-service] # expired then it will be None, which will get replaced with a [e2e-llm-inference-service] # fresh connection during _get_conn. [e2e-llm-inference-service] self._put_conn(conn) [e2e-llm-inference-service] [e2e-llm-inference-service] if not conn: [e2e-llm-inference-service] # Try again [e2e-llm-inference-service] log.warning( [e2e-llm-inference-service] "Retrying (%r) after connection broken by '%r': %s", retries, err, url [e2e-llm-inference-service] ) [e2e-llm-inference-service] > return self.urlopen( [e2e-llm-inference-service] method, [e2e-llm-inference-service] url, [e2e-llm-inference-service] body, [e2e-llm-inference-service] headers, [e2e-llm-inference-service] retries, [e2e-llm-inference-service] redirect, [e2e-llm-inference-service] assert_same_host, [e2e-llm-inference-service] timeout=timeout, [e2e-llm-inference-service] pool_timeout=pool_timeout, [e2e-llm-inference-service] release_conn=release_conn, [e2e-llm-inference-service] chunked=chunked, [e2e-llm-inference-service] body_pos=body_pos, [e2e-llm-inference-service] preload_content=preload_content, [e2e-llm-inference-service] decode_content=decode_content, [e2e-llm-inference-service] **response_kw, [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] ../../python/kserve/.venv/lib64/python3.11/site-packages/urllib3/connectionpool.py:871: [e2e-llm-inference-service] _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ [e2e-llm-inference-service] [e2e-llm-inference-service] self = [e2e-llm-inference-service] method = 'POST', url = '/v1/completions' [e2e-llm-inference-service] body = b'{"model": "publishers/kserve-ci-e2e-test/models/lora-adapter-1", "prompt": "KServe is a", "max_tokens": 20}' [e2e-llm-inference-service] headers = {'User-Agent': 'python-requests/2.32.3', 'Accept-Encoding': 'gzip, deflate', 'Accept': '*/*', 'Connection': 'keep-aliv...lication/json', 'X-Gateway-Model-Name': 'publishers/kserve-ci-e2e-test/models/lora-adapter-1', 'Content-Length': '107'} [e2e-llm-inference-service] retries = Retry(total=0, connect=None, read=None, redirect=None, status=None) [e2e-llm-inference-service] redirect = False, assert_same_host = False [e2e-llm-inference-service] timeout = Timeout(connect=60, read=60, total=None), pool_timeout = None [e2e-llm-inference-service] release_conn = False, chunked = False, body_pos = None, preload_content = False [e2e-llm-inference-service] decode_content = False, response_kw = {} [e2e-llm-inference-service] parsed_url = Url(scheme=None, auth=None, host=None, port=None, path='/v1/completions', query=None, fragment=None) [e2e-llm-inference-service] destination_scheme = None, conn = None, release_this_conn = True [e2e-llm-inference-service] http_tunnel_required = False [e2e-llm-inference-service] err = ReadTimeoutError("HTTPConnectionPool(host='ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com', port=80): Read timed out. (read timeout=60)") [e2e-llm-inference-service] clean_exit = False [e2e-llm-inference-service] [e2e-llm-inference-service] def urlopen( # type: ignore[override] [e2e-llm-inference-service] self, [e2e-llm-inference-service] method: str, [e2e-llm-inference-service] url: str, [e2e-llm-inference-service] body: _TYPE_BODY | None = None, [e2e-llm-inference-service] headers: typing.Mapping[str, str] | None = None, [e2e-llm-inference-service] retries: Retry | bool | int | None = None, [e2e-llm-inference-service] redirect: bool = True, [e2e-llm-inference-service] assert_same_host: bool = True, [e2e-llm-inference-service] timeout: _TYPE_TIMEOUT = _DEFAULT_TIMEOUT, [e2e-llm-inference-service] pool_timeout: int | None = None, [e2e-llm-inference-service] release_conn: bool | None = None, [e2e-llm-inference-service] chunked: bool = False, [e2e-llm-inference-service] body_pos: _TYPE_BODY_POSITION | None = None, [e2e-llm-inference-service] preload_content: bool = True, [e2e-llm-inference-service] decode_content: bool = True, [e2e-llm-inference-service] **response_kw: typing.Any, [e2e-llm-inference-service] ) -> BaseHTTPResponse: [e2e-llm-inference-service] """ [e2e-llm-inference-service] Get a connection from the pool and perform an HTTP request. This is the [e2e-llm-inference-service] lowest level call for making a request, so you'll need to specify all [e2e-llm-inference-service] the raw details. [e2e-llm-inference-service] [e2e-llm-inference-service] .. note:: [e2e-llm-inference-service] [e2e-llm-inference-service] More commonly, it's appropriate to use a convenience method [e2e-llm-inference-service] such as :meth:`request`. [e2e-llm-inference-service] [e2e-llm-inference-service] .. note:: [e2e-llm-inference-service] [e2e-llm-inference-service] `release_conn` will only behave as expected if [e2e-llm-inference-service] `preload_content=False` because we want to make [e2e-llm-inference-service] `preload_content=False` the default behaviour someday soon without [e2e-llm-inference-service] breaking backwards compatibility. [e2e-llm-inference-service] [e2e-llm-inference-service] :param method: [e2e-llm-inference-service] HTTP request method (such as GET, POST, PUT, etc.) [e2e-llm-inference-service] [e2e-llm-inference-service] :param url: [e2e-llm-inference-service] The URL to perform the request on. [e2e-llm-inference-service] [e2e-llm-inference-service] :param body: [e2e-llm-inference-service] Data to send in the request body, either :class:`str`, :class:`bytes`, [e2e-llm-inference-service] an iterable of :class:`str`/:class:`bytes`, or a file-like object. [e2e-llm-inference-service] [e2e-llm-inference-service] :param headers: [e2e-llm-inference-service] Dictionary of custom headers to send, such as User-Agent, [e2e-llm-inference-service] If-None-Match, etc. If None, pool headers are used. If provided, [e2e-llm-inference-service] these headers completely replace any pool-specific headers. [e2e-llm-inference-service] [e2e-llm-inference-service] :param retries: [e2e-llm-inference-service] Configure the number of retries to allow before raising a [e2e-llm-inference-service] :class:`~urllib3.exceptions.MaxRetryError` exception. [e2e-llm-inference-service] [e2e-llm-inference-service] If ``None`` (default) will retry 3 times, see ``Retry.DEFAULT``. Pass a [e2e-llm-inference-service] :class:`~urllib3.util.retry.Retry` object for fine-grained control [e2e-llm-inference-service] over different types of retries. [e2e-llm-inference-service] Pass an integer number to retry connection errors that many times, [e2e-llm-inference-service] but no other types of errors. Pass zero to never retry. [e2e-llm-inference-service] [e2e-llm-inference-service] If ``False``, then retries are disabled and any exception is raised [e2e-llm-inference-service] immediately. Also, instead of raising a MaxRetryError on redirects, [e2e-llm-inference-service] the redirect response will be returned. [e2e-llm-inference-service] [e2e-llm-inference-service] :type retries: :class:`~urllib3.util.retry.Retry`, False, or an int. [e2e-llm-inference-service] [e2e-llm-inference-service] :param redirect: [e2e-llm-inference-service] If True, automatically handle redirects (status codes 301, 302, [e2e-llm-inference-service] 303, 307, 308). Each redirect counts as a retry. Disabling retries [e2e-llm-inference-service] will disable redirect, too. [e2e-llm-inference-service] [e2e-llm-inference-service] :param assert_same_host: [e2e-llm-inference-service] If ``True``, will make sure that the host of the pool requests is [e2e-llm-inference-service] consistent else will raise HostChangedError. When ``False``, you can [e2e-llm-inference-service] use the pool on an HTTP proxy and request foreign hosts. [e2e-llm-inference-service] [e2e-llm-inference-service] :param timeout: [e2e-llm-inference-service] If specified, overrides the default timeout for this one [e2e-llm-inference-service] request. It may be a float (in seconds) or an instance of [e2e-llm-inference-service] :class:`urllib3.util.Timeout`. [e2e-llm-inference-service] [e2e-llm-inference-service] :param pool_timeout: [e2e-llm-inference-service] If set and the pool is set to block=True, then this method will [e2e-llm-inference-service] block for ``pool_timeout`` seconds and raise EmptyPoolError if no [e2e-llm-inference-service] connection is available within the time period. [e2e-llm-inference-service] [e2e-llm-inference-service] :param bool preload_content: [e2e-llm-inference-service] If True, the response's body will be preloaded into memory. [e2e-llm-inference-service] [e2e-llm-inference-service] :param bool decode_content: [e2e-llm-inference-service] If True, will attempt to decode the body based on the [e2e-llm-inference-service] 'content-encoding' header. [e2e-llm-inference-service] [e2e-llm-inference-service] :param release_conn: [e2e-llm-inference-service] If False, then the urlopen call will not release the connection [e2e-llm-inference-service] back into the pool once a response is received (but will release if [e2e-llm-inference-service] you read the entire contents of the response such as when [e2e-llm-inference-service] `preload_content=True`). This is useful if you're not preloading [e2e-llm-inference-service] the response's content immediately. You will need to call [e2e-llm-inference-service] ``r.release_conn()`` on the response ``r`` to return the connection [e2e-llm-inference-service] back into the pool. If None, it takes the value of ``preload_content`` [e2e-llm-inference-service] which defaults to ``True``. [e2e-llm-inference-service] [e2e-llm-inference-service] :param bool chunked: [e2e-llm-inference-service] If True, urllib3 will send the body using chunked transfer [e2e-llm-inference-service] encoding. Otherwise, urllib3 will send the body using the standard [e2e-llm-inference-service] content-length form. Defaults to False. [e2e-llm-inference-service] [e2e-llm-inference-service] :param int body_pos: [e2e-llm-inference-service] Position to seek to in file-like body in the event of a retry or [e2e-llm-inference-service] redirect. Typically this won't need to be set because urllib3 will [e2e-llm-inference-service] auto-populate the value when needed. [e2e-llm-inference-service] """ [e2e-llm-inference-service] parsed_url = parse_url(url) [e2e-llm-inference-service] destination_scheme = parsed_url.scheme [e2e-llm-inference-service] [e2e-llm-inference-service] if headers is None: [e2e-llm-inference-service] headers = self.headers [e2e-llm-inference-service] [e2e-llm-inference-service] if not isinstance(retries, Retry): [e2e-llm-inference-service] retries = Retry.from_int(retries, redirect=redirect, default=self.retries) [e2e-llm-inference-service] [e2e-llm-inference-service] if release_conn is None: [e2e-llm-inference-service] release_conn = preload_content [e2e-llm-inference-service] [e2e-llm-inference-service] # Check host [e2e-llm-inference-service] if assert_same_host and not self.is_same_host(url): [e2e-llm-inference-service] raise HostChangedError(self, url, retries) [e2e-llm-inference-service] [e2e-llm-inference-service] # Ensure that the URL we're connecting to is properly encoded [e2e-llm-inference-service] if url.startswith("/"): [e2e-llm-inference-service] url = to_str(_encode_target(url)) [e2e-llm-inference-service] else: [e2e-llm-inference-service] url = to_str(parsed_url.url) [e2e-llm-inference-service] [e2e-llm-inference-service] conn = None [e2e-llm-inference-service] [e2e-llm-inference-service] # Track whether `conn` needs to be released before [e2e-llm-inference-service] # returning/raising/recursing. Update this variable if necessary, and [e2e-llm-inference-service] # leave `release_conn` constant throughout the function. That way, if [e2e-llm-inference-service] # the function recurses, the original value of `release_conn` will be [e2e-llm-inference-service] # passed down into the recursive call, and its value will be respected. [e2e-llm-inference-service] # [e2e-llm-inference-service] # See issue #651 [1] for details. [e2e-llm-inference-service] # [e2e-llm-inference-service] # [1] [e2e-llm-inference-service] release_this_conn = release_conn [e2e-llm-inference-service] [e2e-llm-inference-service] http_tunnel_required = connection_requires_http_tunnel( [e2e-llm-inference-service] self.proxy, self.proxy_config, destination_scheme [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] # Merge the proxy headers. Only done when not using HTTP CONNECT. We [e2e-llm-inference-service] # have to copy the headers dict so we can safely change it without those [e2e-llm-inference-service] # changes being reflected in anyone else's copy. [e2e-llm-inference-service] if not http_tunnel_required: [e2e-llm-inference-service] headers = headers.copy() # type: ignore[attr-defined] [e2e-llm-inference-service] headers.update(self.proxy_headers) # type: ignore[union-attr] [e2e-llm-inference-service] [e2e-llm-inference-service] # Must keep the exception bound to a separate variable or else Python 3 [e2e-llm-inference-service] # complains about UnboundLocalError. [e2e-llm-inference-service] err = None [e2e-llm-inference-service] [e2e-llm-inference-service] # Keep track of whether we cleanly exited the except block. This [e2e-llm-inference-service] # ensures we do proper cleanup in finally. [e2e-llm-inference-service] clean_exit = False [e2e-llm-inference-service] [e2e-llm-inference-service] # Rewind body position, if needed. Record current position [e2e-llm-inference-service] # for future rewinds in the event of a redirect/retry. [e2e-llm-inference-service] body_pos = set_file_position(body, body_pos) [e2e-llm-inference-service] [e2e-llm-inference-service] try: [e2e-llm-inference-service] # Request a connection from the queue. [e2e-llm-inference-service] timeout_obj = self._get_timeout(timeout) [e2e-llm-inference-service] conn = self._get_conn(timeout=pool_timeout) [e2e-llm-inference-service] [e2e-llm-inference-service] conn.timeout = timeout_obj.connect_timeout # type: ignore[assignment] [e2e-llm-inference-service] [e2e-llm-inference-service] # Is this a closed/new connection that requires CONNECT tunnelling? [e2e-llm-inference-service] if self.proxy is not None and http_tunnel_required and conn.is_closed: [e2e-llm-inference-service] try: [e2e-llm-inference-service] self._prepare_proxy(conn) [e2e-llm-inference-service] except (BaseSSLError, OSError, SocketTimeout) as e: [e2e-llm-inference-service] self._raise_timeout( [e2e-llm-inference-service] err=e, url=self.proxy.url, timeout_value=conn.timeout [e2e-llm-inference-service] ) [e2e-llm-inference-service] raise [e2e-llm-inference-service] [e2e-llm-inference-service] # If we're going to release the connection in ``finally:``, then [e2e-llm-inference-service] # the response doesn't need to know about the connection. Otherwise [e2e-llm-inference-service] # it will also try to release it and we'll have a double-release [e2e-llm-inference-service] # mess. [e2e-llm-inference-service] response_conn = conn if not release_conn else None [e2e-llm-inference-service] [e2e-llm-inference-service] # Make the request on the HTTPConnection object [e2e-llm-inference-service] response = self._make_request( [e2e-llm-inference-service] conn, [e2e-llm-inference-service] method, [e2e-llm-inference-service] url, [e2e-llm-inference-service] timeout=timeout_obj, [e2e-llm-inference-service] body=body, [e2e-llm-inference-service] headers=headers, [e2e-llm-inference-service] chunked=chunked, [e2e-llm-inference-service] retries=retries, [e2e-llm-inference-service] response_conn=response_conn, [e2e-llm-inference-service] preload_content=preload_content, [e2e-llm-inference-service] decode_content=decode_content, [e2e-llm-inference-service] **response_kw, [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] # Everything went great! [e2e-llm-inference-service] clean_exit = True [e2e-llm-inference-service] [e2e-llm-inference-service] except EmptyPoolError: [e2e-llm-inference-service] # Didn't get a connection from the pool, no need to clean up [e2e-llm-inference-service] clean_exit = True [e2e-llm-inference-service] release_this_conn = False [e2e-llm-inference-service] raise [e2e-llm-inference-service] [e2e-llm-inference-service] except ( [e2e-llm-inference-service] TimeoutError, [e2e-llm-inference-service] HTTPException, [e2e-llm-inference-service] OSError, [e2e-llm-inference-service] ProtocolError, [e2e-llm-inference-service] BaseSSLError, [e2e-llm-inference-service] SSLError, [e2e-llm-inference-service] CertificateError, [e2e-llm-inference-service] ProxyError, [e2e-llm-inference-service] ) as e: [e2e-llm-inference-service] # Discard the connection for these exceptions. It will be [e2e-llm-inference-service] # replaced during the next _get_conn() call. [e2e-llm-inference-service] clean_exit = False [e2e-llm-inference-service] new_e: Exception = e [e2e-llm-inference-service] if isinstance(e, (BaseSSLError, CertificateError)): [e2e-llm-inference-service] new_e = SSLError(e) [e2e-llm-inference-service] if isinstance( [e2e-llm-inference-service] new_e, [e2e-llm-inference-service] ( [e2e-llm-inference-service] OSError, [e2e-llm-inference-service] NewConnectionError, [e2e-llm-inference-service] TimeoutError, [e2e-llm-inference-service] SSLError, [e2e-llm-inference-service] HTTPException, [e2e-llm-inference-service] ), [e2e-llm-inference-service] ) and (conn and conn.proxy and not conn.has_connected_to_proxy): [e2e-llm-inference-service] new_e = _wrap_proxy_error(new_e, conn.proxy.scheme) [e2e-llm-inference-service] elif isinstance(new_e, (OSError, HTTPException)): [e2e-llm-inference-service] new_e = ProtocolError("Connection aborted.", new_e) [e2e-llm-inference-service] [e2e-llm-inference-service] retries = retries.increment( [e2e-llm-inference-service] method, url, error=new_e, _pool=self, _stacktrace=sys.exc_info()[2] [e2e-llm-inference-service] ) [e2e-llm-inference-service] retries.sleep() [e2e-llm-inference-service] [e2e-llm-inference-service] # Keep track of the error for the retry warning. [e2e-llm-inference-service] err = e [e2e-llm-inference-service] [e2e-llm-inference-service] finally: [e2e-llm-inference-service] if not clean_exit: [e2e-llm-inference-service] # We hit some kind of exception, handled or otherwise. We need [e2e-llm-inference-service] # to throw the connection away unless explicitly told not to. [e2e-llm-inference-service] # Close the connection, set the variable to None, and make sure [e2e-llm-inference-service] # we put the None back in the pool to avoid leaking it. [e2e-llm-inference-service] if conn: [e2e-llm-inference-service] conn.close() [e2e-llm-inference-service] conn = None [e2e-llm-inference-service] release_this_conn = True [e2e-llm-inference-service] [e2e-llm-inference-service] if release_this_conn: [e2e-llm-inference-service] # Put the connection back to be reused. If the connection is [e2e-llm-inference-service] # expired then it will be None, which will get replaced with a [e2e-llm-inference-service] # fresh connection during _get_conn. [e2e-llm-inference-service] self._put_conn(conn) [e2e-llm-inference-service] [e2e-llm-inference-service] if not conn: [e2e-llm-inference-service] # Try again [e2e-llm-inference-service] log.warning( [e2e-llm-inference-service] "Retrying (%r) after connection broken by '%r': %s", retries, err, url [e2e-llm-inference-service] ) [e2e-llm-inference-service] > return self.urlopen( [e2e-llm-inference-service] method, [e2e-llm-inference-service] url, [e2e-llm-inference-service] body, [e2e-llm-inference-service] headers, [e2e-llm-inference-service] retries, [e2e-llm-inference-service] redirect, [e2e-llm-inference-service] assert_same_host, [e2e-llm-inference-service] timeout=timeout, [e2e-llm-inference-service] pool_timeout=pool_timeout, [e2e-llm-inference-service] release_conn=release_conn, [e2e-llm-inference-service] chunked=chunked, [e2e-llm-inference-service] body_pos=body_pos, [e2e-llm-inference-service] preload_content=preload_content, [e2e-llm-inference-service] decode_content=decode_content, [e2e-llm-inference-service] **response_kw, [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] ../../python/kserve/.venv/lib64/python3.11/site-packages/urllib3/connectionpool.py:871: [e2e-llm-inference-service] _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ [e2e-llm-inference-service] [e2e-llm-inference-service] self = [e2e-llm-inference-service] method = 'POST', url = '/v1/completions' [e2e-llm-inference-service] body = b'{"model": "publishers/kserve-ci-e2e-test/models/lora-adapter-1", "prompt": "KServe is a", "max_tokens": 20}' [e2e-llm-inference-service] headers = {'User-Agent': 'python-requests/2.32.3', 'Accept-Encoding': 'gzip, deflate', 'Accept': '*/*', 'Connection': 'keep-aliv...lication/json', 'X-Gateway-Model-Name': 'publishers/kserve-ci-e2e-test/models/lora-adapter-1', 'Content-Length': '107'} [e2e-llm-inference-service] retries = Retry(total=0, connect=None, read=None, redirect=None, status=None) [e2e-llm-inference-service] redirect = False, assert_same_host = False [e2e-llm-inference-service] timeout = Timeout(connect=60, read=60, total=None), pool_timeout = None [e2e-llm-inference-service] release_conn = False, chunked = False, body_pos = None, preload_content = False [e2e-llm-inference-service] decode_content = False, response_kw = {} [e2e-llm-inference-service] parsed_url = Url(scheme=None, auth=None, host=None, port=None, path='/v1/completions', query=None, fragment=None) [e2e-llm-inference-service] destination_scheme = None, conn = None, release_this_conn = True [e2e-llm-inference-service] http_tunnel_required = False, err = None, clean_exit = False [e2e-llm-inference-service] [e2e-llm-inference-service] def urlopen( # type: ignore[override] [e2e-llm-inference-service] self, [e2e-llm-inference-service] method: str, [e2e-llm-inference-service] url: str, [e2e-llm-inference-service] body: _TYPE_BODY | None = None, [e2e-llm-inference-service] headers: typing.Mapping[str, str] | None = None, [e2e-llm-inference-service] retries: Retry | bool | int | None = None, [e2e-llm-inference-service] redirect: bool = True, [e2e-llm-inference-service] assert_same_host: bool = True, [e2e-llm-inference-service] timeout: _TYPE_TIMEOUT = _DEFAULT_TIMEOUT, [e2e-llm-inference-service] pool_timeout: int | None = None, [e2e-llm-inference-service] release_conn: bool | None = None, [e2e-llm-inference-service] chunked: bool = False, [e2e-llm-inference-service] body_pos: _TYPE_BODY_POSITION | None = None, [e2e-llm-inference-service] preload_content: bool = True, [e2e-llm-inference-service] decode_content: bool = True, [e2e-llm-inference-service] **response_kw: typing.Any, [e2e-llm-inference-service] ) -> BaseHTTPResponse: [e2e-llm-inference-service] """ [e2e-llm-inference-service] Get a connection from the pool and perform an HTTP request. This is the [e2e-llm-inference-service] lowest level call for making a request, so you'll need to specify all [e2e-llm-inference-service] the raw details. [e2e-llm-inference-service] [e2e-llm-inference-service] .. note:: [e2e-llm-inference-service] [e2e-llm-inference-service] More commonly, it's appropriate to use a convenience method [e2e-llm-inference-service] such as :meth:`request`. [e2e-llm-inference-service] [e2e-llm-inference-service] .. note:: [e2e-llm-inference-service] [e2e-llm-inference-service] `release_conn` will only behave as expected if [e2e-llm-inference-service] `preload_content=False` because we want to make [e2e-llm-inference-service] `preload_content=False` the default behaviour someday soon without [e2e-llm-inference-service] breaking backwards compatibility. [e2e-llm-inference-service] [e2e-llm-inference-service] :param method: [e2e-llm-inference-service] HTTP request method (such as GET, POST, PUT, etc.) [e2e-llm-inference-service] [e2e-llm-inference-service] :param url: [e2e-llm-inference-service] The URL to perform the request on. [e2e-llm-inference-service] [e2e-llm-inference-service] :param body: [e2e-llm-inference-service] Data to send in the request body, either :class:`str`, :class:`bytes`, [e2e-llm-inference-service] an iterable of :class:`str`/:class:`bytes`, or a file-like object. [e2e-llm-inference-service] [e2e-llm-inference-service] :param headers: [e2e-llm-inference-service] Dictionary of custom headers to send, such as User-Agent, [e2e-llm-inference-service] If-None-Match, etc. If None, pool headers are used. If provided, [e2e-llm-inference-service] these headers completely replace any pool-specific headers. [e2e-llm-inference-service] [e2e-llm-inference-service] :param retries: [e2e-llm-inference-service] Configure the number of retries to allow before raising a [e2e-llm-inference-service] :class:`~urllib3.exceptions.MaxRetryError` exception. [e2e-llm-inference-service] [e2e-llm-inference-service] If ``None`` (default) will retry 3 times, see ``Retry.DEFAULT``. Pass a [e2e-llm-inference-service] :class:`~urllib3.util.retry.Retry` object for fine-grained control [e2e-llm-inference-service] over different types of retries. [e2e-llm-inference-service] Pass an integer number to retry connection errors that many times, [e2e-llm-inference-service] but no other types of errors. Pass zero to never retry. [e2e-llm-inference-service] [e2e-llm-inference-service] If ``False``, then retries are disabled and any exception is raised [e2e-llm-inference-service] immediately. Also, instead of raising a MaxRetryError on redirects, [e2e-llm-inference-service] the redirect response will be returned. [e2e-llm-inference-service] [e2e-llm-inference-service] :type retries: :class:`~urllib3.util.retry.Retry`, False, or an int. [e2e-llm-inference-service] [e2e-llm-inference-service] :param redirect: [e2e-llm-inference-service] If True, automatically handle redirects (status codes 301, 302, [e2e-llm-inference-service] 303, 307, 308). Each redirect counts as a retry. Disabling retries [e2e-llm-inference-service] will disable redirect, too. [e2e-llm-inference-service] [e2e-llm-inference-service] :param assert_same_host: [e2e-llm-inference-service] If ``True``, will make sure that the host of the pool requests is [e2e-llm-inference-service] consistent else will raise HostChangedError. When ``False``, you can [e2e-llm-inference-service] use the pool on an HTTP proxy and request foreign hosts. [e2e-llm-inference-service] [e2e-llm-inference-service] :param timeout: [e2e-llm-inference-service] If specified, overrides the default timeout for this one [e2e-llm-inference-service] request. It may be a float (in seconds) or an instance of [e2e-llm-inference-service] :class:`urllib3.util.Timeout`. [e2e-llm-inference-service] [e2e-llm-inference-service] :param pool_timeout: [e2e-llm-inference-service] If set and the pool is set to block=True, then this method will [e2e-llm-inference-service] block for ``pool_timeout`` seconds and raise EmptyPoolError if no [e2e-llm-inference-service] connection is available within the time period. [e2e-llm-inference-service] [e2e-llm-inference-service] :param bool preload_content: [e2e-llm-inference-service] If True, the response's body will be preloaded into memory. [e2e-llm-inference-service] [e2e-llm-inference-service] :param bool decode_content: [e2e-llm-inference-service] If True, will attempt to decode the body based on the [e2e-llm-inference-service] 'content-encoding' header. [e2e-llm-inference-service] [e2e-llm-inference-service] :param release_conn: [e2e-llm-inference-service] If False, then the urlopen call will not release the connection [e2e-llm-inference-service] back into the pool once a response is received (but will release if [e2e-llm-inference-service] you read the entire contents of the response such as when [e2e-llm-inference-service] `preload_content=True`). This is useful if you're not preloading [e2e-llm-inference-service] the response's content immediately. You will need to call [e2e-llm-inference-service] ``r.release_conn()`` on the response ``r`` to return the connection [e2e-llm-inference-service] back into the pool. If None, it takes the value of ``preload_content`` [e2e-llm-inference-service] which defaults to ``True``. [e2e-llm-inference-service] [e2e-llm-inference-service] :param bool chunked: [e2e-llm-inference-service] If True, urllib3 will send the body using chunked transfer [e2e-llm-inference-service] encoding. Otherwise, urllib3 will send the body using the standard [e2e-llm-inference-service] content-length form. Defaults to False. [e2e-llm-inference-service] [e2e-llm-inference-service] :param int body_pos: [e2e-llm-inference-service] Position to seek to in file-like body in the event of a retry or [e2e-llm-inference-service] redirect. Typically this won't need to be set because urllib3 will [e2e-llm-inference-service] auto-populate the value when needed. [e2e-llm-inference-service] """ [e2e-llm-inference-service] parsed_url = parse_url(url) [e2e-llm-inference-service] destination_scheme = parsed_url.scheme [e2e-llm-inference-service] [e2e-llm-inference-service] if headers is None: [e2e-llm-inference-service] headers = self.headers [e2e-llm-inference-service] [e2e-llm-inference-service] if not isinstance(retries, Retry): [e2e-llm-inference-service] retries = Retry.from_int(retries, redirect=redirect, default=self.retries) [e2e-llm-inference-service] [e2e-llm-inference-service] if release_conn is None: [e2e-llm-inference-service] release_conn = preload_content [e2e-llm-inference-service] [e2e-llm-inference-service] # Check host [e2e-llm-inference-service] if assert_same_host and not self.is_same_host(url): [e2e-llm-inference-service] raise HostChangedError(self, url, retries) [e2e-llm-inference-service] [e2e-llm-inference-service] # Ensure that the URL we're connecting to is properly encoded [e2e-llm-inference-service] if url.startswith("/"): [e2e-llm-inference-service] url = to_str(_encode_target(url)) [e2e-llm-inference-service] else: [e2e-llm-inference-service] url = to_str(parsed_url.url) [e2e-llm-inference-service] [e2e-llm-inference-service] conn = None [e2e-llm-inference-service] [e2e-llm-inference-service] # Track whether `conn` needs to be released before [e2e-llm-inference-service] # returning/raising/recursing. Update this variable if necessary, and [e2e-llm-inference-service] # leave `release_conn` constant throughout the function. That way, if [e2e-llm-inference-service] # the function recurses, the original value of `release_conn` will be [e2e-llm-inference-service] # passed down into the recursive call, and its value will be respected. [e2e-llm-inference-service] # [e2e-llm-inference-service] # See issue #651 [1] for details. [e2e-llm-inference-service] # [e2e-llm-inference-service] # [1] [e2e-llm-inference-service] release_this_conn = release_conn [e2e-llm-inference-service] [e2e-llm-inference-service] http_tunnel_required = connection_requires_http_tunnel( [e2e-llm-inference-service] self.proxy, self.proxy_config, destination_scheme [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] # Merge the proxy headers. Only done when not using HTTP CONNECT. We [e2e-llm-inference-service] # have to copy the headers dict so we can safely change it without those [e2e-llm-inference-service] # changes being reflected in anyone else's copy. [e2e-llm-inference-service] if not http_tunnel_required: [e2e-llm-inference-service] headers = headers.copy() # type: ignore[attr-defined] [e2e-llm-inference-service] headers.update(self.proxy_headers) # type: ignore[union-attr] [e2e-llm-inference-service] [e2e-llm-inference-service] # Must keep the exception bound to a separate variable or else Python 3 [e2e-llm-inference-service] # complains about UnboundLocalError. [e2e-llm-inference-service] err = None [e2e-llm-inference-service] [e2e-llm-inference-service] # Keep track of whether we cleanly exited the except block. This [e2e-llm-inference-service] # ensures we do proper cleanup in finally. [e2e-llm-inference-service] clean_exit = False [e2e-llm-inference-service] [e2e-llm-inference-service] # Rewind body position, if needed. Record current position [e2e-llm-inference-service] # for future rewinds in the event of a redirect/retry. [e2e-llm-inference-service] body_pos = set_file_position(body, body_pos) [e2e-llm-inference-service] [e2e-llm-inference-service] try: [e2e-llm-inference-service] # Request a connection from the queue. [e2e-llm-inference-service] timeout_obj = self._get_timeout(timeout) [e2e-llm-inference-service] conn = self._get_conn(timeout=pool_timeout) [e2e-llm-inference-service] [e2e-llm-inference-service] conn.timeout = timeout_obj.connect_timeout # type: ignore[assignment] [e2e-llm-inference-service] [e2e-llm-inference-service] # Is this a closed/new connection that requires CONNECT tunnelling? [e2e-llm-inference-service] if self.proxy is not None and http_tunnel_required and conn.is_closed: [e2e-llm-inference-service] try: [e2e-llm-inference-service] self._prepare_proxy(conn) [e2e-llm-inference-service] except (BaseSSLError, OSError, SocketTimeout) as e: [e2e-llm-inference-service] self._raise_timeout( [e2e-llm-inference-service] err=e, url=self.proxy.url, timeout_value=conn.timeout [e2e-llm-inference-service] ) [e2e-llm-inference-service] raise [e2e-llm-inference-service] [e2e-llm-inference-service] # If we're going to release the connection in ``finally:``, then [e2e-llm-inference-service] # the response doesn't need to know about the connection. Otherwise [e2e-llm-inference-service] # it will also try to release it and we'll have a double-release [e2e-llm-inference-service] # mess. [e2e-llm-inference-service] response_conn = conn if not release_conn else None [e2e-llm-inference-service] [e2e-llm-inference-service] # Make the request on the HTTPConnection object [e2e-llm-inference-service] response = self._make_request( [e2e-llm-inference-service] conn, [e2e-llm-inference-service] method, [e2e-llm-inference-service] url, [e2e-llm-inference-service] timeout=timeout_obj, [e2e-llm-inference-service] body=body, [e2e-llm-inference-service] headers=headers, [e2e-llm-inference-service] chunked=chunked, [e2e-llm-inference-service] retries=retries, [e2e-llm-inference-service] response_conn=response_conn, [e2e-llm-inference-service] preload_content=preload_content, [e2e-llm-inference-service] decode_content=decode_content, [e2e-llm-inference-service] **response_kw, [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] # Everything went great! [e2e-llm-inference-service] clean_exit = True [e2e-llm-inference-service] [e2e-llm-inference-service] except EmptyPoolError: [e2e-llm-inference-service] # Didn't get a connection from the pool, no need to clean up [e2e-llm-inference-service] clean_exit = True [e2e-llm-inference-service] release_this_conn = False [e2e-llm-inference-service] raise [e2e-llm-inference-service] [e2e-llm-inference-service] except ( [e2e-llm-inference-service] TimeoutError, [e2e-llm-inference-service] HTTPException, [e2e-llm-inference-service] OSError, [e2e-llm-inference-service] ProtocolError, [e2e-llm-inference-service] BaseSSLError, [e2e-llm-inference-service] SSLError, [e2e-llm-inference-service] CertificateError, [e2e-llm-inference-service] ProxyError, [e2e-llm-inference-service] ) as e: [e2e-llm-inference-service] # Discard the connection for these exceptions. It will be [e2e-llm-inference-service] # replaced during the next _get_conn() call. [e2e-llm-inference-service] clean_exit = False [e2e-llm-inference-service] new_e: Exception = e [e2e-llm-inference-service] if isinstance(e, (BaseSSLError, CertificateError)): [e2e-llm-inference-service] new_e = SSLError(e) [e2e-llm-inference-service] if isinstance( [e2e-llm-inference-service] new_e, [e2e-llm-inference-service] ( [e2e-llm-inference-service] OSError, [e2e-llm-inference-service] NewConnectionError, [e2e-llm-inference-service] TimeoutError, [e2e-llm-inference-service] SSLError, [e2e-llm-inference-service] HTTPException, [e2e-llm-inference-service] ), [e2e-llm-inference-service] ) and (conn and conn.proxy and not conn.has_connected_to_proxy): [e2e-llm-inference-service] new_e = _wrap_proxy_error(new_e, conn.proxy.scheme) [e2e-llm-inference-service] elif isinstance(new_e, (OSError, HTTPException)): [e2e-llm-inference-service] new_e = ProtocolError("Connection aborted.", new_e) [e2e-llm-inference-service] [e2e-llm-inference-service] > retries = retries.increment( [e2e-llm-inference-service] method, url, error=new_e, _pool=self, _stacktrace=sys.exc_info()[2] [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] ../../python/kserve/.venv/lib64/python3.11/site-packages/urllib3/connectionpool.py:841: [e2e-llm-inference-service] _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ [e2e-llm-inference-service] [e2e-llm-inference-service] self = Retry(total=0, connect=None, read=None, redirect=None, status=None) [e2e-llm-inference-service] method = 'POST', url = '/v1/completions', response = None [e2e-llm-inference-service] error = ReadTimeoutError("HTTPConnectionPool(host='ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com', port=80): Read timed out. (read timeout=60)") [e2e-llm-inference-service] _pool = [e2e-llm-inference-service] _stacktrace = [e2e-llm-inference-service] [e2e-llm-inference-service] def increment( [e2e-llm-inference-service] self, [e2e-llm-inference-service] method: str | None = None, [e2e-llm-inference-service] url: str | None = None, [e2e-llm-inference-service] response: BaseHTTPResponse | None = None, [e2e-llm-inference-service] error: Exception | None = None, [e2e-llm-inference-service] _pool: ConnectionPool | None = None, [e2e-llm-inference-service] _stacktrace: TracebackType | None = None, [e2e-llm-inference-service] ) -> Self: [e2e-llm-inference-service] """Return a new Retry object with incremented retry counters. [e2e-llm-inference-service] [e2e-llm-inference-service] :param response: A response object, or None, if the server did not [e2e-llm-inference-service] return a response. [e2e-llm-inference-service] :type response: :class:`~urllib3.response.BaseHTTPResponse` [e2e-llm-inference-service] :param Exception error: An error encountered during the request, or [e2e-llm-inference-service] None if the response was received successfully. [e2e-llm-inference-service] [e2e-llm-inference-service] :return: A new ``Retry`` object. [e2e-llm-inference-service] """ [e2e-llm-inference-service] if self.total is False and error: [e2e-llm-inference-service] # Disabled, indicate to re-raise the error. [e2e-llm-inference-service] raise reraise(type(error), error, _stacktrace) [e2e-llm-inference-service] [e2e-llm-inference-service] total = self.total [e2e-llm-inference-service] if total is not None: [e2e-llm-inference-service] total -= 1 [e2e-llm-inference-service] [e2e-llm-inference-service] connect = self.connect [e2e-llm-inference-service] read = self.read [e2e-llm-inference-service] redirect = self.redirect [e2e-llm-inference-service] status_count = self.status [e2e-llm-inference-service] other = self.other [e2e-llm-inference-service] cause = "unknown" [e2e-llm-inference-service] status = None [e2e-llm-inference-service] redirect_location = None [e2e-llm-inference-service] [e2e-llm-inference-service] if error and self._is_connection_error(error): [e2e-llm-inference-service] # Connect retry? [e2e-llm-inference-service] if connect is False: [e2e-llm-inference-service] raise reraise(type(error), error, _stacktrace) [e2e-llm-inference-service] elif connect is not None: [e2e-llm-inference-service] connect -= 1 [e2e-llm-inference-service] [e2e-llm-inference-service] elif error and self._is_read_error(error): [e2e-llm-inference-service] # Read retry? [e2e-llm-inference-service] if read is False or method is None or not self._is_method_retryable(method): [e2e-llm-inference-service] raise reraise(type(error), error, _stacktrace) [e2e-llm-inference-service] elif read is not None: [e2e-llm-inference-service] read -= 1 [e2e-llm-inference-service] [e2e-llm-inference-service] elif error: [e2e-llm-inference-service] # Other retry? [e2e-llm-inference-service] if other is not None: [e2e-llm-inference-service] other -= 1 [e2e-llm-inference-service] [e2e-llm-inference-service] elif response and response.get_redirect_location(): [e2e-llm-inference-service] # Redirect retry? [e2e-llm-inference-service] if redirect is not None: [e2e-llm-inference-service] redirect -= 1 [e2e-llm-inference-service] cause = "too many redirects" [e2e-llm-inference-service] response_redirect_location = response.get_redirect_location() [e2e-llm-inference-service] if response_redirect_location: [e2e-llm-inference-service] redirect_location = response_redirect_location [e2e-llm-inference-service] status = response.status [e2e-llm-inference-service] [e2e-llm-inference-service] else: [e2e-llm-inference-service] # Incrementing because of a server error like a 500 in [e2e-llm-inference-service] # status_forcelist and the given method is in the allowed_methods [e2e-llm-inference-service] cause = ResponseError.GENERIC_ERROR [e2e-llm-inference-service] if response and response.status: [e2e-llm-inference-service] if status_count is not None: [e2e-llm-inference-service] status_count -= 1 [e2e-llm-inference-service] cause = ResponseError.SPECIFIC_ERROR.format(status_code=response.status) [e2e-llm-inference-service] status = response.status [e2e-llm-inference-service] [e2e-llm-inference-service] history = self.history + ( [e2e-llm-inference-service] RequestHistory(method, url, error, status, redirect_location), [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] new_retry = self.new( [e2e-llm-inference-service] total=total, [e2e-llm-inference-service] connect=connect, [e2e-llm-inference-service] read=read, [e2e-llm-inference-service] redirect=redirect, [e2e-llm-inference-service] status=status_count, [e2e-llm-inference-service] other=other, [e2e-llm-inference-service] history=history, [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] if new_retry.is_exhausted(): [e2e-llm-inference-service] reason = error or ResponseError(cause) [e2e-llm-inference-service] > raise MaxRetryError(_pool, url, reason) from reason # type: ignore[arg-type] [e2e-llm-inference-service] E urllib3.exceptions.MaxRetryError: HTTPConnectionPool(host='ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com', port=80): Max retries exceeded with url: /v1/completions (Caused by ReadTimeoutError("HTTPConnectionPool(host='ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com', port=80): Read timed out. (read timeout=60)")) [e2e-llm-inference-service] [e2e-llm-inference-service] ../../python/kserve/.venv/lib64/python3.11/site-packages/urllib3/util/retry.py:519: MaxRetryError [e2e-llm-inference-service] [e2e-llm-inference-service] During handling of the above exception, another exception occurred: [e2e-llm-inference-service] [e2e-llm-inference-service] def get_successful_response(): [e2e-llm-inference-service] try: [e2e-llm-inference-service] if test_case.url_getter: [e2e-llm-inference-service] service_url = test_case.url_getter(kserve_client, test_case.llm_service) [e2e-llm-inference-service] else: [e2e-llm-inference-service] service_url = get_llm_service_url(kserve_client, test_case.llm_service) [e2e-llm-inference-service] except Exception as e: [e2e-llm-inference-service] raise AssertionError(f"❌ Failed to get service URL: {e}") from e [e2e-llm-inference-service] [e2e-llm-inference-service] model_url = service_url + test_case.endpoint [e2e-llm-inference-service] [e2e-llm-inference-service] headers = {"Content-Type": "application/json"} [e2e-llm-inference-service] if extra_headers: [e2e-llm-inference-service] headers.update(extra_headers) [e2e-llm-inference-service] [e2e-llm-inference-service] if test_case.payload_formatter is not None: [e2e-llm-inference-service] test_payload = test_case.payload_formatter(test_case) [e2e-llm-inference-service] elif test_case.prompt is not None: [e2e-llm-inference-service] test_payload = { [e2e-llm-inference-service] "model": test_case.model_name [e2e-llm-inference-service] if not extra_headers or MODEL_ROUTING_HEADER not in extra_headers [e2e-llm-inference-service] else extra_headers[MODEL_ROUTING_HEADER], [e2e-llm-inference-service] "prompt": test_case.prompt, [e2e-llm-inference-service] "max_tokens": test_case.max_tokens, [e2e-llm-inference-service] } [e2e-llm-inference-service] else: [e2e-llm-inference-service] test_payload = None [e2e-llm-inference-service] [e2e-llm-inference-service] logger.info(f"Calling LLM service at {model_url} with payload {test_payload}") [e2e-llm-inference-service] try: [e2e-llm-inference-service] if test_payload is not None: [e2e-llm-inference-service] > response = post_with_retry( [e2e-llm-inference-service] model_url, [e2e-llm-inference-service] headers=headers, [e2e-llm-inference-service] json_data=test_payload, [e2e-llm-inference-service] timeout=test_case.response_timeout, [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] llmisvc/test_llm_inference_service.py:1095: [e2e-llm-inference-service] _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ [e2e-llm-inference-service] [e2e-llm-inference-service] url = 'http://ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com/v1/completions' [e2e-llm-inference-service] [e2e-llm-inference-service] def post_with_retry( [e2e-llm-inference-service] url: str, [e2e-llm-inference-service] *, [e2e-llm-inference-service] headers: Dict = None, [e2e-llm-inference-service] json_data: Union[Dict, List] = None, [e2e-llm-inference-service] data: Union[str, bytes] = None, [e2e-llm-inference-service] stream: bool = False, [e2e-llm-inference-service] timeout: float = None, [e2e-llm-inference-service] total_retries: int = DEFAULT_RETRY_TOTAL, [e2e-llm-inference-service] backoff_factor: float = DEFAULT_RETRY_BACKOFF_FACTOR, [e2e-llm-inference-service] retry_status_codes=DEFAULT_RETRY_STATUS_CODES, [e2e-llm-inference-service] ) -> requests.Response: [e2e-llm-inference-service] """ [e2e-llm-inference-service] Send POST request with retries for transient HTTP and network failures. [e2e-llm-inference-service] """ [e2e-llm-inference-service] if json_data is not None and data is not None: [e2e-llm-inference-service] raise ValueError("Only one of json_data or data can be provided.") [e2e-llm-inference-service] [e2e-llm-inference-service] with _retry_session( [e2e-llm-inference-service] ["POST"], total_retries, backoff_factor, retry_status_codes [e2e-llm-inference-service] ) as session: [e2e-llm-inference-service] > return session.post( [e2e-llm-inference-service] url, [e2e-llm-inference-service] json=json_data, [e2e-llm-inference-service] data=data, [e2e-llm-inference-service] headers=headers, [e2e-llm-inference-service] stream=stream, [e2e-llm-inference-service] timeout=timeout, [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] common/http_retry.py:70: [e2e-llm-inference-service] _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ [e2e-llm-inference-service] [e2e-llm-inference-service] self = [e2e-llm-inference-service] url = 'http://ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com/v1/completions' [e2e-llm-inference-service] data = None [e2e-llm-inference-service] json = {'max_tokens': 20, 'model': 'publishers/kserve-ci-e2e-test/models/lora-adapter-1', 'prompt': 'KServe is a'} [e2e-llm-inference-service] kwargs = {'headers': {'Content-Type': 'application/json', 'X-Gateway-Model-Name': 'publishers/kserve-ci-e2e-test/models/lora-adapter-1'}, 'stream': False, 'timeout': 60} [e2e-llm-inference-service] [e2e-llm-inference-service] def post(self, url, data=None, json=None, **kwargs): [e2e-llm-inference-service] r"""Sends a POST request. Returns :class:`Response` object. [e2e-llm-inference-service] [e2e-llm-inference-service] :param url: URL for the new :class:`Request` object. [e2e-llm-inference-service] :param data: (optional) Dictionary, list of tuples, bytes, or file-like [e2e-llm-inference-service] object to send in the body of the :class:`Request`. [e2e-llm-inference-service] :param json: (optional) json to send in the body of the :class:`Request`. [e2e-llm-inference-service] :param \*\*kwargs: Optional arguments that ``request`` takes. [e2e-llm-inference-service] :rtype: requests.Response [e2e-llm-inference-service] """ [e2e-llm-inference-service] [e2e-llm-inference-service] > return self.request("POST", url, data=data, json=json, **kwargs) [e2e-llm-inference-service] [e2e-llm-inference-service] ../../python/kserve/.venv/lib64/python3.11/site-packages/requests/sessions.py:637: [e2e-llm-inference-service] _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ [e2e-llm-inference-service] [e2e-llm-inference-service] self = , method = 'POST' [e2e-llm-inference-service] url = 'http://ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com/v1/completions' [e2e-llm-inference-service] params = None, data = None [e2e-llm-inference-service] headers = {'Content-Type': 'application/json', 'X-Gateway-Model-Name': 'publishers/kserve-ci-e2e-test/models/lora-adapter-1'} [e2e-llm-inference-service] cookies = None, files = None, auth = None, timeout = 60, allow_redirects = True [e2e-llm-inference-service] proxies = {}, hooks = None, stream = False, verify = None, cert = None [e2e-llm-inference-service] json = {'max_tokens': 20, 'model': 'publishers/kserve-ci-e2e-test/models/lora-adapter-1', 'prompt': 'KServe is a'} [e2e-llm-inference-service] [e2e-llm-inference-service] def request( [e2e-llm-inference-service] self, [e2e-llm-inference-service] method, [e2e-llm-inference-service] url, [e2e-llm-inference-service] params=None, [e2e-llm-inference-service] data=None, [e2e-llm-inference-service] headers=None, [e2e-llm-inference-service] cookies=None, [e2e-llm-inference-service] files=None, [e2e-llm-inference-service] auth=None, [e2e-llm-inference-service] timeout=None, [e2e-llm-inference-service] allow_redirects=True, [e2e-llm-inference-service] proxies=None, [e2e-llm-inference-service] hooks=None, [e2e-llm-inference-service] stream=None, [e2e-llm-inference-service] verify=None, [e2e-llm-inference-service] cert=None, [e2e-llm-inference-service] json=None, [e2e-llm-inference-service] ): [e2e-llm-inference-service] """Constructs a :class:`Request `, prepares it and sends it. [e2e-llm-inference-service] Returns :class:`Response ` object. [e2e-llm-inference-service] [e2e-llm-inference-service] :param method: method for the new :class:`Request` object. [e2e-llm-inference-service] :param url: URL for the new :class:`Request` object. [e2e-llm-inference-service] :param params: (optional) Dictionary or bytes to be sent in the query [e2e-llm-inference-service] string for the :class:`Request`. [e2e-llm-inference-service] :param data: (optional) Dictionary, list of tuples, bytes, or file-like [e2e-llm-inference-service] object to send in the body of the :class:`Request`. [e2e-llm-inference-service] :param json: (optional) json to send in the body of the [e2e-llm-inference-service] :class:`Request`. [e2e-llm-inference-service] :param headers: (optional) Dictionary of HTTP Headers to send with the [e2e-llm-inference-service] :class:`Request`. [e2e-llm-inference-service] :param cookies: (optional) Dict or CookieJar object to send with the [e2e-llm-inference-service] :class:`Request`. [e2e-llm-inference-service] :param files: (optional) Dictionary of ``'filename': file-like-objects`` [e2e-llm-inference-service] for multipart encoding upload. [e2e-llm-inference-service] :param auth: (optional) Auth tuple or callable to enable [e2e-llm-inference-service] Basic/Digest/Custom HTTP Auth. [e2e-llm-inference-service] :param timeout: (optional) How long to wait for the server to send [e2e-llm-inference-service] data before giving up, as a float, or a :ref:`(connect timeout, [e2e-llm-inference-service] read timeout) ` tuple. [e2e-llm-inference-service] :type timeout: float or tuple [e2e-llm-inference-service] :param allow_redirects: (optional) Set to True by default. [e2e-llm-inference-service] :type allow_redirects: bool [e2e-llm-inference-service] :param proxies: (optional) Dictionary mapping protocol or protocol and [e2e-llm-inference-service] hostname to the URL of the proxy. [e2e-llm-inference-service] :param hooks: (optional) Dictionary mapping hook name to one event or [e2e-llm-inference-service] list of events, event must be callable. [e2e-llm-inference-service] :param stream: (optional) whether to immediately download the response [e2e-llm-inference-service] content. Defaults to ``False``. [e2e-llm-inference-service] :param verify: (optional) Either a boolean, in which case it controls whether we verify [e2e-llm-inference-service] the server's TLS certificate, or a string, in which case it must be a path [e2e-llm-inference-service] to a CA bundle to use. Defaults to ``True``. When set to [e2e-llm-inference-service] ``False``, requests will accept any TLS certificate presented by [e2e-llm-inference-service] the server, and will ignore hostname mismatches and/or expired [e2e-llm-inference-service] certificates, which will make your application vulnerable to [e2e-llm-inference-service] man-in-the-middle (MitM) attacks. Setting verify to ``False`` [e2e-llm-inference-service] may be useful during local development or testing. [e2e-llm-inference-service] :param cert: (optional) if String, path to ssl client cert file (.pem). [e2e-llm-inference-service] If Tuple, ('cert', 'key') pair. [e2e-llm-inference-service] :rtype: requests.Response [e2e-llm-inference-service] """ [e2e-llm-inference-service] # Create the Request. [e2e-llm-inference-service] req = Request( [e2e-llm-inference-service] method=method.upper(), [e2e-llm-inference-service] url=url, [e2e-llm-inference-service] headers=headers, [e2e-llm-inference-service] files=files, [e2e-llm-inference-service] data=data or {}, [e2e-llm-inference-service] json=json, [e2e-llm-inference-service] params=params or {}, [e2e-llm-inference-service] auth=auth, [e2e-llm-inference-service] cookies=cookies, [e2e-llm-inference-service] hooks=hooks, [e2e-llm-inference-service] ) [e2e-llm-inference-service] prep = self.prepare_request(req) [e2e-llm-inference-service] [e2e-llm-inference-service] proxies = proxies or {} [e2e-llm-inference-service] [e2e-llm-inference-service] settings = self.merge_environment_settings( [e2e-llm-inference-service] prep.url, proxies, stream, verify, cert [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] # Send the request. [e2e-llm-inference-service] send_kwargs = { [e2e-llm-inference-service] "timeout": timeout, [e2e-llm-inference-service] "allow_redirects": allow_redirects, [e2e-llm-inference-service] } [e2e-llm-inference-service] send_kwargs.update(settings) [e2e-llm-inference-service] > resp = self.send(prep, **send_kwargs) [e2e-llm-inference-service] [e2e-llm-inference-service] ../../python/kserve/.venv/lib64/python3.11/site-packages/requests/sessions.py:589: [e2e-llm-inference-service] _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ [e2e-llm-inference-service] [e2e-llm-inference-service] self = [e2e-llm-inference-service] request = [e2e-llm-inference-service] kwargs = {'cert': None, 'proxies': OrderedDict(), 'stream': False, 'timeout': 60, ...} [e2e-llm-inference-service] allow_redirects = True, stream = False, hooks = {'response': []} [e2e-llm-inference-service] adapter = [e2e-llm-inference-service] start = 1783434646.1162572 [e2e-llm-inference-service] [e2e-llm-inference-service] def send(self, request, **kwargs): [e2e-llm-inference-service] """Send a given PreparedRequest. [e2e-llm-inference-service] [e2e-llm-inference-service] :rtype: requests.Response [e2e-llm-inference-service] """ [e2e-llm-inference-service] # Set defaults that the hooks can utilize to ensure they always have [e2e-llm-inference-service] # the correct parameters to reproduce the previous request. [e2e-llm-inference-service] kwargs.setdefault("stream", self.stream) [e2e-llm-inference-service] kwargs.setdefault("verify", self.verify) [e2e-llm-inference-service] kwargs.setdefault("cert", self.cert) [e2e-llm-inference-service] if "proxies" not in kwargs: [e2e-llm-inference-service] kwargs["proxies"] = resolve_proxies(request, self.proxies, self.trust_env) [e2e-llm-inference-service] [e2e-llm-inference-service] # It's possible that users might accidentally send a Request object. [e2e-llm-inference-service] # Guard against that specific failure case. [e2e-llm-inference-service] if isinstance(request, Request): [e2e-llm-inference-service] raise ValueError("You can only send PreparedRequests.") [e2e-llm-inference-service] [e2e-llm-inference-service] # Set up variables needed for resolve_redirects and dispatching of hooks [e2e-llm-inference-service] allow_redirects = kwargs.pop("allow_redirects", True) [e2e-llm-inference-service] stream = kwargs.get("stream") [e2e-llm-inference-service] hooks = request.hooks [e2e-llm-inference-service] [e2e-llm-inference-service] # Get the appropriate adapter to use [e2e-llm-inference-service] adapter = self.get_adapter(url=request.url) [e2e-llm-inference-service] [e2e-llm-inference-service] # Start time (approximately) of the request [e2e-llm-inference-service] start = preferred_clock() [e2e-llm-inference-service] [e2e-llm-inference-service] # Send the request [e2e-llm-inference-service] > r = adapter.send(request, **kwargs) [e2e-llm-inference-service] [e2e-llm-inference-service] ../../python/kserve/.venv/lib64/python3.11/site-packages/requests/sessions.py:703: [e2e-llm-inference-service] _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ [e2e-llm-inference-service] [e2e-llm-inference-service] self = [e2e-llm-inference-service] request = , stream = False [e2e-llm-inference-service] timeout = Timeout(connect=60, read=60, total=None), verify = '/tmp/ca.crt' [e2e-llm-inference-service] cert = None, proxies = OrderedDict() [e2e-llm-inference-service] [e2e-llm-inference-service] def send( [e2e-llm-inference-service] self, request, stream=False, timeout=None, verify=True, cert=None, proxies=None [e2e-llm-inference-service] ): [e2e-llm-inference-service] """Sends PreparedRequest object. Returns Response object. [e2e-llm-inference-service] [e2e-llm-inference-service] :param request: The :class:`PreparedRequest ` being sent. [e2e-llm-inference-service] :param stream: (optional) Whether to stream the request content. [e2e-llm-inference-service] :param timeout: (optional) How long to wait for the server to send [e2e-llm-inference-service] data before giving up, as a float, or a :ref:`(connect timeout, [e2e-llm-inference-service] read timeout) ` tuple. [e2e-llm-inference-service] :type timeout: float or tuple or urllib3 Timeout object [e2e-llm-inference-service] :param verify: (optional) Either a boolean, in which case it controls whether [e2e-llm-inference-service] we verify the server's TLS certificate, or a string, in which case it [e2e-llm-inference-service] must be a path to a CA bundle to use [e2e-llm-inference-service] :param cert: (optional) Any user-provided SSL certificate to be trusted. [e2e-llm-inference-service] :param proxies: (optional) The proxies dictionary to apply to the request. [e2e-llm-inference-service] :rtype: requests.Response [e2e-llm-inference-service] """ [e2e-llm-inference-service] [e2e-llm-inference-service] try: [e2e-llm-inference-service] conn = self.get_connection_with_tls_context( [e2e-llm-inference-service] request, verify, proxies=proxies, cert=cert [e2e-llm-inference-service] ) [e2e-llm-inference-service] except LocationValueError as e: [e2e-llm-inference-service] raise InvalidURL(e, request=request) [e2e-llm-inference-service] [e2e-llm-inference-service] self.cert_verify(conn, request.url, verify, cert) [e2e-llm-inference-service] url = self.request_url(request, proxies) [e2e-llm-inference-service] self.add_headers( [e2e-llm-inference-service] request, [e2e-llm-inference-service] stream=stream, [e2e-llm-inference-service] timeout=timeout, [e2e-llm-inference-service] verify=verify, [e2e-llm-inference-service] cert=cert, [e2e-llm-inference-service] proxies=proxies, [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] chunked = not (request.body is None or "Content-Length" in request.headers) [e2e-llm-inference-service] [e2e-llm-inference-service] if isinstance(timeout, tuple): [e2e-llm-inference-service] try: [e2e-llm-inference-service] connect, read = timeout [e2e-llm-inference-service] timeout = TimeoutSauce(connect=connect, read=read) [e2e-llm-inference-service] except ValueError: [e2e-llm-inference-service] raise ValueError( [e2e-llm-inference-service] f"Invalid timeout {timeout}. Pass a (connect, read) timeout tuple, " [e2e-llm-inference-service] f"or a single float to set both timeouts to the same value." [e2e-llm-inference-service] ) [e2e-llm-inference-service] elif isinstance(timeout, TimeoutSauce): [e2e-llm-inference-service] pass [e2e-llm-inference-service] else: [e2e-llm-inference-service] timeout = TimeoutSauce(connect=timeout, read=timeout) [e2e-llm-inference-service] [e2e-llm-inference-service] try: [e2e-llm-inference-service] resp = conn.urlopen( [e2e-llm-inference-service] method=request.method, [e2e-llm-inference-service] url=url, [e2e-llm-inference-service] body=request.body, [e2e-llm-inference-service] headers=request.headers, [e2e-llm-inference-service] redirect=False, [e2e-llm-inference-service] assert_same_host=False, [e2e-llm-inference-service] preload_content=False, [e2e-llm-inference-service] decode_content=False, [e2e-llm-inference-service] retries=self.max_retries, [e2e-llm-inference-service] timeout=timeout, [e2e-llm-inference-service] chunked=chunked, [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] except (ProtocolError, OSError) as err: [e2e-llm-inference-service] raise ConnectionError(err, request=request) [e2e-llm-inference-service] [e2e-llm-inference-service] except MaxRetryError as e: [e2e-llm-inference-service] if isinstance(e.reason, ConnectTimeoutError): [e2e-llm-inference-service] # TODO: Remove this in 3.0.0: see #2811 [e2e-llm-inference-service] if not isinstance(e.reason, NewConnectionError): [e2e-llm-inference-service] raise ConnectTimeout(e, request=request) [e2e-llm-inference-service] [e2e-llm-inference-service] if isinstance(e.reason, ResponseError): [e2e-llm-inference-service] raise RetryError(e, request=request) [e2e-llm-inference-service] [e2e-llm-inference-service] if isinstance(e.reason, _ProxyError): [e2e-llm-inference-service] raise ProxyError(e, request=request) [e2e-llm-inference-service] [e2e-llm-inference-service] if isinstance(e.reason, _SSLError): [e2e-llm-inference-service] # This branch is for urllib3 v1.22 and later. [e2e-llm-inference-service] raise SSLError(e, request=request) [e2e-llm-inference-service] [e2e-llm-inference-service] > raise ConnectionError(e, request=request) [e2e-llm-inference-service] E requests.exceptions.ConnectionError: HTTPConnectionPool(host='ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com', port=80): Max retries exceeded with url: /v1/completions (Caused by ReadTimeoutError("HTTPConnectionPool(host='ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com', port=80): Read timed out. (read timeout=60)")) [e2e-llm-inference-service] [e2e-llm-inference-service] ../../python/kserve/.venv/lib64/python3.11/site-packages/requests/adapters.py:700: ConnectionError [e2e-llm-inference-service] [e2e-llm-inference-service] The above exception was the direct cause of the following exception: [e2e-llm-inference-service] [e2e-llm-inference-service] test_case = TestCase(base_refs=['router-managed', 'workload-single-cpu', 'model-fb-opt-125m-with-lora-hf'], prompt='KServe is a', ...opt-125m-with-lora-hf-a7886ead'}]}, [e2e-llm-inference-service] 'status': None}, model_name='publishers/kserve-ci-e2e-test/models/lora-adapter-1') [e2e-llm-inference-service] [e2e-llm-inference-service] @pytest.mark.llminferenceservice [e2e-llm-inference-service] @pytest.mark.asyncio(loop_scope="session") [e2e-llm-inference-service] @pytest.mark.parametrize( [e2e-llm-inference-service] "test_case", [e2e-llm-inference-service] [ [e2e-llm-inference-service] pytest.param( [e2e-llm-inference-service] TestCase( [e2e-llm-inference-service] base_refs=[ [e2e-llm-inference-service] "router-with-gateway-ref", [e2e-llm-inference-service] "router-with-managed-route", [e2e-llm-inference-service] "model-fb-opt-125m", [e2e-llm-inference-service] "workload-llmd-simulator", [e2e-llm-inference-service] ], [e2e-llm-inference-service] endpoint="/v1/completions", [e2e-llm-inference-service] prompt="KServe is a", [e2e-llm-inference-service] payload_formatter=completions_payload, [e2e-llm-inference-service] response_assertion=create_response_assertion(with_field="choices"), [e2e-llm-inference-service] expected_gateway=ROUTER_GATEWAYS[0], [e2e-llm-inference-service] before_test=[ [e2e-llm-inference-service] lambda: create_router_resources( [e2e-llm-inference-service] gateways=[ROUTER_GATEWAYS[0]], [e2e-llm-inference-service] ) [e2e-llm-inference-service] ], [e2e-llm-inference-service] ), [e2e-llm-inference-service] marks=[ [e2e-llm-inference-service] pytest.mark.cluster_cpu, [e2e-llm-inference-service] pytest.mark.cluster_single_node, [e2e-llm-inference-service] pytest.mark.llmd_simulator, [e2e-llm-inference-service] pytest.mark.custom_gateway, [e2e-llm-inference-service] ], [e2e-llm-inference-service] ), [e2e-llm-inference-service] pytest.param( [e2e-llm-inference-service] TestCase( [e2e-llm-inference-service] base_refs=[ [e2e-llm-inference-service] "router-managed", [e2e-llm-inference-service] "workload-single-cpu", [e2e-llm-inference-service] "model-fb-opt-125m", [e2e-llm-inference-service] ], [e2e-llm-inference-service] prompt="KServe is a", [e2e-llm-inference-service] payload_formatter=completions_payload, [e2e-llm-inference-service] response_assertion=assert_200_with_choices, [e2e-llm-inference-service] ), [e2e-llm-inference-service] marks=[pytest.mark.cluster_cpu, pytest.mark.cluster_single_node], [e2e-llm-inference-service] ), [e2e-llm-inference-service] pytest.param( [e2e-llm-inference-service] TestCase( [e2e-llm-inference-service] base_refs=[ [e2e-llm-inference-service] "router-custom-route-timeout", [e2e-llm-inference-service] "scheduler-managed", [e2e-llm-inference-service] "workload-single-cpu", [e2e-llm-inference-service] "model-fb-opt-125m", [e2e-llm-inference-service] ], [e2e-llm-inference-service] prompt="KServe is a", [e2e-llm-inference-service] service_name="custom-route-timeout-test", [e2e-llm-inference-service] ), [e2e-llm-inference-service] marks=[pytest.mark.cluster_cpu, pytest.mark.cluster_single_node], [e2e-llm-inference-service] ), [e2e-llm-inference-service] pytest.param( [e2e-llm-inference-service] TestCase( [e2e-llm-inference-service] base_refs=[ [e2e-llm-inference-service] "router-with-refs", [e2e-llm-inference-service] "scheduler-managed", [e2e-llm-inference-service] "workload-single-cpu", [e2e-llm-inference-service] "model-fb-opt-125m", [e2e-llm-inference-service] ], [e2e-llm-inference-service] prompt="KServe is a", [e2e-llm-inference-service] service_name="router-with-refs-test", [e2e-llm-inference-service] expected_gateway=ROUTER_GATEWAYS[0], [e2e-llm-inference-service] before_test=[ [e2e-llm-inference-service] lambda: create_router_resources( [e2e-llm-inference-service] gateways=[ROUTER_GATEWAYS[0]], [e2e-llm-inference-service] routes=[ROUTER_ROUTES[0], ROUTER_ROUTES[1]], [e2e-llm-inference-service] ) [e2e-llm-inference-service] ], [e2e-llm-inference-service] ), [e2e-llm-inference-service] marks=[ [e2e-llm-inference-service] pytest.mark.cluster_cpu, [e2e-llm-inference-service] pytest.mark.cluster_single_node, [e2e-llm-inference-service] pytest.mark.custom_gateway, [e2e-llm-inference-service] ], [e2e-llm-inference-service] ), [e2e-llm-inference-service] pytest.param( [e2e-llm-inference-service] TestCase( [e2e-llm-inference-service] base_refs=["router-managed", "workload-pd-cpu", "model-fb-opt-125m"], [e2e-llm-inference-service] prompt="You are an expert in Kubernetes-native machine learning serving platforms, with deep knowledge of the KServe project. " [e2e-llm-inference-service] "Explain the challenges of serving large-scale models, GPU scheduling, and how KServe integrates with capabilities like multi-model serving. " [e2e-llm-inference-service] "Provide a detailed comparison with open source alternatives, focusing on operational trade-offs.", [e2e-llm-inference-service] response_assertion=assert_200_with_choices, [e2e-llm-inference-service] ), [e2e-llm-inference-service] marks=[pytest.mark.cluster_cpu, pytest.mark.cluster_single_node], [e2e-llm-inference-service] ), [e2e-llm-inference-service] pytest.param( [e2e-llm-inference-service] TestCase( [e2e-llm-inference-service] base_refs=[ [e2e-llm-inference-service] "router-custom-route-timeout-pd", [e2e-llm-inference-service] "scheduler-managed", [e2e-llm-inference-service] "workload-pd-cpu", [e2e-llm-inference-service] "model-fb-opt-125m", [e2e-llm-inference-service] ], [e2e-llm-inference-service] prompt="You are an expert in Kubernetes-native machine learning serving platforms, with deep knowledge of the KServe project. " [e2e-llm-inference-service] "Explain the challenges of serving large-scale models, GPU scheduling, and how KServe integrates with capabilities like multi-model serving. " [e2e-llm-inference-service] "Provide a detailed comparison with open source alternatives, focusing on operational trade-offs.", [e2e-llm-inference-service] service_name="custom-route-timeout-pd-test", [e2e-llm-inference-service] response_assertion=assert_200_with_choices, [e2e-llm-inference-service] ), [e2e-llm-inference-service] marks=[pytest.mark.cluster_cpu, pytest.mark.cluster_single_node], [e2e-llm-inference-service] ), [e2e-llm-inference-service] pytest.param( [e2e-llm-inference-service] TestCase( [e2e-llm-inference-service] base_refs=[ [e2e-llm-inference-service] "router-with-refs-pd", [e2e-llm-inference-service] "scheduler-managed", [e2e-llm-inference-service] "workload-pd-cpu", [e2e-llm-inference-service] "model-fb-opt-125m", [e2e-llm-inference-service] ], [e2e-llm-inference-service] prompt="You are an expert in Kubernetes-native machine learning serving platforms, with deep knowledge of the KServe project. " [e2e-llm-inference-service] "Explain the challenges of serving large-scale models, GPU scheduling, and how KServe integrates with capabilities like multi-model serving. " [e2e-llm-inference-service] "Provide a detailed comparison with open source alternatives, focusing on operational trade-offs.", [e2e-llm-inference-service] service_name="router-with-refs-pd-test", [e2e-llm-inference-service] response_assertion=assert_200_with_choices, [e2e-llm-inference-service] expected_gateway=ROUTER_GATEWAYS[1], [e2e-llm-inference-service] before_test=[ [e2e-llm-inference-service] lambda: create_router_resources( [e2e-llm-inference-service] gateways=[ROUTER_GATEWAYS[1]], [e2e-llm-inference-service] routes=[ROUTER_ROUTES[2], ROUTER_ROUTES[3]], [e2e-llm-inference-service] ) [e2e-llm-inference-service] ], [e2e-llm-inference-service] ), [e2e-llm-inference-service] marks=[ [e2e-llm-inference-service] pytest.mark.cluster_cpu, [e2e-llm-inference-service] pytest.mark.cluster_single_node, [e2e-llm-inference-service] pytest.mark.custom_gateway, [e2e-llm-inference-service] ], [e2e-llm-inference-service] ), [e2e-llm-inference-service] pytest.param( [e2e-llm-inference-service] TestCase( [e2e-llm-inference-service] base_refs=[ [e2e-llm-inference-service] "router-managed", [e2e-llm-inference-service] "workload-dp-ep-gpu", [e2e-llm-inference-service] "workload-dp-ep-prefill-gpu", [e2e-llm-inference-service] "model-deepseek-v2-lite", [e2e-llm-inference-service] ], [e2e-llm-inference-service] prompt="Delve into the multifaceted implications of a fully disaggregated cloud architecture, specifically " [e2e-llm-inference-service] "where the compute plane (P) and the data plane (D) are independently deployed and managed for a " [e2e-llm-inference-service] "geographically distributed, high-throughput, low-latency microservices ecosystem. Beyond the " [e2e-llm-inference-service] "fundamental challenges of network latency and data consistency, elaborate on the advanced " [e2e-llm-inference-service] "considerations and trade-offs inherent in such a setup: 1. Network Architecture and Protocols: " [e2e-llm-inference-service] "How would the network fabric and underlying protocols (e.g., RDMA, custom transport layers) need to " [e2e-llm-inference-service] "evolve to support optimal performance and minimize inter-plane communication overhead, especially for " [e2e-llm-inference-service] "synchronous operations? Discuss the role of network programmability (e.g., SDN, P4) in dynamically " [e2e-llm-inference-service] "optimizing routing and traffic flow between P and D. 2. Advanced Data Consistency and Durability: " [e2e-llm-inference-service] "Explore sophisticated data consistency models (e.g., causal consistency, strong eventual consistency) " [e2e-llm-inference-service] "and their applicability in balancing performance and data integrity across a globally distributed data plane. " [e2e-llm-inference-service] "Detail strategies for ensuring data durability and fault tolerance, including multi-region replication, " [e2e-llm-inference-service] "intelligent partitioning, and recovery mechanisms in the event of partial or full plane failures. " [e2e-llm-inference-service] "3. Dynamic Resource Orchestration and Cost Optimization: Analyze how an orchestration layer would intelligently " [e2e-llm-inference-service] "manage the independent scaling of compute (P) and data (D) resources, considering fluctuating workloads, " [e2e-llm-inference-service] "cost efficiency, and performance targets (e.g., using predictive analytics for resource provisioning). " [e2e-llm-inference-service] "Discuss mechanisms for dynamically reallocating compute nodes to different data partitions based on " [e2e-llm-inference-service] "workload patterns and data locality, potentially involving live migration strategies. " [e2e-llm-inference-service] "4. Security and Compliance in a Distributed Landscape: Address the enhanced security perimeter " [e2e-llm-inference-service] "challenges, including securing communication channels between P and D (encryption in transit, mutual TLS), " [e2e-llm-inference-service] "fine-grained access control to data at rest and in motion, and identity management across disaggregated " [e2e-llm-inference-service] "components. Discuss how such an architecture impacts compliance with regulatory frameworks (e.g., GDPR, HIPAA) " [e2e-llm-inference-service] "concerning data sovereignty, privacy, and auditability. 5. Operational Complexity and Observability: " [e2e-llm-inference-service] "Examine the increased complexity in monitoring, logging, and tracing across highly decoupled compute and " [e2e-llm-inference-service] "data planes. What specialized tooling and practices (e.g., distributed tracing with OpenTelemetry, advanced AIOps) " [e2e-llm-inference-service] "would be essential? How would incident response and troubleshooting differ in this disaggregated environment " [e2e-llm-inference-service] "compared to traditional integrated systems? Consider the challenges of pinpointing root causes across " [e2e-llm-inference-service] "independent failures. 6. Real-world Applicability and Future Trends: Identify specific industries " [e2e-llm-inference-service] "or use cases (e.g., high-frequency trading, IoT edge processing, large language model inference) " [e2e-llm-inference-service] "where the benefits of P/D disaggregation would strongly outweigh its complexities. " [e2e-llm-inference-service] "Conclude by speculating on emerging technologies or paradigms (e.g., serverless compute functions " [e2e-llm-inference-service] "directly interacting with object storage, in-memory disaggregation) that could further drive or " [e2e-llm-inference-service] "transform P/D disaggregation in cloud computing.", [e2e-llm-inference-service] max_tokens=2000, [e2e-llm-inference-service] ), [e2e-llm-inference-service] marks=[ [e2e-llm-inference-service] pytest.mark.cluster_gpu, [e2e-llm-inference-service] pytest.mark.cluster_nvidia, [e2e-llm-inference-service] pytest.mark.cluster_nvidia_roce, [e2e-llm-inference-service] ], [e2e-llm-inference-service] ), [e2e-llm-inference-service] pytest.param( [e2e-llm-inference-service] TestCase( [e2e-llm-inference-service] base_refs=[ [e2e-llm-inference-service] "router-no-scheduler", [e2e-llm-inference-service] "workload-single-cpu", [e2e-llm-inference-service] "model-fb-opt-125m", [e2e-llm-inference-service] ], [e2e-llm-inference-service] prompt="What is KServe?", [e2e-llm-inference-service] ), [e2e-llm-inference-service] marks=[ [e2e-llm-inference-service] pytest.mark.cluster_cpu, [e2e-llm-inference-service] pytest.mark.cluster_single_node, [e2e-llm-inference-service] pytest.mark.no_scheduler, [e2e-llm-inference-service] ], [e2e-llm-inference-service] ), [e2e-llm-inference-service] pytest.param( [e2e-llm-inference-service] TestCase( [e2e-llm-inference-service] base_refs=[ [e2e-llm-inference-service] "router-managed", [e2e-llm-inference-service] "workload-simulated-dp-ep-cpu", [e2e-llm-inference-service] "model-fb-opt-125m", [e2e-llm-inference-service] ], [e2e-llm-inference-service] prompt="This test simulates DP+EP that can run on CPU, the idea is to test the LWS-based deployment, " [e2e-llm-inference-service] "but without the resources requirements for DP+EP (GPUs and ROCe/IB).", [e2e-llm-inference-service] ), [e2e-llm-inference-service] marks=[pytest.mark.cluster_cpu, pytest.mark.cluster_multi_node], [e2e-llm-inference-service] ), [e2e-llm-inference-service] # Scheduler config tests [e2e-llm-inference-service] pytest.param( [e2e-llm-inference-service] TestCase( [e2e-llm-inference-service] base_refs=[ [e2e-llm-inference-service] "router-managed", [e2e-llm-inference-service] "scheduler-with-inline-config", [e2e-llm-inference-service] "workload-llmd-simulator", [e2e-llm-inference-service] ], [e2e-llm-inference-service] prompt="KServe is a", [e2e-llm-inference-service] service_name="scheduler-inline-config-test", [e2e-llm-inference-service] ), [e2e-llm-inference-service] marks=[pytest.mark.cluster_cpu, pytest.mark.cluster_single_node], [e2e-llm-inference-service] ), [e2e-llm-inference-service] # Chat completions endpoint coverage [e2e-llm-inference-service] pytest.param( [e2e-llm-inference-service] TestCase( [e2e-llm-inference-service] base_refs=[ [e2e-llm-inference-service] "router-managed", [e2e-llm-inference-service] "workload-llmd-simulator", [e2e-llm-inference-service] "model-qwen2.5-0.5b", [e2e-llm-inference-service] ], [e2e-llm-inference-service] model_name="Qwen/Qwen2.5-0.5B-Instruct", [e2e-llm-inference-service] endpoint="/v1/chat/completions", [e2e-llm-inference-service] prompt="What is KServe?", [e2e-llm-inference-service] payload_formatter=chat_completions_payload, [e2e-llm-inference-service] response_assertion=create_response_assertion(with_field="choices"), [e2e-llm-inference-service] ), [e2e-llm-inference-service] marks=[ [e2e-llm-inference-service] pytest.mark.cluster_cpu, [e2e-llm-inference-service] pytest.mark.cluster_single_node, [e2e-llm-inference-service] pytest.mark.llmd_simulator, [e2e-llm-inference-service] ], [e2e-llm-inference-service] ), [e2e-llm-inference-service] pytest.param( [e2e-llm-inference-service] TestCase( [e2e-llm-inference-service] base_refs=[ [e2e-llm-inference-service] "router-managed", [e2e-llm-inference-service] "scheduler-with-configmap-ref", [e2e-llm-inference-service] "workload-llmd-simulator", [e2e-llm-inference-service] ], [e2e-llm-inference-service] prompt="KServe is a", [e2e-llm-inference-service] service_name="scheduler-configmap-ref-test", [e2e-llm-inference-service] before_test=[create_scheduler_configmap], [e2e-llm-inference-service] after_test=[delete_scheduler_configmap], [e2e-llm-inference-service] ), [e2e-llm-inference-service] marks=[pytest.mark.cluster_cpu, pytest.mark.cluster_single_node], [e2e-llm-inference-service] ), [e2e-llm-inference-service] pytest.param( [e2e-llm-inference-service] TestCase( [e2e-llm-inference-service] base_refs=[ [e2e-llm-inference-service] "router-managed", [e2e-llm-inference-service] "scheduler-with-replicas", [e2e-llm-inference-service] "workload-llmd-simulator", [e2e-llm-inference-service] ], [e2e-llm-inference-service] prompt="KServe is a", [e2e-llm-inference-service] service_name="scheduler-ha-replicas-test", [e2e-llm-inference-service] ), [e2e-llm-inference-service] marks=[pytest.mark.cluster_cpu, pytest.mark.cluster_single_node], [e2e-llm-inference-service] ), [e2e-llm-inference-service] pytest.param( [e2e-llm-inference-service] TestCase( [e2e-llm-inference-service] base_refs=[ [e2e-llm-inference-service] "router-managed", [e2e-llm-inference-service] "scheduler-with-custom-template", [e2e-llm-inference-service] "workload-llmd-simulator", [e2e-llm-inference-service] ], [e2e-llm-inference-service] prompt="KServe is a", [e2e-llm-inference-service] service_name="scheduler-custom-template-test", [e2e-llm-inference-service] ), [e2e-llm-inference-service] marks=[pytest.mark.cluster_cpu, pytest.mark.cluster_single_node], [e2e-llm-inference-service] ), [e2e-llm-inference-service] # Scheduler v0.6 → v0.7 migration tests. [e2e-llm-inference-service] # Deploy v0.6-style configs and verify the controller migrates them [e2e-llm-inference-service] # so the v0.7 scheduler boots successfully. [e2e-llm-inference-service] pytest.param( [e2e-llm-inference-service] TestCase( [e2e-llm-inference-service] base_refs=[ [e2e-llm-inference-service] "router-managed", [e2e-llm-inference-service] "scheduler-v06-pd-config-migration", [e2e-llm-inference-service] "workload-llmd-simulator-pd", [e2e-llm-inference-service] ], [e2e-llm-inference-service] prompt="KServe is a", [e2e-llm-inference-service] service_name="scheduler-v06-pd-migration-test", [e2e-llm-inference-service] response_assertion=assert_200_with_choices, [e2e-llm-inference-service] ), [e2e-llm-inference-service] marks=[ [e2e-llm-inference-service] pytest.mark.cluster_cpu, [e2e-llm-inference-service] pytest.mark.cluster_single_node, [e2e-llm-inference-service] pytest.mark.llmd_simulator, [e2e-llm-inference-service] ], [e2e-llm-inference-service] ), [e2e-llm-inference-service] pytest.param( [e2e-llm-inference-service] TestCase( [e2e-llm-inference-service] base_refs=[ [e2e-llm-inference-service] "router-managed", [e2e-llm-inference-service] "scheduler-v06-nonzero-threshold-migration", [e2e-llm-inference-service] "workload-llmd-simulator-pd", [e2e-llm-inference-service] ], [e2e-llm-inference-service] prompt="KServe is a", [e2e-llm-inference-service] service_name="scheduler-v06-threshold-migration-test", [e2e-llm-inference-service] response_assertion=assert_200_with_choices, [e2e-llm-inference-service] ), [e2e-llm-inference-service] marks=[ [e2e-llm-inference-service] pytest.mark.cluster_cpu, [e2e-llm-inference-service] pytest.mark.cluster_single_node, [e2e-llm-inference-service] pytest.mark.llmd_simulator, [e2e-llm-inference-service] ], [e2e-llm-inference-service] ), [e2e-llm-inference-service] # Precise prefix KV cache routing test [e2e-llm-inference-service] pytest.param( [e2e-llm-inference-service] TestCase( [e2e-llm-inference-service] base_refs=[ [e2e-llm-inference-service] "router-managed", [e2e-llm-inference-service] "scheduler-with-precise-prefix-cache-inline-config", [e2e-llm-inference-service] "workload-llmd-simulator-kvcache", [e2e-llm-inference-service] ], [e2e-llm-inference-service] prompt="KServe is a", [e2e-llm-inference-service] service_name="precise-prefix-cache-test", [e2e-llm-inference-service] ), [e2e-llm-inference-service] marks=[ [e2e-llm-inference-service] pytest.mark.cluster_cpu, [e2e-llm-inference-service] pytest.mark.cluster_single_node, [e2e-llm-inference-service] pytest.mark.llmd_simulator, [e2e-llm-inference-service] ], [e2e-llm-inference-service] ), [e2e-llm-inference-service] # Models endpoint coverage [e2e-llm-inference-service] pytest.param( [e2e-llm-inference-service] TestCase( [e2e-llm-inference-service] base_refs=[ [e2e-llm-inference-service] "router-managed", [e2e-llm-inference-service] "workload-llmd-simulator", [e2e-llm-inference-service] ], [e2e-llm-inference-service] endpoint="/v1/models", [e2e-llm-inference-service] response_assertion=create_response_assertion(with_field="data"), [e2e-llm-inference-service] ), [e2e-llm-inference-service] marks=[ [e2e-llm-inference-service] pytest.mark.cluster_cpu, [e2e-llm-inference-service] pytest.mark.cluster_single_node, [e2e-llm-inference-service] pytest.mark.llmd_simulator, [e2e-llm-inference-service] ], [e2e-llm-inference-service] ), [e2e-llm-inference-service] # Model-based routing via X-Gateway-Model-Name header — /v1/completions [e2e-llm-inference-service] pytest.param( [e2e-llm-inference-service] TestCase( [e2e-llm-inference-service] base_refs=[ [e2e-llm-inference-service] "router-managed", [e2e-llm-inference-service] "workload-llmd-simulator", [e2e-llm-inference-service] ], [e2e-llm-inference-service] endpoint="/v1/completions", [e2e-llm-inference-service] prompt="KServe is a", [e2e-llm-inference-service] payload_formatter=completions_payload, [e2e-llm-inference-service] response_assertion=assert_model_field_matches("facebook/opt-125m"), [e2e-llm-inference-service] url_getter=get_model_routing_url, [e2e-llm-inference-service] extra_headers={ [e2e-llm-inference-service] MODEL_ROUTING_HEADER: f"publishers/{KSERVE_TEST_NAMESPACE}/models/facebook/opt-125m", [e2e-llm-inference-service] }, [e2e-llm-inference-service] peers=[ [e2e-llm-inference-service] TestCase( [e2e-llm-inference-service] base_refs=[ [e2e-llm-inference-service] "router-managed", [e2e-llm-inference-service] "workload-llmd-simulator", [e2e-llm-inference-service] "model-qwen2.5-0.5b", [e2e-llm-inference-service] ], [e2e-llm-inference-service] endpoint="/v1/completions", [e2e-llm-inference-service] prompt="KServe is a", [e2e-llm-inference-service] payload_formatter=completions_payload, [e2e-llm-inference-service] response_assertion=assert_model_field_matches( [e2e-llm-inference-service] "Qwen/Qwen2.5-0.5B-Instruct" [e2e-llm-inference-service] ), [e2e-llm-inference-service] url_getter=get_model_routing_url, [e2e-llm-inference-service] extra_headers={ [e2e-llm-inference-service] MODEL_ROUTING_HEADER: f"publishers/{KSERVE_TEST_NAMESPACE}/models/Qwen/Qwen2.5-0.5B-Instruct", [e2e-llm-inference-service] }, [e2e-llm-inference-service] ), [e2e-llm-inference-service] ], [e2e-llm-inference-service] ), [e2e-llm-inference-service] marks=[ [e2e-llm-inference-service] pytest.mark.cluster_cpu, [e2e-llm-inference-service] pytest.mark.cluster_single_node, [e2e-llm-inference-service] pytest.mark.llmd_simulator, [e2e-llm-inference-service] pytest.mark.model_routing, [e2e-llm-inference-service] ], [e2e-llm-inference-service] ), [e2e-llm-inference-service] # Model-based routing via X-Gateway-Model-Name header — /v1/chat/completions [e2e-llm-inference-service] pytest.param( [e2e-llm-inference-service] TestCase( [e2e-llm-inference-service] base_refs=[ [e2e-llm-inference-service] "router-managed", [e2e-llm-inference-service] "workload-llmd-simulator", [e2e-llm-inference-service] ], [e2e-llm-inference-service] endpoint="/v1/chat/completions", [e2e-llm-inference-service] prompt="What is KServe?", [e2e-llm-inference-service] payload_formatter=chat_completions_payload, [e2e-llm-inference-service] response_assertion=assert_model_field_matches("facebook/opt-125m"), [e2e-llm-inference-service] url_getter=get_model_routing_url, [e2e-llm-inference-service] extra_headers={ [e2e-llm-inference-service] MODEL_ROUTING_HEADER: f"publishers/{KSERVE_TEST_NAMESPACE}/models/facebook/opt-125m", [e2e-llm-inference-service] }, [e2e-llm-inference-service] peers=[ [e2e-llm-inference-service] TestCase( [e2e-llm-inference-service] base_refs=[ [e2e-llm-inference-service] "router-managed", [e2e-llm-inference-service] "workload-llmd-simulator", [e2e-llm-inference-service] "model-qwen2.5-0.5b", [e2e-llm-inference-service] ], [e2e-llm-inference-service] endpoint="/v1/chat/completions", [e2e-llm-inference-service] prompt="What is KServe?", [e2e-llm-inference-service] payload_formatter=chat_completions_payload, [e2e-llm-inference-service] response_assertion=assert_model_field_matches( [e2e-llm-inference-service] "Qwen/Qwen2.5-0.5B-Instruct" [e2e-llm-inference-service] ), [e2e-llm-inference-service] url_getter=get_model_routing_url, [e2e-llm-inference-service] extra_headers={ [e2e-llm-inference-service] MODEL_ROUTING_HEADER: f"publishers/{KSERVE_TEST_NAMESPACE}/models/Qwen/Qwen2.5-0.5B-Instruct", [e2e-llm-inference-service] }, [e2e-llm-inference-service] ), [e2e-llm-inference-service] ], [e2e-llm-inference-service] ), [e2e-llm-inference-service] marks=[ [e2e-llm-inference-service] pytest.mark.cluster_cpu, [e2e-llm-inference-service] pytest.mark.cluster_single_node, [e2e-llm-inference-service] pytest.mark.llmd_simulator, [e2e-llm-inference-service] pytest.mark.model_routing, [e2e-llm-inference-service] ], [e2e-llm-inference-service] ), [e2e-llm-inference-service] # Model-based routing via X-Gateway-Model-Name header — LoRA adapter [e2e-llm-inference-service] pytest.param( [e2e-llm-inference-service] TestCase( [e2e-llm-inference-service] base_refs=[ [e2e-llm-inference-service] "router-managed", [e2e-llm-inference-service] "workload-single-cpu", [e2e-llm-inference-service] "model-fb-opt-125m-with-lora-hf", [e2e-llm-inference-service] ], [e2e-llm-inference-service] endpoint="/v1/completions", [e2e-llm-inference-service] prompt="KServe is a", [e2e-llm-inference-service] model_name=f"publishers/{KSERVE_TEST_NAMESPACE}/models/lora-adapter-1", [e2e-llm-inference-service] payload_formatter=completions_payload, [e2e-llm-inference-service] response_assertion=assert_model_field_matches( [e2e-llm-inference-service] f"publishers/{KSERVE_TEST_NAMESPACE}/models/lora-adapter-1" [e2e-llm-inference-service] ), [e2e-llm-inference-service] url_getter=get_model_routing_url, [e2e-llm-inference-service] extra_headers={ [e2e-llm-inference-service] MODEL_ROUTING_HEADER: f"publishers/{KSERVE_TEST_NAMESPACE}/models/lora-adapter-1", [e2e-llm-inference-service] }, [e2e-llm-inference-service] ), [e2e-llm-inference-service] marks=[ [e2e-llm-inference-service] pytest.mark.cluster_cpu, [e2e-llm-inference-service] pytest.mark.cluster_single_node, [e2e-llm-inference-service] pytest.mark.model_routing, [e2e-llm-inference-service] pytest.mark.lora, [e2e-llm-inference-service] ], [e2e-llm-inference-service] ), [e2e-llm-inference-service] # Model-based routing via X-Gateway-Model-Name header — /v1/models (base + LoRA) [e2e-llm-inference-service] pytest.param( [e2e-llm-inference-service] TestCase( [e2e-llm-inference-service] base_refs=[ [e2e-llm-inference-service] "router-managed", [e2e-llm-inference-service] "workload-single-cpu", [e2e-llm-inference-service] "model-fb-opt-125m-with-lora-hf", [e2e-llm-inference-service] ], [e2e-llm-inference-service] endpoint="/v1/models", [e2e-llm-inference-service] response_assertion=assert_models_contains( [e2e-llm-inference-service] "facebook/opt-125m", [e2e-llm-inference-service] f"publishers/{KSERVE_TEST_NAMESPACE}/models/facebook/opt-125m", [e2e-llm-inference-service] "lora-adapter-1", [e2e-llm-inference-service] f"publishers/{KSERVE_TEST_NAMESPACE}/models/lora-adapter-1", [e2e-llm-inference-service] ), [e2e-llm-inference-service] url_getter=get_model_routing_url, [e2e-llm-inference-service] extra_headers={ [e2e-llm-inference-service] MODEL_ROUTING_HEADER: f"publishers/{KSERVE_TEST_NAMESPACE}/models/facebook/opt-125m", [e2e-llm-inference-service] }, [e2e-llm-inference-service] ), [e2e-llm-inference-service] marks=[ [e2e-llm-inference-service] pytest.mark.cluster_cpu, [e2e-llm-inference-service] pytest.mark.cluster_single_node, [e2e-llm-inference-service] pytest.mark.model_routing, [e2e-llm-inference-service] pytest.mark.lora, [e2e-llm-inference-service] ], [e2e-llm-inference-service] ), [e2e-llm-inference-service] # PVC storage tests -- validate direct PVC volume mount with real vLLM serving [e2e-llm-inference-service] pytest.param( [e2e-llm-inference-service] TestCase( [e2e-llm-inference-service] base_refs=[ [e2e-llm-inference-service] "router-managed", [e2e-llm-inference-service] "workload-single-cpu", [e2e-llm-inference-service] "model-pvc", [e2e-llm-inference-service] ], [e2e-llm-inference-service] prompt="KServe is a", [e2e-llm-inference-service] response_assertion=assert_200_with_choices, [e2e-llm-inference-service] before_test=[ensure_pvc_with_model], [e2e-llm-inference-service] ), [e2e-llm-inference-service] marks=[ [e2e-llm-inference-service] pytest.mark.cluster_cpu, [e2e-llm-inference-service] pytest.mark.cluster_single_node, [e2e-llm-inference-service] pytest.mark.pvc_storage, [e2e-llm-inference-service] ], [e2e-llm-inference-service] ), [e2e-llm-inference-service] pytest.param( [e2e-llm-inference-service] TestCase( [e2e-llm-inference-service] base_refs=[ [e2e-llm-inference-service] "router-managed", [e2e-llm-inference-service] "workload-pd-cpu", [e2e-llm-inference-service] "model-pvc", [e2e-llm-inference-service] ], [e2e-llm-inference-service] prompt="KServe is a", [e2e-llm-inference-service] response_assertion=assert_200_with_choices, [e2e-llm-inference-service] before_test=[ensure_pvc_with_model], [e2e-llm-inference-service] ), [e2e-llm-inference-service] marks=[ [e2e-llm-inference-service] pytest.mark.cluster_cpu, [e2e-llm-inference-service] pytest.mark.cluster_single_node, [e2e-llm-inference-service] pytest.mark.pvc_storage, [e2e-llm-inference-service] ], [e2e-llm-inference-service] ), [e2e-llm-inference-service] pytest.param( [e2e-llm-inference-service] TestCase( [e2e-llm-inference-service] base_refs=[ [e2e-llm-inference-service] "router-managed", [e2e-llm-inference-service] "workload-simulated-dp-ep-cpu", [e2e-llm-inference-service] "model-pvc", [e2e-llm-inference-service] ], [e2e-llm-inference-service] prompt="KServe is a", [e2e-llm-inference-service] before_test=[ensure_pvc_with_model], [e2e-llm-inference-service] ), [e2e-llm-inference-service] marks=[ [e2e-llm-inference-service] pytest.mark.cluster_cpu, [e2e-llm-inference-service] pytest.mark.cluster_multi_node, [e2e-llm-inference-service] pytest.mark.pvc_storage, [e2e-llm-inference-service] ], [e2e-llm-inference-service] ), [e2e-llm-inference-service] ], [e2e-llm-inference-service] indirect=["test_case"], [e2e-llm-inference-service] ids=generate_test_id, [e2e-llm-inference-service] ) [e2e-llm-inference-service] @log_execution [e2e-llm-inference-service] def test_llm_inference_service(test_case: TestCase): # noqa: F811 [e2e-llm-inference-service] inject_k8s_proxy() [e2e-llm-inference-service] [e2e-llm-inference-service] kserve_client = KServeClient( [e2e-llm-inference-service] config_file=os.environ.get("KUBECONFIG", "~/.kube/config"), [e2e-llm-inference-service] client_configuration=client.Configuration(), [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] service_name = test_case.llm_service.metadata.name [e2e-llm-inference-service] if not test_case.llm_service.metadata.annotations: [e2e-llm-inference-service] test_case.llm_service.metadata.annotations = {} [e2e-llm-inference-service] [e2e-llm-inference-service] test_case.llm_service.metadata.annotations[ [e2e-llm-inference-service] "security.opendatahub.io/enable-auth" [e2e-llm-inference-service] ] = "false" [e2e-llm-inference-service] prefix = test_case.log_prefix [e2e-llm-inference-service] [e2e-llm-inference-service] test_failed = False [e2e-llm-inference-service] try: [e2e-llm-inference-service] print(f"{prefix} Creating LLMInferenceService {service_name}") [e2e-llm-inference-service] create_llmisvc(kserve_client, test_case.llm_service) [e2e-llm-inference-service] print(f"{prefix} Waiting for LLMInferenceService {service_name} to be ready") [e2e-llm-inference-service] wait_for_llm_isvc_ready( [e2e-llm-inference-service] kserve_client, test_case.llm_service, test_case.wait_timeout [e2e-llm-inference-service] ) [e2e-llm-inference-service] print(f"{prefix} Waiting for model response from {service_name}") [e2e-llm-inference-service] > wait_for_model_response( [e2e-llm-inference-service] kserve_client, [e2e-llm-inference-service] test_case, [e2e-llm-inference-service] test_case.wait_timeout, [e2e-llm-inference-service] extra_headers=test_case.extra_headers, [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] llmisvc/test_llm_inference_service.py:816: [e2e-llm-inference-service] _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ [e2e-llm-inference-service] [e2e-llm-inference-service] args = (, TestCase(base_refs=['router-managed', 'workload-sin...5m-with-lora-hf-a7886ead'}]}, [e2e-llm-inference-service] 'status': None}, model_name='publishers/kserve-ci-e2e-test/models/lora-adapter-1'), 900) [e2e-llm-inference-service] kwargs = {'extra_headers': {'X-Gateway-Model-Name': 'publishers/kserve-ci-e2e-test/models/lora-adapter-1'}} [e2e-llm-inference-service] func_name = 'wait_for_model_response' [e2e-llm-inference-service] timestamp_start = '2026-07-07T14:30:46.101641', start_time = 1783434646.1019633 [e2e-llm-inference-service] duration = 904.638260602951, timestamp_end = '2026-07-07T14:45:50.740227' [e2e-llm-inference-service] [e2e-llm-inference-service] @functools.wraps(func) [e2e-llm-inference-service] def wrapper(*args, **kwargs): [e2e-llm-inference-service] func_name = func.__name__ [e2e-llm-inference-service] [e2e-llm-inference-service] timestamp_start = datetime.now().isoformat() [e2e-llm-inference-service] logger.info( [e2e-llm-inference-service] f"[{func_name}] [{timestamp_start}] start - args={args}, kwargs={kwargs}" [e2e-llm-inference-service] ) [e2e-llm-inference-service] start_time = time.time() [e2e-llm-inference-service] [e2e-llm-inference-service] try: [e2e-llm-inference-service] > result = func(*args, **kwargs) [e2e-llm-inference-service] [e2e-llm-inference-service] llmisvc/logging.py:40: [e2e-llm-inference-service] _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ [e2e-llm-inference-service] [e2e-llm-inference-service] kserve_client = [e2e-llm-inference-service] test_case = TestCase(base_refs=['router-managed', 'workload-single-cpu', 'model-fb-opt-125m-with-lora-hf'], prompt='KServe is a', ...opt-125m-with-lora-hf-a7886ead'}]}, [e2e-llm-inference-service] 'status': None}, model_name='publishers/kserve-ci-e2e-test/models/lora-adapter-1') [e2e-llm-inference-service] timeout_seconds = 900 [e2e-llm-inference-service] extra_headers = {'X-Gateway-Model-Name': 'publishers/kserve-ci-e2e-test/models/lora-adapter-1'} [e2e-llm-inference-service] [e2e-llm-inference-service] @log_execution [e2e-llm-inference-service] def wait_for_model_response( [e2e-llm-inference-service] kserve_client: KServeClient, [e2e-llm-inference-service] test_case: TestCase, # noqa: F811 [e2e-llm-inference-service] timeout_seconds: int = 900, [e2e-llm-inference-service] extra_headers: Optional[Dict[str, str]] = None, [e2e-llm-inference-service] ) -> str: [e2e-llm-inference-service] def get_successful_response(): [e2e-llm-inference-service] try: [e2e-llm-inference-service] if test_case.url_getter: [e2e-llm-inference-service] service_url = test_case.url_getter(kserve_client, test_case.llm_service) [e2e-llm-inference-service] else: [e2e-llm-inference-service] service_url = get_llm_service_url(kserve_client, test_case.llm_service) [e2e-llm-inference-service] except Exception as e: [e2e-llm-inference-service] raise AssertionError(f"❌ Failed to get service URL: {e}") from e [e2e-llm-inference-service] [e2e-llm-inference-service] model_url = service_url + test_case.endpoint [e2e-llm-inference-service] [e2e-llm-inference-service] headers = {"Content-Type": "application/json"} [e2e-llm-inference-service] if extra_headers: [e2e-llm-inference-service] headers.update(extra_headers) [e2e-llm-inference-service] [e2e-llm-inference-service] if test_case.payload_formatter is not None: [e2e-llm-inference-service] test_payload = test_case.payload_formatter(test_case) [e2e-llm-inference-service] elif test_case.prompt is not None: [e2e-llm-inference-service] test_payload = { [e2e-llm-inference-service] "model": test_case.model_name [e2e-llm-inference-service] if not extra_headers or MODEL_ROUTING_HEADER not in extra_headers [e2e-llm-inference-service] else extra_headers[MODEL_ROUTING_HEADER], [e2e-llm-inference-service] "prompt": test_case.prompt, [e2e-llm-inference-service] "max_tokens": test_case.max_tokens, [e2e-llm-inference-service] } [e2e-llm-inference-service] else: [e2e-llm-inference-service] test_payload = None [e2e-llm-inference-service] [e2e-llm-inference-service] logger.info(f"Calling LLM service at {model_url} with payload {test_payload}") [e2e-llm-inference-service] try: [e2e-llm-inference-service] if test_payload is not None: [e2e-llm-inference-service] response = post_with_retry( [e2e-llm-inference-service] model_url, [e2e-llm-inference-service] headers=headers, [e2e-llm-inference-service] json_data=test_payload, [e2e-llm-inference-service] timeout=test_case.response_timeout, [e2e-llm-inference-service] ) [e2e-llm-inference-service] else: [e2e-llm-inference-service] response = get_with_retry( [e2e-llm-inference-service] model_url, [e2e-llm-inference-service] headers=headers, [e2e-llm-inference-service] timeout=test_case.response_timeout, [e2e-llm-inference-service] ) [e2e-llm-inference-service] except Exception as e: [e2e-llm-inference-service] logger.error(f"❌ Failed to call model: {e}") [e2e-llm-inference-service] raise AssertionError(f"❌ Failed to call model: {e}") from e [e2e-llm-inference-service] [e2e-llm-inference-service] logger.info(f"Model response is {response.status_code}: {response.text[:500]}") [e2e-llm-inference-service] [e2e-llm-inference-service] if 200 <= response.status_code < 300: [e2e-llm-inference-service] return response [e2e-llm-inference-service] raise AssertionError( [e2e-llm-inference-service] f"Service returned {response.status_code}: {response.text}" [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] > response = wait_for(get_successful_response, timeout=timeout_seconds, interval=5.0) [e2e-llm-inference-service] [e2e-llm-inference-service] llmisvc/test_llm_inference_service.py:1119: [e2e-llm-inference-service] _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ [e2e-llm-inference-service] [e2e-llm-inference-service] assertion_fn = .get_successful_response at 0x7f1d17b70cc0> [e2e-llm-inference-service] timeout = 900, interval = 5.0 [e2e-llm-inference-service] [e2e-llm-inference-service] def wait_for( [e2e-llm-inference-service] assertion_fn: Callable[[], Any], timeout: float = 5.0, interval: float = 0.1 [e2e-llm-inference-service] ) -> Any: [e2e-llm-inference-service] """Wait for the assertion to succeed within timeout.""" [e2e-llm-inference-service] deadline = time.time() + timeout [e2e-llm-inference-service] last_msg = None [e2e-llm-inference-service] while True: [e2e-llm-inference-service] try: [e2e-llm-inference-service] > return assertion_fn() [e2e-llm-inference-service] [e2e-llm-inference-service] llmisvc/test_llm_inference_service.py:1215: [e2e-llm-inference-service] _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ [e2e-llm-inference-service] [e2e-llm-inference-service] def get_successful_response(): [e2e-llm-inference-service] try: [e2e-llm-inference-service] if test_case.url_getter: [e2e-llm-inference-service] service_url = test_case.url_getter(kserve_client, test_case.llm_service) [e2e-llm-inference-service] else: [e2e-llm-inference-service] service_url = get_llm_service_url(kserve_client, test_case.llm_service) [e2e-llm-inference-service] except Exception as e: [e2e-llm-inference-service] raise AssertionError(f"❌ Failed to get service URL: {e}") from e [e2e-llm-inference-service] [e2e-llm-inference-service] model_url = service_url + test_case.endpoint [e2e-llm-inference-service] [e2e-llm-inference-service] headers = {"Content-Type": "application/json"} [e2e-llm-inference-service] if extra_headers: [e2e-llm-inference-service] headers.update(extra_headers) [e2e-llm-inference-service] [e2e-llm-inference-service] if test_case.payload_formatter is not None: [e2e-llm-inference-service] test_payload = test_case.payload_formatter(test_case) [e2e-llm-inference-service] elif test_case.prompt is not None: [e2e-llm-inference-service] test_payload = { [e2e-llm-inference-service] "model": test_case.model_name [e2e-llm-inference-service] if not extra_headers or MODEL_ROUTING_HEADER not in extra_headers [e2e-llm-inference-service] else extra_headers[MODEL_ROUTING_HEADER], [e2e-llm-inference-service] "prompt": test_case.prompt, [e2e-llm-inference-service] "max_tokens": test_case.max_tokens, [e2e-llm-inference-service] } [e2e-llm-inference-service] else: [e2e-llm-inference-service] test_payload = None [e2e-llm-inference-service] [e2e-llm-inference-service] logger.info(f"Calling LLM service at {model_url} with payload {test_payload}") [e2e-llm-inference-service] try: [e2e-llm-inference-service] if test_payload is not None: [e2e-llm-inference-service] response = post_with_retry( [e2e-llm-inference-service] model_url, [e2e-llm-inference-service] headers=headers, [e2e-llm-inference-service] json_data=test_payload, [e2e-llm-inference-service] timeout=test_case.response_timeout, [e2e-llm-inference-service] ) [e2e-llm-inference-service] else: [e2e-llm-inference-service] response = get_with_retry( [e2e-llm-inference-service] model_url, [e2e-llm-inference-service] headers=headers, [e2e-llm-inference-service] timeout=test_case.response_timeout, [e2e-llm-inference-service] ) [e2e-llm-inference-service] except Exception as e: [e2e-llm-inference-service] logger.error(f"❌ Failed to call model: {e}") [e2e-llm-inference-service] > raise AssertionError(f"❌ Failed to call model: {e}") from e [e2e-llm-inference-service] E AssertionError: ❌ Failed to call model: HTTPConnectionPool(host='ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com', port=80): Max retries exceeded with url: /v1/completions (Caused by ReadTimeoutError("HTTPConnectionPool(host='ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com', port=80): Read timed out. (read timeout=60)")) [e2e-llm-inference-service] [e2e-llm-inference-service] llmisvc/test_llm_inference_service.py:1109: AssertionError [e2e-llm-inference-service] ------------------------------ Captured log setup ------------------------------ [e2e-llm-inference-service] INFO e2e.llmisvc.logging:fixtures.py:1637 No HTTP proxy configured for k8s client [e2e-llm-inference-service] INFO e2e.llmisvc.logging:fixtures.py:1586 Checking LLMInferenceServiceConfig router-managed-llmisvc-model-fb-98f275aa in namespace kserve-ci-e2e-test [e2e-llm-inference-service] INFO e2e.llmisvc.logging:fixtures.py:1612 Resource not found, creating LLMInferenceServiceConfig router-managed-llmisvc-model-fb-98f275aa [e2e-llm-inference-service] INFO e2e.llmisvc.logging:fixtures.py:1622 ✓ Successfully created LLMInferenceServiceConfig router-managed-llmisvc-model-fb-98f275aa [e2e-llm-inference-service] INFO e2e.llmisvc.logging:fixtures.py:1586 Checking LLMInferenceServiceConfig workload-single-cpu-llmisvc-mod-495991f8 in namespace kserve-ci-e2e-test [e2e-llm-inference-service] INFO e2e.llmisvc.logging:fixtures.py:1612 Resource not found, creating LLMInferenceServiceConfig workload-single-cpu-llmisvc-mod-495991f8 [e2e-llm-inference-service] INFO e2e.llmisvc.logging:fixtures.py:1622 ✓ Successfully created LLMInferenceServiceConfig workload-single-cpu-llmisvc-mod-495991f8 [e2e-llm-inference-service] INFO e2e.llmisvc.logging:fixtures.py:1586 Checking LLMInferenceServiceConfig model-fb-opt-125m-with-lora-hf-a7886ead in namespace kserve-ci-e2e-test [e2e-llm-inference-service] INFO e2e.llmisvc.logging:fixtures.py:1612 Resource not found, creating LLMInferenceServiceConfig model-fb-opt-125m-with-lora-hf-a7886ead [e2e-llm-inference-service] INFO e2e.llmisvc.logging:fixtures.py:1622 ✓ Successfully created LLMInferenceServiceConfig model-fb-opt-125m-with-lora-hf-a7886ead [e2e-llm-inference-service] ------------------------------ Captured log call ------------------------------- [e2e-llm-inference-service] INFO e2e.llmisvc.logging:logging.py:34 [test_llm_inference_service] [2026-07-07T14:29:18.864237] start - args=(), kwargs={'test_case': TestCase(base_refs=['router-managed', 'workload-single-cpu', 'model-fb-opt-125m-with-lora-hf'], prompt='KServe is a', service_name='llmisvc-model-fb-opt-125m-with-7ca60146', endpoint='/v1/completions', max_tokens=20, payload_formatter=, response_assertion=.response_assertion at 0x7f1d1813eac0>, wait_timeout=900, response_timeout=60, extra_headers={'X-Gateway-Model-Name': 'publishers/kserve-ci-e2e-test/models/lora-adapter-1'}, url_getter=, expected_gateway=None, before_test=[], after_test=[], peers=[], llm_service={'api_version': 'serving.kserve.io/v1alpha1', [e2e-llm-inference-service] 'kind': 'LLMInferenceService', [e2e-llm-inference-service] 'metadata': {'annotations': None, [e2e-llm-inference-service] 'creation_timestamp': None, [e2e-llm-inference-service] 'deletion_grace_period_seconds': None, [e2e-llm-inference-service] 'deletion_timestamp': None, [e2e-llm-inference-service] 'finalizers': None, [e2e-llm-inference-service] 'generate_name': None, [e2e-llm-inference-service] 'generation': None, [e2e-llm-inference-service] 'labels': None, [e2e-llm-inference-service] 'managed_fields': None, [e2e-llm-inference-service] 'name': 'llmisvc-model-fb-opt-125m-with-7ca60146', [e2e-llm-inference-service] 'namespace': 'kserve-ci-e2e-test', [e2e-llm-inference-service] 'owner_references': None, [e2e-llm-inference-service] 'resource_version': None, [e2e-llm-inference-service] 'self_link': None, [e2e-llm-inference-service] 'uid': None}, [e2e-llm-inference-service] 'spec': {'baseRefs': [{'name': 'router-managed-llmisvc-model-fb-98f275aa'}, [e2e-llm-inference-service] {'name': 'workload-single-cpu-llmisvc-mod-495991f8'}, [e2e-llm-inference-service] {'name': 'model-fb-opt-125m-with-lora-hf-a7886ead'}]}, [e2e-llm-inference-service] 'status': None}, model_name='publishers/kserve-ci-e2e-test/models/lora-adapter-1')} [e2e-llm-inference-service] INFO e2e.llmisvc.logging:fixtures.py:1637 No HTTP proxy configured for k8s client [e2e-llm-inference-service] INFO e2e.llmisvc.logging:logging.py:34 [create_llmisvc] [2026-07-07T14:29:18.878552] start - args=(, {'api_version': 'serving.kserve.io/v1alpha1', [e2e-llm-inference-service] 'kind': 'LLMInferenceService', [e2e-llm-inference-service] 'metadata': {'annotations': {'security.opendatahub.io/enable-auth': 'false'}, [e2e-llm-inference-service] 'creation_timestamp': None, [e2e-llm-inference-service] 'deletion_grace_period_seconds': None, [e2e-llm-inference-service] 'deletion_timestamp': None, [e2e-llm-inference-service] 'finalizers': None, [e2e-llm-inference-service] 'generate_name': None, [e2e-llm-inference-service] 'generation': None, [e2e-llm-inference-service] 'labels': None, [e2e-llm-inference-service] 'managed_fields': None, [e2e-llm-inference-service] 'name': 'llmisvc-model-fb-opt-125m-with-7ca60146', [e2e-llm-inference-service] 'namespace': 'kserve-ci-e2e-test', [e2e-llm-inference-service] 'owner_references': None, [e2e-llm-inference-service] 'resource_version': None, [e2e-llm-inference-service] 'self_link': None, [e2e-llm-inference-service] 'uid': None}, [e2e-llm-inference-service] 'spec': {'baseRefs': [{'name': 'router-managed-llmisvc-model-fb-98f275aa'}, [e2e-llm-inference-service] {'name': 'workload-single-cpu-llmisvc-mod-495991f8'}, [e2e-llm-inference-service] {'name': 'model-fb-opt-125m-with-lora-hf-a7886ead'}]}, [e2e-llm-inference-service] 'status': None}), kwargs={} [e2e-llm-inference-service] INFO e2e.llmisvc.logging:logging.py:43 [create_llmisvc] [2026-07-07T14:29:18.936286] end - ✅ in 0.057s [e2e-llm-inference-service] INFO e2e.llmisvc.logging:logging.py:34 [wait_for_llm_isvc_ready] [2026-07-07T14:29:18.936447] start - args=(, {'api_version': 'serving.kserve.io/v1alpha1', [e2e-llm-inference-service] 'kind': 'LLMInferenceService', [e2e-llm-inference-service] 'metadata': {'annotations': {'security.opendatahub.io/enable-auth': 'false'}, [e2e-llm-inference-service] 'creation_timestamp': None, [e2e-llm-inference-service] 'deletion_grace_period_seconds': None, [e2e-llm-inference-service] 'deletion_timestamp': None, [e2e-llm-inference-service] 'finalizers': None, [e2e-llm-inference-service] 'generate_name': None, [e2e-llm-inference-service] 'generation': None, [e2e-llm-inference-service] 'labels': None, [e2e-llm-inference-service] 'managed_fields': None, [e2e-llm-inference-service] 'name': 'llmisvc-model-fb-opt-125m-with-7ca60146', [e2e-llm-inference-service] 'namespace': 'kserve-ci-e2e-test', [e2e-llm-inference-service] 'owner_references': None, [e2e-llm-inference-service] 'resource_version': None, [e2e-llm-inference-service] 'self_link': None, [e2e-llm-inference-service] 'uid': None}, [e2e-llm-inference-service] 'spec': {'baseRefs': [{'name': 'router-managed-llmisvc-model-fb-98f275aa'}, [e2e-llm-inference-service] {'name': 'workload-single-cpu-llmisvc-mod-495991f8'}, [e2e-llm-inference-service] {'name': 'model-fb-opt-125m-with-lora-hf-a7886ead'}]}, [e2e-llm-inference-service] 'status': None}, 900), kwargs={} [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1222 Waiting: No conditions found in status [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1222 Waiting: Missing true conditions: {'RouterReady', 'Ready', 'WorkloadsReady'}, expected {'RouterReady', 'Ready', 'WorkloadsReady'}, got [{'lastTransitionTime': '2026-07-07T14:29:26Z', 'message': 'The following HTTPRoutes are not ready: [kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-with-7ca60146-kserve-route: "False" (reason "InvalidKind", message "referencing unsupported backendRef: group \\"inference.networking.x-k8s.io\\" kind \\"InferencePool\\"")]', 'reason': 'HTTPRoutesNotReady', 'severity': 'Info', 'status': 'False', 'type': 'HTTPRoutesReady'}, {'lastTransitionTime': '2026-07-07T14:29:26Z', 'message': 'Inference Pool kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-with-7ca60146-inference-pool exists but no Gateway controller has accepted it yet', 'reason': 'WaitingForGateway', 'severity': 'Info', 'status': 'False', 'type': 'InferencePoolReady'}, {'lastTransitionTime': '2026-07-07T14:29:26Z', 'reason': 'Progressing', 'severity': 'Info', 'status': 'False', 'type': 'MainWorkloadReady'}, {'lastTransitionTime': '2026-07-07T14:29:26Z', 'severity': 'Info', 'status': 'True', 'type': 'PresetsCombined'}, {'lastTransitionTime': '2026-07-07T14:29:26Z', 'message': 'The following HTTPRoutes are not ready: [kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-with-7ca60146-kserve-route: "False" (reason "InvalidKind", message "referencing unsupported backendRef: group \\"inference.networking.x-k8s.io\\" kind \\"InferencePool\\"")]', 'reason': 'HTTPRoutesNotReady', 'status': 'False', 'type': 'Ready'}, {'lastTransitionTime': '2026-07-07T14:29:26Z', 'message': 'The following HTTPRoutes are not ready: [kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-with-7ca60146-kserve-route: "False" (reason "InvalidKind", message "referencing unsupported backendRef: group \\"inference.networking.x-k8s.io\\" kind \\"InferencePool\\"")]', 'reason': 'HTTPRoutesNotReady', 'status': 'False', 'type': 'RouterReady'}, {'lastTransitionTime': '2026-07-07T14:29:26Z', 'message': 'Deployment rollout in progress', 'reason': 'Progressing', 'severity': 'Info', 'status': 'False', 'type': 'SchedulerWorkloadReady'}, {'lastTransitionTime': '2026-07-07T14:29:26Z', 'reason': 'Progressing', 'status': 'False', 'type': 'WorkloadsReady'}] [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1222 Waiting: Missing true conditions: {'RouterReady', 'Ready', 'WorkloadsReady'}, expected {'RouterReady', 'Ready', 'WorkloadsReady'}, got [{'lastTransitionTime': '2026-07-07T14:29:35Z', 'severity': 'Info', 'status': 'True', 'type': 'HTTPRoutesReady'}, {'lastTransitionTime': '2026-07-07T14:29:35Z', 'severity': 'Info', 'status': 'True', 'type': 'InferencePoolReady'}, {'lastTransitionTime': '2026-07-07T14:29:35Z', 'message': 'Deployment does not have minimum availability.', 'reason': 'MinimumReplicasUnavailable', 'severity': 'Info', 'status': 'False', 'type': 'MainWorkloadReady'}, {'lastTransitionTime': '2026-07-07T14:29:26Z', 'severity': 'Info', 'status': 'True', 'type': 'PresetsCombined'}, {'lastTransitionTime': '2026-07-07T14:29:35Z', 'message': 'Deployment does not have minimum availability.', 'reason': 'MinimumReplicasUnavailable', 'status': 'False', 'type': 'Ready'}, {'lastTransitionTime': '2026-07-07T14:29:35Z', 'message': 'Deployment does not have minimum availability.', 'reason': 'MinimumReplicasUnavailable', 'status': 'False', 'type': 'RouterReady'}, {'lastTransitionTime': '2026-07-07T14:29:35Z', 'message': 'Deployment does not have minimum availability.', 'reason': 'MinimumReplicasUnavailable', 'severity': 'Info', 'status': 'False', 'type': 'SchedulerWorkloadReady'}, {'lastTransitionTime': '2026-07-07T14:29:35Z', 'message': 'Deployment does not have minimum availability.', 'reason': 'MinimumReplicasUnavailable', 'status': 'False', 'type': 'WorkloadsReady'}] [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1222 Waiting: Missing true conditions: {'Ready', 'WorkloadsReady'}, expected {'RouterReady', 'Ready', 'WorkloadsReady'}, got [{'lastTransitionTime': '2026-07-07T14:29:35Z', 'severity': 'Info', 'status': 'True', 'type': 'HTTPRoutesReady'}, {'lastTransitionTime': '2026-07-07T14:29:35Z', 'severity': 'Info', 'status': 'True', 'type': 'InferencePoolReady'}, {'lastTransitionTime': '2026-07-07T14:29:35Z', 'message': 'Deployment does not have minimum availability.', 'reason': 'MinimumReplicasUnavailable', 'severity': 'Info', 'status': 'False', 'type': 'MainWorkloadReady'}, {'lastTransitionTime': '2026-07-07T14:29:26Z', 'severity': 'Info', 'status': 'True', 'type': 'PresetsCombined'}, {'lastTransitionTime': '2026-07-07T14:29:35Z', 'message': 'Deployment does not have minimum availability.', 'reason': 'MinimumReplicasUnavailable', 'status': 'False', 'type': 'Ready'}, {'lastTransitionTime': '2026-07-07T14:29:59Z', 'status': 'True', 'type': 'RouterReady'}, {'lastTransitionTime': '2026-07-07T14:29:59Z', 'severity': 'Info', 'status': 'True', 'type': 'SchedulerWorkloadReady'}, {'lastTransitionTime': '2026-07-07T14:29:35Z', 'message': 'Deployment does not have minimum availability.', 'reason': 'MinimumReplicasUnavailable', 'status': 'False', 'type': 'WorkloadsReady'}] [e2e-llm-inference-service] INFO e2e.llmisvc.logging:logging.py:43 [wait_for_llm_isvc_ready] [2026-07-07T14:30:46.101468] end - ✅ in 87.165s [e2e-llm-inference-service] INFO e2e.llmisvc.logging:logging.py:34 [wait_for_model_response] [2026-07-07T14:30:46.101641] start - args=(, TestCase(base_refs=['router-managed', 'workload-single-cpu', 'model-fb-opt-125m-with-lora-hf'], prompt='KServe is a', service_name='llmisvc-model-fb-opt-125m-with-7ca60146', endpoint='/v1/completions', max_tokens=20, payload_formatter=, response_assertion=.response_assertion at 0x7f1d1813eac0>, wait_timeout=900, response_timeout=60, extra_headers={'X-Gateway-Model-Name': 'publishers/kserve-ci-e2e-test/models/lora-adapter-1'}, url_getter=, expected_gateway=None, before_test=[], after_test=[], peers=[], llm_service={'api_version': 'serving.kserve.io/v1alpha1', [e2e-llm-inference-service] 'kind': 'LLMInferenceService', [e2e-llm-inference-service] 'metadata': {'annotations': {'security.opendatahub.io/enable-auth': 'false'}, [e2e-llm-inference-service] 'creation_timestamp': None, [e2e-llm-inference-service] 'deletion_grace_period_seconds': None, [e2e-llm-inference-service] 'deletion_timestamp': None, [e2e-llm-inference-service] 'finalizers': None, [e2e-llm-inference-service] 'generate_name': None, [e2e-llm-inference-service] 'generation': None, [e2e-llm-inference-service] 'labels': None, [e2e-llm-inference-service] 'managed_fields': None, [e2e-llm-inference-service] 'name': 'llmisvc-model-fb-opt-125m-with-7ca60146', [e2e-llm-inference-service] 'namespace': 'kserve-ci-e2e-test', [e2e-llm-inference-service] 'owner_references': None, [e2e-llm-inference-service] 'resource_version': None, [e2e-llm-inference-service] 'self_link': None, [e2e-llm-inference-service] 'uid': None}, [e2e-llm-inference-service] 'spec': {'baseRefs': [{'name': 'router-managed-llmisvc-model-fb-98f275aa'}, [e2e-llm-inference-service] {'name': 'workload-single-cpu-llmisvc-mod-495991f8'}, [e2e-llm-inference-service] {'name': 'model-fb-opt-125m-with-lora-hf-a7886ead'}]}, [e2e-llm-inference-service] 'status': None}, model_name='publishers/kserve-ci-e2e-test/models/lora-adapter-1'), 900), kwargs={'extra_headers': {'X-Gateway-Model-Name': 'publishers/kserve-ci-e2e-test/models/lora-adapter-1'}} [e2e-llm-inference-service] INFO e2e.llmisvc.logging:logging.py:34 [get_model_routing_url] [2026-07-07T14:30:46.102003] start - args=(, {'api_version': 'serving.kserve.io/v1alpha1', [e2e-llm-inference-service] 'kind': 'LLMInferenceService', [e2e-llm-inference-service] 'metadata': {'annotations': {'security.opendatahub.io/enable-auth': 'false'}, [e2e-llm-inference-service] 'creation_timestamp': None, [e2e-llm-inference-service] 'deletion_grace_period_seconds': None, [e2e-llm-inference-service] 'deletion_timestamp': None, [e2e-llm-inference-service] 'finalizers': None, [e2e-llm-inference-service] 'generate_name': None, [e2e-llm-inference-service] 'generation': None, [e2e-llm-inference-service] 'labels': None, [e2e-llm-inference-service] 'managed_fields': None, [e2e-llm-inference-service] 'name': 'llmisvc-model-fb-opt-125m-with-7ca60146', [e2e-llm-inference-service] 'namespace': 'kserve-ci-e2e-test', [e2e-llm-inference-service] 'owner_references': None, [e2e-llm-inference-service] 'resource_version': None, [e2e-llm-inference-service] 'self_link': None, [e2e-llm-inference-service] 'uid': None}, [e2e-llm-inference-service] 'spec': {'baseRefs': [{'name': 'router-managed-llmisvc-model-fb-98f275aa'}, [e2e-llm-inference-service] {'name': 'workload-single-cpu-llmisvc-mod-495991f8'}, [e2e-llm-inference-service] {'name': 'model-fb-opt-125m-with-lora-hf-a7886ead'}]}, [e2e-llm-inference-service] 'status': None}), kwargs={} [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:181 Found model-routing URL for llmisvc-model-fb-opt-125m-with-7ca60146: http://ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com/ (name='gateway-external-model-routing', path='/') [e2e-llm-inference-service] INFO e2e.llmisvc.logging:logging.py:43 [get_model_routing_url] [2026-07-07T14:30:46.114982] end - ✅ in 0.013s [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1092 Calling LLM service at http://ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com/v1/completions with payload {'model': 'publishers/kserve-ci-e2e-test/models/lora-adapter-1', 'prompt': 'KServe is a', 'max_tokens': 20} [e2e-llm-inference-service] WARNING urllib3.connectionpool:connectionpool.py:868 Retrying (Retry(total=7, connect=None, read=None, redirect=None, status=None)) after connection broken by 'ReadTimeoutError("HTTPConnectionPool(host='ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com', port=80): Read timed out. (read timeout=60)")': /v1/completions [e2e-llm-inference-service] WARNING urllib3.connectionpool:connectionpool.py:868 Retrying (Retry(total=6, connect=None, read=None, redirect=None, status=None)) after connection broken by 'ReadTimeoutError("HTTPConnectionPool(host='ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com', port=80): Read timed out. (read timeout=60)")': /v1/completions [e2e-llm-inference-service] WARNING urllib3.connectionpool:connectionpool.py:868 Retrying (Retry(total=5, connect=None, read=None, redirect=None, status=None)) after connection broken by 'ReadTimeoutError("HTTPConnectionPool(host='ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com', port=80): Read timed out. (read timeout=60)")': /v1/completions [e2e-llm-inference-service] WARNING urllib3.connectionpool:connectionpool.py:868 Retrying (Retry(total=4, connect=None, read=None, redirect=None, status=None)) after connection broken by 'ReadTimeoutError("HTTPConnectionPool(host='ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com', port=80): Read timed out. (read timeout=60)")': /v1/completions [e2e-llm-inference-service] WARNING urllib3.connectionpool:connectionpool.py:868 Retrying (Retry(total=3, connect=None, read=None, redirect=None, status=None)) after connection broken by 'ReadTimeoutError("HTTPConnectionPool(host='ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com', port=80): Read timed out. (read timeout=60)")': /v1/completions [e2e-llm-inference-service] WARNING urllib3.connectionpool:connectionpool.py:868 Retrying (Retry(total=2, connect=None, read=None, redirect=None, status=None)) after connection broken by 'ReadTimeoutError("HTTPConnectionPool(host='ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com', port=80): Read timed out. (read timeout=60)")': /v1/completions [e2e-llm-inference-service] WARNING urllib3.connectionpool:connectionpool.py:868 Retrying (Retry(total=1, connect=None, read=None, redirect=None, status=None)) after connection broken by 'ReadTimeoutError("HTTPConnectionPool(host='ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com', port=80): Read timed out. (read timeout=60)")': /v1/completions [e2e-llm-inference-service] WARNING urllib3.connectionpool:connectionpool.py:868 Retrying (Retry(total=0, connect=None, read=None, redirect=None, status=None)) after connection broken by 'ReadTimeoutError("HTTPConnectionPool(host='ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com', port=80): Read timed out. (read timeout=60)")': /v1/completions [e2e-llm-inference-service] ERROR e2e.llmisvc.logging:test_llm_inference_service.py:1108 ❌ Failed to call model: HTTPConnectionPool(host='ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com', port=80): Max retries exceeded with url: /v1/completions (Caused by ReadTimeoutError("HTTPConnectionPool(host='ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com', port=80): Read timed out. (read timeout=60)")) [e2e-llm-inference-service] ERROR e2e.llmisvc.logging:test_llm_inference_service.py:1219 Timed out waiting: ❌ Failed to call model: HTTPConnectionPool(host='ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com', port=80): Max retries exceeded with url: /v1/completions (Caused by ReadTimeoutError("HTTPConnectionPool(host='ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com', port=80): Read timed out. (read timeout=60)")) [e2e-llm-inference-service] ERROR e2e.llmisvc.logging:logging.py:48 [wait_for_model_response] [2026-07-07T14:45:50.740227] end - ❌ 904.638s: ❌ Failed to call model: HTTPConnectionPool(host='ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com', port=80): Max retries exceeded with url: /v1/completions (Caused by ReadTimeoutError("HTTPConnectionPool(host='ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com', port=80): Read timed out. (read timeout=60)")) [e2e-llm-inference-service] ERROR e2e.llmisvc.logging:test_llm_inference_service.py:831 [router-managed-workload-single-cpu-model-fb-opt-125m-with-lora-hf] ❌ ERROR: Failed to call llm inference service llmisvc-model-fb-opt-125m-with-7ca60146: ❌ Failed to call model: HTTPConnectionPool(host='ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com', port=80): Max retries exceeded with url: /v1/completions (Caused by ReadTimeoutError("HTTPConnectionPool(host='ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com', port=80): Read timed out. (read timeout=60)")) [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1240 🔍 # Diagnostics for 'llmisvc-model-fb-opt-125m-with-7ca60146' in 'kserve-ci-e2e-test' [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1241 --- [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1242 # LLMInferenceService llmisvc-model-fb-opt-125m-with-7ca60146 [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1245 apiVersion: serving.kserve.io/v1alpha1 [e2e-llm-inference-service] kind: LLMInferenceService [e2e-llm-inference-service] metadata: [e2e-llm-inference-service] annotations: [e2e-llm-inference-service] security.opendatahub.io/enable-auth: 'false' [e2e-llm-inference-service] creationTimestamp: '2026-07-07T14:29:18Z' [e2e-llm-inference-service] finalizers: [e2e-llm-inference-service] - serving.kserve.io/llmisvc-finalizer [e2e-llm-inference-service] generation: 1 [e2e-llm-inference-service] managedFields: [e2e-llm-inference-service] - apiVersion: serving.kserve.io/v1alpha1 [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:metadata: [e2e-llm-inference-service] f:annotations: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:security.opendatahub.io/enable-auth: {} [e2e-llm-inference-service] f:spec: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:baseRefs: {} [e2e-llm-inference-service] manager: OpenAPI-Generator [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] time: '2026-07-07T14:29:18Z' [e2e-llm-inference-service] - apiVersion: serving.kserve.io/v1alpha2 [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:metadata: [e2e-llm-inference-service] f:finalizers: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] v:"serving.kserve.io/llmisvc-finalizer": {} [e2e-llm-inference-service] manager: manager [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] time: '2026-07-07T14:29:18Z' [e2e-llm-inference-service] - apiVersion: serving.kserve.io/v1alpha2 [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:status: [e2e-llm-inference-service] f:addresses: {} [e2e-llm-inference-service] f:annotations: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:serving.kserve.io/config-llm-decode-template: {} [e2e-llm-inference-service] f:serving.kserve.io/config-llm-decode-worker-data-parallel: {} [e2e-llm-inference-service] f:serving.kserve.io/config-llm-prefill-template: {} [e2e-llm-inference-service] f:serving.kserve.io/config-llm-prefill-worker-data-parallel: {} [e2e-llm-inference-service] f:serving.kserve.io/config-llm-router-route: {} [e2e-llm-inference-service] f:serving.kserve.io/config-llm-scheduler: {} [e2e-llm-inference-service] f:serving.kserve.io/config-llm-template: {} [e2e-llm-inference-service] f:serving.kserve.io/config-llm-tracing: {} [e2e-llm-inference-service] f:serving.kserve.io/config-llm-worker-data-parallel: {} [e2e-llm-inference-service] f:appliedConfigs: {} [e2e-llm-inference-service] f:conditions: {} [e2e-llm-inference-service] f:observedGeneration: {} [e2e-llm-inference-service] f:router: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:gateways: {} [e2e-llm-inference-service] f:scheduler: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:inferencePool: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:group: {} [e2e-llm-inference-service] f:kind: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:service: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:group: {} [e2e-llm-inference-service] f:kind: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:url: {} [e2e-llm-inference-service] f:workloads: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:primary: {} [e2e-llm-inference-service] f:scheduler: {} [e2e-llm-inference-service] f:service: {} [e2e-llm-inference-service] manager: manager [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] subresource: status [e2e-llm-inference-service] time: '2026-07-07T14:30:45Z' [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-7ca60146 [e2e-llm-inference-service] namespace: kserve-ci-e2e-test [e2e-llm-inference-service] resourceVersion: '29829' [e2e-llm-inference-service] uid: e15cfc59-e4a1-4876-b43b-85ce3b1b239b [e2e-llm-inference-service] spec: [e2e-llm-inference-service] baseRefs: [e2e-llm-inference-service] - name: router-managed-llmisvc-model-fb-98f275aa [e2e-llm-inference-service] - name: workload-single-cpu-llmisvc-mod-495991f8 [e2e-llm-inference-service] - name: model-fb-opt-125m-with-lora-hf-a7886ead [e2e-llm-inference-service] model: [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-7ca60146 [e2e-llm-inference-service] uri: '' [e2e-llm-inference-service] status: [e2e-llm-inference-service] addresses: [e2e-llm-inference-service] - name: gateway-external-model-routing [e2e-llm-inference-service] url: http://ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com/ [e2e-llm-inference-service] - name: gateway-external [e2e-llm-inference-service] url: http://ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com/kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-with-7ca60146 [e2e-llm-inference-service] - name: gateway-internal-model-routing [e2e-llm-inference-service] url: http://openshift-ai-inference-openshift-default.openshift-ingress.svc.cluster.local/ [e2e-llm-inference-service] - name: gateway-internal [e2e-llm-inference-service] url: http://openshift-ai-inference-openshift-default.openshift-ingress.svc.cluster.local/kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-with-7ca60146 [e2e-llm-inference-service] annotations: [e2e-llm-inference-service] serving.kserve.io/config-llm-decode-template: kserve-config-llm-decode-template [e2e-llm-inference-service] serving.kserve.io/config-llm-decode-worker-data-parallel: kserve-config-llm-decode-worker-data-parallel [e2e-llm-inference-service] serving.kserve.io/config-llm-prefill-template: kserve-config-llm-prefill-template [e2e-llm-inference-service] serving.kserve.io/config-llm-prefill-worker-data-parallel: kserve-config-llm-prefill-worker-data-parallel [e2e-llm-inference-service] serving.kserve.io/config-llm-router-route: kserve-config-llm-router-route [e2e-llm-inference-service] serving.kserve.io/config-llm-scheduler: kserve-config-llm-scheduler [e2e-llm-inference-service] serving.kserve.io/config-llm-template: kserve-config-llm-template [e2e-llm-inference-service] serving.kserve.io/config-llm-tracing: kserve-config-llm-tracing [e2e-llm-inference-service] serving.kserve.io/config-llm-worker-data-parallel: kserve-config-llm-worker-data-parallel [e2e-llm-inference-service] conditions: [e2e-llm-inference-service] - lastTransitionTime: '2026-07-07T14:29:35Z' [e2e-llm-inference-service] severity: Info [e2e-llm-inference-service] status: 'True' [e2e-llm-inference-service] type: HTTPRoutesReady [e2e-llm-inference-service] - lastTransitionTime: '2026-07-07T14:29:35Z' [e2e-llm-inference-service] severity: Info [e2e-llm-inference-service] status: 'True' [e2e-llm-inference-service] type: InferencePoolReady [e2e-llm-inference-service] - lastTransitionTime: '2026-07-07T14:30:45Z' [e2e-llm-inference-service] severity: Info [e2e-llm-inference-service] status: 'True' [e2e-llm-inference-service] type: MainWorkloadReady [e2e-llm-inference-service] - lastTransitionTime: '2026-07-07T14:29:26Z' [e2e-llm-inference-service] severity: Info [e2e-llm-inference-service] status: 'True' [e2e-llm-inference-service] type: PresetsCombined [e2e-llm-inference-service] - lastTransitionTime: '2026-07-07T14:30:45Z' [e2e-llm-inference-service] status: 'True' [e2e-llm-inference-service] type: Ready [e2e-llm-inference-service] - lastTransitionTime: '2026-07-07T14:29:59Z' [e2e-llm-inference-service] status: 'True' [e2e-llm-inference-service] type: RouterReady [e2e-llm-inference-service] - lastTransitionTime: '2026-07-07T14:29:59Z' [e2e-llm-inference-service] severity: Info [e2e-llm-inference-service] status: 'True' [e2e-llm-inference-service] type: SchedulerWorkloadReady [e2e-llm-inference-service] - lastTransitionTime: '2026-07-07T14:30:45Z' [e2e-llm-inference-service] status: 'True' [e2e-llm-inference-service] type: WorkloadsReady [e2e-llm-inference-service] observedGeneration: 1 [e2e-llm-inference-service] url: http://ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com/kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-with-7ca60146 [e2e-llm-inference-service] [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:44 TIME NAMESPACE SOURCE TYPE REASON MESSAGE [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:45 -------------------------------------------------------------------------------------------------- [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 None kserve-ci-e2e-test Normal Scheduled Successfully assigned kserve-ci-e2e-test/auth-disabled-test-kserve-658fdbccfc-7k7fm to ip-10-0-139-47.ec2.internal [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:28:27 kserve-ci-e2e-test multus Normal AddedInterface Add eth0 [10.134.0.29/23] from ovn-kubernetes [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:28:28 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Pulled Container image "quay.io/opendatahub/kserve-storage-initializer@sha256:13ad765c0f323200e1db7b45c99661287e65c5158d6b54b16139829636d20be4" already present on machine [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:28:28 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Created Created container: storage-initializer [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:28:28 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Started Started container storage-initializer [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:28:31 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Pulled Container image "public.ecr.aws/q9t5s3a7/vllm-cpu-release-repo:v0.19.0" already present on machine [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:28:32 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Created Created container: main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:28:32 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Started Started container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:30:17 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Warning Unhealthy Startup probe failed: Get "https://10.134.0.29:8000/health": dial tcp 10.134.0.29:8000: connect: connection refused [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:30:32 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Killing Stopping container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:30:47 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Warning Unhealthy Readiness probe failed: Get "https://10.134.0.29:8000/health": dial tcp 10.134.0.29:8000: connect: connection refused [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:28:27 kserve-ci-e2e-test replicaset-controller Normal SuccessfulCreate Created pod: auth-disabled-test-kserve-658fdbccfc-7k7fm [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 None kserve-ci-e2e-test Normal Scheduled Successfully assigned kserve-ci-e2e-test/auth-disabled-test-kserve-router-scheduler-58c676b848-mfsnv to ip-10-0-133-25.ec2.internal [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:28:28 kserve-ci-e2e-test multus Normal AddedInterface Add eth0 [10.132.0.51/23] from ovn-kubernetes [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:28:28 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Pulled Container image "quay.io/opendatahub/kserve-storage-initializer@sha256:13ad765c0f323200e1db7b45c99661287e65c5158d6b54b16139829636d20be4" already present on machine [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:28:28 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Created Created container: storage-initializer [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:28:28 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Started Started container storage-initializer [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:28:29 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Pulled Container image "ghcr.io/llm-d/llm-d-router-endpoint-picker:v0.9.0-rc.2" already present on machine [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:28:29 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Created Created container: main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:28:29 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Started Started container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:30:32 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Killing Stopping container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:30:32 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Warning FailedPreStopHook PreStopHook failed [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:28:28 kserve-ci-e2e-test replicaset-controller Normal SuccessfulCreate Created pod: auth-disabled-test-kserve-router-scheduler-58c676b848-mfsnv [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:28:27 kserve-ci-e2e-test deployment-controller Normal ScalingReplicaSet Scaled up replica set auth-disabled-test-kserve-router-scheduler-58c676b848 from 0 to 1 [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:28:27 kserve-ci-e2e-test deployment-controller Normal ScalingReplicaSet Scaled up replica set auth-disabled-test-kserve-658fdbccfc from 0 to 1 [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:28:27 kserve-ci-e2e-test OpenDataHubModelController Warning ReconcileError Failed to reconcile LLMInferenceService: 1 error occurred: * failed to get HTTPRoute for AuthPolicy auth-disabled-test-kserve-route-authn: failed to get HTTPRoute kserve-ci-e2e-test/auth-disabled-test-kserve-route: HTTPRoute.gateway.networking.k8s.io "auth-disabled-test-kserve-route" not found [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:28:27 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Secret kserve-ci-e2e-test/auth-disabled-test-kserve-self-signed-certs [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:28:27 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Deployment kserve-ci-e2e-test/auth-disabled-test-kserve [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:28:27 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Service kserve-ci-e2e-test/auth-disabled-test-kserve-workload-svc [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:28:27 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.ServiceAccount kserve-ci-e2e-test/auth-disabled-test-epp-sa [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:28:27 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.ClusterRoleBinding /kserve-ci-e2e-test-auth-disabled-test-epp-auth-rb [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:28:27 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Role kserve-ci-e2e-test/auth-disabled-test-epp-role [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:28:27 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.RoleBinding kserve-ci-e2e-test/auth-disabled-test-epp-rb [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:28:27 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Deployment kserve-ci-e2e-test/auth-disabled-test-kserve-router-scheduler [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:28:27 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Service kserve-ci-e2e-test/auth-disabled-test-epp-service [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:28:38 kserve-ci-e2e-test LLMInferenceServiceController Normal Created (combined from similar events): Created v1.DestinationRule kserve-ci-e2e-test/auth-disabled-test-kserve-shadow-svc [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:28:38 kserve-ci-e2e-test LLMInferenceServiceController Normal Updated Updated v1.Secret kserve-ci-e2e-test/auth-disabled-test-kserve-self-signed-certs [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:28:38 kserve-ci-e2e-test LLMInferenceServiceController Normal Updated Updated v1.HTTPRoute kserve-ci-e2e-test/auth-disabled-test-kserve-route [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:30:28 kserve-ci-e2e-test LLMInferenceServiceController Normal LLMInferenceServiceReady LLMInferenceService [auth-disabled-test] is Ready [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:30:32 kserve-ci-e2e-test LLMInferenceServiceController Normal Deleted Deleted v1.ClusterRoleBinding /kserve-ci-e2e-test-auth-disabled-test-epp-auth-rb [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 None kserve-ci-e2e-test Normal Scheduled Successfully assigned kserve-ci-e2e-test/auth-enabled-test-kserve-87bd45bf5-zxvsp to ip-10-0-139-47.ec2.internal [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:33 kserve-ci-e2e-test multus Normal AddedInterface Add eth0 [10.134.0.24/23] from ovn-kubernetes [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:33 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Pulled Container image "quay.io/opendatahub/kserve-storage-initializer@sha256:13ad765c0f323200e1db7b45c99661287e65c5158d6b54b16139829636d20be4" already present on machine [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:33 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Created Created container: storage-initializer [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:33 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Started Started container storage-initializer [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:39 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Pulling Pulling image "public.ecr.aws/q9t5s3a7/vllm-cpu-release-repo:v0.19.0" [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:23:08 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Pulled Successfully pulled image "public.ecr.aws/q9t5s3a7/vllm-cpu-release-repo:v0.19.0" in 28.615s (28.615s including waiting). Image size: 3531177328 bytes. [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:23:08 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Created Created container: main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:23:08 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Started Started container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:24:53 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Warning Unhealthy Startup probe failed: Get "https://10.134.0.24:8000/health": dial tcp 10.134.0.24:8000: connect: connection refused [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:10 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Killing Stopping container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:25 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Warning Unhealthy Readiness probe failed: Get "https://10.134.0.24:8000/health": dial tcp 10.134.0.24:8000: connect: connection refused [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:32 kserve-ci-e2e-test replicaset-controller Normal SuccessfulCreate Created pod: auth-enabled-test-kserve-87bd45bf5-zxvsp [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 None kserve-ci-e2e-test Normal Scheduled Successfully assigned kserve-ci-e2e-test/auth-enabled-test-kserve-router-scheduler-6bbc585b87-nf5lr to ip-10-0-133-25.ec2.internal [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:33 kserve-ci-e2e-test multus Normal AddedInterface Add eth0 [10.132.0.47/23] from ovn-kubernetes [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:33 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Pulled Container image "quay.io/opendatahub/kserve-storage-initializer@sha256:13ad765c0f323200e1db7b45c99661287e65c5158d6b54b16139829636d20be4" already present on machine [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:33 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Created Created container: storage-initializer [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:33 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Started Started container storage-initializer [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:34 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Pulling Pulling image "ghcr.io/llm-d/llm-d-router-endpoint-picker:v0.9.0-rc.2" [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:35 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Pulled Successfully pulled image "ghcr.io/llm-d/llm-d-router-endpoint-picker:v0.9.0-rc.2" in 1.131s (1.131s including waiting). Image size: 75073927 bytes. [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:35 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Created Created container: main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:35 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Started Started container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:10 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Killing Stopping container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:10 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Warning FailedPreStopHook PreStopHook failed [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:33 kserve-ci-e2e-test replicaset-controller Normal SuccessfulCreate Created pod: auth-enabled-test-kserve-router-scheduler-6bbc585b87-nf5lr [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:33 kserve-ci-e2e-test deployment-controller Normal ScalingReplicaSet Scaled up replica set auth-enabled-test-kserve-router-scheduler-6bbc585b87 from 0 to 1 [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:32 kserve-ci-e2e-test deployment-controller Normal ScalingReplicaSet Scaled up replica set auth-enabled-test-kserve-87bd45bf5 from 0 to 1 [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:32 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Secret kserve-ci-e2e-test/auth-enabled-test-kserve-self-signed-certs [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:32 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Deployment kserve-ci-e2e-test/auth-enabled-test-kserve [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:32 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Service kserve-ci-e2e-test/auth-enabled-test-kserve-workload-svc [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:32 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.ServiceAccount kserve-ci-e2e-test/auth-enabled-test-epp-sa [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:33 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.ClusterRoleBinding /kserve-ci-e2e-test-auth-enabled-test-epp-auth-rb [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:33 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Role kserve-ci-e2e-test/auth-enabled-test-epp-role [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:33 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.RoleBinding kserve-ci-e2e-test/auth-enabled-test-epp-rb [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:33 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Deployment kserve-ci-e2e-test/auth-enabled-test-kserve-router-scheduler [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:33 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Service kserve-ci-e2e-test/auth-enabled-test-epp-service [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:43 kserve-ci-e2e-test LLMInferenceServiceController Normal Created (combined from similar events): Created v1.DestinationRule kserve-ci-e2e-test/auth-enabled-test-kserve-shadow-svc [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:42 kserve-ci-e2e-test LLMInferenceServiceController Normal Updated Updated v1.Secret kserve-ci-e2e-test/auth-enabled-test-kserve-self-signed-certs [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:43 kserve-ci-e2e-test LLMInferenceServiceController Normal Updated Updated v1.HTTPRoute kserve-ci-e2e-test/auth-enabled-test-kserve-route [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:05 kserve-ci-e2e-test LLMInferenceServiceController Normal LLMInferenceServiceReady LLMInferenceService [auth-enabled-test] is Ready [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:10 kserve-ci-e2e-test LLMInferenceServiceController Normal Deleted Deleted v1.ClusterRoleBinding /kserve-ci-e2e-test-auth-enabled-test-epp-auth-rb [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 None kserve-ci-e2e-test Normal Scheduled Successfully assigned kserve-ci-e2e-test/auth-invalid-token-test-kserve-6467b8fbf9-jhlvl to ip-10-0-139-47.ec2.internal [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:44 kserve-ci-e2e-test multus Normal AddedInterface Add eth0 [10.134.0.27/23] from ovn-kubernetes [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:44 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Pulled Container image "quay.io/opendatahub/kserve-storage-initializer@sha256:13ad765c0f323200e1db7b45c99661287e65c5158d6b54b16139829636d20be4" already present on machine [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:44 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Created Created container: storage-initializer [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:44 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Started Started container storage-initializer [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:49 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Pulled Container image "public.ecr.aws/q9t5s3a7/vllm-cpu-release-repo:v0.19.0" already present on machine [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:49 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Created Created container: main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:49 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Started Started container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:34 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Warning Unhealthy Startup probe failed: Get "https://10.134.0.27:8000/health": dial tcp 10.134.0.27:8000: connect: connection refused [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:46 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Killing Stopping container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:28:03 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Warning Unhealthy Readiness probe failed: Get "https://10.134.0.27:8000/health": context deadline exceeded [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:43 kserve-ci-e2e-test replicaset-controller Normal SuccessfulCreate Created pod: auth-invalid-token-test-kserve-6467b8fbf9-jhlvl [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:43 kserve-ci-e2e-test replicaset-controller Normal SuccessfulCreate Created pod: auth-invalid-token-test-kserve-router-scheduler-7c68b57cd5h6v2p [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 None kserve-ci-e2e-test Normal Scheduled Successfully assigned kserve-ci-e2e-test/auth-invalid-token-test-kserve-router-scheduler-7c68b57cd5h6v2p to ip-10-0-133-25.ec2.internal [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:44 kserve-ci-e2e-test multus Normal AddedInterface Add eth0 [10.132.0.49/23] from ovn-kubernetes [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:44 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Pulled Container image "quay.io/opendatahub/kserve-storage-initializer@sha256:13ad765c0f323200e1db7b45c99661287e65c5158d6b54b16139829636d20be4" already present on machine [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:44 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Created Created container: storage-initializer [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:44 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Started Started container storage-initializer [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:45 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Pulled Container image "ghcr.io/llm-d/llm-d-router-endpoint-picker:v0.9.0-rc.2" already present on machine [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:45 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Created Created container: main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:45 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Started Started container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:46 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Killing Stopping container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:46 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Warning FailedPreStopHook PreStopHook failed [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:43 kserve-ci-e2e-test deployment-controller Normal ScalingReplicaSet Scaled up replica set auth-invalid-token-test-kserve-router-scheduler-7c68b57cd5 from 0 to 1 [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:43 kserve-ci-e2e-test deployment-controller Normal ScalingReplicaSet Scaled up replica set auth-invalid-token-test-kserve-6467b8fbf9 from 0 to 1 [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:43 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Secret kserve-ci-e2e-test/auth-invalid-token-test-kserve-self-signed-certs [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:43 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Deployment kserve-ci-e2e-test/auth-invalid-token-test-kserve [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:43 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Service kserve-ci-e2e-test/auth-invalid-token-test-kserve-workload-svc [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:43 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.ServiceAccount kserve-ci-e2e-test/auth-invalid-token-test-epp-sa [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:43 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.ClusterRoleBinding /kserve-ci-e2e-test-auth-invalid-token-test-epp-auth-rb [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:43 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Role kserve-ci-e2e-test/auth-invalid-token-test-epp-role [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:43 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.RoleBinding kserve-ci-e2e-test/auth-invalid-token-test-epp-rb [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:43 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Deployment kserve-ci-e2e-test/auth-invalid-token-test-kserve-router-scheduler [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:43 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Service kserve-ci-e2e-test/auth-invalid-token-test-epp-service [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:26:10 kserve-ci-e2e-test LLMInferenceServiceController Normal Created (combined from similar events): Created v1.DestinationRule kserve-ci-e2e-test/auth-invalid-token-test-kserve-shadow-svc [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:26:09 kserve-ci-e2e-test LLMInferenceServiceController Normal Updated Updated v1.Secret kserve-ci-e2e-test/auth-invalid-token-test-kserve-self-signed-certs [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:26:09 kserve-ci-e2e-test LLMInferenceServiceController Normal Updated Updated v1.HTTPRoute kserve-ci-e2e-test/auth-invalid-token-test-kserve-route [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:45 kserve-ci-e2e-test LLMInferenceServiceController Normal LLMInferenceServiceReady LLMInferenceService [auth-invalid-token-test] is Ready [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:46 kserve-ci-e2e-test LLMInferenceServiceController Normal Deleted Deleted v1.ClusterRoleBinding /kserve-ci-e2e-test-auth-invalid-token-test-epp-auth-rb [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:45 kserve-ci-e2e-test replicaset-controller Normal SuccessfulCreate Created pod: gw-section-name-router-with-gat-2f0a622e-kserve-7c9c8cffc558295 [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 None kserve-ci-e2e-test Normal Scheduled Successfully assigned kserve-ci-e2e-test/gw-section-name-router-with-gat-2f0a622e-kserve-7c9c8cffc558295 to ip-10-0-139-47.ec2.internal [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:46 kserve-ci-e2e-test multus Normal AddedInterface Add eth0 [10.134.0.21/23] from ovn-kubernetes [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:46 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Pulling Pulling image "ghcr.io/llm-d/llm-d-inference-sim:v0.8.2" [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:47 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Pulled Successfully pulled image "ghcr.io/llm-d/llm-d-inference-sim:v0.8.2" in 1.195s (1.195s including waiting). Image size: 98346788 bytes. [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:47 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Created Created container: main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:47 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Started Started container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:58 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Killing Stopping container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:45 kserve-ci-e2e-test deployment-controller Normal ScalingReplicaSet Scaled up replica set gw-section-name-router-with-gat-2f0a622e-kserve-7c9c8cffc5 from 0 to 1 [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:45 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Secret kserve-ci-e2e-test/gw-sec0c69dceeb48768325d1a53a749e65786-kserve-self-signed-certs [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:45 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Deployment kserve-ci-e2e-test/gw-section-name-router-with-gat-2f0a622e-kserve [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:45 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Service kserve-ci-e2e-test/gw-section-name-router-with-gat-2f0a622e-kserve-workload-svc [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:45 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.HTTPRoute kserve-ci-e2e-test/gw-section-name-router-with-gat-2f0a622e-kserve-route [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:46 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.DestinationRule kserve-ci-e2e-test/gw-section-name-router-with-gat-2f0a622e-kserve-workload-svc [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:10 kserve-ci-e2e-test replicaset-controller Normal SuccessfulCreate Created pod: gw-section-name-router-with-gat-f1d92d0f-kserve-7bc8dd6c5bcr4rf [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 None kserve-ci-e2e-test Normal Scheduled Successfully assigned kserve-ci-e2e-test/gw-section-name-router-with-gat-f1d92d0f-kserve-7bc8dd6c5bcr4rf to ip-10-0-133-25.ec2.internal [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:10 kserve-ci-e2e-test multus Normal AddedInterface Add eth0 [10.132.0.46/23] from ovn-kubernetes [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:10 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Pulled Container image "ghcr.io/llm-d/llm-d-inference-sim:v0.8.2" already present on machine [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:10 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Created Created container: main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:10 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Started Started container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:27 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Killing Stopping container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:10 kserve-ci-e2e-test deployment-controller Normal ScalingReplicaSet Scaled up replica set gw-section-name-router-with-gat-f1d92d0f-kserve-7bc8dd6c5b from 0 to 1 [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:10 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Secret kserve-ci-e2e-test/gw-sec2774c263d49959f50d9eebc552e13bf9-kserve-self-signed-certs [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:10 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Deployment kserve-ci-e2e-test/gw-section-name-router-with-gat-f1d92d0f-kserve [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:10 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Service kserve-ci-e2e-test/gw-section-name-router-with-gat-f1d92d0f-kserve-workload-svc [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:10 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.HTTPRoute kserve-ci-e2e-test/gw-section-name-router-with-gat-f1d92d0f-kserve-route [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:10 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.DestinationRule kserve-ci-e2e-test/gw-section-name-router-with-gat-f1d92d0f-kserve-workload-svc [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:31:40 kserve-ci-e2e-test replicaset-controller Normal SuccessfulCreate Created pod: llmisvc-model-fb-opt-125m-route-87882a8e-kserve-5bd7864c6ctnql9 [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 None kserve-ci-e2e-test Normal Scheduled Successfully assigned kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-route-87882a8e-kserve-5bd7864c6ctnql9 to ip-10-0-139-47.ec2.internal [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:31:40 kserve-ci-e2e-test multus Normal AddedInterface Add eth0 [10.134.0.31/23] from ovn-kubernetes [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:31:40 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Pulled Container image "quay.io/opendatahub/kserve-storage-initializer@sha256:13ad765c0f323200e1db7b45c99661287e65c5158d6b54b16139829636d20be4" already present on machine [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:31:40 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Created Created container: storage-initializer [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:31:40 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Started Started container storage-initializer [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:31:45 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Pulled Container image "public.ecr.aws/q9t5s3a7/vllm-cpu-release-repo:v0.19.0" already present on machine [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:31:45 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Created Created container: main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:31:45 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Started Started container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:33:30 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Warning Unhealthy Startup probe failed: Get "https://10.134.0.31:8000/health": dial tcp 10.134.0.31:8000: connect: connection refused [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:31:40 kserve-ci-e2e-test deployment-controller Normal ScalingReplicaSet Scaled up replica set llmisvc-model-fb-opt-125m-route-87882a8e-kserve-5bd7864c6c from 0 to 1 [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:31:38 kserve-ci-e2e-test OpenDataHubModelController Warning ReconcileError Failed to reconcile LLMInferenceService: 1 error occurred: * failed to get HTTPRoute for AuthPolicy llmisvc-model-fb-opt-125m-route-87882a8e-kserve-route-authn: failed to get HTTPRoute kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-route-87882a8e-kserve-route: HTTPRoute.gateway.networking.k8s.io "llmisvc-model-fb-opt-125m-route-87882a8e-kserve-route" not found [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:31:40 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Secret kserve-ci-e2e-test/llmisva690bbc929faec8bc98c767f16c003c1-kserve-self-signed-certs [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:31:40 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Deployment kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-route-87882a8e-kserve [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:31:40 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Service kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-route-87882a8e-kserve-workload-svc [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:31:40 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.ServiceAccount kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-route-87882a8e-epp-sa [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:31:40 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.ClusterRoleBinding /kserve-ci-e2e-test21fe6730fe484f3a92b1a16afe1bac8f-llmisvc-mode [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:31:40 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Role kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-route-87882a8e-epp-role [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:31:40 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.RoleBinding kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-route-87882a8e-epp-rb [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:31:40 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Deployment kserve-ci-e2e-test/llmisvca690bbc929faec8bc98c767f16c003c1-kserve-router-scheduler [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:31:40 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Service kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-route-87882a8e-epp-service [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:31:47 kserve-ci-e2e-test LLMInferenceServiceController Normal Created (combined from similar events): Created v1.DestinationRule kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-route-87882a8e-kserve-shadow-svc [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:31:45 kserve-ci-e2e-test LLMInferenceServiceController Normal Updated Updated v1.Secret kserve-ci-e2e-test/llmisva690bbc929faec8bc98c767f16c003c1-kserve-self-signed-certs [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:31:46 kserve-ci-e2e-test LLMInferenceServiceController Normal Updated Updated v1.HTTPRoute kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-route-87882a8e-kserve-route [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:33:40 kserve-ci-e2e-test LLMInferenceServiceController Normal LLMInferenceServiceReady LLMInferenceService [llmisvc-model-fb-opt-125m-route-87882a8e] is Ready [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:31:02 kserve-ci-e2e-test replicaset-controller Normal SuccessfulCreate Created pod: llmisvc-model-fb-opt-125m-route-e95b1dc1-kserve-66874c76d46pqbc [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 None kserve-ci-e2e-test Normal Scheduled Successfully assigned kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-route-e95b1dc1-kserve-66874c76d46pqbc to ip-10-0-135-220.ec2.internal [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:31:02 kserve-ci-e2e-test multus Normal AddedInterface Add eth0 [10.133.0.37/23] from ovn-kubernetes [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:31:02 kserve-ci-e2e-test kubelet/ip-10-0-135-220.ec2.internal Normal Pulled Container image "ghcr.io/llm-d/llm-d-inference-sim:v0.8.2" already present on machine [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:31:02 kserve-ci-e2e-test kubelet/ip-10-0-135-220.ec2.internal Normal Created Created container: main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:31:02 kserve-ci-e2e-test kubelet/ip-10-0-135-220.ec2.internal Normal Started Started container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:31:15 kserve-ci-e2e-test kubelet/ip-10-0-135-220.ec2.internal Normal Killing Stopping container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:31:02 kserve-ci-e2e-test deployment-controller Normal ScalingReplicaSet Scaled up replica set llmisvc-model-fb-opt-125m-route-e95b1dc1-kserve-66874c76d4 from 0 to 1 [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:30:58 kserve-ci-e2e-test OpenDataHubModelController Warning ReconcileError Failed to reconcile LLMInferenceService: 1 error occurred: * failed to get HTTPRoute for AuthPolicy llmisvc-model-fb-opt-125m-route-e95b1dc1-kserve-route-authn: failed to get HTTPRoute kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-route-e95b1dc1-kserve-route: HTTPRoute.gateway.networking.k8s.io "llmisvc-model-fb-opt-125m-route-e95b1dc1-kserve-route" not found [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:31:02 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Secret kserve-ci-e2e-test/llmisv122f03714c5bdf915a2917fdf1262b98-kserve-self-signed-certs [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:31:02 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Deployment kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-route-e95b1dc1-kserve [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:31:02 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Service kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-route-e95b1dc1-kserve-workload-svc [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:31:02 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.HTTPRoute kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-route-e95b1dc1-kserve-route [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:31:02 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.DestinationRule kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-route-e95b1dc1-kserve-workload-svc [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:31:14 kserve-ci-e2e-test LLMInferenceServiceController Normal Updated Updated v1.Secret kserve-ci-e2e-test/llmisv122f03714c5bdf915a2917fdf1262b98-kserve-self-signed-certs [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:31:14 kserve-ci-e2e-test LLMInferenceServiceController Normal LLMInferenceServiceReady LLMInferenceService [llmisvc-model-fb-opt-125m-route-e95b1dc1] is Ready [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 None kserve-ci-e2e-test Normal Scheduled Successfully assigned kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-with-7ca60146-kserve-7cfdf49d74-pd9px to ip-10-0-139-47.ec2.internal [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:29:25 kserve-ci-e2e-test multus Normal AddedInterface Add eth0 [10.134.0.30/23] from ovn-kubernetes [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:29:25 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Pulled Container image "quay.io/opendatahub/kserve-storage-initializer@sha256:13ad765c0f323200e1db7b45c99661287e65c5158d6b54b16139829636d20be4" already present on machine [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:29:25 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Created Created container: storage-initializer [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:29:25 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Started Started container storage-initializer [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:29:30 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Pulled Container image "public.ecr.aws/q9t5s3a7/vllm-cpu-release-repo:v0.19.0" already present on machine [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:29:30 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Created Created container: main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:29:30 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Started Started container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:30:34 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Warning Unhealthy Startup probe failed: Get "https://10.134.0.30:8000/health": dial tcp 10.134.0.30:8000: connect: connection refused [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:29:24 kserve-ci-e2e-test replicaset-controller Normal SuccessfulCreate Created pod: llmisvc-model-fb-opt-125m-with-7ca60146-kserve-7cfdf49d74-pd9px [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:29:24 kserve-ci-e2e-test replicaset-controller Normal SuccessfulCreate Created pod: llmisvc-model-fb-opt-125m-with-7ca60146-kserve-router-schenp5mk [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:29:24 kserve-ci-e2e-test deployment-controller Normal ScalingReplicaSet Scaled up replica set llmisvc-model-fb-opt-125m-with-7ca60146-kserve-router-scheduler-778547fbc from 0 to 1 [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 None kserve-ci-e2e-test Normal Scheduled Successfully assigned kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-with-7ca60146-kserve-router-schenp5mk to ip-10-0-133-25.ec2.internal [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:29:25 kserve-ci-e2e-test multus Normal AddedInterface Add eth0 [10.132.0.52/23] from ovn-kubernetes [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:29:25 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Pulled Container image "quay.io/opendatahub/kserve-storage-initializer@sha256:13ad765c0f323200e1db7b45c99661287e65c5158d6b54b16139829636d20be4" already present on machine [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:29:25 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Created Created container: storage-initializer [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:29:25 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Started Started container storage-initializer [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:29:26 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Pulled Container image "ghcr.io/llm-d/llm-d-router-endpoint-picker:v0.9.0-rc.2" already present on machine [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:29:27 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Created Created container: main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:29:27 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Started Started container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:29:24 kserve-ci-e2e-test deployment-controller Normal ScalingReplicaSet Scaled up replica set llmisvc-model-fb-opt-125m-with-7ca60146-kserve-7cfdf49d74 from 0 to 1 [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:29:23 kserve-ci-e2e-test OpenDataHubModelController Warning ReconcileError Failed to reconcile LLMInferenceService: 1 error occurred: * failed to get HTTPRoute for AuthPolicy llmisvc-model-fb-opt-125m-with-7ca60146-kserve-route-authn: failed to get HTTPRoute kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-with-7ca60146-kserve-route: HTTPRoute.gateway.networking.k8s.io "llmisvc-model-fb-opt-125m-with-7ca60146-kserve-route" not found [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:29:24 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Secret kserve-ci-e2e-test/llmisv3e414c2ba058a022dfd694dbcbac5b51-kserve-self-signed-certs [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:29:24 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Deployment kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-with-7ca60146-kserve [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:29:24 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Service kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-with-7ca60146-kserve-workload-svc [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:29:24 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.ServiceAccount kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-with-7ca60146-epp-sa [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:29:24 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.ClusterRoleBinding /kserve-ci-e2e-test5216bfd716f919dc046bc693ceb22e41-llmisvc-mode [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:29:24 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Role kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-with-7ca60146-epp-role [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:29:24 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.RoleBinding kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-with-7ca60146-epp-rb [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:29:24 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Deployment kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-with-7ca60146-kserve-router-scheduler [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:29:24 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Service kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-with-7ca60146-epp-service [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:29:35 kserve-ci-e2e-test LLMInferenceServiceController Normal Created (combined from similar events): Created v1.DestinationRule kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-with-7ca60146-kserve-shadow-svc [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:29:33 kserve-ci-e2e-test LLMInferenceServiceController Normal Updated Updated v1.Secret kserve-ci-e2e-test/llmisv3e414c2ba058a022dfd694dbcbac5b51-kserve-self-signed-certs [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:29:34 kserve-ci-e2e-test LLMInferenceServiceController Normal Updated Updated v1.HTTPRoute kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-with-7ca60146-kserve-route [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:30:45 kserve-ci-e2e-test LLMInferenceServiceController Normal LLMInferenceServiceReady LLMInferenceService [llmisvc-model-fb-opt-125m-with-7ca60146] is Ready [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:54 kserve-ci-e2e-test replicaset-controller Normal SuccessfulCreate Created pod: llmisvc-model-qwen2-5-0-5b-rout-4f8c0978-kserve-5c54ddb98flf892 [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 None kserve-ci-e2e-test Normal Scheduled Successfully assigned kserve-ci-e2e-test/llmisvc-model-qwen2-5-0-5b-rout-4f8c0978-kserve-5c54ddb98flf892 to ip-10-0-135-220.ec2.internal [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:54 kserve-ci-e2e-test multus Normal AddedInterface Add eth0 [10.133.0.36/23] from ovn-kubernetes [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:54 kserve-ci-e2e-test kubelet/ip-10-0-135-220.ec2.internal Normal Pulled Container image "ghcr.io/llm-d/llm-d-inference-sim:v0.8.2" already present on machine [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:54 kserve-ci-e2e-test kubelet/ip-10-0-135-220.ec2.internal Normal Created Created container: main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:54 kserve-ci-e2e-test kubelet/ip-10-0-135-220.ec2.internal Normal Started Started container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:28:40 kserve-ci-e2e-test kubelet/ip-10-0-135-220.ec2.internal Normal Killing Stopping container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:28:56 kserve-ci-e2e-test kubelet/ip-10-0-135-220.ec2.internal Warning Unhealthy Readiness probe failed: Get "https://10.133.0.36:8000/health": context deadline exceeded [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:54 kserve-ci-e2e-test deployment-controller Normal ScalingReplicaSet Scaled up replica set llmisvc-model-qwen2-5-0-5b-rout-4f8c0978-kserve-5c54ddb98f from 0 to 1 [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:53 kserve-ci-e2e-test OpenDataHubModelController Warning ReconcileError Failed to reconcile LLMInferenceService: 1 error occurred: * failed to get HTTPRoute for AuthPolicy llmisvc-model-qwen2-5-0-5b-rout-4f8c0978-kserve-route-authn: failed to get HTTPRoute kserve-ci-e2e-test/llmisvc-model-qwen2-5-0-5b-rout-4f8c0978-kserve-route: HTTPRoute.gateway.networking.k8s.io "llmisvc-model-qwen2-5-0-5b-rout-4f8c0978-kserve-route" not found [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:54 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Secret kserve-ci-e2e-test/llmisv08544b88a8d963ffd553cc1f3ed82d16-kserve-self-signed-certs [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:54 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Deployment kserve-ci-e2e-test/llmisvc-model-qwen2-5-0-5b-rout-4f8c0978-kserve [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:54 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Service kserve-ci-e2e-test/llmisvc-model-qwen2-5-0-5b-rout-4f8c0978-kserve-workload-svc [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:54 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.ServiceAccount kserve-ci-e2e-test/llmisvc-model-qwen2-5-0-5b-rout-4f8c0978-epp-sa [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:54 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.ClusterRoleBinding /kserve-ci-e2e-test76d7f532acb694e4a7bcef75d32cd8a1-llmisvc-mode [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:54 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Role kserve-ci-e2e-test/llmisvc-model-qwen2-5-0-5b-rout-4f8c0978-epp-role [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:54 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.RoleBinding kserve-ci-e2e-test/llmisvc-model-qwen2-5-0-5b-rout-4f8c0978-epp-rb [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:54 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Deployment kserve-ci-e2e-test/llmisvc08544b88a8d963ffd553cc1f3ed82d16-kserve-router-scheduler [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:54 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Service kserve-ci-e2e-test/llmisvc-model-qwen2-5-0-5b-rout-4f8c0978-epp-service [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:28:06 kserve-ci-e2e-test LLMInferenceServiceController Normal Created (combined from similar events): Created v1.DestinationRule kserve-ci-e2e-test/llmisvc-model-qwen2-5-0-5b-rout-4f8c0978-kserve-shadow-svc [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:28:05 kserve-ci-e2e-test LLMInferenceServiceController Normal Updated Updated v1.Secret kserve-ci-e2e-test/llmisv08544b88a8d963ffd553cc1f3ed82d16-kserve-self-signed-certs [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:28:06 kserve-ci-e2e-test LLMInferenceServiceController Normal Updated Updated v1.HTTPRoute kserve-ci-e2e-test/llmisvc-model-qwen2-5-0-5b-rout-4f8c0978-kserve-route [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:28:31 kserve-ci-e2e-test LLMInferenceServiceController Normal LLMInferenceServiceReady LLMInferenceService [llmisvc-model-qwen2-5-0-5b-rout-4f8c0978] is Ready [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:28:40 kserve-ci-e2e-test LLMInferenceServiceController Normal Deleted Deleted v1.ClusterRoleBinding /kserve-ci-e2e-test76d7f532acb694e4a7bcef75d32cd8a1-llmisvc-mode [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:23 kserve-ci-e2e-test replicaset-controller Normal SuccessfulCreate Created pod: llmisvc-model-qwen2-5-0-5b-rout-a50492e9-kserve-5d8ffd58ddqnd8m [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 None kserve-ci-e2e-test Normal Scheduled Successfully assigned kserve-ci-e2e-test/llmisvc-model-qwen2-5-0-5b-rout-a50492e9-kserve-5d8ffd58ddqnd8m to ip-10-0-135-220.ec2.internal [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:23 kserve-ci-e2e-test multus Normal AddedInterface Add eth0 [10.133.0.34/23] from ovn-kubernetes [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:23 kserve-ci-e2e-test kubelet/ip-10-0-135-220.ec2.internal Normal Pulled Container image "ghcr.io/llm-d/llm-d-inference-sim:v0.8.2" already present on machine [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:23 kserve-ci-e2e-test kubelet/ip-10-0-135-220.ec2.internal Normal Created Created container: main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:23 kserve-ci-e2e-test kubelet/ip-10-0-135-220.ec2.internal Normal Started Started container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:26:13 kserve-ci-e2e-test kubelet/ip-10-0-135-220.ec2.internal Normal Killing Stopping container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:23 kserve-ci-e2e-test deployment-controller Normal ScalingReplicaSet Scaled up replica set llmisvc-model-qwen2-5-0-5b-rout-a50492e9-kserve-5d8ffd58dd from 0 to 1 [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:21 kserve-ci-e2e-test OpenDataHubModelController Warning ReconcileError Failed to reconcile LLMInferenceService: 1 error occurred: * failed to get HTTPRoute for AuthPolicy llmisvc-model-qwen2-5-0-5b-rout-a50492e9-kserve-route-authn: failed to get HTTPRoute kserve-ci-e2e-test/llmisvc-model-qwen2-5-0-5b-rout-a50492e9-kserve-route: HTTPRoute.gateway.networking.k8s.io "llmisvc-model-qwen2-5-0-5b-rout-a50492e9-kserve-route" not found [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:23 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Secret kserve-ci-e2e-test/llmisvb11a9c9832b99b016bc8f8e0ea095712-kserve-self-signed-certs [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:23 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Deployment kserve-ci-e2e-test/llmisvc-model-qwen2-5-0-5b-rout-a50492e9-kserve [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:23 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Service kserve-ci-e2e-test/llmisvc-model-qwen2-5-0-5b-rout-a50492e9-kserve-workload-svc [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:23 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.ServiceAccount kserve-ci-e2e-test/llmisvc-model-qwen2-5-0-5b-rout-a50492e9-epp-sa [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:23 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.ClusterRoleBinding /kserve-ci-e2e-testb7025ca4d8a6f8f5b2fd08b5581d2678-llmisvc-mode [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:23 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Role kserve-ci-e2e-test/llmisvc-model-qwen2-5-0-5b-rout-a50492e9-epp-role [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:23 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.RoleBinding kserve-ci-e2e-test/llmisvc-model-qwen2-5-0-5b-rout-a50492e9-epp-rb [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:23 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Deployment kserve-ci-e2e-test/llmisvcb11a9c9832b99b016bc8f8e0ea095712-kserve-router-scheduler [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:23 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Service kserve-ci-e2e-test/llmisvc-model-qwen2-5-0-5b-rout-a50492e9-epp-service [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:34 kserve-ci-e2e-test LLMInferenceServiceController Normal Created (combined from similar events): Created v1.DestinationRule kserve-ci-e2e-test/llmisvc-model-qwen2-5-0-5b-rout-a50492e9-kserve-shadow-svc [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:33 kserve-ci-e2e-test LLMInferenceServiceController Normal Updated Updated v1.Secret kserve-ci-e2e-test/llmisvb11a9c9832b99b016bc8f8e0ea095712-kserve-self-signed-certs [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:33 kserve-ci-e2e-test LLMInferenceServiceController Normal Updated Updated v1.HTTPRoute kserve-ci-e2e-test/llmisvc-model-qwen2-5-0-5b-rout-a50492e9-kserve-route [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:26:12 kserve-ci-e2e-test LLMInferenceServiceController Normal LLMInferenceServiceReady LLMInferenceService [llmisvc-model-qwen2-5-0-5b-rout-a50492e9] is Ready [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:26:13 kserve-ci-e2e-test LLMInferenceServiceController Normal Deleted Deleted v1.ClusterRoleBinding /kserve-ci-e2e-testb7025ca4d8a6f8f5b2fd08b5581d2678-llmisvc-mode [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 None kserve-ci-e2e-test Normal Scheduled Successfully assigned kserve-ci-e2e-test/llmisvc-router-managed-test-llm-4b931143-kserve-bd545d56-jcs4n to ip-10-0-135-220.ec2.internal [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:18 kserve-ci-e2e-test multus Normal AddedInterface Add eth0 [10.133.0.35/23] from ovn-kubernetes [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:18 kserve-ci-e2e-test kubelet/ip-10-0-135-220.ec2.internal Normal Pulled Container image "ghcr.io/llm-d/llm-d-inference-sim:v0.8.2" already present on machine [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:18 kserve-ci-e2e-test kubelet/ip-10-0-135-220.ec2.internal Normal Created Created container: main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:18 kserve-ci-e2e-test kubelet/ip-10-0-135-220.ec2.internal Normal Started Started container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:28:58 kserve-ci-e2e-test kubelet/ip-10-0-135-220.ec2.internal Normal Killing Stopping container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:17 kserve-ci-e2e-test replicaset-controller Normal SuccessfulCreate Created pod: llmisvc-router-managed-test-llm-4b931143-kserve-bd545d56-jcs4n [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:17 kserve-ci-e2e-test deployment-controller Normal ScalingReplicaSet Scaled up replica set llmisvc-router-managed-test-llm-4b931143-kserve-bd545d56 from 0 to 1 [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:14 kserve-ci-e2e-test OpenDataHubModelController Warning ReconcileError Failed to reconcile LLMInferenceService: 1 error occurred: * failed to get HTTPRoute for AuthPolicy llmisvc-router-managed-test-llm-4b931143-kserve-route-authn: failed to get HTTPRoute kserve-ci-e2e-test/llmisvc-router-managed-test-llm-4b931143-kserve-route: HTTPRoute.gateway.networking.k8s.io "llmisvc-router-managed-test-llm-4b931143-kserve-route" not found [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:17 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Secret kserve-ci-e2e-test/llmisvca2d2d7d499abb359505529ebe02c136-kserve-self-signed-certs [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:17 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Deployment kserve-ci-e2e-test/llmisvc-router-managed-test-llm-4b931143-kserve [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:17 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Service kserve-ci-e2e-test/llmisvc-router-managed-test-llm-4b931143-kserve-workload-svc [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:17 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.ServiceAccount kserve-ci-e2e-test/llmisvc-router-managed-test-llm-4b931143-epp-sa [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:17 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.ClusterRoleBinding /kserve-ci-e2e-test8ac8e3d2264ccb939eb021b0b835847c-llmisvc-rout [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:17 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Role kserve-ci-e2e-test/llmisvc-router-managed-test-llm-4b931143-epp-role [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:17 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.RoleBinding kserve-ci-e2e-test/llmisvc-router-managed-test-llm-4b931143-epp-rb [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:17 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Deployment kserve-ci-e2e-test/llmisvcca2d2d7d499abb359505529ebe02c136-kserve-router-scheduler [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:17 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Service kserve-ci-e2e-test/llmisvc-router-managed-test-llm-4b931143-epp-service [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:43 kserve-ci-e2e-test LLMInferenceServiceController Normal Created (combined from similar events): Created v1.DestinationRule kserve-ci-e2e-test/llmisvc-router-managed-test-llm-4b931143-kserve-shadow-svc [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:42 kserve-ci-e2e-test LLMInferenceServiceController Normal Updated Updated v1.Secret kserve-ci-e2e-test/llmisvca2d2d7d499abb359505529ebe02c136-kserve-self-signed-certs [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:43 kserve-ci-e2e-test LLMInferenceServiceController Normal Updated Updated v1.HTTPRoute kserve-ci-e2e-test/llmisvc-router-managed-test-llm-4b931143-kserve-route [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:50 kserve-ci-e2e-test LLMInferenceServiceController Normal LLMInferenceServiceReady LLMInferenceService [llmisvc-router-managed-test-llm-4b931143] is Ready [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:28:58 kserve-ci-e2e-test LLMInferenceServiceController Normal Deleted Deleted v1.ClusterRoleBinding /kserve-ci-e2e-test8ac8e3d2264ccb939eb021b0b835847c-llmisvc-rout [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:24:42 kserve-ci-e2e-test replicaset-controller Normal SuccessfulCreate Created pod: llmisvc-router-managed-test-llm-5b1e8f15-kserve-64df7bddb4xtf2m [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 None kserve-ci-e2e-test Normal Scheduled Successfully assigned kserve-ci-e2e-test/llmisvc-router-managed-test-llm-5b1e8f15-kserve-64df7bddb4xtf2m to ip-10-0-135-220.ec2.internal [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:24:43 kserve-ci-e2e-test multus Normal AddedInterface Add eth0 [10.133.0.33/23] from ovn-kubernetes [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:24:43 kserve-ci-e2e-test kubelet/ip-10-0-135-220.ec2.internal Normal Pulled Container image "ghcr.io/llm-d/llm-d-inference-sim:v0.8.2" already present on machine [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:24:43 kserve-ci-e2e-test kubelet/ip-10-0-135-220.ec2.internal Normal Created Created container: main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:24:43 kserve-ci-e2e-test kubelet/ip-10-0-135-220.ec2.internal Normal Started Started container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:26:33 kserve-ci-e2e-test kubelet/ip-10-0-135-220.ec2.internal Normal Killing Stopping container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:24:42 kserve-ci-e2e-test deployment-controller Normal ScalingReplicaSet Scaled up replica set llmisvc-router-managed-test-llm-5b1e8f15-kserve-64df7bddb4 from 0 to 1 [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:24:40 kserve-ci-e2e-test OpenDataHubModelController Warning ReconcileError Failed to reconcile LLMInferenceService: 1 error occurred: * failed to get HTTPRoute for AuthPolicy llmisvc-router-managed-test-llm-5b1e8f15-kserve-route-authn: failed to get HTTPRoute kserve-ci-e2e-test/llmisvc-router-managed-test-llm-5b1e8f15-kserve-route: HTTPRoute.gateway.networking.k8s.io "llmisvc-router-managed-test-llm-5b1e8f15-kserve-route" not found [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:24:42 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Secret kserve-ci-e2e-test/llmisve55ae740357a3a31a27cdb8b66ffe20f-kserve-self-signed-certs [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:24:42 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Deployment kserve-ci-e2e-test/llmisvc-router-managed-test-llm-5b1e8f15-kserve [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:24:42 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Service kserve-ci-e2e-test/llmisvc-router-managed-test-llm-5b1e8f15-kserve-workload-svc [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:24:42 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.ServiceAccount kserve-ci-e2e-test/llmisvc-router-managed-test-llm-5b1e8f15-epp-sa [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:24:42 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.ClusterRoleBinding /kserve-ci-e2e-test7f54e84970003a6e7372bdbcb574f7ed-llmisvc-rout [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:24:42 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Role kserve-ci-e2e-test/llmisvc-router-managed-test-llm-5b1e8f15-epp-role [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:24:43 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.RoleBinding kserve-ci-e2e-test/llmisvc-router-managed-test-llm-5b1e8f15-epp-rb [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:24:43 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Deployment kserve-ci-e2e-test/llmisvce55ae740357a3a31a27cdb8b66ffe20f-kserve-router-scheduler [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:24:43 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Service kserve-ci-e2e-test/llmisvc-router-managed-test-llm-5b1e8f15-epp-service [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:03 kserve-ci-e2e-test LLMInferenceServiceController Normal Created (combined from similar events): Created v1.DestinationRule kserve-ci-e2e-test/llmisvc-router-managed-test-llm-5b1e8f15-kserve-shadow-svc [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:03 kserve-ci-e2e-test LLMInferenceServiceController Normal Updated Updated v1.Secret kserve-ci-e2e-test/llmisve55ae740357a3a31a27cdb8b66ffe20f-kserve-self-signed-certs [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:03 kserve-ci-e2e-test LLMInferenceServiceController Normal Updated Updated v1.HTTPRoute kserve-ci-e2e-test/llmisvc-router-managed-test-llm-5b1e8f15-kserve-route [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:15 kserve-ci-e2e-test LLMInferenceServiceController Normal LLMInferenceServiceReady LLMInferenceService [llmisvc-router-managed-test-llm-5b1e8f15] is Ready [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:26:33 kserve-ci-e2e-test LLMInferenceServiceController Normal Deleted Deleted v1.ClusterRoleBinding /kserve-ci-e2e-test7f54e84970003a6e7372bdbcb574f7ed-llmisvc-rout [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:23:41 kserve-ci-e2e-test replicaset-controller Normal SuccessfulCreate Created pod: llmisvc-router-managed-test-llm-e45d1f79-kserve-7fdbbd4879ffz96 [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 None kserve-ci-e2e-test Normal Scheduled Successfully assigned kserve-ci-e2e-test/llmisvc-router-managed-test-llm-e45d1f79-kserve-7fdbbd4879ffz96 to ip-10-0-135-220.ec2.internal [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:23:41 kserve-ci-e2e-test multus Normal AddedInterface Add eth0 [10.133.0.32/23] from ovn-kubernetes [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:23:41 kserve-ci-e2e-test kubelet/ip-10-0-135-220.ec2.internal Normal Pulling Pulling image "ghcr.io/llm-d/llm-d-inference-sim:v0.8.2" [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:23:43 kserve-ci-e2e-test kubelet/ip-10-0-135-220.ec2.internal Normal Pulled Successfully pulled image "ghcr.io/llm-d/llm-d-inference-sim:v0.8.2" in 1.407s (1.407s including waiting). Image size: 98346788 bytes. [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:23:43 kserve-ci-e2e-test kubelet/ip-10-0-135-220.ec2.internal Normal Created Created container: main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:23:43 kserve-ci-e2e-test kubelet/ip-10-0-135-220.ec2.internal Normal Started Started container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:24:15 kserve-ci-e2e-test kubelet/ip-10-0-135-220.ec2.internal Normal Killing Stopping container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:23:41 kserve-ci-e2e-test deployment-controller Normal ScalingReplicaSet Scaled up replica set llmisvc-router-managed-test-llm-e45d1f79-kserve-7fdbbd4879 from 0 to 1 [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:23:36 kserve-ci-e2e-test OpenDataHubModelController Warning ReconcileError Failed to reconcile LLMInferenceService: 1 error occurred: * failed to get HTTPRoute for AuthPolicy llmisvc-router-managed-test-llm-e45d1f79-kserve-route-authn: failed to get HTTPRoute kserve-ci-e2e-test/llmisvc-router-managed-test-llm-e45d1f79-kserve-route: HTTPRoute.gateway.networking.k8s.io "llmisvc-router-managed-test-llm-e45d1f79-kserve-route" not found [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:23:41 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Secret kserve-ci-e2e-test/llmisv5c7e67b6c51568d1d6d13829a9337f2a-kserve-self-signed-certs [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:23:41 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Deployment kserve-ci-e2e-test/llmisvc-router-managed-test-llm-e45d1f79-kserve [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:23:41 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Service kserve-ci-e2e-test/llmisvc-router-managed-test-llm-e45d1f79-kserve-workload-svc [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:23:41 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.ServiceAccount kserve-ci-e2e-test/llmisvc-router-managed-test-llm-e45d1f79-epp-sa [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:23:41 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.ClusterRoleBinding /kserve-ci-e2e-testef4d2875be14b30dc1561ed84d0d4bde-llmisvc-rout [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:23:41 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Role kserve-ci-e2e-test/llmisvc-router-managed-test-llm-e45d1f79-epp-role [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:23:41 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.RoleBinding kserve-ci-e2e-test/llmisvc-router-managed-test-llm-e45d1f79-epp-rb [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:23:41 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Deployment kserve-ci-e2e-test/llmisvc5c7e67b6c51568d1d6d13829a9337f2a-kserve-router-scheduler [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:23:41 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Service kserve-ci-e2e-test/llmisvc-router-managed-test-llm-e45d1f79-epp-service [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:23:53 kserve-ci-e2e-test LLMInferenceServiceController Normal Created (combined from similar events): Created v1.DestinationRule kserve-ci-e2e-test/llmisvc-router-managed-test-llm-e45d1f79-kserve-shadow-svc [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:23:52 kserve-ci-e2e-test LLMInferenceServiceController Normal Updated Updated v1.Secret kserve-ci-e2e-test/llmisv5c7e67b6c51568d1d6d13829a9337f2a-kserve-self-signed-certs [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:23:52 kserve-ci-e2e-test LLMInferenceServiceController Normal Updated Updated v1.HTTPRoute kserve-ci-e2e-test/llmisvc-router-managed-test-llm-e45d1f79-kserve-route [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:24:14 kserve-ci-e2e-test LLMInferenceServiceController Normal LLMInferenceServiceReady LLMInferenceService [llmisvc-router-managed-test-llm-e45d1f79] is Ready [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:24:15 kserve-ci-e2e-test LLMInferenceServiceController Normal Deleted Deleted v1.ClusterRoleBinding /kserve-ci-e2e-testef4d2875be14b30dc1561ed84d0d4bde-llmisvc-rout [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:54 kserve-ci-e2e-test replicaset-controller Normal SuccessfulCreate Created pod: llmisvc08544b88a8d963ffd553cc1f3ed82d16-kserve-router-schefvtgh [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:54 kserve-ci-e2e-test deployment-controller Normal ScalingReplicaSet Scaled up replica set llmisvc08544b88a8d963ffd553cc1f3ed82d16-kserve-router-scheduler-6fcb489785 from 0 to 1 [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 None kserve-ci-e2e-test Normal Scheduled Successfully assigned kserve-ci-e2e-test/llmisvc08544b88a8d963ffd553cc1f3ed82d16-kserve-router-schefvtgh to ip-10-0-133-25.ec2.internal [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:55 kserve-ci-e2e-test multus Normal AddedInterface Add eth0 [10.132.0.50/23] from ovn-kubernetes [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:55 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Pulled Container image "ghcr.io/llm-d/llm-d-router-endpoint-picker:v0.9.0-rc.2" already present on machine [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:55 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Created Created container: main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:55 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Started Started container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:28:40 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Killing Stopping container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:28:40 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Warning FailedPreStopHook PreStopHook failed [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:23:41 kserve-ci-e2e-test replicaset-controller Normal SuccessfulCreate Created pod: llmisvc5c7e67b6c51568d1d6d13829a9337f2a-kserve-router-schexv898 [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:23:41 kserve-ci-e2e-test deployment-controller Normal ScalingReplicaSet Scaled up replica set llmisvc5c7e67b6c51568d1d6d13829a9337f2a-kserve-router-scheduler-5dd88bfbb7 from 0 to 1 [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 None kserve-ci-e2e-test Normal Scheduled Successfully assigned kserve-ci-e2e-test/llmisvc5c7e67b6c51568d1d6d13829a9337f2a-kserve-router-schexv898 to ip-10-0-139-47.ec2.internal [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:23:42 kserve-ci-e2e-test multus Normal AddedInterface Add eth0 [10.134.0.25/23] from ovn-kubernetes [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:23:42 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Pulled Container image "ghcr.io/llm-d/llm-d-router-endpoint-picker:v0.9.0-rc.2" already present on machine [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:23:42 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Created Created container: main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:23:42 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Started Started container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:24:15 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Killing Stopping container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:24:15 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Warning FailedPreStopHook PreStopHook failed [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:31:40 kserve-ci-e2e-test replicaset-controller Normal SuccessfulCreate Created pod: llmisvca690bbc929faec8bc98c767f16c003c1-kserve-router-scheq44kv [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:31:40 kserve-ci-e2e-test deployment-controller Normal ScalingReplicaSet Scaled up replica set llmisvca690bbc929faec8bc98c767f16c003c1-kserve-router-scheduler-5d6998448f from 0 to 1 [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 None kserve-ci-e2e-test Normal Scheduled Successfully assigned kserve-ci-e2e-test/llmisvca690bbc929faec8bc98c767f16c003c1-kserve-router-scheq44kv to ip-10-0-133-25.ec2.internal [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:31:40 kserve-ci-e2e-test multus Normal AddedInterface Add eth0 [10.132.0.53/23] from ovn-kubernetes [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:31:41 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Pulled Container image "quay.io/opendatahub/kserve-storage-initializer@sha256:13ad765c0f323200e1db7b45c99661287e65c5158d6b54b16139829636d20be4" already present on machine [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:31:41 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Created Created container: storage-initializer [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:31:41 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Started Started container storage-initializer [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:31:42 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Pulled Container image "ghcr.io/llm-d/llm-d-router-endpoint-picker:v0.9.0-rc.2" already present on machine [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:31:42 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Created Created container: main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:31:42 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Started Started container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:23 kserve-ci-e2e-test replicaset-controller Normal SuccessfulCreate Created pod: llmisvcb11a9c9832b99b016bc8f8e0ea095712-kserve-router-scheknzg7 [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:23 kserve-ci-e2e-test deployment-controller Normal ScalingReplicaSet Scaled up replica set llmisvcb11a9c9832b99b016bc8f8e0ea095712-kserve-router-scheduler-6d86bd4d9d from 0 to 1 [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 None kserve-ci-e2e-test Normal Scheduled Successfully assigned kserve-ci-e2e-test/llmisvcb11a9c9832b99b016bc8f8e0ea095712-kserve-router-scheknzg7 to ip-10-0-133-25.ec2.internal [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:24 kserve-ci-e2e-test multus Normal AddedInterface Add eth0 [10.132.0.48/23] from ovn-kubernetes [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:24 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Pulled Container image "ghcr.io/llm-d/llm-d-router-endpoint-picker:v0.9.0-rc.2" already present on machine [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:24 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Created Created container: main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:24 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Started Started container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:26:13 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Killing Stopping container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:26:13 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Warning FailedPreStopHook PreStopHook failed [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 None kserve-ci-e2e-test Normal Scheduled Successfully assigned kserve-ci-e2e-test/llmisvcca2d2d7d499abb359505529ebe02c136-kserve-router-scheb2glb to ip-10-0-139-47.ec2.internal [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:18 kserve-ci-e2e-test multus Normal AddedInterface Add eth0 [10.134.0.28/23] from ovn-kubernetes [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:18 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Pulled Container image "ghcr.io/llm-d/llm-d-router-endpoint-picker:v0.9.0-rc.2" already present on machine [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:18 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Created Created container: main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:18 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Started Started container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:28:58 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Killing Stopping container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:28:58 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Warning FailedPreStopHook PreStopHook failed [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:17 kserve-ci-e2e-test replicaset-controller Normal SuccessfulCreate Created pod: llmisvcca2d2d7d499abb359505529ebe02c136-kserve-router-scheb2glb [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:17 kserve-ci-e2e-test deployment-controller Normal ScalingReplicaSet Scaled up replica set llmisvcca2d2d7d499abb359505529ebe02c136-kserve-router-scheduler-68cc9685d6 from 0 to 1 [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:24:43 kserve-ci-e2e-test replicaset-controller Normal SuccessfulCreate Created pod: llmisvce55ae740357a3a31a27cdb8b66ffe20f-kserve-router-schex4lqs [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:24:43 kserve-ci-e2e-test deployment-controller Normal ScalingReplicaSet Scaled up replica set llmisvce55ae740357a3a31a27cdb8b66ffe20f-kserve-router-scheduler-749449dbc8 from 0 to 1 [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 None kserve-ci-e2e-test Normal Scheduled Successfully assigned kserve-ci-e2e-test/llmisvce55ae740357a3a31a27cdb8b66ffe20f-kserve-router-schex4lqs to ip-10-0-139-47.ec2.internal [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:24:43 kserve-ci-e2e-test multus Normal AddedInterface Add eth0 [10.134.0.26/23] from ovn-kubernetes [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:24:43 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Pulled Container image "ghcr.io/llm-d/llm-d-router-endpoint-picker:v0.9.0-rc.2" already present on machine [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:24:43 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Created Created container: main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:24:43 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Started Started container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:26:33 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Killing Stopping container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:26:33 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Warning FailedPreStopHook PreStopHook failed [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 None kserve-ci-e2e-test Normal Scheduled Successfully assigned kserve-ci-e2e-test/precise-prefix-cache-test-kserve-5755bc98d9-f88w6 to ip-10-0-139-47.ec2.internal [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:55 kserve-ci-e2e-test multus Normal AddedInterface Add eth0 [10.134.0.22/23] from ovn-kubernetes [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:55 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Pulled Container image "quay.io/opendatahub/kserve-storage-initializer@sha256:13ad765c0f323200e1db7b45c99661287e65c5158d6b54b16139829636d20be4" already present on machine [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:55 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Created Created container: storage-initializer [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:55 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Started Started container storage-initializer [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:00 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Pulled Container image "ghcr.io/llm-d/llm-d-inference-sim:v0.8.2" already present on machine [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:01 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Created Created container: main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:01 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Started Started container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:23:10 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Killing Stopping container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 None kserve-ci-e2e-test Normal Scheduled Successfully assigned kserve-ci-e2e-test/precise-prefix-cache-test-kserve-5755bc98d9-hdpjr to ip-10-0-133-25.ec2.internal [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:55 kserve-ci-e2e-test multus Normal AddedInterface Add eth0 [10.132.0.45/23] from ovn-kubernetes [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:55 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Pulling Pulling image "quay.io/opendatahub/kserve-storage-initializer@sha256:13ad765c0f323200e1db7b45c99661287e65c5158d6b54b16139829636d20be4" [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:00 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Pulled Successfully pulled image "quay.io/opendatahub/kserve-storage-initializer@sha256:13ad765c0f323200e1db7b45c99661287e65c5158d6b54b16139829636d20be4" in 5.546s (5.546s including waiting). Image size: 301879255 bytes. [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:00 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Created Created container: storage-initializer [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:00 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Started Started container storage-initializer [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:07 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Pulling Pulling image "ghcr.io/llm-d/llm-d-inference-sim:v0.8.2" [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:08 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Pulled Successfully pulled image "ghcr.io/llm-d/llm-d-inference-sim:v0.8.2" in 1.434s (1.434s including waiting). Image size: 98346788 bytes. [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:09 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Created Created container: main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:09 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Started Started container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:23:10 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Killing Stopping container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:54 kserve-ci-e2e-test replicaset-controller Normal SuccessfulCreate Created pod: precise-prefix-cache-test-kserve-5755bc98d9-f88w6 [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:54 kserve-ci-e2e-test replicaset-controller Normal SuccessfulCreate Created pod: precise-prefix-cache-test-kserve-5755bc98d9-hdpjr [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:55 kserve-ci-e2e-test replicaset-controller Normal SuccessfulCreate Created pod: precise-prefix-cache-test-kserve-router-scheduler-647bfdf7jdncq [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 None kserve-ci-e2e-test Normal Scheduled Successfully assigned kserve-ci-e2e-test/precise-prefix-cache-test-kserve-router-scheduler-647bfdf7jdncq to ip-10-0-139-47.ec2.internal [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:55 kserve-ci-e2e-test multus Normal AddedInterface Add eth0 [10.134.0.23/23] from ovn-kubernetes [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:55 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Pulled Container image "quay.io/opendatahub/kserve-storage-initializer@sha256:13ad765c0f323200e1db7b45c99661287e65c5158d6b54b16139829636d20be4" already present on machine [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:55 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Created Created container: storage-initializer [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:55 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Started Started container storage-initializer [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:57 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Pulling Pulling image "ghcr.io/llm-d/llm-d-router-endpoint-picker:v0.9.0-rc.2" [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:58 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Pulled Successfully pulled image "ghcr.io/llm-d/llm-d-router-endpoint-picker:v0.9.0-rc.2" in 1.604s (1.604s including waiting). Image size: 75073927 bytes. [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:30 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Created Created container: main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:30 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Started Started container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:58 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Pulling Pulling image "ghcr.io/llm-d/llm-d-uds-tokenizer:vllm-v0.19.1" [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:29 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Pulled Successfully pulled image "ghcr.io/llm-d/llm-d-uds-tokenizer:vllm-v0.19.1" in 30.722s (30.722s including waiting). Image size: 2989890188 bytes. [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:29 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Created Created container: tokenizer [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:29 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Started Started container tokenizer [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:30 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Pulled Container image "ghcr.io/llm-d/llm-d-router-endpoint-picker:v0.9.0-rc.2" already present on machine [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:35 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Warning Unhealthy Startup probe failed: Get "http://10.134.0.23:8082/healthz": dial tcp 10.134.0.23:8082: connect: connection refused [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:46 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Warning Unhealthy Liveness probe failed: timeout: failed to connect service "10.134.0.23:9003" within 1s: context deadline exceeded [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:23:10 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Killing Stopping container tokenizer [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:23:10 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Killing Stopping container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:23:10 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Warning FailedPreStopHook PreStopHook failed [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:55 kserve-ci-e2e-test deployment-controller Normal ScalingReplicaSet Scaled up replica set precise-prefix-cache-test-kserve-router-scheduler-647bfdf79c from 0 to 1 [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:54 kserve-ci-e2e-test deployment-controller Normal ScalingReplicaSet Scaled up replica set precise-prefix-cache-test-kserve-5755bc98d9 from 0 to 2 [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:47 kserve-ci-e2e-test OpenDataHubModelController Warning ReconcileError Failed to reconcile LLMInferenceService: 1 error occurred: * failed to get HTTPRoute for AuthPolicy precise-prefix-cache-test-kserve-route-authn: failed to get HTTPRoute kserve-ci-e2e-test/precise-prefix-cache-test-kserve-route: HTTPRoute.gateway.networking.k8s.io "precise-prefix-cache-test-kserve-route" not found [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:54 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Secret kserve-ci-e2e-test/precise-prefix-cache-test-kserve-self-signed-certs [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:54 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Deployment kserve-ci-e2e-test/precise-prefix-cache-test-kserve [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:54 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Service kserve-ci-e2e-test/precise-prefix-cache-test-kserve-workload-svc [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:54 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.ServiceAccount kserve-ci-e2e-test/precise-prefix-cache-test-epp-sa [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:54 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.ClusterRoleBinding /kserve-ci-e2e-test-precise-prefix-cache-test-epp-auth-rb [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:54 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Role kserve-ci-e2e-test/precise-prefix-cache-test-epp-role [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:55 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.RoleBinding kserve-ci-e2e-test/precise-prefix-cache-test-epp-rb [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:55 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Deployment kserve-ci-e2e-test/precise-prefix-cache-test-kserve-router-scheduler [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:55 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Service kserve-ci-e2e-test/precise-prefix-cache-test-epp-service [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:26 kserve-ci-e2e-test LLMInferenceServiceController Normal Created (combined from similar events): Created v1.DestinationRule kserve-ci-e2e-test/precise-prefix-cache-test-kserve-shadow-svc [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:25 kserve-ci-e2e-test LLMInferenceServiceController Normal Updated Updated v1.Secret kserve-ci-e2e-test/precise-prefix-cache-test-kserve-self-signed-certs [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:26 kserve-ci-e2e-test LLMInferenceServiceController Normal Updated Updated v1.HTTPRoute kserve-ci-e2e-test/precise-prefix-cache-test-kserve-route [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:23:09 kserve-ci-e2e-test LLMInferenceServiceController Normal LLMInferenceServiceReady LLMInferenceService [precise-prefix-cache-test] is Ready [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:23:10 kserve-ci-e2e-test LLMInferenceServiceController Normal Deleted Deleted v1.ClusterRoleBinding /kserve-ci-e2e-test-precise-prefix-cache-test-epp-auth-rb [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 None kserve-ci-e2e-test Normal Scheduled Successfully assigned kserve-ci-e2e-test/router-gateway-1-openshift-default-799f46c59b-vt5f9 to ip-10-0-133-25.ec2.internal [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:36 kserve-ci-e2e-test multus Normal AddedInterface Add eth0 [10.132.0.44/23] from ovn-kubernetes [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:56 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Pulled Container image "registry.redhat.io/openshift-service-mesh/istio-proxyv2-rhel9@sha256:7d15cebf9b62f3f235c0eab5158ac8ff2fda86a1d193490dc94c301402c99da8" already present on machine [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:56 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Created Created container: istio-proxy [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:56 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Started Started container istio-proxy [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:54 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Warning Unhealthy Readiness probe failed: Get "http://10.132.0.44:15021/healthz/ready": context deadline exceeded (Client.Timeout exceeded while awaiting headers) [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:02 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Warning Unhealthy Startup probe failed: Get "http://10.132.0.44:15021/healthz/ready": dial tcp 10.132.0.44:15021: connect: connection refused [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:35 kserve-ci-e2e-test replicaset-controller Normal SuccessfulCreate Created pod: router-gateway-1-openshift-default-799f46c59b-vt5f9 [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:35 kserve-ci-e2e-test deployment-controller Normal ScalingReplicaSet Scaled up replica set router-gateway-1-openshift-default-799f46c59b from 0 to 1 [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:35 kserve-ci-e2e-test service-controller Normal EnsuringLoadBalancer Ensuring load balancer [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:40 kserve-ci-e2e-test service-controller Normal EnsuredLoadBalancer Ensured load balancer [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:30:52 kserve-ci-e2e-test gateway_labeler_controller Normal AddedLabel Added label istio.io/rev=openshift-gateway to gateway router-gateway-1 [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:148 ### Pod llmisvc-model-fb-opt-125m-with-7ca60146-kserve-7cfdf49d74-pd9px (phase=Running) [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:188 #### init-container 'storage-initializer' (restarts=0) [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:201 # -- logs (current) -- [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:202 2026-07-07 14:29:25.382 1 storage.initializer INFO [initializer-entrypoint:():17] Initializing, args: (src_uri, dest_path): [('hf://facebook/opt-125m', '/mnt/models'), ('hf://edbeeching/opt-125m-lora', '/mnt/lora/lora-adapter-1')] [e2e-llm-inference-service] 2026-07-07 14:29:25.382 1 storage.initializer INFO [kserve_storage.py:download():166] Copying contents of hf://facebook/opt-125m to local [e2e-llm-inference-service] 2026-07-07 14:29:29.700 1 storage.initializer INFO [kserve_storage.py:download():234] Successfully copied hf://facebook/opt-125m to /mnt/models [e2e-llm-inference-service] 2026-07-07 14:29:29.701 1 storage.initializer INFO [kserve_storage.py:download():235] Model downloaded in 4.318500545999996 seconds. [e2e-llm-inference-service] 2026-07-07 14:29:29.701 1 storage.initializer INFO [kserve_storage.py:download():166] Copying contents of hf://edbeeching/opt-125m-lora to local [e2e-llm-inference-service] 2026-07-07 14:29:30.025 1 storage.initializer INFO [kserve_storage.py:download():234] Successfully copied hf://edbeeching/opt-125m-lora to /mnt/lora/lora-adapter-1 [e2e-llm-inference-service] 2026-07-07 14:29:30.025 1 storage.initializer INFO [kserve_storage.py:download():235] Model downloaded in 0.3247854470000675 seconds. [e2e-llm-inference-service] [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:188 #### container 'main' (restarts=0) [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:201 # -- logs (current) -- [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:202 (APIServer pid=1) DEBUG 07-07 14:43:04 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:43:04 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:43:05 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:43:06 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:43:07 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:43:08 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:43:09 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:43:10 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:43:11 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:43:12 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:43:13 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:43:14 [v1/metrics/loggers.py:259] Engine 000: Avg prompt throughput: 0.0 tokens/s, Avg generation throughput: 0.0 tokens/s, Running: 0 reqs, Waiting: 0 reqs, GPU KV cache usage: 0.0%, Prefix cache hit rate: 0.0% [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:43:14 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:43:14 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:43:15 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:43:16 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:43:17 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:43:18 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:43:19 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:43:20 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:43:21 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:43:22 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:43:23 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:43:24 [v1/metrics/loggers.py:259] Engine 000: Avg prompt throughput: 0.0 tokens/s, Avg generation throughput: 0.0 tokens/s, Running: 0 reqs, Waiting: 0 reqs, GPU KV cache usage: 0.0%, Prefix cache hit rate: 0.0% [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:43:24 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:43:24 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:43:25 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:43:26 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:43:27 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:43:28 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:43:29 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:43:30 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:43:31 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:43:32 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:43:33 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:43:34 [v1/metrics/loggers.py:259] Engine 000: Avg prompt throughput: 0.0 tokens/s, Avg generation throughput: 0.0 tokens/s, Running: 0 reqs, Waiting: 0 reqs, GPU KV cache usage: 0.0%, Prefix cache hit rate: 0.0% [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:43:34 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:43:34 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:43:35 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:43:36 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:43:37 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:43:38 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:43:39 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:43:40 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:43:41 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:43:42 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:43:43 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:43:44 [v1/metrics/loggers.py:259] Engine 000: Avg prompt throughput: 0.0 tokens/s, Avg generation throughput: 0.0 tokens/s, Running: 0 reqs, Waiting: 0 reqs, GPU KV cache usage: 0.0%, Prefix cache hit rate: 0.0% [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:43:44 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:43:44 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:43:45 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:43:46 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:43:47 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:43:48 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:43:49 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:43:50 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:43:51 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:43:52 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:43:53 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:43:54 [v1/metrics/loggers.py:259] Engine 000: Avg prompt throughput: 0.0 tokens/s, Avg generation throughput: 0.0 tokens/s, Running: 0 reqs, Waiting: 0 reqs, GPU KV cache usage: 0.0%, Prefix cache hit rate: 0.0% [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:43:54 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:43:54 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:43:55 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:43:56 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:43:57 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:43:58 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:43:59 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:44:00 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:44:01 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:44:02 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:44:03 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:44:04 [v1/metrics/loggers.py:259] Engine 000: Avg prompt throughput: 0.0 tokens/s, Avg generation throughput: 0.0 tokens/s, Running: 0 reqs, Waiting: 0 reqs, GPU KV cache usage: 0.0%, Prefix cache hit rate: 0.0% [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:44:04 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:44:04 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:44:05 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:44:06 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:44:07 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:44:08 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:44:09 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:44:10 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:44:11 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:44:12 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:44:13 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:44:14 [v1/metrics/loggers.py:259] Engine 000: Avg prompt throughput: 0.0 tokens/s, Avg generation throughput: 0.0 tokens/s, Running: 0 reqs, Waiting: 0 reqs, GPU KV cache usage: 0.0%, Prefix cache hit rate: 0.0% [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:44:14 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:44:14 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:44:15 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:44:16 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:44:17 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:44:18 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:44:19 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:44:20 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:44:21 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:44:22 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:44:23 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:44:24 [v1/metrics/loggers.py:259] Engine 000: Avg prompt throughput: 0.0 tokens/s, Avg generation throughput: 0.0 tokens/s, Running: 0 reqs, Waiting: 0 reqs, GPU KV cache usage: 0.0%, Prefix cache hit rate: 0.0% [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:44:24 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:44:24 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:44:25 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:44:26 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:44:27 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:44:28 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:44:29 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:44:30 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:44:31 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:44:32 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:44:33 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:44:34 [v1/metrics/loggers.py:259] Engine 000: Avg prompt throughput: 0.0 tokens/s, Avg generation throughput: 0.0 tokens/s, Running: 0 reqs, Waiting: 0 reqs, GPU KV cache usage: 0.0%, Prefix cache hit rate: 0.0% [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:44:34 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:44:34 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:44:35 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:44:36 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:44:37 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:44:38 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:44:39 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:44:40 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:44:41 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:44:42 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:44:43 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:44:44 [v1/metrics/loggers.py:259] Engine 000: Avg prompt throughput: 0.0 tokens/s, Avg generation throughput: 0.0 tokens/s, Running: 0 reqs, Waiting: 0 reqs, GPU KV cache usage: 0.0%, Prefix cache hit rate: 0.0% [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:44:44 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:44:44 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:44:45 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:44:46 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:44:47 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:44:48 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:44:49 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:44:50 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:44:51 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:44:52 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:44:53 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:44:54 [v1/metrics/loggers.py:259] Engine 000: Avg prompt throughput: 0.0 tokens/s, Avg generation throughput: 0.0 tokens/s, Running: 0 reqs, Waiting: 0 reqs, GPU KV cache usage: 0.0%, Prefix cache hit rate: 0.0% [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:44:54 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:44:54 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:44:55 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:44:56 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:44:57 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:44:58 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:44:59 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:45:00 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:45:01 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:45:02 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:45:03 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:45:04 [v1/metrics/loggers.py:259] Engine 000: Avg prompt throughput: 0.0 tokens/s, Avg generation throughput: 0.0 tokens/s, Running: 0 reqs, Waiting: 0 reqs, GPU KV cache usage: 0.0%, Prefix cache hit rate: 0.0% [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:45:04 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:45:04 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:45:05 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:45:06 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:45:07 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:45:08 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:45:09 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:45:10 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:45:11 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:45:12 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:45:13 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:45:14 [v1/metrics/loggers.py:259] Engine 000: Avg prompt throughput: 0.0 tokens/s, Avg generation throughput: 0.0 tokens/s, Running: 0 reqs, Waiting: 0 reqs, GPU KV cache usage: 0.0%, Prefix cache hit rate: 0.0% [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:45:14 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:45:14 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:45:15 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:45:16 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:45:17 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:45:18 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:45:19 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:45:20 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:45:21 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:45:22 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:45:23 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:45:24 [v1/metrics/loggers.py:259] Engine 000: Avg prompt throughput: 0.0 tokens/s, Avg generation throughput: 0.0 tokens/s, Running: 0 reqs, Waiting: 0 reqs, GPU KV cache usage: 0.0%, Prefix cache hit rate: 0.0% [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:45:24 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:45:24 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:45:25 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:45:26 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:45:27 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:45:28 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:45:29 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:45:30 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:45:31 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:45:32 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:45:33 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:45:34 [v1/metrics/loggers.py:259] Engine 000: Avg prompt throughput: 0.0 tokens/s, Avg generation throughput: 0.0 tokens/s, Running: 0 reqs, Waiting: 0 reqs, GPU KV cache usage: 0.0%, Prefix cache hit rate: 0.0% [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:45:34 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:45:34 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:45:35 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:45:36 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:45:37 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:45:38 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:45:39 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:45:40 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:45:41 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:45:42 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:45:43 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:45:44 [v1/metrics/loggers.py:259] Engine 000: Avg prompt throughput: 0.0 tokens/s, Avg generation throughput: 0.0 tokens/s, Running: 0 reqs, Waiting: 0 reqs, GPU KV cache usage: 0.0%, Prefix cache hit rate: 0.0% [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:45:44 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:45:44 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:45:45 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:45:46 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:45:47 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:45:48 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:45:49 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:45:50 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:148 ### Pod llmisvc-model-fb-opt-125m-with-7ca60146-kserve-router-schenp5mk (phase=Running) [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:188 #### init-container 'storage-initializer' (restarts=0) [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:201 # -- logs (current) -- [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:202 2026-07-07 14:29:25.707 1 storage.initializer INFO [initializer-entrypoint:():17] Initializing, args: (src_uri, dest_path): [('hf://facebook/opt-125m', '/mnt/models')] [e2e-llm-inference-service] 2026-07-07 14:29:25.708 1 storage.initializer INFO [kserve_storage.py:download():166] Copying contents of hf://facebook/opt-125m to local [e2e-llm-inference-service] 2026-07-07 14:29:25.708 1 storage.initializer INFO [kserve_storage.py:download():169] Allow patterns: ['tokenizer.json', 'tokenizer_config.json', 'special_tokens_map.json', 'vocab.json', 'merges.txt', 'config.json', 'generation_config.json'] [e2e-llm-inference-service] Could not set the permissions on the file '/mnt/models/.cache/huggingface/download/8_PA_wEVGiVa2goH2H4KQOQpvVY=.b3fb716a3024261980becb2382e31a3780985130.incomplete'. Error: [Errno 13] Permission denied: '/mnt/tmp_8795c940-861d-468b-a470-fb64ae43b818'. [e2e-llm-inference-service] Continuing without setting permissions. [e2e-llm-inference-service] Could not set the permissions on the file '/mnt/models/.cache/huggingface/download/3EVKVggOldJcKSsGjSdoUCN1AyQ=.cf739e3ba86db7791ebab2828cc34b8a5acd3a86.incomplete'. Error: [Errno 13] Permission denied: '/mnt/tmp_5155687b-1dda-4fca-9542-81310489d61b'. [e2e-llm-inference-service] Continuing without setting permissions. [e2e-llm-inference-service] Could not set the permissions on the file '/mnt/models/.cache/huggingface/download/PtHk0z_I45atnj23IIRhTExwT3w=.226b0752cac7789c48f0cb3ec53eda48b7be36cc.incomplete'. Error: [Errno 13] Permission denied: '/mnt/tmp_77e66729-2c36-471e-8a40-e2edec7aea7b'. [e2e-llm-inference-service] Continuing without setting permissions. [e2e-llm-inference-service] Could not set the permissions on the file '/mnt/models/.cache/huggingface/download/ahkChHUJFxEmOdq5GDFEmerRzCY=.5dfa36546b8eddce0e04df3133c30df43fcc3828.incomplete'. Error: [Errno 13] Permission denied: '/mnt/tmp_725c5297-98da-4454-a2f5-a47bfc02380c'. [e2e-llm-inference-service] Continuing without setting permissions. [e2e-llm-inference-service] Could not set the permissions on the file '/mnt/models/.cache/huggingface/download/vzaExXFZNBay89bvlQv-ZcI6BTg=.27c24ca9d908d0b678b20c698aeb9e950c44d865.incomplete'. Error: [Errno 13] Permission denied: '/mnt/tmp_8450c906-d792-4f03-b32d-6ba64f4e1357'. [e2e-llm-inference-service] Continuing without setting permissions. [e2e-llm-inference-service] Could not set the permissions on the file '/mnt/models/.cache/huggingface/download/j3m-Hy6QvBddw8RXA1uSWl1AJ0c=.0a39732b2d8be8e493cab3da68b68cc3e28221de.incomplete'. Error: [Errno 13] Permission denied: '/mnt/tmp_f101ccff-0236-4edc-a0a9-e5c48187f2af'. [e2e-llm-inference-service] Continuing without setting permissions. [e2e-llm-inference-service] 2026-07-07 14:29:26.119 1 storage.initializer INFO [kserve_storage.py:download():234] Successfully copied hf://facebook/opt-125m to /mnt/models [e2e-llm-inference-service] 2026-07-07 14:29:26.119 1 storage.initializer INFO [kserve_storage.py:download():235] Model downloaded in 0.4119187999999667 seconds. [e2e-llm-inference-service] [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:188 #### container 'main' (restarts=0) [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:201 # -- logs (current) -- [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:202 {"level":"info","ts":1783434567.0940168,"logger":"setup","caller":"runner/runner.go:196","msg":"GIE build","commit-sha":"181aa8358916e19b8844ccc752b2d6153d4b2ad6","build-ref":"v0.9.0-rc.2"} [e2e-llm-inference-service] Flag --model-server-metrics-scheme has been deprecated, This flag is deprecated. Configure via EndpointPickerConfig data layer plugin parameters instead. [e2e-llm-inference-service] {"level":"info","ts":1783434567.094195,"logger":"setup","caller":"runner/runner.go:217","msg":"Flags processed","flags":{"cert-path":"/var/run/kserve/tls","config-file":"","config-text":"apiVersion: inference.networking.x-k8s.io/v1alpha1\nkind: EndpointPickerConfig\nplugins:\n- type: single-profile-handler\n- type: queue-scorer\n- type: prefix-cache-scorer\n- type: max-score-picker\n- type: lora-affinity-scorer\nschedulingProfiles:\n- name: default\n plugins:\n - pluginRef: lora-affinity-scorer\n weight: 4\n - pluginRef: queue-scorer\n weight: 2\n - pluginRef: prefix-cache-scorer\n weight: 3\n - pluginRef: max-score-picker\n","disable-endpoint-subset-filter":false,"enable-cert-reload":true,"enable-grpc-stream-metrics":false,"enable-pprof":true,"endpoint-selector":"","endpoint-target-ports":{},"grpc-health-port":9003,"grpc-max-recv-msg-size":"","grpc-max-send-msg-size":"","grpc-port":9002,"ha-enable-leader-election":false,"health-checking":false,"metrics-endpoint-auth":true,"metrics-port":9090,"metrics-staleness-threshold":2000000000,"model-server-metrics-https-insecure-skip-verify":true,"model-server-metrics-path":"/metrics","model-server-metrics-port":0,"model-server-metrics-scheme":"https","pool-group":"inference.networking.k8s.io","pool-name":"llmisvc-model-fb-opt-125m-with-7ca60146-inference-pool","pool-namespace":"kserve-ci-e2e-test","refresh-metrics-interval":50000000,"refresh-prometheus-metrics-interval":5000000000,"secure-serving":true,"tracing":true,"v":2,"zap-devel":{},"zap-encoder":{},"zap-log-level":{},"zap-stacktrace-level":{},"zap-time-encoding":{}}} [e2e-llm-inference-service] {"level":"info","ts":1783434567.0942664,"logger":"setup.trace","caller":"tracing/telemetry.go:123","msg":"init OTel trace exporter","type":"console"} [e2e-llm-inference-service] {"level":"info","ts":1783434567.095781,"caller":"loader/configloader.go:89","msg":"DEPRECATION: apiVersion inference.networking.x-k8s.io/v1alpha1/EndpointPickerConfig is deprecated","replacement":"llm-d.ai/v1alpha1/EndpointPickerConfig"} [e2e-llm-inference-service] {"level":"info","ts":1783434567.0958574,"caller":"loader/configloader.go:121","msg":"Loaded raw configuration","config":"{Plugins: [{Type: single-profile-handler} {Type: queue-scorer} {Type: prefix-cache-scorer} {Type: max-score-picker} {Type: lora-affinity-scorer}], SchedulingProfiles: [{Name: default, Plugins: [{PluginRef: lora-affinity-scorer, Weight: 4.00} {PluginRef: queue-scorer, Weight: 2.00} {PluginRef: prefix-cache-scorer, Weight: 3.00} {PluginRef: max-score-picker}]}]}"} [e2e-llm-inference-service] {"level":"info","ts":1783434567.0958686,"logger":"setup","caller":"runner/runner.go:622","msg":"Data layer: ENABLED"} [e2e-llm-inference-service] {"level":"info","ts":1783434567.096124,"logger":"setup","caller":"runner/runner.go:281","msg":"Raw config after phase one","config":{"apiVersion":"inference.networking.x-k8s.io/v1alpha1","dataLayer":null,"kind":"EndpointPickerConfig","plugins":[{"name":"single-profile-handler","parameters":null,"type":"single-profile-handler"},{"name":"queue-scorer","parameters":null,"type":"queue-scorer"},{"name":"prefix-cache-scorer","parameters":null,"type":"prefix-cache-scorer"},{"name":"max-score-picker","parameters":null,"type":"max-score-picker"},{"name":"lora-affinity-scorer","parameters":null,"type":"lora-affinity-scorer"}],"schedulingProfiles":[{"name":"default","plugins":[{"pluginRef":"lora-affinity-scorer","weight":4},{"pluginRef":"queue-scorer","weight":2},{"pluginRef":"prefix-cache-scorer","weight":3},{"pluginRef":"max-score-picker","weight":null}]}]}} [e2e-llm-inference-service] {"level":"info","ts":1783434567.1143694,"logger":"utilization-detector/utilization-detector","caller":"utilization/detector.go:83","msg":"Creating new UtilizationDetector","queueDepthThreshold":5,"kvCacheUtilThreshold":0.8,"metricsStalenessThreshold":"200ms","headroom":0} [e2e-llm-inference-service] {"level":"info","ts":1783434567.1144679,"caller":"metrics/factories.go:230","msg":"Registered engine mapping","engine":"vllm","mapping":"Mapping{all specs enabled}"} [e2e-llm-inference-service] {"level":"info","ts":1783434567.1145024,"caller":"metrics/factories.go:230","msg":"Registered engine mapping","engine":"sglang","mapping":"Mapping{disabled: [lora]}"} [e2e-llm-inference-service] {"level":"info","ts":1783434567.114539,"caller":"metrics/factories.go:230","msg":"Registered engine mapping","engine":"trtllm-serve","mapping":"Mapping{disabled: [lora, cacheInfo]}"} [e2e-llm-inference-service] {"level":"info","ts":1783434567.1145985,"caller":"metrics/factories.go:230","msg":"Registered engine mapping","engine":"triton-tensorrt-llm","mapping":"Mapping{disabled: [lora, cacheInfo]}"} [e2e-llm-inference-service] {"level":"info","ts":1783434567.1146183,"caller":"metrics/factories.go:230","msg":"Registered engine mapping","engine":"triton","mapping":"Mapping{disabled: [kv, lora, cacheInfo]}"} [e2e-llm-inference-service] {"level":"info","ts":1783434567.114674,"caller":"loader/configloader.go:154","msg":"Instantiated all plugins and applied system defaults. Effective raw configuration","config":"{Plugins: [{Name: single-profile-handler, Type: single-profile-handler} {Name: queue-scorer, Type: queue-scorer} {Name: prefix-cache-scorer, Type: prefix-cache-scorer} {Name: max-score-picker, Type: max-score-picker} {Name: lora-affinity-scorer, Type: lora-affinity-scorer} {Name: fcfs-ordering-policy, Type: fcfs-ordering-policy} {Name: global-strict-fairness-policy, Type: global-strict-fairness-policy} {Name: static-usage-limit-policy, Type: static-usage-limit-policy} {Name: openai-parser, Type: openai-parser} {Name: anthropic-parser, Type: anthropic-parser} {Name: vllmhttp-parser, Type: vllmhttp-parser} {Name: utilization-detector, Type: utilization-detector} {Name: metrics-data-source, Type: metrics-data-source} {Name: core-metrics-extractor, Type: core-metrics-extractor}], SchedulingProfiles: [{Name: default, Plugins: [{PluginRef: lora-affinity-scorer, Weight: 4.00} {PluginRef: queue-scorer, Weight: 2.00} {PluginRef: prefix-cache-scorer, Weight: 3.00} {PluginRef: max-score-picker}]}], DataLayer: {Sources: [{PluginRef: metrics-data-source, Extractors: [{PluginRef: core-metrics-extractor}]}], Discovery: }, FlowControl: {MaxBytes: unlimited, MaxRequests: unlimited, SaturationDetector: {PluginRef: utilization-detector}}, RequestHandler: {Parsers: [{PluginRef: openai-parser}, {PluginRef: anthropic-parser}, {PluginRef: vllmhttp-parser}]}}"} [e2e-llm-inference-service] {"level":"info","ts":1783434567.1147242,"caller":"approximateprefix/plugin.go:88","msg":"Prefix DataProducer initialized","config":{"autoTune":true,"blockSizeTokens":16,"blockSize":0,"maxPrefixBlocksToMatch":2048,"maxPrefixTokensToMatch":131072,"lruCapacityPerServer":31250}} [e2e-llm-inference-service] {"level":"info","ts":1783434567.1147954,"caller":"approximateprefix/plugin.go:111","msg":"WARNING: configured blockSizeTokens is below the recommended minimum, overriding it.","blockSizeTokens":16,"minimum":64,"issue":"https://github.com/llm-d/llm-d-router/issues/1158"} [e2e-llm-inference-service] {"level":"info","ts":1783434567.1148524,"caller":"datalayer/data_graph.go:116","msg":"auto-created default producer","producer":"approx-prefix-cache-producer/approx-prefix-cache-producer","dataKey":"PrefixCacheMatchInfoDataKey/approx-prefix-cache-producer","consumer":"prefix-cache-scorer"} [e2e-llm-inference-service] {"level":"info","ts":1783434567.114884,"caller":"datalayer/data_graph.go:116","msg":"auto-created default producer","producer":"token-producer/token-producer","dataKey":"TokenizedPrompt/token-producer","consumer":"approx-prefix-cache-producer"} [e2e-llm-inference-service] {"level":"info","ts":1783434567.1149914,"caller":"runner/runner.go:685","msg":"loaded configuration from file/text successfully"} [e2e-llm-inference-service] {"level":"info","ts":1783434567.1150022,"logger":"setup","caller":"runner/runner.go:308","msg":"EPP config after phase two","config":"{SchedulerConfig:{ProfileHandler: single-profile-handler/single-profile-handler, Profiles: map[default:{Filters: [], Scorers: [lora-affinity-scorer/lora-affinity-scorer: 4.000000, queue-scorer/queue-scorer: 2.000000, prefix-cache-scorer/prefix-cache-scorer: 3.000000], Picker: max-score-picker/max-score-picker}]} SaturationDetector:0xc000a82580 DataConfig:{Sources:[{Plugin:0xc000a94510 Extractors:[0xc000a82780]}]} FlowControlConfig: ParserRegistry:0xc000a82c40}"} [e2e-llm-inference-service] {"level":"info","ts":1783434567.1351664,"logger":"setup","caller":"runner/runner.go:352","msg":"Setting pprof handlers"} [e2e-llm-inference-service] {"level":"info","ts":1783434567.1351974,"caller":"manager/internal.go:201","msg":"Registering metrics http server extra handler","path":"/debug/pprof/trace"} [e2e-llm-inference-service] {"level":"info","ts":1783434567.1352146,"caller":"manager/internal.go:201","msg":"Registering metrics http server extra handler","path":"/debug/pprof/heap"} [e2e-llm-inference-service] {"level":"info","ts":1783434567.1352217,"caller":"manager/internal.go:201","msg":"Registering metrics http server extra handler","path":"/debug/pprof/goroutine"} [e2e-llm-inference-service] {"level":"info","ts":1783434567.1352277,"caller":"manager/internal.go:201","msg":"Registering metrics http server extra handler","path":"/debug/pprof/allocs"} [e2e-llm-inference-service] {"level":"info","ts":1783434567.1352339,"caller":"manager/internal.go:201","msg":"Registering metrics http server extra handler","path":"/debug/pprof/block"} [e2e-llm-inference-service] {"level":"info","ts":1783434567.13524,"caller":"manager/internal.go:201","msg":"Registering metrics http server extra handler","path":"/debug/pprof/"} [e2e-llm-inference-service] {"level":"info","ts":1783434567.1352475,"caller":"manager/internal.go:201","msg":"Registering metrics http server extra handler","path":"/debug/pprof/symbol"} [e2e-llm-inference-service] {"level":"info","ts":1783434567.1352556,"caller":"manager/internal.go:201","msg":"Registering metrics http server extra handler","path":"/debug/pprof/threadcreate"} [e2e-llm-inference-service] {"level":"info","ts":1783434567.1352654,"caller":"manager/internal.go:201","msg":"Registering metrics http server extra handler","path":"/debug/pprof/mutex"} [e2e-llm-inference-service] {"level":"info","ts":1783434567.1352735,"caller":"manager/internal.go:201","msg":"Registering metrics http server extra handler","path":"/debug/pprof/cmdline"} [e2e-llm-inference-service] {"level":"info","ts":1783434567.1352823,"caller":"manager/internal.go:201","msg":"Registering metrics http server extra handler","path":"/debug/pprof/profile"} [e2e-llm-inference-service] {"level":"info","ts":1783434567.1352952,"caller":"manager/internal.go:201","msg":"Registering metrics http server extra handler","path":"/debug/plugins/state"} [e2e-llm-inference-service] {"level":"info","ts":1783434567.1353037,"logger":"setup","caller":"runner/runner.go:373","msg":"parsed config","scheduler-config":"{ProfileHandler: single-profile-handler/single-profile-handler, Profiles: map[default:{Filters: [], Scorers: [lora-affinity-scorer/lora-affinity-scorer: 4.000000, queue-scorer/queue-scorer: 2.000000, prefix-cache-scorer/prefix-cache-scorer: 3.000000], Picker: max-score-picker/max-score-picker}]}"} [e2e-llm-inference-service] {"level":"info","ts":1783434567.1353376,"logger":"setup","caller":"datalayer/runtime.go:99","msg":"Configuring datalayer runtime","numSources":1} [e2e-llm-inference-service] {"level":"info","ts":1783434567.135349,"logger":"setup","caller":"datalayer/runtime.go:118","msg":"Processing source","source":"metrics-data-source","numExtractors":1} [e2e-llm-inference-service] {"level":"info","ts":1783434567.1353695,"logger":"setup","caller":"datalayer/runtime.go:147","msg":"Source configured","source":"metrics-data-source","extractors":["core-metrics-extractor/core-metrics-extractor"]} [e2e-llm-inference-service] {"level":"info","ts":1783434567.1353898,"logger":"setup","caller":"datalayer/runtime.go:206","msg":"Datalayer runtime configured","pollers":1,"notifiers":0,"endpointSources":0} [e2e-llm-inference-service] {"level":"info","ts":1783434567.135405,"logger":"setup","caller":"runner/runner.go:833","msg":"Experimental Flow Control layer is disabled, using legacy admission control"} [e2e-llm-inference-service] {"level":"info","ts":1783434567.135503,"logger":"setup","caller":"runner/runner.go:721","msg":"ExtProc server runner added to manager."} [e2e-llm-inference-service] {"level":"info","ts":1783434567.1355312,"logger":"setup","caller":"runner/runner.go:260","msg":"Controller manager starting"} [e2e-llm-inference-service] {"level":"info","ts":1783434567.1355705,"logger":"controller-runtime.metrics","caller":"server/server.go:208","msg":"Starting metrics server"} [e2e-llm-inference-service] {"level":"info","ts":1783434567.135832,"caller":"runnable/grpc.go:35","msg":"gRPC server starting","name":"health"} [e2e-llm-inference-service] {"level":"info","ts":1783434567.1359007,"caller":"runnable/grpc.go:43","msg":"gRPC server listening","name":"health","port":9003} [e2e-llm-inference-service] {"level":"info","ts":1783434567.136171,"caller":"controller/controller.go:370","msg":"Starting EventSource","controller":"inferencepool","controllerGroup":"inference.networking.k8s.io","controllerKind":"InferencePool","source":"kind source: *v1.InferencePool"} [e2e-llm-inference-service] {"level":"info","ts":1783434567.1366694,"caller":"runnable/grpc.go:35","msg":"gRPC server starting","name":"ext-proc"} [e2e-llm-inference-service] {"level":"info","ts":1783434567.1367416,"caller":"runnable/grpc.go:43","msg":"gRPC server listening","name":"ext-proc","port":9002} [e2e-llm-inference-service] {"level":"info","ts":1783434567.136742,"logger":"controller-runtime.metrics","caller":"server/server.go:247","msg":"Serving metrics server","bindAddress":":9090","secure":false} [e2e-llm-inference-service] {"level":"info","ts":1783434567.1372473,"caller":"controller/controller.go:370","msg":"Starting EventSource","controller":"inferencemodelrewrite","controllerGroup":"inference.networking.x-k8s.io","controllerKind":"InferenceModelRewrite","source":"kind source: *v1alpha2.InferenceModelRewrite"} [e2e-llm-inference-service] {"level":"info","ts":1783434567.137539,"caller":"controller/controller.go:370","msg":"Starting EventSource","controller":"inferenceobjective","controllerGroup":"inference.networking.x-k8s.io","controllerKind":"InferenceObjective","source":"kind source: *v1alpha2.InferenceObjective"} [e2e-llm-inference-service] {"level":"info","ts":1783434567.1375804,"caller":"controller/controller.go:370","msg":"Starting EventSource","controller":"pod","controllerGroup":"","controllerKind":"Pod","source":"kind source: *v1.Pod"} [e2e-llm-inference-service] {"level":"info","ts":1783434567.1453776,"logger":"controller-runtime.cache","caller":"cache/reflector.go:446","msg":"Caches populated","type":"*v1alpha2.InferenceModelRewrite","reflector":"pkg/mod/k8s.io/client-go@v0.35.6/tools/cache/reflector.go:289"} [e2e-llm-inference-service] {"level":"info","ts":1783434567.145392,"logger":"controller-runtime.cache","caller":"cache/reflector.go:446","msg":"Caches populated","type":"*v1alpha2.InferenceObjective","reflector":"pkg/mod/k8s.io/client-go@v0.35.6/tools/cache/reflector.go:289"} [e2e-llm-inference-service] {"level":"info","ts":1783434567.1463535,"logger":"controller-runtime.cache","caller":"cache/reflector.go:446","msg":"Caches populated","type":"*v1.InferencePool","reflector":"pkg/mod/k8s.io/client-go@v0.35.6/tools/cache/reflector.go:289"} [e2e-llm-inference-service] {"level":"info","ts":1783434567.1478944,"logger":"controller-runtime.cache","caller":"cache/reflector.go:446","msg":"Caches populated","type":"*v1.Pod","reflector":"pkg/mod/k8s.io/client-go@v0.35.6/tools/cache/reflector.go:289"} [e2e-llm-inference-service] {"level":"info","ts":1783434567.238626,"caller":"controller/controller.go:303","msg":"Starting Controller","controller":"inferenceobjective","controllerGroup":"inference.networking.x-k8s.io","controllerKind":"InferenceObjective"} [e2e-llm-inference-service] {"level":"info","ts":1783434567.2386656,"caller":"controller/controller.go:303","msg":"Starting Controller","controller":"pod","controllerGroup":"","controllerKind":"Pod"} [e2e-llm-inference-service] {"level":"info","ts":1783434567.2386837,"caller":"controller/controller.go:306","msg":"Starting workers","controller":"inferenceobjective","controllerGroup":"inference.networking.x-k8s.io","controllerKind":"InferenceObjective","worker count":1} [e2e-llm-inference-service] {"level":"info","ts":1783434567.2386932,"caller":"controller/controller.go:306","msg":"Starting workers","controller":"pod","controllerGroup":"","controllerKind":"Pod","worker count":1} [e2e-llm-inference-service] {"level":"info","ts":1783434567.238706,"caller":"controller/controller.go:303","msg":"Starting Controller","controller":"inferencemodelrewrite","controllerGroup":"inference.networking.x-k8s.io","controllerKind":"InferenceModelRewrite"} [e2e-llm-inference-service] {"level":"info","ts":1783434567.2387292,"caller":"controller/controller.go:306","msg":"Starting workers","controller":"inferencemodelrewrite","controllerGroup":"inference.networking.x-k8s.io","controllerKind":"InferenceModelRewrite","worker count":1} [e2e-llm-inference-service] {"level":"info","ts":1783434567.3369718,"caller":"controller/controller.go:303","msg":"Starting Controller","controller":"inferencepool","controllerGroup":"inference.networking.k8s.io","controllerKind":"InferencePool"} [e2e-llm-inference-service] {"level":"info","ts":1783434567.3370214,"caller":"controller/controller.go:306","msg":"Starting workers","controller":"inferencepool","controllerGroup":"inference.networking.k8s.io","controllerKind":"InferencePool","worker count":1} [e2e-llm-inference-service] {"level":"info","ts":1783434567.3371742,"caller":"controller/inferencepool_reconciler.go:46","msg":"Reconciling InferencePool","controller":"inferencepool","controllerGroup":"inference.networking.k8s.io","controllerKind":"InferencePool","InferencePool":{"name":"llmisvc-model-fb-opt-125m-with-7ca60146-inference-pool","namespace":"kserve-ci-e2e-test"},"namespace":"kserve-ci-e2e-test","name":"llmisvc-model-fb-opt-125m-with-7ca60146-inference-pool","reconcileID":"6276ddd9-1209-4f04-9608-04ee300a7588"} [e2e-llm-inference-service] {"level":"info","ts":1783434574.436039,"caller":"controller/inferencepool_reconciler.go:46","msg":"Reconciling InferencePool","controller":"inferencepool","controllerGroup":"inference.networking.k8s.io","controllerKind":"InferencePool","InferencePool":{"name":"llmisvc-model-fb-opt-125m-with-7ca60146-inference-pool","namespace":"kserve-ci-e2e-test"},"namespace":"kserve-ci-e2e-test","name":"llmisvc-model-fb-opt-125m-with-7ca60146-inference-pool","reconcileID":"ffc50f6a-256d-47db-b084-dd9a34987457"} [e2e-llm-inference-service] {"level":"info","ts":1783434645.0251436,"caller":"controller/pod_reconciler.go:99","msg":"Pod already exists","controller":"pod","controllerGroup":"","controllerKind":"Pod","Pod":{"name":"llmisvc-model-fb-opt-125m-with-7ca60146-kserve-7cfdf49d74-pd9px","namespace":"kserve-ci-e2e-test"},"namespace":"kserve-ci-e2e-test","name":"llmisvc-model-fb-opt-125m-with-7ca60146-kserve-7cfdf49d74-pd9px","reconcileID":"3db7eb34-dddc-48e4-99d7-ac0e09f6cf60"} [e2e-llm-inference-service] [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1254 --- [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1255 metadata: [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-7ca60146-epp-service [e2e-llm-inference-service] namespace: kserve-ci-e2e-test [e2e-llm-inference-service] uid: 35fa58aa-de8f-4015-996d-ff1b93dcdf2b [e2e-llm-inference-service] resourceVersion: '28937' [e2e-llm-inference-service] creationTimestamp: '2026-07-07T14:29:24Z' [e2e-llm-inference-service] labels: [e2e-llm-inference-service] app.kubernetes.io/component: llminferenceservice-router-scheduler [e2e-llm-inference-service] app.kubernetes.io/name: llmisvc-model-fb-opt-125m-with-7ca60146 [e2e-llm-inference-service] app.kubernetes.io/part-of: llminferenceservice [e2e-llm-inference-service] endpoints.kubernetes.io/managed-by: endpoint-controller [e2e-llm-inference-service] annotations: [e2e-llm-inference-service] endpoints.kubernetes.io/last-change-trigger-time: '2026-07-07T14:29:58Z' [e2e-llm-inference-service] managedFields: [e2e-llm-inference-service] - manager: kube-controller-manager [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] apiVersion: v1 [e2e-llm-inference-service] time: '2026-07-07T14:29:59Z' [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:metadata: [e2e-llm-inference-service] f:annotations: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:endpoints.kubernetes.io/last-change-trigger-time: {} [e2e-llm-inference-service] f:labels: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:app.kubernetes.io/component: {} [e2e-llm-inference-service] f:app.kubernetes.io/name: {} [e2e-llm-inference-service] f:app.kubernetes.io/part-of: {} [e2e-llm-inference-service] f:endpoints.kubernetes.io/managed-by: {} [e2e-llm-inference-service] f:subsets: {} [e2e-llm-inference-service] subsets: [e2e-llm-inference-service] - addresses: [e2e-llm-inference-service] - ip: 10.132.0.52 [e2e-llm-inference-service] nodeName: ip-10-0-133-25.ec2.internal [e2e-llm-inference-service] targetRef: [e2e-llm-inference-service] kind: Pod [e2e-llm-inference-service] namespace: kserve-ci-e2e-test [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-7ca60146-kserve-router-schenp5mk [e2e-llm-inference-service] uid: ce797511-9ac0-4d16-8bf9-6f34aeec1585 [e2e-llm-inference-service] ports: [e2e-llm-inference-service] - name: grpc-health [e2e-llm-inference-service] port: 9003 [e2e-llm-inference-service] protocol: TCP [e2e-llm-inference-service] - name: metrics [e2e-llm-inference-service] port: 9090 [e2e-llm-inference-service] protocol: TCP [e2e-llm-inference-service] - name: zmq [e2e-llm-inference-service] port: 5557 [e2e-llm-inference-service] protocol: TCP [e2e-llm-inference-service] - name: grpc [e2e-llm-inference-service] port: 9002 [e2e-llm-inference-service] protocol: TCP [e2e-llm-inference-service] apiVersion: v1 [e2e-llm-inference-service] kind: Endpoints [e2e-llm-inference-service] [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1254 --- [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1255 metadata: [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-7ca60146-kserve-workload-svc [e2e-llm-inference-service] namespace: kserve-ci-e2e-test [e2e-llm-inference-service] uid: 13368d4a-4ad5-4654-a6b3-95c4f2098900 [e2e-llm-inference-service] resourceVersion: '29815' [e2e-llm-inference-service] creationTimestamp: '2026-07-07T14:29:24Z' [e2e-llm-inference-service] labels: [e2e-llm-inference-service] app.kubernetes.io/component: llminferenceservice-workload [e2e-llm-inference-service] app.kubernetes.io/name: llmisvc-model-fb-opt-125m-with-7ca60146 [e2e-llm-inference-service] app.kubernetes.io/part-of: llminferenceservice [e2e-llm-inference-service] endpoints.kubernetes.io/managed-by: endpoint-controller [e2e-llm-inference-service] annotations: [e2e-llm-inference-service] endpoints.kubernetes.io/last-change-trigger-time: '2026-07-07T14:30:44Z' [e2e-llm-inference-service] managedFields: [e2e-llm-inference-service] - manager: kube-controller-manager [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] apiVersion: v1 [e2e-llm-inference-service] time: '2026-07-07T14:30:45Z' [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:metadata: [e2e-llm-inference-service] f:annotations: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:endpoints.kubernetes.io/last-change-trigger-time: {} [e2e-llm-inference-service] f:labels: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:app.kubernetes.io/component: {} [e2e-llm-inference-service] f:app.kubernetes.io/name: {} [e2e-llm-inference-service] f:app.kubernetes.io/part-of: {} [e2e-llm-inference-service] f:endpoints.kubernetes.io/managed-by: {} [e2e-llm-inference-service] f:subsets: {} [e2e-llm-inference-service] subsets: [e2e-llm-inference-service] - addresses: [e2e-llm-inference-service] - ip: 10.134.0.30 [e2e-llm-inference-service] nodeName: ip-10-0-139-47.ec2.internal [e2e-llm-inference-service] targetRef: [e2e-llm-inference-service] kind: Pod [e2e-llm-inference-service] namespace: kserve-ci-e2e-test [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-7ca60146-kserve-7cfdf49d74-pd9px [e2e-llm-inference-service] uid: dae89c51-1d5b-4d9c-a3cc-24917b85ad0b [e2e-llm-inference-service] ports: [e2e-llm-inference-service] - name: https [e2e-llm-inference-service] port: 8000 [e2e-llm-inference-service] protocol: TCP [e2e-llm-inference-service] appProtocol: https [e2e-llm-inference-service] apiVersion: v1 [e2e-llm-inference-service] kind: Endpoints [e2e-llm-inference-service] [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1254 --- [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1255 metadata: [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-7ca60146-kserve-7cfdf49d74-pd9px [e2e-llm-inference-service] generateName: llmisvc-model-fb-opt-125m-with-7ca60146-kserve-7cfdf49d74- [e2e-llm-inference-service] namespace: kserve-ci-e2e-test [e2e-llm-inference-service] uid: dae89c51-1d5b-4d9c-a3cc-24917b85ad0b [e2e-llm-inference-service] resourceVersion: '29812' [e2e-llm-inference-service] generation: 1 [e2e-llm-inference-service] creationTimestamp: '2026-07-07T14:29:24Z' [e2e-llm-inference-service] labels: [e2e-llm-inference-service] app.kubernetes.io/component: llminferenceservice-workload [e2e-llm-inference-service] app.kubernetes.io/name: llmisvc-model-fb-opt-125m-with-7ca60146 [e2e-llm-inference-service] app.kubernetes.io/part-of: llminferenceservice [e2e-llm-inference-service] kserve.io/component: workload [e2e-llm-inference-service] llm-d.ai/role: both [e2e-llm-inference-service] pod-template-hash: 7cfdf49d74 [e2e-llm-inference-service] annotations: [e2e-llm-inference-service] k8s.ovn.org/pod-networks: '{"default":{"ip_addresses":["10.134.0.30/23"],"mac_address":"0a:58:0a:86:00:1e","gateway_ips":["10.134.0.1"],"routes":[{"dest":"10.132.0.0/14","nextHop":"10.134.0.1"},{"dest":"172.31.0.0/16","nextHop":"10.134.0.1"},{"dest":"169.254.0.5/32","nextHop":"10.134.0.1"},{"dest":"100.64.0.0/16","nextHop":"10.134.0.1"}],"ip_address":"10.134.0.30/23","gateway_ip":"10.134.0.1","role":"primary"}}' [e2e-llm-inference-service] k8s.v1.cni.cncf.io/network-status: "[{\n \"name\": \"ovn-kubernetes\",\n \ [e2e-llm-inference-service] \ \"interface\": \"eth0\",\n \"ips\": [\n \"10.134.0.30\"\n ],\n\ [e2e-llm-inference-service] \ \"mac\": \"0a:58:0a:86:00:1e\",\n \"default\": true,\n \"dns\": {}\n\ [e2e-llm-inference-service] }]" [e2e-llm-inference-service] openshift.io/scc: restricted-v2 [e2e-llm-inference-service] seccomp.security.alpha.kubernetes.io/pod: runtime/default [e2e-llm-inference-service] security.openshift.io/validated-scc-subject-type: user [e2e-llm-inference-service] ownerReferences: [e2e-llm-inference-service] - apiVersion: apps/v1 [e2e-llm-inference-service] kind: ReplicaSet [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-7ca60146-kserve-7cfdf49d74 [e2e-llm-inference-service] uid: 4c32cb94-4745-4bf7-b4d3-d2be2635a15d [e2e-llm-inference-service] controller: true [e2e-llm-inference-service] blockOwnerDeletion: true [e2e-llm-inference-service] managedFields: [e2e-llm-inference-service] - manager: ip-10-0-139-47 [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] apiVersion: v1 [e2e-llm-inference-service] time: '2026-07-07T14:29:24Z' [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:metadata: [e2e-llm-inference-service] f:annotations: [e2e-llm-inference-service] f:k8s.ovn.org/pod-networks: {} [e2e-llm-inference-service] subresource: status [e2e-llm-inference-service] - manager: kube-controller-manager [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] apiVersion: v1 [e2e-llm-inference-service] time: '2026-07-07T14:29:24Z' [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:metadata: [e2e-llm-inference-service] f:generateName: {} [e2e-llm-inference-service] f:labels: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:app.kubernetes.io/component: {} [e2e-llm-inference-service] f:app.kubernetes.io/name: {} [e2e-llm-inference-service] f:app.kubernetes.io/part-of: {} [e2e-llm-inference-service] f:kserve.io/component: {} [e2e-llm-inference-service] f:llm-d.ai/role: {} [e2e-llm-inference-service] f:pod-template-hash: {} [e2e-llm-inference-service] f:ownerReferences: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"uid":"4c32cb94-4745-4bf7-b4d3-d2be2635a15d"}: {} [e2e-llm-inference-service] f:spec: [e2e-llm-inference-service] f:containers: [e2e-llm-inference-service] k:{"name":"main"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:args: {} [e2e-llm-inference-service] f:command: {} [e2e-llm-inference-service] f:env: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"name":"HF_HUB_CACHE"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"HOME"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"TORCHINDUCTOR_CACHE_DIR"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"USER"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"VLLM_CPU_KVCACHE_SPACE"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"VLLM_ENABLE_V1_MULTIPROCESSING"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"VLLM_LOGGING_LEVEL"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] f:image: {} [e2e-llm-inference-service] f:imagePullPolicy: {} [e2e-llm-inference-service] f:lifecycle: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:preStop: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:exec: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:command: {} [e2e-llm-inference-service] f:livenessProbe: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:failureThreshold: {} [e2e-llm-inference-service] f:httpGet: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:path: {} [e2e-llm-inference-service] f:port: {} [e2e-llm-inference-service] f:scheme: {} [e2e-llm-inference-service] f:periodSeconds: {} [e2e-llm-inference-service] f:successThreshold: {} [e2e-llm-inference-service] f:timeoutSeconds: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:ports: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"containerPort":8000,"protocol":"TCP"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:containerPort: {} [e2e-llm-inference-service] f:protocol: {} [e2e-llm-inference-service] f:readinessProbe: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:failureThreshold: {} [e2e-llm-inference-service] f:httpGet: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:path: {} [e2e-llm-inference-service] f:port: {} [e2e-llm-inference-service] f:scheme: {} [e2e-llm-inference-service] f:periodSeconds: {} [e2e-llm-inference-service] f:successThreshold: {} [e2e-llm-inference-service] f:timeoutSeconds: {} [e2e-llm-inference-service] f:resources: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:limits: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:cpu: {} [e2e-llm-inference-service] f:memory: {} [e2e-llm-inference-service] f:requests: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:cpu: {} [e2e-llm-inference-service] f:memory: {} [e2e-llm-inference-service] f:securityContext: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:allowPrivilegeEscalation: {} [e2e-llm-inference-service] f:capabilities: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:drop: {} [e2e-llm-inference-service] f:readOnlyRootFilesystem: {} [e2e-llm-inference-service] f:runAsNonRoot: {} [e2e-llm-inference-service] f:seccompProfile: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:type: {} [e2e-llm-inference-service] f:startupProbe: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:failureThreshold: {} [e2e-llm-inference-service] f:httpGet: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:path: {} [e2e-llm-inference-service] f:port: {} [e2e-llm-inference-service] f:scheme: {} [e2e-llm-inference-service] f:periodSeconds: {} [e2e-llm-inference-service] f:successThreshold: {} [e2e-llm-inference-service] f:timeoutSeconds: {} [e2e-llm-inference-service] f:terminationMessagePath: {} [e2e-llm-inference-service] f:terminationMessagePolicy: {} [e2e-llm-inference-service] f:volumeMounts: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"mountPath":"/dev/shm"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:mountPath: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] k:{"mountPath":"/home"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:mountPath: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] k:{"mountPath":"/mnt"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:mountPath: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:readOnly: {} [e2e-llm-inference-service] k:{"mountPath":"/models"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:mountPath: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] k:{"mountPath":"/tmp"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:mountPath: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] k:{"mountPath":"/var/run/kserve/tls"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:mountPath: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:readOnly: {} [e2e-llm-inference-service] f:dnsPolicy: {} [e2e-llm-inference-service] f:enableServiceLinks: {} [e2e-llm-inference-service] f:initContainers: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"name":"storage-initializer"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:args: {} [e2e-llm-inference-service] f:env: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"name":"AWS_ACCESS_KEY_ID"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:valueFrom: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:secretKeyRef: {} [e2e-llm-inference-service] k:{"name":"AWS_CA_BUNDLE"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"AWS_CA_BUNDLE_CONFIGMAP"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"AWS_ENDPOINT_URL"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"AWS_SECRET_ACCESS_KEY"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:valueFrom: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:secretKeyRef: {} [e2e-llm-inference-service] k:{"name":"CA_BUNDLE_CONFIGMAP_NAME"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"CA_BUNDLE_VOLUME_MOUNT_POINT"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"HF_HUB_ENABLE_HF_TRANSFER"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"HF_XET_HIGH_PERFORMANCE"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"HF_XET_NUM_CONCURRENT_RANGE_GETS"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"S3_ENDPOINT"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"S3_USE_HTTPS"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"S3_VERIFY_SSL"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] f:image: {} [e2e-llm-inference-service] f:imagePullPolicy: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:resources: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:limits: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:cpu: {} [e2e-llm-inference-service] f:memory: {} [e2e-llm-inference-service] f:requests: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:cpu: {} [e2e-llm-inference-service] f:memory: {} [e2e-llm-inference-service] f:terminationMessagePath: {} [e2e-llm-inference-service] f:terminationMessagePolicy: {} [e2e-llm-inference-service] f:volumeMounts: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"mountPath":"/etc/ssl/custom-certs"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:mountPath: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:readOnly: {} [e2e-llm-inference-service] k:{"mountPath":"/mnt"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:mountPath: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:restartPolicy: {} [e2e-llm-inference-service] f:schedulerName: {} [e2e-llm-inference-service] f:securityContext: {} [e2e-llm-inference-service] f:terminationGracePeriodSeconds: {} [e2e-llm-inference-service] f:volumes: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"name":"cabundle-cert"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:configMap: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:defaultMode: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] k:{"name":"dshm"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:emptyDir: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:medium: {} [e2e-llm-inference-service] f:sizeLimit: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] k:{"name":"home"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:emptyDir: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] k:{"name":"kserve-provision-location"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:emptyDir: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] k:{"name":"model-cache"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:emptyDir: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] k:{"name":"tls-certs"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:secret: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:defaultMode: {} [e2e-llm-inference-service] f:secretName: {} [e2e-llm-inference-service] k:{"name":"tmp-dir"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:emptyDir: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] - manager: multus-daemon [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] apiVersion: v1 [e2e-llm-inference-service] time: '2026-07-07T14:29:25Z' [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:metadata: [e2e-llm-inference-service] f:annotations: [e2e-llm-inference-service] f:k8s.v1.cni.cncf.io/network-status: {} [e2e-llm-inference-service] subresource: status [e2e-llm-inference-service] - manager: kubelet [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] apiVersion: v1 [e2e-llm-inference-service] time: '2026-07-07T14:30:45Z' [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:status: [e2e-llm-inference-service] f:conditions: [e2e-llm-inference-service] k:{"type":"ContainersReady"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:lastProbeTime: {} [e2e-llm-inference-service] f:lastTransitionTime: {} [e2e-llm-inference-service] f:observedGeneration: {} [e2e-llm-inference-service] f:status: {} [e2e-llm-inference-service] f:type: {} [e2e-llm-inference-service] k:{"type":"Initialized"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:lastProbeTime: {} [e2e-llm-inference-service] f:lastTransitionTime: {} [e2e-llm-inference-service] f:observedGeneration: {} [e2e-llm-inference-service] f:status: {} [e2e-llm-inference-service] f:type: {} [e2e-llm-inference-service] k:{"type":"PodReadyToStartContainers"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:lastProbeTime: {} [e2e-llm-inference-service] f:lastTransitionTime: {} [e2e-llm-inference-service] f:observedGeneration: {} [e2e-llm-inference-service] f:status: {} [e2e-llm-inference-service] f:type: {} [e2e-llm-inference-service] k:{"type":"PodScheduled"}: [e2e-llm-inference-service] f:observedGeneration: {} [e2e-llm-inference-service] k:{"type":"Ready"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:lastProbeTime: {} [e2e-llm-inference-service] f:lastTransitionTime: {} [e2e-llm-inference-service] f:observedGeneration: {} [e2e-llm-inference-service] f:status: {} [e2e-llm-inference-service] f:type: {} [e2e-llm-inference-service] f:containerStatuses: {} [e2e-llm-inference-service] f:hostIP: {} [e2e-llm-inference-service] f:hostIPs: {} [e2e-llm-inference-service] f:initContainerStatuses: {} [e2e-llm-inference-service] f:observedGeneration: {} [e2e-llm-inference-service] f:phase: {} [e2e-llm-inference-service] f:podIP: {} [e2e-llm-inference-service] f:podIPs: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"ip":"10.134.0.30"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:ip: {} [e2e-llm-inference-service] f:startTime: {} [e2e-llm-inference-service] subresource: status [e2e-llm-inference-service] spec: [e2e-llm-inference-service] volumes: [e2e-llm-inference-service] - name: home [e2e-llm-inference-service] emptyDir: {} [e2e-llm-inference-service] - name: dshm [e2e-llm-inference-service] emptyDir: [e2e-llm-inference-service] medium: Memory [e2e-llm-inference-service] sizeLimit: 1Gi [e2e-llm-inference-service] - name: model-cache [e2e-llm-inference-service] emptyDir: {} [e2e-llm-inference-service] - name: tmp-dir [e2e-llm-inference-service] emptyDir: {} [e2e-llm-inference-service] - name: tls-certs [e2e-llm-inference-service] secret: [e2e-llm-inference-service] secretName: llmisv3e414c2ba058a022dfd694dbcbac5b51-kserve-self-signed-certs [e2e-llm-inference-service] defaultMode: 420 [e2e-llm-inference-service] - name: kserve-provision-location [e2e-llm-inference-service] emptyDir: {} [e2e-llm-inference-service] - name: cabundle-cert [e2e-llm-inference-service] configMap: [e2e-llm-inference-service] name: odh-kserve-custom-ca-bundle [e2e-llm-inference-service] defaultMode: 420 [e2e-llm-inference-service] - name: kube-api-access-p7n7g [e2e-llm-inference-service] projected: [e2e-llm-inference-service] sources: [e2e-llm-inference-service] - serviceAccountToken: [e2e-llm-inference-service] expirationSeconds: 3607 [e2e-llm-inference-service] path: token [e2e-llm-inference-service] - configMap: [e2e-llm-inference-service] name: kube-root-ca.crt [e2e-llm-inference-service] items: [e2e-llm-inference-service] - key: ca.crt [e2e-llm-inference-service] path: ca.crt [e2e-llm-inference-service] - downwardAPI: [e2e-llm-inference-service] items: [e2e-llm-inference-service] - path: namespace [e2e-llm-inference-service] fieldRef: [e2e-llm-inference-service] apiVersion: v1 [e2e-llm-inference-service] fieldPath: metadata.namespace [e2e-llm-inference-service] - configMap: [e2e-llm-inference-service] name: openshift-service-ca.crt [e2e-llm-inference-service] items: [e2e-llm-inference-service] - key: service-ca.crt [e2e-llm-inference-service] path: service-ca.crt [e2e-llm-inference-service] defaultMode: 420 [e2e-llm-inference-service] initContainers: [e2e-llm-inference-service] - name: storage-initializer [e2e-llm-inference-service] image: quay.io/opendatahub/kserve-storage-initializer@sha256:13ad765c0f323200e1db7b45c99661287e65c5158d6b54b16139829636d20be4 [e2e-llm-inference-service] args: [e2e-llm-inference-service] - hf://facebook/opt-125m [e2e-llm-inference-service] - /mnt/models [e2e-llm-inference-service] - hf://edbeeching/opt-125m-lora [e2e-llm-inference-service] - /mnt/lora/lora-adapter-1 [e2e-llm-inference-service] env: [e2e-llm-inference-service] - name: AWS_ACCESS_KEY_ID [e2e-llm-inference-service] valueFrom: [e2e-llm-inference-service] secretKeyRef: [e2e-llm-inference-service] name: seaweedfs-s3-creds [e2e-llm-inference-service] key: AWS_ACCESS_KEY_ID [e2e-llm-inference-service] - name: AWS_SECRET_ACCESS_KEY [e2e-llm-inference-service] valueFrom: [e2e-llm-inference-service] secretKeyRef: [e2e-llm-inference-service] name: seaweedfs-s3-creds [e2e-llm-inference-service] key: AWS_SECRET_ACCESS_KEY [e2e-llm-inference-service] - name: S3_USE_HTTPS [e2e-llm-inference-service] value: '0' [e2e-llm-inference-service] - name: S3_ENDPOINT [e2e-llm-inference-service] value: s3-service.kserve:8333 [e2e-llm-inference-service] - name: AWS_ENDPOINT_URL [e2e-llm-inference-service] value: http://s3-service.kserve:8333 [e2e-llm-inference-service] - name: S3_VERIFY_SSL [e2e-llm-inference-service] value: '0' [e2e-llm-inference-service] - name: AWS_CA_BUNDLE [e2e-llm-inference-service] value: /etc/ssl/custom-certs/cabundle.crt [e2e-llm-inference-service] - name: AWS_CA_BUNDLE_CONFIGMAP [e2e-llm-inference-service] value: odh-kserve-custom-ca-bundle [e2e-llm-inference-service] - name: HF_HUB_ENABLE_HF_TRANSFER [e2e-llm-inference-service] value: '1' [e2e-llm-inference-service] - name: HF_XET_HIGH_PERFORMANCE [e2e-llm-inference-service] value: '1' [e2e-llm-inference-service] - name: HF_XET_NUM_CONCURRENT_RANGE_GETS [e2e-llm-inference-service] value: '8' [e2e-llm-inference-service] - name: CA_BUNDLE_CONFIGMAP_NAME [e2e-llm-inference-service] value: odh-kserve-custom-ca-bundle [e2e-llm-inference-service] - name: CA_BUNDLE_VOLUME_MOUNT_POINT [e2e-llm-inference-service] value: /etc/ssl/custom-certs [e2e-llm-inference-service] resources: [e2e-llm-inference-service] limits: [e2e-llm-inference-service] cpu: '1' [e2e-llm-inference-service] memory: 24Gi [e2e-llm-inference-service] requests: [e2e-llm-inference-service] cpu: 100m [e2e-llm-inference-service] memory: 100Mi [e2e-llm-inference-service] volumeMounts: [e2e-llm-inference-service] - name: kserve-provision-location [e2e-llm-inference-service] mountPath: /mnt [e2e-llm-inference-service] - name: cabundle-cert [e2e-llm-inference-service] readOnly: true [e2e-llm-inference-service] mountPath: /etc/ssl/custom-certs [e2e-llm-inference-service] - name: kube-api-access-p7n7g [e2e-llm-inference-service] readOnly: true [e2e-llm-inference-service] mountPath: /var/run/secrets/kubernetes.io/serviceaccount [e2e-llm-inference-service] terminationMessagePath: /dev/termination-log [e2e-llm-inference-service] terminationMessagePolicy: FallbackToLogsOnError [e2e-llm-inference-service] imagePullPolicy: IfNotPresent [e2e-llm-inference-service] securityContext: [e2e-llm-inference-service] capabilities: [e2e-llm-inference-service] drop: [e2e-llm-inference-service] - ALL [e2e-llm-inference-service] runAsUser: 1000690000 [e2e-llm-inference-service] runAsNonRoot: true [e2e-llm-inference-service] allowPrivilegeEscalation: false [e2e-llm-inference-service] containers: [e2e-llm-inference-service] - name: main [e2e-llm-inference-service] image: public.ecr.aws/q9t5s3a7/vllm-cpu-release-repo:v0.19.0 [e2e-llm-inference-service] command: [e2e-llm-inference-service] - /bin/bash [e2e-llm-inference-service] - -c [e2e-llm-inference-service] - "if [ -f /etc/profile.d/ibm-aiu-setup.sh ]; then\n source /etc/profile.d/ibm-aiu-setup.sh\n\ [e2e-llm-inference-service] fi\n\nif [ \"$KSERVE_INFER_ROCE\" = \"true\" ]; then\n echo \"Trying to infer\ [e2e-llm-inference-service] \ RoCE configs ... \"\n grep -H . /sys/class/infiniband/*/ports/*/gids/* 2>/dev/null\n\ [e2e-llm-inference-service] \ grep -H . /sys/class/infiniband/*/ports/*/gid_attrs/types/* 2>/dev/null\n\ [e2e-llm-inference-service] \n cat /proc/driver/nvidia/params\n\n KSERVE_INFER_IB_GID_INDEX_GREP=${KSERVE_INFER_IB_GID_INDEX_GREP:-\"\ [e2e-llm-inference-service] RoCE v2\"}\n\n echo \"[Infer RoCE] Discovering active HCAs ...\"\n active_hcas=()\n\ [e2e-llm-inference-service] \ # Loop through all mlx5 devices found in sysfs\n for hca_dir in /sys/class/infiniband/mlx5_*;\ [e2e-llm-inference-service] \ do\n # Ensure it's a directory before proceeding\n if [ -d \"$hca_dir\"\ [e2e-llm-inference-service] \ ]; then\n hca_name=$(basename \"$hca_dir\")\n port_state_file=\"\ [e2e-llm-inference-service] $hca_dir/ports/1/state\" # Assume port 1\n type_file=\"$hca_dir/ports/1/gid_attrs/types/*\"\ [e2e-llm-inference-service] \n\n echo \"[Infer RoCE] Check if the port state file ${port_state_file}\ [e2e-llm-inference-service] \ exists and contains 'ACTIVE'\"\n if [ -f \"$port_state_file\" ] &&\ [e2e-llm-inference-service] \ grep -q \"ACTIVE\" \"$port_state_file\" && grep -q \"${KSERVE_INFER_IB_GID_INDEX_GREP}\"\ [e2e-llm-inference-service] \ ${type_file} 2>/dev/null; then\n echo \"[Infer RoCE] Found active\ [e2e-llm-inference-service] \ HCA: $hca_name\"\n active_hcas+=(\"$hca_name\")\n else\n\ [e2e-llm-inference-service] \ echo \"[Infer RoCE] Skipping inactive or down HCA: $hca_name\"\ [e2e-llm-inference-service] \n fi\n fi\n done\n\n # Check if we found any active HCAs\n\ [e2e-llm-inference-service] \ if [ ${#active_hcas[@]} -gt 0 ]; then\n # Join the array elements with\ [e2e-llm-inference-service] \ a comma\n hca_port_pairs=()\n for hca in \"${active_hcas[@]}\";\ [e2e-llm-inference-service] \ do\n hca_port_pairs+=(\"${hca}:1\")\n done\n\n active_hca_list=$(IFS=,;\ [e2e-llm-inference-service] \ echo \"${active_hcas[*]}\")\n hca_port_pairs_list=$(IFS=,; echo \"${hca_port_pairs[*]}\"\ [e2e-llm-inference-service] )\n echo \"[Infer RoCE] Setting active HCAs: ${active_hca_list}\"\n \ [e2e-llm-inference-service] \ export NCCL_IB_HCA=${NCCL_IB_HCA:-${active_hca_list}}\n export NVSHMEM_HCA_LIST=${NVSHMEM_HCA_LIST:-${hca_port_pairs_list}}\n\ [e2e-llm-inference-service] \ export UCX_NET_DEVICES=${UCX_NET_DEVICES:-${hca_port_pairs_list}}\n\n\ [e2e-llm-inference-service] \ echo \"[Infer RoCE] NCCL_IB_HCA=${NCCL_IB_HCA}\"\n echo \"[Infer\ [e2e-llm-inference-service] \ RoCE] NVSHMEM_HCA_LIST=${NVSHMEM_HCA_LIST}\"\n echo \"[Infer RoCE] UCX_NET_DEVICES=${UCX_NET_DEVICES}\"\ [e2e-llm-inference-service] \n else\n echo \"[Infer RoCE] WARNING: No active RoCE HCAs found. NCCL_IB_HCA\ [e2e-llm-inference-service] \ will not be set.\"\n fi\n\n if [ ${#active_hcas[@]} -gt 0 ]; then\n \ [e2e-llm-inference-service] \ echo \"[Infer RoCE] Finding GID_INDEX for each active HCA (SR-IOV compatible)...\"\ [e2e-llm-inference-service] \n\n # For SR-IOV environments, find the most common IPv4 RoCE v2 GID index\ [e2e-llm-inference-service] \ across all HCAs\n declare -A gid_index_count\n declare -A hca_gid_index\n\ [e2e-llm-inference-service] \n for hca_name in \"${active_hcas[@]}\"; do\n echo \"[Infer RoCE]\ [e2e-llm-inference-service] \ Processing HCA: ${hca_name}\"\n\n # Find all RoCE v2 IPv4 GIDs for\ [e2e-llm-inference-service] \ this HCA and count by index\n for tpath in /sys/class/infiniband/${hca_name}/ports/1/gid_attrs/types/*;\ [e2e-llm-inference-service] \ do\n if grep -q \"${KSERVE_INFER_IB_GID_INDEX_GREP}\" \"$tpath\"\ [e2e-llm-inference-service] \ 2>/dev/null; then\n idx=$(basename \"$tpath\")\n \ [e2e-llm-inference-service] \ gid_file=\"/sys/class/infiniband/${hca_name}/ports/1/gids/${idx}\"\ [e2e-llm-inference-service] \n # Check for IPv4 GID (contains ffff:)\n \ [e2e-llm-inference-service] \ if [ -f \"$gid_file\" ] && grep -q \"ffff:\" \"$gid_file\"; then\n \ [e2e-llm-inference-service] \ gid_value=$(cat \"$gid_file\" 2>/dev/null || echo \"\")\n \ [e2e-llm-inference-service] \ echo \"[Infer RoCE] Found IPv4 RoCE v2 GID for ${hca_name}:\ [e2e-llm-inference-service] \ index=${idx}, gid=${gid_value}\"\n hca_gid_index[\"${hca_name}\"\ [e2e-llm-inference-service] ]=\"${idx}\"\n gid_index_count[\"${idx}\"]=$((${gid_index_count[\"\ [e2e-llm-inference-service] ${idx}\"]} + 1))\n break # Use first found IPv4 GID per\ [e2e-llm-inference-service] \ HCA\n fi\n fi\n done\n done\n\n\ [e2e-llm-inference-service] \ # Find the most common GID index (most likely to be consistent across\ [e2e-llm-inference-service] \ nodes)\n best_gid_index=\"\"\n max_count=0\n for idx in \"\ [e2e-llm-inference-service] ${!gid_index_count[@]}\"; do\n count=${gid_index_count[\"${idx}\"]}\n\ [e2e-llm-inference-service] \ echo \"[Infer RoCE] GID_INDEX ${idx} found on ${count} HCAs\"\n \ [e2e-llm-inference-service] \ if [ $count -gt $max_count ]; then\n max_count=$count\n\ [e2e-llm-inference-service] \ best_gid_index=\"$idx\"\n fi\n done\n\n #\ [e2e-llm-inference-service] \ Use deterministic fallback if tied - prefer index 3 (SR-IOV standard)\n \ [e2e-llm-inference-service] \ if [ ${#gid_index_count[@]} -gt 1 ]; then\n echo \"[Infer RoCE]\ [e2e-llm-inference-service] \ Multiple GID indices found, selecting most common: ${best_gid_index}\"\n \ [e2e-llm-inference-service] \ # If there's a tie, prefer index 3 as it's most common in SR-IOV setups\n\ [e2e-llm-inference-service] \ if [ -n \"${gid_index_count['3']}\" ] && [ \"${gid_index_count['3']}\"\ [e2e-llm-inference-service] \ -eq \"$max_count\" ]; then\n best_gid_index=\"3\"\n \ [e2e-llm-inference-service] \ echo \"[Infer RoCE] Using deterministic fallback: GID_INDEX=3 (SR-IOV\ [e2e-llm-inference-service] \ standard)\"\n fi\n fi\n\n # Check if GID_INDEX is already\ [e2e-llm-inference-service] \ set via environment variables\n if [ -n \"${NCCL_IB_GID_INDEX}\" ]; then\n\ [e2e-llm-inference-service] \ echo \"[Infer RoCE] Using pre-configured NCCL_IB_GID_INDEX=${NCCL_IB_GID_INDEX}\ [e2e-llm-inference-service] \ from environment\"\n export NVSHMEM_IB_GID_INDEX=${NVSHMEM_IB_GID_INDEX:-$NCCL_IB_GID_INDEX}\n\ [e2e-llm-inference-service] \ export UCX_IB_GID_INDEX=${UCX_IB_GID_INDEX:-$NCCL_IB_GID_INDEX}\n\ [e2e-llm-inference-service] \ echo \"[Infer RoCE] Using pre-configured GID_INDEX=${NCCL_IB_GID_INDEX}\ [e2e-llm-inference-service] \ for NCCL, NVSHMEM, and UCX\"\n elif [ -n \"$best_gid_index\" ]; then\n\ [e2e-llm-inference-service] \ echo \"[Infer RoCE] Selected GID_INDEX: ${best_gid_index} (found\ [e2e-llm-inference-service] \ on ${max_count} HCAs)\"\n\n export NCCL_IB_GID_INDEX=${NCCL_IB_GID_INDEX:-$best_gid_index}\n\ [e2e-llm-inference-service] \ export NVSHMEM_IB_GID_INDEX=${NVSHMEM_IB_GID_INDEX:-$best_gid_index}\n\ [e2e-llm-inference-service] \ export UCX_IB_GID_INDEX=${UCX_IB_GID_INDEX:-$best_gid_index}\n\n\ [e2e-llm-inference-service] \ echo \"[Infer RoCE] Exported GID_INDEX=${best_gid_index} for NCCL,\ [e2e-llm-inference-service] \ NVSHMEM, and UCX\"\n else\n echo \"[Infer RoCE] ERROR: No valid\ [e2e-llm-inference-service] \ IPv4 ${KSERVE_INFER_IB_GID_INDEX_GREP} GID_INDEX found on any HCA.\"\n \ [e2e-llm-inference-service] \ fi\n else\n echo \"[Infer RoCE] No active HCAs found, skipping GID_INDEX\ [e2e-llm-inference-service] \ inference.\"\n fi\nfi\n\n# --disable-access-log-for-endpoints landed in vLLM\ [e2e-llm-inference-service] \ 0.16.0 (vllm-project/vllm#30011).\n# Older versions still need the blanket\ [e2e-llm-inference-service] \ --disable-uvicorn-access-log.\nACCESS_LOG_ARGS=\"--disable-uvicorn-access-log\"\ [e2e-llm-inference-service] \nVLLM_VERSION=$(vllm --version 2>/dev/null | tail -1 | awk '{print $NF}')\n\ [e2e-llm-inference-service] echo \"[access-log-detect] vllm version='${VLLM_VERSION}'\"\nif [[ \"$VLLM_VERSION\"\ [e2e-llm-inference-service] \ =~ ^[0-9]+\\.[0-9]+ ]] && [ \"$(printf '%s\\n%s\\n' \"0.16.0\" \"${VLLM_VERSION}\"\ [e2e-llm-inference-service] \ | sort -V | head -1)\" = \"0.16.0\" ]; then\n ACCESS_LOG_ARGS=\"--disable-access-log-for-endpoints\ [e2e-llm-inference-service] \ /health,/metrics,/ping\"\nfi\necho \"[access-log-detect] selected ACCESS_LOG_ARGS='${ACCESS_LOG_ARGS}'\"\ [e2e-llm-inference-service] \n\n# --shutdown-timeout landed in vLLM 0.18.0 (vllm-project/vllm#36666).\n\ [e2e-llm-inference-service] SHUTDOWN_TIMEOUT_ARGS=\"\"\nif [[ \"$VLLM_VERSION\" =~ ^[0-9]+\\.[0-9]+ ]] &&\ [e2e-llm-inference-service] \ [ \"$(printf '%s\\n%s\\n' \"0.18.0\" \"${VLLM_VERSION}\" | sort -V | head\ [e2e-llm-inference-service] \ -1)\" = \"0.18.0\" ]; then\n SHUTDOWN_TIMEOUT_ARGS=\"--shutdown-timeout 40\"\ [e2e-llm-inference-service] \nfi\n\neval \"exec vllm serve /mnt/models \\\n --served-model-name \"facebook/opt-125m\"\ [e2e-llm-inference-service] \ \"publishers/kserve-ci-e2e-test/models/facebook/opt-125m\" \\\n --port 8000\ [e2e-llm-inference-service] \ \\\n ${ACCESS_LOG_ARGS} \\\n ${SHUTDOWN_TIMEOUT_ARGS} \\\n --enable-ssl-refresh\ [e2e-llm-inference-service] \ \\\n --ssl-certfile /var/run/kserve/tls/tls.crt \\\n --ssl-keyfile /var/run/kserve/tls/tls.key\ [e2e-llm-inference-service] \ \\\n ${VLLM_ADDITIONAL_ARGS} \\\n $@\"" [e2e-llm-inference-service] - -- [e2e-llm-inference-service] args: [e2e-llm-inference-service] - --enable-lora [e2e-llm-inference-service] - --lora-modules [e2e-llm-inference-service] - '''{"name":"lora-adapter-1","path":"/mnt/lora/lora-adapter-1"}''' [e2e-llm-inference-service] - '''{"name":"publishers/kserve-ci-e2e-test/models/lora-adapter-1","path":"/mnt/lora/lora-adapter-1"}''' [e2e-llm-inference-service] ports: [e2e-llm-inference-service] - containerPort: 8000 [e2e-llm-inference-service] protocol: TCP [e2e-llm-inference-service] env: [e2e-llm-inference-service] - name: HOME [e2e-llm-inference-service] value: /home [e2e-llm-inference-service] - name: VLLM_LOGGING_LEVEL [e2e-llm-inference-service] value: DEBUG [e2e-llm-inference-service] - name: VLLM_CPU_KVCACHE_SPACE [e2e-llm-inference-service] value: '1' [e2e-llm-inference-service] - name: VLLM_ENABLE_V1_MULTIPROCESSING [e2e-llm-inference-service] value: '0' [e2e-llm-inference-service] - name: USER [e2e-llm-inference-service] value: nonroot [e2e-llm-inference-service] - name: TORCHINDUCTOR_CACHE_DIR [e2e-llm-inference-service] value: /tmp/torchinductor-cache [e2e-llm-inference-service] - name: HF_HUB_CACHE [e2e-llm-inference-service] value: /models [e2e-llm-inference-service] resources: [e2e-llm-inference-service] limits: [e2e-llm-inference-service] cpu: '2' [e2e-llm-inference-service] memory: 7Gi [e2e-llm-inference-service] requests: [e2e-llm-inference-service] cpu: 200m [e2e-llm-inference-service] memory: 2Gi [e2e-llm-inference-service] volumeMounts: [e2e-llm-inference-service] - name: home [e2e-llm-inference-service] mountPath: /home [e2e-llm-inference-service] - name: tmp-dir [e2e-llm-inference-service] mountPath: /tmp [e2e-llm-inference-service] - name: dshm [e2e-llm-inference-service] mountPath: /dev/shm [e2e-llm-inference-service] - name: model-cache [e2e-llm-inference-service] mountPath: /models [e2e-llm-inference-service] - name: tls-certs [e2e-llm-inference-service] readOnly: true [e2e-llm-inference-service] mountPath: /var/run/kserve/tls [e2e-llm-inference-service] - name: kserve-provision-location [e2e-llm-inference-service] readOnly: true [e2e-llm-inference-service] mountPath: /mnt [e2e-llm-inference-service] - name: kube-api-access-p7n7g [e2e-llm-inference-service] readOnly: true [e2e-llm-inference-service] mountPath: /var/run/secrets/kubernetes.io/serviceaccount [e2e-llm-inference-service] livenessProbe: [e2e-llm-inference-service] httpGet: [e2e-llm-inference-service] path: /health [e2e-llm-inference-service] port: 8000 [e2e-llm-inference-service] scheme: HTTPS [e2e-llm-inference-service] timeoutSeconds: 1 [e2e-llm-inference-service] periodSeconds: 10 [e2e-llm-inference-service] successThreshold: 1 [e2e-llm-inference-service] failureThreshold: 10 [e2e-llm-inference-service] readinessProbe: [e2e-llm-inference-service] httpGet: [e2e-llm-inference-service] path: /health [e2e-llm-inference-service] port: 8000 [e2e-llm-inference-service] scheme: HTTPS [e2e-llm-inference-service] timeoutSeconds: 1 [e2e-llm-inference-service] periodSeconds: 1 [e2e-llm-inference-service] successThreshold: 1 [e2e-llm-inference-service] failureThreshold: 2 [e2e-llm-inference-service] startupProbe: [e2e-llm-inference-service] httpGet: [e2e-llm-inference-service] path: /health [e2e-llm-inference-service] port: 8000 [e2e-llm-inference-service] scheme: HTTPS [e2e-llm-inference-service] timeoutSeconds: 1 [e2e-llm-inference-service] periodSeconds: 10 [e2e-llm-inference-service] successThreshold: 1 [e2e-llm-inference-service] failureThreshold: 60 [e2e-llm-inference-service] lifecycle: [e2e-llm-inference-service] preStop: [e2e-llm-inference-service] exec: [e2e-llm-inference-service] command: [e2e-llm-inference-service] - /bin/sleep [e2e-llm-inference-service] - '15' [e2e-llm-inference-service] terminationMessagePath: /dev/termination-log [e2e-llm-inference-service] terminationMessagePolicy: FallbackToLogsOnError [e2e-llm-inference-service] imagePullPolicy: IfNotPresent [e2e-llm-inference-service] securityContext: [e2e-llm-inference-service] capabilities: [e2e-llm-inference-service] drop: [e2e-llm-inference-service] - ALL [e2e-llm-inference-service] runAsUser: 1000690000 [e2e-llm-inference-service] runAsNonRoot: true [e2e-llm-inference-service] readOnlyRootFilesystem: true [e2e-llm-inference-service] allowPrivilegeEscalation: false [e2e-llm-inference-service] seccompProfile: [e2e-llm-inference-service] type: RuntimeDefault [e2e-llm-inference-service] restartPolicy: Always [e2e-llm-inference-service] terminationGracePeriodSeconds: 60 [e2e-llm-inference-service] dnsPolicy: ClusterFirst [e2e-llm-inference-service] serviceAccountName: default [e2e-llm-inference-service] serviceAccount: default [e2e-llm-inference-service] nodeName: ip-10-0-139-47.ec2.internal [e2e-llm-inference-service] securityContext: [e2e-llm-inference-service] seLinuxOptions: [e2e-llm-inference-service] level: s0:c26,c20 [e2e-llm-inference-service] fsGroup: 1000690000 [e2e-llm-inference-service] seccompProfile: [e2e-llm-inference-service] type: RuntimeDefault [e2e-llm-inference-service] imagePullSecrets: [e2e-llm-inference-service] - name: default-dockercfg-tbbgr [e2e-llm-inference-service] schedulerName: default-scheduler [e2e-llm-inference-service] tolerations: [e2e-llm-inference-service] - key: node.kubernetes.io/not-ready [e2e-llm-inference-service] operator: Exists [e2e-llm-inference-service] effect: NoExecute [e2e-llm-inference-service] tolerationSeconds: 300 [e2e-llm-inference-service] - key: node.kubernetes.io/unreachable [e2e-llm-inference-service] operator: Exists [e2e-llm-inference-service] effect: NoExecute [e2e-llm-inference-service] tolerationSeconds: 300 [e2e-llm-inference-service] - key: node.kubernetes.io/memory-pressure [e2e-llm-inference-service] operator: Exists [e2e-llm-inference-service] effect: NoSchedule [e2e-llm-inference-service] priority: 0 [e2e-llm-inference-service] enableServiceLinks: true [e2e-llm-inference-service] preemptionPolicy: PreemptLowerPriority [e2e-llm-inference-service] status: [e2e-llm-inference-service] observedGeneration: 1 [e2e-llm-inference-service] phase: Running [e2e-llm-inference-service] conditions: [e2e-llm-inference-service] - type: PodReadyToStartContainers [e2e-llm-inference-service] observedGeneration: 1 [e2e-llm-inference-service] status: 'True' [e2e-llm-inference-service] lastProbeTime: null [e2e-llm-inference-service] lastTransitionTime: '2026-07-07T14:29:25Z' [e2e-llm-inference-service] - type: Initialized [e2e-llm-inference-service] observedGeneration: 1 [e2e-llm-inference-service] status: 'True' [e2e-llm-inference-service] lastProbeTime: null [e2e-llm-inference-service] lastTransitionTime: '2026-07-07T14:29:30Z' [e2e-llm-inference-service] - type: Ready [e2e-llm-inference-service] observedGeneration: 1 [e2e-llm-inference-service] status: 'True' [e2e-llm-inference-service] lastProbeTime: null [e2e-llm-inference-service] lastTransitionTime: '2026-07-07T14:30:44Z' [e2e-llm-inference-service] - type: ContainersReady [e2e-llm-inference-service] observedGeneration: 1 [e2e-llm-inference-service] status: 'True' [e2e-llm-inference-service] lastProbeTime: null [e2e-llm-inference-service] lastTransitionTime: '2026-07-07T14:30:44Z' [e2e-llm-inference-service] - type: PodScheduled [e2e-llm-inference-service] observedGeneration: 1 [e2e-llm-inference-service] status: 'True' [e2e-llm-inference-service] lastProbeTime: null [e2e-llm-inference-service] lastTransitionTime: '2026-07-07T14:29:24Z' [e2e-llm-inference-service] hostIP: 10.0.139.47 [e2e-llm-inference-service] hostIPs: [e2e-llm-inference-service] - ip: 10.0.139.47 [e2e-llm-inference-service] podIP: 10.134.0.30 [e2e-llm-inference-service] podIPs: [e2e-llm-inference-service] - ip: 10.134.0.30 [e2e-llm-inference-service] startTime: '2026-07-07T14:29:24Z' [e2e-llm-inference-service] initContainerStatuses: [e2e-llm-inference-service] - name: storage-initializer [e2e-llm-inference-service] state: [e2e-llm-inference-service] terminated: [e2e-llm-inference-service] exitCode: 0 [e2e-llm-inference-service] reason: Completed [e2e-llm-inference-service] startedAt: '2026-07-07T14:29:25Z' [e2e-llm-inference-service] finishedAt: '2026-07-07T14:29:30Z' [e2e-llm-inference-service] containerID: cri-o://7cebcb04222774aaba8c29354d9f18f01b6023cbd4163fb16882c5ca80e0a13d [e2e-llm-inference-service] lastState: {} [e2e-llm-inference-service] ready: true [e2e-llm-inference-service] restartCount: 0 [e2e-llm-inference-service] image: quay.io/opendatahub/kserve-storage-initializer@sha256:13ad765c0f323200e1db7b45c99661287e65c5158d6b54b16139829636d20be4 [e2e-llm-inference-service] imageID: quay.io/opendatahub/kserve-storage-initializer@sha256:13ad765c0f323200e1db7b45c99661287e65c5158d6b54b16139829636d20be4 [e2e-llm-inference-service] containerID: cri-o://7cebcb04222774aaba8c29354d9f18f01b6023cbd4163fb16882c5ca80e0a13d [e2e-llm-inference-service] started: false [e2e-llm-inference-service] allocatedResources: [e2e-llm-inference-service] cpu: 100m [e2e-llm-inference-service] memory: 100Mi [e2e-llm-inference-service] resources: [e2e-llm-inference-service] limits: [e2e-llm-inference-service] cpu: '1' [e2e-llm-inference-service] memory: 24Gi [e2e-llm-inference-service] requests: [e2e-llm-inference-service] cpu: 100m [e2e-llm-inference-service] memory: 100Mi [e2e-llm-inference-service] volumeMounts: [e2e-llm-inference-service] - name: kserve-provision-location [e2e-llm-inference-service] mountPath: /mnt [e2e-llm-inference-service] - name: cabundle-cert [e2e-llm-inference-service] mountPath: /etc/ssl/custom-certs [e2e-llm-inference-service] readOnly: true [e2e-llm-inference-service] recursiveReadOnly: Disabled [e2e-llm-inference-service] - name: kube-api-access-p7n7g [e2e-llm-inference-service] mountPath: /var/run/secrets/kubernetes.io/serviceaccount [e2e-llm-inference-service] readOnly: true [e2e-llm-inference-service] recursiveReadOnly: Disabled [e2e-llm-inference-service] user: [e2e-llm-inference-service] linux: [e2e-llm-inference-service] uid: 1000690000 [e2e-llm-inference-service] gid: 0 [e2e-llm-inference-service] supplementalGroups: [e2e-llm-inference-service] - 0 [e2e-llm-inference-service] - 1000690000 [e2e-llm-inference-service] containerStatuses: [e2e-llm-inference-service] - name: main [e2e-llm-inference-service] state: [e2e-llm-inference-service] running: [e2e-llm-inference-service] startedAt: '2026-07-07T14:29:30Z' [e2e-llm-inference-service] lastState: {} [e2e-llm-inference-service] ready: true [e2e-llm-inference-service] restartCount: 0 [e2e-llm-inference-service] image: public.ecr.aws/q9t5s3a7/vllm-cpu-release-repo:v0.19.0 [e2e-llm-inference-service] imageID: public.ecr.aws/q9t5s3a7/vllm-cpu-release-repo@sha256:afb39fca138b51d019d986229d546531b45a2a3deb73bcf59bd42406e13fbba0 [e2e-llm-inference-service] containerID: cri-o://35d8a33929fbfc11201f69efa769b944a570e61a916bcc19157f10b8e371f778 [e2e-llm-inference-service] started: true [e2e-llm-inference-service] allocatedResources: [e2e-llm-inference-service] cpu: 200m [e2e-llm-inference-service] memory: 2Gi [e2e-llm-inference-service] resources: [e2e-llm-inference-service] limits: [e2e-llm-inference-service] cpu: '2' [e2e-llm-inference-service] memory: 7Gi [e2e-llm-inference-service] requests: [e2e-llm-inference-service] cpu: 200m [e2e-llm-inference-service] memory: 2Gi [e2e-llm-inference-service] volumeMounts: [e2e-llm-inference-service] - name: home [e2e-llm-inference-service] mountPath: /home [e2e-llm-inference-service] - name: tmp-dir [e2e-llm-inference-service] mountPath: /tmp [e2e-llm-inference-service] - name: dshm [e2e-llm-inference-service] mountPath: /dev/shm [e2e-llm-inference-service] - name: model-cache [e2e-llm-inference-service] mountPath: /models [e2e-llm-inference-service] - name: tls-certs [e2e-llm-inference-service] mountPath: /var/run/kserve/tls [e2e-llm-inference-service] readOnly: true [e2e-llm-inference-service] recursiveReadOnly: Disabled [e2e-llm-inference-service] - name: kserve-provision-location [e2e-llm-inference-service] mountPath: /mnt [e2e-llm-inference-service] readOnly: true [e2e-llm-inference-service] recursiveReadOnly: Disabled [e2e-llm-inference-service] - name: kube-api-access-p7n7g [e2e-llm-inference-service] mountPath: /var/run/secrets/kubernetes.io/serviceaccount [e2e-llm-inference-service] readOnly: true [e2e-llm-inference-service] recursiveReadOnly: Disabled [e2e-llm-inference-service] user: [e2e-llm-inference-service] linux: [e2e-llm-inference-service] uid: 1000690000 [e2e-llm-inference-service] gid: 0 [e2e-llm-inference-service] supplementalGroups: [e2e-llm-inference-service] - 0 [e2e-llm-inference-service] - 1000690000 [e2e-llm-inference-service] qosClass: Burstable [e2e-llm-inference-service] apiVersion: v1 [e2e-llm-inference-service] kind: Pod [e2e-llm-inference-service] [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1254 --- [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1255 metadata: [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-7ca60146-kserve-router-schenp5mk [e2e-llm-inference-service] generateName: llmisvc-model-fb-opt-125m-with-7ca60146-kserve-router-scheduler-778547fbc- [e2e-llm-inference-service] namespace: kserve-ci-e2e-test [e2e-llm-inference-service] uid: ce797511-9ac0-4d16-8bf9-6f34aeec1585 [e2e-llm-inference-service] resourceVersion: '28935' [e2e-llm-inference-service] generation: 1 [e2e-llm-inference-service] creationTimestamp: '2026-07-07T14:29:24Z' [e2e-llm-inference-service] labels: [e2e-llm-inference-service] app.kubernetes.io/component: llminferenceservice-router-scheduler [e2e-llm-inference-service] app.kubernetes.io/name: llmisvc-model-fb-opt-125m-with-7ca60146 [e2e-llm-inference-service] app.kubernetes.io/part-of: llminferenceservice [e2e-llm-inference-service] pod-template-hash: 778547fbc [e2e-llm-inference-service] annotations: [e2e-llm-inference-service] app.kubernetes.io/version: 0.9.0 [e2e-llm-inference-service] certificates.kserve.io/expiration-v2: 'true' [e2e-llm-inference-service] k8s.ovn.org/pod-networks: '{"default":{"ip_addresses":["10.132.0.52/23"],"mac_address":"0a:58:0a:84:00:34","gateway_ips":["10.132.0.1"],"routes":[{"dest":"10.132.0.0/14","nextHop":"10.132.0.1"},{"dest":"172.31.0.0/16","nextHop":"10.132.0.1"},{"dest":"169.254.0.5/32","nextHop":"10.132.0.1"},{"dest":"100.64.0.0/16","nextHop":"10.132.0.1"}],"ip_address":"10.132.0.52/23","gateway_ip":"10.132.0.1","role":"primary"}}' [e2e-llm-inference-service] k8s.v1.cni.cncf.io/network-status: "[{\n \"name\": \"ovn-kubernetes\",\n \ [e2e-llm-inference-service] \ \"interface\": \"eth0\",\n \"ips\": [\n \"10.132.0.52\"\n ],\n\ [e2e-llm-inference-service] \ \"mac\": \"0a:58:0a:84:00:34\",\n \"default\": true,\n \"dns\": {}\n\ [e2e-llm-inference-service] }]" [e2e-llm-inference-service] openshift.io/scc: restricted-v2 [e2e-llm-inference-service] seccomp.security.alpha.kubernetes.io/pod: runtime/default [e2e-llm-inference-service] security.openshift.io/validated-scc-subject-type: user [e2e-llm-inference-service] ownerReferences: [e2e-llm-inference-service] - apiVersion: apps/v1 [e2e-llm-inference-service] kind: ReplicaSet [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-7ca60146-kserve-router-scheduler-778547fbc [e2e-llm-inference-service] uid: e375700c-4e9f-4122-abe2-116603aa60cd [e2e-llm-inference-service] controller: true [e2e-llm-inference-service] blockOwnerDeletion: true [e2e-llm-inference-service] managedFields: [e2e-llm-inference-service] - manager: ip-10-0-133-25 [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] apiVersion: v1 [e2e-llm-inference-service] time: '2026-07-07T14:29:24Z' [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:metadata: [e2e-llm-inference-service] f:annotations: [e2e-llm-inference-service] f:k8s.ovn.org/pod-networks: {} [e2e-llm-inference-service] subresource: status [e2e-llm-inference-service] - manager: kube-controller-manager [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] apiVersion: v1 [e2e-llm-inference-service] time: '2026-07-07T14:29:24Z' [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:metadata: [e2e-llm-inference-service] f:annotations: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:app.kubernetes.io/version: {} [e2e-llm-inference-service] f:certificates.kserve.io/expiration-v2: {} [e2e-llm-inference-service] f:generateName: {} [e2e-llm-inference-service] f:labels: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:app.kubernetes.io/component: {} [e2e-llm-inference-service] f:app.kubernetes.io/name: {} [e2e-llm-inference-service] f:app.kubernetes.io/part-of: {} [e2e-llm-inference-service] f:pod-template-hash: {} [e2e-llm-inference-service] f:ownerReferences: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"uid":"e375700c-4e9f-4122-abe2-116603aa60cd"}: {} [e2e-llm-inference-service] f:spec: [e2e-llm-inference-service] f:containers: [e2e-llm-inference-service] k:{"name":"main"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:args: {} [e2e-llm-inference-service] f:command: {} [e2e-llm-inference-service] f:env: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"name":"SSL_CERT_DIR"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] f:image: {} [e2e-llm-inference-service] f:imagePullPolicy: {} [e2e-llm-inference-service] f:lifecycle: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:preStop: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:exec: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:command: {} [e2e-llm-inference-service] f:livenessProbe: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:failureThreshold: {} [e2e-llm-inference-service] f:grpc: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:port: {} [e2e-llm-inference-service] f:service: {} [e2e-llm-inference-service] f:initialDelaySeconds: {} [e2e-llm-inference-service] f:periodSeconds: {} [e2e-llm-inference-service] f:successThreshold: {} [e2e-llm-inference-service] f:timeoutSeconds: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:ports: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"containerPort":5557,"protocol":"TCP"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:containerPort: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:protocol: {} [e2e-llm-inference-service] k:{"containerPort":9002,"protocol":"TCP"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:containerPort: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:protocol: {} [e2e-llm-inference-service] k:{"containerPort":9003,"protocol":"TCP"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:containerPort: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:protocol: {} [e2e-llm-inference-service] k:{"containerPort":9090,"protocol":"TCP"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:containerPort: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:protocol: {} [e2e-llm-inference-service] f:readinessProbe: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:failureThreshold: {} [e2e-llm-inference-service] f:grpc: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:port: {} [e2e-llm-inference-service] f:service: {} [e2e-llm-inference-service] f:initialDelaySeconds: {} [e2e-llm-inference-service] f:periodSeconds: {} [e2e-llm-inference-service] f:successThreshold: {} [e2e-llm-inference-service] f:timeoutSeconds: {} [e2e-llm-inference-service] f:resources: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:requests: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:cpu: {} [e2e-llm-inference-service] f:memory: {} [e2e-llm-inference-service] f:securityContext: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:allowPrivilegeEscalation: {} [e2e-llm-inference-service] f:capabilities: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:drop: {} [e2e-llm-inference-service] f:readOnlyRootFilesystem: {} [e2e-llm-inference-service] f:runAsNonRoot: {} [e2e-llm-inference-service] f:seccompProfile: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:type: {} [e2e-llm-inference-service] f:terminationMessagePath: {} [e2e-llm-inference-service] f:terminationMessagePolicy: {} [e2e-llm-inference-service] f:volumeMounts: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"mountPath":"/tmp/tokenizer"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:mountPath: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] k:{"mountPath":"/var/run/kserve/tls"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:mountPath: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:readOnly: {} [e2e-llm-inference-service] f:dnsPolicy: {} [e2e-llm-inference-service] f:enableServiceLinks: {} [e2e-llm-inference-service] f:initContainers: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"name":"storage-initializer"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:args: {} [e2e-llm-inference-service] f:env: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"name":"AWS_ACCESS_KEY_ID"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:valueFrom: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:secretKeyRef: {} [e2e-llm-inference-service] k:{"name":"AWS_CA_BUNDLE"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"AWS_CA_BUNDLE_CONFIGMAP"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"AWS_ENDPOINT_URL"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"AWS_SECRET_ACCESS_KEY"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:valueFrom: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:secretKeyRef: {} [e2e-llm-inference-service] k:{"name":"HF_HUB_ENABLE_HF_TRANSFER"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"HF_XET_HIGH_PERFORMANCE"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"HF_XET_NUM_CONCURRENT_RANGE_GETS"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"S3_ENDPOINT"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"S3_USE_HTTPS"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"S3_VERIFY_SSL"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"STORAGE_ALLOW_PATTERNS"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] f:image: {} [e2e-llm-inference-service] f:imagePullPolicy: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:resources: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:limits: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:cpu: {} [e2e-llm-inference-service] f:memory: {} [e2e-llm-inference-service] f:requests: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:cpu: {} [e2e-llm-inference-service] f:memory: {} [e2e-llm-inference-service] f:terminationMessagePath: {} [e2e-llm-inference-service] f:terminationMessagePolicy: {} [e2e-llm-inference-service] f:volumeMounts: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"mountPath":"/mnt/models"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:mountPath: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:restartPolicy: {} [e2e-llm-inference-service] f:schedulerName: {} [e2e-llm-inference-service] f:securityContext: {} [e2e-llm-inference-service] f:serviceAccount: {} [e2e-llm-inference-service] f:serviceAccountName: {} [e2e-llm-inference-service] f:terminationGracePeriodSeconds: {} [e2e-llm-inference-service] f:volumes: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"name":"kserve-provision-location"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:emptyDir: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] k:{"name":"tls-certs"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:secret: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:defaultMode: {} [e2e-llm-inference-service] f:secretName: {} [e2e-llm-inference-service] k:{"name":"tokenizer-cache"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:emptyDir: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] k:{"name":"tokenizer-tmp"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:emptyDir: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] k:{"name":"tokenizer-uds"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:emptyDir: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] - manager: multus-daemon [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] apiVersion: v1 [e2e-llm-inference-service] time: '2026-07-07T14:29:25Z' [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:metadata: [e2e-llm-inference-service] f:annotations: [e2e-llm-inference-service] f:k8s.v1.cni.cncf.io/network-status: {} [e2e-llm-inference-service] subresource: status [e2e-llm-inference-service] - manager: kubelet [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] apiVersion: v1 [e2e-llm-inference-service] time: '2026-07-07T14:29:59Z' [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:status: [e2e-llm-inference-service] f:conditions: [e2e-llm-inference-service] k:{"type":"ContainersReady"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:lastProbeTime: {} [e2e-llm-inference-service] f:lastTransitionTime: {} [e2e-llm-inference-service] f:observedGeneration: {} [e2e-llm-inference-service] f:status: {} [e2e-llm-inference-service] f:type: {} [e2e-llm-inference-service] k:{"type":"Initialized"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:lastProbeTime: {} [e2e-llm-inference-service] f:lastTransitionTime: {} [e2e-llm-inference-service] f:observedGeneration: {} [e2e-llm-inference-service] f:status: {} [e2e-llm-inference-service] f:type: {} [e2e-llm-inference-service] k:{"type":"PodReadyToStartContainers"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:lastProbeTime: {} [e2e-llm-inference-service] f:lastTransitionTime: {} [e2e-llm-inference-service] f:observedGeneration: {} [e2e-llm-inference-service] f:status: {} [e2e-llm-inference-service] f:type: {} [e2e-llm-inference-service] k:{"type":"PodScheduled"}: [e2e-llm-inference-service] f:observedGeneration: {} [e2e-llm-inference-service] k:{"type":"Ready"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:lastProbeTime: {} [e2e-llm-inference-service] f:lastTransitionTime: {} [e2e-llm-inference-service] f:observedGeneration: {} [e2e-llm-inference-service] f:status: {} [e2e-llm-inference-service] f:type: {} [e2e-llm-inference-service] f:containerStatuses: {} [e2e-llm-inference-service] f:hostIP: {} [e2e-llm-inference-service] f:hostIPs: {} [e2e-llm-inference-service] f:initContainerStatuses: {} [e2e-llm-inference-service] f:observedGeneration: {} [e2e-llm-inference-service] f:phase: {} [e2e-llm-inference-service] f:podIP: {} [e2e-llm-inference-service] f:podIPs: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"ip":"10.132.0.52"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:ip: {} [e2e-llm-inference-service] f:startTime: {} [e2e-llm-inference-service] subresource: status [e2e-llm-inference-service] spec: [e2e-llm-inference-service] volumes: [e2e-llm-inference-service] - name: tls-certs [e2e-llm-inference-service] secret: [e2e-llm-inference-service] secretName: llmisv3e414c2ba058a022dfd694dbcbac5b51-kserve-self-signed-certs [e2e-llm-inference-service] defaultMode: 420 [e2e-llm-inference-service] - name: tokenizer-uds [e2e-llm-inference-service] emptyDir: {} [e2e-llm-inference-service] - name: tokenizer-tmp [e2e-llm-inference-service] emptyDir: {} [e2e-llm-inference-service] - name: tokenizer-cache [e2e-llm-inference-service] emptyDir: {} [e2e-llm-inference-service] - name: kserve-provision-location [e2e-llm-inference-service] emptyDir: {} [e2e-llm-inference-service] - name: kube-api-access-k2cvv [e2e-llm-inference-service] projected: [e2e-llm-inference-service] sources: [e2e-llm-inference-service] - serviceAccountToken: [e2e-llm-inference-service] expirationSeconds: 3607 [e2e-llm-inference-service] path: token [e2e-llm-inference-service] - configMap: [e2e-llm-inference-service] name: kube-root-ca.crt [e2e-llm-inference-service] items: [e2e-llm-inference-service] - key: ca.crt [e2e-llm-inference-service] path: ca.crt [e2e-llm-inference-service] - downwardAPI: [e2e-llm-inference-service] items: [e2e-llm-inference-service] - path: namespace [e2e-llm-inference-service] fieldRef: [e2e-llm-inference-service] apiVersion: v1 [e2e-llm-inference-service] fieldPath: metadata.namespace [e2e-llm-inference-service] - configMap: [e2e-llm-inference-service] name: openshift-service-ca.crt [e2e-llm-inference-service] items: [e2e-llm-inference-service] - key: service-ca.crt [e2e-llm-inference-service] path: service-ca.crt [e2e-llm-inference-service] defaultMode: 420 [e2e-llm-inference-service] initContainers: [e2e-llm-inference-service] - name: storage-initializer [e2e-llm-inference-service] image: quay.io/opendatahub/kserve-storage-initializer@sha256:13ad765c0f323200e1db7b45c99661287e65c5158d6b54b16139829636d20be4 [e2e-llm-inference-service] args: [e2e-llm-inference-service] - hf://facebook/opt-125m [e2e-llm-inference-service] - /mnt/models [e2e-llm-inference-service] env: [e2e-llm-inference-service] - name: AWS_ACCESS_KEY_ID [e2e-llm-inference-service] valueFrom: [e2e-llm-inference-service] secretKeyRef: [e2e-llm-inference-service] name: seaweedfs-s3-creds [e2e-llm-inference-service] key: AWS_ACCESS_KEY_ID [e2e-llm-inference-service] - name: AWS_SECRET_ACCESS_KEY [e2e-llm-inference-service] valueFrom: [e2e-llm-inference-service] secretKeyRef: [e2e-llm-inference-service] name: seaweedfs-s3-creds [e2e-llm-inference-service] key: AWS_SECRET_ACCESS_KEY [e2e-llm-inference-service] - name: S3_USE_HTTPS [e2e-llm-inference-service] value: '0' [e2e-llm-inference-service] - name: S3_ENDPOINT [e2e-llm-inference-service] value: s3-service.kserve:8333 [e2e-llm-inference-service] - name: AWS_ENDPOINT_URL [e2e-llm-inference-service] value: http://s3-service.kserve:8333 [e2e-llm-inference-service] - name: S3_VERIFY_SSL [e2e-llm-inference-service] value: '0' [e2e-llm-inference-service] - name: AWS_CA_BUNDLE [e2e-llm-inference-service] value: /etc/ssl/custom-certs/cabundle.crt [e2e-llm-inference-service] - name: AWS_CA_BUNDLE_CONFIGMAP [e2e-llm-inference-service] value: odh-kserve-custom-ca-bundle [e2e-llm-inference-service] - name: HF_HUB_ENABLE_HF_TRANSFER [e2e-llm-inference-service] value: '1' [e2e-llm-inference-service] - name: HF_XET_HIGH_PERFORMANCE [e2e-llm-inference-service] value: '1' [e2e-llm-inference-service] - name: HF_XET_NUM_CONCURRENT_RANGE_GETS [e2e-llm-inference-service] value: '8' [e2e-llm-inference-service] - name: STORAGE_ALLOW_PATTERNS [e2e-llm-inference-service] value: '["tokenizer.json", "tokenizer_config.json", "special_tokens_map.json", [e2e-llm-inference-service] "vocab.json", "merges.txt", "config.json", "generation_config.json"]' [e2e-llm-inference-service] resources: [e2e-llm-inference-service] limits: [e2e-llm-inference-service] cpu: '1' [e2e-llm-inference-service] memory: 24Gi [e2e-llm-inference-service] requests: [e2e-llm-inference-service] cpu: 100m [e2e-llm-inference-service] memory: 100Mi [e2e-llm-inference-service] volumeMounts: [e2e-llm-inference-service] - name: kserve-provision-location [e2e-llm-inference-service] mountPath: /mnt/models [e2e-llm-inference-service] - name: kube-api-access-k2cvv [e2e-llm-inference-service] readOnly: true [e2e-llm-inference-service] mountPath: /var/run/secrets/kubernetes.io/serviceaccount [e2e-llm-inference-service] terminationMessagePath: /dev/termination-log [e2e-llm-inference-service] terminationMessagePolicy: FallbackToLogsOnError [e2e-llm-inference-service] imagePullPolicy: IfNotPresent [e2e-llm-inference-service] securityContext: [e2e-llm-inference-service] capabilities: [e2e-llm-inference-service] drop: [e2e-llm-inference-service] - ALL [e2e-llm-inference-service] runAsUser: 1000690000 [e2e-llm-inference-service] runAsNonRoot: true [e2e-llm-inference-service] allowPrivilegeEscalation: false [e2e-llm-inference-service] containers: [e2e-llm-inference-service] - name: main [e2e-llm-inference-service] image: ghcr.io/llm-d/llm-d-router-endpoint-picker:v0.9.0-rc.2 [e2e-llm-inference-service] command: [e2e-llm-inference-service] - /app/epp [e2e-llm-inference-service] - --pool-name [e2e-llm-inference-service] - llmisvc-model-fb-opt-125m-with-7ca60146-inference-pool [e2e-llm-inference-service] - --pool-namespace [e2e-llm-inference-service] - kserve-ci-e2e-test [e2e-llm-inference-service] - --zap-encoder [e2e-llm-inference-service] - json [e2e-llm-inference-service] - --grpc-port [e2e-llm-inference-service] - '9002' [e2e-llm-inference-service] - --grpc-health-port [e2e-llm-inference-service] - '9003' [e2e-llm-inference-service] - --enable-cert-reload=true [e2e-llm-inference-service] - --secure-serving=true [e2e-llm-inference-service] - --model-server-metrics-scheme=https [e2e-llm-inference-service] - --cert-path=/var/run/kserve/tls [e2e-llm-inference-service] args: [e2e-llm-inference-service] - --config-text [e2e-llm-inference-service] - "apiVersion: inference.networking.x-k8s.io/v1alpha1\nkind: EndpointPickerConfig\n\ [e2e-llm-inference-service] plugins:\n- type: single-profile-handler\n- type: queue-scorer\n- type: prefix-cache-scorer\n\ [e2e-llm-inference-service] - type: max-score-picker\n- type: lora-affinity-scorer\nschedulingProfiles:\n\ [e2e-llm-inference-service] - name: default\n plugins:\n - pluginRef: lora-affinity-scorer\n weight:\ [e2e-llm-inference-service] \ 4\n - pluginRef: queue-scorer\n weight: 2\n - pluginRef: prefix-cache-scorer\n\ [e2e-llm-inference-service] \ weight: 3\n - pluginRef: max-score-picker\n" [e2e-llm-inference-service] ports: [e2e-llm-inference-service] - name: grpc [e2e-llm-inference-service] containerPort: 9002 [e2e-llm-inference-service] protocol: TCP [e2e-llm-inference-service] - name: grpc-health [e2e-llm-inference-service] containerPort: 9003 [e2e-llm-inference-service] protocol: TCP [e2e-llm-inference-service] - name: metrics [e2e-llm-inference-service] containerPort: 9090 [e2e-llm-inference-service] protocol: TCP [e2e-llm-inference-service] - name: zmq [e2e-llm-inference-service] containerPort: 5557 [e2e-llm-inference-service] protocol: TCP [e2e-llm-inference-service] env: [e2e-llm-inference-service] - name: SSL_CERT_DIR [e2e-llm-inference-service] value: /var/run/kserve/tls:/var/run/secrets/kubernetes.io/serviceaccount:/etc/pki/tls/certs [e2e-llm-inference-service] resources: [e2e-llm-inference-service] requests: [e2e-llm-inference-service] cpu: 256m [e2e-llm-inference-service] memory: 500Mi [e2e-llm-inference-service] volumeMounts: [e2e-llm-inference-service] - name: tls-certs [e2e-llm-inference-service] readOnly: true [e2e-llm-inference-service] mountPath: /var/run/kserve/tls [e2e-llm-inference-service] - name: tokenizer-uds [e2e-llm-inference-service] mountPath: /tmp/tokenizer [e2e-llm-inference-service] - name: kube-api-access-k2cvv [e2e-llm-inference-service] readOnly: true [e2e-llm-inference-service] mountPath: /var/run/secrets/kubernetes.io/serviceaccount [e2e-llm-inference-service] livenessProbe: [e2e-llm-inference-service] grpc: [e2e-llm-inference-service] port: 9003 [e2e-llm-inference-service] service: liveness [e2e-llm-inference-service] initialDelaySeconds: 5 [e2e-llm-inference-service] timeoutSeconds: 1 [e2e-llm-inference-service] periodSeconds: 10 [e2e-llm-inference-service] successThreshold: 1 [e2e-llm-inference-service] failureThreshold: 3 [e2e-llm-inference-service] readinessProbe: [e2e-llm-inference-service] grpc: [e2e-llm-inference-service] port: 9003 [e2e-llm-inference-service] service: readiness [e2e-llm-inference-service] initialDelaySeconds: 30 [e2e-llm-inference-service] timeoutSeconds: 1 [e2e-llm-inference-service] periodSeconds: 10 [e2e-llm-inference-service] successThreshold: 1 [e2e-llm-inference-service] failureThreshold: 3 [e2e-llm-inference-service] lifecycle: [e2e-llm-inference-service] preStop: [e2e-llm-inference-service] exec: [e2e-llm-inference-service] command: [e2e-llm-inference-service] - /bin/sleep [e2e-llm-inference-service] - '15' [e2e-llm-inference-service] terminationMessagePath: /dev/termination-log [e2e-llm-inference-service] terminationMessagePolicy: FallbackToLogsOnError [e2e-llm-inference-service] imagePullPolicy: IfNotPresent [e2e-llm-inference-service] securityContext: [e2e-llm-inference-service] capabilities: [e2e-llm-inference-service] drop: [e2e-llm-inference-service] - ALL [e2e-llm-inference-service] runAsUser: 1000690000 [e2e-llm-inference-service] runAsNonRoot: true [e2e-llm-inference-service] readOnlyRootFilesystem: true [e2e-llm-inference-service] allowPrivilegeEscalation: false [e2e-llm-inference-service] seccompProfile: [e2e-llm-inference-service] type: RuntimeDefault [e2e-llm-inference-service] restartPolicy: Always [e2e-llm-inference-service] terminationGracePeriodSeconds: 60 [e2e-llm-inference-service] dnsPolicy: ClusterFirst [e2e-llm-inference-service] serviceAccountName: llmisvc-model-fb-opt-125m-with-7ca60146-epp-sa [e2e-llm-inference-service] serviceAccount: llmisvc-model-fb-opt-125m-with-7ca60146-epp-sa [e2e-llm-inference-service] nodeName: ip-10-0-133-25.ec2.internal [e2e-llm-inference-service] securityContext: [e2e-llm-inference-service] seLinuxOptions: [e2e-llm-inference-service] level: s0:c26,c20 [e2e-llm-inference-service] fsGroup: 1000690000 [e2e-llm-inference-service] seccompProfile: [e2e-llm-inference-service] type: RuntimeDefault [e2e-llm-inference-service] imagePullSecrets: [e2e-llm-inference-service] - name: default-dockercfg-tbbgr [e2e-llm-inference-service] - name: llmisvc-model-fb-opt-125m-with-7ca60146-epp-sa-dockercfg-hfb5n [e2e-llm-inference-service] schedulerName: default-scheduler [e2e-llm-inference-service] tolerations: [e2e-llm-inference-service] - key: node.kubernetes.io/not-ready [e2e-llm-inference-service] operator: Exists [e2e-llm-inference-service] effect: NoExecute [e2e-llm-inference-service] tolerationSeconds: 300 [e2e-llm-inference-service] - key: node.kubernetes.io/unreachable [e2e-llm-inference-service] operator: Exists [e2e-llm-inference-service] effect: NoExecute [e2e-llm-inference-service] tolerationSeconds: 300 [e2e-llm-inference-service] - key: node.kubernetes.io/memory-pressure [e2e-llm-inference-service] operator: Exists [e2e-llm-inference-service] effect: NoSchedule [e2e-llm-inference-service] priority: 0 [e2e-llm-inference-service] enableServiceLinks: true [e2e-llm-inference-service] preemptionPolicy: PreemptLowerPriority [e2e-llm-inference-service] status: [e2e-llm-inference-service] observedGeneration: 1 [e2e-llm-inference-service] phase: Running [e2e-llm-inference-service] conditions: [e2e-llm-inference-service] - type: PodReadyToStartContainers [e2e-llm-inference-service] observedGeneration: 1 [e2e-llm-inference-service] status: 'True' [e2e-llm-inference-service] lastProbeTime: null [e2e-llm-inference-service] lastTransitionTime: '2026-07-07T14:29:25Z' [e2e-llm-inference-service] - type: Initialized [e2e-llm-inference-service] observedGeneration: 1 [e2e-llm-inference-service] status: 'True' [e2e-llm-inference-service] lastProbeTime: null [e2e-llm-inference-service] lastTransitionTime: '2026-07-07T14:29:26Z' [e2e-llm-inference-service] - type: Ready [e2e-llm-inference-service] observedGeneration: 1 [e2e-llm-inference-service] status: 'True' [e2e-llm-inference-service] lastProbeTime: null [e2e-llm-inference-service] lastTransitionTime: '2026-07-07T14:29:58Z' [e2e-llm-inference-service] - type: ContainersReady [e2e-llm-inference-service] observedGeneration: 1 [e2e-llm-inference-service] status: 'True' [e2e-llm-inference-service] lastProbeTime: null [e2e-llm-inference-service] lastTransitionTime: '2026-07-07T14:29:58Z' [e2e-llm-inference-service] - type: PodScheduled [e2e-llm-inference-service] observedGeneration: 1 [e2e-llm-inference-service] status: 'True' [e2e-llm-inference-service] lastProbeTime: null [e2e-llm-inference-service] lastTransitionTime: '2026-07-07T14:29:24Z' [e2e-llm-inference-service] hostIP: 10.0.133.25 [e2e-llm-inference-service] hostIPs: [e2e-llm-inference-service] - ip: 10.0.133.25 [e2e-llm-inference-service] podIP: 10.132.0.52 [e2e-llm-inference-service] podIPs: [e2e-llm-inference-service] - ip: 10.132.0.52 [e2e-llm-inference-service] startTime: '2026-07-07T14:29:24Z' [e2e-llm-inference-service] initContainerStatuses: [e2e-llm-inference-service] - name: storage-initializer [e2e-llm-inference-service] state: [e2e-llm-inference-service] terminated: [e2e-llm-inference-service] exitCode: 0 [e2e-llm-inference-service] reason: Completed [e2e-llm-inference-service] startedAt: '2026-07-07T14:29:25Z' [e2e-llm-inference-service] finishedAt: '2026-07-07T14:29:26Z' [e2e-llm-inference-service] containerID: cri-o://70df7e4193c1735925d1fec0d070580c17b863ce3d5f780acb6d71bd663ffb02 [e2e-llm-inference-service] lastState: {} [e2e-llm-inference-service] ready: true [e2e-llm-inference-service] restartCount: 0 [e2e-llm-inference-service] image: quay.io/opendatahub/kserve-storage-initializer@sha256:13ad765c0f323200e1db7b45c99661287e65c5158d6b54b16139829636d20be4 [e2e-llm-inference-service] imageID: quay.io/opendatahub/kserve-storage-initializer@sha256:13ad765c0f323200e1db7b45c99661287e65c5158d6b54b16139829636d20be4 [e2e-llm-inference-service] containerID: cri-o://70df7e4193c1735925d1fec0d070580c17b863ce3d5f780acb6d71bd663ffb02 [e2e-llm-inference-service] started: false [e2e-llm-inference-service] allocatedResources: [e2e-llm-inference-service] cpu: 100m [e2e-llm-inference-service] memory: 100Mi [e2e-llm-inference-service] resources: [e2e-llm-inference-service] limits: [e2e-llm-inference-service] cpu: '1' [e2e-llm-inference-service] memory: 24Gi [e2e-llm-inference-service] requests: [e2e-llm-inference-service] cpu: 100m [e2e-llm-inference-service] memory: 100Mi [e2e-llm-inference-service] volumeMounts: [e2e-llm-inference-service] - name: kserve-provision-location [e2e-llm-inference-service] mountPath: /mnt/models [e2e-llm-inference-service] - name: kube-api-access-k2cvv [e2e-llm-inference-service] mountPath: /var/run/secrets/kubernetes.io/serviceaccount [e2e-llm-inference-service] readOnly: true [e2e-llm-inference-service] recursiveReadOnly: Disabled [e2e-llm-inference-service] user: [e2e-llm-inference-service] linux: [e2e-llm-inference-service] uid: 1000690000 [e2e-llm-inference-service] gid: 0 [e2e-llm-inference-service] supplementalGroups: [e2e-llm-inference-service] - 0 [e2e-llm-inference-service] - 1000690000 [e2e-llm-inference-service] containerStatuses: [e2e-llm-inference-service] - name: main [e2e-llm-inference-service] state: [e2e-llm-inference-service] running: [e2e-llm-inference-service] startedAt: '2026-07-07T14:29:27Z' [e2e-llm-inference-service] lastState: {} [e2e-llm-inference-service] ready: true [e2e-llm-inference-service] restartCount: 0 [e2e-llm-inference-service] image: ghcr.io/llm-d/llm-d-router-endpoint-picker:v0.9.0-rc.2 [e2e-llm-inference-service] imageID: ghcr.io/llm-d/llm-d-router-endpoint-picker@sha256:06b6c75d77afd0e07053402752a9736c2dfbc12a306d0d37d963aac4c1d4e6a6 [e2e-llm-inference-service] containerID: cri-o://c6e587be09e96b381ef41aee8da704fc79917f3eda718ca8aedab288c6790d21 [e2e-llm-inference-service] started: true [e2e-llm-inference-service] allocatedResources: [e2e-llm-inference-service] cpu: 256m [e2e-llm-inference-service] memory: 500Mi [e2e-llm-inference-service] resources: [e2e-llm-inference-service] requests: [e2e-llm-inference-service] cpu: 256m [e2e-llm-inference-service] memory: 500Mi [e2e-llm-inference-service] volumeMounts: [e2e-llm-inference-service] - name: tls-certs [e2e-llm-inference-service] mountPath: /var/run/kserve/tls [e2e-llm-inference-service] readOnly: true [e2e-llm-inference-service] recursiveReadOnly: Disabled [e2e-llm-inference-service] - name: tokenizer-uds [e2e-llm-inference-service] mountPath: /tmp/tokenizer [e2e-llm-inference-service] - name: kube-api-access-k2cvv [e2e-llm-inference-service] mountPath: /var/run/secrets/kubernetes.io/serviceaccount [e2e-llm-inference-service] readOnly: true [e2e-llm-inference-service] recursiveReadOnly: Disabled [e2e-llm-inference-service] user: [e2e-llm-inference-service] linux: [e2e-llm-inference-service] uid: 1000690000 [e2e-llm-inference-service] gid: 0 [e2e-llm-inference-service] supplementalGroups: [e2e-llm-inference-service] - 0 [e2e-llm-inference-service] - 1000690000 [e2e-llm-inference-service] qosClass: Burstable [e2e-llm-inference-service] apiVersion: v1 [e2e-llm-inference-service] kind: Pod [e2e-llm-inference-service] [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1254 --- [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1255 metadata: [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-7ca60146-epp-sa [e2e-llm-inference-service] namespace: kserve-ci-e2e-test [e2e-llm-inference-service] uid: a6b15943-23c8-476a-ad32-55c3be1d812d [e2e-llm-inference-service] resourceVersion: '28229' [e2e-llm-inference-service] creationTimestamp: '2026-07-07T14:29:24Z' [e2e-llm-inference-service] labels: [e2e-llm-inference-service] app.kubernetes.io/component: llminferenceservice-router-scheduler [e2e-llm-inference-service] app.kubernetes.io/name: llmisvc-model-fb-opt-125m-with-7ca60146 [e2e-llm-inference-service] app.kubernetes.io/part-of: llminferenceservice [e2e-llm-inference-service] annotations: [e2e-llm-inference-service] openshift.io/internal-registry-pull-secret-ref: llmisvc-model-fb-opt-125m-with-7ca60146-epp-sa-dockercfg-hfb5n [e2e-llm-inference-service] ownerReferences: [e2e-llm-inference-service] - apiVersion: serving.kserve.io/v1alpha2 [e2e-llm-inference-service] kind: LLMInferenceService [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-7ca60146 [e2e-llm-inference-service] uid: e15cfc59-e4a1-4876-b43b-85ce3b1b239b [e2e-llm-inference-service] controller: true [e2e-llm-inference-service] blockOwnerDeletion: true [e2e-llm-inference-service] managedFields: [e2e-llm-inference-service] - manager: openshift.io/image-registry-pull-secrets_service-account-controller [e2e-llm-inference-service] operation: Apply [e2e-llm-inference-service] apiVersion: v1 [e2e-llm-inference-service] time: '2026-07-07T14:29:24Z' [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:imagePullSecrets: {} [e2e-llm-inference-service] f:metadata: [e2e-llm-inference-service] f:annotations: [e2e-llm-inference-service] f:openshift.io/internal-registry-pull-secret-ref: {} [e2e-llm-inference-service] f:secrets: [e2e-llm-inference-service] k:{"name":"llmisvc-model-fb-opt-125m-with-7ca60146-epp-sa-dockercfg-hfb5n"}: {} [e2e-llm-inference-service] - manager: manager [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] apiVersion: v1 [e2e-llm-inference-service] time: '2026-07-07T14:29:24Z' [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:metadata: [e2e-llm-inference-service] f:labels: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:app.kubernetes.io/component: {} [e2e-llm-inference-service] f:app.kubernetes.io/name: {} [e2e-llm-inference-service] f:app.kubernetes.io/part-of: {} [e2e-llm-inference-service] f:ownerReferences: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"uid":"e15cfc59-e4a1-4876-b43b-85ce3b1b239b"}: {} [e2e-llm-inference-service] f:secrets: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"name":"default-dockercfg-tbbgr"}: {} [e2e-llm-inference-service] k:{"name":"seaweedfs-s3-creds"}: {} [e2e-llm-inference-service] secrets: [e2e-llm-inference-service] - name: default-dockercfg-tbbgr [e2e-llm-inference-service] - name: seaweedfs-s3-creds [e2e-llm-inference-service] - name: llmisvc-model-fb-opt-125m-with-7ca60146-epp-sa-dockercfg-hfb5n [e2e-llm-inference-service] imagePullSecrets: [e2e-llm-inference-service] - name: default-dockercfg-tbbgr [e2e-llm-inference-service] - name: llmisvc-model-fb-opt-125m-with-7ca60146-epp-sa-dockercfg-hfb5n [e2e-llm-inference-service] apiVersion: v1 [e2e-llm-inference-service] kind: ServiceAccount [e2e-llm-inference-service] [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1254 --- [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1255 metadata: [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-7ca60146-epp-service [e2e-llm-inference-service] namespace: kserve-ci-e2e-test [e2e-llm-inference-service] uid: f144c5eb-55c5-451e-b7ed-1e7b6e6d4ebf [e2e-llm-inference-service] resourceVersion: '28247' [e2e-llm-inference-service] creationTimestamp: '2026-07-07T14:29:24Z' [e2e-llm-inference-service] labels: [e2e-llm-inference-service] app.kubernetes.io/component: llminferenceservice-router-scheduler [e2e-llm-inference-service] app.kubernetes.io/name: llmisvc-model-fb-opt-125m-with-7ca60146 [e2e-llm-inference-service] app.kubernetes.io/part-of: llminferenceservice [e2e-llm-inference-service] ownerReferences: [e2e-llm-inference-service] - apiVersion: serving.kserve.io/v1alpha2 [e2e-llm-inference-service] kind: LLMInferenceService [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-7ca60146 [e2e-llm-inference-service] uid: e15cfc59-e4a1-4876-b43b-85ce3b1b239b [e2e-llm-inference-service] controller: true [e2e-llm-inference-service] blockOwnerDeletion: true [e2e-llm-inference-service] managedFields: [e2e-llm-inference-service] - manager: manager [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] apiVersion: v1 [e2e-llm-inference-service] time: '2026-07-07T14:29:24Z' [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:metadata: [e2e-llm-inference-service] f:labels: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:app.kubernetes.io/component: {} [e2e-llm-inference-service] f:app.kubernetes.io/name: {} [e2e-llm-inference-service] f:app.kubernetes.io/part-of: {} [e2e-llm-inference-service] f:ownerReferences: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"uid":"e15cfc59-e4a1-4876-b43b-85ce3b1b239b"}: {} [e2e-llm-inference-service] f:spec: [e2e-llm-inference-service] f:internalTrafficPolicy: {} [e2e-llm-inference-service] f:ports: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"port":5557,"protocol":"TCP"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:port: {} [e2e-llm-inference-service] f:protocol: {} [e2e-llm-inference-service] f:targetPort: {} [e2e-llm-inference-service] k:{"port":9002,"protocol":"TCP"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:port: {} [e2e-llm-inference-service] f:protocol: {} [e2e-llm-inference-service] f:targetPort: {} [e2e-llm-inference-service] k:{"port":9003,"protocol":"TCP"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:port: {} [e2e-llm-inference-service] f:protocol: {} [e2e-llm-inference-service] f:targetPort: {} [e2e-llm-inference-service] k:{"port":9090,"protocol":"TCP"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:port: {} [e2e-llm-inference-service] f:protocol: {} [e2e-llm-inference-service] f:targetPort: {} [e2e-llm-inference-service] f:selector: {} [e2e-llm-inference-service] f:sessionAffinity: {} [e2e-llm-inference-service] f:type: {} [e2e-llm-inference-service] spec: [e2e-llm-inference-service] ports: [e2e-llm-inference-service] - name: grpc [e2e-llm-inference-service] protocol: TCP [e2e-llm-inference-service] port: 9002 [e2e-llm-inference-service] targetPort: grpc [e2e-llm-inference-service] - name: grpc-health [e2e-llm-inference-service] protocol: TCP [e2e-llm-inference-service] port: 9003 [e2e-llm-inference-service] targetPort: grpc-health [e2e-llm-inference-service] - name: metrics [e2e-llm-inference-service] protocol: TCP [e2e-llm-inference-service] port: 9090 [e2e-llm-inference-service] targetPort: metrics [e2e-llm-inference-service] - name: zmq [e2e-llm-inference-service] protocol: TCP [e2e-llm-inference-service] port: 5557 [e2e-llm-inference-service] targetPort: zmq [e2e-llm-inference-service] selector: [e2e-llm-inference-service] app.kubernetes.io/component: llminferenceservice-router-scheduler [e2e-llm-inference-service] app.kubernetes.io/name: llmisvc-model-fb-opt-125m-with-7ca60146 [e2e-llm-inference-service] app.kubernetes.io/part-of: llminferenceservice [e2e-llm-inference-service] clusterIP: 172.31.213.19 [e2e-llm-inference-service] clusterIPs: [e2e-llm-inference-service] - 172.31.213.19 [e2e-llm-inference-service] type: ClusterIP [e2e-llm-inference-service] sessionAffinity: None [e2e-llm-inference-service] ipFamilies: [e2e-llm-inference-service] - IPv4 [e2e-llm-inference-service] ipFamilyPolicy: SingleStack [e2e-llm-inference-service] internalTrafficPolicy: Cluster [e2e-llm-inference-service] status: [e2e-llm-inference-service] loadBalancer: {} [e2e-llm-inference-service] apiVersion: v1 [e2e-llm-inference-service] kind: Service [e2e-llm-inference-service] [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1254 --- [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1255 metadata: [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-7ca60146-kserve-workload-svc [e2e-llm-inference-service] namespace: kserve-ci-e2e-test [e2e-llm-inference-service] uid: 8fe30c1f-242c-4ca3-aad7-a4b6836297ea [e2e-llm-inference-service] resourceVersion: '28215' [e2e-llm-inference-service] creationTimestamp: '2026-07-07T14:29:24Z' [e2e-llm-inference-service] labels: [e2e-llm-inference-service] app.kubernetes.io/component: llminferenceservice-workload [e2e-llm-inference-service] app.kubernetes.io/name: llmisvc-model-fb-opt-125m-with-7ca60146 [e2e-llm-inference-service] app.kubernetes.io/part-of: llminferenceservice [e2e-llm-inference-service] ownerReferences: [e2e-llm-inference-service] - apiVersion: serving.kserve.io/v1alpha2 [e2e-llm-inference-service] kind: LLMInferenceService [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-7ca60146 [e2e-llm-inference-service] uid: e15cfc59-e4a1-4876-b43b-85ce3b1b239b [e2e-llm-inference-service] controller: true [e2e-llm-inference-service] blockOwnerDeletion: true [e2e-llm-inference-service] managedFields: [e2e-llm-inference-service] - manager: manager [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] apiVersion: v1 [e2e-llm-inference-service] time: '2026-07-07T14:29:24Z' [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:metadata: [e2e-llm-inference-service] f:labels: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:app.kubernetes.io/component: {} [e2e-llm-inference-service] f:app.kubernetes.io/name: {} [e2e-llm-inference-service] f:app.kubernetes.io/part-of: {} [e2e-llm-inference-service] f:ownerReferences: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"uid":"e15cfc59-e4a1-4876-b43b-85ce3b1b239b"}: {} [e2e-llm-inference-service] f:spec: [e2e-llm-inference-service] f:internalTrafficPolicy: {} [e2e-llm-inference-service] f:ports: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"port":8000,"protocol":"TCP"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:appProtocol: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:port: {} [e2e-llm-inference-service] f:protocol: {} [e2e-llm-inference-service] f:targetPort: {} [e2e-llm-inference-service] f:selector: {} [e2e-llm-inference-service] f:sessionAffinity: {} [e2e-llm-inference-service] f:type: {} [e2e-llm-inference-service] spec: [e2e-llm-inference-service] ports: [e2e-llm-inference-service] - name: https [e2e-llm-inference-service] protocol: TCP [e2e-llm-inference-service] appProtocol: https [e2e-llm-inference-service] port: 8000 [e2e-llm-inference-service] targetPort: 8000 [e2e-llm-inference-service] selector: [e2e-llm-inference-service] app.kubernetes.io/name: llmisvc-model-fb-opt-125m-with-7ca60146 [e2e-llm-inference-service] app.kubernetes.io/part-of: llminferenceservice [e2e-llm-inference-service] kserve.io/component: workload [e2e-llm-inference-service] clusterIP: 172.31.156.182 [e2e-llm-inference-service] clusterIPs: [e2e-llm-inference-service] - 172.31.156.182 [e2e-llm-inference-service] type: ClusterIP [e2e-llm-inference-service] sessionAffinity: None [e2e-llm-inference-service] ipFamilies: [e2e-llm-inference-service] - IPv4 [e2e-llm-inference-service] ipFamilyPolicy: SingleStack [e2e-llm-inference-service] internalTrafficPolicy: Cluster [e2e-llm-inference-service] status: [e2e-llm-inference-service] loadBalancer: {} [e2e-llm-inference-service] apiVersion: v1 [e2e-llm-inference-service] kind: Service [e2e-llm-inference-service] [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1254 --- [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1255 metadata: [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-7ca60146-kserve [e2e-llm-inference-service] namespace: kserve-ci-e2e-test [e2e-llm-inference-service] uid: 0e103aa1-bb8a-4024-977f-462552a80dcf [e2e-llm-inference-service] resourceVersion: '29818' [e2e-llm-inference-service] generation: 1 [e2e-llm-inference-service] creationTimestamp: '2026-07-07T14:29:24Z' [e2e-llm-inference-service] labels: [e2e-llm-inference-service] app.kubernetes.io/component: llminferenceservice-workload [e2e-llm-inference-service] app.kubernetes.io/name: llmisvc-model-fb-opt-125m-with-7ca60146 [e2e-llm-inference-service] app.kubernetes.io/part-of: llminferenceservice [e2e-llm-inference-service] kserve.io/component: workload [e2e-llm-inference-service] llm-d.ai/role: both [e2e-llm-inference-service] annotations: [e2e-llm-inference-service] deployment.kubernetes.io/revision: '1' [e2e-llm-inference-service] ownerReferences: [e2e-llm-inference-service] - apiVersion: serving.kserve.io/v1alpha2 [e2e-llm-inference-service] kind: LLMInferenceService [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-7ca60146 [e2e-llm-inference-service] uid: e15cfc59-e4a1-4876-b43b-85ce3b1b239b [e2e-llm-inference-service] controller: true [e2e-llm-inference-service] blockOwnerDeletion: true [e2e-llm-inference-service] managedFields: [e2e-llm-inference-service] - manager: manager [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] apiVersion: apps/v1 [e2e-llm-inference-service] time: '2026-07-07T14:29:24Z' [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:metadata: [e2e-llm-inference-service] f:labels: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:app.kubernetes.io/component: {} [e2e-llm-inference-service] f:app.kubernetes.io/name: {} [e2e-llm-inference-service] f:app.kubernetes.io/part-of: {} [e2e-llm-inference-service] f:kserve.io/component: {} [e2e-llm-inference-service] f:llm-d.ai/role: {} [e2e-llm-inference-service] f:ownerReferences: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"uid":"e15cfc59-e4a1-4876-b43b-85ce3b1b239b"}: {} [e2e-llm-inference-service] f:spec: [e2e-llm-inference-service] f:progressDeadlineSeconds: {} [e2e-llm-inference-service] f:replicas: {} [e2e-llm-inference-service] f:revisionHistoryLimit: {} [e2e-llm-inference-service] f:selector: {} [e2e-llm-inference-service] f:strategy: [e2e-llm-inference-service] f:rollingUpdate: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:maxSurge: {} [e2e-llm-inference-service] f:maxUnavailable: {} [e2e-llm-inference-service] f:type: {} [e2e-llm-inference-service] f:template: [e2e-llm-inference-service] f:metadata: [e2e-llm-inference-service] f:labels: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:app.kubernetes.io/component: {} [e2e-llm-inference-service] f:app.kubernetes.io/name: {} [e2e-llm-inference-service] f:app.kubernetes.io/part-of: {} [e2e-llm-inference-service] f:kserve.io/component: {} [e2e-llm-inference-service] f:llm-d.ai/role: {} [e2e-llm-inference-service] f:spec: [e2e-llm-inference-service] f:containers: [e2e-llm-inference-service] k:{"name":"main"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:args: {} [e2e-llm-inference-service] f:command: {} [e2e-llm-inference-service] f:env: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"name":"HF_HUB_CACHE"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"HOME"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"TORCHINDUCTOR_CACHE_DIR"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"USER"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"VLLM_CPU_KVCACHE_SPACE"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"VLLM_ENABLE_V1_MULTIPROCESSING"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"VLLM_LOGGING_LEVEL"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] f:image: {} [e2e-llm-inference-service] f:imagePullPolicy: {} [e2e-llm-inference-service] f:lifecycle: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:preStop: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:exec: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:command: {} [e2e-llm-inference-service] f:livenessProbe: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:failureThreshold: {} [e2e-llm-inference-service] f:httpGet: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:path: {} [e2e-llm-inference-service] f:port: {} [e2e-llm-inference-service] f:scheme: {} [e2e-llm-inference-service] f:periodSeconds: {} [e2e-llm-inference-service] f:successThreshold: {} [e2e-llm-inference-service] f:timeoutSeconds: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:ports: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"containerPort":8000,"protocol":"TCP"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:containerPort: {} [e2e-llm-inference-service] f:protocol: {} [e2e-llm-inference-service] f:readinessProbe: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:failureThreshold: {} [e2e-llm-inference-service] f:httpGet: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:path: {} [e2e-llm-inference-service] f:port: {} [e2e-llm-inference-service] f:scheme: {} [e2e-llm-inference-service] f:periodSeconds: {} [e2e-llm-inference-service] f:successThreshold: {} [e2e-llm-inference-service] f:timeoutSeconds: {} [e2e-llm-inference-service] f:resources: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:limits: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:cpu: {} [e2e-llm-inference-service] f:memory: {} [e2e-llm-inference-service] f:requests: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:cpu: {} [e2e-llm-inference-service] f:memory: {} [e2e-llm-inference-service] f:securityContext: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:allowPrivilegeEscalation: {} [e2e-llm-inference-service] f:capabilities: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:drop: {} [e2e-llm-inference-service] f:readOnlyRootFilesystem: {} [e2e-llm-inference-service] f:runAsNonRoot: {} [e2e-llm-inference-service] f:seccompProfile: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:type: {} [e2e-llm-inference-service] f:startupProbe: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:failureThreshold: {} [e2e-llm-inference-service] f:httpGet: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:path: {} [e2e-llm-inference-service] f:port: {} [e2e-llm-inference-service] f:scheme: {} [e2e-llm-inference-service] f:periodSeconds: {} [e2e-llm-inference-service] f:successThreshold: {} [e2e-llm-inference-service] f:timeoutSeconds: {} [e2e-llm-inference-service] f:terminationMessagePath: {} [e2e-llm-inference-service] f:terminationMessagePolicy: {} [e2e-llm-inference-service] f:volumeMounts: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"mountPath":"/dev/shm"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:mountPath: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] k:{"mountPath":"/home"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:mountPath: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] k:{"mountPath":"/mnt"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:mountPath: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:readOnly: {} [e2e-llm-inference-service] k:{"mountPath":"/models"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:mountPath: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] k:{"mountPath":"/tmp"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:mountPath: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] k:{"mountPath":"/var/run/kserve/tls"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:mountPath: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:readOnly: {} [e2e-llm-inference-service] f:dnsPolicy: {} [e2e-llm-inference-service] f:initContainers: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"name":"storage-initializer"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:args: {} [e2e-llm-inference-service] f:env: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"name":"AWS_ACCESS_KEY_ID"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:valueFrom: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:secretKeyRef: {} [e2e-llm-inference-service] k:{"name":"AWS_CA_BUNDLE"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"AWS_CA_BUNDLE_CONFIGMAP"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"AWS_ENDPOINT_URL"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"AWS_SECRET_ACCESS_KEY"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:valueFrom: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:secretKeyRef: {} [e2e-llm-inference-service] k:{"name":"CA_BUNDLE_CONFIGMAP_NAME"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"CA_BUNDLE_VOLUME_MOUNT_POINT"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"HF_HUB_ENABLE_HF_TRANSFER"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"HF_XET_HIGH_PERFORMANCE"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"HF_XET_NUM_CONCURRENT_RANGE_GETS"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"S3_ENDPOINT"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"S3_USE_HTTPS"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"S3_VERIFY_SSL"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] f:image: {} [e2e-llm-inference-service] f:imagePullPolicy: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:resources: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:limits: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:cpu: {} [e2e-llm-inference-service] f:memory: {} [e2e-llm-inference-service] f:requests: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:cpu: {} [e2e-llm-inference-service] f:memory: {} [e2e-llm-inference-service] f:terminationMessagePath: {} [e2e-llm-inference-service] f:terminationMessagePolicy: {} [e2e-llm-inference-service] f:volumeMounts: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"mountPath":"/etc/ssl/custom-certs"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:mountPath: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:readOnly: {} [e2e-llm-inference-service] k:{"mountPath":"/mnt"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:mountPath: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:restartPolicy: {} [e2e-llm-inference-service] f:schedulerName: {} [e2e-llm-inference-service] f:securityContext: {} [e2e-llm-inference-service] f:terminationGracePeriodSeconds: {} [e2e-llm-inference-service] f:volumes: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"name":"cabundle-cert"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:configMap: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:defaultMode: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] k:{"name":"dshm"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:emptyDir: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:medium: {} [e2e-llm-inference-service] f:sizeLimit: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] k:{"name":"home"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:emptyDir: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] k:{"name":"kserve-provision-location"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:emptyDir: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] k:{"name":"model-cache"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:emptyDir: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] k:{"name":"tls-certs"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:secret: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:defaultMode: {} [e2e-llm-inference-service] f:secretName: {} [e2e-llm-inference-service] k:{"name":"tmp-dir"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:emptyDir: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] - manager: kube-controller-manager [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] apiVersion: apps/v1 [e2e-llm-inference-service] time: '2026-07-07T14:30:45Z' [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:metadata: [e2e-llm-inference-service] f:annotations: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:deployment.kubernetes.io/revision: {} [e2e-llm-inference-service] f:status: [e2e-llm-inference-service] f:availableReplicas: {} [e2e-llm-inference-service] f:conditions: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"type":"Available"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:lastTransitionTime: {} [e2e-llm-inference-service] f:lastUpdateTime: {} [e2e-llm-inference-service] f:message: {} [e2e-llm-inference-service] f:reason: {} [e2e-llm-inference-service] f:status: {} [e2e-llm-inference-service] f:type: {} [e2e-llm-inference-service] k:{"type":"Progressing"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:lastTransitionTime: {} [e2e-llm-inference-service] f:lastUpdateTime: {} [e2e-llm-inference-service] f:message: {} [e2e-llm-inference-service] f:reason: {} [e2e-llm-inference-service] f:status: {} [e2e-llm-inference-service] f:type: {} [e2e-llm-inference-service] f:observedGeneration: {} [e2e-llm-inference-service] f:readyReplicas: {} [e2e-llm-inference-service] f:replicas: {} [e2e-llm-inference-service] f:updatedReplicas: {} [e2e-llm-inference-service] subresource: status [e2e-llm-inference-service] spec: [e2e-llm-inference-service] replicas: 1 [e2e-llm-inference-service] selector: [e2e-llm-inference-service] matchLabels: [e2e-llm-inference-service] app.kubernetes.io/component: llminferenceservice-workload [e2e-llm-inference-service] app.kubernetes.io/name: llmisvc-model-fb-opt-125m-with-7ca60146 [e2e-llm-inference-service] app.kubernetes.io/part-of: llminferenceservice [e2e-llm-inference-service] kserve.io/component: workload [e2e-llm-inference-service] llm-d.ai/role: both [e2e-llm-inference-service] template: [e2e-llm-inference-service] metadata: [e2e-llm-inference-service] labels: [e2e-llm-inference-service] app.kubernetes.io/component: llminferenceservice-workload [e2e-llm-inference-service] app.kubernetes.io/name: llmisvc-model-fb-opt-125m-with-7ca60146 [e2e-llm-inference-service] app.kubernetes.io/part-of: llminferenceservice [e2e-llm-inference-service] kserve.io/component: workload [e2e-llm-inference-service] llm-d.ai/role: both [e2e-llm-inference-service] spec: [e2e-llm-inference-service] volumes: [e2e-llm-inference-service] - name: home [e2e-llm-inference-service] emptyDir: {} [e2e-llm-inference-service] - name: dshm [e2e-llm-inference-service] emptyDir: [e2e-llm-inference-service] medium: Memory [e2e-llm-inference-service] sizeLimit: 1Gi [e2e-llm-inference-service] - name: model-cache [e2e-llm-inference-service] emptyDir: {} [e2e-llm-inference-service] - name: tmp-dir [e2e-llm-inference-service] emptyDir: {} [e2e-llm-inference-service] - name: tls-certs [e2e-llm-inference-service] secret: [e2e-llm-inference-service] secretName: llmisv3e414c2ba058a022dfd694dbcbac5b51-kserve-self-signed-certs [e2e-llm-inference-service] defaultMode: 420 [e2e-llm-inference-service] - name: kserve-provision-location [e2e-llm-inference-service] emptyDir: {} [e2e-llm-inference-service] - name: cabundle-cert [e2e-llm-inference-service] configMap: [e2e-llm-inference-service] name: odh-kserve-custom-ca-bundle [e2e-llm-inference-service] defaultMode: 420 [e2e-llm-inference-service] initContainers: [e2e-llm-inference-service] - name: storage-initializer [e2e-llm-inference-service] image: quay.io/opendatahub/kserve-storage-initializer@sha256:13ad765c0f323200e1db7b45c99661287e65c5158d6b54b16139829636d20be4 [e2e-llm-inference-service] args: [e2e-llm-inference-service] - hf://facebook/opt-125m [e2e-llm-inference-service] - /mnt/models [e2e-llm-inference-service] - hf://edbeeching/opt-125m-lora [e2e-llm-inference-service] - /mnt/lora/lora-adapter-1 [e2e-llm-inference-service] env: [e2e-llm-inference-service] - name: AWS_ACCESS_KEY_ID [e2e-llm-inference-service] valueFrom: [e2e-llm-inference-service] secretKeyRef: [e2e-llm-inference-service] name: seaweedfs-s3-creds [e2e-llm-inference-service] key: AWS_ACCESS_KEY_ID [e2e-llm-inference-service] - name: AWS_SECRET_ACCESS_KEY [e2e-llm-inference-service] valueFrom: [e2e-llm-inference-service] secretKeyRef: [e2e-llm-inference-service] name: seaweedfs-s3-creds [e2e-llm-inference-service] key: AWS_SECRET_ACCESS_KEY [e2e-llm-inference-service] - name: S3_USE_HTTPS [e2e-llm-inference-service] value: '0' [e2e-llm-inference-service] - name: S3_ENDPOINT [e2e-llm-inference-service] value: s3-service.kserve:8333 [e2e-llm-inference-service] - name: AWS_ENDPOINT_URL [e2e-llm-inference-service] value: http://s3-service.kserve:8333 [e2e-llm-inference-service] - name: S3_VERIFY_SSL [e2e-llm-inference-service] value: '0' [e2e-llm-inference-service] - name: AWS_CA_BUNDLE [e2e-llm-inference-service] value: /etc/ssl/custom-certs/cabundle.crt [e2e-llm-inference-service] - name: AWS_CA_BUNDLE_CONFIGMAP [e2e-llm-inference-service] value: odh-kserve-custom-ca-bundle [e2e-llm-inference-service] - name: HF_HUB_ENABLE_HF_TRANSFER [e2e-llm-inference-service] value: '1' [e2e-llm-inference-service] - name: HF_XET_HIGH_PERFORMANCE [e2e-llm-inference-service] value: '1' [e2e-llm-inference-service] - name: HF_XET_NUM_CONCURRENT_RANGE_GETS [e2e-llm-inference-service] value: '8' [e2e-llm-inference-service] - name: CA_BUNDLE_CONFIGMAP_NAME [e2e-llm-inference-service] value: odh-kserve-custom-ca-bundle [e2e-llm-inference-service] - name: CA_BUNDLE_VOLUME_MOUNT_POINT [e2e-llm-inference-service] value: /etc/ssl/custom-certs [e2e-llm-inference-service] resources: [e2e-llm-inference-service] limits: [e2e-llm-inference-service] cpu: '1' [e2e-llm-inference-service] memory: 24Gi [e2e-llm-inference-service] requests: [e2e-llm-inference-service] cpu: 100m [e2e-llm-inference-service] memory: 100Mi [e2e-llm-inference-service] volumeMounts: [e2e-llm-inference-service] - name: kserve-provision-location [e2e-llm-inference-service] mountPath: /mnt [e2e-llm-inference-service] - name: cabundle-cert [e2e-llm-inference-service] readOnly: true [e2e-llm-inference-service] mountPath: /etc/ssl/custom-certs [e2e-llm-inference-service] terminationMessagePath: /dev/termination-log [e2e-llm-inference-service] terminationMessagePolicy: FallbackToLogsOnError [e2e-llm-inference-service] imagePullPolicy: IfNotPresent [e2e-llm-inference-service] containers: [e2e-llm-inference-service] - name: main [e2e-llm-inference-service] image: public.ecr.aws/q9t5s3a7/vllm-cpu-release-repo:v0.19.0 [e2e-llm-inference-service] command: [e2e-llm-inference-service] - /bin/bash [e2e-llm-inference-service] - -c [e2e-llm-inference-service] - "if [ -f /etc/profile.d/ibm-aiu-setup.sh ]; then\n source /etc/profile.d/ibm-aiu-setup.sh\n\ [e2e-llm-inference-service] fi\n\nif [ \"$KSERVE_INFER_ROCE\" = \"true\" ]; then\n echo \"Trying to\ [e2e-llm-inference-service] \ infer RoCE configs ... \"\n grep -H . /sys/class/infiniband/*/ports/*/gids/*\ [e2e-llm-inference-service] \ 2>/dev/null\n grep -H . /sys/class/infiniband/*/ports/*/gid_attrs/types/*\ [e2e-llm-inference-service] \ 2>/dev/null\n\n cat /proc/driver/nvidia/params\n\n KSERVE_INFER_IB_GID_INDEX_GREP=${KSERVE_INFER_IB_GID_INDEX_GREP:-\"\ [e2e-llm-inference-service] RoCE v2\"}\n\n echo \"[Infer RoCE] Discovering active HCAs ...\"\n active_hcas=()\n\ [e2e-llm-inference-service] \ # Loop through all mlx5 devices found in sysfs\n for hca_dir in /sys/class/infiniband/mlx5_*;\ [e2e-llm-inference-service] \ do\n # Ensure it's a directory before proceeding\n if [ -d \"\ [e2e-llm-inference-service] $hca_dir\" ]; then\n hca_name=$(basename \"$hca_dir\")\n \ [e2e-llm-inference-service] \ port_state_file=\"$hca_dir/ports/1/state\" # Assume port 1\n \ [e2e-llm-inference-service] \ type_file=\"$hca_dir/ports/1/gid_attrs/types/*\"\n\n echo\ [e2e-llm-inference-service] \ \"[Infer RoCE] Check if the port state file ${port_state_file} exists\ [e2e-llm-inference-service] \ and contains 'ACTIVE'\"\n if [ -f \"$port_state_file\" ] && grep\ [e2e-llm-inference-service] \ -q \"ACTIVE\" \"$port_state_file\" && grep -q \"${KSERVE_INFER_IB_GID_INDEX_GREP}\"\ [e2e-llm-inference-service] \ ${type_file} 2>/dev/null; then\n echo \"[Infer RoCE] Found\ [e2e-llm-inference-service] \ active HCA: $hca_name\"\n active_hcas+=(\"$hca_name\")\n\ [e2e-llm-inference-service] \ else\n echo \"[Infer RoCE] Skipping inactive or\ [e2e-llm-inference-service] \ down HCA: $hca_name\"\n fi\n fi\n done\n\n # Check if\ [e2e-llm-inference-service] \ we found any active HCAs\n if [ ${#active_hcas[@]} -gt 0 ]; then\n \ [e2e-llm-inference-service] \ # Join the array elements with a comma\n hca_port_pairs=()\n \ [e2e-llm-inference-service] \ for hca in \"${active_hcas[@]}\"; do\n hca_port_pairs+=(\"\ [e2e-llm-inference-service] ${hca}:1\")\n done\n\n active_hca_list=$(IFS=,; echo \"${active_hcas[*]}\"\ [e2e-llm-inference-service] )\n hca_port_pairs_list=$(IFS=,; echo \"${hca_port_pairs[*]}\")\n \ [e2e-llm-inference-service] \ echo \"[Infer RoCE] Setting active HCAs: ${active_hca_list}\"\n \ [e2e-llm-inference-service] \ export NCCL_IB_HCA=${NCCL_IB_HCA:-${active_hca_list}}\n export\ [e2e-llm-inference-service] \ NVSHMEM_HCA_LIST=${NVSHMEM_HCA_LIST:-${hca_port_pairs_list}}\n export\ [e2e-llm-inference-service] \ UCX_NET_DEVICES=${UCX_NET_DEVICES:-${hca_port_pairs_list}}\n\n echo\ [e2e-llm-inference-service] \ \"[Infer RoCE] NCCL_IB_HCA=${NCCL_IB_HCA}\"\n echo \"[Infer RoCE]\ [e2e-llm-inference-service] \ NVSHMEM_HCA_LIST=${NVSHMEM_HCA_LIST}\"\n echo \"[Infer RoCE] UCX_NET_DEVICES=${UCX_NET_DEVICES}\"\ [e2e-llm-inference-service] \n else\n echo \"[Infer RoCE] WARNING: No active RoCE HCAs found.\ [e2e-llm-inference-service] \ NCCL_IB_HCA will not be set.\"\n fi\n\n if [ ${#active_hcas[@]} -gt\ [e2e-llm-inference-service] \ 0 ]; then\n echo \"[Infer RoCE] Finding GID_INDEX for each active\ [e2e-llm-inference-service] \ HCA (SR-IOV compatible)...\"\n\n # For SR-IOV environments, find\ [e2e-llm-inference-service] \ the most common IPv4 RoCE v2 GID index across all HCAs\n declare\ [e2e-llm-inference-service] \ -A gid_index_count\n declare -A hca_gid_index\n\n for hca_name\ [e2e-llm-inference-service] \ in \"${active_hcas[@]}\"; do\n echo \"[Infer RoCE] Processing\ [e2e-llm-inference-service] \ HCA: ${hca_name}\"\n\n # Find all RoCE v2 IPv4 GIDs for this\ [e2e-llm-inference-service] \ HCA and count by index\n for tpath in /sys/class/infiniband/${hca_name}/ports/1/gid_attrs/types/*;\ [e2e-llm-inference-service] \ do\n if grep -q \"${KSERVE_INFER_IB_GID_INDEX_GREP}\" \"\ [e2e-llm-inference-service] $tpath\" 2>/dev/null; then\n idx=$(basename \"$tpath\"\ [e2e-llm-inference-service] )\n gid_file=\"/sys/class/infiniband/${hca_name}/ports/1/gids/${idx}\"\ [e2e-llm-inference-service] \n # Check for IPv4 GID (contains ffff:)\n \ [e2e-llm-inference-service] \ if [ -f \"$gid_file\" ] && grep -q \"ffff:\" \"$gid_file\"; then\n\ [e2e-llm-inference-service] \ gid_value=$(cat \"$gid_file\" 2>/dev/null || echo\ [e2e-llm-inference-service] \ \"\")\n echo \"[Infer RoCE] Found IPv4 RoCE v2 GID\ [e2e-llm-inference-service] \ for ${hca_name}: index=${idx}, gid=${gid_value}\"\n \ [e2e-llm-inference-service] \ hca_gid_index[\"${hca_name}\"]=\"${idx}\"\n gid_index_count[\"\ [e2e-llm-inference-service] ${idx}\"]=$((${gid_index_count[\"${idx}\"]} + 1))\n \ [e2e-llm-inference-service] \ break # Use first found IPv4 GID per HCA\n fi\n \ [e2e-llm-inference-service] \ fi\n done\n done\n\n # Find the most common\ [e2e-llm-inference-service] \ GID index (most likely to be consistent across nodes)\n best_gid_index=\"\ [e2e-llm-inference-service] \"\n max_count=0\n for idx in \"${!gid_index_count[@]}\"; do\n\ [e2e-llm-inference-service] \ count=${gid_index_count[\"${idx}\"]}\n echo \"[Infer\ [e2e-llm-inference-service] \ RoCE] GID_INDEX ${idx} found on ${count} HCAs\"\n if [ $count\ [e2e-llm-inference-service] \ -gt $max_count ]; then\n max_count=$count\n \ [e2e-llm-inference-service] \ best_gid_index=\"$idx\"\n fi\n done\n\n # Use deterministic\ [e2e-llm-inference-service] \ fallback if tied - prefer index 3 (SR-IOV standard)\n if [ ${#gid_index_count[@]}\ [e2e-llm-inference-service] \ -gt 1 ]; then\n echo \"[Infer RoCE] Multiple GID indices found,\ [e2e-llm-inference-service] \ selecting most common: ${best_gid_index}\"\n # If there's a tie,\ [e2e-llm-inference-service] \ prefer index 3 as it's most common in SR-IOV setups\n if [ -n\ [e2e-llm-inference-service] \ \"${gid_index_count['3']}\" ] && [ \"${gid_index_count['3']}\" -eq \"\ [e2e-llm-inference-service] $max_count\" ]; then\n best_gid_index=\"3\"\n \ [e2e-llm-inference-service] \ echo \"[Infer RoCE] Using deterministic fallback: GID_INDEX=3 (SR-IOV\ [e2e-llm-inference-service] \ standard)\"\n fi\n fi\n\n # Check if GID_INDEX is already\ [e2e-llm-inference-service] \ set via environment variables\n if [ -n \"${NCCL_IB_GID_INDEX}\"\ [e2e-llm-inference-service] \ ]; then\n echo \"[Infer RoCE] Using pre-configured NCCL_IB_GID_INDEX=${NCCL_IB_GID_INDEX}\ [e2e-llm-inference-service] \ from environment\"\n export NVSHMEM_IB_GID_INDEX=${NVSHMEM_IB_GID_INDEX:-$NCCL_IB_GID_INDEX}\n\ [e2e-llm-inference-service] \ export UCX_IB_GID_INDEX=${UCX_IB_GID_INDEX:-$NCCL_IB_GID_INDEX}\n\ [e2e-llm-inference-service] \ echo \"[Infer RoCE] Using pre-configured GID_INDEX=${NCCL_IB_GID_INDEX}\ [e2e-llm-inference-service] \ for NCCL, NVSHMEM, and UCX\"\n elif [ -n \"$best_gid_index\" ]; then\n\ [e2e-llm-inference-service] \ echo \"[Infer RoCE] Selected GID_INDEX: ${best_gid_index} (found\ [e2e-llm-inference-service] \ on ${max_count} HCAs)\"\n\n export NCCL_IB_GID_INDEX=${NCCL_IB_GID_INDEX:-$best_gid_index}\n\ [e2e-llm-inference-service] \ export NVSHMEM_IB_GID_INDEX=${NVSHMEM_IB_GID_INDEX:-$best_gid_index}\n\ [e2e-llm-inference-service] \ export UCX_IB_GID_INDEX=${UCX_IB_GID_INDEX:-$best_gid_index}\n\ [e2e-llm-inference-service] \n echo \"[Infer RoCE] Exported GID_INDEX=${best_gid_index} for\ [e2e-llm-inference-service] \ NCCL, NVSHMEM, and UCX\"\n else\n echo \"[Infer RoCE] ERROR:\ [e2e-llm-inference-service] \ No valid IPv4 ${KSERVE_INFER_IB_GID_INDEX_GREP} GID_INDEX found on any\ [e2e-llm-inference-service] \ HCA.\"\n fi\n else\n echo \"[Infer RoCE] No active HCAs found,\ [e2e-llm-inference-service] \ skipping GID_INDEX inference.\"\n fi\nfi\n\n# --disable-access-log-for-endpoints\ [e2e-llm-inference-service] \ landed in vLLM 0.16.0 (vllm-project/vllm#30011).\n# Older versions still\ [e2e-llm-inference-service] \ need the blanket --disable-uvicorn-access-log.\nACCESS_LOG_ARGS=\"--disable-uvicorn-access-log\"\ [e2e-llm-inference-service] \nVLLM_VERSION=$(vllm --version 2>/dev/null | tail -1 | awk '{print $NF}')\n\ [e2e-llm-inference-service] echo \"[access-log-detect] vllm version='${VLLM_VERSION}'\"\nif [[ \"$VLLM_VERSION\"\ [e2e-llm-inference-service] \ =~ ^[0-9]+\\.[0-9]+ ]] && [ \"$(printf '%s\\n%s\\n' \"0.16.0\" \"${VLLM_VERSION}\"\ [e2e-llm-inference-service] \ | sort -V | head -1)\" = \"0.16.0\" ]; then\n ACCESS_LOG_ARGS=\"--disable-access-log-for-endpoints\ [e2e-llm-inference-service] \ /health,/metrics,/ping\"\nfi\necho \"[access-log-detect] selected ACCESS_LOG_ARGS='${ACCESS_LOG_ARGS}'\"\ [e2e-llm-inference-service] \n\n# --shutdown-timeout landed in vLLM 0.18.0 (vllm-project/vllm#36666).\n\ [e2e-llm-inference-service] SHUTDOWN_TIMEOUT_ARGS=\"\"\nif [[ \"$VLLM_VERSION\" =~ ^[0-9]+\\.[0-9]+\ [e2e-llm-inference-service] \ ]] && [ \"$(printf '%s\\n%s\\n' \"0.18.0\" \"${VLLM_VERSION}\" | sort\ [e2e-llm-inference-service] \ -V | head -1)\" = \"0.18.0\" ]; then\n SHUTDOWN_TIMEOUT_ARGS=\"--shutdown-timeout\ [e2e-llm-inference-service] \ 40\"\nfi\n\neval \"exec vllm serve /mnt/models \\\n --served-model-name\ [e2e-llm-inference-service] \ \"facebook/opt-125m\" \"publishers/kserve-ci-e2e-test/models/facebook/opt-125m\"\ [e2e-llm-inference-service] \ \\\n --port 8000 \\\n ${ACCESS_LOG_ARGS} \\\n ${SHUTDOWN_TIMEOUT_ARGS}\ [e2e-llm-inference-service] \ \\\n --enable-ssl-refresh \\\n --ssl-certfile /var/run/kserve/tls/tls.crt\ [e2e-llm-inference-service] \ \\\n --ssl-keyfile /var/run/kserve/tls/tls.key \\\n ${VLLM_ADDITIONAL_ARGS}\ [e2e-llm-inference-service] \ \\\n $@\"" [e2e-llm-inference-service] - -- [e2e-llm-inference-service] args: [e2e-llm-inference-service] - --enable-lora [e2e-llm-inference-service] - --lora-modules [e2e-llm-inference-service] - '''{"name":"lora-adapter-1","path":"/mnt/lora/lora-adapter-1"}''' [e2e-llm-inference-service] - '''{"name":"publishers/kserve-ci-e2e-test/models/lora-adapter-1","path":"/mnt/lora/lora-adapter-1"}''' [e2e-llm-inference-service] ports: [e2e-llm-inference-service] - containerPort: 8000 [e2e-llm-inference-service] protocol: TCP [e2e-llm-inference-service] env: [e2e-llm-inference-service] - name: HOME [e2e-llm-inference-service] value: /home [e2e-llm-inference-service] - name: VLLM_LOGGING_LEVEL [e2e-llm-inference-service] value: DEBUG [e2e-llm-inference-service] - name: VLLM_CPU_KVCACHE_SPACE [e2e-llm-inference-service] value: '1' [e2e-llm-inference-service] - name: VLLM_ENABLE_V1_MULTIPROCESSING [e2e-llm-inference-service] value: '0' [e2e-llm-inference-service] - name: USER [e2e-llm-inference-service] value: nonroot [e2e-llm-inference-service] - name: TORCHINDUCTOR_CACHE_DIR [e2e-llm-inference-service] value: /tmp/torchinductor-cache [e2e-llm-inference-service] - name: HF_HUB_CACHE [e2e-llm-inference-service] value: /models [e2e-llm-inference-service] resources: [e2e-llm-inference-service] limits: [e2e-llm-inference-service] cpu: '2' [e2e-llm-inference-service] memory: 7Gi [e2e-llm-inference-service] requests: [e2e-llm-inference-service] cpu: 200m [e2e-llm-inference-service] memory: 2Gi [e2e-llm-inference-service] volumeMounts: [e2e-llm-inference-service] - name: home [e2e-llm-inference-service] mountPath: /home [e2e-llm-inference-service] - name: tmp-dir [e2e-llm-inference-service] mountPath: /tmp [e2e-llm-inference-service] - name: dshm [e2e-llm-inference-service] mountPath: /dev/shm [e2e-llm-inference-service] - name: model-cache [e2e-llm-inference-service] mountPath: /models [e2e-llm-inference-service] - name: tls-certs [e2e-llm-inference-service] readOnly: true [e2e-llm-inference-service] mountPath: /var/run/kserve/tls [e2e-llm-inference-service] - name: kserve-provision-location [e2e-llm-inference-service] readOnly: true [e2e-llm-inference-service] mountPath: /mnt [e2e-llm-inference-service] livenessProbe: [e2e-llm-inference-service] httpGet: [e2e-llm-inference-service] path: /health [e2e-llm-inference-service] port: 8000 [e2e-llm-inference-service] scheme: HTTPS [e2e-llm-inference-service] timeoutSeconds: 1 [e2e-llm-inference-service] periodSeconds: 10 [e2e-llm-inference-service] successThreshold: 1 [e2e-llm-inference-service] failureThreshold: 10 [e2e-llm-inference-service] readinessProbe: [e2e-llm-inference-service] httpGet: [e2e-llm-inference-service] path: /health [e2e-llm-inference-service] port: 8000 [e2e-llm-inference-service] scheme: HTTPS [e2e-llm-inference-service] timeoutSeconds: 1 [e2e-llm-inference-service] periodSeconds: 1 [e2e-llm-inference-service] successThreshold: 1 [e2e-llm-inference-service] failureThreshold: 2 [e2e-llm-inference-service] startupProbe: [e2e-llm-inference-service] httpGet: [e2e-llm-inference-service] path: /health [e2e-llm-inference-service] port: 8000 [e2e-llm-inference-service] scheme: HTTPS [e2e-llm-inference-service] timeoutSeconds: 1 [e2e-llm-inference-service] periodSeconds: 10 [e2e-llm-inference-service] successThreshold: 1 [e2e-llm-inference-service] failureThreshold: 60 [e2e-llm-inference-service] lifecycle: [e2e-llm-inference-service] preStop: [e2e-llm-inference-service] exec: [e2e-llm-inference-service] command: [e2e-llm-inference-service] - /bin/sleep [e2e-llm-inference-service] - '15' [e2e-llm-inference-service] terminationMessagePath: /dev/termination-log [e2e-llm-inference-service] terminationMessagePolicy: FallbackToLogsOnError [e2e-llm-inference-service] imagePullPolicy: IfNotPresent [e2e-llm-inference-service] securityContext: [e2e-llm-inference-service] capabilities: [e2e-llm-inference-service] drop: [e2e-llm-inference-service] - ALL [e2e-llm-inference-service] runAsNonRoot: true [e2e-llm-inference-service] readOnlyRootFilesystem: true [e2e-llm-inference-service] allowPrivilegeEscalation: false [e2e-llm-inference-service] seccompProfile: [e2e-llm-inference-service] type: RuntimeDefault [e2e-llm-inference-service] restartPolicy: Always [e2e-llm-inference-service] terminationGracePeriodSeconds: 60 [e2e-llm-inference-service] dnsPolicy: ClusterFirst [e2e-llm-inference-service] securityContext: {} [e2e-llm-inference-service] schedulerName: default-scheduler [e2e-llm-inference-service] strategy: [e2e-llm-inference-service] type: RollingUpdate [e2e-llm-inference-service] rollingUpdate: [e2e-llm-inference-service] maxUnavailable: 25% [e2e-llm-inference-service] maxSurge: 25% [e2e-llm-inference-service] revisionHistoryLimit: 10 [e2e-llm-inference-service] progressDeadlineSeconds: 600 [e2e-llm-inference-service] status: [e2e-llm-inference-service] observedGeneration: 1 [e2e-llm-inference-service] replicas: 1 [e2e-llm-inference-service] updatedReplicas: 1 [e2e-llm-inference-service] readyReplicas: 1 [e2e-llm-inference-service] availableReplicas: 1 [e2e-llm-inference-service] conditions: [e2e-llm-inference-service] - type: Available [e2e-llm-inference-service] status: 'True' [e2e-llm-inference-service] lastUpdateTime: '2026-07-07T14:30:45Z' [e2e-llm-inference-service] lastTransitionTime: '2026-07-07T14:30:45Z' [e2e-llm-inference-service] reason: MinimumReplicasAvailable [e2e-llm-inference-service] message: Deployment has minimum availability. [e2e-llm-inference-service] - type: Progressing [e2e-llm-inference-service] status: 'True' [e2e-llm-inference-service] lastUpdateTime: '2026-07-07T14:30:45Z' [e2e-llm-inference-service] lastTransitionTime: '2026-07-07T14:29:24Z' [e2e-llm-inference-service] reason: NewReplicaSetAvailable [e2e-llm-inference-service] message: ReplicaSet "llmisvc-model-fb-opt-125m-with-7ca60146-kserve-7cfdf49d74" [e2e-llm-inference-service] has successfully progressed. [e2e-llm-inference-service] apiVersion: apps/v1 [e2e-llm-inference-service] kind: Deployment [e2e-llm-inference-service] [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1254 --- [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1255 metadata: [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-7ca60146-kserve-router-scheduler [e2e-llm-inference-service] namespace: kserve-ci-e2e-test [e2e-llm-inference-service] uid: ef37f35e-12b2-4fac-8d62-d818466fcbff [e2e-llm-inference-service] resourceVersion: '28939' [e2e-llm-inference-service] generation: 1 [e2e-llm-inference-service] creationTimestamp: '2026-07-07T14:29:24Z' [e2e-llm-inference-service] labels: [e2e-llm-inference-service] app.kubernetes.io/component: llminferenceservice-router-scheduler [e2e-llm-inference-service] app.kubernetes.io/name: llmisvc-model-fb-opt-125m-with-7ca60146 [e2e-llm-inference-service] app.kubernetes.io/part-of: llminferenceservice [e2e-llm-inference-service] annotations: [e2e-llm-inference-service] deployment.kubernetes.io/revision: '1' [e2e-llm-inference-service] ownerReferences: [e2e-llm-inference-service] - apiVersion: serving.kserve.io/v1alpha2 [e2e-llm-inference-service] kind: LLMInferenceService [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-7ca60146 [e2e-llm-inference-service] uid: e15cfc59-e4a1-4876-b43b-85ce3b1b239b [e2e-llm-inference-service] controller: true [e2e-llm-inference-service] blockOwnerDeletion: true [e2e-llm-inference-service] managedFields: [e2e-llm-inference-service] - manager: manager [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] apiVersion: apps/v1 [e2e-llm-inference-service] time: '2026-07-07T14:29:24Z' [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:metadata: [e2e-llm-inference-service] f:labels: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:app.kubernetes.io/component: {} [e2e-llm-inference-service] f:app.kubernetes.io/name: {} [e2e-llm-inference-service] f:app.kubernetes.io/part-of: {} [e2e-llm-inference-service] f:ownerReferences: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"uid":"e15cfc59-e4a1-4876-b43b-85ce3b1b239b"}: {} [e2e-llm-inference-service] f:spec: [e2e-llm-inference-service] f:progressDeadlineSeconds: {} [e2e-llm-inference-service] f:replicas: {} [e2e-llm-inference-service] f:revisionHistoryLimit: {} [e2e-llm-inference-service] f:selector: {} [e2e-llm-inference-service] f:strategy: [e2e-llm-inference-service] f:type: {} [e2e-llm-inference-service] f:template: [e2e-llm-inference-service] f:metadata: [e2e-llm-inference-service] f:annotations: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:app.kubernetes.io/version: {} [e2e-llm-inference-service] f:certificates.kserve.io/expiration-v2: {} [e2e-llm-inference-service] f:labels: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:app.kubernetes.io/component: {} [e2e-llm-inference-service] f:app.kubernetes.io/name: {} [e2e-llm-inference-service] f:app.kubernetes.io/part-of: {} [e2e-llm-inference-service] f:spec: [e2e-llm-inference-service] f:containers: [e2e-llm-inference-service] k:{"name":"main"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:args: {} [e2e-llm-inference-service] f:command: {} [e2e-llm-inference-service] f:env: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"name":"SSL_CERT_DIR"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] f:image: {} [e2e-llm-inference-service] f:imagePullPolicy: {} [e2e-llm-inference-service] f:lifecycle: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:preStop: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:exec: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:command: {} [e2e-llm-inference-service] f:livenessProbe: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:failureThreshold: {} [e2e-llm-inference-service] f:grpc: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:port: {} [e2e-llm-inference-service] f:service: {} [e2e-llm-inference-service] f:initialDelaySeconds: {} [e2e-llm-inference-service] f:periodSeconds: {} [e2e-llm-inference-service] f:successThreshold: {} [e2e-llm-inference-service] f:timeoutSeconds: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:ports: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"containerPort":5557,"protocol":"TCP"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:containerPort: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:protocol: {} [e2e-llm-inference-service] k:{"containerPort":9002,"protocol":"TCP"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:containerPort: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:protocol: {} [e2e-llm-inference-service] k:{"containerPort":9003,"protocol":"TCP"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:containerPort: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:protocol: {} [e2e-llm-inference-service] k:{"containerPort":9090,"protocol":"TCP"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:containerPort: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:protocol: {} [e2e-llm-inference-service] f:readinessProbe: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:failureThreshold: {} [e2e-llm-inference-service] f:grpc: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:port: {} [e2e-llm-inference-service] f:service: {} [e2e-llm-inference-service] f:initialDelaySeconds: {} [e2e-llm-inference-service] f:periodSeconds: {} [e2e-llm-inference-service] f:successThreshold: {} [e2e-llm-inference-service] f:timeoutSeconds: {} [e2e-llm-inference-service] f:resources: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:requests: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:cpu: {} [e2e-llm-inference-service] f:memory: {} [e2e-llm-inference-service] f:securityContext: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:allowPrivilegeEscalation: {} [e2e-llm-inference-service] f:capabilities: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:drop: {} [e2e-llm-inference-service] f:readOnlyRootFilesystem: {} [e2e-llm-inference-service] f:runAsNonRoot: {} [e2e-llm-inference-service] f:seccompProfile: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:type: {} [e2e-llm-inference-service] f:terminationMessagePath: {} [e2e-llm-inference-service] f:terminationMessagePolicy: {} [e2e-llm-inference-service] f:volumeMounts: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"mountPath":"/tmp/tokenizer"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:mountPath: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] k:{"mountPath":"/var/run/kserve/tls"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:mountPath: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:readOnly: {} [e2e-llm-inference-service] f:dnsPolicy: {} [e2e-llm-inference-service] f:initContainers: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"name":"storage-initializer"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:args: {} [e2e-llm-inference-service] f:env: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"name":"AWS_ACCESS_KEY_ID"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:valueFrom: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:secretKeyRef: {} [e2e-llm-inference-service] k:{"name":"AWS_CA_BUNDLE"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"AWS_CA_BUNDLE_CONFIGMAP"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"AWS_ENDPOINT_URL"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"AWS_SECRET_ACCESS_KEY"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:valueFrom: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:secretKeyRef: {} [e2e-llm-inference-service] k:{"name":"HF_HUB_ENABLE_HF_TRANSFER"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"HF_XET_HIGH_PERFORMANCE"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"HF_XET_NUM_CONCURRENT_RANGE_GETS"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"S3_ENDPOINT"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"S3_USE_HTTPS"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"S3_VERIFY_SSL"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"STORAGE_ALLOW_PATTERNS"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] f:image: {} [e2e-llm-inference-service] f:imagePullPolicy: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:resources: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:limits: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:cpu: {} [e2e-llm-inference-service] f:memory: {} [e2e-llm-inference-service] f:requests: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:cpu: {} [e2e-llm-inference-service] f:memory: {} [e2e-llm-inference-service] f:terminationMessagePath: {} [e2e-llm-inference-service] f:terminationMessagePolicy: {} [e2e-llm-inference-service] f:volumeMounts: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"mountPath":"/mnt/models"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:mountPath: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:restartPolicy: {} [e2e-llm-inference-service] f:schedulerName: {} [e2e-llm-inference-service] f:securityContext: {} [e2e-llm-inference-service] f:serviceAccount: {} [e2e-llm-inference-service] f:serviceAccountName: {} [e2e-llm-inference-service] f:terminationGracePeriodSeconds: {} [e2e-llm-inference-service] f:volumes: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"name":"kserve-provision-location"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:emptyDir: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] k:{"name":"tls-certs"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:secret: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:defaultMode: {} [e2e-llm-inference-service] f:secretName: {} [e2e-llm-inference-service] k:{"name":"tokenizer-cache"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:emptyDir: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] k:{"name":"tokenizer-tmp"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:emptyDir: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] k:{"name":"tokenizer-uds"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:emptyDir: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] - manager: kube-controller-manager [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] apiVersion: apps/v1 [e2e-llm-inference-service] time: '2026-07-07T14:29:59Z' [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:metadata: [e2e-llm-inference-service] f:annotations: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:deployment.kubernetes.io/revision: {} [e2e-llm-inference-service] f:status: [e2e-llm-inference-service] f:availableReplicas: {} [e2e-llm-inference-service] f:conditions: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"type":"Available"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:lastTransitionTime: {} [e2e-llm-inference-service] f:lastUpdateTime: {} [e2e-llm-inference-service] f:message: {} [e2e-llm-inference-service] f:reason: {} [e2e-llm-inference-service] f:status: {} [e2e-llm-inference-service] f:type: {} [e2e-llm-inference-service] k:{"type":"Progressing"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:lastTransitionTime: {} [e2e-llm-inference-service] f:lastUpdateTime: {} [e2e-llm-inference-service] f:message: {} [e2e-llm-inference-service] f:reason: {} [e2e-llm-inference-service] f:status: {} [e2e-llm-inference-service] f:type: {} [e2e-llm-inference-service] f:observedGeneration: {} [e2e-llm-inference-service] f:readyReplicas: {} [e2e-llm-inference-service] f:replicas: {} [e2e-llm-inference-service] f:updatedReplicas: {} [e2e-llm-inference-service] subresource: status [e2e-llm-inference-service] spec: [e2e-llm-inference-service] replicas: 1 [e2e-llm-inference-service] selector: [e2e-llm-inference-service] matchLabels: [e2e-llm-inference-service] app.kubernetes.io/component: llminferenceservice-router-scheduler [e2e-llm-inference-service] app.kubernetes.io/name: llmisvc-model-fb-opt-125m-with-7ca60146 [e2e-llm-inference-service] app.kubernetes.io/part-of: llminferenceservice [e2e-llm-inference-service] template: [e2e-llm-inference-service] metadata: [e2e-llm-inference-service] labels: [e2e-llm-inference-service] app.kubernetes.io/component: llminferenceservice-router-scheduler [e2e-llm-inference-service] app.kubernetes.io/name: llmisvc-model-fb-opt-125m-with-7ca60146 [e2e-llm-inference-service] app.kubernetes.io/part-of: llminferenceservice [e2e-llm-inference-service] annotations: [e2e-llm-inference-service] app.kubernetes.io/version: 0.9.0 [e2e-llm-inference-service] certificates.kserve.io/expiration-v2: 'true' [e2e-llm-inference-service] spec: [e2e-llm-inference-service] volumes: [e2e-llm-inference-service] - name: tls-certs [e2e-llm-inference-service] secret: [e2e-llm-inference-service] secretName: llmisv3e414c2ba058a022dfd694dbcbac5b51-kserve-self-signed-certs [e2e-llm-inference-service] defaultMode: 420 [e2e-llm-inference-service] - name: tokenizer-uds [e2e-llm-inference-service] emptyDir: {} [e2e-llm-inference-service] - name: tokenizer-tmp [e2e-llm-inference-service] emptyDir: {} [e2e-llm-inference-service] - name: tokenizer-cache [e2e-llm-inference-service] emptyDir: {} [e2e-llm-inference-service] - name: kserve-provision-location [e2e-llm-inference-service] emptyDir: {} [e2e-llm-inference-service] initContainers: [e2e-llm-inference-service] - name: storage-initializer [e2e-llm-inference-service] image: quay.io/opendatahub/kserve-storage-initializer@sha256:13ad765c0f323200e1db7b45c99661287e65c5158d6b54b16139829636d20be4 [e2e-llm-inference-service] args: [e2e-llm-inference-service] - hf://facebook/opt-125m [e2e-llm-inference-service] - /mnt/models [e2e-llm-inference-service] env: [e2e-llm-inference-service] - name: AWS_ACCESS_KEY_ID [e2e-llm-inference-service] valueFrom: [e2e-llm-inference-service] secretKeyRef: [e2e-llm-inference-service] name: seaweedfs-s3-creds [e2e-llm-inference-service] key: AWS_ACCESS_KEY_ID [e2e-llm-inference-service] - name: AWS_SECRET_ACCESS_KEY [e2e-llm-inference-service] valueFrom: [e2e-llm-inference-service] secretKeyRef: [e2e-llm-inference-service] name: seaweedfs-s3-creds [e2e-llm-inference-service] key: AWS_SECRET_ACCESS_KEY [e2e-llm-inference-service] - name: S3_USE_HTTPS [e2e-llm-inference-service] value: '0' [e2e-llm-inference-service] - name: S3_ENDPOINT [e2e-llm-inference-service] value: s3-service.kserve:8333 [e2e-llm-inference-service] - name: AWS_ENDPOINT_URL [e2e-llm-inference-service] value: http://s3-service.kserve:8333 [e2e-llm-inference-service] - name: S3_VERIFY_SSL [e2e-llm-inference-service] value: '0' [e2e-llm-inference-service] - name: AWS_CA_BUNDLE [e2e-llm-inference-service] value: /etc/ssl/custom-certs/cabundle.crt [e2e-llm-inference-service] - name: AWS_CA_BUNDLE_CONFIGMAP [e2e-llm-inference-service] value: odh-kserve-custom-ca-bundle [e2e-llm-inference-service] - name: HF_HUB_ENABLE_HF_TRANSFER [e2e-llm-inference-service] value: '1' [e2e-llm-inference-service] - name: HF_XET_HIGH_PERFORMANCE [e2e-llm-inference-service] value: '1' [e2e-llm-inference-service] - name: HF_XET_NUM_CONCURRENT_RANGE_GETS [e2e-llm-inference-service] value: '8' [e2e-llm-inference-service] - name: STORAGE_ALLOW_PATTERNS [e2e-llm-inference-service] value: '["tokenizer.json", "tokenizer_config.json", "special_tokens_map.json", [e2e-llm-inference-service] "vocab.json", "merges.txt", "config.json", "generation_config.json"]' [e2e-llm-inference-service] resources: [e2e-llm-inference-service] limits: [e2e-llm-inference-service] cpu: '1' [e2e-llm-inference-service] memory: 24Gi [e2e-llm-inference-service] requests: [e2e-llm-inference-service] cpu: 100m [e2e-llm-inference-service] memory: 100Mi [e2e-llm-inference-service] volumeMounts: [e2e-llm-inference-service] - name: kserve-provision-location [e2e-llm-inference-service] mountPath: /mnt/models [e2e-llm-inference-service] terminationMessagePath: /dev/termination-log [e2e-llm-inference-service] terminationMessagePolicy: FallbackToLogsOnError [e2e-llm-inference-service] imagePullPolicy: IfNotPresent [e2e-llm-inference-service] containers: [e2e-llm-inference-service] - name: main [e2e-llm-inference-service] image: ghcr.io/llm-d/llm-d-router-endpoint-picker:v0.9.0-rc.2 [e2e-llm-inference-service] command: [e2e-llm-inference-service] - /app/epp [e2e-llm-inference-service] - --pool-name [e2e-llm-inference-service] - llmisvc-model-fb-opt-125m-with-7ca60146-inference-pool [e2e-llm-inference-service] - --pool-namespace [e2e-llm-inference-service] - kserve-ci-e2e-test [e2e-llm-inference-service] - --zap-encoder [e2e-llm-inference-service] - json [e2e-llm-inference-service] - --grpc-port [e2e-llm-inference-service] - '9002' [e2e-llm-inference-service] - --grpc-health-port [e2e-llm-inference-service] - '9003' [e2e-llm-inference-service] - --enable-cert-reload=true [e2e-llm-inference-service] - --secure-serving=true [e2e-llm-inference-service] - --model-server-metrics-scheme=https [e2e-llm-inference-service] - --cert-path=/var/run/kserve/tls [e2e-llm-inference-service] args: [e2e-llm-inference-service] - --config-text [e2e-llm-inference-service] - "apiVersion: inference.networking.x-k8s.io/v1alpha1\nkind: EndpointPickerConfig\n\ [e2e-llm-inference-service] plugins:\n- type: single-profile-handler\n- type: queue-scorer\n- type:\ [e2e-llm-inference-service] \ prefix-cache-scorer\n- type: max-score-picker\n- type: lora-affinity-scorer\n\ [e2e-llm-inference-service] schedulingProfiles:\n- name: default\n plugins:\n - pluginRef: lora-affinity-scorer\n\ [e2e-llm-inference-service] \ weight: 4\n - pluginRef: queue-scorer\n weight: 2\n - pluginRef:\ [e2e-llm-inference-service] \ prefix-cache-scorer\n weight: 3\n - pluginRef: max-score-picker\n" [e2e-llm-inference-service] ports: [e2e-llm-inference-service] - name: grpc [e2e-llm-inference-service] containerPort: 9002 [e2e-llm-inference-service] protocol: TCP [e2e-llm-inference-service] - name: grpc-health [e2e-llm-inference-service] containerPort: 9003 [e2e-llm-inference-service] protocol: TCP [e2e-llm-inference-service] - name: metrics [e2e-llm-inference-service] containerPort: 9090 [e2e-llm-inference-service] protocol: TCP [e2e-llm-inference-service] - name: zmq [e2e-llm-inference-service] containerPort: 5557 [e2e-llm-inference-service] protocol: TCP [e2e-llm-inference-service] env: [e2e-llm-inference-service] - name: SSL_CERT_DIR [e2e-llm-inference-service] value: /var/run/kserve/tls:/var/run/secrets/kubernetes.io/serviceaccount:/etc/pki/tls/certs [e2e-llm-inference-service] resources: [e2e-llm-inference-service] requests: [e2e-llm-inference-service] cpu: 256m [e2e-llm-inference-service] memory: 500Mi [e2e-llm-inference-service] volumeMounts: [e2e-llm-inference-service] - name: tls-certs [e2e-llm-inference-service] readOnly: true [e2e-llm-inference-service] mountPath: /var/run/kserve/tls [e2e-llm-inference-service] - name: tokenizer-uds [e2e-llm-inference-service] mountPath: /tmp/tokenizer [e2e-llm-inference-service] livenessProbe: [e2e-llm-inference-service] grpc: [e2e-llm-inference-service] port: 9003 [e2e-llm-inference-service] service: liveness [e2e-llm-inference-service] initialDelaySeconds: 5 [e2e-llm-inference-service] timeoutSeconds: 1 [e2e-llm-inference-service] periodSeconds: 10 [e2e-llm-inference-service] successThreshold: 1 [e2e-llm-inference-service] failureThreshold: 3 [e2e-llm-inference-service] readinessProbe: [e2e-llm-inference-service] grpc: [e2e-llm-inference-service] port: 9003 [e2e-llm-inference-service] service: readiness [e2e-llm-inference-service] initialDelaySeconds: 30 [e2e-llm-inference-service] timeoutSeconds: 1 [e2e-llm-inference-service] periodSeconds: 10 [e2e-llm-inference-service] successThreshold: 1 [e2e-llm-inference-service] failureThreshold: 3 [e2e-llm-inference-service] lifecycle: [e2e-llm-inference-service] preStop: [e2e-llm-inference-service] exec: [e2e-llm-inference-service] command: [e2e-llm-inference-service] - /bin/sleep [e2e-llm-inference-service] - '15' [e2e-llm-inference-service] terminationMessagePath: /dev/termination-log [e2e-llm-inference-service] terminationMessagePolicy: FallbackToLogsOnError [e2e-llm-inference-service] imagePullPolicy: IfNotPresent [e2e-llm-inference-service] securityContext: [e2e-llm-inference-service] capabilities: [e2e-llm-inference-service] drop: [e2e-llm-inference-service] - ALL [e2e-llm-inference-service] runAsNonRoot: true [e2e-llm-inference-service] readOnlyRootFilesystem: true [e2e-llm-inference-service] allowPrivilegeEscalation: false [e2e-llm-inference-service] seccompProfile: [e2e-llm-inference-service] type: RuntimeDefault [e2e-llm-inference-service] restartPolicy: Always [e2e-llm-inference-service] terminationGracePeriodSeconds: 60 [e2e-llm-inference-service] dnsPolicy: ClusterFirst [e2e-llm-inference-service] serviceAccountName: llmisvc-model-fb-opt-125m-with-7ca60146-epp-sa [e2e-llm-inference-service] serviceAccount: llmisvc-model-fb-opt-125m-with-7ca60146-epp-sa [e2e-llm-inference-service] securityContext: {} [e2e-llm-inference-service] schedulerName: default-scheduler [e2e-llm-inference-service] strategy: [e2e-llm-inference-service] type: Recreate [e2e-llm-inference-service] revisionHistoryLimit: 10 [e2e-llm-inference-service] progressDeadlineSeconds: 600 [e2e-llm-inference-service] status: [e2e-llm-inference-service] observedGeneration: 1 [e2e-llm-inference-service] replicas: 1 [e2e-llm-inference-service] updatedReplicas: 1 [e2e-llm-inference-service] readyReplicas: 1 [e2e-llm-inference-service] availableReplicas: 1 [e2e-llm-inference-service] conditions: [e2e-llm-inference-service] - type: Available [e2e-llm-inference-service] status: 'True' [e2e-llm-inference-service] lastUpdateTime: '2026-07-07T14:29:59Z' [e2e-llm-inference-service] lastTransitionTime: '2026-07-07T14:29:59Z' [e2e-llm-inference-service] reason: MinimumReplicasAvailable [e2e-llm-inference-service] message: Deployment has minimum availability. [e2e-llm-inference-service] - type: Progressing [e2e-llm-inference-service] status: 'True' [e2e-llm-inference-service] lastUpdateTime: '2026-07-07T14:29:59Z' [e2e-llm-inference-service] lastTransitionTime: '2026-07-07T14:29:24Z' [e2e-llm-inference-service] reason: NewReplicaSetAvailable [e2e-llm-inference-service] message: ReplicaSet "llmisvc-model-fb-opt-125m-with-7ca60146-kserve-router-scheduler-778547fbc" [e2e-llm-inference-service] has successfully progressed. [e2e-llm-inference-service] apiVersion: apps/v1 [e2e-llm-inference-service] kind: Deployment [e2e-llm-inference-service] [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1254 --- [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1255 metadata: [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-7ca60146-kserve-7cfdf49d74 [e2e-llm-inference-service] namespace: kserve-ci-e2e-test [e2e-llm-inference-service] uid: 4c32cb94-4745-4bf7-b4d3-d2be2635a15d [e2e-llm-inference-service] resourceVersion: '29816' [e2e-llm-inference-service] generation: 1 [e2e-llm-inference-service] creationTimestamp: '2026-07-07T14:29:24Z' [e2e-llm-inference-service] labels: [e2e-llm-inference-service] app.kubernetes.io/component: llminferenceservice-workload [e2e-llm-inference-service] app.kubernetes.io/name: llmisvc-model-fb-opt-125m-with-7ca60146 [e2e-llm-inference-service] app.kubernetes.io/part-of: llminferenceservice [e2e-llm-inference-service] kserve.io/component: workload [e2e-llm-inference-service] llm-d.ai/role: both [e2e-llm-inference-service] pod-template-hash: 7cfdf49d74 [e2e-llm-inference-service] annotations: [e2e-llm-inference-service] deployment.kubernetes.io/desired-replicas: '1' [e2e-llm-inference-service] deployment.kubernetes.io/max-replicas: '2' [e2e-llm-inference-service] deployment.kubernetes.io/revision: '1' [e2e-llm-inference-service] ownerReferences: [e2e-llm-inference-service] - apiVersion: apps/v1 [e2e-llm-inference-service] kind: Deployment [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-7ca60146-kserve [e2e-llm-inference-service] uid: 0e103aa1-bb8a-4024-977f-462552a80dcf [e2e-llm-inference-service] controller: true [e2e-llm-inference-service] blockOwnerDeletion: true [e2e-llm-inference-service] managedFields: [e2e-llm-inference-service] - manager: kube-controller-manager [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] apiVersion: apps/v1 [e2e-llm-inference-service] time: '2026-07-07T14:29:24Z' [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:metadata: [e2e-llm-inference-service] f:annotations: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:deployment.kubernetes.io/desired-replicas: {} [e2e-llm-inference-service] f:deployment.kubernetes.io/max-replicas: {} [e2e-llm-inference-service] f:deployment.kubernetes.io/revision: {} [e2e-llm-inference-service] f:labels: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:app.kubernetes.io/component: {} [e2e-llm-inference-service] f:app.kubernetes.io/name: {} [e2e-llm-inference-service] f:app.kubernetes.io/part-of: {} [e2e-llm-inference-service] f:kserve.io/component: {} [e2e-llm-inference-service] f:llm-d.ai/role: {} [e2e-llm-inference-service] f:pod-template-hash: {} [e2e-llm-inference-service] f:ownerReferences: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"uid":"0e103aa1-bb8a-4024-977f-462552a80dcf"}: {} [e2e-llm-inference-service] f:spec: [e2e-llm-inference-service] f:replicas: {} [e2e-llm-inference-service] f:selector: {} [e2e-llm-inference-service] f:template: [e2e-llm-inference-service] f:metadata: [e2e-llm-inference-service] f:labels: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:app.kubernetes.io/component: {} [e2e-llm-inference-service] f:app.kubernetes.io/name: {} [e2e-llm-inference-service] f:app.kubernetes.io/part-of: {} [e2e-llm-inference-service] f:kserve.io/component: {} [e2e-llm-inference-service] f:llm-d.ai/role: {} [e2e-llm-inference-service] f:pod-template-hash: {} [e2e-llm-inference-service] f:spec: [e2e-llm-inference-service] f:containers: [e2e-llm-inference-service] k:{"name":"main"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:args: {} [e2e-llm-inference-service] f:command: {} [e2e-llm-inference-service] f:env: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"name":"HF_HUB_CACHE"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"HOME"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"TORCHINDUCTOR_CACHE_DIR"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"USER"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"VLLM_CPU_KVCACHE_SPACE"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"VLLM_ENABLE_V1_MULTIPROCESSING"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"VLLM_LOGGING_LEVEL"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] f:image: {} [e2e-llm-inference-service] f:imagePullPolicy: {} [e2e-llm-inference-service] f:lifecycle: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:preStop: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:exec: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:command: {} [e2e-llm-inference-service] f:livenessProbe: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:failureThreshold: {} [e2e-llm-inference-service] f:httpGet: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:path: {} [e2e-llm-inference-service] f:port: {} [e2e-llm-inference-service] f:scheme: {} [e2e-llm-inference-service] f:periodSeconds: {} [e2e-llm-inference-service] f:successThreshold: {} [e2e-llm-inference-service] f:timeoutSeconds: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:ports: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"containerPort":8000,"protocol":"TCP"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:containerPort: {} [e2e-llm-inference-service] f:protocol: {} [e2e-llm-inference-service] f:readinessProbe: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:failureThreshold: {} [e2e-llm-inference-service] f:httpGet: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:path: {} [e2e-llm-inference-service] f:port: {} [e2e-llm-inference-service] f:scheme: {} [e2e-llm-inference-service] f:periodSeconds: {} [e2e-llm-inference-service] f:successThreshold: {} [e2e-llm-inference-service] f:timeoutSeconds: {} [e2e-llm-inference-service] f:resources: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:limits: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:cpu: {} [e2e-llm-inference-service] f:memory: {} [e2e-llm-inference-service] f:requests: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:cpu: {} [e2e-llm-inference-service] f:memory: {} [e2e-llm-inference-service] f:securityContext: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:allowPrivilegeEscalation: {} [e2e-llm-inference-service] f:capabilities: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:drop: {} [e2e-llm-inference-service] f:readOnlyRootFilesystem: {} [e2e-llm-inference-service] f:runAsNonRoot: {} [e2e-llm-inference-service] f:seccompProfile: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:type: {} [e2e-llm-inference-service] f:startupProbe: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:failureThreshold: {} [e2e-llm-inference-service] f:httpGet: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:path: {} [e2e-llm-inference-service] f:port: {} [e2e-llm-inference-service] f:scheme: {} [e2e-llm-inference-service] f:periodSeconds: {} [e2e-llm-inference-service] f:successThreshold: {} [e2e-llm-inference-service] f:timeoutSeconds: {} [e2e-llm-inference-service] f:terminationMessagePath: {} [e2e-llm-inference-service] f:terminationMessagePolicy: {} [e2e-llm-inference-service] f:volumeMounts: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"mountPath":"/dev/shm"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:mountPath: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] k:{"mountPath":"/home"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:mountPath: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] k:{"mountPath":"/mnt"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:mountPath: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:readOnly: {} [e2e-llm-inference-service] k:{"mountPath":"/models"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:mountPath: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] k:{"mountPath":"/tmp"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:mountPath: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] k:{"mountPath":"/var/run/kserve/tls"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:mountPath: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:readOnly: {} [e2e-llm-inference-service] f:dnsPolicy: {} [e2e-llm-inference-service] f:initContainers: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"name":"storage-initializer"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:args: {} [e2e-llm-inference-service] f:env: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"name":"AWS_ACCESS_KEY_ID"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:valueFrom: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:secretKeyRef: {} [e2e-llm-inference-service] k:{"name":"AWS_CA_BUNDLE"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"AWS_CA_BUNDLE_CONFIGMAP"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"AWS_ENDPOINT_URL"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"AWS_SECRET_ACCESS_KEY"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:valueFrom: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:secretKeyRef: {} [e2e-llm-inference-service] k:{"name":"CA_BUNDLE_CONFIGMAP_NAME"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"CA_BUNDLE_VOLUME_MOUNT_POINT"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"HF_HUB_ENABLE_HF_TRANSFER"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"HF_XET_HIGH_PERFORMANCE"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"HF_XET_NUM_CONCURRENT_RANGE_GETS"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"S3_ENDPOINT"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"S3_USE_HTTPS"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"S3_VERIFY_SSL"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] f:image: {} [e2e-llm-inference-service] f:imagePullPolicy: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:resources: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:limits: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:cpu: {} [e2e-llm-inference-service] f:memory: {} [e2e-llm-inference-service] f:requests: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:cpu: {} [e2e-llm-inference-service] f:memory: {} [e2e-llm-inference-service] f:terminationMessagePath: {} [e2e-llm-inference-service] f:terminationMessagePolicy: {} [e2e-llm-inference-service] f:volumeMounts: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"mountPath":"/etc/ssl/custom-certs"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:mountPath: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:readOnly: {} [e2e-llm-inference-service] k:{"mountPath":"/mnt"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:mountPath: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:restartPolicy: {} [e2e-llm-inference-service] f:schedulerName: {} [e2e-llm-inference-service] f:securityContext: {} [e2e-llm-inference-service] f:terminationGracePeriodSeconds: {} [e2e-llm-inference-service] f:volumes: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"name":"cabundle-cert"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:configMap: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:defaultMode: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] k:{"name":"dshm"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:emptyDir: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:medium: {} [e2e-llm-inference-service] f:sizeLimit: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] k:{"name":"home"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:emptyDir: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] k:{"name":"kserve-provision-location"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:emptyDir: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] k:{"name":"model-cache"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:emptyDir: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] k:{"name":"tls-certs"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:secret: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:defaultMode: {} [e2e-llm-inference-service] f:secretName: {} [e2e-llm-inference-service] k:{"name":"tmp-dir"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:emptyDir: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] - manager: kube-controller-manager [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] apiVersion: apps/v1 [e2e-llm-inference-service] time: '2026-07-07T14:30:45Z' [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:status: [e2e-llm-inference-service] f:availableReplicas: {} [e2e-llm-inference-service] f:fullyLabeledReplicas: {} [e2e-llm-inference-service] f:observedGeneration: {} [e2e-llm-inference-service] f:readyReplicas: {} [e2e-llm-inference-service] f:replicas: {} [e2e-llm-inference-service] subresource: status [e2e-llm-inference-service] spec: [e2e-llm-inference-service] replicas: 1 [e2e-llm-inference-service] selector: [e2e-llm-inference-service] matchLabels: [e2e-llm-inference-service] app.kubernetes.io/component: llminferenceservice-workload [e2e-llm-inference-service] app.kubernetes.io/name: llmisvc-model-fb-opt-125m-with-7ca60146 [e2e-llm-inference-service] app.kubernetes.io/part-of: llminferenceservice [e2e-llm-inference-service] kserve.io/component: workload [e2e-llm-inference-service] llm-d.ai/role: both [e2e-llm-inference-service] pod-template-hash: 7cfdf49d74 [e2e-llm-inference-service] template: [e2e-llm-inference-service] metadata: [e2e-llm-inference-service] labels: [e2e-llm-inference-service] app.kubernetes.io/component: llminferenceservice-workload [e2e-llm-inference-service] app.kubernetes.io/name: llmisvc-model-fb-opt-125m-with-7ca60146 [e2e-llm-inference-service] app.kubernetes.io/part-of: llminferenceservice [e2e-llm-inference-service] kserve.io/component: workload [e2e-llm-inference-service] llm-d.ai/role: both [e2e-llm-inference-service] pod-template-hash: 7cfdf49d74 [e2e-llm-inference-service] spec: [e2e-llm-inference-service] volumes: [e2e-llm-inference-service] - name: home [e2e-llm-inference-service] emptyDir: {} [e2e-llm-inference-service] - name: dshm [e2e-llm-inference-service] emptyDir: [e2e-llm-inference-service] medium: Memory [e2e-llm-inference-service] sizeLimit: 1Gi [e2e-llm-inference-service] - name: model-cache [e2e-llm-inference-service] emptyDir: {} [e2e-llm-inference-service] - name: tmp-dir [e2e-llm-inference-service] emptyDir: {} [e2e-llm-inference-service] - name: tls-certs [e2e-llm-inference-service] secret: [e2e-llm-inference-service] secretName: llmisv3e414c2ba058a022dfd694dbcbac5b51-kserve-self-signed-certs [e2e-llm-inference-service] defaultMode: 420 [e2e-llm-inference-service] - name: kserve-provision-location [e2e-llm-inference-service] emptyDir: {} [e2e-llm-inference-service] - name: cabundle-cert [e2e-llm-inference-service] configMap: [e2e-llm-inference-service] name: odh-kserve-custom-ca-bundle [e2e-llm-inference-service] defaultMode: 420 [e2e-llm-inference-service] initContainers: [e2e-llm-inference-service] - name: storage-initializer [e2e-llm-inference-service] image: quay.io/opendatahub/kserve-storage-initializer@sha256:13ad765c0f323200e1db7b45c99661287e65c5158d6b54b16139829636d20be4 [e2e-llm-inference-service] args: [e2e-llm-inference-service] - hf://facebook/opt-125m [e2e-llm-inference-service] - /mnt/models [e2e-llm-inference-service] - hf://edbeeching/opt-125m-lora [e2e-llm-inference-service] - /mnt/lora/lora-adapter-1 [e2e-llm-inference-service] env: [e2e-llm-inference-service] - name: AWS_ACCESS_KEY_ID [e2e-llm-inference-service] valueFrom: [e2e-llm-inference-service] secretKeyRef: [e2e-llm-inference-service] name: seaweedfs-s3-creds [e2e-llm-inference-service] key: AWS_ACCESS_KEY_ID [e2e-llm-inference-service] - name: AWS_SECRET_ACCESS_KEY [e2e-llm-inference-service] valueFrom: [e2e-llm-inference-service] secretKeyRef: [e2e-llm-inference-service] name: seaweedfs-s3-creds [e2e-llm-inference-service] key: AWS_SECRET_ACCESS_KEY [e2e-llm-inference-service] - name: S3_USE_HTTPS [e2e-llm-inference-service] value: '0' [e2e-llm-inference-service] - name: S3_ENDPOINT [e2e-llm-inference-service] value: s3-service.kserve:8333 [e2e-llm-inference-service] - name: AWS_ENDPOINT_URL [e2e-llm-inference-service] value: http://s3-service.kserve:8333 [e2e-llm-inference-service] - name: S3_VERIFY_SSL [e2e-llm-inference-service] value: '0' [e2e-llm-inference-service] - name: AWS_CA_BUNDLE [e2e-llm-inference-service] value: /etc/ssl/custom-certs/cabundle.crt [e2e-llm-inference-service] - name: AWS_CA_BUNDLE_CONFIGMAP [e2e-llm-inference-service] value: odh-kserve-custom-ca-bundle [e2e-llm-inference-service] - name: HF_HUB_ENABLE_HF_TRANSFER [e2e-llm-inference-service] value: '1' [e2e-llm-inference-service] - name: HF_XET_HIGH_PERFORMANCE [e2e-llm-inference-service] value: '1' [e2e-llm-inference-service] - name: HF_XET_NUM_CONCURRENT_RANGE_GETS [e2e-llm-inference-service] value: '8' [e2e-llm-inference-service] - name: CA_BUNDLE_CONFIGMAP_NAME [e2e-llm-inference-service] value: odh-kserve-custom-ca-bundle [e2e-llm-inference-service] - name: CA_BUNDLE_VOLUME_MOUNT_POINT [e2e-llm-inference-service] value: /etc/ssl/custom-certs [e2e-llm-inference-service] resources: [e2e-llm-inference-service] limits: [e2e-llm-inference-service] cpu: '1' [e2e-llm-inference-service] memory: 24Gi [e2e-llm-inference-service] requests: [e2e-llm-inference-service] cpu: 100m [e2e-llm-inference-service] memory: 100Mi [e2e-llm-inference-service] volumeMounts: [e2e-llm-inference-service] - name: kserve-provision-location [e2e-llm-inference-service] mountPath: /mnt [e2e-llm-inference-service] - name: cabundle-cert [e2e-llm-inference-service] readOnly: true [e2e-llm-inference-service] mountPath: /etc/ssl/custom-certs [e2e-llm-inference-service] terminationMessagePath: /dev/termination-log [e2e-llm-inference-service] terminationMessagePolicy: FallbackToLogsOnError [e2e-llm-inference-service] imagePullPolicy: IfNotPresent [e2e-llm-inference-service] containers: [e2e-llm-inference-service] - name: main [e2e-llm-inference-service] image: public.ecr.aws/q9t5s3a7/vllm-cpu-release-repo:v0.19.0 [e2e-llm-inference-service] command: [e2e-llm-inference-service] - /bin/bash [e2e-llm-inference-service] - -c [e2e-llm-inference-service] - "if [ -f /etc/profile.d/ibm-aiu-setup.sh ]; then\n source /etc/profile.d/ibm-aiu-setup.sh\n\ [e2e-llm-inference-service] fi\n\nif [ \"$KSERVE_INFER_ROCE\" = \"true\" ]; then\n echo \"Trying to\ [e2e-llm-inference-service] \ infer RoCE configs ... \"\n grep -H . /sys/class/infiniband/*/ports/*/gids/*\ [e2e-llm-inference-service] \ 2>/dev/null\n grep -H . /sys/class/infiniband/*/ports/*/gid_attrs/types/*\ [e2e-llm-inference-service] \ 2>/dev/null\n\n cat /proc/driver/nvidia/params\n\n KSERVE_INFER_IB_GID_INDEX_GREP=${KSERVE_INFER_IB_GID_INDEX_GREP:-\"\ [e2e-llm-inference-service] RoCE v2\"}\n\n echo \"[Infer RoCE] Discovering active HCAs ...\"\n active_hcas=()\n\ [e2e-llm-inference-service] \ # Loop through all mlx5 devices found in sysfs\n for hca_dir in /sys/class/infiniband/mlx5_*;\ [e2e-llm-inference-service] \ do\n # Ensure it's a directory before proceeding\n if [ -d \"\ [e2e-llm-inference-service] $hca_dir\" ]; then\n hca_name=$(basename \"$hca_dir\")\n \ [e2e-llm-inference-service] \ port_state_file=\"$hca_dir/ports/1/state\" # Assume port 1\n \ [e2e-llm-inference-service] \ type_file=\"$hca_dir/ports/1/gid_attrs/types/*\"\n\n echo\ [e2e-llm-inference-service] \ \"[Infer RoCE] Check if the port state file ${port_state_file} exists\ [e2e-llm-inference-service] \ and contains 'ACTIVE'\"\n if [ -f \"$port_state_file\" ] && grep\ [e2e-llm-inference-service] \ -q \"ACTIVE\" \"$port_state_file\" && grep -q \"${KSERVE_INFER_IB_GID_INDEX_GREP}\"\ [e2e-llm-inference-service] \ ${type_file} 2>/dev/null; then\n echo \"[Infer RoCE] Found\ [e2e-llm-inference-service] \ active HCA: $hca_name\"\n active_hcas+=(\"$hca_name\")\n\ [e2e-llm-inference-service] \ else\n echo \"[Infer RoCE] Skipping inactive or\ [e2e-llm-inference-service] \ down HCA: $hca_name\"\n fi\n fi\n done\n\n # Check if\ [e2e-llm-inference-service] \ we found any active HCAs\n if [ ${#active_hcas[@]} -gt 0 ]; then\n \ [e2e-llm-inference-service] \ # Join the array elements with a comma\n hca_port_pairs=()\n \ [e2e-llm-inference-service] \ for hca in \"${active_hcas[@]}\"; do\n hca_port_pairs+=(\"\ [e2e-llm-inference-service] ${hca}:1\")\n done\n\n active_hca_list=$(IFS=,; echo \"${active_hcas[*]}\"\ [e2e-llm-inference-service] )\n hca_port_pairs_list=$(IFS=,; echo \"${hca_port_pairs[*]}\")\n \ [e2e-llm-inference-service] \ echo \"[Infer RoCE] Setting active HCAs: ${active_hca_list}\"\n \ [e2e-llm-inference-service] \ export NCCL_IB_HCA=${NCCL_IB_HCA:-${active_hca_list}}\n export\ [e2e-llm-inference-service] \ NVSHMEM_HCA_LIST=${NVSHMEM_HCA_LIST:-${hca_port_pairs_list}}\n export\ [e2e-llm-inference-service] \ UCX_NET_DEVICES=${UCX_NET_DEVICES:-${hca_port_pairs_list}}\n\n echo\ [e2e-llm-inference-service] \ \"[Infer RoCE] NCCL_IB_HCA=${NCCL_IB_HCA}\"\n echo \"[Infer RoCE]\ [e2e-llm-inference-service] \ NVSHMEM_HCA_LIST=${NVSHMEM_HCA_LIST}\"\n echo \"[Infer RoCE] UCX_NET_DEVICES=${UCX_NET_DEVICES}\"\ [e2e-llm-inference-service] \n else\n echo \"[Infer RoCE] WARNING: No active RoCE HCAs found.\ [e2e-llm-inference-service] \ NCCL_IB_HCA will not be set.\"\n fi\n\n if [ ${#active_hcas[@]} -gt\ [e2e-llm-inference-service] \ 0 ]; then\n echo \"[Infer RoCE] Finding GID_INDEX for each active\ [e2e-llm-inference-service] \ HCA (SR-IOV compatible)...\"\n\n # For SR-IOV environments, find\ [e2e-llm-inference-service] \ the most common IPv4 RoCE v2 GID index across all HCAs\n declare\ [e2e-llm-inference-service] \ -A gid_index_count\n declare -A hca_gid_index\n\n for hca_name\ [e2e-llm-inference-service] \ in \"${active_hcas[@]}\"; do\n echo \"[Infer RoCE] Processing\ [e2e-llm-inference-service] \ HCA: ${hca_name}\"\n\n # Find all RoCE v2 IPv4 GIDs for this\ [e2e-llm-inference-service] \ HCA and count by index\n for tpath in /sys/class/infiniband/${hca_name}/ports/1/gid_attrs/types/*;\ [e2e-llm-inference-service] \ do\n if grep -q \"${KSERVE_INFER_IB_GID_INDEX_GREP}\" \"\ [e2e-llm-inference-service] $tpath\" 2>/dev/null; then\n idx=$(basename \"$tpath\"\ [e2e-llm-inference-service] )\n gid_file=\"/sys/class/infiniband/${hca_name}/ports/1/gids/${idx}\"\ [e2e-llm-inference-service] \n # Check for IPv4 GID (contains ffff:)\n \ [e2e-llm-inference-service] \ if [ -f \"$gid_file\" ] && grep -q \"ffff:\" \"$gid_file\"; then\n\ [e2e-llm-inference-service] \ gid_value=$(cat \"$gid_file\" 2>/dev/null || echo\ [e2e-llm-inference-service] \ \"\")\n echo \"[Infer RoCE] Found IPv4 RoCE v2 GID\ [e2e-llm-inference-service] \ for ${hca_name}: index=${idx}, gid=${gid_value}\"\n \ [e2e-llm-inference-service] \ hca_gid_index[\"${hca_name}\"]=\"${idx}\"\n gid_index_count[\"\ [e2e-llm-inference-service] ${idx}\"]=$((${gid_index_count[\"${idx}\"]} + 1))\n \ [e2e-llm-inference-service] \ break # Use first found IPv4 GID per HCA\n fi\n \ [e2e-llm-inference-service] \ fi\n done\n done\n\n # Find the most common\ [e2e-llm-inference-service] \ GID index (most likely to be consistent across nodes)\n best_gid_index=\"\ [e2e-llm-inference-service] \"\n max_count=0\n for idx in \"${!gid_index_count[@]}\"; do\n\ [e2e-llm-inference-service] \ count=${gid_index_count[\"${idx}\"]}\n echo \"[Infer\ [e2e-llm-inference-service] \ RoCE] GID_INDEX ${idx} found on ${count} HCAs\"\n if [ $count\ [e2e-llm-inference-service] \ -gt $max_count ]; then\n max_count=$count\n \ [e2e-llm-inference-service] \ best_gid_index=\"$idx\"\n fi\n done\n\n # Use deterministic\ [e2e-llm-inference-service] \ fallback if tied - prefer index 3 (SR-IOV standard)\n if [ ${#gid_index_count[@]}\ [e2e-llm-inference-service] \ -gt 1 ]; then\n echo \"[Infer RoCE] Multiple GID indices found,\ [e2e-llm-inference-service] \ selecting most common: ${best_gid_index}\"\n # If there's a tie,\ [e2e-llm-inference-service] \ prefer index 3 as it's most common in SR-IOV setups\n if [ -n\ [e2e-llm-inference-service] \ \"${gid_index_count['3']}\" ] && [ \"${gid_index_count['3']}\" -eq \"\ [e2e-llm-inference-service] $max_count\" ]; then\n best_gid_index=\"3\"\n \ [e2e-llm-inference-service] \ echo \"[Infer RoCE] Using deterministic fallback: GID_INDEX=3 (SR-IOV\ [e2e-llm-inference-service] \ standard)\"\n fi\n fi\n\n # Check if GID_INDEX is already\ [e2e-llm-inference-service] \ set via environment variables\n if [ -n \"${NCCL_IB_GID_INDEX}\"\ [e2e-llm-inference-service] \ ]; then\n echo \"[Infer RoCE] Using pre-configured NCCL_IB_GID_INDEX=${NCCL_IB_GID_INDEX}\ [e2e-llm-inference-service] \ from environment\"\n export NVSHMEM_IB_GID_INDEX=${NVSHMEM_IB_GID_INDEX:-$NCCL_IB_GID_INDEX}\n\ [e2e-llm-inference-service] \ export UCX_IB_GID_INDEX=${UCX_IB_GID_INDEX:-$NCCL_IB_GID_INDEX}\n\ [e2e-llm-inference-service] \ echo \"[Infer RoCE] Using pre-configured GID_INDEX=${NCCL_IB_GID_INDEX}\ [e2e-llm-inference-service] \ for NCCL, NVSHMEM, and UCX\"\n elif [ -n \"$best_gid_index\" ]; then\n\ [e2e-llm-inference-service] \ echo \"[Infer RoCE] Selected GID_INDEX: ${best_gid_index} (found\ [e2e-llm-inference-service] \ on ${max_count} HCAs)\"\n\n export NCCL_IB_GID_INDEX=${NCCL_IB_GID_INDEX:-$best_gid_index}\n\ [e2e-llm-inference-service] \ export NVSHMEM_IB_GID_INDEX=${NVSHMEM_IB_GID_INDEX:-$best_gid_index}\n\ [e2e-llm-inference-service] \ export UCX_IB_GID_INDEX=${UCX_IB_GID_INDEX:-$best_gid_index}\n\ [e2e-llm-inference-service] \n echo \"[Infer RoCE] Exported GID_INDEX=${best_gid_index} for\ [e2e-llm-inference-service] \ NCCL, NVSHMEM, and UCX\"\n else\n echo \"[Infer RoCE] ERROR:\ [e2e-llm-inference-service] \ No valid IPv4 ${KSERVE_INFER_IB_GID_INDEX_GREP} GID_INDEX found on any\ [e2e-llm-inference-service] \ HCA.\"\n fi\n else\n echo \"[Infer RoCE] No active HCAs found,\ [e2e-llm-inference-service] \ skipping GID_INDEX inference.\"\n fi\nfi\n\n# --disable-access-log-for-endpoints\ [e2e-llm-inference-service] \ landed in vLLM 0.16.0 (vllm-project/vllm#30011).\n# Older versions still\ [e2e-llm-inference-service] \ need the blanket --disable-uvicorn-access-log.\nACCESS_LOG_ARGS=\"--disable-uvicorn-access-log\"\ [e2e-llm-inference-service] \nVLLM_VERSION=$(vllm --version 2>/dev/null | tail -1 | awk '{print $NF}')\n\ [e2e-llm-inference-service] echo \"[access-log-detect] vllm version='${VLLM_VERSION}'\"\nif [[ \"$VLLM_VERSION\"\ [e2e-llm-inference-service] \ =~ ^[0-9]+\\.[0-9]+ ]] && [ \"$(printf '%s\\n%s\\n' \"0.16.0\" \"${VLLM_VERSION}\"\ [e2e-llm-inference-service] \ | sort -V | head -1)\" = \"0.16.0\" ]; then\n ACCESS_LOG_ARGS=\"--disable-access-log-for-endpoints\ [e2e-llm-inference-service] \ /health,/metrics,/ping\"\nfi\necho \"[access-log-detect] selected ACCESS_LOG_ARGS='${ACCESS_LOG_ARGS}'\"\ [e2e-llm-inference-service] \n\n# --shutdown-timeout landed in vLLM 0.18.0 (vllm-project/vllm#36666).\n\ [e2e-llm-inference-service] SHUTDOWN_TIMEOUT_ARGS=\"\"\nif [[ \"$VLLM_VERSION\" =~ ^[0-9]+\\.[0-9]+\ [e2e-llm-inference-service] \ ]] && [ \"$(printf '%s\\n%s\\n' \"0.18.0\" \"${VLLM_VERSION}\" | sort\ [e2e-llm-inference-service] \ -V | head -1)\" = \"0.18.0\" ]; then\n SHUTDOWN_TIMEOUT_ARGS=\"--shutdown-timeout\ [e2e-llm-inference-service] \ 40\"\nfi\n\neval \"exec vllm serve /mnt/models \\\n --served-model-name\ [e2e-llm-inference-service] \ \"facebook/opt-125m\" \"publishers/kserve-ci-e2e-test/models/facebook/opt-125m\"\ [e2e-llm-inference-service] \ \\\n --port 8000 \\\n ${ACCESS_LOG_ARGS} \\\n ${SHUTDOWN_TIMEOUT_ARGS}\ [e2e-llm-inference-service] \ \\\n --enable-ssl-refresh \\\n --ssl-certfile /var/run/kserve/tls/tls.crt\ [e2e-llm-inference-service] \ \\\n --ssl-keyfile /var/run/kserve/tls/tls.key \\\n ${VLLM_ADDITIONAL_ARGS}\ [e2e-llm-inference-service] \ \\\n $@\"" [e2e-llm-inference-service] - -- [e2e-llm-inference-service] args: [e2e-llm-inference-service] - --enable-lora [e2e-llm-inference-service] - --lora-modules [e2e-llm-inference-service] - '''{"name":"lora-adapter-1","path":"/mnt/lora/lora-adapter-1"}''' [e2e-llm-inference-service] - '''{"name":"publishers/kserve-ci-e2e-test/models/lora-adapter-1","path":"/mnt/lora/lora-adapter-1"}''' [e2e-llm-inference-service] ports: [e2e-llm-inference-service] - containerPort: 8000 [e2e-llm-inference-service] protocol: TCP [e2e-llm-inference-service] env: [e2e-llm-inference-service] - name: HOME [e2e-llm-inference-service] value: /home [e2e-llm-inference-service] - name: VLLM_LOGGING_LEVEL [e2e-llm-inference-service] value: DEBUG [e2e-llm-inference-service] - name: VLLM_CPU_KVCACHE_SPACE [e2e-llm-inference-service] value: '1' [e2e-llm-inference-service] - name: VLLM_ENABLE_V1_MULTIPROCESSING [e2e-llm-inference-service] value: '0' [e2e-llm-inference-service] - name: USER [e2e-llm-inference-service] value: nonroot [e2e-llm-inference-service] - name: TORCHINDUCTOR_CACHE_DIR [e2e-llm-inference-service] value: /tmp/torchinductor-cache [e2e-llm-inference-service] - name: HF_HUB_CACHE [e2e-llm-inference-service] value: /models [e2e-llm-inference-service] resources: [e2e-llm-inference-service] limits: [e2e-llm-inference-service] cpu: '2' [e2e-llm-inference-service] memory: 7Gi [e2e-llm-inference-service] requests: [e2e-llm-inference-service] cpu: 200m [e2e-llm-inference-service] memory: 2Gi [e2e-llm-inference-service] volumeMounts: [e2e-llm-inference-service] - name: home [e2e-llm-inference-service] mountPath: /home [e2e-llm-inference-service] - name: tmp-dir [e2e-llm-inference-service] mountPath: /tmp [e2e-llm-inference-service] - name: dshm [e2e-llm-inference-service] mountPath: /dev/shm [e2e-llm-inference-service] - name: model-cache [e2e-llm-inference-service] mountPath: /models [e2e-llm-inference-service] - name: tls-certs [e2e-llm-inference-service] readOnly: true [e2e-llm-inference-service] mountPath: /var/run/kserve/tls [e2e-llm-inference-service] - name: kserve-provision-location [e2e-llm-inference-service] readOnly: true [e2e-llm-inference-service] mountPath: /mnt [e2e-llm-inference-service] livenessProbe: [e2e-llm-inference-service] httpGet: [e2e-llm-inference-service] path: /health [e2e-llm-inference-service] port: 8000 [e2e-llm-inference-service] scheme: HTTPS [e2e-llm-inference-service] timeoutSeconds: 1 [e2e-llm-inference-service] periodSeconds: 10 [e2e-llm-inference-service] successThreshold: 1 [e2e-llm-inference-service] failureThreshold: 10 [e2e-llm-inference-service] readinessProbe: [e2e-llm-inference-service] httpGet: [e2e-llm-inference-service] path: /health [e2e-llm-inference-service] port: 8000 [e2e-llm-inference-service] scheme: HTTPS [e2e-llm-inference-service] timeoutSeconds: 1 [e2e-llm-inference-service] periodSeconds: 1 [e2e-llm-inference-service] successThreshold: 1 [e2e-llm-inference-service] failureThreshold: 2 [e2e-llm-inference-service] startupProbe: [e2e-llm-inference-service] httpGet: [e2e-llm-inference-service] path: /health [e2e-llm-inference-service] port: 8000 [e2e-llm-inference-service] scheme: HTTPS [e2e-llm-inference-service] timeoutSeconds: 1 [e2e-llm-inference-service] periodSeconds: 10 [e2e-llm-inference-service] successThreshold: 1 [e2e-llm-inference-service] failureThreshold: 60 [e2e-llm-inference-service] lifecycle: [e2e-llm-inference-service] preStop: [e2e-llm-inference-service] exec: [e2e-llm-inference-service] command: [e2e-llm-inference-service] - /bin/sleep [e2e-llm-inference-service] - '15' [e2e-llm-inference-service] terminationMessagePath: /dev/termination-log [e2e-llm-inference-service] terminationMessagePolicy: FallbackToLogsOnError [e2e-llm-inference-service] imagePullPolicy: IfNotPresent [e2e-llm-inference-service] securityContext: [e2e-llm-inference-service] capabilities: [e2e-llm-inference-service] drop: [e2e-llm-inference-service] - ALL [e2e-llm-inference-service] runAsNonRoot: true [e2e-llm-inference-service] readOnlyRootFilesystem: true [e2e-llm-inference-service] allowPrivilegeEscalation: false [e2e-llm-inference-service] seccompProfile: [e2e-llm-inference-service] type: RuntimeDefault [e2e-llm-inference-service] restartPolicy: Always [e2e-llm-inference-service] terminationGracePeriodSeconds: 60 [e2e-llm-inference-service] dnsPolicy: ClusterFirst [e2e-llm-inference-service] securityContext: {} [e2e-llm-inference-service] schedulerName: default-scheduler [e2e-llm-inference-service] status: [e2e-llm-inference-service] replicas: 1 [e2e-llm-inference-service] fullyLabeledReplicas: 1 [e2e-llm-inference-service] readyReplicas: 1 [e2e-llm-inference-service] availableReplicas: 1 [e2e-llm-inference-service] observedGeneration: 1 [e2e-llm-inference-service] apiVersion: apps/v1 [e2e-llm-inference-service] kind: ReplicaSet [e2e-llm-inference-service] [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1254 --- [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1255 metadata: [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-7ca60146-kserve-router-scheduler-778547fbc [e2e-llm-inference-service] namespace: kserve-ci-e2e-test [e2e-llm-inference-service] uid: e375700c-4e9f-4122-abe2-116603aa60cd [e2e-llm-inference-service] resourceVersion: '28938' [e2e-llm-inference-service] generation: 1 [e2e-llm-inference-service] creationTimestamp: '2026-07-07T14:29:24Z' [e2e-llm-inference-service] labels: [e2e-llm-inference-service] app.kubernetes.io/component: llminferenceservice-router-scheduler [e2e-llm-inference-service] app.kubernetes.io/name: llmisvc-model-fb-opt-125m-with-7ca60146 [e2e-llm-inference-service] app.kubernetes.io/part-of: llminferenceservice [e2e-llm-inference-service] pod-template-hash: 778547fbc [e2e-llm-inference-service] annotations: [e2e-llm-inference-service] deployment.kubernetes.io/desired-replicas: '1' [e2e-llm-inference-service] deployment.kubernetes.io/max-replicas: '1' [e2e-llm-inference-service] deployment.kubernetes.io/revision: '1' [e2e-llm-inference-service] ownerReferences: [e2e-llm-inference-service] - apiVersion: apps/v1 [e2e-llm-inference-service] kind: Deployment [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-7ca60146-kserve-router-scheduler [e2e-llm-inference-service] uid: ef37f35e-12b2-4fac-8d62-d818466fcbff [e2e-llm-inference-service] controller: true [e2e-llm-inference-service] blockOwnerDeletion: true [e2e-llm-inference-service] managedFields: [e2e-llm-inference-service] - manager: kube-controller-manager [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] apiVersion: apps/v1 [e2e-llm-inference-service] time: '2026-07-07T14:29:24Z' [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:metadata: [e2e-llm-inference-service] f:annotations: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:deployment.kubernetes.io/desired-replicas: {} [e2e-llm-inference-service] f:deployment.kubernetes.io/max-replicas: {} [e2e-llm-inference-service] f:deployment.kubernetes.io/revision: {} [e2e-llm-inference-service] f:labels: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:app.kubernetes.io/component: {} [e2e-llm-inference-service] f:app.kubernetes.io/name: {} [e2e-llm-inference-service] f:app.kubernetes.io/part-of: {} [e2e-llm-inference-service] f:pod-template-hash: {} [e2e-llm-inference-service] f:ownerReferences: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"uid":"ef37f35e-12b2-4fac-8d62-d818466fcbff"}: {} [e2e-llm-inference-service] f:spec: [e2e-llm-inference-service] f:replicas: {} [e2e-llm-inference-service] f:selector: {} [e2e-llm-inference-service] f:template: [e2e-llm-inference-service] f:metadata: [e2e-llm-inference-service] f:annotations: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:app.kubernetes.io/version: {} [e2e-llm-inference-service] f:certificates.kserve.io/expiration-v2: {} [e2e-llm-inference-service] f:labels: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:app.kubernetes.io/component: {} [e2e-llm-inference-service] f:app.kubernetes.io/name: {} [e2e-llm-inference-service] f:app.kubernetes.io/part-of: {} [e2e-llm-inference-service] f:pod-template-hash: {} [e2e-llm-inference-service] f:spec: [e2e-llm-inference-service] f:containers: [e2e-llm-inference-service] k:{"name":"main"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:args: {} [e2e-llm-inference-service] f:command: {} [e2e-llm-inference-service] f:env: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"name":"SSL_CERT_DIR"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] f:image: {} [e2e-llm-inference-service] f:imagePullPolicy: {} [e2e-llm-inference-service] f:lifecycle: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:preStop: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:exec: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:command: {} [e2e-llm-inference-service] f:livenessProbe: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:failureThreshold: {} [e2e-llm-inference-service] f:grpc: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:port: {} [e2e-llm-inference-service] f:service: {} [e2e-llm-inference-service] f:initialDelaySeconds: {} [e2e-llm-inference-service] f:periodSeconds: {} [e2e-llm-inference-service] f:successThreshold: {} [e2e-llm-inference-service] f:timeoutSeconds: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:ports: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"containerPort":5557,"protocol":"TCP"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:containerPort: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:protocol: {} [e2e-llm-inference-service] k:{"containerPort":9002,"protocol":"TCP"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:containerPort: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:protocol: {} [e2e-llm-inference-service] k:{"containerPort":9003,"protocol":"TCP"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:containerPort: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:protocol: {} [e2e-llm-inference-service] k:{"containerPort":9090,"protocol":"TCP"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:containerPort: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:protocol: {} [e2e-llm-inference-service] f:readinessProbe: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:failureThreshold: {} [e2e-llm-inference-service] f:grpc: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:port: {} [e2e-llm-inference-service] f:service: {} [e2e-llm-inference-service] f:initialDelaySeconds: {} [e2e-llm-inference-service] f:periodSeconds: {} [e2e-llm-inference-service] f:successThreshold: {} [e2e-llm-inference-service] f:timeoutSeconds: {} [e2e-llm-inference-service] f:resources: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:requests: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:cpu: {} [e2e-llm-inference-service] f:memory: {} [e2e-llm-inference-service] f:securityContext: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:allowPrivilegeEscalation: {} [e2e-llm-inference-service] f:capabilities: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:drop: {} [e2e-llm-inference-service] f:readOnlyRootFilesystem: {} [e2e-llm-inference-service] f:runAsNonRoot: {} [e2e-llm-inference-service] f:seccompProfile: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:type: {} [e2e-llm-inference-service] f:terminationMessagePath: {} [e2e-llm-inference-service] f:terminationMessagePolicy: {} [e2e-llm-inference-service] f:volumeMounts: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"mountPath":"/tmp/tokenizer"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:mountPath: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] k:{"mountPath":"/var/run/kserve/tls"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:mountPath: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:readOnly: {} [e2e-llm-inference-service] f:dnsPolicy: {} [e2e-llm-inference-service] f:initContainers: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"name":"storage-initializer"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:args: {} [e2e-llm-inference-service] f:env: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"name":"AWS_ACCESS_KEY_ID"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:valueFrom: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:secretKeyRef: {} [e2e-llm-inference-service] k:{"name":"AWS_CA_BUNDLE"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"AWS_CA_BUNDLE_CONFIGMAP"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"AWS_ENDPOINT_URL"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"AWS_SECRET_ACCESS_KEY"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:valueFrom: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:secretKeyRef: {} [e2e-llm-inference-service] k:{"name":"HF_HUB_ENABLE_HF_TRANSFER"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"HF_XET_HIGH_PERFORMANCE"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"HF_XET_NUM_CONCURRENT_RANGE_GETS"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"S3_ENDPOINT"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"S3_USE_HTTPS"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"S3_VERIFY_SSL"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"STORAGE_ALLOW_PATTERNS"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] f:image: {} [e2e-llm-inference-service] f:imagePullPolicy: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:resources: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:limits: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:cpu: {} [e2e-llm-inference-service] f:memory: {} [e2e-llm-inference-service] f:requests: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:cpu: {} [e2e-llm-inference-service] f:memory: {} [e2e-llm-inference-service] f:terminationMessagePath: {} [e2e-llm-inference-service] f:terminationMessagePolicy: {} [e2e-llm-inference-service] f:volumeMounts: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"mountPath":"/mnt/models"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:mountPath: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:restartPolicy: {} [e2e-llm-inference-service] f:schedulerName: {} [e2e-llm-inference-service] f:securityContext: {} [e2e-llm-inference-service] f:serviceAccount: {} [e2e-llm-inference-service] f:serviceAccountName: {} [e2e-llm-inference-service] f:terminationGracePeriodSeconds: {} [e2e-llm-inference-service] f:volumes: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"name":"kserve-provision-location"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:emptyDir: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] k:{"name":"tls-certs"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:secret: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:defaultMode: {} [e2e-llm-inference-service] f:secretName: {} [e2e-llm-inference-service] k:{"name":"tokenizer-cache"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:emptyDir: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] k:{"name":"tokenizer-tmp"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:emptyDir: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] k:{"name":"tokenizer-uds"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:emptyDir: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] - manager: kube-controller-manager [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] apiVersion: apps/v1 [e2e-llm-inference-service] time: '2026-07-07T14:29:59Z' [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:status: [e2e-llm-inference-service] f:availableReplicas: {} [e2e-llm-inference-service] f:fullyLabeledReplicas: {} [e2e-llm-inference-service] f:observedGeneration: {} [e2e-llm-inference-service] f:readyReplicas: {} [e2e-llm-inference-service] f:replicas: {} [e2e-llm-inference-service] subresource: status [e2e-llm-inference-service] spec: [e2e-llm-inference-service] replicas: 1 [e2e-llm-inference-service] selector: [e2e-llm-inference-service] matchLabels: [e2e-llm-inference-service] app.kubernetes.io/component: llminferenceservice-router-scheduler [e2e-llm-inference-service] app.kubernetes.io/name: llmisvc-model-fb-opt-125m-with-7ca60146 [e2e-llm-inference-service] app.kubernetes.io/part-of: llminferenceservice [e2e-llm-inference-service] pod-template-hash: 778547fbc [e2e-llm-inference-service] template: [e2e-llm-inference-service] metadata: [e2e-llm-inference-service] labels: [e2e-llm-inference-service] app.kubernetes.io/component: llminferenceservice-router-scheduler [e2e-llm-inference-service] app.kubernetes.io/name: llmisvc-model-fb-opt-125m-with-7ca60146 [e2e-llm-inference-service] app.kubernetes.io/part-of: llminferenceservice [e2e-llm-inference-service] pod-template-hash: 778547fbc [e2e-llm-inference-service] annotations: [e2e-llm-inference-service] app.kubernetes.io/version: 0.9.0 [e2e-llm-inference-service] certificates.kserve.io/expiration-v2: 'true' [e2e-llm-inference-service] spec: [e2e-llm-inference-service] volumes: [e2e-llm-inference-service] - name: tls-certs [e2e-llm-inference-service] secret: [e2e-llm-inference-service] secretName: llmisv3e414c2ba058a022dfd694dbcbac5b51-kserve-self-signed-certs [e2e-llm-inference-service] defaultMode: 420 [e2e-llm-inference-service] - name: tokenizer-uds [e2e-llm-inference-service] emptyDir: {} [e2e-llm-inference-service] - name: tokenizer-tmp [e2e-llm-inference-service] emptyDir: {} [e2e-llm-inference-service] - name: tokenizer-cache [e2e-llm-inference-service] emptyDir: {} [e2e-llm-inference-service] - name: kserve-provision-location [e2e-llm-inference-service] emptyDir: {} [e2e-llm-inference-service] initContainers: [e2e-llm-inference-service] - name: storage-initializer [e2e-llm-inference-service] image: quay.io/opendatahub/kserve-storage-initializer@sha256:13ad765c0f323200e1db7b45c99661287e65c5158d6b54b16139829636d20be4 [e2e-llm-inference-service] args: [e2e-llm-inference-service] - hf://facebook/opt-125m [e2e-llm-inference-service] - /mnt/models [e2e-llm-inference-service] env: [e2e-llm-inference-service] - name: AWS_ACCESS_KEY_ID [e2e-llm-inference-service] valueFrom: [e2e-llm-inference-service] secretKeyRef: [e2e-llm-inference-service] name: seaweedfs-s3-creds [e2e-llm-inference-service] key: AWS_ACCESS_KEY_ID [e2e-llm-inference-service] - name: AWS_SECRET_ACCESS_KEY [e2e-llm-inference-service] valueFrom: [e2e-llm-inference-service] secretKeyRef: [e2e-llm-inference-service] name: seaweedfs-s3-creds [e2e-llm-inference-service] key: AWS_SECRET_ACCESS_KEY [e2e-llm-inference-service] - name: S3_USE_HTTPS [e2e-llm-inference-service] value: '0' [e2e-llm-inference-service] - name: S3_ENDPOINT [e2e-llm-inference-service] value: s3-service.kserve:8333 [e2e-llm-inference-service] - name: AWS_ENDPOINT_URL [e2e-llm-inference-service] value: http://s3-service.kserve:8333 [e2e-llm-inference-service] - name: S3_VERIFY_SSL [e2e-llm-inference-service] value: '0' [e2e-llm-inference-service] - name: AWS_CA_BUNDLE [e2e-llm-inference-service] value: /etc/ssl/custom-certs/cabundle.crt [e2e-llm-inference-service] - name: AWS_CA_BUNDLE_CONFIGMAP [e2e-llm-inference-service] value: odh-kserve-custom-ca-bundle [e2e-llm-inference-service] - name: HF_HUB_ENABLE_HF_TRANSFER [e2e-llm-inference-service] value: '1' [e2e-llm-inference-service] - name: HF_XET_HIGH_PERFORMANCE [e2e-llm-inference-service] value: '1' [e2e-llm-inference-service] - name: HF_XET_NUM_CONCURRENT_RANGE_GETS [e2e-llm-inference-service] value: '8' [e2e-llm-inference-service] - name: STORAGE_ALLOW_PATTERNS [e2e-llm-inference-service] value: '["tokenizer.json", "tokenizer_config.json", "special_tokens_map.json", [e2e-llm-inference-service] "vocab.json", "merges.txt", "config.json", "generation_config.json"]' [e2e-llm-inference-service] resources: [e2e-llm-inference-service] limits: [e2e-llm-inference-service] cpu: '1' [e2e-llm-inference-service] memory: 24Gi [e2e-llm-inference-service] requests: [e2e-llm-inference-service] cpu: 100m [e2e-llm-inference-service] memory: 100Mi [e2e-llm-inference-service] volumeMounts: [e2e-llm-inference-service] - name: kserve-provision-location [e2e-llm-inference-service] mountPath: /mnt/models [e2e-llm-inference-service] terminationMessagePath: /dev/termination-log [e2e-llm-inference-service] terminationMessagePolicy: FallbackToLogsOnError [e2e-llm-inference-service] imagePullPolicy: IfNotPresent [e2e-llm-inference-service] containers: [e2e-llm-inference-service] - name: main [e2e-llm-inference-service] image: ghcr.io/llm-d/llm-d-router-endpoint-picker:v0.9.0-rc.2 [e2e-llm-inference-service] command: [e2e-llm-inference-service] - /app/epp [e2e-llm-inference-service] - --pool-name [e2e-llm-inference-service] - llmisvc-model-fb-opt-125m-with-7ca60146-inference-pool [e2e-llm-inference-service] - --pool-namespace [e2e-llm-inference-service] - kserve-ci-e2e-test [e2e-llm-inference-service] - --zap-encoder [e2e-llm-inference-service] - json [e2e-llm-inference-service] - --grpc-port [e2e-llm-inference-service] - '9002' [e2e-llm-inference-service] - --grpc-health-port [e2e-llm-inference-service] - '9003' [e2e-llm-inference-service] - --enable-cert-reload=true [e2e-llm-inference-service] - --secure-serving=true [e2e-llm-inference-service] - --model-server-metrics-scheme=https [e2e-llm-inference-service] - --cert-path=/var/run/kserve/tls [e2e-llm-inference-service] args: [e2e-llm-inference-service] - --config-text [e2e-llm-inference-service] - "apiVersion: inference.networking.x-k8s.io/v1alpha1\nkind: EndpointPickerConfig\n\ [e2e-llm-inference-service] plugins:\n- type: single-profile-handler\n- type: queue-scorer\n- type:\ [e2e-llm-inference-service] \ prefix-cache-scorer\n- type: max-score-picker\n- type: lora-affinity-scorer\n\ [e2e-llm-inference-service] schedulingProfiles:\n- name: default\n plugins:\n - pluginRef: lora-affinity-scorer\n\ [e2e-llm-inference-service] \ weight: 4\n - pluginRef: queue-scorer\n weight: 2\n - pluginRef:\ [e2e-llm-inference-service] \ prefix-cache-scorer\n weight: 3\n - pluginRef: max-score-picker\n" [e2e-llm-inference-service] ports: [e2e-llm-inference-service] - name: grpc [e2e-llm-inference-service] containerPort: 9002 [e2e-llm-inference-service] protocol: TCP [e2e-llm-inference-service] - name: grpc-health [e2e-llm-inference-service] containerPort: 9003 [e2e-llm-inference-service] protocol: TCP [e2e-llm-inference-service] - name: metrics [e2e-llm-inference-service] containerPort: 9090 [e2e-llm-inference-service] protocol: TCP [e2e-llm-inference-service] - name: zmq [e2e-llm-inference-service] containerPort: 5557 [e2e-llm-inference-service] protocol: TCP [e2e-llm-inference-service] env: [e2e-llm-inference-service] - name: SSL_CERT_DIR [e2e-llm-inference-service] value: /var/run/kserve/tls:/var/run/secrets/kubernetes.io/serviceaccount:/etc/pki/tls/certs [e2e-llm-inference-service] resources: [e2e-llm-inference-service] requests: [e2e-llm-inference-service] cpu: 256m [e2e-llm-inference-service] memory: 500Mi [e2e-llm-inference-service] volumeMounts: [e2e-llm-inference-service] - name: tls-certs [e2e-llm-inference-service] readOnly: true [e2e-llm-inference-service] mountPath: /var/run/kserve/tls [e2e-llm-inference-service] - name: tokenizer-uds [e2e-llm-inference-service] mountPath: /tmp/tokenizer [e2e-llm-inference-service] livenessProbe: [e2e-llm-inference-service] grpc: [e2e-llm-inference-service] port: 9003 [e2e-llm-inference-service] service: liveness [e2e-llm-inference-service] initialDelaySeconds: 5 [e2e-llm-inference-service] timeoutSeconds: 1 [e2e-llm-inference-service] periodSeconds: 10 [e2e-llm-inference-service] successThreshold: 1 [e2e-llm-inference-service] failureThreshold: 3 [e2e-llm-inference-service] readinessProbe: [e2e-llm-inference-service] grpc: [e2e-llm-inference-service] port: 9003 [e2e-llm-inference-service] service: readiness [e2e-llm-inference-service] initialDelaySeconds: 30 [e2e-llm-inference-service] timeoutSeconds: 1 [e2e-llm-inference-service] periodSeconds: 10 [e2e-llm-inference-service] successThreshold: 1 [e2e-llm-inference-service] failureThreshold: 3 [e2e-llm-inference-service] lifecycle: [e2e-llm-inference-service] preStop: [e2e-llm-inference-service] exec: [e2e-llm-inference-service] command: [e2e-llm-inference-service] - /bin/sleep [e2e-llm-inference-service] - '15' [e2e-llm-inference-service] terminationMessagePath: /dev/termination-log [e2e-llm-inference-service] terminationMessagePolicy: FallbackToLogsOnError [e2e-llm-inference-service] imagePullPolicy: IfNotPresent [e2e-llm-inference-service] securityContext: [e2e-llm-inference-service] capabilities: [e2e-llm-inference-service] drop: [e2e-llm-inference-service] - ALL [e2e-llm-inference-service] runAsNonRoot: true [e2e-llm-inference-service] readOnlyRootFilesystem: true [e2e-llm-inference-service] allowPrivilegeEscalation: false [e2e-llm-inference-service] seccompProfile: [e2e-llm-inference-service] type: RuntimeDefault [e2e-llm-inference-service] restartPolicy: Always [e2e-llm-inference-service] terminationGracePeriodSeconds: 60 [e2e-llm-inference-service] dnsPolicy: ClusterFirst [e2e-llm-inference-service] serviceAccountName: llmisvc-model-fb-opt-125m-with-7ca60146-epp-sa [e2e-llm-inference-service] serviceAccount: llmisvc-model-fb-opt-125m-with-7ca60146-epp-sa [e2e-llm-inference-service] securityContext: {} [e2e-llm-inference-service] schedulerName: default-scheduler [e2e-llm-inference-service] status: [e2e-llm-inference-service] replicas: 1 [e2e-llm-inference-service] fullyLabeledReplicas: 1 [e2e-llm-inference-service] readyReplicas: 1 [e2e-llm-inference-service] availableReplicas: 1 [e2e-llm-inference-service] observedGeneration: 1 [e2e-llm-inference-service] apiVersion: apps/v1 [e2e-llm-inference-service] kind: ReplicaSet [e2e-llm-inference-service] [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1254 --- [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1255 metadata: [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-7ca60146-epp-rb [e2e-llm-inference-service] namespace: kserve-ci-e2e-test [e2e-llm-inference-service] uid: 5dec14fd-be15-4f60-ade8-0954fd32be81 [e2e-llm-inference-service] resourceVersion: '28241' [e2e-llm-inference-service] creationTimestamp: '2026-07-07T14:29:24Z' [e2e-llm-inference-service] labels: [e2e-llm-inference-service] app.kubernetes.io/component: llminferenceservice-router-scheduler [e2e-llm-inference-service] app.kubernetes.io/name: llmisvc-model-fb-opt-125m-with-7ca60146 [e2e-llm-inference-service] app.kubernetes.io/part-of: llminferenceservice [e2e-llm-inference-service] ownerReferences: [e2e-llm-inference-service] - apiVersion: serving.kserve.io/v1alpha2 [e2e-llm-inference-service] kind: LLMInferenceService [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-7ca60146 [e2e-llm-inference-service] uid: e15cfc59-e4a1-4876-b43b-85ce3b1b239b [e2e-llm-inference-service] controller: true [e2e-llm-inference-service] blockOwnerDeletion: true [e2e-llm-inference-service] managedFields: [e2e-llm-inference-service] - manager: manager [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] apiVersion: rbac.authorization.k8s.io/v1 [e2e-llm-inference-service] time: '2026-07-07T14:29:24Z' [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:metadata: [e2e-llm-inference-service] f:labels: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:app.kubernetes.io/component: {} [e2e-llm-inference-service] f:app.kubernetes.io/name: {} [e2e-llm-inference-service] f:app.kubernetes.io/part-of: {} [e2e-llm-inference-service] f:ownerReferences: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"uid":"e15cfc59-e4a1-4876-b43b-85ce3b1b239b"}: {} [e2e-llm-inference-service] f:roleRef: {} [e2e-llm-inference-service] f:subjects: {} [e2e-llm-inference-service] subjects: [e2e-llm-inference-service] - kind: ServiceAccount [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-7ca60146-epp-sa [e2e-llm-inference-service] namespace: kserve-ci-e2e-test [e2e-llm-inference-service] roleRef: [e2e-llm-inference-service] apiGroup: rbac.authorization.k8s.io [e2e-llm-inference-service] kind: Role [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-7ca60146-epp-role [e2e-llm-inference-service] apiVersion: rbac.authorization.k8s.io/v1 [e2e-llm-inference-service] kind: RoleBinding [e2e-llm-inference-service] [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1254 --- [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1255 metadata: [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-7ca60146-epp-role [e2e-llm-inference-service] namespace: kserve-ci-e2e-test [e2e-llm-inference-service] uid: 13c554f4-7989-49ec-8acf-0b4132cbd414 [e2e-llm-inference-service] resourceVersion: '28236' [e2e-llm-inference-service] creationTimestamp: '2026-07-07T14:29:24Z' [e2e-llm-inference-service] labels: [e2e-llm-inference-service] app.kubernetes.io/component: llminferenceservice-router-scheduler [e2e-llm-inference-service] app.kubernetes.io/name: llmisvc-model-fb-opt-125m-with-7ca60146 [e2e-llm-inference-service] app.kubernetes.io/part-of: llminferenceservice [e2e-llm-inference-service] ownerReferences: [e2e-llm-inference-service] - apiVersion: serving.kserve.io/v1alpha2 [e2e-llm-inference-service] kind: LLMInferenceService [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-7ca60146 [e2e-llm-inference-service] uid: e15cfc59-e4a1-4876-b43b-85ce3b1b239b [e2e-llm-inference-service] controller: true [e2e-llm-inference-service] blockOwnerDeletion: true [e2e-llm-inference-service] managedFields: [e2e-llm-inference-service] - manager: manager [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] apiVersion: rbac.authorization.k8s.io/v1 [e2e-llm-inference-service] time: '2026-07-07T14:29:24Z' [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:metadata: [e2e-llm-inference-service] f:labels: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:app.kubernetes.io/component: {} [e2e-llm-inference-service] f:app.kubernetes.io/name: {} [e2e-llm-inference-service] f:app.kubernetes.io/part-of: {} [e2e-llm-inference-service] f:ownerReferences: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"uid":"e15cfc59-e4a1-4876-b43b-85ce3b1b239b"}: {} [e2e-llm-inference-service] f:rules: {} [e2e-llm-inference-service] rules: [e2e-llm-inference-service] - verbs: [e2e-llm-inference-service] - get [e2e-llm-inference-service] - list [e2e-llm-inference-service] - watch [e2e-llm-inference-service] apiGroups: [e2e-llm-inference-service] - '' [e2e-llm-inference-service] resources: [e2e-llm-inference-service] - pods [e2e-llm-inference-service] - verbs: [e2e-llm-inference-service] - get [e2e-llm-inference-service] - list [e2e-llm-inference-service] - watch [e2e-llm-inference-service] apiGroups: [e2e-llm-inference-service] - inference.networking.k8s.io [e2e-llm-inference-service] - inference.networking.x-k8s.io [e2e-llm-inference-service] resources: [e2e-llm-inference-service] - inferencepools [e2e-llm-inference-service] - inferenceobjectives [e2e-llm-inference-service] - inferencemodels [e2e-llm-inference-service] - verbs: [e2e-llm-inference-service] - get [e2e-llm-inference-service] - list [e2e-llm-inference-service] - watch [e2e-llm-inference-service] apiGroups: [e2e-llm-inference-service] - inference.networking.x-k8s.io [e2e-llm-inference-service] resources: [e2e-llm-inference-service] - inferencemodelrewrites [e2e-llm-inference-service] - inferencepoolimports [e2e-llm-inference-service] - verbs: [e2e-llm-inference-service] - get [e2e-llm-inference-service] - list [e2e-llm-inference-service] - watch [e2e-llm-inference-service] apiGroups: [e2e-llm-inference-service] - discovery.k8s.io [e2e-llm-inference-service] resources: [e2e-llm-inference-service] - endpointslices [e2e-llm-inference-service] - verbs: [e2e-llm-inference-service] - get [e2e-llm-inference-service] - list [e2e-llm-inference-service] - watch [e2e-llm-inference-service] - create [e2e-llm-inference-service] - update [e2e-llm-inference-service] - patch [e2e-llm-inference-service] - delete [e2e-llm-inference-service] apiGroups: [e2e-llm-inference-service] - coordination.k8s.io [e2e-llm-inference-service] resources: [e2e-llm-inference-service] - leases [e2e-llm-inference-service] apiVersion: rbac.authorization.k8s.io/v1 [e2e-llm-inference-service] kind: Role [e2e-llm-inference-service] [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1254 --- [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1255 metadata: [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-7ca60146-epp-service-f85gk [e2e-llm-inference-service] generateName: llmisvc-model-fb-opt-125m-with-7ca60146-epp-service- [e2e-llm-inference-service] namespace: kserve-ci-e2e-test [e2e-llm-inference-service] uid: 3c547c7f-6dd2-41d3-b450-58d49f8d0d2e [e2e-llm-inference-service] resourceVersion: '28936' [e2e-llm-inference-service] generation: 3 [e2e-llm-inference-service] creationTimestamp: '2026-07-07T14:29:24Z' [e2e-llm-inference-service] labels: [e2e-llm-inference-service] app.kubernetes.io/component: llminferenceservice-router-scheduler [e2e-llm-inference-service] app.kubernetes.io/name: llmisvc-model-fb-opt-125m-with-7ca60146 [e2e-llm-inference-service] app.kubernetes.io/part-of: llminferenceservice [e2e-llm-inference-service] endpointslice.kubernetes.io/managed-by: endpointslice-controller.k8s.io [e2e-llm-inference-service] kubernetes.io/service-name: llmisvc-model-fb-opt-125m-with-7ca60146-epp-service [e2e-llm-inference-service] annotations: [e2e-llm-inference-service] endpoints.kubernetes.io/last-change-trigger-time: '2026-07-07T14:29:58Z' [e2e-llm-inference-service] ownerReferences: [e2e-llm-inference-service] - apiVersion: v1 [e2e-llm-inference-service] kind: Service [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-7ca60146-epp-service [e2e-llm-inference-service] uid: f144c5eb-55c5-451e-b7ed-1e7b6e6d4ebf [e2e-llm-inference-service] controller: true [e2e-llm-inference-service] blockOwnerDeletion: true [e2e-llm-inference-service] managedFields: [e2e-llm-inference-service] - manager: kube-controller-manager [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] apiVersion: discovery.k8s.io/v1 [e2e-llm-inference-service] time: '2026-07-07T14:29:59Z' [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:addressType: {} [e2e-llm-inference-service] f:endpoints: {} [e2e-llm-inference-service] f:metadata: [e2e-llm-inference-service] f:annotations: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:endpoints.kubernetes.io/last-change-trigger-time: {} [e2e-llm-inference-service] f:generateName: {} [e2e-llm-inference-service] f:labels: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:app.kubernetes.io/component: {} [e2e-llm-inference-service] f:app.kubernetes.io/name: {} [e2e-llm-inference-service] f:app.kubernetes.io/part-of: {} [e2e-llm-inference-service] f:endpointslice.kubernetes.io/managed-by: {} [e2e-llm-inference-service] f:kubernetes.io/service-name: {} [e2e-llm-inference-service] f:ownerReferences: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"uid":"f144c5eb-55c5-451e-b7ed-1e7b6e6d4ebf"}: {} [e2e-llm-inference-service] f:ports: {} [e2e-llm-inference-service] addressType: IPv4 [e2e-llm-inference-service] endpoints: [e2e-llm-inference-service] - addresses: [e2e-llm-inference-service] - 10.132.0.52 [e2e-llm-inference-service] conditions: [e2e-llm-inference-service] ready: true [e2e-llm-inference-service] serving: true [e2e-llm-inference-service] terminating: false [e2e-llm-inference-service] targetRef: [e2e-llm-inference-service] kind: Pod [e2e-llm-inference-service] namespace: kserve-ci-e2e-test [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-7ca60146-kserve-router-schenp5mk [e2e-llm-inference-service] uid: ce797511-9ac0-4d16-8bf9-6f34aeec1585 [e2e-llm-inference-service] nodeName: ip-10-0-133-25.ec2.internal [e2e-llm-inference-service] zone: us-east-1a [e2e-llm-inference-service] ports: [e2e-llm-inference-service] - name: grpc [e2e-llm-inference-service] protocol: TCP [e2e-llm-inference-service] port: 9002 [e2e-llm-inference-service] - name: grpc-health [e2e-llm-inference-service] protocol: TCP [e2e-llm-inference-service] port: 9003 [e2e-llm-inference-service] - name: metrics [e2e-llm-inference-service] protocol: TCP [e2e-llm-inference-service] port: 9090 [e2e-llm-inference-service] - name: zmq [e2e-llm-inference-service] protocol: TCP [e2e-llm-inference-service] port: 5557 [e2e-llm-inference-service] apiVersion: discovery.k8s.io/v1 [e2e-llm-inference-service] kind: EndpointSlice [e2e-llm-inference-service] [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1254 --- [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1255 metadata: [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-7ca60146-kserve-workload-svn7kd2 [e2e-llm-inference-service] generateName: llmisvc-model-fb-opt-125m-with-7ca60146-kserve-workload-svc- [e2e-llm-inference-service] namespace: kserve-ci-e2e-test [e2e-llm-inference-service] uid: 8c54ca53-129d-4a9e-921b-21688770e9bc [e2e-llm-inference-service] resourceVersion: '29813' [e2e-llm-inference-service] generation: 3 [e2e-llm-inference-service] creationTimestamp: '2026-07-07T14:29:24Z' [e2e-llm-inference-service] labels: [e2e-llm-inference-service] app.kubernetes.io/component: llminferenceservice-workload [e2e-llm-inference-service] app.kubernetes.io/name: llmisvc-model-fb-opt-125m-with-7ca60146 [e2e-llm-inference-service] app.kubernetes.io/part-of: llminferenceservice [e2e-llm-inference-service] endpointslice.kubernetes.io/managed-by: endpointslice-controller.k8s.io [e2e-llm-inference-service] kubernetes.io/service-name: llmisvc-model-fb-opt-125m-with-7ca60146-kserve-workload-svc [e2e-llm-inference-service] annotations: [e2e-llm-inference-service] endpoints.kubernetes.io/last-change-trigger-time: '2026-07-07T14:30:44Z' [e2e-llm-inference-service] ownerReferences: [e2e-llm-inference-service] - apiVersion: v1 [e2e-llm-inference-service] kind: Service [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-7ca60146-kserve-workload-svc [e2e-llm-inference-service] uid: 8fe30c1f-242c-4ca3-aad7-a4b6836297ea [e2e-llm-inference-service] controller: true [e2e-llm-inference-service] blockOwnerDeletion: true [e2e-llm-inference-service] managedFields: [e2e-llm-inference-service] - manager: kube-controller-manager [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] apiVersion: discovery.k8s.io/v1 [e2e-llm-inference-service] time: '2026-07-07T14:30:45Z' [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:addressType: {} [e2e-llm-inference-service] f:endpoints: {} [e2e-llm-inference-service] f:metadata: [e2e-llm-inference-service] f:annotations: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:endpoints.kubernetes.io/last-change-trigger-time: {} [e2e-llm-inference-service] f:generateName: {} [e2e-llm-inference-service] f:labels: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:app.kubernetes.io/component: {} [e2e-llm-inference-service] f:app.kubernetes.io/name: {} [e2e-llm-inference-service] f:app.kubernetes.io/part-of: {} [e2e-llm-inference-service] f:endpointslice.kubernetes.io/managed-by: {} [e2e-llm-inference-service] f:kubernetes.io/service-name: {} [e2e-llm-inference-service] f:ownerReferences: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"uid":"8fe30c1f-242c-4ca3-aad7-a4b6836297ea"}: {} [e2e-llm-inference-service] f:ports: {} [e2e-llm-inference-service] addressType: IPv4 [e2e-llm-inference-service] endpoints: [e2e-llm-inference-service] - addresses: [e2e-llm-inference-service] - 10.134.0.30 [e2e-llm-inference-service] conditions: [e2e-llm-inference-service] ready: true [e2e-llm-inference-service] serving: true [e2e-llm-inference-service] terminating: false [e2e-llm-inference-service] targetRef: [e2e-llm-inference-service] kind: Pod [e2e-llm-inference-service] namespace: kserve-ci-e2e-test [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-7ca60146-kserve-7cfdf49d74-pd9px [e2e-llm-inference-service] uid: dae89c51-1d5b-4d9c-a3cc-24917b85ad0b [e2e-llm-inference-service] nodeName: ip-10-0-139-47.ec2.internal [e2e-llm-inference-service] zone: us-east-1a [e2e-llm-inference-service] ports: [e2e-llm-inference-service] - name: https [e2e-llm-inference-service] protocol: TCP [e2e-llm-inference-service] port: 8000 [e2e-llm-inference-service] appProtocol: https [e2e-llm-inference-service] apiVersion: discovery.k8s.io/v1 [e2e-llm-inference-service] kind: EndpointSlice [e2e-llm-inference-service] [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1254 --- [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1255 metadata: [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-7ca60146-epp-rb [e2e-llm-inference-service] namespace: kserve-ci-e2e-test [e2e-llm-inference-service] uid: 5dec14fd-be15-4f60-ade8-0954fd32be81 [e2e-llm-inference-service] resourceVersion: '28241' [e2e-llm-inference-service] creationTimestamp: '2026-07-07T14:29:24Z' [e2e-llm-inference-service] labels: [e2e-llm-inference-service] app.kubernetes.io/component: llminferenceservice-router-scheduler [e2e-llm-inference-service] app.kubernetes.io/name: llmisvc-model-fb-opt-125m-with-7ca60146 [e2e-llm-inference-service] app.kubernetes.io/part-of: llminferenceservice [e2e-llm-inference-service] ownerReferences: [e2e-llm-inference-service] - apiVersion: serving.kserve.io/v1alpha2 [e2e-llm-inference-service] kind: LLMInferenceService [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-7ca60146 [e2e-llm-inference-service] uid: e15cfc59-e4a1-4876-b43b-85ce3b1b239b [e2e-llm-inference-service] controller: true [e2e-llm-inference-service] blockOwnerDeletion: true [e2e-llm-inference-service] managedFields: [e2e-llm-inference-service] - manager: manager [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] apiVersion: rbac.authorization.k8s.io/v1 [e2e-llm-inference-service] time: '2026-07-07T14:29:24Z' [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:metadata: [e2e-llm-inference-service] f:labels: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:app.kubernetes.io/component: {} [e2e-llm-inference-service] f:app.kubernetes.io/name: {} [e2e-llm-inference-service] f:app.kubernetes.io/part-of: {} [e2e-llm-inference-service] f:ownerReferences: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"uid":"e15cfc59-e4a1-4876-b43b-85ce3b1b239b"}: {} [e2e-llm-inference-service] f:roleRef: {} [e2e-llm-inference-service] f:subjects: {} [e2e-llm-inference-service] userNames: [e2e-llm-inference-service] - system:serviceaccount:kserve-ci-e2e-test:llmisvc-model-fb-opt-125m-with-7ca60146-epp-sa [e2e-llm-inference-service] groupNames: null [e2e-llm-inference-service] subjects: [e2e-llm-inference-service] - kind: ServiceAccount [e2e-llm-inference-service] namespace: kserve-ci-e2e-test [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-7ca60146-epp-sa [e2e-llm-inference-service] roleRef: [e2e-llm-inference-service] namespace: kserve-ci-e2e-test [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-7ca60146-epp-role [e2e-llm-inference-service] apiVersion: authorization.openshift.io/v1 [e2e-llm-inference-service] kind: RoleBinding [e2e-llm-inference-service] [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1254 --- [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1255 metadata: [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-7ca60146-epp-role [e2e-llm-inference-service] namespace: kserve-ci-e2e-test [e2e-llm-inference-service] uid: 13c554f4-7989-49ec-8acf-0b4132cbd414 [e2e-llm-inference-service] resourceVersion: '28236' [e2e-llm-inference-service] creationTimestamp: '2026-07-07T14:29:24Z' [e2e-llm-inference-service] labels: [e2e-llm-inference-service] app.kubernetes.io/component: llminferenceservice-router-scheduler [e2e-llm-inference-service] app.kubernetes.io/name: llmisvc-model-fb-opt-125m-with-7ca60146 [e2e-llm-inference-service] app.kubernetes.io/part-of: llminferenceservice [e2e-llm-inference-service] ownerReferences: [e2e-llm-inference-service] - apiVersion: serving.kserve.io/v1alpha2 [e2e-llm-inference-service] kind: LLMInferenceService [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-7ca60146 [e2e-llm-inference-service] uid: e15cfc59-e4a1-4876-b43b-85ce3b1b239b [e2e-llm-inference-service] controller: true [e2e-llm-inference-service] blockOwnerDeletion: true [e2e-llm-inference-service] managedFields: [e2e-llm-inference-service] - manager: manager [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] apiVersion: rbac.authorization.k8s.io/v1 [e2e-llm-inference-service] time: '2026-07-07T14:29:24Z' [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:metadata: [e2e-llm-inference-service] f:labels: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:app.kubernetes.io/component: {} [e2e-llm-inference-service] f:app.kubernetes.io/name: {} [e2e-llm-inference-service] f:app.kubernetes.io/part-of: {} [e2e-llm-inference-service] f:ownerReferences: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"uid":"e15cfc59-e4a1-4876-b43b-85ce3b1b239b"}: {} [e2e-llm-inference-service] f:rules: {} [e2e-llm-inference-service] rules: [e2e-llm-inference-service] - verbs: [e2e-llm-inference-service] - get [e2e-llm-inference-service] - list [e2e-llm-inference-service] - watch [e2e-llm-inference-service] attributeRestrictions: null [e2e-llm-inference-service] apiGroups: [e2e-llm-inference-service] - '' [e2e-llm-inference-service] resources: [e2e-llm-inference-service] - pods [e2e-llm-inference-service] - verbs: [e2e-llm-inference-service] - get [e2e-llm-inference-service] - list [e2e-llm-inference-service] - watch [e2e-llm-inference-service] attributeRestrictions: null [e2e-llm-inference-service] apiGroups: [e2e-llm-inference-service] - inference.networking.k8s.io [e2e-llm-inference-service] - inference.networking.x-k8s.io [e2e-llm-inference-service] resources: [e2e-llm-inference-service] - inferencemodels [e2e-llm-inference-service] - inferenceobjectives [e2e-llm-inference-service] - inferencepools [e2e-llm-inference-service] - verbs: [e2e-llm-inference-service] - get [e2e-llm-inference-service] - list [e2e-llm-inference-service] - watch [e2e-llm-inference-service] attributeRestrictions: null [e2e-llm-inference-service] apiGroups: [e2e-llm-inference-service] - inference.networking.x-k8s.io [e2e-llm-inference-service] resources: [e2e-llm-inference-service] - inferencemodelrewrites [e2e-llm-inference-service] - inferencepoolimports [e2e-llm-inference-service] - verbs: [e2e-llm-inference-service] - get [e2e-llm-inference-service] - list [e2e-llm-inference-service] - watch [e2e-llm-inference-service] attributeRestrictions: null [e2e-llm-inference-service] apiGroups: [e2e-llm-inference-service] - discovery.k8s.io [e2e-llm-inference-service] resources: [e2e-llm-inference-service] - endpointslices [e2e-llm-inference-service] - verbs: [e2e-llm-inference-service] - create [e2e-llm-inference-service] - delete [e2e-llm-inference-service] - get [e2e-llm-inference-service] - list [e2e-llm-inference-service] - patch [e2e-llm-inference-service] - update [e2e-llm-inference-service] - watch [e2e-llm-inference-service] attributeRestrictions: null [e2e-llm-inference-service] apiGroups: [e2e-llm-inference-service] - coordination.k8s.io [e2e-llm-inference-service] resources: [e2e-llm-inference-service] - leases [e2e-llm-inference-service] apiVersion: authorization.openshift.io/v1 [e2e-llm-inference-service] kind: Role [e2e-llm-inference-service] [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1254 --- [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1255 apiVersion: gateway.networking.k8s.io/v1 [e2e-llm-inference-service] kind: HTTPRoute [e2e-llm-inference-service] metadata: [e2e-llm-inference-service] annotations: [e2e-llm-inference-service] serving.kserve.io/inference-pool-migrated: v1 [e2e-llm-inference-service] creationTimestamp: '2026-07-07T14:29:25Z' [e2e-llm-inference-service] generation: 2 [e2e-llm-inference-service] labels: [e2e-llm-inference-service] app.kubernetes.io/component: llminferenceservice-router [e2e-llm-inference-service] app.kubernetes.io/name: llmisvc-model-fb-opt-125m-with-7ca60146 [e2e-llm-inference-service] app.kubernetes.io/part-of: llminferenceservice [e2e-llm-inference-service] managedFields: [e2e-llm-inference-service] - apiVersion: gateway.networking.k8s.io/v1 [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:metadata: [e2e-llm-inference-service] f:annotations: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:serving.kserve.io/inference-pool-migrated: {} [e2e-llm-inference-service] f:labels: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:app.kubernetes.io/component: {} [e2e-llm-inference-service] f:app.kubernetes.io/name: {} [e2e-llm-inference-service] f:app.kubernetes.io/part-of: {} [e2e-llm-inference-service] f:ownerReferences: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"uid":"e15cfc59-e4a1-4876-b43b-85ce3b1b239b"}: {} [e2e-llm-inference-service] f:spec: {} [e2e-llm-inference-service] manager: manager [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] time: '2026-07-07T14:29:34Z' [e2e-llm-inference-service] - apiVersion: gateway.networking.k8s.io/v1beta1 [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:status: {} [e2e-llm-inference-service] manager: pilot-discovery [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] subresource: status [e2e-llm-inference-service] time: '2026-07-07T14:29:34Z' [e2e-llm-inference-service] - apiVersion: gateway.networking.k8s.io/v1 [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:status: [e2e-llm-inference-service] f:parents: {} [e2e-llm-inference-service] manager: manager [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] subresource: status [e2e-llm-inference-service] time: '2026-07-07T14:29:35Z' [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-7ca60146-kserve-route [e2e-llm-inference-service] namespace: kserve-ci-e2e-test [e2e-llm-inference-service] ownerReferences: [e2e-llm-inference-service] - apiVersion: serving.kserve.io/v1alpha2 [e2e-llm-inference-service] blockOwnerDeletion: true [e2e-llm-inference-service] controller: true [e2e-llm-inference-service] kind: LLMInferenceService [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-7ca60146 [e2e-llm-inference-service] uid: e15cfc59-e4a1-4876-b43b-85ce3b1b239b [e2e-llm-inference-service] resourceVersion: '28532' [e2e-llm-inference-service] uid: 9b1d5526-1cec-4a64-afb1-757c31d00a4d [e2e-llm-inference-service] spec: [e2e-llm-inference-service] parentRefs: [e2e-llm-inference-service] - group: gateway.networking.k8s.io [e2e-llm-inference-service] kind: Gateway [e2e-llm-inference-service] name: openshift-ai-inference [e2e-llm-inference-service] namespace: openshift-ingress [e2e-llm-inference-service] rules: [e2e-llm-inference-service] - backendRefs: [e2e-llm-inference-service] - group: inference.networking.k8s.io [e2e-llm-inference-service] kind: InferencePool [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-7ca60146-inference-pool [e2e-llm-inference-service] port: 8000 [e2e-llm-inference-service] weight: 1 [e2e-llm-inference-service] filters: [e2e-llm-inference-service] - type: URLRewrite [e2e-llm-inference-service] urlRewrite: [e2e-llm-inference-service] path: [e2e-llm-inference-service] replacePrefixMatch: /v1/completions [e2e-llm-inference-service] type: ReplacePrefixMatch [e2e-llm-inference-service] matches: [e2e-llm-inference-service] - path: [e2e-llm-inference-service] type: PathPrefix [e2e-llm-inference-service] value: /kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-with-7ca60146/v1/completions [e2e-llm-inference-service] timeouts: [e2e-llm-inference-service] backendRequest: 0s [e2e-llm-inference-service] request: 0s [e2e-llm-inference-service] - backendRefs: [e2e-llm-inference-service] - group: inference.networking.k8s.io [e2e-llm-inference-service] kind: InferencePool [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-7ca60146-inference-pool [e2e-llm-inference-service] port: 8000 [e2e-llm-inference-service] weight: 1 [e2e-llm-inference-service] matches: [e2e-llm-inference-service] - headers: [e2e-llm-inference-service] - name: X-Gateway-Model-Name [e2e-llm-inference-service] type: Exact [e2e-llm-inference-service] value: publishers/kserve-ci-e2e-test/models/facebook/opt-125m [e2e-llm-inference-service] path: [e2e-llm-inference-service] type: Exact [e2e-llm-inference-service] value: /v1/completions [e2e-llm-inference-service] - headers: [e2e-llm-inference-service] - name: X-Gateway-Model-Name [e2e-llm-inference-service] type: Exact [e2e-llm-inference-service] value: publishers/kserve-ci-e2e-test/models/facebook/opt-125m [e2e-llm-inference-service] path: [e2e-llm-inference-service] type: Exact [e2e-llm-inference-service] value: /v1/completions/ [e2e-llm-inference-service] - headers: [e2e-llm-inference-service] - name: X-Gateway-Model-Name [e2e-llm-inference-service] type: Exact [e2e-llm-inference-service] value: publishers/kserve-ci-e2e-test/models/lora-adapter-1 [e2e-llm-inference-service] path: [e2e-llm-inference-service] type: Exact [e2e-llm-inference-service] value: /v1/completions [e2e-llm-inference-service] - headers: [e2e-llm-inference-service] - name: X-Gateway-Model-Name [e2e-llm-inference-service] type: Exact [e2e-llm-inference-service] value: publishers/kserve-ci-e2e-test/models/lora-adapter-1 [e2e-llm-inference-service] path: [e2e-llm-inference-service] type: Exact [e2e-llm-inference-service] value: /v1/completions/ [e2e-llm-inference-service] timeouts: [e2e-llm-inference-service] backendRequest: 0s [e2e-llm-inference-service] request: 0s [e2e-llm-inference-service] - backendRefs: [e2e-llm-inference-service] - group: inference.networking.k8s.io [e2e-llm-inference-service] kind: InferencePool [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-7ca60146-inference-pool [e2e-llm-inference-service] port: 8000 [e2e-llm-inference-service] weight: 1 [e2e-llm-inference-service] filters: [e2e-llm-inference-service] - type: URLRewrite [e2e-llm-inference-service] urlRewrite: [e2e-llm-inference-service] path: [e2e-llm-inference-service] replacePrefixMatch: /v1/chat/completions [e2e-llm-inference-service] type: ReplacePrefixMatch [e2e-llm-inference-service] matches: [e2e-llm-inference-service] - path: [e2e-llm-inference-service] type: PathPrefix [e2e-llm-inference-service] value: /kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-with-7ca60146/v1/chat/completions [e2e-llm-inference-service] timeouts: [e2e-llm-inference-service] backendRequest: 0s [e2e-llm-inference-service] request: 0s [e2e-llm-inference-service] - backendRefs: [e2e-llm-inference-service] - group: inference.networking.k8s.io [e2e-llm-inference-service] kind: InferencePool [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-7ca60146-inference-pool [e2e-llm-inference-service] port: 8000 [e2e-llm-inference-service] weight: 1 [e2e-llm-inference-service] matches: [e2e-llm-inference-service] - headers: [e2e-llm-inference-service] - name: X-Gateway-Model-Name [e2e-llm-inference-service] type: Exact [e2e-llm-inference-service] value: publishers/kserve-ci-e2e-test/models/facebook/opt-125m [e2e-llm-inference-service] path: [e2e-llm-inference-service] type: Exact [e2e-llm-inference-service] value: /v1/chat/completions [e2e-llm-inference-service] - headers: [e2e-llm-inference-service] - name: X-Gateway-Model-Name [e2e-llm-inference-service] type: Exact [e2e-llm-inference-service] value: publishers/kserve-ci-e2e-test/models/facebook/opt-125m [e2e-llm-inference-service] path: [e2e-llm-inference-service] type: Exact [e2e-llm-inference-service] value: /v1/chat/completions/ [e2e-llm-inference-service] - headers: [e2e-llm-inference-service] - name: X-Gateway-Model-Name [e2e-llm-inference-service] type: Exact [e2e-llm-inference-service] value: publishers/kserve-ci-e2e-test/models/lora-adapter-1 [e2e-llm-inference-service] path: [e2e-llm-inference-service] type: Exact [e2e-llm-inference-service] value: /v1/chat/completions [e2e-llm-inference-service] - headers: [e2e-llm-inference-service] - name: X-Gateway-Model-Name [e2e-llm-inference-service] type: Exact [e2e-llm-inference-service] value: publishers/kserve-ci-e2e-test/models/lora-adapter-1 [e2e-llm-inference-service] path: [e2e-llm-inference-service] type: Exact [e2e-llm-inference-service] value: /v1/chat/completions/ [e2e-llm-inference-service] timeouts: [e2e-llm-inference-service] backendRequest: 0s [e2e-llm-inference-service] request: 0s [e2e-llm-inference-service] - backendRefs: [e2e-llm-inference-service] - group: inference.networking.k8s.io [e2e-llm-inference-service] kind: InferencePool [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-7ca60146-inference-pool [e2e-llm-inference-service] port: 8000 [e2e-llm-inference-service] weight: 1 [e2e-llm-inference-service] filters: [e2e-llm-inference-service] - type: URLRewrite [e2e-llm-inference-service] urlRewrite: [e2e-llm-inference-service] path: [e2e-llm-inference-service] replacePrefixMatch: /v1/responses [e2e-llm-inference-service] type: ReplacePrefixMatch [e2e-llm-inference-service] matches: [e2e-llm-inference-service] - path: [e2e-llm-inference-service] type: PathPrefix [e2e-llm-inference-service] value: /kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-with-7ca60146/v1/responses [e2e-llm-inference-service] timeouts: [e2e-llm-inference-service] backendRequest: 0s [e2e-llm-inference-service] request: 0s [e2e-llm-inference-service] - backendRefs: [e2e-llm-inference-service] - group: inference.networking.k8s.io [e2e-llm-inference-service] kind: InferencePool [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-7ca60146-inference-pool [e2e-llm-inference-service] port: 8000 [e2e-llm-inference-service] weight: 1 [e2e-llm-inference-service] matches: [e2e-llm-inference-service] - headers: [e2e-llm-inference-service] - name: X-Gateway-Model-Name [e2e-llm-inference-service] type: Exact [e2e-llm-inference-service] value: publishers/kserve-ci-e2e-test/models/facebook/opt-125m [e2e-llm-inference-service] path: [e2e-llm-inference-service] type: Exact [e2e-llm-inference-service] value: /v1/responses [e2e-llm-inference-service] - headers: [e2e-llm-inference-service] - name: X-Gateway-Model-Name [e2e-llm-inference-service] type: Exact [e2e-llm-inference-service] value: publishers/kserve-ci-e2e-test/models/facebook/opt-125m [e2e-llm-inference-service] path: [e2e-llm-inference-service] type: Exact [e2e-llm-inference-service] value: /v1/responses/ [e2e-llm-inference-service] - headers: [e2e-llm-inference-service] - name: X-Gateway-Model-Name [e2e-llm-inference-service] type: Exact [e2e-llm-inference-service] value: publishers/kserve-ci-e2e-test/models/lora-adapter-1 [e2e-llm-inference-service] path: [e2e-llm-inference-service] type: Exact [e2e-llm-inference-service] value: /v1/responses [e2e-llm-inference-service] - headers: [e2e-llm-inference-service] - name: X-Gateway-Model-Name [e2e-llm-inference-service] type: Exact [e2e-llm-inference-service] value: publishers/kserve-ci-e2e-test/models/lora-adapter-1 [e2e-llm-inference-service] path: [e2e-llm-inference-service] type: Exact [e2e-llm-inference-service] value: /v1/responses/ [e2e-llm-inference-service] timeouts: [e2e-llm-inference-service] backendRequest: 0s [e2e-llm-inference-service] request: 0s [e2e-llm-inference-service] - backendRefs: [e2e-llm-inference-service] - group: inference.networking.k8s.io [e2e-llm-inference-service] kind: InferencePool [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-7ca60146-inference-pool [e2e-llm-inference-service] port: 8000 [e2e-llm-inference-service] weight: 1 [e2e-llm-inference-service] filters: [e2e-llm-inference-service] - type: URLRewrite [e2e-llm-inference-service] urlRewrite: [e2e-llm-inference-service] path: [e2e-llm-inference-service] replacePrefixMatch: /v1/messages [e2e-llm-inference-service] type: ReplacePrefixMatch [e2e-llm-inference-service] matches: [e2e-llm-inference-service] - path: [e2e-llm-inference-service] type: PathPrefix [e2e-llm-inference-service] value: /kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-with-7ca60146/v1/messages [e2e-llm-inference-service] timeouts: [e2e-llm-inference-service] backendRequest: 0s [e2e-llm-inference-service] request: 0s [e2e-llm-inference-service] - backendRefs: [e2e-llm-inference-service] - group: inference.networking.k8s.io [e2e-llm-inference-service] kind: InferencePool [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-7ca60146-inference-pool [e2e-llm-inference-service] port: 8000 [e2e-llm-inference-service] weight: 1 [e2e-llm-inference-service] matches: [e2e-llm-inference-service] - headers: [e2e-llm-inference-service] - name: X-Gateway-Model-Name [e2e-llm-inference-service] type: Exact [e2e-llm-inference-service] value: publishers/kserve-ci-e2e-test/models/facebook/opt-125m [e2e-llm-inference-service] path: [e2e-llm-inference-service] type: Exact [e2e-llm-inference-service] value: /v1/messages [e2e-llm-inference-service] - headers: [e2e-llm-inference-service] - name: X-Gateway-Model-Name [e2e-llm-inference-service] type: Exact [e2e-llm-inference-service] value: publishers/kserve-ci-e2e-test/models/facebook/opt-125m [e2e-llm-inference-service] path: [e2e-llm-inference-service] type: Exact [e2e-llm-inference-service] value: /v1/messages/ [e2e-llm-inference-service] - headers: [e2e-llm-inference-service] - name: X-Gateway-Model-Name [e2e-llm-inference-service] type: Exact [e2e-llm-inference-service] value: publishers/kserve-ci-e2e-test/models/lora-adapter-1 [e2e-llm-inference-service] path: [e2e-llm-inference-service] type: Exact [e2e-llm-inference-service] value: /v1/messages [e2e-llm-inference-service] - headers: [e2e-llm-inference-service] - name: X-Gateway-Model-Name [e2e-llm-inference-service] type: Exact [e2e-llm-inference-service] value: publishers/kserve-ci-e2e-test/models/lora-adapter-1 [e2e-llm-inference-service] path: [e2e-llm-inference-service] type: Exact [e2e-llm-inference-service] value: /v1/messages/ [e2e-llm-inference-service] timeouts: [e2e-llm-inference-service] backendRequest: 0s [e2e-llm-inference-service] request: 0s [e2e-llm-inference-service] - backendRefs: [e2e-llm-inference-service] - group: '' [e2e-llm-inference-service] kind: Service [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-7ca60146-kserve-workload-svc [e2e-llm-inference-service] port: 8000 [e2e-llm-inference-service] weight: 1 [e2e-llm-inference-service] filters: [e2e-llm-inference-service] - type: URLRewrite [e2e-llm-inference-service] urlRewrite: [e2e-llm-inference-service] path: [e2e-llm-inference-service] replacePrefixMatch: / [e2e-llm-inference-service] type: ReplacePrefixMatch [e2e-llm-inference-service] matches: [e2e-llm-inference-service] - path: [e2e-llm-inference-service] type: PathPrefix [e2e-llm-inference-service] value: /kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-with-7ca60146 [e2e-llm-inference-service] timeouts: [e2e-llm-inference-service] backendRequest: 0s [e2e-llm-inference-service] request: 0s [e2e-llm-inference-service] - backendRefs: [e2e-llm-inference-service] - group: '' [e2e-llm-inference-service] kind: Service [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-7ca60146-kserve-workload-svc [e2e-llm-inference-service] port: 8000 [e2e-llm-inference-service] weight: 1 [e2e-llm-inference-service] matches: [e2e-llm-inference-service] - headers: [e2e-llm-inference-service] - name: X-Gateway-Model-Name [e2e-llm-inference-service] type: Exact [e2e-llm-inference-service] value: publishers/kserve-ci-e2e-test/models/facebook/opt-125m [e2e-llm-inference-service] path: [e2e-llm-inference-service] type: PathPrefix [e2e-llm-inference-service] value: / [e2e-llm-inference-service] - headers: [e2e-llm-inference-service] - name: X-Gateway-Model-Name [e2e-llm-inference-service] type: Exact [e2e-llm-inference-service] value: publishers/kserve-ci-e2e-test/models/lora-adapter-1 [e2e-llm-inference-service] path: [e2e-llm-inference-service] type: PathPrefix [e2e-llm-inference-service] value: / [e2e-llm-inference-service] timeouts: [e2e-llm-inference-service] backendRequest: 0s [e2e-llm-inference-service] request: 0s [e2e-llm-inference-service] status: [e2e-llm-inference-service] parents: [e2e-llm-inference-service] - conditions: [e2e-llm-inference-service] - lastTransitionTime: '2026-07-07T14:29:25Z' [e2e-llm-inference-service] message: Route was valid [e2e-llm-inference-service] observedGeneration: 2 [e2e-llm-inference-service] reason: Accepted [e2e-llm-inference-service] status: 'True' [e2e-llm-inference-service] type: Accepted [e2e-llm-inference-service] - lastTransitionTime: '2026-07-07T14:29:34Z' [e2e-llm-inference-service] message: All references resolved [e2e-llm-inference-service] observedGeneration: 2 [e2e-llm-inference-service] reason: ResolvedRefs [e2e-llm-inference-service] status: 'True' [e2e-llm-inference-service] type: ResolvedRefs [e2e-llm-inference-service] controllerName: openshift.io/gateway-controller/v1 [e2e-llm-inference-service] parentRef: [e2e-llm-inference-service] group: gateway.networking.k8s.io [e2e-llm-inference-service] kind: Gateway [e2e-llm-inference-service] name: openshift-ai-inference [e2e-llm-inference-service] namespace: openshift-ingress [e2e-llm-inference-service] - conditions: [e2e-llm-inference-service] - lastTransitionTime: '2026-07-07T14:29:25Z' [e2e-llm-inference-service] message: Object affected by AuthPolicy [kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-with-7ca60146-kserve-route-authn [e2e-llm-inference-service] openshift-ingress/openshift-ai-inference-authn] [e2e-llm-inference-service] observedGeneration: 2 [e2e-llm-inference-service] reason: Accepted [e2e-llm-inference-service] status: 'True' [e2e-llm-inference-service] type: kuadrant.io/AuthPolicyAffected [e2e-llm-inference-service] controllerName: kuadrant.io/policy-controller [e2e-llm-inference-service] parentRef: [e2e-llm-inference-service] group: gateway.networking.k8s.io [e2e-llm-inference-service] kind: Gateway [e2e-llm-inference-service] name: openshift-ai-inference [e2e-llm-inference-service] namespace: openshift-ingress [e2e-llm-inference-service] [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1254 --- [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1255 apiVersion: gateway.networking.k8s.io/v1beta1 [e2e-llm-inference-service] kind: HTTPRoute [e2e-llm-inference-service] metadata: [e2e-llm-inference-service] annotations: [e2e-llm-inference-service] serving.kserve.io/inference-pool-migrated: v1 [e2e-llm-inference-service] creationTimestamp: '2026-07-07T14:29:25Z' [e2e-llm-inference-service] generation: 2 [e2e-llm-inference-service] labels: [e2e-llm-inference-service] app.kubernetes.io/component: llminferenceservice-router [e2e-llm-inference-service] app.kubernetes.io/name: llmisvc-model-fb-opt-125m-with-7ca60146 [e2e-llm-inference-service] app.kubernetes.io/part-of: llminferenceservice [e2e-llm-inference-service] managedFields: [e2e-llm-inference-service] - apiVersion: gateway.networking.k8s.io/v1 [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:metadata: [e2e-llm-inference-service] f:annotations: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:serving.kserve.io/inference-pool-migrated: {} [e2e-llm-inference-service] f:labels: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:app.kubernetes.io/component: {} [e2e-llm-inference-service] f:app.kubernetes.io/name: {} [e2e-llm-inference-service] f:app.kubernetes.io/part-of: {} [e2e-llm-inference-service] f:ownerReferences: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"uid":"e15cfc59-e4a1-4876-b43b-85ce3b1b239b"}: {} [e2e-llm-inference-service] f:spec: {} [e2e-llm-inference-service] manager: manager [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] time: '2026-07-07T14:29:34Z' [e2e-llm-inference-service] - apiVersion: gateway.networking.k8s.io/v1beta1 [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:status: {} [e2e-llm-inference-service] manager: pilot-discovery [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] subresource: status [e2e-llm-inference-service] time: '2026-07-07T14:29:34Z' [e2e-llm-inference-service] - apiVersion: gateway.networking.k8s.io/v1 [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:status: [e2e-llm-inference-service] f:parents: {} [e2e-llm-inference-service] manager: manager [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] subresource: status [e2e-llm-inference-service] time: '2026-07-07T14:29:35Z' [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-7ca60146-kserve-route [e2e-llm-inference-service] namespace: kserve-ci-e2e-test [e2e-llm-inference-service] ownerReferences: [e2e-llm-inference-service] - apiVersion: serving.kserve.io/v1alpha2 [e2e-llm-inference-service] blockOwnerDeletion: true [e2e-llm-inference-service] controller: true [e2e-llm-inference-service] kind: LLMInferenceService [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-7ca60146 [e2e-llm-inference-service] uid: e15cfc59-e4a1-4876-b43b-85ce3b1b239b [e2e-llm-inference-service] resourceVersion: '28532' [e2e-llm-inference-service] uid: 9b1d5526-1cec-4a64-afb1-757c31d00a4d [e2e-llm-inference-service] spec: [e2e-llm-inference-service] parentRefs: [e2e-llm-inference-service] - group: gateway.networking.k8s.io [e2e-llm-inference-service] kind: Gateway [e2e-llm-inference-service] name: openshift-ai-inference [e2e-llm-inference-service] namespace: openshift-ingress [e2e-llm-inference-service] rules: [e2e-llm-inference-service] - backendRefs: [e2e-llm-inference-service] - group: inference.networking.k8s.io [e2e-llm-inference-service] kind: InferencePool [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-7ca60146-inference-pool [e2e-llm-inference-service] port: 8000 [e2e-llm-inference-service] weight: 1 [e2e-llm-inference-service] filters: [e2e-llm-inference-service] - type: URLRewrite [e2e-llm-inference-service] urlRewrite: [e2e-llm-inference-service] path: [e2e-llm-inference-service] replacePrefixMatch: /v1/completions [e2e-llm-inference-service] type: ReplacePrefixMatch [e2e-llm-inference-service] matches: [e2e-llm-inference-service] - path: [e2e-llm-inference-service] type: PathPrefix [e2e-llm-inference-service] value: /kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-with-7ca60146/v1/completions [e2e-llm-inference-service] timeouts: [e2e-llm-inference-service] backendRequest: 0s [e2e-llm-inference-service] request: 0s [e2e-llm-inference-service] - backendRefs: [e2e-llm-inference-service] - group: inference.networking.k8s.io [e2e-llm-inference-service] kind: InferencePool [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-7ca60146-inference-pool [e2e-llm-inference-service] port: 8000 [e2e-llm-inference-service] weight: 1 [e2e-llm-inference-service] matches: [e2e-llm-inference-service] - headers: [e2e-llm-inference-service] - name: X-Gateway-Model-Name [e2e-llm-inference-service] type: Exact [e2e-llm-inference-service] value: publishers/kserve-ci-e2e-test/models/facebook/opt-125m [e2e-llm-inference-service] path: [e2e-llm-inference-service] type: Exact [e2e-llm-inference-service] value: /v1/completions [e2e-llm-inference-service] - headers: [e2e-llm-inference-service] - name: X-Gateway-Model-Name [e2e-llm-inference-service] type: Exact [e2e-llm-inference-service] value: publishers/kserve-ci-e2e-test/models/facebook/opt-125m [e2e-llm-inference-service] path: [e2e-llm-inference-service] type: Exact [e2e-llm-inference-service] value: /v1/completions/ [e2e-llm-inference-service] - headers: [e2e-llm-inference-service] - name: X-Gateway-Model-Name [e2e-llm-inference-service] type: Exact [e2e-llm-inference-service] value: publishers/kserve-ci-e2e-test/models/lora-adapter-1 [e2e-llm-inference-service] path: [e2e-llm-inference-service] type: Exact [e2e-llm-inference-service] value: /v1/completions [e2e-llm-inference-service] - headers: [e2e-llm-inference-service] - name: X-Gateway-Model-Name [e2e-llm-inference-service] type: Exact [e2e-llm-inference-service] value: publishers/kserve-ci-e2e-test/models/lora-adapter-1 [e2e-llm-inference-service] path: [e2e-llm-inference-service] type: Exact [e2e-llm-inference-service] value: /v1/completions/ [e2e-llm-inference-service] timeouts: [e2e-llm-inference-service] backendRequest: 0s [e2e-llm-inference-service] request: 0s [e2e-llm-inference-service] - backendRefs: [e2e-llm-inference-service] - group: inference.networking.k8s.io [e2e-llm-inference-service] kind: InferencePool [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-7ca60146-inference-pool [e2e-llm-inference-service] port: 8000 [e2e-llm-inference-service] weight: 1 [e2e-llm-inference-service] filters: [e2e-llm-inference-service] - type: URLRewrite [e2e-llm-inference-service] urlRewrite: [e2e-llm-inference-service] path: [e2e-llm-inference-service] replacePrefixMatch: /v1/chat/completions [e2e-llm-inference-service] type: ReplacePrefixMatch [e2e-llm-inference-service] matches: [e2e-llm-inference-service] - path: [e2e-llm-inference-service] type: PathPrefix [e2e-llm-inference-service] value: /kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-with-7ca60146/v1/chat/completions [e2e-llm-inference-service] timeouts: [e2e-llm-inference-service] backendRequest: 0s [e2e-llm-inference-service] request: 0s [e2e-llm-inference-service] - backendRefs: [e2e-llm-inference-service] - group: inference.networking.k8s.io [e2e-llm-inference-service] kind: InferencePool [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-7ca60146-inference-pool [e2e-llm-inference-service] port: 8000 [e2e-llm-inference-service] weight: 1 [e2e-llm-inference-service] matches: [e2e-llm-inference-service] - headers: [e2e-llm-inference-service] - name: X-Gateway-Model-Name [e2e-llm-inference-service] type: Exact [e2e-llm-inference-service] value: publishers/kserve-ci-e2e-test/models/facebook/opt-125m [e2e-llm-inference-service] path: [e2e-llm-inference-service] type: Exact [e2e-llm-inference-service] value: /v1/chat/completions [e2e-llm-inference-service] - headers: [e2e-llm-inference-service] - name: X-Gateway-Model-Name [e2e-llm-inference-service] type: Exact [e2e-llm-inference-service] value: publishers/kserve-ci-e2e-test/models/facebook/opt-125m [e2e-llm-inference-service] path: [e2e-llm-inference-service] type: Exact [e2e-llm-inference-service] value: /v1/chat/completions/ [e2e-llm-inference-service] - headers: [e2e-llm-inference-service] - name: X-Gateway-Model-Name [e2e-llm-inference-service] type: Exact [e2e-llm-inference-service] value: publishers/kserve-ci-e2e-test/models/lora-adapter-1 [e2e-llm-inference-service] path: [e2e-llm-inference-service] type: Exact [e2e-llm-inference-service] value: /v1/chat/completions [e2e-llm-inference-service] - headers: [e2e-llm-inference-service] - name: X-Gateway-Model-Name [e2e-llm-inference-service] type: Exact [e2e-llm-inference-service] value: publishers/kserve-ci-e2e-test/models/lora-adapter-1 [e2e-llm-inference-service] path: [e2e-llm-inference-service] type: Exact [e2e-llm-inference-service] value: /v1/chat/completions/ [e2e-llm-inference-service] timeouts: [e2e-llm-inference-service] backendRequest: 0s [e2e-llm-inference-service] request: 0s [e2e-llm-inference-service] - backendRefs: [e2e-llm-inference-service] - group: inference.networking.k8s.io [e2e-llm-inference-service] kind: InferencePool [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-7ca60146-inference-pool [e2e-llm-inference-service] port: 8000 [e2e-llm-inference-service] weight: 1 [e2e-llm-inference-service] filters: [e2e-llm-inference-service] - type: URLRewrite [e2e-llm-inference-service] urlRewrite: [e2e-llm-inference-service] path: [e2e-llm-inference-service] replacePrefixMatch: /v1/responses [e2e-llm-inference-service] type: ReplacePrefixMatch [e2e-llm-inference-service] matches: [e2e-llm-inference-service] - path: [e2e-llm-inference-service] type: PathPrefix [e2e-llm-inference-service] value: /kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-with-7ca60146/v1/responses [e2e-llm-inference-service] timeouts: [e2e-llm-inference-service] backendRequest: 0s [e2e-llm-inference-service] request: 0s [e2e-llm-inference-service] - backendRefs: [e2e-llm-inference-service] - group: inference.networking.k8s.io [e2e-llm-inference-service] kind: InferencePool [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-7ca60146-inference-pool [e2e-llm-inference-service] port: 8000 [e2e-llm-inference-service] weight: 1 [e2e-llm-inference-service] matches: [e2e-llm-inference-service] - headers: [e2e-llm-inference-service] - name: X-Gateway-Model-Name [e2e-llm-inference-service] type: Exact [e2e-llm-inference-service] value: publishers/kserve-ci-e2e-test/models/facebook/opt-125m [e2e-llm-inference-service] path: [e2e-llm-inference-service] type: Exact [e2e-llm-inference-service] value: /v1/responses [e2e-llm-inference-service] - headers: [e2e-llm-inference-service] - name: X-Gateway-Model-Name [e2e-llm-inference-service] type: Exact [e2e-llm-inference-service] value: publishers/kserve-ci-e2e-test/models/facebook/opt-125m [e2e-llm-inference-service] path: [e2e-llm-inference-service] type: Exact [e2e-llm-inference-service] value: /v1/responses/ [e2e-llm-inference-service] - headers: [e2e-llm-inference-service] - name: X-Gateway-Model-Name [e2e-llm-inference-service] type: Exact [e2e-llm-inference-service] value: publishers/kserve-ci-e2e-test/models/lora-adapter-1 [e2e-llm-inference-service] path: [e2e-llm-inference-service] type: Exact [e2e-llm-inference-service] value: /v1/responses [e2e-llm-inference-service] - headers: [e2e-llm-inference-service] - name: X-Gateway-Model-Name [e2e-llm-inference-service] type: Exact [e2e-llm-inference-service] value: publishers/kserve-ci-e2e-test/models/lora-adapter-1 [e2e-llm-inference-service] path: [e2e-llm-inference-service] type: Exact [e2e-llm-inference-service] value: /v1/responses/ [e2e-llm-inference-service] timeouts: [e2e-llm-inference-service] backendRequest: 0s [e2e-llm-inference-service] request: 0s [e2e-llm-inference-service] - backendRefs: [e2e-llm-inference-service] - group: inference.networking.k8s.io [e2e-llm-inference-service] kind: InferencePool [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-7ca60146-inference-pool [e2e-llm-inference-service] port: 8000 [e2e-llm-inference-service] weight: 1 [e2e-llm-inference-service] filters: [e2e-llm-inference-service] - type: URLRewrite [e2e-llm-inference-service] urlRewrite: [e2e-llm-inference-service] path: [e2e-llm-inference-service] replacePrefixMatch: /v1/messages [e2e-llm-inference-service] type: ReplacePrefixMatch [e2e-llm-inference-service] matches: [e2e-llm-inference-service] - path: [e2e-llm-inference-service] type: PathPrefix [e2e-llm-inference-service] value: /kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-with-7ca60146/v1/messages [e2e-llm-inference-service] timeouts: [e2e-llm-inference-service] backendRequest: 0s [e2e-llm-inference-service] request: 0s [e2e-llm-inference-service] - backendRefs: [e2e-llm-inference-service] - group: inference.networking.k8s.io [e2e-llm-inference-service] kind: InferencePool [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-7ca60146-inference-pool [e2e-llm-inference-service] port: 8000 [e2e-llm-inference-service] weight: 1 [e2e-llm-inference-service] matches: [e2e-llm-inference-service] - headers: [e2e-llm-inference-service] - name: X-Gateway-Model-Name [e2e-llm-inference-service] type: Exact [e2e-llm-inference-service] value: publishers/kserve-ci-e2e-test/models/facebook/opt-125m [e2e-llm-inference-service] path: [e2e-llm-inference-service] type: Exact [e2e-llm-inference-service] value: /v1/messages [e2e-llm-inference-service] - headers: [e2e-llm-inference-service] - name: X-Gateway-Model-Name [e2e-llm-inference-service] type: Exact [e2e-llm-inference-service] value: publishers/kserve-ci-e2e-test/models/facebook/opt-125m [e2e-llm-inference-service] path: [e2e-llm-inference-service] type: Exact [e2e-llm-inference-service] value: /v1/messages/ [e2e-llm-inference-service] - headers: [e2e-llm-inference-service] - name: X-Gateway-Model-Name [e2e-llm-inference-service] type: Exact [e2e-llm-inference-service] value: publishers/kserve-ci-e2e-test/models/lora-adapter-1 [e2e-llm-inference-service] path: [e2e-llm-inference-service] type: Exact [e2e-llm-inference-service] value: /v1/messages [e2e-llm-inference-service] - headers: [e2e-llm-inference-service] - name: X-Gateway-Model-Name [e2e-llm-inference-service] type: Exact [e2e-llm-inference-service] value: publishers/kserve-ci-e2e-test/models/lora-adapter-1 [e2e-llm-inference-service] path: [e2e-llm-inference-service] type: Exact [e2e-llm-inference-service] value: /v1/messages/ [e2e-llm-inference-service] timeouts: [e2e-llm-inference-service] backendRequest: 0s [e2e-llm-inference-service] request: 0s [e2e-llm-inference-service] - backendRefs: [e2e-llm-inference-service] - group: '' [e2e-llm-inference-service] kind: Service [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-7ca60146-kserve-workload-svc [e2e-llm-inference-service] port: 8000 [e2e-llm-inference-service] weight: 1 [e2e-llm-inference-service] filters: [e2e-llm-inference-service] - type: URLRewrite [e2e-llm-inference-service] urlRewrite: [e2e-llm-inference-service] path: [e2e-llm-inference-service] replacePrefixMatch: / [e2e-llm-inference-service] type: ReplacePrefixMatch [e2e-llm-inference-service] matches: [e2e-llm-inference-service] - path: [e2e-llm-inference-service] type: PathPrefix [e2e-llm-inference-service] value: /kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-with-7ca60146 [e2e-llm-inference-service] timeouts: [e2e-llm-inference-service] backendRequest: 0s [e2e-llm-inference-service] request: 0s [e2e-llm-inference-service] - backendRefs: [e2e-llm-inference-service] - group: '' [e2e-llm-inference-service] kind: Service [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-7ca60146-kserve-workload-svc [e2e-llm-inference-service] port: 8000 [e2e-llm-inference-service] weight: 1 [e2e-llm-inference-service] matches: [e2e-llm-inference-service] - headers: [e2e-llm-inference-service] - name: X-Gateway-Model-Name [e2e-llm-inference-service] type: Exact [e2e-llm-inference-service] value: publishers/kserve-ci-e2e-test/models/facebook/opt-125m [e2e-llm-inference-service] path: [e2e-llm-inference-service] type: PathPrefix [e2e-llm-inference-service] value: / [e2e-llm-inference-service] - headers: [e2e-llm-inference-service] - name: X-Gateway-Model-Name [e2e-llm-inference-service] type: Exact [e2e-llm-inference-service] value: publishers/kserve-ci-e2e-test/models/lora-adapter-1 [e2e-llm-inference-service] path: [e2e-llm-inference-service] type: PathPrefix [e2e-llm-inference-service] value: / [e2e-llm-inference-service] timeouts: [e2e-llm-inference-service] backendRequest: 0s [e2e-llm-inference-service] request: 0s [e2e-llm-inference-service] status: [e2e-llm-inference-service] parents: [e2e-llm-inference-service] - conditions: [e2e-llm-inference-service] - lastTransitionTime: '2026-07-07T14:29:25Z' [e2e-llm-inference-service] message: Route was valid [e2e-llm-inference-service] observedGeneration: 2 [e2e-llm-inference-service] reason: Accepted [e2e-llm-inference-service] status: 'True' [e2e-llm-inference-service] type: Accepted [e2e-llm-inference-service] - lastTransitionTime: '2026-07-07T14:29:34Z' [e2e-llm-inference-service] message: All references resolved [e2e-llm-inference-service] observedGeneration: 2 [e2e-llm-inference-service] reason: ResolvedRefs [e2e-llm-inference-service] status: 'True' [e2e-llm-inference-service] type: ResolvedRefs [e2e-llm-inference-service] controllerName: openshift.io/gateway-controller/v1 [e2e-llm-inference-service] parentRef: [e2e-llm-inference-service] group: gateway.networking.k8s.io [e2e-llm-inference-service] kind: Gateway [e2e-llm-inference-service] name: openshift-ai-inference [e2e-llm-inference-service] namespace: openshift-ingress [e2e-llm-inference-service] - conditions: [e2e-llm-inference-service] - lastTransitionTime: '2026-07-07T14:29:25Z' [e2e-llm-inference-service] message: Object affected by AuthPolicy [kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-with-7ca60146-kserve-route-authn [e2e-llm-inference-service] openshift-ingress/openshift-ai-inference-authn] [e2e-llm-inference-service] observedGeneration: 2 [e2e-llm-inference-service] reason: Accepted [e2e-llm-inference-service] status: 'True' [e2e-llm-inference-service] type: kuadrant.io/AuthPolicyAffected [e2e-llm-inference-service] controllerName: kuadrant.io/policy-controller [e2e-llm-inference-service] parentRef: [e2e-llm-inference-service] group: gateway.networking.k8s.io [e2e-llm-inference-service] kind: Gateway [e2e-llm-inference-service] name: openshift-ai-inference [e2e-llm-inference-service] namespace: openshift-ingress [e2e-llm-inference-service] [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1254 --- [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1255 apiVersion: inference.networking.k8s.io/v1 [e2e-llm-inference-service] kind: InferencePool [e2e-llm-inference-service] metadata: [e2e-llm-inference-service] creationTimestamp: '2026-07-07T14:29:24Z' [e2e-llm-inference-service] generation: 1 [e2e-llm-inference-service] labels: [e2e-llm-inference-service] app.kubernetes.io/component: llminferenceservice-router-scheduler [e2e-llm-inference-service] app.kubernetes.io/name: llmisvc-model-fb-opt-125m-with-7ca60146 [e2e-llm-inference-service] app.kubernetes.io/part-of: llminferenceservice [e2e-llm-inference-service] managedFields: [e2e-llm-inference-service] - apiVersion: inference.networking.k8s.io/v1 [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:metadata: [e2e-llm-inference-service] f:labels: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:app.kubernetes.io/component: {} [e2e-llm-inference-service] f:app.kubernetes.io/name: {} [e2e-llm-inference-service] f:app.kubernetes.io/part-of: {} [e2e-llm-inference-service] f:ownerReferences: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"uid":"e15cfc59-e4a1-4876-b43b-85ce3b1b239b"}: {} [e2e-llm-inference-service] f:spec: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:appProtocol: {} [e2e-llm-inference-service] f:endpointPickerRef: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:failureMode: {} [e2e-llm-inference-service] f:group: {} [e2e-llm-inference-service] f:kind: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:port: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:number: {} [e2e-llm-inference-service] f:selector: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:matchLabels: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:app.kubernetes.io/name: {} [e2e-llm-inference-service] f:app.kubernetes.io/part-of: {} [e2e-llm-inference-service] f:kserve.io/component: {} [e2e-llm-inference-service] f:targetPorts: {} [e2e-llm-inference-service] manager: manager [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] time: '2026-07-07T14:29:24Z' [e2e-llm-inference-service] - apiVersion: inference.networking.k8s.io/v1 [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:status: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:parents: {} [e2e-llm-inference-service] manager: pilot-discovery [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] subresource: status [e2e-llm-inference-service] time: '2026-07-07T14:29:34Z' [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-7ca60146-inference-pool [e2e-llm-inference-service] namespace: kserve-ci-e2e-test [e2e-llm-inference-service] ownerReferences: [e2e-llm-inference-service] - apiVersion: serving.kserve.io/v1alpha2 [e2e-llm-inference-service] blockOwnerDeletion: true [e2e-llm-inference-service] controller: true [e2e-llm-inference-service] kind: LLMInferenceService [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-7ca60146 [e2e-llm-inference-service] uid: e15cfc59-e4a1-4876-b43b-85ce3b1b239b [e2e-llm-inference-service] resourceVersion: '28513' [e2e-llm-inference-service] uid: af92b2a2-60e2-426b-8bc7-99693e01e6f6 [e2e-llm-inference-service] spec: [e2e-llm-inference-service] appProtocol: http [e2e-llm-inference-service] endpointPickerRef: [e2e-llm-inference-service] failureMode: FailOpen [e2e-llm-inference-service] group: '' [e2e-llm-inference-service] kind: Service [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-7ca60146-epp-service [e2e-llm-inference-service] port: [e2e-llm-inference-service] number: 9002 [e2e-llm-inference-service] selector: [e2e-llm-inference-service] matchLabels: [e2e-llm-inference-service] app.kubernetes.io/name: llmisvc-model-fb-opt-125m-with-7ca60146 [e2e-llm-inference-service] app.kubernetes.io/part-of: llminferenceservice [e2e-llm-inference-service] kserve.io/component: workload [e2e-llm-inference-service] targetPorts: [e2e-llm-inference-service] - number: 8000 [e2e-llm-inference-service] status: [e2e-llm-inference-service] parents: [e2e-llm-inference-service] - conditions: [e2e-llm-inference-service] - lastTransitionTime: '2026-07-07T14:29:34Z' [e2e-llm-inference-service] message: Referenced by an HTTPRoute accepted by the parentRef Gateway [e2e-llm-inference-service] observedGeneration: 1 [e2e-llm-inference-service] reason: Accepted [e2e-llm-inference-service] status: 'True' [e2e-llm-inference-service] type: Accepted [e2e-llm-inference-service] - lastTransitionTime: '2026-07-07T14:29:34Z' [e2e-llm-inference-service] message: Referenced ExtensionRef resolved successfully [e2e-llm-inference-service] observedGeneration: 1 [e2e-llm-inference-service] reason: ResolvedRefs [e2e-llm-inference-service] status: 'True' [e2e-llm-inference-service] type: ResolvedRefs [e2e-llm-inference-service] parentRef: [e2e-llm-inference-service] group: networking.istio.io [e2e-llm-inference-service] kind: Gateway [e2e-llm-inference-service] name: openshift-ai-inference [e2e-llm-inference-service] namespace: openshift-ingress [e2e-llm-inference-service] [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1254 --- [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1255 apiVersion: kuadrant.io/v1 [e2e-llm-inference-service] kind: AuthPolicy [e2e-llm-inference-service] metadata: [e2e-llm-inference-service] creationTimestamp: '2026-07-07T14:29:26Z' [e2e-llm-inference-service] generation: 1 [e2e-llm-inference-service] labels: [e2e-llm-inference-service] app.kubernetes.io/component: llminferenceservice-policies [e2e-llm-inference-service] app.kubernetes.io/managed-by: odh-model-controller [e2e-llm-inference-service] app.kubernetes.io/name: llmisvc-model-fb-opt-125m-with-7ca60146 [e2e-llm-inference-service] app.kubernetes.io/part-of: llminferenceservice [e2e-llm-inference-service] managedFields: [e2e-llm-inference-service] - apiVersion: kuadrant.io/v1 [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:metadata: [e2e-llm-inference-service] f:labels: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:app.kubernetes.io/component: {} [e2e-llm-inference-service] f:app.kubernetes.io/managed-by: {} [e2e-llm-inference-service] f:app.kubernetes.io/name: {} [e2e-llm-inference-service] f:app.kubernetes.io/part-of: {} [e2e-llm-inference-service] f:ownerReferences: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"uid":"e15cfc59-e4a1-4876-b43b-85ce3b1b239b"}: {} [e2e-llm-inference-service] f:spec: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:rules: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:authentication: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:public: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:anonymous: {} [e2e-llm-inference-service] f:credentials: {} [e2e-llm-inference-service] f:metrics: {} [e2e-llm-inference-service] f:overrides: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:fairness: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] f:objective: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] f:priority: {} [e2e-llm-inference-service] f:response: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:success: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:headers: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:x-gateway-inference-fairness-id: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:metrics: {} [e2e-llm-inference-service] f:plain: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:expression: {} [e2e-llm-inference-service] f:priority: {} [e2e-llm-inference-service] f:x-gateway-inference-objective: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:metrics: {} [e2e-llm-inference-service] f:plain: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:expression: {} [e2e-llm-inference-service] f:priority: {} [e2e-llm-inference-service] f:targetRef: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:group: {} [e2e-llm-inference-service] f:kind: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] manager: manager [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] time: '2026-07-07T14:29:26Z' [e2e-llm-inference-service] - apiVersion: kuadrant.io/v1 [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:status: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:conditions: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"type":"Accepted"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:lastTransitionTime: {} [e2e-llm-inference-service] f:message: {} [e2e-llm-inference-service] f:reason: {} [e2e-llm-inference-service] f:status: {} [e2e-llm-inference-service] f:type: {} [e2e-llm-inference-service] k:{"type":"Enforced"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:lastTransitionTime: {} [e2e-llm-inference-service] f:message: {} [e2e-llm-inference-service] f:reason: {} [e2e-llm-inference-service] f:status: {} [e2e-llm-inference-service] f:type: {} [e2e-llm-inference-service] f:observedGeneration: {} [e2e-llm-inference-service] manager: manager [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] subresource: status [e2e-llm-inference-service] time: '2026-07-07T14:29:28Z' [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-7ca60146-kserve-route-authn [e2e-llm-inference-service] namespace: kserve-ci-e2e-test [e2e-llm-inference-service] ownerReferences: [e2e-llm-inference-service] - apiVersion: serving.kserve.io/v1alpha2 [e2e-llm-inference-service] blockOwnerDeletion: true [e2e-llm-inference-service] controller: true [e2e-llm-inference-service] kind: LLMInferenceService [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-7ca60146 [e2e-llm-inference-service] uid: e15cfc59-e4a1-4876-b43b-85ce3b1b239b [e2e-llm-inference-service] resourceVersion: '28405' [e2e-llm-inference-service] uid: e4a846e1-103b-47b7-a4f2-bebaf66de2e5 [e2e-llm-inference-service] spec: [e2e-llm-inference-service] rules: [e2e-llm-inference-service] authentication: [e2e-llm-inference-service] public: [e2e-llm-inference-service] anonymous: {} [e2e-llm-inference-service] credentials: {} [e2e-llm-inference-service] metrics: false [e2e-llm-inference-service] overrides: [e2e-llm-inference-service] fairness: [e2e-llm-inference-service] value: unauthenticated [e2e-llm-inference-service] objective: [e2e-llm-inference-service] value: unauthenticated [e2e-llm-inference-service] priority: 0 [e2e-llm-inference-service] response: [e2e-llm-inference-service] success: [e2e-llm-inference-service] headers: [e2e-llm-inference-service] x-gateway-inference-fairness-id: [e2e-llm-inference-service] metrics: false [e2e-llm-inference-service] plain: [e2e-llm-inference-service] expression: auth.identity.fairness [e2e-llm-inference-service] priority: 0 [e2e-llm-inference-service] x-gateway-inference-objective: [e2e-llm-inference-service] metrics: false [e2e-llm-inference-service] plain: [e2e-llm-inference-service] expression: auth.identity.objective [e2e-llm-inference-service] priority: 0 [e2e-llm-inference-service] targetRef: [e2e-llm-inference-service] group: gateway.networking.k8s.io [e2e-llm-inference-service] kind: HTTPRoute [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-7ca60146-kserve-route [e2e-llm-inference-service] status: [e2e-llm-inference-service] conditions: [e2e-llm-inference-service] - lastTransitionTime: '2026-07-07T14:29:27Z' [e2e-llm-inference-service] message: AuthPolicy has been accepted [e2e-llm-inference-service] reason: Accepted [e2e-llm-inference-service] status: 'True' [e2e-llm-inference-service] type: Accepted [e2e-llm-inference-service] - lastTransitionTime: '2026-07-07T14:29:28Z' [e2e-llm-inference-service] message: AuthPolicy has been successfully enforced [e2e-llm-inference-service] reason: Enforced [e2e-llm-inference-service] status: 'True' [e2e-llm-inference-service] type: Enforced [e2e-llm-inference-service] observedGeneration: 1 [e2e-llm-inference-service] [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1254 --- [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1255 apiVersion: networking.istio.io/v1 [e2e-llm-inference-service] kind: DestinationRule [e2e-llm-inference-service] metadata: [e2e-llm-inference-service] creationTimestamp: '2026-07-07T14:29:25Z' [e2e-llm-inference-service] generation: 1 [e2e-llm-inference-service] labels: [e2e-llm-inference-service] app.kubernetes.io/component: llminferenceservice-router-scheduler [e2e-llm-inference-service] app.kubernetes.io/name: llmisvc-model-fb-opt-125m-with-7ca60146 [e2e-llm-inference-service] app.kubernetes.io/part-of: llminferenceservice [e2e-llm-inference-service] llm-d.ai/managed: 'true' [e2e-llm-inference-service] managedFields: [e2e-llm-inference-service] - apiVersion: networking.istio.io/v1 [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:metadata: [e2e-llm-inference-service] f:labels: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:app.kubernetes.io/component: {} [e2e-llm-inference-service] f:app.kubernetes.io/name: {} [e2e-llm-inference-service] f:app.kubernetes.io/part-of: {} [e2e-llm-inference-service] f:llm-d.ai/managed: {} [e2e-llm-inference-service] f:ownerReferences: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"uid":"e15cfc59-e4a1-4876-b43b-85ce3b1b239b"}: {} [e2e-llm-inference-service] f:spec: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:exportTo: {} [e2e-llm-inference-service] f:host: {} [e2e-llm-inference-service] f:trafficPolicy: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:tls: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:insecureSkipVerify: {} [e2e-llm-inference-service] f:mode: {} [e2e-llm-inference-service] f:sni: {} [e2e-llm-inference-service] manager: manager [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] time: '2026-07-07T14:29:25Z' [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-7ca60146-kserve-scheduler [e2e-llm-inference-service] namespace: kserve-ci-e2e-test [e2e-llm-inference-service] ownerReferences: [e2e-llm-inference-service] - apiVersion: serving.kserve.io/v1alpha2 [e2e-llm-inference-service] blockOwnerDeletion: true [e2e-llm-inference-service] controller: true [e2e-llm-inference-service] kind: LLMInferenceService [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-7ca60146 [e2e-llm-inference-service] uid: e15cfc59-e4a1-4876-b43b-85ce3b1b239b [e2e-llm-inference-service] resourceVersion: '28290' [e2e-llm-inference-service] uid: fce433b3-6aa1-468a-b634-3316fe5fcbbc [e2e-llm-inference-service] spec: [e2e-llm-inference-service] exportTo: [e2e-llm-inference-service] - '*' [e2e-llm-inference-service] host: llmisvc-model-fb-opt-125m-with-7ca60146-epp-service.kserve-ci-e2e-test.svc.cluster.local [e2e-llm-inference-service] trafficPolicy: [e2e-llm-inference-service] tls: [e2e-llm-inference-service] insecureSkipVerify: true [e2e-llm-inference-service] mode: SIMPLE [e2e-llm-inference-service] sni: llmisvc-model-fb-opt-125m-with-7ca60146-epp-service.kserve-ci-e2e-test.svc.cluster.local [e2e-llm-inference-service] [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1254 --- [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1255 apiVersion: networking.istio.io/v1 [e2e-llm-inference-service] kind: DestinationRule [e2e-llm-inference-service] metadata: [e2e-llm-inference-service] creationTimestamp: '2026-07-07T14:29:34Z' [e2e-llm-inference-service] generation: 1 [e2e-llm-inference-service] labels: [e2e-llm-inference-service] app.kubernetes.io/component: llminferenceservice-shadow-service [e2e-llm-inference-service] app.kubernetes.io/name: llmisvc-model-fb-opt-125m-with-7ca60146 [e2e-llm-inference-service] app.kubernetes.io/part-of: llminferenceservice [e2e-llm-inference-service] llm-d.ai/managed: 'true' [e2e-llm-inference-service] managedFields: [e2e-llm-inference-service] - apiVersion: networking.istio.io/v1 [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:metadata: [e2e-llm-inference-service] f:labels: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:app.kubernetes.io/component: {} [e2e-llm-inference-service] f:app.kubernetes.io/name: {} [e2e-llm-inference-service] f:app.kubernetes.io/part-of: {} [e2e-llm-inference-service] f:llm-d.ai/managed: {} [e2e-llm-inference-service] f:ownerReferences: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"uid":"e15cfc59-e4a1-4876-b43b-85ce3b1b239b"}: {} [e2e-llm-inference-service] f:spec: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:exportTo: {} [e2e-llm-inference-service] f:host: {} [e2e-llm-inference-service] f:trafficPolicy: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:tls: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:insecureSkipVerify: {} [e2e-llm-inference-service] f:mode: {} [e2e-llm-inference-service] f:sni: {} [e2e-llm-inference-service] manager: manager [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] time: '2026-07-07T14:29:34Z' [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-7ca60146-kserve-shadow-svc [e2e-llm-inference-service] namespace: kserve-ci-e2e-test [e2e-llm-inference-service] ownerReferences: [e2e-llm-inference-service] - apiVersion: serving.kserve.io/v1alpha2 [e2e-llm-inference-service] blockOwnerDeletion: true [e2e-llm-inference-service] controller: true [e2e-llm-inference-service] kind: LLMInferenceService [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-7ca60146 [e2e-llm-inference-service] uid: e15cfc59-e4a1-4876-b43b-85ce3b1b239b [e2e-llm-inference-service] resourceVersion: '28525' [e2e-llm-inference-service] uid: 50be281a-9e88-4284-b624-9ee35607afef [e2e-llm-inference-service] spec: [e2e-llm-inference-service] exportTo: [e2e-llm-inference-service] - '*' [e2e-llm-inference-service] host: llmisvc-model-fb-opt-125m-with-7ca60146-inference-p-ip-16c62f55.kserve-ci-e2e-test.svc.cluster.local [e2e-llm-inference-service] trafficPolicy: [e2e-llm-inference-service] tls: [e2e-llm-inference-service] insecureSkipVerify: true [e2e-llm-inference-service] mode: SIMPLE [e2e-llm-inference-service] sni: llmisvc-model-fb-opt-125m-with-7ca60146-kserve-workload-svc.kserve-ci-e2e-test.svc.cluster.local [e2e-llm-inference-service] [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1254 --- [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1255 apiVersion: networking.istio.io/v1 [e2e-llm-inference-service] kind: DestinationRule [e2e-llm-inference-service] metadata: [e2e-llm-inference-service] creationTimestamp: '2026-07-07T14:29:25Z' [e2e-llm-inference-service] generation: 1 [e2e-llm-inference-service] labels: [e2e-llm-inference-service] app.kubernetes.io/component: llminferenceservice-workload [e2e-llm-inference-service] app.kubernetes.io/name: llmisvc-model-fb-opt-125m-with-7ca60146 [e2e-llm-inference-service] app.kubernetes.io/part-of: llminferenceservice [e2e-llm-inference-service] llm-d.ai/managed: 'true' [e2e-llm-inference-service] managedFields: [e2e-llm-inference-service] - apiVersion: networking.istio.io/v1 [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:metadata: [e2e-llm-inference-service] f:labels: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:app.kubernetes.io/component: {} [e2e-llm-inference-service] f:app.kubernetes.io/name: {} [e2e-llm-inference-service] f:app.kubernetes.io/part-of: {} [e2e-llm-inference-service] f:llm-d.ai/managed: {} [e2e-llm-inference-service] f:ownerReferences: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"uid":"e15cfc59-e4a1-4876-b43b-85ce3b1b239b"}: {} [e2e-llm-inference-service] f:spec: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:exportTo: {} [e2e-llm-inference-service] f:host: {} [e2e-llm-inference-service] f:trafficPolicy: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:tls: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:caCertificates: {} [e2e-llm-inference-service] f:insecureSkipVerify: {} [e2e-llm-inference-service] f:mode: {} [e2e-llm-inference-service] f:sni: {} [e2e-llm-inference-service] manager: manager [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] time: '2026-07-07T14:29:25Z' [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-7ca60146-kserve-workload-svc [e2e-llm-inference-service] namespace: kserve-ci-e2e-test [e2e-llm-inference-service] ownerReferences: [e2e-llm-inference-service] - apiVersion: serving.kserve.io/v1alpha2 [e2e-llm-inference-service] blockOwnerDeletion: true [e2e-llm-inference-service] controller: true [e2e-llm-inference-service] kind: LLMInferenceService [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-7ca60146 [e2e-llm-inference-service] uid: e15cfc59-e4a1-4876-b43b-85ce3b1b239b [e2e-llm-inference-service] resourceVersion: '28322' [e2e-llm-inference-service] uid: eb5b5cb3-209a-4843-b33b-5dbfc7721e67 [e2e-llm-inference-service] spec: [e2e-llm-inference-service] exportTo: [e2e-llm-inference-service] - '*' [e2e-llm-inference-service] host: llmisvc-model-fb-opt-125m-with-7ca60146-kserve-workload-svc.kserve-ci-e2e-test.svc.cluster.local [e2e-llm-inference-service] trafficPolicy: [e2e-llm-inference-service] tls: [e2e-llm-inference-service] caCertificates: /var/run/secrets/kubernetes.io/serviceaccount/service-ca.crt [e2e-llm-inference-service] insecureSkipVerify: false [e2e-llm-inference-service] mode: SIMPLE [e2e-llm-inference-service] sni: llmisvc-model-fb-opt-125m-with-7ca60146-kserve-workload-svc.kserve-ci-e2e-test.svc.cluster.local [e2e-llm-inference-service] [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1254 --- [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1255 apiVersion: networking.istio.io/v1beta1 [e2e-llm-inference-service] kind: DestinationRule [e2e-llm-inference-service] metadata: [e2e-llm-inference-service] creationTimestamp: '2026-07-07T14:29:25Z' [e2e-llm-inference-service] generation: 1 [e2e-llm-inference-service] labels: [e2e-llm-inference-service] app.kubernetes.io/component: llminferenceservice-router-scheduler [e2e-llm-inference-service] app.kubernetes.io/name: llmisvc-model-fb-opt-125m-with-7ca60146 [e2e-llm-inference-service] app.kubernetes.io/part-of: llminferenceservice [e2e-llm-inference-service] llm-d.ai/managed: 'true' [e2e-llm-inference-service] managedFields: [e2e-llm-inference-service] - apiVersion: networking.istio.io/v1 [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:metadata: [e2e-llm-inference-service] f:labels: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:app.kubernetes.io/component: {} [e2e-llm-inference-service] f:app.kubernetes.io/name: {} [e2e-llm-inference-service] f:app.kubernetes.io/part-of: {} [e2e-llm-inference-service] f:llm-d.ai/managed: {} [e2e-llm-inference-service] f:ownerReferences: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"uid":"e15cfc59-e4a1-4876-b43b-85ce3b1b239b"}: {} [e2e-llm-inference-service] f:spec: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:exportTo: {} [e2e-llm-inference-service] f:host: {} [e2e-llm-inference-service] f:trafficPolicy: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:tls: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:insecureSkipVerify: {} [e2e-llm-inference-service] f:mode: {} [e2e-llm-inference-service] f:sni: {} [e2e-llm-inference-service] manager: manager [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] time: '2026-07-07T14:29:25Z' [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-7ca60146-kserve-scheduler [e2e-llm-inference-service] namespace: kserve-ci-e2e-test [e2e-llm-inference-service] ownerReferences: [e2e-llm-inference-service] - apiVersion: serving.kserve.io/v1alpha2 [e2e-llm-inference-service] blockOwnerDeletion: true [e2e-llm-inference-service] controller: true [e2e-llm-inference-service] kind: LLMInferenceService [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-7ca60146 [e2e-llm-inference-service] uid: e15cfc59-e4a1-4876-b43b-85ce3b1b239b [e2e-llm-inference-service] resourceVersion: '28290' [e2e-llm-inference-service] uid: fce433b3-6aa1-468a-b634-3316fe5fcbbc [e2e-llm-inference-service] spec: [e2e-llm-inference-service] exportTo: [e2e-llm-inference-service] - '*' [e2e-llm-inference-service] host: llmisvc-model-fb-opt-125m-with-7ca60146-epp-service.kserve-ci-e2e-test.svc.cluster.local [e2e-llm-inference-service] trafficPolicy: [e2e-llm-inference-service] tls: [e2e-llm-inference-service] insecureSkipVerify: true [e2e-llm-inference-service] mode: SIMPLE [e2e-llm-inference-service] sni: llmisvc-model-fb-opt-125m-with-7ca60146-epp-service.kserve-ci-e2e-test.svc.cluster.local [e2e-llm-inference-service] [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1254 --- [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1255 apiVersion: networking.istio.io/v1beta1 [e2e-llm-inference-service] kind: DestinationRule [e2e-llm-inference-service] metadata: [e2e-llm-inference-service] creationTimestamp: '2026-07-07T14:29:34Z' [e2e-llm-inference-service] generation: 1 [e2e-llm-inference-service] labels: [e2e-llm-inference-service] app.kubernetes.io/component: llminferenceservice-shadow-service [e2e-llm-inference-service] app.kubernetes.io/name: llmisvc-model-fb-opt-125m-with-7ca60146 [e2e-llm-inference-service] app.kubernetes.io/part-of: llminferenceservice [e2e-llm-inference-service] llm-d.ai/managed: 'true' [e2e-llm-inference-service] managedFields: [e2e-llm-inference-service] - apiVersion: networking.istio.io/v1 [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:metadata: [e2e-llm-inference-service] f:labels: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:app.kubernetes.io/component: {} [e2e-llm-inference-service] f:app.kubernetes.io/name: {} [e2e-llm-inference-service] f:app.kubernetes.io/part-of: {} [e2e-llm-inference-service] f:llm-d.ai/managed: {} [e2e-llm-inference-service] f:ownerReferences: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"uid":"e15cfc59-e4a1-4876-b43b-85ce3b1b239b"}: {} [e2e-llm-inference-service] f:spec: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:exportTo: {} [e2e-llm-inference-service] f:host: {} [e2e-llm-inference-service] f:trafficPolicy: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:tls: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:insecureSkipVerify: {} [e2e-llm-inference-service] f:mode: {} [e2e-llm-inference-service] f:sni: {} [e2e-llm-inference-service] manager: manager [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] time: '2026-07-07T14:29:34Z' [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-7ca60146-kserve-shadow-svc [e2e-llm-inference-service] namespace: kserve-ci-e2e-test [e2e-llm-inference-service] ownerReferences: [e2e-llm-inference-service] - apiVersion: serving.kserve.io/v1alpha2 [e2e-llm-inference-service] blockOwnerDeletion: true [e2e-llm-inference-service] controller: true [e2e-llm-inference-service] kind: LLMInferenceService [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-7ca60146 [e2e-llm-inference-service] uid: e15cfc59-e4a1-4876-b43b-85ce3b1b239b [e2e-llm-inference-service] resourceVersion: '28525' [e2e-llm-inference-service] uid: 50be281a-9e88-4284-b624-9ee35607afef [e2e-llm-inference-service] spec: [e2e-llm-inference-service] exportTo: [e2e-llm-inference-service] - '*' [e2e-llm-inference-service] host: llmisvc-model-fb-opt-125m-with-7ca60146-inference-p-ip-16c62f55.kserve-ci-e2e-test.svc.cluster.local [e2e-llm-inference-service] trafficPolicy: [e2e-llm-inference-service] tls: [e2e-llm-inference-service] insecureSkipVerify: true [e2e-llm-inference-service] mode: SIMPLE [e2e-llm-inference-service] sni: llmisvc-model-fb-opt-125m-with-7ca60146-kserve-workload-svc.kserve-ci-e2e-test.svc.cluster.local [e2e-llm-inference-service] [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1254 --- [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1255 apiVersion: networking.istio.io/v1beta1 [e2e-llm-inference-service] kind: DestinationRule [e2e-llm-inference-service] metadata: [e2e-llm-inference-service] creationTimestamp: '2026-07-07T14:29:25Z' [e2e-llm-inference-service] generation: 1 [e2e-llm-inference-service] labels: [e2e-llm-inference-service] app.kubernetes.io/component: llminferenceservice-workload [e2e-llm-inference-service] app.kubernetes.io/name: llmisvc-model-fb-opt-125m-with-7ca60146 [e2e-llm-inference-service] app.kubernetes.io/part-of: llminferenceservice [e2e-llm-inference-service] llm-d.ai/managed: 'true' [e2e-llm-inference-service] managedFields: [e2e-llm-inference-service] - apiVersion: networking.istio.io/v1 [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:metadata: [e2e-llm-inference-service] f:labels: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:app.kubernetes.io/component: {} [e2e-llm-inference-service] f:app.kubernetes.io/name: {} [e2e-llm-inference-service] f:app.kubernetes.io/part-of: {} [e2e-llm-inference-service] f:llm-d.ai/managed: {} [e2e-llm-inference-service] f:ownerReferences: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"uid":"e15cfc59-e4a1-4876-b43b-85ce3b1b239b"}: {} [e2e-llm-inference-service] f:spec: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:exportTo: {} [e2e-llm-inference-service] f:host: {} [e2e-llm-inference-service] f:trafficPolicy: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:tls: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:caCertificates: {} [e2e-llm-inference-service] f:insecureSkipVerify: {} [e2e-llm-inference-service] f:mode: {} [e2e-llm-inference-service] f:sni: {} [e2e-llm-inference-service] manager: manager [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] time: '2026-07-07T14:29:25Z' [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-7ca60146-kserve-workload-svc [e2e-llm-inference-service] namespace: kserve-ci-e2e-test [e2e-llm-inference-service] ownerReferences: [e2e-llm-inference-service] - apiVersion: serving.kserve.io/v1alpha2 [e2e-llm-inference-service] blockOwnerDeletion: true [e2e-llm-inference-service] controller: true [e2e-llm-inference-service] kind: LLMInferenceService [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-7ca60146 [e2e-llm-inference-service] uid: e15cfc59-e4a1-4876-b43b-85ce3b1b239b [e2e-llm-inference-service] resourceVersion: '28322' [e2e-llm-inference-service] uid: eb5b5cb3-209a-4843-b33b-5dbfc7721e67 [e2e-llm-inference-service] spec: [e2e-llm-inference-service] exportTo: [e2e-llm-inference-service] - '*' [e2e-llm-inference-service] host: llmisvc-model-fb-opt-125m-with-7ca60146-kserve-workload-svc.kserve-ci-e2e-test.svc.cluster.local [e2e-llm-inference-service] trafficPolicy: [e2e-llm-inference-service] tls: [e2e-llm-inference-service] caCertificates: /var/run/secrets/kubernetes.io/serviceaccount/service-ca.crt [e2e-llm-inference-service] insecureSkipVerify: false [e2e-llm-inference-service] mode: SIMPLE [e2e-llm-inference-service] sni: llmisvc-model-fb-opt-125m-with-7ca60146-kserve-workload-svc.kserve-ci-e2e-test.svc.cluster.local [e2e-llm-inference-service] [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1254 --- [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1255 apiVersion: networking.istio.io/v1alpha3 [e2e-llm-inference-service] kind: DestinationRule [e2e-llm-inference-service] metadata: [e2e-llm-inference-service] creationTimestamp: '2026-07-07T14:29:25Z' [e2e-llm-inference-service] generation: 1 [e2e-llm-inference-service] labels: [e2e-llm-inference-service] app.kubernetes.io/component: llminferenceservice-router-scheduler [e2e-llm-inference-service] app.kubernetes.io/name: llmisvc-model-fb-opt-125m-with-7ca60146 [e2e-llm-inference-service] app.kubernetes.io/part-of: llminferenceservice [e2e-llm-inference-service] llm-d.ai/managed: 'true' [e2e-llm-inference-service] managedFields: [e2e-llm-inference-service] - apiVersion: networking.istio.io/v1 [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:metadata: [e2e-llm-inference-service] f:labels: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:app.kubernetes.io/component: {} [e2e-llm-inference-service] f:app.kubernetes.io/name: {} [e2e-llm-inference-service] f:app.kubernetes.io/part-of: {} [e2e-llm-inference-service] f:llm-d.ai/managed: {} [e2e-llm-inference-service] f:ownerReferences: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"uid":"e15cfc59-e4a1-4876-b43b-85ce3b1b239b"}: {} [e2e-llm-inference-service] f:spec: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:exportTo: {} [e2e-llm-inference-service] f:host: {} [e2e-llm-inference-service] f:trafficPolicy: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:tls: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:insecureSkipVerify: {} [e2e-llm-inference-service] f:mode: {} [e2e-llm-inference-service] f:sni: {} [e2e-llm-inference-service] manager: manager [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] time: '2026-07-07T14:29:25Z' [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-7ca60146-kserve-scheduler [e2e-llm-inference-service] namespace: kserve-ci-e2e-test [e2e-llm-inference-service] ownerReferences: [e2e-llm-inference-service] - apiVersion: serving.kserve.io/v1alpha2 [e2e-llm-inference-service] blockOwnerDeletion: true [e2e-llm-inference-service] controller: true [e2e-llm-inference-service] kind: LLMInferenceService [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-7ca60146 [e2e-llm-inference-service] uid: e15cfc59-e4a1-4876-b43b-85ce3b1b239b [e2e-llm-inference-service] resourceVersion: '28290' [e2e-llm-inference-service] uid: fce433b3-6aa1-468a-b634-3316fe5fcbbc [e2e-llm-inference-service] spec: [e2e-llm-inference-service] exportTo: [e2e-llm-inference-service] - '*' [e2e-llm-inference-service] host: llmisvc-model-fb-opt-125m-with-7ca60146-epp-service.kserve-ci-e2e-test.svc.cluster.local [e2e-llm-inference-service] trafficPolicy: [e2e-llm-inference-service] tls: [e2e-llm-inference-service] insecureSkipVerify: true [e2e-llm-inference-service] mode: SIMPLE [e2e-llm-inference-service] sni: llmisvc-model-fb-opt-125m-with-7ca60146-epp-service.kserve-ci-e2e-test.svc.cluster.local [e2e-llm-inference-service] [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1254 --- [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1255 apiVersion: networking.istio.io/v1alpha3 [e2e-llm-inference-service] kind: DestinationRule [e2e-llm-inference-service] metadata: [e2e-llm-inference-service] creationTimestamp: '2026-07-07T14:29:34Z' [e2e-llm-inference-service] generation: 1 [e2e-llm-inference-service] labels: [e2e-llm-inference-service] app.kubernetes.io/component: llminferenceservice-shadow-service [e2e-llm-inference-service] app.kubernetes.io/name: llmisvc-model-fb-opt-125m-with-7ca60146 [e2e-llm-inference-service] app.kubernetes.io/part-of: llminferenceservice [e2e-llm-inference-service] llm-d.ai/managed: 'true' [e2e-llm-inference-service] managedFields: [e2e-llm-inference-service] - apiVersion: networking.istio.io/v1 [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:metadata: [e2e-llm-inference-service] f:labels: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:app.kubernetes.io/component: {} [e2e-llm-inference-service] f:app.kubernetes.io/name: {} [e2e-llm-inference-service] f:app.kubernetes.io/part-of: {} [e2e-llm-inference-service] f:llm-d.ai/managed: {} [e2e-llm-inference-service] f:ownerReferences: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"uid":"e15cfc59-e4a1-4876-b43b-85ce3b1b239b"}: {} [e2e-llm-inference-service] f:spec: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:exportTo: {} [e2e-llm-inference-service] f:host: {} [e2e-llm-inference-service] f:trafficPolicy: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:tls: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:insecureSkipVerify: {} [e2e-llm-inference-service] f:mode: {} [e2e-llm-inference-service] f:sni: {} [e2e-llm-inference-service] manager: manager [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] time: '2026-07-07T14:29:34Z' [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-7ca60146-kserve-shadow-svc [e2e-llm-inference-service] namespace: kserve-ci-e2e-test [e2e-llm-inference-service] ownerReferences: [e2e-llm-inference-service] - apiVersion: serving.kserve.io/v1alpha2 [e2e-llm-inference-service] blockOwnerDeletion: true [e2e-llm-inference-service] controller: true [e2e-llm-inference-service] kind: LLMInferenceService [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-7ca60146 [e2e-llm-inference-service] uid: e15cfc59-e4a1-4876-b43b-85ce3b1b239b [e2e-llm-inference-service] resourceVersion: '28525' [e2e-llm-inference-service] uid: 50be281a-9e88-4284-b624-9ee35607afef [e2e-llm-inference-service] spec: [e2e-llm-inference-service] exportTo: [e2e-llm-inference-service] - '*' [e2e-llm-inference-service] host: llmisvc-model-fb-opt-125m-with-7ca60146-inference-p-ip-16c62f55.kserve-ci-e2e-test.svc.cluster.local [e2e-llm-inference-service] trafficPolicy: [e2e-llm-inference-service] tls: [e2e-llm-inference-service] insecureSkipVerify: true [e2e-llm-inference-service] mode: SIMPLE [e2e-llm-inference-service] sni: llmisvc-model-fb-opt-125m-with-7ca60146-kserve-workload-svc.kserve-ci-e2e-test.svc.cluster.local [e2e-llm-inference-service] [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1254 --- [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1255 apiVersion: networking.istio.io/v1alpha3 [e2e-llm-inference-service] kind: DestinationRule [e2e-llm-inference-service] metadata: [e2e-llm-inference-service] creationTimestamp: '2026-07-07T14:29:25Z' [e2e-llm-inference-service] generation: 1 [e2e-llm-inference-service] labels: [e2e-llm-inference-service] app.kubernetes.io/component: llminferenceservice-workload [e2e-llm-inference-service] app.kubernetes.io/name: llmisvc-model-fb-opt-125m-with-7ca60146 [e2e-llm-inference-service] app.kubernetes.io/part-of: llminferenceservice [e2e-llm-inference-service] llm-d.ai/managed: 'true' [e2e-llm-inference-service] managedFields: [e2e-llm-inference-service] - apiVersion: networking.istio.io/v1 [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:metadata: [e2e-llm-inference-service] f:labels: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:app.kubernetes.io/component: {} [e2e-llm-inference-service] f:app.kubernetes.io/name: {} [e2e-llm-inference-service] f:app.kubernetes.io/part-of: {} [e2e-llm-inference-service] f:llm-d.ai/managed: {} [e2e-llm-inference-service] f:ownerReferences: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"uid":"e15cfc59-e4a1-4876-b43b-85ce3b1b239b"}: {} [e2e-llm-inference-service] f:spec: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:exportTo: {} [e2e-llm-inference-service] f:host: {} [e2e-llm-inference-service] f:trafficPolicy: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:tls: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:caCertificates: {} [e2e-llm-inference-service] f:insecureSkipVerify: {} [e2e-llm-inference-service] f:mode: {} [e2e-llm-inference-service] f:sni: {} [e2e-llm-inference-service] manager: manager [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] time: '2026-07-07T14:29:25Z' [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-7ca60146-kserve-workload-svc [e2e-llm-inference-service] namespace: kserve-ci-e2e-test [e2e-llm-inference-service] ownerReferences: [e2e-llm-inference-service] - apiVersion: serving.kserve.io/v1alpha2 [e2e-llm-inference-service] blockOwnerDeletion: true [e2e-llm-inference-service] controller: true [e2e-llm-inference-service] kind: LLMInferenceService [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-7ca60146 [e2e-llm-inference-service] uid: e15cfc59-e4a1-4876-b43b-85ce3b1b239b [e2e-llm-inference-service] resourceVersion: '28322' [e2e-llm-inference-service] uid: eb5b5cb3-209a-4843-b33b-5dbfc7721e67 [e2e-llm-inference-service] spec: [e2e-llm-inference-service] exportTo: [e2e-llm-inference-service] - '*' [e2e-llm-inference-service] host: llmisvc-model-fb-opt-125m-with-7ca60146-kserve-workload-svc.kserve-ci-e2e-test.svc.cluster.local [e2e-llm-inference-service] trafficPolicy: [e2e-llm-inference-service] tls: [e2e-llm-inference-service] caCertificates: /var/run/secrets/kubernetes.io/serviceaccount/service-ca.crt [e2e-llm-inference-service] insecureSkipVerify: false [e2e-llm-inference-service] mode: SIMPLE [e2e-llm-inference-service] sni: llmisvc-model-fb-opt-125m-with-7ca60146-kserve-workload-svc.kserve-ci-e2e-test.svc.cluster.local [e2e-llm-inference-service] [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1254 --- [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1255 apiVersion: inference.networking.x-k8s.io/v1alpha2 [e2e-llm-inference-service] kind: InferencePool [e2e-llm-inference-service] metadata: [e2e-llm-inference-service] creationTimestamp: '2026-07-07T14:29:25Z' [e2e-llm-inference-service] generation: 1 [e2e-llm-inference-service] labels: [e2e-llm-inference-service] app.kubernetes.io/component: llminferenceservice-router-scheduler [e2e-llm-inference-service] app.kubernetes.io/name: llmisvc-model-fb-opt-125m-with-7ca60146 [e2e-llm-inference-service] app.kubernetes.io/part-of: llminferenceservice [e2e-llm-inference-service] managedFields: [e2e-llm-inference-service] - apiVersion: inference.networking.x-k8s.io/v1alpha2 [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:metadata: [e2e-llm-inference-service] f:labels: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:app.kubernetes.io/component: {} [e2e-llm-inference-service] f:app.kubernetes.io/name: {} [e2e-llm-inference-service] f:app.kubernetes.io/part-of: {} [e2e-llm-inference-service] f:ownerReferences: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"uid":"e15cfc59-e4a1-4876-b43b-85ce3b1b239b"}: {} [e2e-llm-inference-service] f:spec: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:extensionRef: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:failureMode: {} [e2e-llm-inference-service] f:group: {} [e2e-llm-inference-service] f:kind: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:portNumber: {} [e2e-llm-inference-service] f:selector: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:app.kubernetes.io/name: {} [e2e-llm-inference-service] f:app.kubernetes.io/part-of: {} [e2e-llm-inference-service] f:kserve.io/component: {} [e2e-llm-inference-service] f:targetPortNumber: {} [e2e-llm-inference-service] manager: manager [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] time: '2026-07-07T14:29:25Z' [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-7ca60146-inference-pool [e2e-llm-inference-service] namespace: kserve-ci-e2e-test [e2e-llm-inference-service] ownerReferences: [e2e-llm-inference-service] - apiVersion: serving.kserve.io/v1alpha2 [e2e-llm-inference-service] blockOwnerDeletion: true [e2e-llm-inference-service] controller: true [e2e-llm-inference-service] kind: LLMInferenceService [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-7ca60146 [e2e-llm-inference-service] uid: e15cfc59-e4a1-4876-b43b-85ce3b1b239b [e2e-llm-inference-service] resourceVersion: '28271' [e2e-llm-inference-service] uid: e7b02240-94eb-4fd1-a286-d0b6ef8ba9b2 [e2e-llm-inference-service] spec: [e2e-llm-inference-service] extensionRef: [e2e-llm-inference-service] failureMode: FailOpen [e2e-llm-inference-service] group: '' [e2e-llm-inference-service] kind: Service [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-7ca60146-epp-service [e2e-llm-inference-service] portNumber: 9002 [e2e-llm-inference-service] selector: [e2e-llm-inference-service] app.kubernetes.io/name: llmisvc-model-fb-opt-125m-with-7ca60146 [e2e-llm-inference-service] app.kubernetes.io/part-of: llminferenceservice [e2e-llm-inference-service] kserve.io/component: workload [e2e-llm-inference-service] targetPortNumber: 8000 [e2e-llm-inference-service] status: [e2e-llm-inference-service] parent: [e2e-llm-inference-service] - conditions: [e2e-llm-inference-service] - lastTransitionTime: '1970-01-01T00:00:00Z' [e2e-llm-inference-service] message: Waiting for controller [e2e-llm-inference-service] reason: Pending [e2e-llm-inference-service] status: Unknown [e2e-llm-inference-service] type: Accepted [e2e-llm-inference-service] parentRef: [e2e-llm-inference-service] group: gateway.networking.k8s.io [e2e-llm-inference-service] kind: Status [e2e-llm-inference-service] name: default [e2e-llm-inference-service] [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1254 --- [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1255 metadata: [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-7ca60146-kserve-7cfdf49d74-pd9px [e2e-llm-inference-service] namespace: kserve-ci-e2e-test [e2e-llm-inference-service] creationTimestamp: '2026-07-07T14:45:52Z' [e2e-llm-inference-service] labels: [e2e-llm-inference-service] app.kubernetes.io/component: llminferenceservice-workload [e2e-llm-inference-service] app.kubernetes.io/name: llmisvc-model-fb-opt-125m-with-7ca60146 [e2e-llm-inference-service] app.kubernetes.io/part-of: llminferenceservice [e2e-llm-inference-service] kserve.io/component: workload [e2e-llm-inference-service] llm-d.ai/role: both [e2e-llm-inference-service] pod-template-hash: 7cfdf49d74 [e2e-llm-inference-service] timestamp: '2026-07-07T14:45:32Z' [e2e-llm-inference-service] window: 15.663s [e2e-llm-inference-service] containers: [e2e-llm-inference-service] - name: main [e2e-llm-inference-service] usage: [e2e-llm-inference-service] cpu: 98897848n [e2e-llm-inference-service] memory: 2319216Ki [e2e-llm-inference-service] apiVersion: metrics.k8s.io/v1beta1 [e2e-llm-inference-service] kind: PodMetrics [e2e-llm-inference-service] [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1254 --- [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1255 metadata: [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-7ca60146-kserve-router-schenp5mk [e2e-llm-inference-service] namespace: kserve-ci-e2e-test [e2e-llm-inference-service] creationTimestamp: '2026-07-07T14:45:52Z' [e2e-llm-inference-service] labels: [e2e-llm-inference-service] app.kubernetes.io/component: llminferenceservice-router-scheduler [e2e-llm-inference-service] app.kubernetes.io/name: llmisvc-model-fb-opt-125m-with-7ca60146 [e2e-llm-inference-service] app.kubernetes.io/part-of: llminferenceservice [e2e-llm-inference-service] pod-template-hash: 778547fbc [e2e-llm-inference-service] timestamp: '2026-07-07T14:45:26Z' [e2e-llm-inference-service] window: 24.785s [e2e-llm-inference-service] containers: [e2e-llm-inference-service] - name: main [e2e-llm-inference-service] usage: [e2e-llm-inference-service] cpu: 58241436n [e2e-llm-inference-service] memory: 31540Ki [e2e-llm-inference-service] apiVersion: metrics.k8s.io/v1beta1 [e2e-llm-inference-service] kind: PodMetrics [e2e-llm-inference-service] [e2e-llm-inference-service] ERROR e2e.llmisvc.logging:logging.py:48 [test_llm_inference_service] [2026-07-07T14:45:52.811776] end - ❌ 993.947s: ❌ Failed to call model: HTTPConnectionPool(host='ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com', port=80): Max retries exceeded with url: /v1/completions (Caused by ReadTimeoutError("HTTPConnectionPool(host='ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com', port=80): Read timed out. (read timeout=60)")) [e2e-llm-inference-service] _ test_llm_inference_service[router-managed-workload-single-cpu-model-fb-opt-125m] _ [e2e-llm-inference-service] [gw0] linux -- Python 3.11.13 /workspace/source/python/kserve/.venv/bin/python [e2e-llm-inference-service] [e2e-llm-inference-service] self = [e2e-llm-inference-service] conn = [e2e-llm-inference-service] method = 'POST' [e2e-llm-inference-service] url = '/kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-route-87882a8e/v1/completions' [e2e-llm-inference-service] body = b'{"model": "facebook/opt-125m", "prompt": "KServe is a", "max_tokens": 20}' [e2e-llm-inference-service] headers = {'User-Agent': 'python-requests/2.32.3', 'Accept-Encoding': 'gzip, deflate', 'Accept': '*/*', 'Connection': 'keep-alive', 'Content-Type': 'application/json', 'Content-Length': '73'} [e2e-llm-inference-service] retries = Retry(total=0, connect=None, read=None, redirect=None, status=None) [e2e-llm-inference-service] timeout = Timeout(connect=60, read=60, total=None), chunked = False [e2e-llm-inference-service] response_conn = [e2e-llm-inference-service] preload_content = False, decode_content = False, enforce_content_length = True [e2e-llm-inference-service] [e2e-llm-inference-service] def _make_request( [e2e-llm-inference-service] self, [e2e-llm-inference-service] conn: BaseHTTPConnection, [e2e-llm-inference-service] method: str, [e2e-llm-inference-service] url: str, [e2e-llm-inference-service] body: _TYPE_BODY | None = None, [e2e-llm-inference-service] headers: typing.Mapping[str, str] | None = None, [e2e-llm-inference-service] retries: Retry | None = None, [e2e-llm-inference-service] timeout: _TYPE_TIMEOUT = _DEFAULT_TIMEOUT, [e2e-llm-inference-service] chunked: bool = False, [e2e-llm-inference-service] response_conn: BaseHTTPConnection | None = None, [e2e-llm-inference-service] preload_content: bool = True, [e2e-llm-inference-service] decode_content: bool = True, [e2e-llm-inference-service] enforce_content_length: bool = True, [e2e-llm-inference-service] ) -> BaseHTTPResponse: [e2e-llm-inference-service] """ [e2e-llm-inference-service] Perform a request on a given urllib connection object taken from our [e2e-llm-inference-service] pool. [e2e-llm-inference-service] [e2e-llm-inference-service] :param conn: [e2e-llm-inference-service] a connection from one of our connection pools [e2e-llm-inference-service] [e2e-llm-inference-service] :param method: [e2e-llm-inference-service] HTTP request method (such as GET, POST, PUT, etc.) [e2e-llm-inference-service] [e2e-llm-inference-service] :param url: [e2e-llm-inference-service] The URL to perform the request on. [e2e-llm-inference-service] [e2e-llm-inference-service] :param body: [e2e-llm-inference-service] Data to send in the request body, either :class:`str`, :class:`bytes`, [e2e-llm-inference-service] an iterable of :class:`str`/:class:`bytes`, or a file-like object. [e2e-llm-inference-service] [e2e-llm-inference-service] :param headers: [e2e-llm-inference-service] Dictionary of custom headers to send, such as User-Agent, [e2e-llm-inference-service] If-None-Match, etc. If None, pool headers are used. If provided, [e2e-llm-inference-service] these headers completely replace any pool-specific headers. [e2e-llm-inference-service] [e2e-llm-inference-service] :param retries: [e2e-llm-inference-service] Configure the number of retries to allow before raising a [e2e-llm-inference-service] :class:`~urllib3.exceptions.MaxRetryError` exception. [e2e-llm-inference-service] [e2e-llm-inference-service] Pass ``None`` to retry until you receive a response. Pass a [e2e-llm-inference-service] :class:`~urllib3.util.retry.Retry` object for fine-grained control [e2e-llm-inference-service] over different types of retries. [e2e-llm-inference-service] Pass an integer number to retry connection errors that many times, [e2e-llm-inference-service] but no other types of errors. Pass zero to never retry. [e2e-llm-inference-service] [e2e-llm-inference-service] If ``False``, then retries are disabled and any exception is raised [e2e-llm-inference-service] immediately. Also, instead of raising a MaxRetryError on redirects, [e2e-llm-inference-service] the redirect response will be returned. [e2e-llm-inference-service] [e2e-llm-inference-service] :type retries: :class:`~urllib3.util.retry.Retry`, False, or an int. [e2e-llm-inference-service] [e2e-llm-inference-service] :param timeout: [e2e-llm-inference-service] If specified, overrides the default timeout for this one [e2e-llm-inference-service] request. It may be a float (in seconds) or an instance of [e2e-llm-inference-service] :class:`urllib3.util.Timeout`. [e2e-llm-inference-service] [e2e-llm-inference-service] :param chunked: [e2e-llm-inference-service] If True, urllib3 will send the body using chunked transfer [e2e-llm-inference-service] encoding. Otherwise, urllib3 will send the body using the standard [e2e-llm-inference-service] content-length form. Defaults to False. [e2e-llm-inference-service] [e2e-llm-inference-service] :param response_conn: [e2e-llm-inference-service] Set this to ``None`` if you will handle releasing the connection or [e2e-llm-inference-service] set the connection to have the response release it. [e2e-llm-inference-service] [e2e-llm-inference-service] :param preload_content: [e2e-llm-inference-service] If True, the response's body will be preloaded during construction. [e2e-llm-inference-service] [e2e-llm-inference-service] :param decode_content: [e2e-llm-inference-service] If True, will attempt to decode the body based on the [e2e-llm-inference-service] 'content-encoding' header. [e2e-llm-inference-service] [e2e-llm-inference-service] :param enforce_content_length: [e2e-llm-inference-service] Enforce content length checking. Body returned by server must match [e2e-llm-inference-service] value of Content-Length header, if present. Otherwise, raise error. [e2e-llm-inference-service] """ [e2e-llm-inference-service] self.num_requests += 1 [e2e-llm-inference-service] [e2e-llm-inference-service] timeout_obj = self._get_timeout(timeout) [e2e-llm-inference-service] timeout_obj.start_connect() [e2e-llm-inference-service] conn.timeout = Timeout.resolve_default_timeout(timeout_obj.connect_timeout) [e2e-llm-inference-service] [e2e-llm-inference-service] try: [e2e-llm-inference-service] # Trigger any extra validation we need to do. [e2e-llm-inference-service] try: [e2e-llm-inference-service] self._validate_conn(conn) [e2e-llm-inference-service] except (SocketTimeout, BaseSSLError) as e: [e2e-llm-inference-service] self._raise_timeout(err=e, url=url, timeout_value=conn.timeout) [e2e-llm-inference-service] raise [e2e-llm-inference-service] [e2e-llm-inference-service] # _validate_conn() starts the connection to an HTTPS proxy [e2e-llm-inference-service] # so we need to wrap errors with 'ProxyError' here too. [e2e-llm-inference-service] except ( [e2e-llm-inference-service] OSError, [e2e-llm-inference-service] NewConnectionError, [e2e-llm-inference-service] TimeoutError, [e2e-llm-inference-service] BaseSSLError, [e2e-llm-inference-service] CertificateError, [e2e-llm-inference-service] SSLError, [e2e-llm-inference-service] ) as e: [e2e-llm-inference-service] new_e: Exception = e [e2e-llm-inference-service] if isinstance(e, (BaseSSLError, CertificateError)): [e2e-llm-inference-service] new_e = SSLError(e) [e2e-llm-inference-service] # If the connection didn't successfully connect to it's proxy [e2e-llm-inference-service] # then there [e2e-llm-inference-service] if isinstance( [e2e-llm-inference-service] new_e, (OSError, NewConnectionError, TimeoutError, SSLError) [e2e-llm-inference-service] ) and (conn and conn.proxy and not conn.has_connected_to_proxy): [e2e-llm-inference-service] new_e = _wrap_proxy_error(new_e, conn.proxy.scheme) [e2e-llm-inference-service] raise new_e [e2e-llm-inference-service] [e2e-llm-inference-service] # conn.request() calls http.client.*.request, not the method in [e2e-llm-inference-service] # urllib3.request. It also calls makefile (recv) on the socket. [e2e-llm-inference-service] try: [e2e-llm-inference-service] conn.request( [e2e-llm-inference-service] method, [e2e-llm-inference-service] url, [e2e-llm-inference-service] body=body, [e2e-llm-inference-service] headers=headers, [e2e-llm-inference-service] chunked=chunked, [e2e-llm-inference-service] preload_content=preload_content, [e2e-llm-inference-service] decode_content=decode_content, [e2e-llm-inference-service] enforce_content_length=enforce_content_length, [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] # We are swallowing BrokenPipeError (errno.EPIPE) since the server is [e2e-llm-inference-service] # legitimately able to close the connection after sending a valid response. [e2e-llm-inference-service] # With this behaviour, the received response is still readable. [e2e-llm-inference-service] except BrokenPipeError: [e2e-llm-inference-service] pass [e2e-llm-inference-service] except OSError as e: [e2e-llm-inference-service] # MacOS/Linux [e2e-llm-inference-service] # EPROTOTYPE and ECONNRESET are needed on macOS [e2e-llm-inference-service] # https://erickt.github.io/blog/2014/11/19/adventures-in-debugging-a-potential-osx-kernel-bug/ [e2e-llm-inference-service] # Condition changed later to emit ECONNRESET instead of only EPROTOTYPE. [e2e-llm-inference-service] if e.errno != errno.EPROTOTYPE and e.errno != errno.ECONNRESET: [e2e-llm-inference-service] raise [e2e-llm-inference-service] [e2e-llm-inference-service] # Reset the timeout for the recv() on the socket [e2e-llm-inference-service] read_timeout = timeout_obj.read_timeout [e2e-llm-inference-service] [e2e-llm-inference-service] if not conn.is_closed: [e2e-llm-inference-service] # In Python 3 socket.py will catch EAGAIN and return None when you [e2e-llm-inference-service] # try and read into the file pointer created by http.client, which [e2e-llm-inference-service] # instead raises a BadStatusLine exception. Instead of catching [e2e-llm-inference-service] # the exception and assuming all BadStatusLine exceptions are read [e2e-llm-inference-service] # timeouts, check for a zero timeout before making the request. [e2e-llm-inference-service] if read_timeout == 0: [e2e-llm-inference-service] raise ReadTimeoutError( [e2e-llm-inference-service] self, url, f"Read timed out. (read timeout={read_timeout})" [e2e-llm-inference-service] ) [e2e-llm-inference-service] conn.timeout = read_timeout [e2e-llm-inference-service] [e2e-llm-inference-service] # Receive the response from the server [e2e-llm-inference-service] try: [e2e-llm-inference-service] > response = conn.getresponse() [e2e-llm-inference-service] [e2e-llm-inference-service] ../../python/kserve/.venv/lib64/python3.11/site-packages/urllib3/connectionpool.py:534: [e2e-llm-inference-service] _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ [e2e-llm-inference-service] [e2e-llm-inference-service] self = [e2e-llm-inference-service] [e2e-llm-inference-service] def getresponse( # type: ignore[override] [e2e-llm-inference-service] self, [e2e-llm-inference-service] ) -> HTTPResponse: [e2e-llm-inference-service] """ [e2e-llm-inference-service] Get the response from the server. [e2e-llm-inference-service] [e2e-llm-inference-service] If the HTTPConnection is in the correct state, returns an instance of HTTPResponse or of whatever object is returned by the response_class variable. [e2e-llm-inference-service] [e2e-llm-inference-service] If a request has not been sent or if a previous response has not be handled, ResponseNotReady is raised. If the HTTP response indicates that the connection should be closed, then it will be closed before the response is returned. When the connection is closed, the underlying socket is closed. [e2e-llm-inference-service] """ [e2e-llm-inference-service] # Raise the same error as http.client.HTTPConnection [e2e-llm-inference-service] if self._response_options is None: [e2e-llm-inference-service] raise ResponseNotReady() [e2e-llm-inference-service] [e2e-llm-inference-service] # Reset this attribute for being used again. [e2e-llm-inference-service] resp_options = self._response_options [e2e-llm-inference-service] self._response_options = None [e2e-llm-inference-service] [e2e-llm-inference-service] # Since the connection's timeout value may have been updated [e2e-llm-inference-service] # we need to set the timeout on the socket. [e2e-llm-inference-service] self.sock.settimeout(self.timeout) [e2e-llm-inference-service] [e2e-llm-inference-service] # This is needed here to avoid circular import errors [e2e-llm-inference-service] from .response import HTTPResponse [e2e-llm-inference-service] [e2e-llm-inference-service] # Save a reference to the shutdown function before ownership is passed [e2e-llm-inference-service] # to httplib_response [e2e-llm-inference-service] # TODO should we implement it everywhere? [e2e-llm-inference-service] _shutdown = getattr(self.sock, "shutdown", None) [e2e-llm-inference-service] [e2e-llm-inference-service] # Get the response from http.client.HTTPConnection [e2e-llm-inference-service] > httplib_response = super().getresponse() [e2e-llm-inference-service] [e2e-llm-inference-service] ../../python/kserve/.venv/lib64/python3.11/site-packages/urllib3/connection.py:571: [e2e-llm-inference-service] _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ [e2e-llm-inference-service] [e2e-llm-inference-service] self = [e2e-llm-inference-service] [e2e-llm-inference-service] def getresponse(self): [e2e-llm-inference-service] """Get the response from the server. [e2e-llm-inference-service] [e2e-llm-inference-service] If the HTTPConnection is in the correct state, returns an [e2e-llm-inference-service] instance of HTTPResponse or of whatever object is returned by [e2e-llm-inference-service] the response_class variable. [e2e-llm-inference-service] [e2e-llm-inference-service] If a request has not been sent or if a previous response has [e2e-llm-inference-service] not be handled, ResponseNotReady is raised. If the HTTP [e2e-llm-inference-service] response indicates that the connection should be closed, then [e2e-llm-inference-service] it will be closed before the response is returned. When the [e2e-llm-inference-service] connection is closed, the underlying socket is closed. [e2e-llm-inference-service] """ [e2e-llm-inference-service] [e2e-llm-inference-service] # if a prior response has been completed, then forget about it. [e2e-llm-inference-service] if self.__response and self.__response.isclosed(): [e2e-llm-inference-service] self.__response = None [e2e-llm-inference-service] [e2e-llm-inference-service] # if a prior response exists, then it must be completed (otherwise, we [e2e-llm-inference-service] # cannot read this response's header to determine the connection-close [e2e-llm-inference-service] # behavior) [e2e-llm-inference-service] # [e2e-llm-inference-service] # note: if a prior response existed, but was connection-close, then the [e2e-llm-inference-service] # socket and response were made independent of this HTTPConnection [e2e-llm-inference-service] # object since a new request requires that we open a whole new [e2e-llm-inference-service] # connection [e2e-llm-inference-service] # [e2e-llm-inference-service] # this means the prior response had one of two states: [e2e-llm-inference-service] # 1) will_close: this connection was reset and the prior socket and [e2e-llm-inference-service] # response operate independently [e2e-llm-inference-service] # 2) persistent: the response was retained and we await its [e2e-llm-inference-service] # isclosed() status to become true. [e2e-llm-inference-service] # [e2e-llm-inference-service] if self.__state != _CS_REQ_SENT or self.__response: [e2e-llm-inference-service] raise ResponseNotReady(self.__state) [e2e-llm-inference-service] [e2e-llm-inference-service] if self.debuglevel > 0: [e2e-llm-inference-service] response = self.response_class(self.sock, self.debuglevel, [e2e-llm-inference-service] method=self._method) [e2e-llm-inference-service] else: [e2e-llm-inference-service] response = self.response_class(self.sock, method=self._method) [e2e-llm-inference-service] [e2e-llm-inference-service] try: [e2e-llm-inference-service] try: [e2e-llm-inference-service] > response.begin() [e2e-llm-inference-service] [e2e-llm-inference-service] /usr/lib64/python3.11/http/client.py:1395: [e2e-llm-inference-service] _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ [e2e-llm-inference-service] [e2e-llm-inference-service] self = [e2e-llm-inference-service] [e2e-llm-inference-service] def begin(self): [e2e-llm-inference-service] if self.headers is not None: [e2e-llm-inference-service] # we've already started reading the response [e2e-llm-inference-service] return [e2e-llm-inference-service] [e2e-llm-inference-service] # read until we get a non-100 response [e2e-llm-inference-service] while True: [e2e-llm-inference-service] > version, status, reason = self._read_status() [e2e-llm-inference-service] [e2e-llm-inference-service] /usr/lib64/python3.11/http/client.py:325: [e2e-llm-inference-service] _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ [e2e-llm-inference-service] [e2e-llm-inference-service] self = [e2e-llm-inference-service] [e2e-llm-inference-service] def _read_status(self): [e2e-llm-inference-service] > line = str(self.fp.readline(_MAXLINE + 1), "iso-8859-1") [e2e-llm-inference-service] [e2e-llm-inference-service] /usr/lib64/python3.11/http/client.py:286: [e2e-llm-inference-service] _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ [e2e-llm-inference-service] [e2e-llm-inference-service] self = [e2e-llm-inference-service] b = [e2e-llm-inference-service] [e2e-llm-inference-service] def readinto(self, b): [e2e-llm-inference-service] """Read up to len(b) bytes into the writable buffer *b* and return [e2e-llm-inference-service] the number of bytes read. If the socket is non-blocking and no bytes [e2e-llm-inference-service] are available, None is returned. [e2e-llm-inference-service] [e2e-llm-inference-service] If *b* is non-empty, a 0 return value indicates that the connection [e2e-llm-inference-service] was shutdown at the other end. [e2e-llm-inference-service] """ [e2e-llm-inference-service] self._checkClosed() [e2e-llm-inference-service] self._checkReadable() [e2e-llm-inference-service] if self._timeout_occurred: [e2e-llm-inference-service] raise OSError("cannot read from timed out object") [e2e-llm-inference-service] while True: [e2e-llm-inference-service] try: [e2e-llm-inference-service] > return self._sock.recv_into(b) [e2e-llm-inference-service] E TimeoutError: timed out [e2e-llm-inference-service] [e2e-llm-inference-service] /usr/lib64/python3.11/socket.py:718: TimeoutError [e2e-llm-inference-service] [e2e-llm-inference-service] The above exception was the direct cause of the following exception: [e2e-llm-inference-service] [e2e-llm-inference-service] self = [e2e-llm-inference-service] method = 'POST' [e2e-llm-inference-service] url = '/kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-route-87882a8e/v1/completions' [e2e-llm-inference-service] body = b'{"model": "facebook/opt-125m", "prompt": "KServe is a", "max_tokens": 20}' [e2e-llm-inference-service] headers = {'User-Agent': 'python-requests/2.32.3', 'Accept-Encoding': 'gzip, deflate', 'Accept': '*/*', 'Connection': 'keep-alive', 'Content-Type': 'application/json', 'Content-Length': '73'} [e2e-llm-inference-service] retries = Retry(total=0, connect=None, read=None, redirect=None, status=None) [e2e-llm-inference-service] redirect = False, assert_same_host = False [e2e-llm-inference-service] timeout = Timeout(connect=60, read=60, total=None), pool_timeout = None [e2e-llm-inference-service] release_conn = False, chunked = False, body_pos = None, preload_content = False [e2e-llm-inference-service] decode_content = False, response_kw = {} [e2e-llm-inference-service] parsed_url = Url(scheme=None, auth=None, host=None, port=None, path='/kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-route-87882a8e/v1/completions', query=None, fragment=None) [e2e-llm-inference-service] destination_scheme = None, conn = None, release_this_conn = True [e2e-llm-inference-service] http_tunnel_required = False, err = None, clean_exit = False [e2e-llm-inference-service] [e2e-llm-inference-service] def urlopen( # type: ignore[override] [e2e-llm-inference-service] self, [e2e-llm-inference-service] method: str, [e2e-llm-inference-service] url: str, [e2e-llm-inference-service] body: _TYPE_BODY | None = None, [e2e-llm-inference-service] headers: typing.Mapping[str, str] | None = None, [e2e-llm-inference-service] retries: Retry | bool | int | None = None, [e2e-llm-inference-service] redirect: bool = True, [e2e-llm-inference-service] assert_same_host: bool = True, [e2e-llm-inference-service] timeout: _TYPE_TIMEOUT = _DEFAULT_TIMEOUT, [e2e-llm-inference-service] pool_timeout: int | None = None, [e2e-llm-inference-service] release_conn: bool | None = None, [e2e-llm-inference-service] chunked: bool = False, [e2e-llm-inference-service] body_pos: _TYPE_BODY_POSITION | None = None, [e2e-llm-inference-service] preload_content: bool = True, [e2e-llm-inference-service] decode_content: bool = True, [e2e-llm-inference-service] **response_kw: typing.Any, [e2e-llm-inference-service] ) -> BaseHTTPResponse: [e2e-llm-inference-service] """ [e2e-llm-inference-service] Get a connection from the pool and perform an HTTP request. This is the [e2e-llm-inference-service] lowest level call for making a request, so you'll need to specify all [e2e-llm-inference-service] the raw details. [e2e-llm-inference-service] [e2e-llm-inference-service] .. note:: [e2e-llm-inference-service] [e2e-llm-inference-service] More commonly, it's appropriate to use a convenience method [e2e-llm-inference-service] such as :meth:`request`. [e2e-llm-inference-service] [e2e-llm-inference-service] .. note:: [e2e-llm-inference-service] [e2e-llm-inference-service] `release_conn` will only behave as expected if [e2e-llm-inference-service] `preload_content=False` because we want to make [e2e-llm-inference-service] `preload_content=False` the default behaviour someday soon without [e2e-llm-inference-service] breaking backwards compatibility. [e2e-llm-inference-service] [e2e-llm-inference-service] :param method: [e2e-llm-inference-service] HTTP request method (such as GET, POST, PUT, etc.) [e2e-llm-inference-service] [e2e-llm-inference-service] :param url: [e2e-llm-inference-service] The URL to perform the request on. [e2e-llm-inference-service] [e2e-llm-inference-service] :param body: [e2e-llm-inference-service] Data to send in the request body, either :class:`str`, :class:`bytes`, [e2e-llm-inference-service] an iterable of :class:`str`/:class:`bytes`, or a file-like object. [e2e-llm-inference-service] [e2e-llm-inference-service] :param headers: [e2e-llm-inference-service] Dictionary of custom headers to send, such as User-Agent, [e2e-llm-inference-service] If-None-Match, etc. If None, pool headers are used. If provided, [e2e-llm-inference-service] these headers completely replace any pool-specific headers. [e2e-llm-inference-service] [e2e-llm-inference-service] :param retries: [e2e-llm-inference-service] Configure the number of retries to allow before raising a [e2e-llm-inference-service] :class:`~urllib3.exceptions.MaxRetryError` exception. [e2e-llm-inference-service] [e2e-llm-inference-service] If ``None`` (default) will retry 3 times, see ``Retry.DEFAULT``. Pass a [e2e-llm-inference-service] :class:`~urllib3.util.retry.Retry` object for fine-grained control [e2e-llm-inference-service] over different types of retries. [e2e-llm-inference-service] Pass an integer number to retry connection errors that many times, [e2e-llm-inference-service] but no other types of errors. Pass zero to never retry. [e2e-llm-inference-service] [e2e-llm-inference-service] If ``False``, then retries are disabled and any exception is raised [e2e-llm-inference-service] immediately. Also, instead of raising a MaxRetryError on redirects, [e2e-llm-inference-service] the redirect response will be returned. [e2e-llm-inference-service] [e2e-llm-inference-service] :type retries: :class:`~urllib3.util.retry.Retry`, False, or an int. [e2e-llm-inference-service] [e2e-llm-inference-service] :param redirect: [e2e-llm-inference-service] If True, automatically handle redirects (status codes 301, 302, [e2e-llm-inference-service] 303, 307, 308). Each redirect counts as a retry. Disabling retries [e2e-llm-inference-service] will disable redirect, too. [e2e-llm-inference-service] [e2e-llm-inference-service] :param assert_same_host: [e2e-llm-inference-service] If ``True``, will make sure that the host of the pool requests is [e2e-llm-inference-service] consistent else will raise HostChangedError. When ``False``, you can [e2e-llm-inference-service] use the pool on an HTTP proxy and request foreign hosts. [e2e-llm-inference-service] [e2e-llm-inference-service] :param timeout: [e2e-llm-inference-service] If specified, overrides the default timeout for this one [e2e-llm-inference-service] request. It may be a float (in seconds) or an instance of [e2e-llm-inference-service] :class:`urllib3.util.Timeout`. [e2e-llm-inference-service] [e2e-llm-inference-service] :param pool_timeout: [e2e-llm-inference-service] If set and the pool is set to block=True, then this method will [e2e-llm-inference-service] block for ``pool_timeout`` seconds and raise EmptyPoolError if no [e2e-llm-inference-service] connection is available within the time period. [e2e-llm-inference-service] [e2e-llm-inference-service] :param bool preload_content: [e2e-llm-inference-service] If True, the response's body will be preloaded into memory. [e2e-llm-inference-service] [e2e-llm-inference-service] :param bool decode_content: [e2e-llm-inference-service] If True, will attempt to decode the body based on the [e2e-llm-inference-service] 'content-encoding' header. [e2e-llm-inference-service] [e2e-llm-inference-service] :param release_conn: [e2e-llm-inference-service] If False, then the urlopen call will not release the connection [e2e-llm-inference-service] back into the pool once a response is received (but will release if [e2e-llm-inference-service] you read the entire contents of the response such as when [e2e-llm-inference-service] `preload_content=True`). This is useful if you're not preloading [e2e-llm-inference-service] the response's content immediately. You will need to call [e2e-llm-inference-service] ``r.release_conn()`` on the response ``r`` to return the connection [e2e-llm-inference-service] back into the pool. If None, it takes the value of ``preload_content`` [e2e-llm-inference-service] which defaults to ``True``. [e2e-llm-inference-service] [e2e-llm-inference-service] :param bool chunked: [e2e-llm-inference-service] If True, urllib3 will send the body using chunked transfer [e2e-llm-inference-service] encoding. Otherwise, urllib3 will send the body using the standard [e2e-llm-inference-service] content-length form. Defaults to False. [e2e-llm-inference-service] [e2e-llm-inference-service] :param int body_pos: [e2e-llm-inference-service] Position to seek to in file-like body in the event of a retry or [e2e-llm-inference-service] redirect. Typically this won't need to be set because urllib3 will [e2e-llm-inference-service] auto-populate the value when needed. [e2e-llm-inference-service] """ [e2e-llm-inference-service] parsed_url = parse_url(url) [e2e-llm-inference-service] destination_scheme = parsed_url.scheme [e2e-llm-inference-service] [e2e-llm-inference-service] if headers is None: [e2e-llm-inference-service] headers = self.headers [e2e-llm-inference-service] [e2e-llm-inference-service] if not isinstance(retries, Retry): [e2e-llm-inference-service] retries = Retry.from_int(retries, redirect=redirect, default=self.retries) [e2e-llm-inference-service] [e2e-llm-inference-service] if release_conn is None: [e2e-llm-inference-service] release_conn = preload_content [e2e-llm-inference-service] [e2e-llm-inference-service] # Check host [e2e-llm-inference-service] if assert_same_host and not self.is_same_host(url): [e2e-llm-inference-service] raise HostChangedError(self, url, retries) [e2e-llm-inference-service] [e2e-llm-inference-service] # Ensure that the URL we're connecting to is properly encoded [e2e-llm-inference-service] if url.startswith("/"): [e2e-llm-inference-service] url = to_str(_encode_target(url)) [e2e-llm-inference-service] else: [e2e-llm-inference-service] url = to_str(parsed_url.url) [e2e-llm-inference-service] [e2e-llm-inference-service] conn = None [e2e-llm-inference-service] [e2e-llm-inference-service] # Track whether `conn` needs to be released before [e2e-llm-inference-service] # returning/raising/recursing. Update this variable if necessary, and [e2e-llm-inference-service] # leave `release_conn` constant throughout the function. That way, if [e2e-llm-inference-service] # the function recurses, the original value of `release_conn` will be [e2e-llm-inference-service] # passed down into the recursive call, and its value will be respected. [e2e-llm-inference-service] # [e2e-llm-inference-service] # See issue #651 [1] for details. [e2e-llm-inference-service] # [e2e-llm-inference-service] # [1] [e2e-llm-inference-service] release_this_conn = release_conn [e2e-llm-inference-service] [e2e-llm-inference-service] http_tunnel_required = connection_requires_http_tunnel( [e2e-llm-inference-service] self.proxy, self.proxy_config, destination_scheme [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] # Merge the proxy headers. Only done when not using HTTP CONNECT. We [e2e-llm-inference-service] # have to copy the headers dict so we can safely change it without those [e2e-llm-inference-service] # changes being reflected in anyone else's copy. [e2e-llm-inference-service] if not http_tunnel_required: [e2e-llm-inference-service] headers = headers.copy() # type: ignore[attr-defined] [e2e-llm-inference-service] headers.update(self.proxy_headers) # type: ignore[union-attr] [e2e-llm-inference-service] [e2e-llm-inference-service] # Must keep the exception bound to a separate variable or else Python 3 [e2e-llm-inference-service] # complains about UnboundLocalError. [e2e-llm-inference-service] err = None [e2e-llm-inference-service] [e2e-llm-inference-service] # Keep track of whether we cleanly exited the except block. This [e2e-llm-inference-service] # ensures we do proper cleanup in finally. [e2e-llm-inference-service] clean_exit = False [e2e-llm-inference-service] [e2e-llm-inference-service] # Rewind body position, if needed. Record current position [e2e-llm-inference-service] # for future rewinds in the event of a redirect/retry. [e2e-llm-inference-service] body_pos = set_file_position(body, body_pos) [e2e-llm-inference-service] [e2e-llm-inference-service] try: [e2e-llm-inference-service] # Request a connection from the queue. [e2e-llm-inference-service] timeout_obj = self._get_timeout(timeout) [e2e-llm-inference-service] conn = self._get_conn(timeout=pool_timeout) [e2e-llm-inference-service] [e2e-llm-inference-service] conn.timeout = timeout_obj.connect_timeout # type: ignore[assignment] [e2e-llm-inference-service] [e2e-llm-inference-service] # Is this a closed/new connection that requires CONNECT tunnelling? [e2e-llm-inference-service] if self.proxy is not None and http_tunnel_required and conn.is_closed: [e2e-llm-inference-service] try: [e2e-llm-inference-service] self._prepare_proxy(conn) [e2e-llm-inference-service] except (BaseSSLError, OSError, SocketTimeout) as e: [e2e-llm-inference-service] self._raise_timeout( [e2e-llm-inference-service] err=e, url=self.proxy.url, timeout_value=conn.timeout [e2e-llm-inference-service] ) [e2e-llm-inference-service] raise [e2e-llm-inference-service] [e2e-llm-inference-service] # If we're going to release the connection in ``finally:``, then [e2e-llm-inference-service] # the response doesn't need to know about the connection. Otherwise [e2e-llm-inference-service] # it will also try to release it and we'll have a double-release [e2e-llm-inference-service] # mess. [e2e-llm-inference-service] response_conn = conn if not release_conn else None [e2e-llm-inference-service] [e2e-llm-inference-service] # Make the request on the HTTPConnection object [e2e-llm-inference-service] > response = self._make_request( [e2e-llm-inference-service] conn, [e2e-llm-inference-service] method, [e2e-llm-inference-service] url, [e2e-llm-inference-service] timeout=timeout_obj, [e2e-llm-inference-service] body=body, [e2e-llm-inference-service] headers=headers, [e2e-llm-inference-service] chunked=chunked, [e2e-llm-inference-service] retries=retries, [e2e-llm-inference-service] response_conn=response_conn, [e2e-llm-inference-service] preload_content=preload_content, [e2e-llm-inference-service] decode_content=decode_content, [e2e-llm-inference-service] **response_kw, [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] ../../python/kserve/.venv/lib64/python3.11/site-packages/urllib3/connectionpool.py:787: [e2e-llm-inference-service] _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ [e2e-llm-inference-service] [e2e-llm-inference-service] self = [e2e-llm-inference-service] conn = [e2e-llm-inference-service] method = 'POST' [e2e-llm-inference-service] url = '/kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-route-87882a8e/v1/completions' [e2e-llm-inference-service] body = b'{"model": "facebook/opt-125m", "prompt": "KServe is a", "max_tokens": 20}' [e2e-llm-inference-service] headers = {'User-Agent': 'python-requests/2.32.3', 'Accept-Encoding': 'gzip, deflate', 'Accept': '*/*', 'Connection': 'keep-alive', 'Content-Type': 'application/json', 'Content-Length': '73'} [e2e-llm-inference-service] retries = Retry(total=0, connect=None, read=None, redirect=None, status=None) [e2e-llm-inference-service] timeout = Timeout(connect=60, read=60, total=None), chunked = False [e2e-llm-inference-service] response_conn = [e2e-llm-inference-service] preload_content = False, decode_content = False, enforce_content_length = True [e2e-llm-inference-service] [e2e-llm-inference-service] def _make_request( [e2e-llm-inference-service] self, [e2e-llm-inference-service] conn: BaseHTTPConnection, [e2e-llm-inference-service] method: str, [e2e-llm-inference-service] url: str, [e2e-llm-inference-service] body: _TYPE_BODY | None = None, [e2e-llm-inference-service] headers: typing.Mapping[str, str] | None = None, [e2e-llm-inference-service] retries: Retry | None = None, [e2e-llm-inference-service] timeout: _TYPE_TIMEOUT = _DEFAULT_TIMEOUT, [e2e-llm-inference-service] chunked: bool = False, [e2e-llm-inference-service] response_conn: BaseHTTPConnection | None = None, [e2e-llm-inference-service] preload_content: bool = True, [e2e-llm-inference-service] decode_content: bool = True, [e2e-llm-inference-service] enforce_content_length: bool = True, [e2e-llm-inference-service] ) -> BaseHTTPResponse: [e2e-llm-inference-service] """ [e2e-llm-inference-service] Perform a request on a given urllib connection object taken from our [e2e-llm-inference-service] pool. [e2e-llm-inference-service] [e2e-llm-inference-service] :param conn: [e2e-llm-inference-service] a connection from one of our connection pools [e2e-llm-inference-service] [e2e-llm-inference-service] :param method: [e2e-llm-inference-service] HTTP request method (such as GET, POST, PUT, etc.) [e2e-llm-inference-service] [e2e-llm-inference-service] :param url: [e2e-llm-inference-service] The URL to perform the request on. [e2e-llm-inference-service] [e2e-llm-inference-service] :param body: [e2e-llm-inference-service] Data to send in the request body, either :class:`str`, :class:`bytes`, [e2e-llm-inference-service] an iterable of :class:`str`/:class:`bytes`, or a file-like object. [e2e-llm-inference-service] [e2e-llm-inference-service] :param headers: [e2e-llm-inference-service] Dictionary of custom headers to send, such as User-Agent, [e2e-llm-inference-service] If-None-Match, etc. If None, pool headers are used. If provided, [e2e-llm-inference-service] these headers completely replace any pool-specific headers. [e2e-llm-inference-service] [e2e-llm-inference-service] :param retries: [e2e-llm-inference-service] Configure the number of retries to allow before raising a [e2e-llm-inference-service] :class:`~urllib3.exceptions.MaxRetryError` exception. [e2e-llm-inference-service] [e2e-llm-inference-service] Pass ``None`` to retry until you receive a response. Pass a [e2e-llm-inference-service] :class:`~urllib3.util.retry.Retry` object for fine-grained control [e2e-llm-inference-service] over different types of retries. [e2e-llm-inference-service] Pass an integer number to retry connection errors that many times, [e2e-llm-inference-service] but no other types of errors. Pass zero to never retry. [e2e-llm-inference-service] [e2e-llm-inference-service] If ``False``, then retries are disabled and any exception is raised [e2e-llm-inference-service] immediately. Also, instead of raising a MaxRetryError on redirects, [e2e-llm-inference-service] the redirect response will be returned. [e2e-llm-inference-service] [e2e-llm-inference-service] :type retries: :class:`~urllib3.util.retry.Retry`, False, or an int. [e2e-llm-inference-service] [e2e-llm-inference-service] :param timeout: [e2e-llm-inference-service] If specified, overrides the default timeout for this one [e2e-llm-inference-service] request. It may be a float (in seconds) or an instance of [e2e-llm-inference-service] :class:`urllib3.util.Timeout`. [e2e-llm-inference-service] [e2e-llm-inference-service] :param chunked: [e2e-llm-inference-service] If True, urllib3 will send the body using chunked transfer [e2e-llm-inference-service] encoding. Otherwise, urllib3 will send the body using the standard [e2e-llm-inference-service] content-length form. Defaults to False. [e2e-llm-inference-service] [e2e-llm-inference-service] :param response_conn: [e2e-llm-inference-service] Set this to ``None`` if you will handle releasing the connection or [e2e-llm-inference-service] set the connection to have the response release it. [e2e-llm-inference-service] [e2e-llm-inference-service] :param preload_content: [e2e-llm-inference-service] If True, the response's body will be preloaded during construction. [e2e-llm-inference-service] [e2e-llm-inference-service] :param decode_content: [e2e-llm-inference-service] If True, will attempt to decode the body based on the [e2e-llm-inference-service] 'content-encoding' header. [e2e-llm-inference-service] [e2e-llm-inference-service] :param enforce_content_length: [e2e-llm-inference-service] Enforce content length checking. Body returned by server must match [e2e-llm-inference-service] value of Content-Length header, if present. Otherwise, raise error. [e2e-llm-inference-service] """ [e2e-llm-inference-service] self.num_requests += 1 [e2e-llm-inference-service] [e2e-llm-inference-service] timeout_obj = self._get_timeout(timeout) [e2e-llm-inference-service] timeout_obj.start_connect() [e2e-llm-inference-service] conn.timeout = Timeout.resolve_default_timeout(timeout_obj.connect_timeout) [e2e-llm-inference-service] [e2e-llm-inference-service] try: [e2e-llm-inference-service] # Trigger any extra validation we need to do. [e2e-llm-inference-service] try: [e2e-llm-inference-service] self._validate_conn(conn) [e2e-llm-inference-service] except (SocketTimeout, BaseSSLError) as e: [e2e-llm-inference-service] self._raise_timeout(err=e, url=url, timeout_value=conn.timeout) [e2e-llm-inference-service] raise [e2e-llm-inference-service] [e2e-llm-inference-service] # _validate_conn() starts the connection to an HTTPS proxy [e2e-llm-inference-service] # so we need to wrap errors with 'ProxyError' here too. [e2e-llm-inference-service] except ( [e2e-llm-inference-service] OSError, [e2e-llm-inference-service] NewConnectionError, [e2e-llm-inference-service] TimeoutError, [e2e-llm-inference-service] BaseSSLError, [e2e-llm-inference-service] CertificateError, [e2e-llm-inference-service] SSLError, [e2e-llm-inference-service] ) as e: [e2e-llm-inference-service] new_e: Exception = e [e2e-llm-inference-service] if isinstance(e, (BaseSSLError, CertificateError)): [e2e-llm-inference-service] new_e = SSLError(e) [e2e-llm-inference-service] # If the connection didn't successfully connect to it's proxy [e2e-llm-inference-service] # then there [e2e-llm-inference-service] if isinstance( [e2e-llm-inference-service] new_e, (OSError, NewConnectionError, TimeoutError, SSLError) [e2e-llm-inference-service] ) and (conn and conn.proxy and not conn.has_connected_to_proxy): [e2e-llm-inference-service] new_e = _wrap_proxy_error(new_e, conn.proxy.scheme) [e2e-llm-inference-service] raise new_e [e2e-llm-inference-service] [e2e-llm-inference-service] # conn.request() calls http.client.*.request, not the method in [e2e-llm-inference-service] # urllib3.request. It also calls makefile (recv) on the socket. [e2e-llm-inference-service] try: [e2e-llm-inference-service] conn.request( [e2e-llm-inference-service] method, [e2e-llm-inference-service] url, [e2e-llm-inference-service] body=body, [e2e-llm-inference-service] headers=headers, [e2e-llm-inference-service] chunked=chunked, [e2e-llm-inference-service] preload_content=preload_content, [e2e-llm-inference-service] decode_content=decode_content, [e2e-llm-inference-service] enforce_content_length=enforce_content_length, [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] # We are swallowing BrokenPipeError (errno.EPIPE) since the server is [e2e-llm-inference-service] # legitimately able to close the connection after sending a valid response. [e2e-llm-inference-service] # With this behaviour, the received response is still readable. [e2e-llm-inference-service] except BrokenPipeError: [e2e-llm-inference-service] pass [e2e-llm-inference-service] except OSError as e: [e2e-llm-inference-service] # MacOS/Linux [e2e-llm-inference-service] # EPROTOTYPE and ECONNRESET are needed on macOS [e2e-llm-inference-service] # https://erickt.github.io/blog/2014/11/19/adventures-in-debugging-a-potential-osx-kernel-bug/ [e2e-llm-inference-service] # Condition changed later to emit ECONNRESET instead of only EPROTOTYPE. [e2e-llm-inference-service] if e.errno != errno.EPROTOTYPE and e.errno != errno.ECONNRESET: [e2e-llm-inference-service] raise [e2e-llm-inference-service] [e2e-llm-inference-service] # Reset the timeout for the recv() on the socket [e2e-llm-inference-service] read_timeout = timeout_obj.read_timeout [e2e-llm-inference-service] [e2e-llm-inference-service] if not conn.is_closed: [e2e-llm-inference-service] # In Python 3 socket.py will catch EAGAIN and return None when you [e2e-llm-inference-service] # try and read into the file pointer created by http.client, which [e2e-llm-inference-service] # instead raises a BadStatusLine exception. Instead of catching [e2e-llm-inference-service] # the exception and assuming all BadStatusLine exceptions are read [e2e-llm-inference-service] # timeouts, check for a zero timeout before making the request. [e2e-llm-inference-service] if read_timeout == 0: [e2e-llm-inference-service] raise ReadTimeoutError( [e2e-llm-inference-service] self, url, f"Read timed out. (read timeout={read_timeout})" [e2e-llm-inference-service] ) [e2e-llm-inference-service] conn.timeout = read_timeout [e2e-llm-inference-service] [e2e-llm-inference-service] # Receive the response from the server [e2e-llm-inference-service] try: [e2e-llm-inference-service] response = conn.getresponse() [e2e-llm-inference-service] except (BaseSSLError, OSError) as e: [e2e-llm-inference-service] > self._raise_timeout(err=e, url=url, timeout_value=read_timeout) [e2e-llm-inference-service] [e2e-llm-inference-service] ../../python/kserve/.venv/lib64/python3.11/site-packages/urllib3/connectionpool.py:536: [e2e-llm-inference-service] _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ [e2e-llm-inference-service] [e2e-llm-inference-service] self = [e2e-llm-inference-service] err = TimeoutError('timed out') [e2e-llm-inference-service] url = '/kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-route-87882a8e/v1/completions' [e2e-llm-inference-service] timeout_value = 60 [e2e-llm-inference-service] [e2e-llm-inference-service] def _raise_timeout( [e2e-llm-inference-service] self, [e2e-llm-inference-service] err: BaseSSLError | OSError | SocketTimeout, [e2e-llm-inference-service] url: str, [e2e-llm-inference-service] timeout_value: _TYPE_TIMEOUT | None, [e2e-llm-inference-service] ) -> None: [e2e-llm-inference-service] """Is the error actually a timeout? Will raise a ReadTimeout or pass""" [e2e-llm-inference-service] [e2e-llm-inference-service] if isinstance(err, SocketTimeout): [e2e-llm-inference-service] > raise ReadTimeoutError( [e2e-llm-inference-service] self, url, f"Read timed out. (read timeout={timeout_value})" [e2e-llm-inference-service] ) from err [e2e-llm-inference-service] E urllib3.exceptions.ReadTimeoutError: HTTPConnectionPool(host='ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com', port=80): Read timed out. (read timeout=60) [e2e-llm-inference-service] [e2e-llm-inference-service] ../../python/kserve/.venv/lib64/python3.11/site-packages/urllib3/connectionpool.py:367: ReadTimeoutError [e2e-llm-inference-service] [e2e-llm-inference-service] The above exception was the direct cause of the following exception: [e2e-llm-inference-service] [e2e-llm-inference-service] self = [e2e-llm-inference-service] request = , stream = False [e2e-llm-inference-service] timeout = Timeout(connect=60, read=60, total=None), verify = '/tmp/ca.crt' [e2e-llm-inference-service] cert = None, proxies = OrderedDict() [e2e-llm-inference-service] [e2e-llm-inference-service] def send( [e2e-llm-inference-service] self, request, stream=False, timeout=None, verify=True, cert=None, proxies=None [e2e-llm-inference-service] ): [e2e-llm-inference-service] """Sends PreparedRequest object. Returns Response object. [e2e-llm-inference-service] [e2e-llm-inference-service] :param request: The :class:`PreparedRequest ` being sent. [e2e-llm-inference-service] :param stream: (optional) Whether to stream the request content. [e2e-llm-inference-service] :param timeout: (optional) How long to wait for the server to send [e2e-llm-inference-service] data before giving up, as a float, or a :ref:`(connect timeout, [e2e-llm-inference-service] read timeout) ` tuple. [e2e-llm-inference-service] :type timeout: float or tuple or urllib3 Timeout object [e2e-llm-inference-service] :param verify: (optional) Either a boolean, in which case it controls whether [e2e-llm-inference-service] we verify the server's TLS certificate, or a string, in which case it [e2e-llm-inference-service] must be a path to a CA bundle to use [e2e-llm-inference-service] :param cert: (optional) Any user-provided SSL certificate to be trusted. [e2e-llm-inference-service] :param proxies: (optional) The proxies dictionary to apply to the request. [e2e-llm-inference-service] :rtype: requests.Response [e2e-llm-inference-service] """ [e2e-llm-inference-service] [e2e-llm-inference-service] try: [e2e-llm-inference-service] conn = self.get_connection_with_tls_context( [e2e-llm-inference-service] request, verify, proxies=proxies, cert=cert [e2e-llm-inference-service] ) [e2e-llm-inference-service] except LocationValueError as e: [e2e-llm-inference-service] raise InvalidURL(e, request=request) [e2e-llm-inference-service] [e2e-llm-inference-service] self.cert_verify(conn, request.url, verify, cert) [e2e-llm-inference-service] url = self.request_url(request, proxies) [e2e-llm-inference-service] self.add_headers( [e2e-llm-inference-service] request, [e2e-llm-inference-service] stream=stream, [e2e-llm-inference-service] timeout=timeout, [e2e-llm-inference-service] verify=verify, [e2e-llm-inference-service] cert=cert, [e2e-llm-inference-service] proxies=proxies, [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] chunked = not (request.body is None or "Content-Length" in request.headers) [e2e-llm-inference-service] [e2e-llm-inference-service] if isinstance(timeout, tuple): [e2e-llm-inference-service] try: [e2e-llm-inference-service] connect, read = timeout [e2e-llm-inference-service] timeout = TimeoutSauce(connect=connect, read=read) [e2e-llm-inference-service] except ValueError: [e2e-llm-inference-service] raise ValueError( [e2e-llm-inference-service] f"Invalid timeout {timeout}. Pass a (connect, read) timeout tuple, " [e2e-llm-inference-service] f"or a single float to set both timeouts to the same value." [e2e-llm-inference-service] ) [e2e-llm-inference-service] elif isinstance(timeout, TimeoutSauce): [e2e-llm-inference-service] pass [e2e-llm-inference-service] else: [e2e-llm-inference-service] timeout = TimeoutSauce(connect=timeout, read=timeout) [e2e-llm-inference-service] [e2e-llm-inference-service] try: [e2e-llm-inference-service] > resp = conn.urlopen( [e2e-llm-inference-service] method=request.method, [e2e-llm-inference-service] url=url, [e2e-llm-inference-service] body=request.body, [e2e-llm-inference-service] headers=request.headers, [e2e-llm-inference-service] redirect=False, [e2e-llm-inference-service] assert_same_host=False, [e2e-llm-inference-service] preload_content=False, [e2e-llm-inference-service] decode_content=False, [e2e-llm-inference-service] retries=self.max_retries, [e2e-llm-inference-service] timeout=timeout, [e2e-llm-inference-service] chunked=chunked, [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] ../../python/kserve/.venv/lib64/python3.11/site-packages/requests/adapters.py:667: [e2e-llm-inference-service] _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ [e2e-llm-inference-service] [e2e-llm-inference-service] self = [e2e-llm-inference-service] method = 'POST' [e2e-llm-inference-service] url = '/kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-route-87882a8e/v1/completions' [e2e-llm-inference-service] body = b'{"model": "facebook/opt-125m", "prompt": "KServe is a", "max_tokens": 20}' [e2e-llm-inference-service] headers = {'User-Agent': 'python-requests/2.32.3', 'Accept-Encoding': 'gzip, deflate', 'Accept': '*/*', 'Connection': 'keep-alive', 'Content-Type': 'application/json', 'Content-Length': '73'} [e2e-llm-inference-service] retries = Retry(total=7, connect=None, read=None, redirect=None, status=None) [e2e-llm-inference-service] redirect = False, assert_same_host = False [e2e-llm-inference-service] timeout = Timeout(connect=60, read=60, total=None), pool_timeout = None [e2e-llm-inference-service] release_conn = False, chunked = False, body_pos = None, preload_content = False [e2e-llm-inference-service] decode_content = False, response_kw = {} [e2e-llm-inference-service] parsed_url = Url(scheme=None, auth=None, host=None, port=None, path='/kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-route-87882a8e/v1/completions', query=None, fragment=None) [e2e-llm-inference-service] destination_scheme = None, conn = None, release_this_conn = True [e2e-llm-inference-service] http_tunnel_required = False [e2e-llm-inference-service] err = ReadTimeoutError("HTTPConnectionPool(host='ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com', port=80): Read timed out. (read timeout=60)") [e2e-llm-inference-service] clean_exit = False [e2e-llm-inference-service] [e2e-llm-inference-service] def urlopen( # type: ignore[override] [e2e-llm-inference-service] self, [e2e-llm-inference-service] method: str, [e2e-llm-inference-service] url: str, [e2e-llm-inference-service] body: _TYPE_BODY | None = None, [e2e-llm-inference-service] headers: typing.Mapping[str, str] | None = None, [e2e-llm-inference-service] retries: Retry | bool | int | None = None, [e2e-llm-inference-service] redirect: bool = True, [e2e-llm-inference-service] assert_same_host: bool = True, [e2e-llm-inference-service] timeout: _TYPE_TIMEOUT = _DEFAULT_TIMEOUT, [e2e-llm-inference-service] pool_timeout: int | None = None, [e2e-llm-inference-service] release_conn: bool | None = None, [e2e-llm-inference-service] chunked: bool = False, [e2e-llm-inference-service] body_pos: _TYPE_BODY_POSITION | None = None, [e2e-llm-inference-service] preload_content: bool = True, [e2e-llm-inference-service] decode_content: bool = True, [e2e-llm-inference-service] **response_kw: typing.Any, [e2e-llm-inference-service] ) -> BaseHTTPResponse: [e2e-llm-inference-service] """ [e2e-llm-inference-service] Get a connection from the pool and perform an HTTP request. This is the [e2e-llm-inference-service] lowest level call for making a request, so you'll need to specify all [e2e-llm-inference-service] the raw details. [e2e-llm-inference-service] [e2e-llm-inference-service] .. note:: [e2e-llm-inference-service] [e2e-llm-inference-service] More commonly, it's appropriate to use a convenience method [e2e-llm-inference-service] such as :meth:`request`. [e2e-llm-inference-service] [e2e-llm-inference-service] .. note:: [e2e-llm-inference-service] [e2e-llm-inference-service] `release_conn` will only behave as expected if [e2e-llm-inference-service] `preload_content=False` because we want to make [e2e-llm-inference-service] `preload_content=False` the default behaviour someday soon without [e2e-llm-inference-service] breaking backwards compatibility. [e2e-llm-inference-service] [e2e-llm-inference-service] :param method: [e2e-llm-inference-service] HTTP request method (such as GET, POST, PUT, etc.) [e2e-llm-inference-service] [e2e-llm-inference-service] :param url: [e2e-llm-inference-service] The URL to perform the request on. [e2e-llm-inference-service] [e2e-llm-inference-service] :param body: [e2e-llm-inference-service] Data to send in the request body, either :class:`str`, :class:`bytes`, [e2e-llm-inference-service] an iterable of :class:`str`/:class:`bytes`, or a file-like object. [e2e-llm-inference-service] [e2e-llm-inference-service] :param headers: [e2e-llm-inference-service] Dictionary of custom headers to send, such as User-Agent, [e2e-llm-inference-service] If-None-Match, etc. If None, pool headers are used. If provided, [e2e-llm-inference-service] these headers completely replace any pool-specific headers. [e2e-llm-inference-service] [e2e-llm-inference-service] :param retries: [e2e-llm-inference-service] Configure the number of retries to allow before raising a [e2e-llm-inference-service] :class:`~urllib3.exceptions.MaxRetryError` exception. [e2e-llm-inference-service] [e2e-llm-inference-service] If ``None`` (default) will retry 3 times, see ``Retry.DEFAULT``. Pass a [e2e-llm-inference-service] :class:`~urllib3.util.retry.Retry` object for fine-grained control [e2e-llm-inference-service] over different types of retries. [e2e-llm-inference-service] Pass an integer number to retry connection errors that many times, [e2e-llm-inference-service] but no other types of errors. Pass zero to never retry. [e2e-llm-inference-service] [e2e-llm-inference-service] If ``False``, then retries are disabled and any exception is raised [e2e-llm-inference-service] immediately. Also, instead of raising a MaxRetryError on redirects, [e2e-llm-inference-service] the redirect response will be returned. [e2e-llm-inference-service] [e2e-llm-inference-service] :type retries: :class:`~urllib3.util.retry.Retry`, False, or an int. [e2e-llm-inference-service] [e2e-llm-inference-service] :param redirect: [e2e-llm-inference-service] If True, automatically handle redirects (status codes 301, 302, [e2e-llm-inference-service] 303, 307, 308). Each redirect counts as a retry. Disabling retries [e2e-llm-inference-service] will disable redirect, too. [e2e-llm-inference-service] [e2e-llm-inference-service] :param assert_same_host: [e2e-llm-inference-service] If ``True``, will make sure that the host of the pool requests is [e2e-llm-inference-service] consistent else will raise HostChangedError. When ``False``, you can [e2e-llm-inference-service] use the pool on an HTTP proxy and request foreign hosts. [e2e-llm-inference-service] [e2e-llm-inference-service] :param timeout: [e2e-llm-inference-service] If specified, overrides the default timeout for this one [e2e-llm-inference-service] request. It may be a float (in seconds) or an instance of [e2e-llm-inference-service] :class:`urllib3.util.Timeout`. [e2e-llm-inference-service] [e2e-llm-inference-service] :param pool_timeout: [e2e-llm-inference-service] If set and the pool is set to block=True, then this method will [e2e-llm-inference-service] block for ``pool_timeout`` seconds and raise EmptyPoolError if no [e2e-llm-inference-service] connection is available within the time period. [e2e-llm-inference-service] [e2e-llm-inference-service] :param bool preload_content: [e2e-llm-inference-service] If True, the response's body will be preloaded into memory. [e2e-llm-inference-service] [e2e-llm-inference-service] :param bool decode_content: [e2e-llm-inference-service] If True, will attempt to decode the body based on the [e2e-llm-inference-service] 'content-encoding' header. [e2e-llm-inference-service] [e2e-llm-inference-service] :param release_conn: [e2e-llm-inference-service] If False, then the urlopen call will not release the connection [e2e-llm-inference-service] back into the pool once a response is received (but will release if [e2e-llm-inference-service] you read the entire contents of the response such as when [e2e-llm-inference-service] `preload_content=True`). This is useful if you're not preloading [e2e-llm-inference-service] the response's content immediately. You will need to call [e2e-llm-inference-service] ``r.release_conn()`` on the response ``r`` to return the connection [e2e-llm-inference-service] back into the pool. If None, it takes the value of ``preload_content`` [e2e-llm-inference-service] which defaults to ``True``. [e2e-llm-inference-service] [e2e-llm-inference-service] :param bool chunked: [e2e-llm-inference-service] If True, urllib3 will send the body using chunked transfer [e2e-llm-inference-service] encoding. Otherwise, urllib3 will send the body using the standard [e2e-llm-inference-service] content-length form. Defaults to False. [e2e-llm-inference-service] [e2e-llm-inference-service] :param int body_pos: [e2e-llm-inference-service] Position to seek to in file-like body in the event of a retry or [e2e-llm-inference-service] redirect. Typically this won't need to be set because urllib3 will [e2e-llm-inference-service] auto-populate the value when needed. [e2e-llm-inference-service] """ [e2e-llm-inference-service] parsed_url = parse_url(url) [e2e-llm-inference-service] destination_scheme = parsed_url.scheme [e2e-llm-inference-service] [e2e-llm-inference-service] if headers is None: [e2e-llm-inference-service] headers = self.headers [e2e-llm-inference-service] [e2e-llm-inference-service] if not isinstance(retries, Retry): [e2e-llm-inference-service] retries = Retry.from_int(retries, redirect=redirect, default=self.retries) [e2e-llm-inference-service] [e2e-llm-inference-service] if release_conn is None: [e2e-llm-inference-service] release_conn = preload_content [e2e-llm-inference-service] [e2e-llm-inference-service] # Check host [e2e-llm-inference-service] if assert_same_host and not self.is_same_host(url): [e2e-llm-inference-service] raise HostChangedError(self, url, retries) [e2e-llm-inference-service] [e2e-llm-inference-service] # Ensure that the URL we're connecting to is properly encoded [e2e-llm-inference-service] if url.startswith("/"): [e2e-llm-inference-service] url = to_str(_encode_target(url)) [e2e-llm-inference-service] else: [e2e-llm-inference-service] url = to_str(parsed_url.url) [e2e-llm-inference-service] [e2e-llm-inference-service] conn = None [e2e-llm-inference-service] [e2e-llm-inference-service] # Track whether `conn` needs to be released before [e2e-llm-inference-service] # returning/raising/recursing. Update this variable if necessary, and [e2e-llm-inference-service] # leave `release_conn` constant throughout the function. That way, if [e2e-llm-inference-service] # the function recurses, the original value of `release_conn` will be [e2e-llm-inference-service] # passed down into the recursive call, and its value will be respected. [e2e-llm-inference-service] # [e2e-llm-inference-service] # See issue #651 [1] for details. [e2e-llm-inference-service] # [e2e-llm-inference-service] # [1] [e2e-llm-inference-service] release_this_conn = release_conn [e2e-llm-inference-service] [e2e-llm-inference-service] http_tunnel_required = connection_requires_http_tunnel( [e2e-llm-inference-service] self.proxy, self.proxy_config, destination_scheme [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] # Merge the proxy headers. Only done when not using HTTP CONNECT. We [e2e-llm-inference-service] # have to copy the headers dict so we can safely change it without those [e2e-llm-inference-service] # changes being reflected in anyone else's copy. [e2e-llm-inference-service] if not http_tunnel_required: [e2e-llm-inference-service] headers = headers.copy() # type: ignore[attr-defined] [e2e-llm-inference-service] headers.update(self.proxy_headers) # type: ignore[union-attr] [e2e-llm-inference-service] [e2e-llm-inference-service] # Must keep the exception bound to a separate variable or else Python 3 [e2e-llm-inference-service] # complains about UnboundLocalError. [e2e-llm-inference-service] err = None [e2e-llm-inference-service] [e2e-llm-inference-service] # Keep track of whether we cleanly exited the except block. This [e2e-llm-inference-service] # ensures we do proper cleanup in finally. [e2e-llm-inference-service] clean_exit = False [e2e-llm-inference-service] [e2e-llm-inference-service] # Rewind body position, if needed. Record current position [e2e-llm-inference-service] # for future rewinds in the event of a redirect/retry. [e2e-llm-inference-service] body_pos = set_file_position(body, body_pos) [e2e-llm-inference-service] [e2e-llm-inference-service] try: [e2e-llm-inference-service] # Request a connection from the queue. [e2e-llm-inference-service] timeout_obj = self._get_timeout(timeout) [e2e-llm-inference-service] conn = self._get_conn(timeout=pool_timeout) [e2e-llm-inference-service] [e2e-llm-inference-service] conn.timeout = timeout_obj.connect_timeout # type: ignore[assignment] [e2e-llm-inference-service] [e2e-llm-inference-service] # Is this a closed/new connection that requires CONNECT tunnelling? [e2e-llm-inference-service] if self.proxy is not None and http_tunnel_required and conn.is_closed: [e2e-llm-inference-service] try: [e2e-llm-inference-service] self._prepare_proxy(conn) [e2e-llm-inference-service] except (BaseSSLError, OSError, SocketTimeout) as e: [e2e-llm-inference-service] self._raise_timeout( [e2e-llm-inference-service] err=e, url=self.proxy.url, timeout_value=conn.timeout [e2e-llm-inference-service] ) [e2e-llm-inference-service] raise [e2e-llm-inference-service] [e2e-llm-inference-service] # If we're going to release the connection in ``finally:``, then [e2e-llm-inference-service] # the response doesn't need to know about the connection. Otherwise [e2e-llm-inference-service] # it will also try to release it and we'll have a double-release [e2e-llm-inference-service] # mess. [e2e-llm-inference-service] response_conn = conn if not release_conn else None [e2e-llm-inference-service] [e2e-llm-inference-service] # Make the request on the HTTPConnection object [e2e-llm-inference-service] response = self._make_request( [e2e-llm-inference-service] conn, [e2e-llm-inference-service] method, [e2e-llm-inference-service] url, [e2e-llm-inference-service] timeout=timeout_obj, [e2e-llm-inference-service] body=body, [e2e-llm-inference-service] headers=headers, [e2e-llm-inference-service] chunked=chunked, [e2e-llm-inference-service] retries=retries, [e2e-llm-inference-service] response_conn=response_conn, [e2e-llm-inference-service] preload_content=preload_content, [e2e-llm-inference-service] decode_content=decode_content, [e2e-llm-inference-service] **response_kw, [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] # Everything went great! [e2e-llm-inference-service] clean_exit = True [e2e-llm-inference-service] [e2e-llm-inference-service] except EmptyPoolError: [e2e-llm-inference-service] # Didn't get a connection from the pool, no need to clean up [e2e-llm-inference-service] clean_exit = True [e2e-llm-inference-service] release_this_conn = False [e2e-llm-inference-service] raise [e2e-llm-inference-service] [e2e-llm-inference-service] except ( [e2e-llm-inference-service] TimeoutError, [e2e-llm-inference-service] HTTPException, [e2e-llm-inference-service] OSError, [e2e-llm-inference-service] ProtocolError, [e2e-llm-inference-service] BaseSSLError, [e2e-llm-inference-service] SSLError, [e2e-llm-inference-service] CertificateError, [e2e-llm-inference-service] ProxyError, [e2e-llm-inference-service] ) as e: [e2e-llm-inference-service] # Discard the connection for these exceptions. It will be [e2e-llm-inference-service] # replaced during the next _get_conn() call. [e2e-llm-inference-service] clean_exit = False [e2e-llm-inference-service] new_e: Exception = e [e2e-llm-inference-service] if isinstance(e, (BaseSSLError, CertificateError)): [e2e-llm-inference-service] new_e = SSLError(e) [e2e-llm-inference-service] if isinstance( [e2e-llm-inference-service] new_e, [e2e-llm-inference-service] ( [e2e-llm-inference-service] OSError, [e2e-llm-inference-service] NewConnectionError, [e2e-llm-inference-service] TimeoutError, [e2e-llm-inference-service] SSLError, [e2e-llm-inference-service] HTTPException, [e2e-llm-inference-service] ), [e2e-llm-inference-service] ) and (conn and conn.proxy and not conn.has_connected_to_proxy): [e2e-llm-inference-service] new_e = _wrap_proxy_error(new_e, conn.proxy.scheme) [e2e-llm-inference-service] elif isinstance(new_e, (OSError, HTTPException)): [e2e-llm-inference-service] new_e = ProtocolError("Connection aborted.", new_e) [e2e-llm-inference-service] [e2e-llm-inference-service] retries = retries.increment( [e2e-llm-inference-service] method, url, error=new_e, _pool=self, _stacktrace=sys.exc_info()[2] [e2e-llm-inference-service] ) [e2e-llm-inference-service] retries.sleep() [e2e-llm-inference-service] [e2e-llm-inference-service] # Keep track of the error for the retry warning. [e2e-llm-inference-service] err = e [e2e-llm-inference-service] [e2e-llm-inference-service] finally: [e2e-llm-inference-service] if not clean_exit: [e2e-llm-inference-service] # We hit some kind of exception, handled or otherwise. We need [e2e-llm-inference-service] # to throw the connection away unless explicitly told not to. [e2e-llm-inference-service] # Close the connection, set the variable to None, and make sure [e2e-llm-inference-service] # we put the None back in the pool to avoid leaking it. [e2e-llm-inference-service] if conn: [e2e-llm-inference-service] conn.close() [e2e-llm-inference-service] conn = None [e2e-llm-inference-service] release_this_conn = True [e2e-llm-inference-service] [e2e-llm-inference-service] if release_this_conn: [e2e-llm-inference-service] # Put the connection back to be reused. If the connection is [e2e-llm-inference-service] # expired then it will be None, which will get replaced with a [e2e-llm-inference-service] # fresh connection during _get_conn. [e2e-llm-inference-service] self._put_conn(conn) [e2e-llm-inference-service] [e2e-llm-inference-service] if not conn: [e2e-llm-inference-service] # Try again [e2e-llm-inference-service] log.warning( [e2e-llm-inference-service] "Retrying (%r) after connection broken by '%r': %s", retries, err, url [e2e-llm-inference-service] ) [e2e-llm-inference-service] > return self.urlopen( [e2e-llm-inference-service] method, [e2e-llm-inference-service] url, [e2e-llm-inference-service] body, [e2e-llm-inference-service] headers, [e2e-llm-inference-service] retries, [e2e-llm-inference-service] redirect, [e2e-llm-inference-service] assert_same_host, [e2e-llm-inference-service] timeout=timeout, [e2e-llm-inference-service] pool_timeout=pool_timeout, [e2e-llm-inference-service] release_conn=release_conn, [e2e-llm-inference-service] chunked=chunked, [e2e-llm-inference-service] body_pos=body_pos, [e2e-llm-inference-service] preload_content=preload_content, [e2e-llm-inference-service] decode_content=decode_content, [e2e-llm-inference-service] **response_kw, [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] ../../python/kserve/.venv/lib64/python3.11/site-packages/urllib3/connectionpool.py:871: [e2e-llm-inference-service] _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ [e2e-llm-inference-service] [e2e-llm-inference-service] self = [e2e-llm-inference-service] method = 'POST' [e2e-llm-inference-service] url = '/kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-route-87882a8e/v1/completions' [e2e-llm-inference-service] body = b'{"model": "facebook/opt-125m", "prompt": "KServe is a", "max_tokens": 20}' [e2e-llm-inference-service] headers = {'User-Agent': 'python-requests/2.32.3', 'Accept-Encoding': 'gzip, deflate', 'Accept': '*/*', 'Connection': 'keep-alive', 'Content-Type': 'application/json', 'Content-Length': '73'} [e2e-llm-inference-service] retries = Retry(total=6, connect=None, read=None, redirect=None, status=None) [e2e-llm-inference-service] redirect = False, assert_same_host = False [e2e-llm-inference-service] timeout = Timeout(connect=60, read=60, total=None), pool_timeout = None [e2e-llm-inference-service] release_conn = False, chunked = False, body_pos = None, preload_content = False [e2e-llm-inference-service] decode_content = False, response_kw = {} [e2e-llm-inference-service] parsed_url = Url(scheme=None, auth=None, host=None, port=None, path='/kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-route-87882a8e/v1/completions', query=None, fragment=None) [e2e-llm-inference-service] destination_scheme = None, conn = None, release_this_conn = True [e2e-llm-inference-service] http_tunnel_required = False [e2e-llm-inference-service] err = ReadTimeoutError("HTTPConnectionPool(host='ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com', port=80): Read timed out. (read timeout=60)") [e2e-llm-inference-service] clean_exit = False [e2e-llm-inference-service] [e2e-llm-inference-service] def urlopen( # type: ignore[override] [e2e-llm-inference-service] self, [e2e-llm-inference-service] method: str, [e2e-llm-inference-service] url: str, [e2e-llm-inference-service] body: _TYPE_BODY | None = None, [e2e-llm-inference-service] headers: typing.Mapping[str, str] | None = None, [e2e-llm-inference-service] retries: Retry | bool | int | None = None, [e2e-llm-inference-service] redirect: bool = True, [e2e-llm-inference-service] assert_same_host: bool = True, [e2e-llm-inference-service] timeout: _TYPE_TIMEOUT = _DEFAULT_TIMEOUT, [e2e-llm-inference-service] pool_timeout: int | None = None, [e2e-llm-inference-service] release_conn: bool | None = None, [e2e-llm-inference-service] chunked: bool = False, [e2e-llm-inference-service] body_pos: _TYPE_BODY_POSITION | None = None, [e2e-llm-inference-service] preload_content: bool = True, [e2e-llm-inference-service] decode_content: bool = True, [e2e-llm-inference-service] **response_kw: typing.Any, [e2e-llm-inference-service] ) -> BaseHTTPResponse: [e2e-llm-inference-service] """ [e2e-llm-inference-service] Get a connection from the pool and perform an HTTP request. This is the [e2e-llm-inference-service] lowest level call for making a request, so you'll need to specify all [e2e-llm-inference-service] the raw details. [e2e-llm-inference-service] [e2e-llm-inference-service] .. note:: [e2e-llm-inference-service] [e2e-llm-inference-service] More commonly, it's appropriate to use a convenience method [e2e-llm-inference-service] such as :meth:`request`. [e2e-llm-inference-service] [e2e-llm-inference-service] .. note:: [e2e-llm-inference-service] [e2e-llm-inference-service] `release_conn` will only behave as expected if [e2e-llm-inference-service] `preload_content=False` because we want to make [e2e-llm-inference-service] `preload_content=False` the default behaviour someday soon without [e2e-llm-inference-service] breaking backwards compatibility. [e2e-llm-inference-service] [e2e-llm-inference-service] :param method: [e2e-llm-inference-service] HTTP request method (such as GET, POST, PUT, etc.) [e2e-llm-inference-service] [e2e-llm-inference-service] :param url: [e2e-llm-inference-service] The URL to perform the request on. [e2e-llm-inference-service] [e2e-llm-inference-service] :param body: [e2e-llm-inference-service] Data to send in the request body, either :class:`str`, :class:`bytes`, [e2e-llm-inference-service] an iterable of :class:`str`/:class:`bytes`, or a file-like object. [e2e-llm-inference-service] [e2e-llm-inference-service] :param headers: [e2e-llm-inference-service] Dictionary of custom headers to send, such as User-Agent, [e2e-llm-inference-service] If-None-Match, etc. If None, pool headers are used. If provided, [e2e-llm-inference-service] these headers completely replace any pool-specific headers. [e2e-llm-inference-service] [e2e-llm-inference-service] :param retries: [e2e-llm-inference-service] Configure the number of retries to allow before raising a [e2e-llm-inference-service] :class:`~urllib3.exceptions.MaxRetryError` exception. [e2e-llm-inference-service] [e2e-llm-inference-service] If ``None`` (default) will retry 3 times, see ``Retry.DEFAULT``. Pass a [e2e-llm-inference-service] :class:`~urllib3.util.retry.Retry` object for fine-grained control [e2e-llm-inference-service] over different types of retries. [e2e-llm-inference-service] Pass an integer number to retry connection errors that many times, [e2e-llm-inference-service] but no other types of errors. Pass zero to never retry. [e2e-llm-inference-service] [e2e-llm-inference-service] If ``False``, then retries are disabled and any exception is raised [e2e-llm-inference-service] immediately. Also, instead of raising a MaxRetryError on redirects, [e2e-llm-inference-service] the redirect response will be returned. [e2e-llm-inference-service] [e2e-llm-inference-service] :type retries: :class:`~urllib3.util.retry.Retry`, False, or an int. [e2e-llm-inference-service] [e2e-llm-inference-service] :param redirect: [e2e-llm-inference-service] If True, automatically handle redirects (status codes 301, 302, [e2e-llm-inference-service] 303, 307, 308). Each redirect counts as a retry. Disabling retries [e2e-llm-inference-service] will disable redirect, too. [e2e-llm-inference-service] [e2e-llm-inference-service] :param assert_same_host: [e2e-llm-inference-service] If ``True``, will make sure that the host of the pool requests is [e2e-llm-inference-service] consistent else will raise HostChangedError. When ``False``, you can [e2e-llm-inference-service] use the pool on an HTTP proxy and request foreign hosts. [e2e-llm-inference-service] [e2e-llm-inference-service] :param timeout: [e2e-llm-inference-service] If specified, overrides the default timeout for this one [e2e-llm-inference-service] request. It may be a float (in seconds) or an instance of [e2e-llm-inference-service] :class:`urllib3.util.Timeout`. [e2e-llm-inference-service] [e2e-llm-inference-service] :param pool_timeout: [e2e-llm-inference-service] If set and the pool is set to block=True, then this method will [e2e-llm-inference-service] block for ``pool_timeout`` seconds and raise EmptyPoolError if no [e2e-llm-inference-service] connection is available within the time period. [e2e-llm-inference-service] [e2e-llm-inference-service] :param bool preload_content: [e2e-llm-inference-service] If True, the response's body will be preloaded into memory. [e2e-llm-inference-service] [e2e-llm-inference-service] :param bool decode_content: [e2e-llm-inference-service] If True, will attempt to decode the body based on the [e2e-llm-inference-service] 'content-encoding' header. [e2e-llm-inference-service] [e2e-llm-inference-service] :param release_conn: [e2e-llm-inference-service] If False, then the urlopen call will not release the connection [e2e-llm-inference-service] back into the pool once a response is received (but will release if [e2e-llm-inference-service] you read the entire contents of the response such as when [e2e-llm-inference-service] `preload_content=True`). This is useful if you're not preloading [e2e-llm-inference-service] the response's content immediately. You will need to call [e2e-llm-inference-service] ``r.release_conn()`` on the response ``r`` to return the connection [e2e-llm-inference-service] back into the pool. If None, it takes the value of ``preload_content`` [e2e-llm-inference-service] which defaults to ``True``. [e2e-llm-inference-service] [e2e-llm-inference-service] :param bool chunked: [e2e-llm-inference-service] If True, urllib3 will send the body using chunked transfer [e2e-llm-inference-service] encoding. Otherwise, urllib3 will send the body using the standard [e2e-llm-inference-service] content-length form. Defaults to False. [e2e-llm-inference-service] [e2e-llm-inference-service] :param int body_pos: [e2e-llm-inference-service] Position to seek to in file-like body in the event of a retry or [e2e-llm-inference-service] redirect. Typically this won't need to be set because urllib3 will [e2e-llm-inference-service] auto-populate the value when needed. [e2e-llm-inference-service] """ [e2e-llm-inference-service] parsed_url = parse_url(url) [e2e-llm-inference-service] destination_scheme = parsed_url.scheme [e2e-llm-inference-service] [e2e-llm-inference-service] if headers is None: [e2e-llm-inference-service] headers = self.headers [e2e-llm-inference-service] [e2e-llm-inference-service] if not isinstance(retries, Retry): [e2e-llm-inference-service] retries = Retry.from_int(retries, redirect=redirect, default=self.retries) [e2e-llm-inference-service] [e2e-llm-inference-service] if release_conn is None: [e2e-llm-inference-service] release_conn = preload_content [e2e-llm-inference-service] [e2e-llm-inference-service] # Check host [e2e-llm-inference-service] if assert_same_host and not self.is_same_host(url): [e2e-llm-inference-service] raise HostChangedError(self, url, retries) [e2e-llm-inference-service] [e2e-llm-inference-service] # Ensure that the URL we're connecting to is properly encoded [e2e-llm-inference-service] if url.startswith("/"): [e2e-llm-inference-service] url = to_str(_encode_target(url)) [e2e-llm-inference-service] else: [e2e-llm-inference-service] url = to_str(parsed_url.url) [e2e-llm-inference-service] [e2e-llm-inference-service] conn = None [e2e-llm-inference-service] [e2e-llm-inference-service] # Track whether `conn` needs to be released before [e2e-llm-inference-service] # returning/raising/recursing. Update this variable if necessary, and [e2e-llm-inference-service] # leave `release_conn` constant throughout the function. That way, if [e2e-llm-inference-service] # the function recurses, the original value of `release_conn` will be [e2e-llm-inference-service] # passed down into the recursive call, and its value will be respected. [e2e-llm-inference-service] # [e2e-llm-inference-service] # See issue #651 [1] for details. [e2e-llm-inference-service] # [e2e-llm-inference-service] # [1] [e2e-llm-inference-service] release_this_conn = release_conn [e2e-llm-inference-service] [e2e-llm-inference-service] http_tunnel_required = connection_requires_http_tunnel( [e2e-llm-inference-service] self.proxy, self.proxy_config, destination_scheme [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] # Merge the proxy headers. Only done when not using HTTP CONNECT. We [e2e-llm-inference-service] # have to copy the headers dict so we can safely change it without those [e2e-llm-inference-service] # changes being reflected in anyone else's copy. [e2e-llm-inference-service] if not http_tunnel_required: [e2e-llm-inference-service] headers = headers.copy() # type: ignore[attr-defined] [e2e-llm-inference-service] headers.update(self.proxy_headers) # type: ignore[union-attr] [e2e-llm-inference-service] [e2e-llm-inference-service] # Must keep the exception bound to a separate variable or else Python 3 [e2e-llm-inference-service] # complains about UnboundLocalError. [e2e-llm-inference-service] err = None [e2e-llm-inference-service] [e2e-llm-inference-service] # Keep track of whether we cleanly exited the except block. This [e2e-llm-inference-service] # ensures we do proper cleanup in finally. [e2e-llm-inference-service] clean_exit = False [e2e-llm-inference-service] [e2e-llm-inference-service] # Rewind body position, if needed. Record current position [e2e-llm-inference-service] # for future rewinds in the event of a redirect/retry. [e2e-llm-inference-service] body_pos = set_file_position(body, body_pos) [e2e-llm-inference-service] [e2e-llm-inference-service] try: [e2e-llm-inference-service] # Request a connection from the queue. [e2e-llm-inference-service] timeout_obj = self._get_timeout(timeout) [e2e-llm-inference-service] conn = self._get_conn(timeout=pool_timeout) [e2e-llm-inference-service] [e2e-llm-inference-service] conn.timeout = timeout_obj.connect_timeout # type: ignore[assignment] [e2e-llm-inference-service] [e2e-llm-inference-service] # Is this a closed/new connection that requires CONNECT tunnelling? [e2e-llm-inference-service] if self.proxy is not None and http_tunnel_required and conn.is_closed: [e2e-llm-inference-service] try: [e2e-llm-inference-service] self._prepare_proxy(conn) [e2e-llm-inference-service] except (BaseSSLError, OSError, SocketTimeout) as e: [e2e-llm-inference-service] self._raise_timeout( [e2e-llm-inference-service] err=e, url=self.proxy.url, timeout_value=conn.timeout [e2e-llm-inference-service] ) [e2e-llm-inference-service] raise [e2e-llm-inference-service] [e2e-llm-inference-service] # If we're going to release the connection in ``finally:``, then [e2e-llm-inference-service] # the response doesn't need to know about the connection. Otherwise [e2e-llm-inference-service] # it will also try to release it and we'll have a double-release [e2e-llm-inference-service] # mess. [e2e-llm-inference-service] response_conn = conn if not release_conn else None [e2e-llm-inference-service] [e2e-llm-inference-service] # Make the request on the HTTPConnection object [e2e-llm-inference-service] response = self._make_request( [e2e-llm-inference-service] conn, [e2e-llm-inference-service] method, [e2e-llm-inference-service] url, [e2e-llm-inference-service] timeout=timeout_obj, [e2e-llm-inference-service] body=body, [e2e-llm-inference-service] headers=headers, [e2e-llm-inference-service] chunked=chunked, [e2e-llm-inference-service] retries=retries, [e2e-llm-inference-service] response_conn=response_conn, [e2e-llm-inference-service] preload_content=preload_content, [e2e-llm-inference-service] decode_content=decode_content, [e2e-llm-inference-service] **response_kw, [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] # Everything went great! [e2e-llm-inference-service] clean_exit = True [e2e-llm-inference-service] [e2e-llm-inference-service] except EmptyPoolError: [e2e-llm-inference-service] # Didn't get a connection from the pool, no need to clean up [e2e-llm-inference-service] clean_exit = True [e2e-llm-inference-service] release_this_conn = False [e2e-llm-inference-service] raise [e2e-llm-inference-service] [e2e-llm-inference-service] except ( [e2e-llm-inference-service] TimeoutError, [e2e-llm-inference-service] HTTPException, [e2e-llm-inference-service] OSError, [e2e-llm-inference-service] ProtocolError, [e2e-llm-inference-service] BaseSSLError, [e2e-llm-inference-service] SSLError, [e2e-llm-inference-service] CertificateError, [e2e-llm-inference-service] ProxyError, [e2e-llm-inference-service] ) as e: [e2e-llm-inference-service] # Discard the connection for these exceptions. It will be [e2e-llm-inference-service] # replaced during the next _get_conn() call. [e2e-llm-inference-service] clean_exit = False [e2e-llm-inference-service] new_e: Exception = e [e2e-llm-inference-service] if isinstance(e, (BaseSSLError, CertificateError)): [e2e-llm-inference-service] new_e = SSLError(e) [e2e-llm-inference-service] if isinstance( [e2e-llm-inference-service] new_e, [e2e-llm-inference-service] ( [e2e-llm-inference-service] OSError, [e2e-llm-inference-service] NewConnectionError, [e2e-llm-inference-service] TimeoutError, [e2e-llm-inference-service] SSLError, [e2e-llm-inference-service] HTTPException, [e2e-llm-inference-service] ), [e2e-llm-inference-service] ) and (conn and conn.proxy and not conn.has_connected_to_proxy): [e2e-llm-inference-service] new_e = _wrap_proxy_error(new_e, conn.proxy.scheme) [e2e-llm-inference-service] elif isinstance(new_e, (OSError, HTTPException)): [e2e-llm-inference-service] new_e = ProtocolError("Connection aborted.", new_e) [e2e-llm-inference-service] [e2e-llm-inference-service] retries = retries.increment( [e2e-llm-inference-service] method, url, error=new_e, _pool=self, _stacktrace=sys.exc_info()[2] [e2e-llm-inference-service] ) [e2e-llm-inference-service] retries.sleep() [e2e-llm-inference-service] [e2e-llm-inference-service] # Keep track of the error for the retry warning. [e2e-llm-inference-service] err = e [e2e-llm-inference-service] [e2e-llm-inference-service] finally: [e2e-llm-inference-service] if not clean_exit: [e2e-llm-inference-service] # We hit some kind of exception, handled or otherwise. We need [e2e-llm-inference-service] # to throw the connection away unless explicitly told not to. [e2e-llm-inference-service] # Close the connection, set the variable to None, and make sure [e2e-llm-inference-service] # we put the None back in the pool to avoid leaking it. [e2e-llm-inference-service] if conn: [e2e-llm-inference-service] conn.close() [e2e-llm-inference-service] conn = None [e2e-llm-inference-service] release_this_conn = True [e2e-llm-inference-service] [e2e-llm-inference-service] if release_this_conn: [e2e-llm-inference-service] # Put the connection back to be reused. If the connection is [e2e-llm-inference-service] # expired then it will be None, which will get replaced with a [e2e-llm-inference-service] # fresh connection during _get_conn. [e2e-llm-inference-service] self._put_conn(conn) [e2e-llm-inference-service] [e2e-llm-inference-service] if not conn: [e2e-llm-inference-service] # Try again [e2e-llm-inference-service] log.warning( [e2e-llm-inference-service] "Retrying (%r) after connection broken by '%r': %s", retries, err, url [e2e-llm-inference-service] ) [e2e-llm-inference-service] > return self.urlopen( [e2e-llm-inference-service] method, [e2e-llm-inference-service] url, [e2e-llm-inference-service] body, [e2e-llm-inference-service] headers, [e2e-llm-inference-service] retries, [e2e-llm-inference-service] redirect, [e2e-llm-inference-service] assert_same_host, [e2e-llm-inference-service] timeout=timeout, [e2e-llm-inference-service] pool_timeout=pool_timeout, [e2e-llm-inference-service] release_conn=release_conn, [e2e-llm-inference-service] chunked=chunked, [e2e-llm-inference-service] body_pos=body_pos, [e2e-llm-inference-service] preload_content=preload_content, [e2e-llm-inference-service] decode_content=decode_content, [e2e-llm-inference-service] **response_kw, [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] ../../python/kserve/.venv/lib64/python3.11/site-packages/urllib3/connectionpool.py:871: [e2e-llm-inference-service] _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ [e2e-llm-inference-service] [e2e-llm-inference-service] self = [e2e-llm-inference-service] method = 'POST' [e2e-llm-inference-service] url = '/kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-route-87882a8e/v1/completions' [e2e-llm-inference-service] body = b'{"model": "facebook/opt-125m", "prompt": "KServe is a", "max_tokens": 20}' [e2e-llm-inference-service] headers = {'User-Agent': 'python-requests/2.32.3', 'Accept-Encoding': 'gzip, deflate', 'Accept': '*/*', 'Connection': 'keep-alive', 'Content-Type': 'application/json', 'Content-Length': '73'} [e2e-llm-inference-service] retries = Retry(total=5, connect=None, read=None, redirect=None, status=None) [e2e-llm-inference-service] redirect = False, assert_same_host = False [e2e-llm-inference-service] timeout = Timeout(connect=60, read=60, total=None), pool_timeout = None [e2e-llm-inference-service] release_conn = False, chunked = False, body_pos = None, preload_content = False [e2e-llm-inference-service] decode_content = False, response_kw = {} [e2e-llm-inference-service] parsed_url = Url(scheme=None, auth=None, host=None, port=None, path='/kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-route-87882a8e/v1/completions', query=None, fragment=None) [e2e-llm-inference-service] destination_scheme = None, conn = None, release_this_conn = True [e2e-llm-inference-service] http_tunnel_required = False [e2e-llm-inference-service] err = ReadTimeoutError("HTTPConnectionPool(host='ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com', port=80): Read timed out. (read timeout=60)") [e2e-llm-inference-service] clean_exit = False [e2e-llm-inference-service] [e2e-llm-inference-service] def urlopen( # type: ignore[override] [e2e-llm-inference-service] self, [e2e-llm-inference-service] method: str, [e2e-llm-inference-service] url: str, [e2e-llm-inference-service] body: _TYPE_BODY | None = None, [e2e-llm-inference-service] headers: typing.Mapping[str, str] | None = None, [e2e-llm-inference-service] retries: Retry | bool | int | None = None, [e2e-llm-inference-service] redirect: bool = True, [e2e-llm-inference-service] assert_same_host: bool = True, [e2e-llm-inference-service] timeout: _TYPE_TIMEOUT = _DEFAULT_TIMEOUT, [e2e-llm-inference-service] pool_timeout: int | None = None, [e2e-llm-inference-service] release_conn: bool | None = None, [e2e-llm-inference-service] chunked: bool = False, [e2e-llm-inference-service] body_pos: _TYPE_BODY_POSITION | None = None, [e2e-llm-inference-service] preload_content: bool = True, [e2e-llm-inference-service] decode_content: bool = True, [e2e-llm-inference-service] **response_kw: typing.Any, [e2e-llm-inference-service] ) -> BaseHTTPResponse: [e2e-llm-inference-service] """ [e2e-llm-inference-service] Get a connection from the pool and perform an HTTP request. This is the [e2e-llm-inference-service] lowest level call for making a request, so you'll need to specify all [e2e-llm-inference-service] the raw details. [e2e-llm-inference-service] [e2e-llm-inference-service] .. note:: [e2e-llm-inference-service] [e2e-llm-inference-service] More commonly, it's appropriate to use a convenience method [e2e-llm-inference-service] such as :meth:`request`. [e2e-llm-inference-service] [e2e-llm-inference-service] .. note:: [e2e-llm-inference-service] [e2e-llm-inference-service] `release_conn` will only behave as expected if [e2e-llm-inference-service] `preload_content=False` because we want to make [e2e-llm-inference-service] `preload_content=False` the default behaviour someday soon without [e2e-llm-inference-service] breaking backwards compatibility. [e2e-llm-inference-service] [e2e-llm-inference-service] :param method: [e2e-llm-inference-service] HTTP request method (such as GET, POST, PUT, etc.) [e2e-llm-inference-service] [e2e-llm-inference-service] :param url: [e2e-llm-inference-service] The URL to perform the request on. [e2e-llm-inference-service] [e2e-llm-inference-service] :param body: [e2e-llm-inference-service] Data to send in the request body, either :class:`str`, :class:`bytes`, [e2e-llm-inference-service] an iterable of :class:`str`/:class:`bytes`, or a file-like object. [e2e-llm-inference-service] [e2e-llm-inference-service] :param headers: [e2e-llm-inference-service] Dictionary of custom headers to send, such as User-Agent, [e2e-llm-inference-service] If-None-Match, etc. If None, pool headers are used. If provided, [e2e-llm-inference-service] these headers completely replace any pool-specific headers. [e2e-llm-inference-service] [e2e-llm-inference-service] :param retries: [e2e-llm-inference-service] Configure the number of retries to allow before raising a [e2e-llm-inference-service] :class:`~urllib3.exceptions.MaxRetryError` exception. [e2e-llm-inference-service] [e2e-llm-inference-service] If ``None`` (default) will retry 3 times, see ``Retry.DEFAULT``. Pass a [e2e-llm-inference-service] :class:`~urllib3.util.retry.Retry` object for fine-grained control [e2e-llm-inference-service] over different types of retries. [e2e-llm-inference-service] Pass an integer number to retry connection errors that many times, [e2e-llm-inference-service] but no other types of errors. Pass zero to never retry. [e2e-llm-inference-service] [e2e-llm-inference-service] If ``False``, then retries are disabled and any exception is raised [e2e-llm-inference-service] immediately. Also, instead of raising a MaxRetryError on redirects, [e2e-llm-inference-service] the redirect response will be returned. [e2e-llm-inference-service] [e2e-llm-inference-service] :type retries: :class:`~urllib3.util.retry.Retry`, False, or an int. [e2e-llm-inference-service] [e2e-llm-inference-service] :param redirect: [e2e-llm-inference-service] If True, automatically handle redirects (status codes 301, 302, [e2e-llm-inference-service] 303, 307, 308). Each redirect counts as a retry. Disabling retries [e2e-llm-inference-service] will disable redirect, too. [e2e-llm-inference-service] [e2e-llm-inference-service] :param assert_same_host: [e2e-llm-inference-service] If ``True``, will make sure that the host of the pool requests is [e2e-llm-inference-service] consistent else will raise HostChangedError. When ``False``, you can [e2e-llm-inference-service] use the pool on an HTTP proxy and request foreign hosts. [e2e-llm-inference-service] [e2e-llm-inference-service] :param timeout: [e2e-llm-inference-service] If specified, overrides the default timeout for this one [e2e-llm-inference-service] request. It may be a float (in seconds) or an instance of [e2e-llm-inference-service] :class:`urllib3.util.Timeout`. [e2e-llm-inference-service] [e2e-llm-inference-service] :param pool_timeout: [e2e-llm-inference-service] If set and the pool is set to block=True, then this method will [e2e-llm-inference-service] block for ``pool_timeout`` seconds and raise EmptyPoolError if no [e2e-llm-inference-service] connection is available within the time period. [e2e-llm-inference-service] [e2e-llm-inference-service] :param bool preload_content: [e2e-llm-inference-service] If True, the response's body will be preloaded into memory. [e2e-llm-inference-service] [e2e-llm-inference-service] :param bool decode_content: [e2e-llm-inference-service] If True, will attempt to decode the body based on the [e2e-llm-inference-service] 'content-encoding' header. [e2e-llm-inference-service] [e2e-llm-inference-service] :param release_conn: [e2e-llm-inference-service] If False, then the urlopen call will not release the connection [e2e-llm-inference-service] back into the pool once a response is received (but will release if [e2e-llm-inference-service] you read the entire contents of the response such as when [e2e-llm-inference-service] `preload_content=True`). This is useful if you're not preloading [e2e-llm-inference-service] the response's content immediately. You will need to call [e2e-llm-inference-service] ``r.release_conn()`` on the response ``r`` to return the connection [e2e-llm-inference-service] back into the pool. If None, it takes the value of ``preload_content`` [e2e-llm-inference-service] which defaults to ``True``. [e2e-llm-inference-service] [e2e-llm-inference-service] :param bool chunked: [e2e-llm-inference-service] If True, urllib3 will send the body using chunked transfer [e2e-llm-inference-service] encoding. Otherwise, urllib3 will send the body using the standard [e2e-llm-inference-service] content-length form. Defaults to False. [e2e-llm-inference-service] [e2e-llm-inference-service] :param int body_pos: [e2e-llm-inference-service] Position to seek to in file-like body in the event of a retry or [e2e-llm-inference-service] redirect. Typically this won't need to be set because urllib3 will [e2e-llm-inference-service] auto-populate the value when needed. [e2e-llm-inference-service] """ [e2e-llm-inference-service] parsed_url = parse_url(url) [e2e-llm-inference-service] destination_scheme = parsed_url.scheme [e2e-llm-inference-service] [e2e-llm-inference-service] if headers is None: [e2e-llm-inference-service] headers = self.headers [e2e-llm-inference-service] [e2e-llm-inference-service] if not isinstance(retries, Retry): [e2e-llm-inference-service] retries = Retry.from_int(retries, redirect=redirect, default=self.retries) [e2e-llm-inference-service] [e2e-llm-inference-service] if release_conn is None: [e2e-llm-inference-service] release_conn = preload_content [e2e-llm-inference-service] [e2e-llm-inference-service] # Check host [e2e-llm-inference-service] if assert_same_host and not self.is_same_host(url): [e2e-llm-inference-service] raise HostChangedError(self, url, retries) [e2e-llm-inference-service] [e2e-llm-inference-service] # Ensure that the URL we're connecting to is properly encoded [e2e-llm-inference-service] if url.startswith("/"): [e2e-llm-inference-service] url = to_str(_encode_target(url)) [e2e-llm-inference-service] else: [e2e-llm-inference-service] url = to_str(parsed_url.url) [e2e-llm-inference-service] [e2e-llm-inference-service] conn = None [e2e-llm-inference-service] [e2e-llm-inference-service] # Track whether `conn` needs to be released before [e2e-llm-inference-service] # returning/raising/recursing. Update this variable if necessary, and [e2e-llm-inference-service] # leave `release_conn` constant throughout the function. That way, if [e2e-llm-inference-service] # the function recurses, the original value of `release_conn` will be [e2e-llm-inference-service] # passed down into the recursive call, and its value will be respected. [e2e-llm-inference-service] # [e2e-llm-inference-service] # See issue #651 [1] for details. [e2e-llm-inference-service] # [e2e-llm-inference-service] # [1] [e2e-llm-inference-service] release_this_conn = release_conn [e2e-llm-inference-service] [e2e-llm-inference-service] http_tunnel_required = connection_requires_http_tunnel( [e2e-llm-inference-service] self.proxy, self.proxy_config, destination_scheme [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] # Merge the proxy headers. Only done when not using HTTP CONNECT. We [e2e-llm-inference-service] # have to copy the headers dict so we can safely change it without those [e2e-llm-inference-service] # changes being reflected in anyone else's copy. [e2e-llm-inference-service] if not http_tunnel_required: [e2e-llm-inference-service] headers = headers.copy() # type: ignore[attr-defined] [e2e-llm-inference-service] headers.update(self.proxy_headers) # type: ignore[union-attr] [e2e-llm-inference-service] [e2e-llm-inference-service] # Must keep the exception bound to a separate variable or else Python 3 [e2e-llm-inference-service] # complains about UnboundLocalError. [e2e-llm-inference-service] err = None [e2e-llm-inference-service] [e2e-llm-inference-service] # Keep track of whether we cleanly exited the except block. This [e2e-llm-inference-service] # ensures we do proper cleanup in finally. [e2e-llm-inference-service] clean_exit = False [e2e-llm-inference-service] [e2e-llm-inference-service] # Rewind body position, if needed. Record current position [e2e-llm-inference-service] # for future rewinds in the event of a redirect/retry. [e2e-llm-inference-service] body_pos = set_file_position(body, body_pos) [e2e-llm-inference-service] [e2e-llm-inference-service] try: [e2e-llm-inference-service] # Request a connection from the queue. [e2e-llm-inference-service] timeout_obj = self._get_timeout(timeout) [e2e-llm-inference-service] conn = self._get_conn(timeout=pool_timeout) [e2e-llm-inference-service] [e2e-llm-inference-service] conn.timeout = timeout_obj.connect_timeout # type: ignore[assignment] [e2e-llm-inference-service] [e2e-llm-inference-service] # Is this a closed/new connection that requires CONNECT tunnelling? [e2e-llm-inference-service] if self.proxy is not None and http_tunnel_required and conn.is_closed: [e2e-llm-inference-service] try: [e2e-llm-inference-service] self._prepare_proxy(conn) [e2e-llm-inference-service] except (BaseSSLError, OSError, SocketTimeout) as e: [e2e-llm-inference-service] self._raise_timeout( [e2e-llm-inference-service] err=e, url=self.proxy.url, timeout_value=conn.timeout [e2e-llm-inference-service] ) [e2e-llm-inference-service] raise [e2e-llm-inference-service] [e2e-llm-inference-service] # If we're going to release the connection in ``finally:``, then [e2e-llm-inference-service] # the response doesn't need to know about the connection. Otherwise [e2e-llm-inference-service] # it will also try to release it and we'll have a double-release [e2e-llm-inference-service] # mess. [e2e-llm-inference-service] response_conn = conn if not release_conn else None [e2e-llm-inference-service] [e2e-llm-inference-service] # Make the request on the HTTPConnection object [e2e-llm-inference-service] response = self._make_request( [e2e-llm-inference-service] conn, [e2e-llm-inference-service] method, [e2e-llm-inference-service] url, [e2e-llm-inference-service] timeout=timeout_obj, [e2e-llm-inference-service] body=body, [e2e-llm-inference-service] headers=headers, [e2e-llm-inference-service] chunked=chunked, [e2e-llm-inference-service] retries=retries, [e2e-llm-inference-service] response_conn=response_conn, [e2e-llm-inference-service] preload_content=preload_content, [e2e-llm-inference-service] decode_content=decode_content, [e2e-llm-inference-service] **response_kw, [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] # Everything went great! [e2e-llm-inference-service] clean_exit = True [e2e-llm-inference-service] [e2e-llm-inference-service] except EmptyPoolError: [e2e-llm-inference-service] # Didn't get a connection from the pool, no need to clean up [e2e-llm-inference-service] clean_exit = True [e2e-llm-inference-service] release_this_conn = False [e2e-llm-inference-service] raise [e2e-llm-inference-service] [e2e-llm-inference-service] except ( [e2e-llm-inference-service] TimeoutError, [e2e-llm-inference-service] HTTPException, [e2e-llm-inference-service] OSError, [e2e-llm-inference-service] ProtocolError, [e2e-llm-inference-service] BaseSSLError, [e2e-llm-inference-service] SSLError, [e2e-llm-inference-service] CertificateError, [e2e-llm-inference-service] ProxyError, [e2e-llm-inference-service] ) as e: [e2e-llm-inference-service] # Discard the connection for these exceptions. It will be [e2e-llm-inference-service] # replaced during the next _get_conn() call. [e2e-llm-inference-service] clean_exit = False [e2e-llm-inference-service] new_e: Exception = e [e2e-llm-inference-service] if isinstance(e, (BaseSSLError, CertificateError)): [e2e-llm-inference-service] new_e = SSLError(e) [e2e-llm-inference-service] if isinstance( [e2e-llm-inference-service] new_e, [e2e-llm-inference-service] ( [e2e-llm-inference-service] OSError, [e2e-llm-inference-service] NewConnectionError, [e2e-llm-inference-service] TimeoutError, [e2e-llm-inference-service] SSLError, [e2e-llm-inference-service] HTTPException, [e2e-llm-inference-service] ), [e2e-llm-inference-service] ) and (conn and conn.proxy and not conn.has_connected_to_proxy): [e2e-llm-inference-service] new_e = _wrap_proxy_error(new_e, conn.proxy.scheme) [e2e-llm-inference-service] elif isinstance(new_e, (OSError, HTTPException)): [e2e-llm-inference-service] new_e = ProtocolError("Connection aborted.", new_e) [e2e-llm-inference-service] [e2e-llm-inference-service] retries = retries.increment( [e2e-llm-inference-service] method, url, error=new_e, _pool=self, _stacktrace=sys.exc_info()[2] [e2e-llm-inference-service] ) [e2e-llm-inference-service] retries.sleep() [e2e-llm-inference-service] [e2e-llm-inference-service] # Keep track of the error for the retry warning. [e2e-llm-inference-service] err = e [e2e-llm-inference-service] [e2e-llm-inference-service] finally: [e2e-llm-inference-service] if not clean_exit: [e2e-llm-inference-service] # We hit some kind of exception, handled or otherwise. We need [e2e-llm-inference-service] # to throw the connection away unless explicitly told not to. [e2e-llm-inference-service] # Close the connection, set the variable to None, and make sure [e2e-llm-inference-service] # we put the None back in the pool to avoid leaking it. [e2e-llm-inference-service] if conn: [e2e-llm-inference-service] conn.close() [e2e-llm-inference-service] conn = None [e2e-llm-inference-service] release_this_conn = True [e2e-llm-inference-service] [e2e-llm-inference-service] if release_this_conn: [e2e-llm-inference-service] # Put the connection back to be reused. If the connection is [e2e-llm-inference-service] # expired then it will be None, which will get replaced with a [e2e-llm-inference-service] # fresh connection during _get_conn. [e2e-llm-inference-service] self._put_conn(conn) [e2e-llm-inference-service] [e2e-llm-inference-service] if not conn: [e2e-llm-inference-service] # Try again [e2e-llm-inference-service] log.warning( [e2e-llm-inference-service] "Retrying (%r) after connection broken by '%r': %s", retries, err, url [e2e-llm-inference-service] ) [e2e-llm-inference-service] > return self.urlopen( [e2e-llm-inference-service] method, [e2e-llm-inference-service] url, [e2e-llm-inference-service] body, [e2e-llm-inference-service] headers, [e2e-llm-inference-service] retries, [e2e-llm-inference-service] redirect, [e2e-llm-inference-service] assert_same_host, [e2e-llm-inference-service] timeout=timeout, [e2e-llm-inference-service] pool_timeout=pool_timeout, [e2e-llm-inference-service] release_conn=release_conn, [e2e-llm-inference-service] chunked=chunked, [e2e-llm-inference-service] body_pos=body_pos, [e2e-llm-inference-service] preload_content=preload_content, [e2e-llm-inference-service] decode_content=decode_content, [e2e-llm-inference-service] **response_kw, [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] ../../python/kserve/.venv/lib64/python3.11/site-packages/urllib3/connectionpool.py:871: [e2e-llm-inference-service] _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ [e2e-llm-inference-service] [e2e-llm-inference-service] self = [e2e-llm-inference-service] method = 'POST' [e2e-llm-inference-service] url = '/kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-route-87882a8e/v1/completions' [e2e-llm-inference-service] body = b'{"model": "facebook/opt-125m", "prompt": "KServe is a", "max_tokens": 20}' [e2e-llm-inference-service] headers = {'User-Agent': 'python-requests/2.32.3', 'Accept-Encoding': 'gzip, deflate', 'Accept': '*/*', 'Connection': 'keep-alive', 'Content-Type': 'application/json', 'Content-Length': '73'} [e2e-llm-inference-service] retries = Retry(total=4, connect=None, read=None, redirect=None, status=None) [e2e-llm-inference-service] redirect = False, assert_same_host = False [e2e-llm-inference-service] timeout = Timeout(connect=60, read=60, total=None), pool_timeout = None [e2e-llm-inference-service] release_conn = False, chunked = False, body_pos = None, preload_content = False [e2e-llm-inference-service] decode_content = False, response_kw = {} [e2e-llm-inference-service] parsed_url = Url(scheme=None, auth=None, host=None, port=None, path='/kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-route-87882a8e/v1/completions', query=None, fragment=None) [e2e-llm-inference-service] destination_scheme = None, conn = None, release_this_conn = True [e2e-llm-inference-service] http_tunnel_required = False [e2e-llm-inference-service] err = ReadTimeoutError("HTTPConnectionPool(host='ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com', port=80): Read timed out. (read timeout=60)") [e2e-llm-inference-service] clean_exit = False [e2e-llm-inference-service] [e2e-llm-inference-service] def urlopen( # type: ignore[override] [e2e-llm-inference-service] self, [e2e-llm-inference-service] method: str, [e2e-llm-inference-service] url: str, [e2e-llm-inference-service] body: _TYPE_BODY | None = None, [e2e-llm-inference-service] headers: typing.Mapping[str, str] | None = None, [e2e-llm-inference-service] retries: Retry | bool | int | None = None, [e2e-llm-inference-service] redirect: bool = True, [e2e-llm-inference-service] assert_same_host: bool = True, [e2e-llm-inference-service] timeout: _TYPE_TIMEOUT = _DEFAULT_TIMEOUT, [e2e-llm-inference-service] pool_timeout: int | None = None, [e2e-llm-inference-service] release_conn: bool | None = None, [e2e-llm-inference-service] chunked: bool = False, [e2e-llm-inference-service] body_pos: _TYPE_BODY_POSITION | None = None, [e2e-llm-inference-service] preload_content: bool = True, [e2e-llm-inference-service] decode_content: bool = True, [e2e-llm-inference-service] **response_kw: typing.Any, [e2e-llm-inference-service] ) -> BaseHTTPResponse: [e2e-llm-inference-service] """ [e2e-llm-inference-service] Get a connection from the pool and perform an HTTP request. This is the [e2e-llm-inference-service] lowest level call for making a request, so you'll need to specify all [e2e-llm-inference-service] the raw details. [e2e-llm-inference-service] [e2e-llm-inference-service] .. note:: [e2e-llm-inference-service] [e2e-llm-inference-service] More commonly, it's appropriate to use a convenience method [e2e-llm-inference-service] such as :meth:`request`. [e2e-llm-inference-service] [e2e-llm-inference-service] .. note:: [e2e-llm-inference-service] [e2e-llm-inference-service] `release_conn` will only behave as expected if [e2e-llm-inference-service] `preload_content=False` because we want to make [e2e-llm-inference-service] `preload_content=False` the default behaviour someday soon without [e2e-llm-inference-service] breaking backwards compatibility. [e2e-llm-inference-service] [e2e-llm-inference-service] :param method: [e2e-llm-inference-service] HTTP request method (such as GET, POST, PUT, etc.) [e2e-llm-inference-service] [e2e-llm-inference-service] :param url: [e2e-llm-inference-service] The URL to perform the request on. [e2e-llm-inference-service] [e2e-llm-inference-service] :param body: [e2e-llm-inference-service] Data to send in the request body, either :class:`str`, :class:`bytes`, [e2e-llm-inference-service] an iterable of :class:`str`/:class:`bytes`, or a file-like object. [e2e-llm-inference-service] [e2e-llm-inference-service] :param headers: [e2e-llm-inference-service] Dictionary of custom headers to send, such as User-Agent, [e2e-llm-inference-service] If-None-Match, etc. If None, pool headers are used. If provided, [e2e-llm-inference-service] these headers completely replace any pool-specific headers. [e2e-llm-inference-service] [e2e-llm-inference-service] :param retries: [e2e-llm-inference-service] Configure the number of retries to allow before raising a [e2e-llm-inference-service] :class:`~urllib3.exceptions.MaxRetryError` exception. [e2e-llm-inference-service] [e2e-llm-inference-service] If ``None`` (default) will retry 3 times, see ``Retry.DEFAULT``. Pass a [e2e-llm-inference-service] :class:`~urllib3.util.retry.Retry` object for fine-grained control [e2e-llm-inference-service] over different types of retries. [e2e-llm-inference-service] Pass an integer number to retry connection errors that many times, [e2e-llm-inference-service] but no other types of errors. Pass zero to never retry. [e2e-llm-inference-service] [e2e-llm-inference-service] If ``False``, then retries are disabled and any exception is raised [e2e-llm-inference-service] immediately. Also, instead of raising a MaxRetryError on redirects, [e2e-llm-inference-service] the redirect response will be returned. [e2e-llm-inference-service] [e2e-llm-inference-service] :type retries: :class:`~urllib3.util.retry.Retry`, False, or an int. [e2e-llm-inference-service] [e2e-llm-inference-service] :param redirect: [e2e-llm-inference-service] If True, automatically handle redirects (status codes 301, 302, [e2e-llm-inference-service] 303, 307, 308). Each redirect counts as a retry. Disabling retries [e2e-llm-inference-service] will disable redirect, too. [e2e-llm-inference-service] [e2e-llm-inference-service] :param assert_same_host: [e2e-llm-inference-service] If ``True``, will make sure that the host of the pool requests is [e2e-llm-inference-service] consistent else will raise HostChangedError. When ``False``, you can [e2e-llm-inference-service] use the pool on an HTTP proxy and request foreign hosts. [e2e-llm-inference-service] [e2e-llm-inference-service] :param timeout: [e2e-llm-inference-service] If specified, overrides the default timeout for this one [e2e-llm-inference-service] request. It may be a float (in seconds) or an instance of [e2e-llm-inference-service] :class:`urllib3.util.Timeout`. [e2e-llm-inference-service] [e2e-llm-inference-service] :param pool_timeout: [e2e-llm-inference-service] If set and the pool is set to block=True, then this method will [e2e-llm-inference-service] block for ``pool_timeout`` seconds and raise EmptyPoolError if no [e2e-llm-inference-service] connection is available within the time period. [e2e-llm-inference-service] [e2e-llm-inference-service] :param bool preload_content: [e2e-llm-inference-service] If True, the response's body will be preloaded into memory. [e2e-llm-inference-service] [e2e-llm-inference-service] :param bool decode_content: [e2e-llm-inference-service] If True, will attempt to decode the body based on the [e2e-llm-inference-service] 'content-encoding' header. [e2e-llm-inference-service] [e2e-llm-inference-service] :param release_conn: [e2e-llm-inference-service] If False, then the urlopen call will not release the connection [e2e-llm-inference-service] back into the pool once a response is received (but will release if [e2e-llm-inference-service] you read the entire contents of the response such as when [e2e-llm-inference-service] `preload_content=True`). This is useful if you're not preloading [e2e-llm-inference-service] the response's content immediately. You will need to call [e2e-llm-inference-service] ``r.release_conn()`` on the response ``r`` to return the connection [e2e-llm-inference-service] back into the pool. If None, it takes the value of ``preload_content`` [e2e-llm-inference-service] which defaults to ``True``. [e2e-llm-inference-service] [e2e-llm-inference-service] :param bool chunked: [e2e-llm-inference-service] If True, urllib3 will send the body using chunked transfer [e2e-llm-inference-service] encoding. Otherwise, urllib3 will send the body using the standard [e2e-llm-inference-service] content-length form. Defaults to False. [e2e-llm-inference-service] [e2e-llm-inference-service] :param int body_pos: [e2e-llm-inference-service] Position to seek to in file-like body in the event of a retry or [e2e-llm-inference-service] redirect. Typically this won't need to be set because urllib3 will [e2e-llm-inference-service] auto-populate the value when needed. [e2e-llm-inference-service] """ [e2e-llm-inference-service] parsed_url = parse_url(url) [e2e-llm-inference-service] destination_scheme = parsed_url.scheme [e2e-llm-inference-service] [e2e-llm-inference-service] if headers is None: [e2e-llm-inference-service] headers = self.headers [e2e-llm-inference-service] [e2e-llm-inference-service] if not isinstance(retries, Retry): [e2e-llm-inference-service] retries = Retry.from_int(retries, redirect=redirect, default=self.retries) [e2e-llm-inference-service] [e2e-llm-inference-service] if release_conn is None: [e2e-llm-inference-service] release_conn = preload_content [e2e-llm-inference-service] [e2e-llm-inference-service] # Check host [e2e-llm-inference-service] if assert_same_host and not self.is_same_host(url): [e2e-llm-inference-service] raise HostChangedError(self, url, retries) [e2e-llm-inference-service] [e2e-llm-inference-service] # Ensure that the URL we're connecting to is properly encoded [e2e-llm-inference-service] if url.startswith("/"): [e2e-llm-inference-service] url = to_str(_encode_target(url)) [e2e-llm-inference-service] else: [e2e-llm-inference-service] url = to_str(parsed_url.url) [e2e-llm-inference-service] [e2e-llm-inference-service] conn = None [e2e-llm-inference-service] [e2e-llm-inference-service] # Track whether `conn` needs to be released before [e2e-llm-inference-service] # returning/raising/recursing. Update this variable if necessary, and [e2e-llm-inference-service] # leave `release_conn` constant throughout the function. That way, if [e2e-llm-inference-service] # the function recurses, the original value of `release_conn` will be [e2e-llm-inference-service] # passed down into the recursive call, and its value will be respected. [e2e-llm-inference-service] # [e2e-llm-inference-service] # See issue #651 [1] for details. [e2e-llm-inference-service] # [e2e-llm-inference-service] # [1] [e2e-llm-inference-service] release_this_conn = release_conn [e2e-llm-inference-service] [e2e-llm-inference-service] http_tunnel_required = connection_requires_http_tunnel( [e2e-llm-inference-service] self.proxy, self.proxy_config, destination_scheme [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] # Merge the proxy headers. Only done when not using HTTP CONNECT. We [e2e-llm-inference-service] # have to copy the headers dict so we can safely change it without those [e2e-llm-inference-service] # changes being reflected in anyone else's copy. [e2e-llm-inference-service] if not http_tunnel_required: [e2e-llm-inference-service] headers = headers.copy() # type: ignore[attr-defined] [e2e-llm-inference-service] headers.update(self.proxy_headers) # type: ignore[union-attr] [e2e-llm-inference-service] [e2e-llm-inference-service] # Must keep the exception bound to a separate variable or else Python 3 [e2e-llm-inference-service] # complains about UnboundLocalError. [e2e-llm-inference-service] err = None [e2e-llm-inference-service] [e2e-llm-inference-service] # Keep track of whether we cleanly exited the except block. This [e2e-llm-inference-service] # ensures we do proper cleanup in finally. [e2e-llm-inference-service] clean_exit = False [e2e-llm-inference-service] [e2e-llm-inference-service] # Rewind body position, if needed. Record current position [e2e-llm-inference-service] # for future rewinds in the event of a redirect/retry. [e2e-llm-inference-service] body_pos = set_file_position(body, body_pos) [e2e-llm-inference-service] [e2e-llm-inference-service] try: [e2e-llm-inference-service] # Request a connection from the queue. [e2e-llm-inference-service] timeout_obj = self._get_timeout(timeout) [e2e-llm-inference-service] conn = self._get_conn(timeout=pool_timeout) [e2e-llm-inference-service] [e2e-llm-inference-service] conn.timeout = timeout_obj.connect_timeout # type: ignore[assignment] [e2e-llm-inference-service] [e2e-llm-inference-service] # Is this a closed/new connection that requires CONNECT tunnelling? [e2e-llm-inference-service] if self.proxy is not None and http_tunnel_required and conn.is_closed: [e2e-llm-inference-service] try: [e2e-llm-inference-service] self._prepare_proxy(conn) [e2e-llm-inference-service] except (BaseSSLError, OSError, SocketTimeout) as e: [e2e-llm-inference-service] self._raise_timeout( [e2e-llm-inference-service] err=e, url=self.proxy.url, timeout_value=conn.timeout [e2e-llm-inference-service] ) [e2e-llm-inference-service] raise [e2e-llm-inference-service] [e2e-llm-inference-service] # If we're going to release the connection in ``finally:``, then [e2e-llm-inference-service] # the response doesn't need to know about the connection. Otherwise [e2e-llm-inference-service] # it will also try to release it and we'll have a double-release [e2e-llm-inference-service] # mess. [e2e-llm-inference-service] response_conn = conn if not release_conn else None [e2e-llm-inference-service] [e2e-llm-inference-service] # Make the request on the HTTPConnection object [e2e-llm-inference-service] response = self._make_request( [e2e-llm-inference-service] conn, [e2e-llm-inference-service] method, [e2e-llm-inference-service] url, [e2e-llm-inference-service] timeout=timeout_obj, [e2e-llm-inference-service] body=body, [e2e-llm-inference-service] headers=headers, [e2e-llm-inference-service] chunked=chunked, [e2e-llm-inference-service] retries=retries, [e2e-llm-inference-service] response_conn=response_conn, [e2e-llm-inference-service] preload_content=preload_content, [e2e-llm-inference-service] decode_content=decode_content, [e2e-llm-inference-service] **response_kw, [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] # Everything went great! [e2e-llm-inference-service] clean_exit = True [e2e-llm-inference-service] [e2e-llm-inference-service] except EmptyPoolError: [e2e-llm-inference-service] # Didn't get a connection from the pool, no need to clean up [e2e-llm-inference-service] clean_exit = True [e2e-llm-inference-service] release_this_conn = False [e2e-llm-inference-service] raise [e2e-llm-inference-service] [e2e-llm-inference-service] except ( [e2e-llm-inference-service] TimeoutError, [e2e-llm-inference-service] HTTPException, [e2e-llm-inference-service] OSError, [e2e-llm-inference-service] ProtocolError, [e2e-llm-inference-service] BaseSSLError, [e2e-llm-inference-service] SSLError, [e2e-llm-inference-service] CertificateError, [e2e-llm-inference-service] ProxyError, [e2e-llm-inference-service] ) as e: [e2e-llm-inference-service] # Discard the connection for these exceptions. It will be [e2e-llm-inference-service] # replaced during the next _get_conn() call. [e2e-llm-inference-service] clean_exit = False [e2e-llm-inference-service] new_e: Exception = e [e2e-llm-inference-service] if isinstance(e, (BaseSSLError, CertificateError)): [e2e-llm-inference-service] new_e = SSLError(e) [e2e-llm-inference-service] if isinstance( [e2e-llm-inference-service] new_e, [e2e-llm-inference-service] ( [e2e-llm-inference-service] OSError, [e2e-llm-inference-service] NewConnectionError, [e2e-llm-inference-service] TimeoutError, [e2e-llm-inference-service] SSLError, [e2e-llm-inference-service] HTTPException, [e2e-llm-inference-service] ), [e2e-llm-inference-service] ) and (conn and conn.proxy and not conn.has_connected_to_proxy): [e2e-llm-inference-service] new_e = _wrap_proxy_error(new_e, conn.proxy.scheme) [e2e-llm-inference-service] elif isinstance(new_e, (OSError, HTTPException)): [e2e-llm-inference-service] new_e = ProtocolError("Connection aborted.", new_e) [e2e-llm-inference-service] [e2e-llm-inference-service] retries = retries.increment( [e2e-llm-inference-service] method, url, error=new_e, _pool=self, _stacktrace=sys.exc_info()[2] [e2e-llm-inference-service] ) [e2e-llm-inference-service] retries.sleep() [e2e-llm-inference-service] [e2e-llm-inference-service] # Keep track of the error for the retry warning. [e2e-llm-inference-service] err = e [e2e-llm-inference-service] [e2e-llm-inference-service] finally: [e2e-llm-inference-service] if not clean_exit: [e2e-llm-inference-service] # We hit some kind of exception, handled or otherwise. We need [e2e-llm-inference-service] # to throw the connection away unless explicitly told not to. [e2e-llm-inference-service] # Close the connection, set the variable to None, and make sure [e2e-llm-inference-service] # we put the None back in the pool to avoid leaking it. [e2e-llm-inference-service] if conn: [e2e-llm-inference-service] conn.close() [e2e-llm-inference-service] conn = None [e2e-llm-inference-service] release_this_conn = True [e2e-llm-inference-service] [e2e-llm-inference-service] if release_this_conn: [e2e-llm-inference-service] # Put the connection back to be reused. If the connection is [e2e-llm-inference-service] # expired then it will be None, which will get replaced with a [e2e-llm-inference-service] # fresh connection during _get_conn. [e2e-llm-inference-service] self._put_conn(conn) [e2e-llm-inference-service] [e2e-llm-inference-service] if not conn: [e2e-llm-inference-service] # Try again [e2e-llm-inference-service] log.warning( [e2e-llm-inference-service] "Retrying (%r) after connection broken by '%r': %s", retries, err, url [e2e-llm-inference-service] ) [e2e-llm-inference-service] > return self.urlopen( [e2e-llm-inference-service] method, [e2e-llm-inference-service] url, [e2e-llm-inference-service] body, [e2e-llm-inference-service] headers, [e2e-llm-inference-service] retries, [e2e-llm-inference-service] redirect, [e2e-llm-inference-service] assert_same_host, [e2e-llm-inference-service] timeout=timeout, [e2e-llm-inference-service] pool_timeout=pool_timeout, [e2e-llm-inference-service] release_conn=release_conn, [e2e-llm-inference-service] chunked=chunked, [e2e-llm-inference-service] body_pos=body_pos, [e2e-llm-inference-service] preload_content=preload_content, [e2e-llm-inference-service] decode_content=decode_content, [e2e-llm-inference-service] **response_kw, [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] ../../python/kserve/.venv/lib64/python3.11/site-packages/urllib3/connectionpool.py:871: [e2e-llm-inference-service] _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ [e2e-llm-inference-service] [e2e-llm-inference-service] self = [e2e-llm-inference-service] method = 'POST' [e2e-llm-inference-service] url = '/kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-route-87882a8e/v1/completions' [e2e-llm-inference-service] body = b'{"model": "facebook/opt-125m", "prompt": "KServe is a", "max_tokens": 20}' [e2e-llm-inference-service] headers = {'User-Agent': 'python-requests/2.32.3', 'Accept-Encoding': 'gzip, deflate', 'Accept': '*/*', 'Connection': 'keep-alive', 'Content-Type': 'application/json', 'Content-Length': '73'} [e2e-llm-inference-service] retries = Retry(total=3, connect=None, read=None, redirect=None, status=None) [e2e-llm-inference-service] redirect = False, assert_same_host = False [e2e-llm-inference-service] timeout = Timeout(connect=60, read=60, total=None), pool_timeout = None [e2e-llm-inference-service] release_conn = False, chunked = False, body_pos = None, preload_content = False [e2e-llm-inference-service] decode_content = False, response_kw = {} [e2e-llm-inference-service] parsed_url = Url(scheme=None, auth=None, host=None, port=None, path='/kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-route-87882a8e/v1/completions', query=None, fragment=None) [e2e-llm-inference-service] destination_scheme = None, conn = None, release_this_conn = True [e2e-llm-inference-service] http_tunnel_required = False [e2e-llm-inference-service] err = ReadTimeoutError("HTTPConnectionPool(host='ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com', port=80): Read timed out. (read timeout=60)") [e2e-llm-inference-service] clean_exit = False [e2e-llm-inference-service] [e2e-llm-inference-service] def urlopen( # type: ignore[override] [e2e-llm-inference-service] self, [e2e-llm-inference-service] method: str, [e2e-llm-inference-service] url: str, [e2e-llm-inference-service] body: _TYPE_BODY | None = None, [e2e-llm-inference-service] headers: typing.Mapping[str, str] | None = None, [e2e-llm-inference-service] retries: Retry | bool | int | None = None, [e2e-llm-inference-service] redirect: bool = True, [e2e-llm-inference-service] assert_same_host: bool = True, [e2e-llm-inference-service] timeout: _TYPE_TIMEOUT = _DEFAULT_TIMEOUT, [e2e-llm-inference-service] pool_timeout: int | None = None, [e2e-llm-inference-service] release_conn: bool | None = None, [e2e-llm-inference-service] chunked: bool = False, [e2e-llm-inference-service] body_pos: _TYPE_BODY_POSITION | None = None, [e2e-llm-inference-service] preload_content: bool = True, [e2e-llm-inference-service] decode_content: bool = True, [e2e-llm-inference-service] **response_kw: typing.Any, [e2e-llm-inference-service] ) -> BaseHTTPResponse: [e2e-llm-inference-service] """ [e2e-llm-inference-service] Get a connection from the pool and perform an HTTP request. This is the [e2e-llm-inference-service] lowest level call for making a request, so you'll need to specify all [e2e-llm-inference-service] the raw details. [e2e-llm-inference-service] [e2e-llm-inference-service] .. note:: [e2e-llm-inference-service] [e2e-llm-inference-service] More commonly, it's appropriate to use a convenience method [e2e-llm-inference-service] such as :meth:`request`. [e2e-llm-inference-service] [e2e-llm-inference-service] .. note:: [e2e-llm-inference-service] [e2e-llm-inference-service] `release_conn` will only behave as expected if [e2e-llm-inference-service] `preload_content=False` because we want to make [e2e-llm-inference-service] `preload_content=False` the default behaviour someday soon without [e2e-llm-inference-service] breaking backwards compatibility. [e2e-llm-inference-service] [e2e-llm-inference-service] :param method: [e2e-llm-inference-service] HTTP request method (such as GET, POST, PUT, etc.) [e2e-llm-inference-service] [e2e-llm-inference-service] :param url: [e2e-llm-inference-service] The URL to perform the request on. [e2e-llm-inference-service] [e2e-llm-inference-service] :param body: [e2e-llm-inference-service] Data to send in the request body, either :class:`str`, :class:`bytes`, [e2e-llm-inference-service] an iterable of :class:`str`/:class:`bytes`, or a file-like object. [e2e-llm-inference-service] [e2e-llm-inference-service] :param headers: [e2e-llm-inference-service] Dictionary of custom headers to send, such as User-Agent, [e2e-llm-inference-service] If-None-Match, etc. If None, pool headers are used. If provided, [e2e-llm-inference-service] these headers completely replace any pool-specific headers. [e2e-llm-inference-service] [e2e-llm-inference-service] :param retries: [e2e-llm-inference-service] Configure the number of retries to allow before raising a [e2e-llm-inference-service] :class:`~urllib3.exceptions.MaxRetryError` exception. [e2e-llm-inference-service] [e2e-llm-inference-service] If ``None`` (default) will retry 3 times, see ``Retry.DEFAULT``. Pass a [e2e-llm-inference-service] :class:`~urllib3.util.retry.Retry` object for fine-grained control [e2e-llm-inference-service] over different types of retries. [e2e-llm-inference-service] Pass an integer number to retry connection errors that many times, [e2e-llm-inference-service] but no other types of errors. Pass zero to never retry. [e2e-llm-inference-service] [e2e-llm-inference-service] If ``False``, then retries are disabled and any exception is raised [e2e-llm-inference-service] immediately. Also, instead of raising a MaxRetryError on redirects, [e2e-llm-inference-service] the redirect response will be returned. [e2e-llm-inference-service] [e2e-llm-inference-service] :type retries: :class:`~urllib3.util.retry.Retry`, False, or an int. [e2e-llm-inference-service] [e2e-llm-inference-service] :param redirect: [e2e-llm-inference-service] If True, automatically handle redirects (status codes 301, 302, [e2e-llm-inference-service] 303, 307, 308). Each redirect counts as a retry. Disabling retries [e2e-llm-inference-service] will disable redirect, too. [e2e-llm-inference-service] [e2e-llm-inference-service] :param assert_same_host: [e2e-llm-inference-service] If ``True``, will make sure that the host of the pool requests is [e2e-llm-inference-service] consistent else will raise HostChangedError. When ``False``, you can [e2e-llm-inference-service] use the pool on an HTTP proxy and request foreign hosts. [e2e-llm-inference-service] [e2e-llm-inference-service] :param timeout: [e2e-llm-inference-service] If specified, overrides the default timeout for this one [e2e-llm-inference-service] request. It may be a float (in seconds) or an instance of [e2e-llm-inference-service] :class:`urllib3.util.Timeout`. [e2e-llm-inference-service] [e2e-llm-inference-service] :param pool_timeout: [e2e-llm-inference-service] If set and the pool is set to block=True, then this method will [e2e-llm-inference-service] block for ``pool_timeout`` seconds and raise EmptyPoolError if no [e2e-llm-inference-service] connection is available within the time period. [e2e-llm-inference-service] [e2e-llm-inference-service] :param bool preload_content: [e2e-llm-inference-service] If True, the response's body will be preloaded into memory. [e2e-llm-inference-service] [e2e-llm-inference-service] :param bool decode_content: [e2e-llm-inference-service] If True, will attempt to decode the body based on the [e2e-llm-inference-service] 'content-encoding' header. [e2e-llm-inference-service] [e2e-llm-inference-service] :param release_conn: [e2e-llm-inference-service] If False, then the urlopen call will not release the connection [e2e-llm-inference-service] back into the pool once a response is received (but will release if [e2e-llm-inference-service] you read the entire contents of the response such as when [e2e-llm-inference-service] `preload_content=True`). This is useful if you're not preloading [e2e-llm-inference-service] the response's content immediately. You will need to call [e2e-llm-inference-service] ``r.release_conn()`` on the response ``r`` to return the connection [e2e-llm-inference-service] back into the pool. If None, it takes the value of ``preload_content`` [e2e-llm-inference-service] which defaults to ``True``. [e2e-llm-inference-service] [e2e-llm-inference-service] :param bool chunked: [e2e-llm-inference-service] If True, urllib3 will send the body using chunked transfer [e2e-llm-inference-service] encoding. Otherwise, urllib3 will send the body using the standard [e2e-llm-inference-service] content-length form. Defaults to False. [e2e-llm-inference-service] [e2e-llm-inference-service] :param int body_pos: [e2e-llm-inference-service] Position to seek to in file-like body in the event of a retry or [e2e-llm-inference-service] redirect. Typically this won't need to be set because urllib3 will [e2e-llm-inference-service] auto-populate the value when needed. [e2e-llm-inference-service] """ [e2e-llm-inference-service] parsed_url = parse_url(url) [e2e-llm-inference-service] destination_scheme = parsed_url.scheme [e2e-llm-inference-service] [e2e-llm-inference-service] if headers is None: [e2e-llm-inference-service] headers = self.headers [e2e-llm-inference-service] [e2e-llm-inference-service] if not isinstance(retries, Retry): [e2e-llm-inference-service] retries = Retry.from_int(retries, redirect=redirect, default=self.retries) [e2e-llm-inference-service] [e2e-llm-inference-service] if release_conn is None: [e2e-llm-inference-service] release_conn = preload_content [e2e-llm-inference-service] [e2e-llm-inference-service] # Check host [e2e-llm-inference-service] if assert_same_host and not self.is_same_host(url): [e2e-llm-inference-service] raise HostChangedError(self, url, retries) [e2e-llm-inference-service] [e2e-llm-inference-service] # Ensure that the URL we're connecting to is properly encoded [e2e-llm-inference-service] if url.startswith("/"): [e2e-llm-inference-service] url = to_str(_encode_target(url)) [e2e-llm-inference-service] else: [e2e-llm-inference-service] url = to_str(parsed_url.url) [e2e-llm-inference-service] [e2e-llm-inference-service] conn = None [e2e-llm-inference-service] [e2e-llm-inference-service] # Track whether `conn` needs to be released before [e2e-llm-inference-service] # returning/raising/recursing. Update this variable if necessary, and [e2e-llm-inference-service] # leave `release_conn` constant throughout the function. That way, if [e2e-llm-inference-service] # the function recurses, the original value of `release_conn` will be [e2e-llm-inference-service] # passed down into the recursive call, and its value will be respected. [e2e-llm-inference-service] # [e2e-llm-inference-service] # See issue #651 [1] for details. [e2e-llm-inference-service] # [e2e-llm-inference-service] # [1] [e2e-llm-inference-service] release_this_conn = release_conn [e2e-llm-inference-service] [e2e-llm-inference-service] http_tunnel_required = connection_requires_http_tunnel( [e2e-llm-inference-service] self.proxy, self.proxy_config, destination_scheme [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] # Merge the proxy headers. Only done when not using HTTP CONNECT. We [e2e-llm-inference-service] # have to copy the headers dict so we can safely change it without those [e2e-llm-inference-service] # changes being reflected in anyone else's copy. [e2e-llm-inference-service] if not http_tunnel_required: [e2e-llm-inference-service] headers = headers.copy() # type: ignore[attr-defined] [e2e-llm-inference-service] headers.update(self.proxy_headers) # type: ignore[union-attr] [e2e-llm-inference-service] [e2e-llm-inference-service] # Must keep the exception bound to a separate variable or else Python 3 [e2e-llm-inference-service] # complains about UnboundLocalError. [e2e-llm-inference-service] err = None [e2e-llm-inference-service] [e2e-llm-inference-service] # Keep track of whether we cleanly exited the except block. This [e2e-llm-inference-service] # ensures we do proper cleanup in finally. [e2e-llm-inference-service] clean_exit = False [e2e-llm-inference-service] [e2e-llm-inference-service] # Rewind body position, if needed. Record current position [e2e-llm-inference-service] # for future rewinds in the event of a redirect/retry. [e2e-llm-inference-service] body_pos = set_file_position(body, body_pos) [e2e-llm-inference-service] [e2e-llm-inference-service] try: [e2e-llm-inference-service] # Request a connection from the queue. [e2e-llm-inference-service] timeout_obj = self._get_timeout(timeout) [e2e-llm-inference-service] conn = self._get_conn(timeout=pool_timeout) [e2e-llm-inference-service] [e2e-llm-inference-service] conn.timeout = timeout_obj.connect_timeout # type: ignore[assignment] [e2e-llm-inference-service] [e2e-llm-inference-service] # Is this a closed/new connection that requires CONNECT tunnelling? [e2e-llm-inference-service] if self.proxy is not None and http_tunnel_required and conn.is_closed: [e2e-llm-inference-service] try: [e2e-llm-inference-service] self._prepare_proxy(conn) [e2e-llm-inference-service] except (BaseSSLError, OSError, SocketTimeout) as e: [e2e-llm-inference-service] self._raise_timeout( [e2e-llm-inference-service] err=e, url=self.proxy.url, timeout_value=conn.timeout [e2e-llm-inference-service] ) [e2e-llm-inference-service] raise [e2e-llm-inference-service] [e2e-llm-inference-service] # If we're going to release the connection in ``finally:``, then [e2e-llm-inference-service] # the response doesn't need to know about the connection. Otherwise [e2e-llm-inference-service] # it will also try to release it and we'll have a double-release [e2e-llm-inference-service] # mess. [e2e-llm-inference-service] response_conn = conn if not release_conn else None [e2e-llm-inference-service] [e2e-llm-inference-service] # Make the request on the HTTPConnection object [e2e-llm-inference-service] response = self._make_request( [e2e-llm-inference-service] conn, [e2e-llm-inference-service] method, [e2e-llm-inference-service] url, [e2e-llm-inference-service] timeout=timeout_obj, [e2e-llm-inference-service] body=body, [e2e-llm-inference-service] headers=headers, [e2e-llm-inference-service] chunked=chunked, [e2e-llm-inference-service] retries=retries, [e2e-llm-inference-service] response_conn=response_conn, [e2e-llm-inference-service] preload_content=preload_content, [e2e-llm-inference-service] decode_content=decode_content, [e2e-llm-inference-service] **response_kw, [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] # Everything went great! [e2e-llm-inference-service] clean_exit = True [e2e-llm-inference-service] [e2e-llm-inference-service] except EmptyPoolError: [e2e-llm-inference-service] # Didn't get a connection from the pool, no need to clean up [e2e-llm-inference-service] clean_exit = True [e2e-llm-inference-service] release_this_conn = False [e2e-llm-inference-service] raise [e2e-llm-inference-service] [e2e-llm-inference-service] except ( [e2e-llm-inference-service] TimeoutError, [e2e-llm-inference-service] HTTPException, [e2e-llm-inference-service] OSError, [e2e-llm-inference-service] ProtocolError, [e2e-llm-inference-service] BaseSSLError, [e2e-llm-inference-service] SSLError, [e2e-llm-inference-service] CertificateError, [e2e-llm-inference-service] ProxyError, [e2e-llm-inference-service] ) as e: [e2e-llm-inference-service] # Discard the connection for these exceptions. It will be [e2e-llm-inference-service] # replaced during the next _get_conn() call. [e2e-llm-inference-service] clean_exit = False [e2e-llm-inference-service] new_e: Exception = e [e2e-llm-inference-service] if isinstance(e, (BaseSSLError, CertificateError)): [e2e-llm-inference-service] new_e = SSLError(e) [e2e-llm-inference-service] if isinstance( [e2e-llm-inference-service] new_e, [e2e-llm-inference-service] ( [e2e-llm-inference-service] OSError, [e2e-llm-inference-service] NewConnectionError, [e2e-llm-inference-service] TimeoutError, [e2e-llm-inference-service] SSLError, [e2e-llm-inference-service] HTTPException, [e2e-llm-inference-service] ), [e2e-llm-inference-service] ) and (conn and conn.proxy and not conn.has_connected_to_proxy): [e2e-llm-inference-service] new_e = _wrap_proxy_error(new_e, conn.proxy.scheme) [e2e-llm-inference-service] elif isinstance(new_e, (OSError, HTTPException)): [e2e-llm-inference-service] new_e = ProtocolError("Connection aborted.", new_e) [e2e-llm-inference-service] [e2e-llm-inference-service] retries = retries.increment( [e2e-llm-inference-service] method, url, error=new_e, _pool=self, _stacktrace=sys.exc_info()[2] [e2e-llm-inference-service] ) [e2e-llm-inference-service] retries.sleep() [e2e-llm-inference-service] [e2e-llm-inference-service] # Keep track of the error for the retry warning. [e2e-llm-inference-service] err = e [e2e-llm-inference-service] [e2e-llm-inference-service] finally: [e2e-llm-inference-service] if not clean_exit: [e2e-llm-inference-service] # We hit some kind of exception, handled or otherwise. We need [e2e-llm-inference-service] # to throw the connection away unless explicitly told not to. [e2e-llm-inference-service] # Close the connection, set the variable to None, and make sure [e2e-llm-inference-service] # we put the None back in the pool to avoid leaking it. [e2e-llm-inference-service] if conn: [e2e-llm-inference-service] conn.close() [e2e-llm-inference-service] conn = None [e2e-llm-inference-service] release_this_conn = True [e2e-llm-inference-service] [e2e-llm-inference-service] if release_this_conn: [e2e-llm-inference-service] # Put the connection back to be reused. If the connection is [e2e-llm-inference-service] # expired then it will be None, which will get replaced with a [e2e-llm-inference-service] # fresh connection during _get_conn. [e2e-llm-inference-service] self._put_conn(conn) [e2e-llm-inference-service] [e2e-llm-inference-service] if not conn: [e2e-llm-inference-service] # Try again [e2e-llm-inference-service] log.warning( [e2e-llm-inference-service] "Retrying (%r) after connection broken by '%r': %s", retries, err, url [e2e-llm-inference-service] ) [e2e-llm-inference-service] > return self.urlopen( [e2e-llm-inference-service] method, [e2e-llm-inference-service] url, [e2e-llm-inference-service] body, [e2e-llm-inference-service] headers, [e2e-llm-inference-service] retries, [e2e-llm-inference-service] redirect, [e2e-llm-inference-service] assert_same_host, [e2e-llm-inference-service] timeout=timeout, [e2e-llm-inference-service] pool_timeout=pool_timeout, [e2e-llm-inference-service] release_conn=release_conn, [e2e-llm-inference-service] chunked=chunked, [e2e-llm-inference-service] body_pos=body_pos, [e2e-llm-inference-service] preload_content=preload_content, [e2e-llm-inference-service] decode_content=decode_content, [e2e-llm-inference-service] **response_kw, [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] ../../python/kserve/.venv/lib64/python3.11/site-packages/urllib3/connectionpool.py:871: [e2e-llm-inference-service] _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ [e2e-llm-inference-service] [e2e-llm-inference-service] self = [e2e-llm-inference-service] method = 'POST' [e2e-llm-inference-service] url = '/kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-route-87882a8e/v1/completions' [e2e-llm-inference-service] body = b'{"model": "facebook/opt-125m", "prompt": "KServe is a", "max_tokens": 20}' [e2e-llm-inference-service] headers = {'User-Agent': 'python-requests/2.32.3', 'Accept-Encoding': 'gzip, deflate', 'Accept': '*/*', 'Connection': 'keep-alive', 'Content-Type': 'application/json', 'Content-Length': '73'} [e2e-llm-inference-service] retries = Retry(total=2, connect=None, read=None, redirect=None, status=None) [e2e-llm-inference-service] redirect = False, assert_same_host = False [e2e-llm-inference-service] timeout = Timeout(connect=60, read=60, total=None), pool_timeout = None [e2e-llm-inference-service] release_conn = False, chunked = False, body_pos = None, preload_content = False [e2e-llm-inference-service] decode_content = False, response_kw = {} [e2e-llm-inference-service] parsed_url = Url(scheme=None, auth=None, host=None, port=None, path='/kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-route-87882a8e/v1/completions', query=None, fragment=None) [e2e-llm-inference-service] destination_scheme = None, conn = None, release_this_conn = True [e2e-llm-inference-service] http_tunnel_required = False [e2e-llm-inference-service] err = RemoteDisconnected('Remote end closed connection without response') [e2e-llm-inference-service] clean_exit = False [e2e-llm-inference-service] [e2e-llm-inference-service] def urlopen( # type: ignore[override] [e2e-llm-inference-service] self, [e2e-llm-inference-service] method: str, [e2e-llm-inference-service] url: str, [e2e-llm-inference-service] body: _TYPE_BODY | None = None, [e2e-llm-inference-service] headers: typing.Mapping[str, str] | None = None, [e2e-llm-inference-service] retries: Retry | bool | int | None = None, [e2e-llm-inference-service] redirect: bool = True, [e2e-llm-inference-service] assert_same_host: bool = True, [e2e-llm-inference-service] timeout: _TYPE_TIMEOUT = _DEFAULT_TIMEOUT, [e2e-llm-inference-service] pool_timeout: int | None = None, [e2e-llm-inference-service] release_conn: bool | None = None, [e2e-llm-inference-service] chunked: bool = False, [e2e-llm-inference-service] body_pos: _TYPE_BODY_POSITION | None = None, [e2e-llm-inference-service] preload_content: bool = True, [e2e-llm-inference-service] decode_content: bool = True, [e2e-llm-inference-service] **response_kw: typing.Any, [e2e-llm-inference-service] ) -> BaseHTTPResponse: [e2e-llm-inference-service] """ [e2e-llm-inference-service] Get a connection from the pool and perform an HTTP request. This is the [e2e-llm-inference-service] lowest level call for making a request, so you'll need to specify all [e2e-llm-inference-service] the raw details. [e2e-llm-inference-service] [e2e-llm-inference-service] .. note:: [e2e-llm-inference-service] [e2e-llm-inference-service] More commonly, it's appropriate to use a convenience method [e2e-llm-inference-service] such as :meth:`request`. [e2e-llm-inference-service] [e2e-llm-inference-service] .. note:: [e2e-llm-inference-service] [e2e-llm-inference-service] `release_conn` will only behave as expected if [e2e-llm-inference-service] `preload_content=False` because we want to make [e2e-llm-inference-service] `preload_content=False` the default behaviour someday soon without [e2e-llm-inference-service] breaking backwards compatibility. [e2e-llm-inference-service] [e2e-llm-inference-service] :param method: [e2e-llm-inference-service] HTTP request method (such as GET, POST, PUT, etc.) [e2e-llm-inference-service] [e2e-llm-inference-service] :param url: [e2e-llm-inference-service] The URL to perform the request on. [e2e-llm-inference-service] [e2e-llm-inference-service] :param body: [e2e-llm-inference-service] Data to send in the request body, either :class:`str`, :class:`bytes`, [e2e-llm-inference-service] an iterable of :class:`str`/:class:`bytes`, or a file-like object. [e2e-llm-inference-service] [e2e-llm-inference-service] :param headers: [e2e-llm-inference-service] Dictionary of custom headers to send, such as User-Agent, [e2e-llm-inference-service] If-None-Match, etc. If None, pool headers are used. If provided, [e2e-llm-inference-service] these headers completely replace any pool-specific headers. [e2e-llm-inference-service] [e2e-llm-inference-service] :param retries: [e2e-llm-inference-service] Configure the number of retries to allow before raising a [e2e-llm-inference-service] :class:`~urllib3.exceptions.MaxRetryError` exception. [e2e-llm-inference-service] [e2e-llm-inference-service] If ``None`` (default) will retry 3 times, see ``Retry.DEFAULT``. Pass a [e2e-llm-inference-service] :class:`~urllib3.util.retry.Retry` object for fine-grained control [e2e-llm-inference-service] over different types of retries. [e2e-llm-inference-service] Pass an integer number to retry connection errors that many times, [e2e-llm-inference-service] but no other types of errors. Pass zero to never retry. [e2e-llm-inference-service] [e2e-llm-inference-service] If ``False``, then retries are disabled and any exception is raised [e2e-llm-inference-service] immediately. Also, instead of raising a MaxRetryError on redirects, [e2e-llm-inference-service] the redirect response will be returned. [e2e-llm-inference-service] [e2e-llm-inference-service] :type retries: :class:`~urllib3.util.retry.Retry`, False, or an int. [e2e-llm-inference-service] [e2e-llm-inference-service] :param redirect: [e2e-llm-inference-service] If True, automatically handle redirects (status codes 301, 302, [e2e-llm-inference-service] 303, 307, 308). Each redirect counts as a retry. Disabling retries [e2e-llm-inference-service] will disable redirect, too. [e2e-llm-inference-service] [e2e-llm-inference-service] :param assert_same_host: [e2e-llm-inference-service] If ``True``, will make sure that the host of the pool requests is [e2e-llm-inference-service] consistent else will raise HostChangedError. When ``False``, you can [e2e-llm-inference-service] use the pool on an HTTP proxy and request foreign hosts. [e2e-llm-inference-service] [e2e-llm-inference-service] :param timeout: [e2e-llm-inference-service] If specified, overrides the default timeout for this one [e2e-llm-inference-service] request. It may be a float (in seconds) or an instance of [e2e-llm-inference-service] :class:`urllib3.util.Timeout`. [e2e-llm-inference-service] [e2e-llm-inference-service] :param pool_timeout: [e2e-llm-inference-service] If set and the pool is set to block=True, then this method will [e2e-llm-inference-service] block for ``pool_timeout`` seconds and raise EmptyPoolError if no [e2e-llm-inference-service] connection is available within the time period. [e2e-llm-inference-service] [e2e-llm-inference-service] :param bool preload_content: [e2e-llm-inference-service] If True, the response's body will be preloaded into memory. [e2e-llm-inference-service] [e2e-llm-inference-service] :param bool decode_content: [e2e-llm-inference-service] If True, will attempt to decode the body based on the [e2e-llm-inference-service] 'content-encoding' header. [e2e-llm-inference-service] [e2e-llm-inference-service] :param release_conn: [e2e-llm-inference-service] If False, then the urlopen call will not release the connection [e2e-llm-inference-service] back into the pool once a response is received (but will release if [e2e-llm-inference-service] you read the entire contents of the response such as when [e2e-llm-inference-service] `preload_content=True`). This is useful if you're not preloading [e2e-llm-inference-service] the response's content immediately. You will need to call [e2e-llm-inference-service] ``r.release_conn()`` on the response ``r`` to return the connection [e2e-llm-inference-service] back into the pool. If None, it takes the value of ``preload_content`` [e2e-llm-inference-service] which defaults to ``True``. [e2e-llm-inference-service] [e2e-llm-inference-service] :param bool chunked: [e2e-llm-inference-service] If True, urllib3 will send the body using chunked transfer [e2e-llm-inference-service] encoding. Otherwise, urllib3 will send the body using the standard [e2e-llm-inference-service] content-length form. Defaults to False. [e2e-llm-inference-service] [e2e-llm-inference-service] :param int body_pos: [e2e-llm-inference-service] Position to seek to in file-like body in the event of a retry or [e2e-llm-inference-service] redirect. Typically this won't need to be set because urllib3 will [e2e-llm-inference-service] auto-populate the value when needed. [e2e-llm-inference-service] """ [e2e-llm-inference-service] parsed_url = parse_url(url) [e2e-llm-inference-service] destination_scheme = parsed_url.scheme [e2e-llm-inference-service] [e2e-llm-inference-service] if headers is None: [e2e-llm-inference-service] headers = self.headers [e2e-llm-inference-service] [e2e-llm-inference-service] if not isinstance(retries, Retry): [e2e-llm-inference-service] retries = Retry.from_int(retries, redirect=redirect, default=self.retries) [e2e-llm-inference-service] [e2e-llm-inference-service] if release_conn is None: [e2e-llm-inference-service] release_conn = preload_content [e2e-llm-inference-service] [e2e-llm-inference-service] # Check host [e2e-llm-inference-service] if assert_same_host and not self.is_same_host(url): [e2e-llm-inference-service] raise HostChangedError(self, url, retries) [e2e-llm-inference-service] [e2e-llm-inference-service] # Ensure that the URL we're connecting to is properly encoded [e2e-llm-inference-service] if url.startswith("/"): [e2e-llm-inference-service] url = to_str(_encode_target(url)) [e2e-llm-inference-service] else: [e2e-llm-inference-service] url = to_str(parsed_url.url) [e2e-llm-inference-service] [e2e-llm-inference-service] conn = None [e2e-llm-inference-service] [e2e-llm-inference-service] # Track whether `conn` needs to be released before [e2e-llm-inference-service] # returning/raising/recursing. Update this variable if necessary, and [e2e-llm-inference-service] # leave `release_conn` constant throughout the function. That way, if [e2e-llm-inference-service] # the function recurses, the original value of `release_conn` will be [e2e-llm-inference-service] # passed down into the recursive call, and its value will be respected. [e2e-llm-inference-service] # [e2e-llm-inference-service] # See issue #651 [1] for details. [e2e-llm-inference-service] # [e2e-llm-inference-service] # [1] [e2e-llm-inference-service] release_this_conn = release_conn [e2e-llm-inference-service] [e2e-llm-inference-service] http_tunnel_required = connection_requires_http_tunnel( [e2e-llm-inference-service] self.proxy, self.proxy_config, destination_scheme [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] # Merge the proxy headers. Only done when not using HTTP CONNECT. We [e2e-llm-inference-service] # have to copy the headers dict so we can safely change it without those [e2e-llm-inference-service] # changes being reflected in anyone else's copy. [e2e-llm-inference-service] if not http_tunnel_required: [e2e-llm-inference-service] headers = headers.copy() # type: ignore[attr-defined] [e2e-llm-inference-service] headers.update(self.proxy_headers) # type: ignore[union-attr] [e2e-llm-inference-service] [e2e-llm-inference-service] # Must keep the exception bound to a separate variable or else Python 3 [e2e-llm-inference-service] # complains about UnboundLocalError. [e2e-llm-inference-service] err = None [e2e-llm-inference-service] [e2e-llm-inference-service] # Keep track of whether we cleanly exited the except block. This [e2e-llm-inference-service] # ensures we do proper cleanup in finally. [e2e-llm-inference-service] clean_exit = False [e2e-llm-inference-service] [e2e-llm-inference-service] # Rewind body position, if needed. Record current position [e2e-llm-inference-service] # for future rewinds in the event of a redirect/retry. [e2e-llm-inference-service] body_pos = set_file_position(body, body_pos) [e2e-llm-inference-service] [e2e-llm-inference-service] try: [e2e-llm-inference-service] # Request a connection from the queue. [e2e-llm-inference-service] timeout_obj = self._get_timeout(timeout) [e2e-llm-inference-service] conn = self._get_conn(timeout=pool_timeout) [e2e-llm-inference-service] [e2e-llm-inference-service] conn.timeout = timeout_obj.connect_timeout # type: ignore[assignment] [e2e-llm-inference-service] [e2e-llm-inference-service] # Is this a closed/new connection that requires CONNECT tunnelling? [e2e-llm-inference-service] if self.proxy is not None and http_tunnel_required and conn.is_closed: [e2e-llm-inference-service] try: [e2e-llm-inference-service] self._prepare_proxy(conn) [e2e-llm-inference-service] except (BaseSSLError, OSError, SocketTimeout) as e: [e2e-llm-inference-service] self._raise_timeout( [e2e-llm-inference-service] err=e, url=self.proxy.url, timeout_value=conn.timeout [e2e-llm-inference-service] ) [e2e-llm-inference-service] raise [e2e-llm-inference-service] [e2e-llm-inference-service] # If we're going to release the connection in ``finally:``, then [e2e-llm-inference-service] # the response doesn't need to know about the connection. Otherwise [e2e-llm-inference-service] # it will also try to release it and we'll have a double-release [e2e-llm-inference-service] # mess. [e2e-llm-inference-service] response_conn = conn if not release_conn else None [e2e-llm-inference-service] [e2e-llm-inference-service] # Make the request on the HTTPConnection object [e2e-llm-inference-service] response = self._make_request( [e2e-llm-inference-service] conn, [e2e-llm-inference-service] method, [e2e-llm-inference-service] url, [e2e-llm-inference-service] timeout=timeout_obj, [e2e-llm-inference-service] body=body, [e2e-llm-inference-service] headers=headers, [e2e-llm-inference-service] chunked=chunked, [e2e-llm-inference-service] retries=retries, [e2e-llm-inference-service] response_conn=response_conn, [e2e-llm-inference-service] preload_content=preload_content, [e2e-llm-inference-service] decode_content=decode_content, [e2e-llm-inference-service] **response_kw, [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] # Everything went great! [e2e-llm-inference-service] clean_exit = True [e2e-llm-inference-service] [e2e-llm-inference-service] except EmptyPoolError: [e2e-llm-inference-service] # Didn't get a connection from the pool, no need to clean up [e2e-llm-inference-service] clean_exit = True [e2e-llm-inference-service] release_this_conn = False [e2e-llm-inference-service] raise [e2e-llm-inference-service] [e2e-llm-inference-service] except ( [e2e-llm-inference-service] TimeoutError, [e2e-llm-inference-service] HTTPException, [e2e-llm-inference-service] OSError, [e2e-llm-inference-service] ProtocolError, [e2e-llm-inference-service] BaseSSLError, [e2e-llm-inference-service] SSLError, [e2e-llm-inference-service] CertificateError, [e2e-llm-inference-service] ProxyError, [e2e-llm-inference-service] ) as e: [e2e-llm-inference-service] # Discard the connection for these exceptions. It will be [e2e-llm-inference-service] # replaced during the next _get_conn() call. [e2e-llm-inference-service] clean_exit = False [e2e-llm-inference-service] new_e: Exception = e [e2e-llm-inference-service] if isinstance(e, (BaseSSLError, CertificateError)): [e2e-llm-inference-service] new_e = SSLError(e) [e2e-llm-inference-service] if isinstance( [e2e-llm-inference-service] new_e, [e2e-llm-inference-service] ( [e2e-llm-inference-service] OSError, [e2e-llm-inference-service] NewConnectionError, [e2e-llm-inference-service] TimeoutError, [e2e-llm-inference-service] SSLError, [e2e-llm-inference-service] HTTPException, [e2e-llm-inference-service] ), [e2e-llm-inference-service] ) and (conn and conn.proxy and not conn.has_connected_to_proxy): [e2e-llm-inference-service] new_e = _wrap_proxy_error(new_e, conn.proxy.scheme) [e2e-llm-inference-service] elif isinstance(new_e, (OSError, HTTPException)): [e2e-llm-inference-service] new_e = ProtocolError("Connection aborted.", new_e) [e2e-llm-inference-service] [e2e-llm-inference-service] retries = retries.increment( [e2e-llm-inference-service] method, url, error=new_e, _pool=self, _stacktrace=sys.exc_info()[2] [e2e-llm-inference-service] ) [e2e-llm-inference-service] retries.sleep() [e2e-llm-inference-service] [e2e-llm-inference-service] # Keep track of the error for the retry warning. [e2e-llm-inference-service] err = e [e2e-llm-inference-service] [e2e-llm-inference-service] finally: [e2e-llm-inference-service] if not clean_exit: [e2e-llm-inference-service] # We hit some kind of exception, handled or otherwise. We need [e2e-llm-inference-service] # to throw the connection away unless explicitly told not to. [e2e-llm-inference-service] # Close the connection, set the variable to None, and make sure [e2e-llm-inference-service] # we put the None back in the pool to avoid leaking it. [e2e-llm-inference-service] if conn: [e2e-llm-inference-service] conn.close() [e2e-llm-inference-service] conn = None [e2e-llm-inference-service] release_this_conn = True [e2e-llm-inference-service] [e2e-llm-inference-service] if release_this_conn: [e2e-llm-inference-service] # Put the connection back to be reused. If the connection is [e2e-llm-inference-service] # expired then it will be None, which will get replaced with a [e2e-llm-inference-service] # fresh connection during _get_conn. [e2e-llm-inference-service] self._put_conn(conn) [e2e-llm-inference-service] [e2e-llm-inference-service] if not conn: [e2e-llm-inference-service] # Try again [e2e-llm-inference-service] log.warning( [e2e-llm-inference-service] "Retrying (%r) after connection broken by '%r': %s", retries, err, url [e2e-llm-inference-service] ) [e2e-llm-inference-service] > return self.urlopen( [e2e-llm-inference-service] method, [e2e-llm-inference-service] url, [e2e-llm-inference-service] body, [e2e-llm-inference-service] headers, [e2e-llm-inference-service] retries, [e2e-llm-inference-service] redirect, [e2e-llm-inference-service] assert_same_host, [e2e-llm-inference-service] timeout=timeout, [e2e-llm-inference-service] pool_timeout=pool_timeout, [e2e-llm-inference-service] release_conn=release_conn, [e2e-llm-inference-service] chunked=chunked, [e2e-llm-inference-service] body_pos=body_pos, [e2e-llm-inference-service] preload_content=preload_content, [e2e-llm-inference-service] decode_content=decode_content, [e2e-llm-inference-service] **response_kw, [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] ../../python/kserve/.venv/lib64/python3.11/site-packages/urllib3/connectionpool.py:871: [e2e-llm-inference-service] _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ [e2e-llm-inference-service] [e2e-llm-inference-service] self = [e2e-llm-inference-service] method = 'POST' [e2e-llm-inference-service] url = '/kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-route-87882a8e/v1/completions' [e2e-llm-inference-service] body = b'{"model": "facebook/opt-125m", "prompt": "KServe is a", "max_tokens": 20}' [e2e-llm-inference-service] headers = {'User-Agent': 'python-requests/2.32.3', 'Accept-Encoding': 'gzip, deflate', 'Accept': '*/*', 'Connection': 'keep-alive', 'Content-Type': 'application/json', 'Content-Length': '73'} [e2e-llm-inference-service] retries = Retry(total=1, connect=None, read=None, redirect=None, status=None) [e2e-llm-inference-service] redirect = False, assert_same_host = False [e2e-llm-inference-service] timeout = Timeout(connect=60, read=60, total=None), pool_timeout = None [e2e-llm-inference-service] release_conn = False, chunked = False, body_pos = None, preload_content = False [e2e-llm-inference-service] decode_content = False, response_kw = {} [e2e-llm-inference-service] parsed_url = Url(scheme=None, auth=None, host=None, port=None, path='/kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-route-87882a8e/v1/completions', query=None, fragment=None) [e2e-llm-inference-service] destination_scheme = None, conn = None, release_this_conn = True [e2e-llm-inference-service] http_tunnel_required = False [e2e-llm-inference-service] err = ReadTimeoutError("HTTPConnectionPool(host='ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com', port=80): Read timed out. (read timeout=60)") [e2e-llm-inference-service] clean_exit = False [e2e-llm-inference-service] [e2e-llm-inference-service] def urlopen( # type: ignore[override] [e2e-llm-inference-service] self, [e2e-llm-inference-service] method: str, [e2e-llm-inference-service] url: str, [e2e-llm-inference-service] body: _TYPE_BODY | None = None, [e2e-llm-inference-service] headers: typing.Mapping[str, str] | None = None, [e2e-llm-inference-service] retries: Retry | bool | int | None = None, [e2e-llm-inference-service] redirect: bool = True, [e2e-llm-inference-service] assert_same_host: bool = True, [e2e-llm-inference-service] timeout: _TYPE_TIMEOUT = _DEFAULT_TIMEOUT, [e2e-llm-inference-service] pool_timeout: int | None = None, [e2e-llm-inference-service] release_conn: bool | None = None, [e2e-llm-inference-service] chunked: bool = False, [e2e-llm-inference-service] body_pos: _TYPE_BODY_POSITION | None = None, [e2e-llm-inference-service] preload_content: bool = True, [e2e-llm-inference-service] decode_content: bool = True, [e2e-llm-inference-service] **response_kw: typing.Any, [e2e-llm-inference-service] ) -> BaseHTTPResponse: [e2e-llm-inference-service] """ [e2e-llm-inference-service] Get a connection from the pool and perform an HTTP request. This is the [e2e-llm-inference-service] lowest level call for making a request, so you'll need to specify all [e2e-llm-inference-service] the raw details. [e2e-llm-inference-service] [e2e-llm-inference-service] .. note:: [e2e-llm-inference-service] [e2e-llm-inference-service] More commonly, it's appropriate to use a convenience method [e2e-llm-inference-service] such as :meth:`request`. [e2e-llm-inference-service] [e2e-llm-inference-service] .. note:: [e2e-llm-inference-service] [e2e-llm-inference-service] `release_conn` will only behave as expected if [e2e-llm-inference-service] `preload_content=False` because we want to make [e2e-llm-inference-service] `preload_content=False` the default behaviour someday soon without [e2e-llm-inference-service] breaking backwards compatibility. [e2e-llm-inference-service] [e2e-llm-inference-service] :param method: [e2e-llm-inference-service] HTTP request method (such as GET, POST, PUT, etc.) [e2e-llm-inference-service] [e2e-llm-inference-service] :param url: [e2e-llm-inference-service] The URL to perform the request on. [e2e-llm-inference-service] [e2e-llm-inference-service] :param body: [e2e-llm-inference-service] Data to send in the request body, either :class:`str`, :class:`bytes`, [e2e-llm-inference-service] an iterable of :class:`str`/:class:`bytes`, or a file-like object. [e2e-llm-inference-service] [e2e-llm-inference-service] :param headers: [e2e-llm-inference-service] Dictionary of custom headers to send, such as User-Agent, [e2e-llm-inference-service] If-None-Match, etc. If None, pool headers are used. If provided, [e2e-llm-inference-service] these headers completely replace any pool-specific headers. [e2e-llm-inference-service] [e2e-llm-inference-service] :param retries: [e2e-llm-inference-service] Configure the number of retries to allow before raising a [e2e-llm-inference-service] :class:`~urllib3.exceptions.MaxRetryError` exception. [e2e-llm-inference-service] [e2e-llm-inference-service] If ``None`` (default) will retry 3 times, see ``Retry.DEFAULT``. Pass a [e2e-llm-inference-service] :class:`~urllib3.util.retry.Retry` object for fine-grained control [e2e-llm-inference-service] over different types of retries. [e2e-llm-inference-service] Pass an integer number to retry connection errors that many times, [e2e-llm-inference-service] but no other types of errors. Pass zero to never retry. [e2e-llm-inference-service] [e2e-llm-inference-service] If ``False``, then retries are disabled and any exception is raised [e2e-llm-inference-service] immediately. Also, instead of raising a MaxRetryError on redirects, [e2e-llm-inference-service] the redirect response will be returned. [e2e-llm-inference-service] [e2e-llm-inference-service] :type retries: :class:`~urllib3.util.retry.Retry`, False, or an int. [e2e-llm-inference-service] [e2e-llm-inference-service] :param redirect: [e2e-llm-inference-service] If True, automatically handle redirects (status codes 301, 302, [e2e-llm-inference-service] 303, 307, 308). Each redirect counts as a retry. Disabling retries [e2e-llm-inference-service] will disable redirect, too. [e2e-llm-inference-service] [e2e-llm-inference-service] :param assert_same_host: [e2e-llm-inference-service] If ``True``, will make sure that the host of the pool requests is [e2e-llm-inference-service] consistent else will raise HostChangedError. When ``False``, you can [e2e-llm-inference-service] use the pool on an HTTP proxy and request foreign hosts. [e2e-llm-inference-service] [e2e-llm-inference-service] :param timeout: [e2e-llm-inference-service] If specified, overrides the default timeout for this one [e2e-llm-inference-service] request. It may be a float (in seconds) or an instance of [e2e-llm-inference-service] :class:`urllib3.util.Timeout`. [e2e-llm-inference-service] [e2e-llm-inference-service] :param pool_timeout: [e2e-llm-inference-service] If set and the pool is set to block=True, then this method will [e2e-llm-inference-service] block for ``pool_timeout`` seconds and raise EmptyPoolError if no [e2e-llm-inference-service] connection is available within the time period. [e2e-llm-inference-service] [e2e-llm-inference-service] :param bool preload_content: [e2e-llm-inference-service] If True, the response's body will be preloaded into memory. [e2e-llm-inference-service] [e2e-llm-inference-service] :param bool decode_content: [e2e-llm-inference-service] If True, will attempt to decode the body based on the [e2e-llm-inference-service] 'content-encoding' header. [e2e-llm-inference-service] [e2e-llm-inference-service] :param release_conn: [e2e-llm-inference-service] If False, then the urlopen call will not release the connection [e2e-llm-inference-service] back into the pool once a response is received (but will release if [e2e-llm-inference-service] you read the entire contents of the response such as when [e2e-llm-inference-service] `preload_content=True`). This is useful if you're not preloading [e2e-llm-inference-service] the response's content immediately. You will need to call [e2e-llm-inference-service] ``r.release_conn()`` on the response ``r`` to return the connection [e2e-llm-inference-service] back into the pool. If None, it takes the value of ``preload_content`` [e2e-llm-inference-service] which defaults to ``True``. [e2e-llm-inference-service] [e2e-llm-inference-service] :param bool chunked: [e2e-llm-inference-service] If True, urllib3 will send the body using chunked transfer [e2e-llm-inference-service] encoding. Otherwise, urllib3 will send the body using the standard [e2e-llm-inference-service] content-length form. Defaults to False. [e2e-llm-inference-service] [e2e-llm-inference-service] :param int body_pos: [e2e-llm-inference-service] Position to seek to in file-like body in the event of a retry or [e2e-llm-inference-service] redirect. Typically this won't need to be set because urllib3 will [e2e-llm-inference-service] auto-populate the value when needed. [e2e-llm-inference-service] """ [e2e-llm-inference-service] parsed_url = parse_url(url) [e2e-llm-inference-service] destination_scheme = parsed_url.scheme [e2e-llm-inference-service] [e2e-llm-inference-service] if headers is None: [e2e-llm-inference-service] headers = self.headers [e2e-llm-inference-service] [e2e-llm-inference-service] if not isinstance(retries, Retry): [e2e-llm-inference-service] retries = Retry.from_int(retries, redirect=redirect, default=self.retries) [e2e-llm-inference-service] [e2e-llm-inference-service] if release_conn is None: [e2e-llm-inference-service] release_conn = preload_content [e2e-llm-inference-service] [e2e-llm-inference-service] # Check host [e2e-llm-inference-service] if assert_same_host and not self.is_same_host(url): [e2e-llm-inference-service] raise HostChangedError(self, url, retries) [e2e-llm-inference-service] [e2e-llm-inference-service] # Ensure that the URL we're connecting to is properly encoded [e2e-llm-inference-service] if url.startswith("/"): [e2e-llm-inference-service] url = to_str(_encode_target(url)) [e2e-llm-inference-service] else: [e2e-llm-inference-service] url = to_str(parsed_url.url) [e2e-llm-inference-service] [e2e-llm-inference-service] conn = None [e2e-llm-inference-service] [e2e-llm-inference-service] # Track whether `conn` needs to be released before [e2e-llm-inference-service] # returning/raising/recursing. Update this variable if necessary, and [e2e-llm-inference-service] # leave `release_conn` constant throughout the function. That way, if [e2e-llm-inference-service] # the function recurses, the original value of `release_conn` will be [e2e-llm-inference-service] # passed down into the recursive call, and its value will be respected. [e2e-llm-inference-service] # [e2e-llm-inference-service] # See issue #651 [1] for details. [e2e-llm-inference-service] # [e2e-llm-inference-service] # [1] [e2e-llm-inference-service] release_this_conn = release_conn [e2e-llm-inference-service] [e2e-llm-inference-service] http_tunnel_required = connection_requires_http_tunnel( [e2e-llm-inference-service] self.proxy, self.proxy_config, destination_scheme [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] # Merge the proxy headers. Only done when not using HTTP CONNECT. We [e2e-llm-inference-service] # have to copy the headers dict so we can safely change it without those [e2e-llm-inference-service] # changes being reflected in anyone else's copy. [e2e-llm-inference-service] if not http_tunnel_required: [e2e-llm-inference-service] headers = headers.copy() # type: ignore[attr-defined] [e2e-llm-inference-service] headers.update(self.proxy_headers) # type: ignore[union-attr] [e2e-llm-inference-service] [e2e-llm-inference-service] # Must keep the exception bound to a separate variable or else Python 3 [e2e-llm-inference-service] # complains about UnboundLocalError. [e2e-llm-inference-service] err = None [e2e-llm-inference-service] [e2e-llm-inference-service] # Keep track of whether we cleanly exited the except block. This [e2e-llm-inference-service] # ensures we do proper cleanup in finally. [e2e-llm-inference-service] clean_exit = False [e2e-llm-inference-service] [e2e-llm-inference-service] # Rewind body position, if needed. Record current position [e2e-llm-inference-service] # for future rewinds in the event of a redirect/retry. [e2e-llm-inference-service] body_pos = set_file_position(body, body_pos) [e2e-llm-inference-service] [e2e-llm-inference-service] try: [e2e-llm-inference-service] # Request a connection from the queue. [e2e-llm-inference-service] timeout_obj = self._get_timeout(timeout) [e2e-llm-inference-service] conn = self._get_conn(timeout=pool_timeout) [e2e-llm-inference-service] [e2e-llm-inference-service] conn.timeout = timeout_obj.connect_timeout # type: ignore[assignment] [e2e-llm-inference-service] [e2e-llm-inference-service] # Is this a closed/new connection that requires CONNECT tunnelling? [e2e-llm-inference-service] if self.proxy is not None and http_tunnel_required and conn.is_closed: [e2e-llm-inference-service] try: [e2e-llm-inference-service] self._prepare_proxy(conn) [e2e-llm-inference-service] except (BaseSSLError, OSError, SocketTimeout) as e: [e2e-llm-inference-service] self._raise_timeout( [e2e-llm-inference-service] err=e, url=self.proxy.url, timeout_value=conn.timeout [e2e-llm-inference-service] ) [e2e-llm-inference-service] raise [e2e-llm-inference-service] [e2e-llm-inference-service] # If we're going to release the connection in ``finally:``, then [e2e-llm-inference-service] # the response doesn't need to know about the connection. Otherwise [e2e-llm-inference-service] # it will also try to release it and we'll have a double-release [e2e-llm-inference-service] # mess. [e2e-llm-inference-service] response_conn = conn if not release_conn else None [e2e-llm-inference-service] [e2e-llm-inference-service] # Make the request on the HTTPConnection object [e2e-llm-inference-service] response = self._make_request( [e2e-llm-inference-service] conn, [e2e-llm-inference-service] method, [e2e-llm-inference-service] url, [e2e-llm-inference-service] timeout=timeout_obj, [e2e-llm-inference-service] body=body, [e2e-llm-inference-service] headers=headers, [e2e-llm-inference-service] chunked=chunked, [e2e-llm-inference-service] retries=retries, [e2e-llm-inference-service] response_conn=response_conn, [e2e-llm-inference-service] preload_content=preload_content, [e2e-llm-inference-service] decode_content=decode_content, [e2e-llm-inference-service] **response_kw, [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] # Everything went great! [e2e-llm-inference-service] clean_exit = True [e2e-llm-inference-service] [e2e-llm-inference-service] except EmptyPoolError: [e2e-llm-inference-service] # Didn't get a connection from the pool, no need to clean up [e2e-llm-inference-service] clean_exit = True [e2e-llm-inference-service] release_this_conn = False [e2e-llm-inference-service] raise [e2e-llm-inference-service] [e2e-llm-inference-service] except ( [e2e-llm-inference-service] TimeoutError, [e2e-llm-inference-service] HTTPException, [e2e-llm-inference-service] OSError, [e2e-llm-inference-service] ProtocolError, [e2e-llm-inference-service] BaseSSLError, [e2e-llm-inference-service] SSLError, [e2e-llm-inference-service] CertificateError, [e2e-llm-inference-service] ProxyError, [e2e-llm-inference-service] ) as e: [e2e-llm-inference-service] # Discard the connection for these exceptions. It will be [e2e-llm-inference-service] # replaced during the next _get_conn() call. [e2e-llm-inference-service] clean_exit = False [e2e-llm-inference-service] new_e: Exception = e [e2e-llm-inference-service] if isinstance(e, (BaseSSLError, CertificateError)): [e2e-llm-inference-service] new_e = SSLError(e) [e2e-llm-inference-service] if isinstance( [e2e-llm-inference-service] new_e, [e2e-llm-inference-service] ( [e2e-llm-inference-service] OSError, [e2e-llm-inference-service] NewConnectionError, [e2e-llm-inference-service] TimeoutError, [e2e-llm-inference-service] SSLError, [e2e-llm-inference-service] HTTPException, [e2e-llm-inference-service] ), [e2e-llm-inference-service] ) and (conn and conn.proxy and not conn.has_connected_to_proxy): [e2e-llm-inference-service] new_e = _wrap_proxy_error(new_e, conn.proxy.scheme) [e2e-llm-inference-service] elif isinstance(new_e, (OSError, HTTPException)): [e2e-llm-inference-service] new_e = ProtocolError("Connection aborted.", new_e) [e2e-llm-inference-service] [e2e-llm-inference-service] retries = retries.increment( [e2e-llm-inference-service] method, url, error=new_e, _pool=self, _stacktrace=sys.exc_info()[2] [e2e-llm-inference-service] ) [e2e-llm-inference-service] retries.sleep() [e2e-llm-inference-service] [e2e-llm-inference-service] # Keep track of the error for the retry warning. [e2e-llm-inference-service] err = e [e2e-llm-inference-service] [e2e-llm-inference-service] finally: [e2e-llm-inference-service] if not clean_exit: [e2e-llm-inference-service] # We hit some kind of exception, handled or otherwise. We need [e2e-llm-inference-service] # to throw the connection away unless explicitly told not to. [e2e-llm-inference-service] # Close the connection, set the variable to None, and make sure [e2e-llm-inference-service] # we put the None back in the pool to avoid leaking it. [e2e-llm-inference-service] if conn: [e2e-llm-inference-service] conn.close() [e2e-llm-inference-service] conn = None [e2e-llm-inference-service] release_this_conn = True [e2e-llm-inference-service] [e2e-llm-inference-service] if release_this_conn: [e2e-llm-inference-service] # Put the connection back to be reused. If the connection is [e2e-llm-inference-service] # expired then it will be None, which will get replaced with a [e2e-llm-inference-service] # fresh connection during _get_conn. [e2e-llm-inference-service] self._put_conn(conn) [e2e-llm-inference-service] [e2e-llm-inference-service] if not conn: [e2e-llm-inference-service] # Try again [e2e-llm-inference-service] log.warning( [e2e-llm-inference-service] "Retrying (%r) after connection broken by '%r': %s", retries, err, url [e2e-llm-inference-service] ) [e2e-llm-inference-service] > return self.urlopen( [e2e-llm-inference-service] method, [e2e-llm-inference-service] url, [e2e-llm-inference-service] body, [e2e-llm-inference-service] headers, [e2e-llm-inference-service] retries, [e2e-llm-inference-service] redirect, [e2e-llm-inference-service] assert_same_host, [e2e-llm-inference-service] timeout=timeout, [e2e-llm-inference-service] pool_timeout=pool_timeout, [e2e-llm-inference-service] release_conn=release_conn, [e2e-llm-inference-service] chunked=chunked, [e2e-llm-inference-service] body_pos=body_pos, [e2e-llm-inference-service] preload_content=preload_content, [e2e-llm-inference-service] decode_content=decode_content, [e2e-llm-inference-service] **response_kw, [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] ../../python/kserve/.venv/lib64/python3.11/site-packages/urllib3/connectionpool.py:871: [e2e-llm-inference-service] _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ [e2e-llm-inference-service] [e2e-llm-inference-service] self = [e2e-llm-inference-service] method = 'POST' [e2e-llm-inference-service] url = '/kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-route-87882a8e/v1/completions' [e2e-llm-inference-service] body = b'{"model": "facebook/opt-125m", "prompt": "KServe is a", "max_tokens": 20}' [e2e-llm-inference-service] headers = {'User-Agent': 'python-requests/2.32.3', 'Accept-Encoding': 'gzip, deflate', 'Accept': '*/*', 'Connection': 'keep-alive', 'Content-Type': 'application/json', 'Content-Length': '73'} [e2e-llm-inference-service] retries = Retry(total=0, connect=None, read=None, redirect=None, status=None) [e2e-llm-inference-service] redirect = False, assert_same_host = False [e2e-llm-inference-service] timeout = Timeout(connect=60, read=60, total=None), pool_timeout = None [e2e-llm-inference-service] release_conn = False, chunked = False, body_pos = None, preload_content = False [e2e-llm-inference-service] decode_content = False, response_kw = {} [e2e-llm-inference-service] parsed_url = Url(scheme=None, auth=None, host=None, port=None, path='/kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-route-87882a8e/v1/completions', query=None, fragment=None) [e2e-llm-inference-service] destination_scheme = None, conn = None, release_this_conn = True [e2e-llm-inference-service] http_tunnel_required = False [e2e-llm-inference-service] err = ReadTimeoutError("HTTPConnectionPool(host='ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com', port=80): Read timed out. (read timeout=60)") [e2e-llm-inference-service] clean_exit = False [e2e-llm-inference-service] [e2e-llm-inference-service] def urlopen( # type: ignore[override] [e2e-llm-inference-service] self, [e2e-llm-inference-service] method: str, [e2e-llm-inference-service] url: str, [e2e-llm-inference-service] body: _TYPE_BODY | None = None, [e2e-llm-inference-service] headers: typing.Mapping[str, str] | None = None, [e2e-llm-inference-service] retries: Retry | bool | int | None = None, [e2e-llm-inference-service] redirect: bool = True, [e2e-llm-inference-service] assert_same_host: bool = True, [e2e-llm-inference-service] timeout: _TYPE_TIMEOUT = _DEFAULT_TIMEOUT, [e2e-llm-inference-service] pool_timeout: int | None = None, [e2e-llm-inference-service] release_conn: bool | None = None, [e2e-llm-inference-service] chunked: bool = False, [e2e-llm-inference-service] body_pos: _TYPE_BODY_POSITION | None = None, [e2e-llm-inference-service] preload_content: bool = True, [e2e-llm-inference-service] decode_content: bool = True, [e2e-llm-inference-service] **response_kw: typing.Any, [e2e-llm-inference-service] ) -> BaseHTTPResponse: [e2e-llm-inference-service] """ [e2e-llm-inference-service] Get a connection from the pool and perform an HTTP request. This is the [e2e-llm-inference-service] lowest level call for making a request, so you'll need to specify all [e2e-llm-inference-service] the raw details. [e2e-llm-inference-service] [e2e-llm-inference-service] .. note:: [e2e-llm-inference-service] [e2e-llm-inference-service] More commonly, it's appropriate to use a convenience method [e2e-llm-inference-service] such as :meth:`request`. [e2e-llm-inference-service] [e2e-llm-inference-service] .. note:: [e2e-llm-inference-service] [e2e-llm-inference-service] `release_conn` will only behave as expected if [e2e-llm-inference-service] `preload_content=False` because we want to make [e2e-llm-inference-service] `preload_content=False` the default behaviour someday soon without [e2e-llm-inference-service] breaking backwards compatibility. [e2e-llm-inference-service] [e2e-llm-inference-service] :param method: [e2e-llm-inference-service] HTTP request method (such as GET, POST, PUT, etc.) [e2e-llm-inference-service] [e2e-llm-inference-service] :param url: [e2e-llm-inference-service] The URL to perform the request on. [e2e-llm-inference-service] [e2e-llm-inference-service] :param body: [e2e-llm-inference-service] Data to send in the request body, either :class:`str`, :class:`bytes`, [e2e-llm-inference-service] an iterable of :class:`str`/:class:`bytes`, or a file-like object. [e2e-llm-inference-service] [e2e-llm-inference-service] :param headers: [e2e-llm-inference-service] Dictionary of custom headers to send, such as User-Agent, [e2e-llm-inference-service] If-None-Match, etc. If None, pool headers are used. If provided, [e2e-llm-inference-service] these headers completely replace any pool-specific headers. [e2e-llm-inference-service] [e2e-llm-inference-service] :param retries: [e2e-llm-inference-service] Configure the number of retries to allow before raising a [e2e-llm-inference-service] :class:`~urllib3.exceptions.MaxRetryError` exception. [e2e-llm-inference-service] [e2e-llm-inference-service] If ``None`` (default) will retry 3 times, see ``Retry.DEFAULT``. Pass a [e2e-llm-inference-service] :class:`~urllib3.util.retry.Retry` object for fine-grained control [e2e-llm-inference-service] over different types of retries. [e2e-llm-inference-service] Pass an integer number to retry connection errors that many times, [e2e-llm-inference-service] but no other types of errors. Pass zero to never retry. [e2e-llm-inference-service] [e2e-llm-inference-service] If ``False``, then retries are disabled and any exception is raised [e2e-llm-inference-service] immediately. Also, instead of raising a MaxRetryError on redirects, [e2e-llm-inference-service] the redirect response will be returned. [e2e-llm-inference-service] [e2e-llm-inference-service] :type retries: :class:`~urllib3.util.retry.Retry`, False, or an int. [e2e-llm-inference-service] [e2e-llm-inference-service] :param redirect: [e2e-llm-inference-service] If True, automatically handle redirects (status codes 301, 302, [e2e-llm-inference-service] 303, 307, 308). Each redirect counts as a retry. Disabling retries [e2e-llm-inference-service] will disable redirect, too. [e2e-llm-inference-service] [e2e-llm-inference-service] :param assert_same_host: [e2e-llm-inference-service] If ``True``, will make sure that the host of the pool requests is [e2e-llm-inference-service] consistent else will raise HostChangedError. When ``False``, you can [e2e-llm-inference-service] use the pool on an HTTP proxy and request foreign hosts. [e2e-llm-inference-service] [e2e-llm-inference-service] :param timeout: [e2e-llm-inference-service] If specified, overrides the default timeout for this one [e2e-llm-inference-service] request. It may be a float (in seconds) or an instance of [e2e-llm-inference-service] :class:`urllib3.util.Timeout`. [e2e-llm-inference-service] [e2e-llm-inference-service] :param pool_timeout: [e2e-llm-inference-service] If set and the pool is set to block=True, then this method will [e2e-llm-inference-service] block for ``pool_timeout`` seconds and raise EmptyPoolError if no [e2e-llm-inference-service] connection is available within the time period. [e2e-llm-inference-service] [e2e-llm-inference-service] :param bool preload_content: [e2e-llm-inference-service] If True, the response's body will be preloaded into memory. [e2e-llm-inference-service] [e2e-llm-inference-service] :param bool decode_content: [e2e-llm-inference-service] If True, will attempt to decode the body based on the [e2e-llm-inference-service] 'content-encoding' header. [e2e-llm-inference-service] [e2e-llm-inference-service] :param release_conn: [e2e-llm-inference-service] If False, then the urlopen call will not release the connection [e2e-llm-inference-service] back into the pool once a response is received (but will release if [e2e-llm-inference-service] you read the entire contents of the response such as when [e2e-llm-inference-service] `preload_content=True`). This is useful if you're not preloading [e2e-llm-inference-service] the response's content immediately. You will need to call [e2e-llm-inference-service] ``r.release_conn()`` on the response ``r`` to return the connection [e2e-llm-inference-service] back into the pool. If None, it takes the value of ``preload_content`` [e2e-llm-inference-service] which defaults to ``True``. [e2e-llm-inference-service] [e2e-llm-inference-service] :param bool chunked: [e2e-llm-inference-service] If True, urllib3 will send the body using chunked transfer [e2e-llm-inference-service] encoding. Otherwise, urllib3 will send the body using the standard [e2e-llm-inference-service] content-length form. Defaults to False. [e2e-llm-inference-service] [e2e-llm-inference-service] :param int body_pos: [e2e-llm-inference-service] Position to seek to in file-like body in the event of a retry or [e2e-llm-inference-service] redirect. Typically this won't need to be set because urllib3 will [e2e-llm-inference-service] auto-populate the value when needed. [e2e-llm-inference-service] """ [e2e-llm-inference-service] parsed_url = parse_url(url) [e2e-llm-inference-service] destination_scheme = parsed_url.scheme [e2e-llm-inference-service] [e2e-llm-inference-service] if headers is None: [e2e-llm-inference-service] headers = self.headers [e2e-llm-inference-service] [e2e-llm-inference-service] if not isinstance(retries, Retry): [e2e-llm-inference-service] retries = Retry.from_int(retries, redirect=redirect, default=self.retries) [e2e-llm-inference-service] [e2e-llm-inference-service] if release_conn is None: [e2e-llm-inference-service] release_conn = preload_content [e2e-llm-inference-service] [e2e-llm-inference-service] # Check host [e2e-llm-inference-service] if assert_same_host and not self.is_same_host(url): [e2e-llm-inference-service] raise HostChangedError(self, url, retries) [e2e-llm-inference-service] [e2e-llm-inference-service] # Ensure that the URL we're connecting to is properly encoded [e2e-llm-inference-service] if url.startswith("/"): [e2e-llm-inference-service] url = to_str(_encode_target(url)) [e2e-llm-inference-service] else: [e2e-llm-inference-service] url = to_str(parsed_url.url) [e2e-llm-inference-service] [e2e-llm-inference-service] conn = None [e2e-llm-inference-service] [e2e-llm-inference-service] # Track whether `conn` needs to be released before [e2e-llm-inference-service] # returning/raising/recursing. Update this variable if necessary, and [e2e-llm-inference-service] # leave `release_conn` constant throughout the function. That way, if [e2e-llm-inference-service] # the function recurses, the original value of `release_conn` will be [e2e-llm-inference-service] # passed down into the recursive call, and its value will be respected. [e2e-llm-inference-service] # [e2e-llm-inference-service] # See issue #651 [1] for details. [e2e-llm-inference-service] # [e2e-llm-inference-service] # [1] [e2e-llm-inference-service] release_this_conn = release_conn [e2e-llm-inference-service] [e2e-llm-inference-service] http_tunnel_required = connection_requires_http_tunnel( [e2e-llm-inference-service] self.proxy, self.proxy_config, destination_scheme [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] # Merge the proxy headers. Only done when not using HTTP CONNECT. We [e2e-llm-inference-service] # have to copy the headers dict so we can safely change it without those [e2e-llm-inference-service] # changes being reflected in anyone else's copy. [e2e-llm-inference-service] if not http_tunnel_required: [e2e-llm-inference-service] headers = headers.copy() # type: ignore[attr-defined] [e2e-llm-inference-service] headers.update(self.proxy_headers) # type: ignore[union-attr] [e2e-llm-inference-service] [e2e-llm-inference-service] # Must keep the exception bound to a separate variable or else Python 3 [e2e-llm-inference-service] # complains about UnboundLocalError. [e2e-llm-inference-service] err = None [e2e-llm-inference-service] [e2e-llm-inference-service] # Keep track of whether we cleanly exited the except block. This [e2e-llm-inference-service] # ensures we do proper cleanup in finally. [e2e-llm-inference-service] clean_exit = False [e2e-llm-inference-service] [e2e-llm-inference-service] # Rewind body position, if needed. Record current position [e2e-llm-inference-service] # for future rewinds in the event of a redirect/retry. [e2e-llm-inference-service] body_pos = set_file_position(body, body_pos) [e2e-llm-inference-service] [e2e-llm-inference-service] try: [e2e-llm-inference-service] # Request a connection from the queue. [e2e-llm-inference-service] timeout_obj = self._get_timeout(timeout) [e2e-llm-inference-service] conn = self._get_conn(timeout=pool_timeout) [e2e-llm-inference-service] [e2e-llm-inference-service] conn.timeout = timeout_obj.connect_timeout # type: ignore[assignment] [e2e-llm-inference-service] [e2e-llm-inference-service] # Is this a closed/new connection that requires CONNECT tunnelling? [e2e-llm-inference-service] if self.proxy is not None and http_tunnel_required and conn.is_closed: [e2e-llm-inference-service] try: [e2e-llm-inference-service] self._prepare_proxy(conn) [e2e-llm-inference-service] except (BaseSSLError, OSError, SocketTimeout) as e: [e2e-llm-inference-service] self._raise_timeout( [e2e-llm-inference-service] err=e, url=self.proxy.url, timeout_value=conn.timeout [e2e-llm-inference-service] ) [e2e-llm-inference-service] raise [e2e-llm-inference-service] [e2e-llm-inference-service] # If we're going to release the connection in ``finally:``, then [e2e-llm-inference-service] # the response doesn't need to know about the connection. Otherwise [e2e-llm-inference-service] # it will also try to release it and we'll have a double-release [e2e-llm-inference-service] # mess. [e2e-llm-inference-service] response_conn = conn if not release_conn else None [e2e-llm-inference-service] [e2e-llm-inference-service] # Make the request on the HTTPConnection object [e2e-llm-inference-service] response = self._make_request( [e2e-llm-inference-service] conn, [e2e-llm-inference-service] method, [e2e-llm-inference-service] url, [e2e-llm-inference-service] timeout=timeout_obj, [e2e-llm-inference-service] body=body, [e2e-llm-inference-service] headers=headers, [e2e-llm-inference-service] chunked=chunked, [e2e-llm-inference-service] retries=retries, [e2e-llm-inference-service] response_conn=response_conn, [e2e-llm-inference-service] preload_content=preload_content, [e2e-llm-inference-service] decode_content=decode_content, [e2e-llm-inference-service] **response_kw, [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] # Everything went great! [e2e-llm-inference-service] clean_exit = True [e2e-llm-inference-service] [e2e-llm-inference-service] except EmptyPoolError: [e2e-llm-inference-service] # Didn't get a connection from the pool, no need to clean up [e2e-llm-inference-service] clean_exit = True [e2e-llm-inference-service] release_this_conn = False [e2e-llm-inference-service] raise [e2e-llm-inference-service] [e2e-llm-inference-service] except ( [e2e-llm-inference-service] TimeoutError, [e2e-llm-inference-service] HTTPException, [e2e-llm-inference-service] OSError, [e2e-llm-inference-service] ProtocolError, [e2e-llm-inference-service] BaseSSLError, [e2e-llm-inference-service] SSLError, [e2e-llm-inference-service] CertificateError, [e2e-llm-inference-service] ProxyError, [e2e-llm-inference-service] ) as e: [e2e-llm-inference-service] # Discard the connection for these exceptions. It will be [e2e-llm-inference-service] # replaced during the next _get_conn() call. [e2e-llm-inference-service] clean_exit = False [e2e-llm-inference-service] new_e: Exception = e [e2e-llm-inference-service] if isinstance(e, (BaseSSLError, CertificateError)): [e2e-llm-inference-service] new_e = SSLError(e) [e2e-llm-inference-service] if isinstance( [e2e-llm-inference-service] new_e, [e2e-llm-inference-service] ( [e2e-llm-inference-service] OSError, [e2e-llm-inference-service] NewConnectionError, [e2e-llm-inference-service] TimeoutError, [e2e-llm-inference-service] SSLError, [e2e-llm-inference-service] HTTPException, [e2e-llm-inference-service] ), [e2e-llm-inference-service] ) and (conn and conn.proxy and not conn.has_connected_to_proxy): [e2e-llm-inference-service] new_e = _wrap_proxy_error(new_e, conn.proxy.scheme) [e2e-llm-inference-service] elif isinstance(new_e, (OSError, HTTPException)): [e2e-llm-inference-service] new_e = ProtocolError("Connection aborted.", new_e) [e2e-llm-inference-service] [e2e-llm-inference-service] retries = retries.increment( [e2e-llm-inference-service] method, url, error=new_e, _pool=self, _stacktrace=sys.exc_info()[2] [e2e-llm-inference-service] ) [e2e-llm-inference-service] retries.sleep() [e2e-llm-inference-service] [e2e-llm-inference-service] # Keep track of the error for the retry warning. [e2e-llm-inference-service] err = e [e2e-llm-inference-service] [e2e-llm-inference-service] finally: [e2e-llm-inference-service] if not clean_exit: [e2e-llm-inference-service] # We hit some kind of exception, handled or otherwise. We need [e2e-llm-inference-service] # to throw the connection away unless explicitly told not to. [e2e-llm-inference-service] # Close the connection, set the variable to None, and make sure [e2e-llm-inference-service] # we put the None back in the pool to avoid leaking it. [e2e-llm-inference-service] if conn: [e2e-llm-inference-service] conn.close() [e2e-llm-inference-service] conn = None [e2e-llm-inference-service] release_this_conn = True [e2e-llm-inference-service] [e2e-llm-inference-service] if release_this_conn: [e2e-llm-inference-service] # Put the connection back to be reused. If the connection is [e2e-llm-inference-service] # expired then it will be None, which will get replaced with a [e2e-llm-inference-service] # fresh connection during _get_conn. [e2e-llm-inference-service] self._put_conn(conn) [e2e-llm-inference-service] [e2e-llm-inference-service] if not conn: [e2e-llm-inference-service] # Try again [e2e-llm-inference-service] log.warning( [e2e-llm-inference-service] "Retrying (%r) after connection broken by '%r': %s", retries, err, url [e2e-llm-inference-service] ) [e2e-llm-inference-service] > return self.urlopen( [e2e-llm-inference-service] method, [e2e-llm-inference-service] url, [e2e-llm-inference-service] body, [e2e-llm-inference-service] headers, [e2e-llm-inference-service] retries, [e2e-llm-inference-service] redirect, [e2e-llm-inference-service] assert_same_host, [e2e-llm-inference-service] timeout=timeout, [e2e-llm-inference-service] pool_timeout=pool_timeout, [e2e-llm-inference-service] release_conn=release_conn, [e2e-llm-inference-service] chunked=chunked, [e2e-llm-inference-service] body_pos=body_pos, [e2e-llm-inference-service] preload_content=preload_content, [e2e-llm-inference-service] decode_content=decode_content, [e2e-llm-inference-service] **response_kw, [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] ../../python/kserve/.venv/lib64/python3.11/site-packages/urllib3/connectionpool.py:871: [e2e-llm-inference-service] _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ [e2e-llm-inference-service] [e2e-llm-inference-service] self = [e2e-llm-inference-service] method = 'POST' [e2e-llm-inference-service] url = '/kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-route-87882a8e/v1/completions' [e2e-llm-inference-service] body = b'{"model": "facebook/opt-125m", "prompt": "KServe is a", "max_tokens": 20}' [e2e-llm-inference-service] headers = {'User-Agent': 'python-requests/2.32.3', 'Accept-Encoding': 'gzip, deflate', 'Accept': '*/*', 'Connection': 'keep-alive', 'Content-Type': 'application/json', 'Content-Length': '73'} [e2e-llm-inference-service] retries = Retry(total=0, connect=None, read=None, redirect=None, status=None) [e2e-llm-inference-service] redirect = False, assert_same_host = False [e2e-llm-inference-service] timeout = Timeout(connect=60, read=60, total=None), pool_timeout = None [e2e-llm-inference-service] release_conn = False, chunked = False, body_pos = None, preload_content = False [e2e-llm-inference-service] decode_content = False, response_kw = {} [e2e-llm-inference-service] parsed_url = Url(scheme=None, auth=None, host=None, port=None, path='/kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-route-87882a8e/v1/completions', query=None, fragment=None) [e2e-llm-inference-service] destination_scheme = None, conn = None, release_this_conn = True [e2e-llm-inference-service] http_tunnel_required = False, err = None, clean_exit = False [e2e-llm-inference-service] [e2e-llm-inference-service] def urlopen( # type: ignore[override] [e2e-llm-inference-service] self, [e2e-llm-inference-service] method: str, [e2e-llm-inference-service] url: str, [e2e-llm-inference-service] body: _TYPE_BODY | None = None, [e2e-llm-inference-service] headers: typing.Mapping[str, str] | None = None, [e2e-llm-inference-service] retries: Retry | bool | int | None = None, [e2e-llm-inference-service] redirect: bool = True, [e2e-llm-inference-service] assert_same_host: bool = True, [e2e-llm-inference-service] timeout: _TYPE_TIMEOUT = _DEFAULT_TIMEOUT, [e2e-llm-inference-service] pool_timeout: int | None = None, [e2e-llm-inference-service] release_conn: bool | None = None, [e2e-llm-inference-service] chunked: bool = False, [e2e-llm-inference-service] body_pos: _TYPE_BODY_POSITION | None = None, [e2e-llm-inference-service] preload_content: bool = True, [e2e-llm-inference-service] decode_content: bool = True, [e2e-llm-inference-service] **response_kw: typing.Any, [e2e-llm-inference-service] ) -> BaseHTTPResponse: [e2e-llm-inference-service] """ [e2e-llm-inference-service] Get a connection from the pool and perform an HTTP request. This is the [e2e-llm-inference-service] lowest level call for making a request, so you'll need to specify all [e2e-llm-inference-service] the raw details. [e2e-llm-inference-service] [e2e-llm-inference-service] .. note:: [e2e-llm-inference-service] [e2e-llm-inference-service] More commonly, it's appropriate to use a convenience method [e2e-llm-inference-service] such as :meth:`request`. [e2e-llm-inference-service] [e2e-llm-inference-service] .. note:: [e2e-llm-inference-service] [e2e-llm-inference-service] `release_conn` will only behave as expected if [e2e-llm-inference-service] `preload_content=False` because we want to make [e2e-llm-inference-service] `preload_content=False` the default behaviour someday soon without [e2e-llm-inference-service] breaking backwards compatibility. [e2e-llm-inference-service] [e2e-llm-inference-service] :param method: [e2e-llm-inference-service] HTTP request method (such as GET, POST, PUT, etc.) [e2e-llm-inference-service] [e2e-llm-inference-service] :param url: [e2e-llm-inference-service] The URL to perform the request on. [e2e-llm-inference-service] [e2e-llm-inference-service] :param body: [e2e-llm-inference-service] Data to send in the request body, either :class:`str`, :class:`bytes`, [e2e-llm-inference-service] an iterable of :class:`str`/:class:`bytes`, or a file-like object. [e2e-llm-inference-service] [e2e-llm-inference-service] :param headers: [e2e-llm-inference-service] Dictionary of custom headers to send, such as User-Agent, [e2e-llm-inference-service] If-None-Match, etc. If None, pool headers are used. If provided, [e2e-llm-inference-service] these headers completely replace any pool-specific headers. [e2e-llm-inference-service] [e2e-llm-inference-service] :param retries: [e2e-llm-inference-service] Configure the number of retries to allow before raising a [e2e-llm-inference-service] :class:`~urllib3.exceptions.MaxRetryError` exception. [e2e-llm-inference-service] [e2e-llm-inference-service] If ``None`` (default) will retry 3 times, see ``Retry.DEFAULT``. Pass a [e2e-llm-inference-service] :class:`~urllib3.util.retry.Retry` object for fine-grained control [e2e-llm-inference-service] over different types of retries. [e2e-llm-inference-service] Pass an integer number to retry connection errors that many times, [e2e-llm-inference-service] but no other types of errors. Pass zero to never retry. [e2e-llm-inference-service] [e2e-llm-inference-service] If ``False``, then retries are disabled and any exception is raised [e2e-llm-inference-service] immediately. Also, instead of raising a MaxRetryError on redirects, [e2e-llm-inference-service] the redirect response will be returned. [e2e-llm-inference-service] [e2e-llm-inference-service] :type retries: :class:`~urllib3.util.retry.Retry`, False, or an int. [e2e-llm-inference-service] [e2e-llm-inference-service] :param redirect: [e2e-llm-inference-service] If True, automatically handle redirects (status codes 301, 302, [e2e-llm-inference-service] 303, 307, 308). Each redirect counts as a retry. Disabling retries [e2e-llm-inference-service] will disable redirect, too. [e2e-llm-inference-service] [e2e-llm-inference-service] :param assert_same_host: [e2e-llm-inference-service] If ``True``, will make sure that the host of the pool requests is [e2e-llm-inference-service] consistent else will raise HostChangedError. When ``False``, you can [e2e-llm-inference-service] use the pool on an HTTP proxy and request foreign hosts. [e2e-llm-inference-service] [e2e-llm-inference-service] :param timeout: [e2e-llm-inference-service] If specified, overrides the default timeout for this one [e2e-llm-inference-service] request. It may be a float (in seconds) or an instance of [e2e-llm-inference-service] :class:`urllib3.util.Timeout`. [e2e-llm-inference-service] [e2e-llm-inference-service] :param pool_timeout: [e2e-llm-inference-service] If set and the pool is set to block=True, then this method will [e2e-llm-inference-service] block for ``pool_timeout`` seconds and raise EmptyPoolError if no [e2e-llm-inference-service] connection is available within the time period. [e2e-llm-inference-service] [e2e-llm-inference-service] :param bool preload_content: [e2e-llm-inference-service] If True, the response's body will be preloaded into memory. [e2e-llm-inference-service] [e2e-llm-inference-service] :param bool decode_content: [e2e-llm-inference-service] If True, will attempt to decode the body based on the [e2e-llm-inference-service] 'content-encoding' header. [e2e-llm-inference-service] [e2e-llm-inference-service] :param release_conn: [e2e-llm-inference-service] If False, then the urlopen call will not release the connection [e2e-llm-inference-service] back into the pool once a response is received (but will release if [e2e-llm-inference-service] you read the entire contents of the response such as when [e2e-llm-inference-service] `preload_content=True`). This is useful if you're not preloading [e2e-llm-inference-service] the response's content immediately. You will need to call [e2e-llm-inference-service] ``r.release_conn()`` on the response ``r`` to return the connection [e2e-llm-inference-service] back into the pool. If None, it takes the value of ``preload_content`` [e2e-llm-inference-service] which defaults to ``True``. [e2e-llm-inference-service] [e2e-llm-inference-service] :param bool chunked: [e2e-llm-inference-service] If True, urllib3 will send the body using chunked transfer [e2e-llm-inference-service] encoding. Otherwise, urllib3 will send the body using the standard [e2e-llm-inference-service] content-length form. Defaults to False. [e2e-llm-inference-service] [e2e-llm-inference-service] :param int body_pos: [e2e-llm-inference-service] Position to seek to in file-like body in the event of a retry or [e2e-llm-inference-service] redirect. Typically this won't need to be set because urllib3 will [e2e-llm-inference-service] auto-populate the value when needed. [e2e-llm-inference-service] """ [e2e-llm-inference-service] parsed_url = parse_url(url) [e2e-llm-inference-service] destination_scheme = parsed_url.scheme [e2e-llm-inference-service] [e2e-llm-inference-service] if headers is None: [e2e-llm-inference-service] headers = self.headers [e2e-llm-inference-service] [e2e-llm-inference-service] if not isinstance(retries, Retry): [e2e-llm-inference-service] retries = Retry.from_int(retries, redirect=redirect, default=self.retries) [e2e-llm-inference-service] [e2e-llm-inference-service] if release_conn is None: [e2e-llm-inference-service] release_conn = preload_content [e2e-llm-inference-service] [e2e-llm-inference-service] # Check host [e2e-llm-inference-service] if assert_same_host and not self.is_same_host(url): [e2e-llm-inference-service] raise HostChangedError(self, url, retries) [e2e-llm-inference-service] [e2e-llm-inference-service] # Ensure that the URL we're connecting to is properly encoded [e2e-llm-inference-service] if url.startswith("/"): [e2e-llm-inference-service] url = to_str(_encode_target(url)) [e2e-llm-inference-service] else: [e2e-llm-inference-service] url = to_str(parsed_url.url) [e2e-llm-inference-service] [e2e-llm-inference-service] conn = None [e2e-llm-inference-service] [e2e-llm-inference-service] # Track whether `conn` needs to be released before [e2e-llm-inference-service] # returning/raising/recursing. Update this variable if necessary, and [e2e-llm-inference-service] # leave `release_conn` constant throughout the function. That way, if [e2e-llm-inference-service] # the function recurses, the original value of `release_conn` will be [e2e-llm-inference-service] # passed down into the recursive call, and its value will be respected. [e2e-llm-inference-service] # [e2e-llm-inference-service] # See issue #651 [1] for details. [e2e-llm-inference-service] # [e2e-llm-inference-service] # [1] [e2e-llm-inference-service] release_this_conn = release_conn [e2e-llm-inference-service] [e2e-llm-inference-service] http_tunnel_required = connection_requires_http_tunnel( [e2e-llm-inference-service] self.proxy, self.proxy_config, destination_scheme [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] # Merge the proxy headers. Only done when not using HTTP CONNECT. We [e2e-llm-inference-service] # have to copy the headers dict so we can safely change it without those [e2e-llm-inference-service] # changes being reflected in anyone else's copy. [e2e-llm-inference-service] if not http_tunnel_required: [e2e-llm-inference-service] headers = headers.copy() # type: ignore[attr-defined] [e2e-llm-inference-service] headers.update(self.proxy_headers) # type: ignore[union-attr] [e2e-llm-inference-service] [e2e-llm-inference-service] # Must keep the exception bound to a separate variable or else Python 3 [e2e-llm-inference-service] # complains about UnboundLocalError. [e2e-llm-inference-service] err = None [e2e-llm-inference-service] [e2e-llm-inference-service] # Keep track of whether we cleanly exited the except block. This [e2e-llm-inference-service] # ensures we do proper cleanup in finally. [e2e-llm-inference-service] clean_exit = False [e2e-llm-inference-service] [e2e-llm-inference-service] # Rewind body position, if needed. Record current position [e2e-llm-inference-service] # for future rewinds in the event of a redirect/retry. [e2e-llm-inference-service] body_pos = set_file_position(body, body_pos) [e2e-llm-inference-service] [e2e-llm-inference-service] try: [e2e-llm-inference-service] # Request a connection from the queue. [e2e-llm-inference-service] timeout_obj = self._get_timeout(timeout) [e2e-llm-inference-service] conn = self._get_conn(timeout=pool_timeout) [e2e-llm-inference-service] [e2e-llm-inference-service] conn.timeout = timeout_obj.connect_timeout # type: ignore[assignment] [e2e-llm-inference-service] [e2e-llm-inference-service] # Is this a closed/new connection that requires CONNECT tunnelling? [e2e-llm-inference-service] if self.proxy is not None and http_tunnel_required and conn.is_closed: [e2e-llm-inference-service] try: [e2e-llm-inference-service] self._prepare_proxy(conn) [e2e-llm-inference-service] except (BaseSSLError, OSError, SocketTimeout) as e: [e2e-llm-inference-service] self._raise_timeout( [e2e-llm-inference-service] err=e, url=self.proxy.url, timeout_value=conn.timeout [e2e-llm-inference-service] ) [e2e-llm-inference-service] raise [e2e-llm-inference-service] [e2e-llm-inference-service] # If we're going to release the connection in ``finally:``, then [e2e-llm-inference-service] # the response doesn't need to know about the connection. Otherwise [e2e-llm-inference-service] # it will also try to release it and we'll have a double-release [e2e-llm-inference-service] # mess. [e2e-llm-inference-service] response_conn = conn if not release_conn else None [e2e-llm-inference-service] [e2e-llm-inference-service] # Make the request on the HTTPConnection object [e2e-llm-inference-service] response = self._make_request( [e2e-llm-inference-service] conn, [e2e-llm-inference-service] method, [e2e-llm-inference-service] url, [e2e-llm-inference-service] timeout=timeout_obj, [e2e-llm-inference-service] body=body, [e2e-llm-inference-service] headers=headers, [e2e-llm-inference-service] chunked=chunked, [e2e-llm-inference-service] retries=retries, [e2e-llm-inference-service] response_conn=response_conn, [e2e-llm-inference-service] preload_content=preload_content, [e2e-llm-inference-service] decode_content=decode_content, [e2e-llm-inference-service] **response_kw, [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] # Everything went great! [e2e-llm-inference-service] clean_exit = True [e2e-llm-inference-service] [e2e-llm-inference-service] except EmptyPoolError: [e2e-llm-inference-service] # Didn't get a connection from the pool, no need to clean up [e2e-llm-inference-service] clean_exit = True [e2e-llm-inference-service] release_this_conn = False [e2e-llm-inference-service] raise [e2e-llm-inference-service] [e2e-llm-inference-service] except ( [e2e-llm-inference-service] TimeoutError, [e2e-llm-inference-service] HTTPException, [e2e-llm-inference-service] OSError, [e2e-llm-inference-service] ProtocolError, [e2e-llm-inference-service] BaseSSLError, [e2e-llm-inference-service] SSLError, [e2e-llm-inference-service] CertificateError, [e2e-llm-inference-service] ProxyError, [e2e-llm-inference-service] ) as e: [e2e-llm-inference-service] # Discard the connection for these exceptions. It will be [e2e-llm-inference-service] # replaced during the next _get_conn() call. [e2e-llm-inference-service] clean_exit = False [e2e-llm-inference-service] new_e: Exception = e [e2e-llm-inference-service] if isinstance(e, (BaseSSLError, CertificateError)): [e2e-llm-inference-service] new_e = SSLError(e) [e2e-llm-inference-service] if isinstance( [e2e-llm-inference-service] new_e, [e2e-llm-inference-service] ( [e2e-llm-inference-service] OSError, [e2e-llm-inference-service] NewConnectionError, [e2e-llm-inference-service] TimeoutError, [e2e-llm-inference-service] SSLError, [e2e-llm-inference-service] HTTPException, [e2e-llm-inference-service] ), [e2e-llm-inference-service] ) and (conn and conn.proxy and not conn.has_connected_to_proxy): [e2e-llm-inference-service] new_e = _wrap_proxy_error(new_e, conn.proxy.scheme) [e2e-llm-inference-service] elif isinstance(new_e, (OSError, HTTPException)): [e2e-llm-inference-service] new_e = ProtocolError("Connection aborted.", new_e) [e2e-llm-inference-service] [e2e-llm-inference-service] > retries = retries.increment( [e2e-llm-inference-service] method, url, error=new_e, _pool=self, _stacktrace=sys.exc_info()[2] [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] ../../python/kserve/.venv/lib64/python3.11/site-packages/urllib3/connectionpool.py:841: [e2e-llm-inference-service] _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ [e2e-llm-inference-service] [e2e-llm-inference-service] self = Retry(total=0, connect=None, read=None, redirect=None, status=None) [e2e-llm-inference-service] method = 'POST' [e2e-llm-inference-service] url = '/kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-route-87882a8e/v1/completions' [e2e-llm-inference-service] response = None [e2e-llm-inference-service] error = ReadTimeoutError("HTTPConnectionPool(host='ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com', port=80): Read timed out. (read timeout=60)") [e2e-llm-inference-service] _pool = [e2e-llm-inference-service] _stacktrace = [e2e-llm-inference-service] [e2e-llm-inference-service] def increment( [e2e-llm-inference-service] self, [e2e-llm-inference-service] method: str | None = None, [e2e-llm-inference-service] url: str | None = None, [e2e-llm-inference-service] response: BaseHTTPResponse | None = None, [e2e-llm-inference-service] error: Exception | None = None, [e2e-llm-inference-service] _pool: ConnectionPool | None = None, [e2e-llm-inference-service] _stacktrace: TracebackType | None = None, [e2e-llm-inference-service] ) -> Self: [e2e-llm-inference-service] """Return a new Retry object with incremented retry counters. [e2e-llm-inference-service] [e2e-llm-inference-service] :param response: A response object, or None, if the server did not [e2e-llm-inference-service] return a response. [e2e-llm-inference-service] :type response: :class:`~urllib3.response.BaseHTTPResponse` [e2e-llm-inference-service] :param Exception error: An error encountered during the request, or [e2e-llm-inference-service] None if the response was received successfully. [e2e-llm-inference-service] [e2e-llm-inference-service] :return: A new ``Retry`` object. [e2e-llm-inference-service] """ [e2e-llm-inference-service] if self.total is False and error: [e2e-llm-inference-service] # Disabled, indicate to re-raise the error. [e2e-llm-inference-service] raise reraise(type(error), error, _stacktrace) [e2e-llm-inference-service] [e2e-llm-inference-service] total = self.total [e2e-llm-inference-service] if total is not None: [e2e-llm-inference-service] total -= 1 [e2e-llm-inference-service] [e2e-llm-inference-service] connect = self.connect [e2e-llm-inference-service] read = self.read [e2e-llm-inference-service] redirect = self.redirect [e2e-llm-inference-service] status_count = self.status [e2e-llm-inference-service] other = self.other [e2e-llm-inference-service] cause = "unknown" [e2e-llm-inference-service] status = None [e2e-llm-inference-service] redirect_location = None [e2e-llm-inference-service] [e2e-llm-inference-service] if error and self._is_connection_error(error): [e2e-llm-inference-service] # Connect retry? [e2e-llm-inference-service] if connect is False: [e2e-llm-inference-service] raise reraise(type(error), error, _stacktrace) [e2e-llm-inference-service] elif connect is not None: [e2e-llm-inference-service] connect -= 1 [e2e-llm-inference-service] [e2e-llm-inference-service] elif error and self._is_read_error(error): [e2e-llm-inference-service] # Read retry? [e2e-llm-inference-service] if read is False or method is None or not self._is_method_retryable(method): [e2e-llm-inference-service] raise reraise(type(error), error, _stacktrace) [e2e-llm-inference-service] elif read is not None: [e2e-llm-inference-service] read -= 1 [e2e-llm-inference-service] [e2e-llm-inference-service] elif error: [e2e-llm-inference-service] # Other retry? [e2e-llm-inference-service] if other is not None: [e2e-llm-inference-service] other -= 1 [e2e-llm-inference-service] [e2e-llm-inference-service] elif response and response.get_redirect_location(): [e2e-llm-inference-service] # Redirect retry? [e2e-llm-inference-service] if redirect is not None: [e2e-llm-inference-service] redirect -= 1 [e2e-llm-inference-service] cause = "too many redirects" [e2e-llm-inference-service] response_redirect_location = response.get_redirect_location() [e2e-llm-inference-service] if response_redirect_location: [e2e-llm-inference-service] redirect_location = response_redirect_location [e2e-llm-inference-service] status = response.status [e2e-llm-inference-service] [e2e-llm-inference-service] else: [e2e-llm-inference-service] # Incrementing because of a server error like a 500 in [e2e-llm-inference-service] # status_forcelist and the given method is in the allowed_methods [e2e-llm-inference-service] cause = ResponseError.GENERIC_ERROR [e2e-llm-inference-service] if response and response.status: [e2e-llm-inference-service] if status_count is not None: [e2e-llm-inference-service] status_count -= 1 [e2e-llm-inference-service] cause = ResponseError.SPECIFIC_ERROR.format(status_code=response.status) [e2e-llm-inference-service] status = response.status [e2e-llm-inference-service] [e2e-llm-inference-service] history = self.history + ( [e2e-llm-inference-service] RequestHistory(method, url, error, status, redirect_location), [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] new_retry = self.new( [e2e-llm-inference-service] total=total, [e2e-llm-inference-service] connect=connect, [e2e-llm-inference-service] read=read, [e2e-llm-inference-service] redirect=redirect, [e2e-llm-inference-service] status=status_count, [e2e-llm-inference-service] other=other, [e2e-llm-inference-service] history=history, [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] if new_retry.is_exhausted(): [e2e-llm-inference-service] reason = error or ResponseError(cause) [e2e-llm-inference-service] > raise MaxRetryError(_pool, url, reason) from reason # type: ignore[arg-type] [e2e-llm-inference-service] E urllib3.exceptions.MaxRetryError: HTTPConnectionPool(host='ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com', port=80): Max retries exceeded with url: /kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-route-87882a8e/v1/completions (Caused by ReadTimeoutError("HTTPConnectionPool(host='ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com', port=80): Read timed out. (read timeout=60)")) [e2e-llm-inference-service] [e2e-llm-inference-service] ../../python/kserve/.venv/lib64/python3.11/site-packages/urllib3/util/retry.py:519: MaxRetryError [e2e-llm-inference-service] [e2e-llm-inference-service] During handling of the above exception, another exception occurred: [e2e-llm-inference-service] [e2e-llm-inference-service] def get_successful_response(): [e2e-llm-inference-service] try: [e2e-llm-inference-service] if test_case.url_getter: [e2e-llm-inference-service] service_url = test_case.url_getter(kserve_client, test_case.llm_service) [e2e-llm-inference-service] else: [e2e-llm-inference-service] service_url = get_llm_service_url(kserve_client, test_case.llm_service) [e2e-llm-inference-service] except Exception as e: [e2e-llm-inference-service] raise AssertionError(f"❌ Failed to get service URL: {e}") from e [e2e-llm-inference-service] [e2e-llm-inference-service] model_url = service_url + test_case.endpoint [e2e-llm-inference-service] [e2e-llm-inference-service] headers = {"Content-Type": "application/json"} [e2e-llm-inference-service] if extra_headers: [e2e-llm-inference-service] headers.update(extra_headers) [e2e-llm-inference-service] [e2e-llm-inference-service] if test_case.payload_formatter is not None: [e2e-llm-inference-service] test_payload = test_case.payload_formatter(test_case) [e2e-llm-inference-service] elif test_case.prompt is not None: [e2e-llm-inference-service] test_payload = { [e2e-llm-inference-service] "model": test_case.model_name [e2e-llm-inference-service] if not extra_headers or MODEL_ROUTING_HEADER not in extra_headers [e2e-llm-inference-service] else extra_headers[MODEL_ROUTING_HEADER], [e2e-llm-inference-service] "prompt": test_case.prompt, [e2e-llm-inference-service] "max_tokens": test_case.max_tokens, [e2e-llm-inference-service] } [e2e-llm-inference-service] else: [e2e-llm-inference-service] test_payload = None [e2e-llm-inference-service] [e2e-llm-inference-service] logger.info(f"Calling LLM service at {model_url} with payload {test_payload}") [e2e-llm-inference-service] try: [e2e-llm-inference-service] if test_payload is not None: [e2e-llm-inference-service] > response = post_with_retry( [e2e-llm-inference-service] model_url, [e2e-llm-inference-service] headers=headers, [e2e-llm-inference-service] json_data=test_payload, [e2e-llm-inference-service] timeout=test_case.response_timeout, [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] llmisvc/test_llm_inference_service.py:1095: [e2e-llm-inference-service] _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ [e2e-llm-inference-service] [e2e-llm-inference-service] url = 'http://ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com/kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-route-87882a8e/v1/completions' [e2e-llm-inference-service] [e2e-llm-inference-service] def post_with_retry( [e2e-llm-inference-service] url: str, [e2e-llm-inference-service] *, [e2e-llm-inference-service] headers: Dict = None, [e2e-llm-inference-service] json_data: Union[Dict, List] = None, [e2e-llm-inference-service] data: Union[str, bytes] = None, [e2e-llm-inference-service] stream: bool = False, [e2e-llm-inference-service] timeout: float = None, [e2e-llm-inference-service] total_retries: int = DEFAULT_RETRY_TOTAL, [e2e-llm-inference-service] backoff_factor: float = DEFAULT_RETRY_BACKOFF_FACTOR, [e2e-llm-inference-service] retry_status_codes=DEFAULT_RETRY_STATUS_CODES, [e2e-llm-inference-service] ) -> requests.Response: [e2e-llm-inference-service] """ [e2e-llm-inference-service] Send POST request with retries for transient HTTP and network failures. [e2e-llm-inference-service] """ [e2e-llm-inference-service] if json_data is not None and data is not None: [e2e-llm-inference-service] raise ValueError("Only one of json_data or data can be provided.") [e2e-llm-inference-service] [e2e-llm-inference-service] with _retry_session( [e2e-llm-inference-service] ["POST"], total_retries, backoff_factor, retry_status_codes [e2e-llm-inference-service] ) as session: [e2e-llm-inference-service] > return session.post( [e2e-llm-inference-service] url, [e2e-llm-inference-service] json=json_data, [e2e-llm-inference-service] data=data, [e2e-llm-inference-service] headers=headers, [e2e-llm-inference-service] stream=stream, [e2e-llm-inference-service] timeout=timeout, [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] common/http_retry.py:70: [e2e-llm-inference-service] _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ [e2e-llm-inference-service] [e2e-llm-inference-service] self = [e2e-llm-inference-service] url = 'http://ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com/kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-route-87882a8e/v1/completions' [e2e-llm-inference-service] data = None [e2e-llm-inference-service] json = {'max_tokens': 20, 'model': 'facebook/opt-125m', 'prompt': 'KServe is a'} [e2e-llm-inference-service] kwargs = {'headers': {'Content-Type': 'application/json'}, 'stream': False, 'timeout': 60} [e2e-llm-inference-service] [e2e-llm-inference-service] def post(self, url, data=None, json=None, **kwargs): [e2e-llm-inference-service] r"""Sends a POST request. Returns :class:`Response` object. [e2e-llm-inference-service] [e2e-llm-inference-service] :param url: URL for the new :class:`Request` object. [e2e-llm-inference-service] :param data: (optional) Dictionary, list of tuples, bytes, or file-like [e2e-llm-inference-service] object to send in the body of the :class:`Request`. [e2e-llm-inference-service] :param json: (optional) json to send in the body of the :class:`Request`. [e2e-llm-inference-service] :param \*\*kwargs: Optional arguments that ``request`` takes. [e2e-llm-inference-service] :rtype: requests.Response [e2e-llm-inference-service] """ [e2e-llm-inference-service] [e2e-llm-inference-service] > return self.request("POST", url, data=data, json=json, **kwargs) [e2e-llm-inference-service] [e2e-llm-inference-service] ../../python/kserve/.venv/lib64/python3.11/site-packages/requests/sessions.py:637: [e2e-llm-inference-service] _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ [e2e-llm-inference-service] [e2e-llm-inference-service] self = , method = 'POST' [e2e-llm-inference-service] url = 'http://ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com/kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-route-87882a8e/v1/completions' [e2e-llm-inference-service] params = None, data = None, headers = {'Content-Type': 'application/json'} [e2e-llm-inference-service] cookies = None, files = None, auth = None, timeout = 60, allow_redirects = True [e2e-llm-inference-service] proxies = {}, hooks = None, stream = False, verify = None, cert = None [e2e-llm-inference-service] json = {'max_tokens': 20, 'model': 'facebook/opt-125m', 'prompt': 'KServe is a'} [e2e-llm-inference-service] [e2e-llm-inference-service] def request( [e2e-llm-inference-service] self, [e2e-llm-inference-service] method, [e2e-llm-inference-service] url, [e2e-llm-inference-service] params=None, [e2e-llm-inference-service] data=None, [e2e-llm-inference-service] headers=None, [e2e-llm-inference-service] cookies=None, [e2e-llm-inference-service] files=None, [e2e-llm-inference-service] auth=None, [e2e-llm-inference-service] timeout=None, [e2e-llm-inference-service] allow_redirects=True, [e2e-llm-inference-service] proxies=None, [e2e-llm-inference-service] hooks=None, [e2e-llm-inference-service] stream=None, [e2e-llm-inference-service] verify=None, [e2e-llm-inference-service] cert=None, [e2e-llm-inference-service] json=None, [e2e-llm-inference-service] ): [e2e-llm-inference-service] """Constructs a :class:`Request `, prepares it and sends it. [e2e-llm-inference-service] Returns :class:`Response ` object. [e2e-llm-inference-service] [e2e-llm-inference-service] :param method: method for the new :class:`Request` object. [e2e-llm-inference-service] :param url: URL for the new :class:`Request` object. [e2e-llm-inference-service] :param params: (optional) Dictionary or bytes to be sent in the query [e2e-llm-inference-service] string for the :class:`Request`. [e2e-llm-inference-service] :param data: (optional) Dictionary, list of tuples, bytes, or file-like [e2e-llm-inference-service] object to send in the body of the :class:`Request`. [e2e-llm-inference-service] :param json: (optional) json to send in the body of the [e2e-llm-inference-service] :class:`Request`. [e2e-llm-inference-service] :param headers: (optional) Dictionary of HTTP Headers to send with the [e2e-llm-inference-service] :class:`Request`. [e2e-llm-inference-service] :param cookies: (optional) Dict or CookieJar object to send with the [e2e-llm-inference-service] :class:`Request`. [e2e-llm-inference-service] :param files: (optional) Dictionary of ``'filename': file-like-objects`` [e2e-llm-inference-service] for multipart encoding upload. [e2e-llm-inference-service] :param auth: (optional) Auth tuple or callable to enable [e2e-llm-inference-service] Basic/Digest/Custom HTTP Auth. [e2e-llm-inference-service] :param timeout: (optional) How long to wait for the server to send [e2e-llm-inference-service] data before giving up, as a float, or a :ref:`(connect timeout, [e2e-llm-inference-service] read timeout) ` tuple. [e2e-llm-inference-service] :type timeout: float or tuple [e2e-llm-inference-service] :param allow_redirects: (optional) Set to True by default. [e2e-llm-inference-service] :type allow_redirects: bool [e2e-llm-inference-service] :param proxies: (optional) Dictionary mapping protocol or protocol and [e2e-llm-inference-service] hostname to the URL of the proxy. [e2e-llm-inference-service] :param hooks: (optional) Dictionary mapping hook name to one event or [e2e-llm-inference-service] list of events, event must be callable. [e2e-llm-inference-service] :param stream: (optional) whether to immediately download the response [e2e-llm-inference-service] content. Defaults to ``False``. [e2e-llm-inference-service] :param verify: (optional) Either a boolean, in which case it controls whether we verify [e2e-llm-inference-service] the server's TLS certificate, or a string, in which case it must be a path [e2e-llm-inference-service] to a CA bundle to use. Defaults to ``True``. When set to [e2e-llm-inference-service] ``False``, requests will accept any TLS certificate presented by [e2e-llm-inference-service] the server, and will ignore hostname mismatches and/or expired [e2e-llm-inference-service] certificates, which will make your application vulnerable to [e2e-llm-inference-service] man-in-the-middle (MitM) attacks. Setting verify to ``False`` [e2e-llm-inference-service] may be useful during local development or testing. [e2e-llm-inference-service] :param cert: (optional) if String, path to ssl client cert file (.pem). [e2e-llm-inference-service] If Tuple, ('cert', 'key') pair. [e2e-llm-inference-service] :rtype: requests.Response [e2e-llm-inference-service] """ [e2e-llm-inference-service] # Create the Request. [e2e-llm-inference-service] req = Request( [e2e-llm-inference-service] method=method.upper(), [e2e-llm-inference-service] url=url, [e2e-llm-inference-service] headers=headers, [e2e-llm-inference-service] files=files, [e2e-llm-inference-service] data=data or {}, [e2e-llm-inference-service] json=json, [e2e-llm-inference-service] params=params or {}, [e2e-llm-inference-service] auth=auth, [e2e-llm-inference-service] cookies=cookies, [e2e-llm-inference-service] hooks=hooks, [e2e-llm-inference-service] ) [e2e-llm-inference-service] prep = self.prepare_request(req) [e2e-llm-inference-service] [e2e-llm-inference-service] proxies = proxies or {} [e2e-llm-inference-service] [e2e-llm-inference-service] settings = self.merge_environment_settings( [e2e-llm-inference-service] prep.url, proxies, stream, verify, cert [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] # Send the request. [e2e-llm-inference-service] send_kwargs = { [e2e-llm-inference-service] "timeout": timeout, [e2e-llm-inference-service] "allow_redirects": allow_redirects, [e2e-llm-inference-service] } [e2e-llm-inference-service] send_kwargs.update(settings) [e2e-llm-inference-service] > resp = self.send(prep, **send_kwargs) [e2e-llm-inference-service] [e2e-llm-inference-service] ../../python/kserve/.venv/lib64/python3.11/site-packages/requests/sessions.py:589: [e2e-llm-inference-service] _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ [e2e-llm-inference-service] [e2e-llm-inference-service] self = [e2e-llm-inference-service] request = [e2e-llm-inference-service] kwargs = {'cert': None, 'proxies': OrderedDict(), 'stream': False, 'timeout': 60, ...} [e2e-llm-inference-service] allow_redirects = True, stream = False, hooks = {'response': []} [e2e-llm-inference-service] adapter = [e2e-llm-inference-service] start = 1783434821.7591457 [e2e-llm-inference-service] [e2e-llm-inference-service] def send(self, request, **kwargs): [e2e-llm-inference-service] """Send a given PreparedRequest. [e2e-llm-inference-service] [e2e-llm-inference-service] :rtype: requests.Response [e2e-llm-inference-service] """ [e2e-llm-inference-service] # Set defaults that the hooks can utilize to ensure they always have [e2e-llm-inference-service] # the correct parameters to reproduce the previous request. [e2e-llm-inference-service] kwargs.setdefault("stream", self.stream) [e2e-llm-inference-service] kwargs.setdefault("verify", self.verify) [e2e-llm-inference-service] kwargs.setdefault("cert", self.cert) [e2e-llm-inference-service] if "proxies" not in kwargs: [e2e-llm-inference-service] kwargs["proxies"] = resolve_proxies(request, self.proxies, self.trust_env) [e2e-llm-inference-service] [e2e-llm-inference-service] # It's possible that users might accidentally send a Request object. [e2e-llm-inference-service] # Guard against that specific failure case. [e2e-llm-inference-service] if isinstance(request, Request): [e2e-llm-inference-service] raise ValueError("You can only send PreparedRequests.") [e2e-llm-inference-service] [e2e-llm-inference-service] # Set up variables needed for resolve_redirects and dispatching of hooks [e2e-llm-inference-service] allow_redirects = kwargs.pop("allow_redirects", True) [e2e-llm-inference-service] stream = kwargs.get("stream") [e2e-llm-inference-service] hooks = request.hooks [e2e-llm-inference-service] [e2e-llm-inference-service] # Get the appropriate adapter to use [e2e-llm-inference-service] adapter = self.get_adapter(url=request.url) [e2e-llm-inference-service] [e2e-llm-inference-service] # Start time (approximately) of the request [e2e-llm-inference-service] start = preferred_clock() [e2e-llm-inference-service] [e2e-llm-inference-service] # Send the request [e2e-llm-inference-service] > r = adapter.send(request, **kwargs) [e2e-llm-inference-service] [e2e-llm-inference-service] ../../python/kserve/.venv/lib64/python3.11/site-packages/requests/sessions.py:703: [e2e-llm-inference-service] _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ [e2e-llm-inference-service] [e2e-llm-inference-service] self = [e2e-llm-inference-service] request = , stream = False [e2e-llm-inference-service] timeout = Timeout(connect=60, read=60, total=None), verify = '/tmp/ca.crt' [e2e-llm-inference-service] cert = None, proxies = OrderedDict() [e2e-llm-inference-service] [e2e-llm-inference-service] def send( [e2e-llm-inference-service] self, request, stream=False, timeout=None, verify=True, cert=None, proxies=None [e2e-llm-inference-service] ): [e2e-llm-inference-service] """Sends PreparedRequest object. Returns Response object. [e2e-llm-inference-service] [e2e-llm-inference-service] :param request: The :class:`PreparedRequest ` being sent. [e2e-llm-inference-service] :param stream: (optional) Whether to stream the request content. [e2e-llm-inference-service] :param timeout: (optional) How long to wait for the server to send [e2e-llm-inference-service] data before giving up, as a float, or a :ref:`(connect timeout, [e2e-llm-inference-service] read timeout) ` tuple. [e2e-llm-inference-service] :type timeout: float or tuple or urllib3 Timeout object [e2e-llm-inference-service] :param verify: (optional) Either a boolean, in which case it controls whether [e2e-llm-inference-service] we verify the server's TLS certificate, or a string, in which case it [e2e-llm-inference-service] must be a path to a CA bundle to use [e2e-llm-inference-service] :param cert: (optional) Any user-provided SSL certificate to be trusted. [e2e-llm-inference-service] :param proxies: (optional) The proxies dictionary to apply to the request. [e2e-llm-inference-service] :rtype: requests.Response [e2e-llm-inference-service] """ [e2e-llm-inference-service] [e2e-llm-inference-service] try: [e2e-llm-inference-service] conn = self.get_connection_with_tls_context( [e2e-llm-inference-service] request, verify, proxies=proxies, cert=cert [e2e-llm-inference-service] ) [e2e-llm-inference-service] except LocationValueError as e: [e2e-llm-inference-service] raise InvalidURL(e, request=request) [e2e-llm-inference-service] [e2e-llm-inference-service] self.cert_verify(conn, request.url, verify, cert) [e2e-llm-inference-service] url = self.request_url(request, proxies) [e2e-llm-inference-service] self.add_headers( [e2e-llm-inference-service] request, [e2e-llm-inference-service] stream=stream, [e2e-llm-inference-service] timeout=timeout, [e2e-llm-inference-service] verify=verify, [e2e-llm-inference-service] cert=cert, [e2e-llm-inference-service] proxies=proxies, [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] chunked = not (request.body is None or "Content-Length" in request.headers) [e2e-llm-inference-service] [e2e-llm-inference-service] if isinstance(timeout, tuple): [e2e-llm-inference-service] try: [e2e-llm-inference-service] connect, read = timeout [e2e-llm-inference-service] timeout = TimeoutSauce(connect=connect, read=read) [e2e-llm-inference-service] except ValueError: [e2e-llm-inference-service] raise ValueError( [e2e-llm-inference-service] f"Invalid timeout {timeout}. Pass a (connect, read) timeout tuple, " [e2e-llm-inference-service] f"or a single float to set both timeouts to the same value." [e2e-llm-inference-service] ) [e2e-llm-inference-service] elif isinstance(timeout, TimeoutSauce): [e2e-llm-inference-service] pass [e2e-llm-inference-service] else: [e2e-llm-inference-service] timeout = TimeoutSauce(connect=timeout, read=timeout) [e2e-llm-inference-service] [e2e-llm-inference-service] try: [e2e-llm-inference-service] resp = conn.urlopen( [e2e-llm-inference-service] method=request.method, [e2e-llm-inference-service] url=url, [e2e-llm-inference-service] body=request.body, [e2e-llm-inference-service] headers=request.headers, [e2e-llm-inference-service] redirect=False, [e2e-llm-inference-service] assert_same_host=False, [e2e-llm-inference-service] preload_content=False, [e2e-llm-inference-service] decode_content=False, [e2e-llm-inference-service] retries=self.max_retries, [e2e-llm-inference-service] timeout=timeout, [e2e-llm-inference-service] chunked=chunked, [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] except (ProtocolError, OSError) as err: [e2e-llm-inference-service] raise ConnectionError(err, request=request) [e2e-llm-inference-service] [e2e-llm-inference-service] except MaxRetryError as e: [e2e-llm-inference-service] if isinstance(e.reason, ConnectTimeoutError): [e2e-llm-inference-service] # TODO: Remove this in 3.0.0: see #2811 [e2e-llm-inference-service] if not isinstance(e.reason, NewConnectionError): [e2e-llm-inference-service] raise ConnectTimeout(e, request=request) [e2e-llm-inference-service] [e2e-llm-inference-service] if isinstance(e.reason, ResponseError): [e2e-llm-inference-service] raise RetryError(e, request=request) [e2e-llm-inference-service] [e2e-llm-inference-service] if isinstance(e.reason, _ProxyError): [e2e-llm-inference-service] raise ProxyError(e, request=request) [e2e-llm-inference-service] [e2e-llm-inference-service] if isinstance(e.reason, _SSLError): [e2e-llm-inference-service] # This branch is for urllib3 v1.22 and later. [e2e-llm-inference-service] raise SSLError(e, request=request) [e2e-llm-inference-service] [e2e-llm-inference-service] > raise ConnectionError(e, request=request) [e2e-llm-inference-service] E requests.exceptions.ConnectionError: HTTPConnectionPool(host='ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com', port=80): Max retries exceeded with url: /kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-route-87882a8e/v1/completions (Caused by ReadTimeoutError("HTTPConnectionPool(host='ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com', port=80): Read timed out. (read timeout=60)")) [e2e-llm-inference-service] [e2e-llm-inference-service] ../../python/kserve/.venv/lib64/python3.11/site-packages/requests/adapters.py:700: ConnectionError [e2e-llm-inference-service] [e2e-llm-inference-service] The above exception was the direct cause of the following exception: [e2e-llm-inference-service] [e2e-llm-inference-service] test_case = TestCase(base_refs=['router-managed', 'workload-single-cpu', 'model-fb-opt-125m'], prompt='KServe is a', service_name=... {'name': 'model-fb-opt-125m-llmisvc-model-2a331009'}]}, [e2e-llm-inference-service] 'status': None}, model_name='facebook/opt-125m') [e2e-llm-inference-service] [e2e-llm-inference-service] @pytest.mark.llminferenceservice [e2e-llm-inference-service] @pytest.mark.asyncio(loop_scope="session") [e2e-llm-inference-service] @pytest.mark.parametrize( [e2e-llm-inference-service] "test_case", [e2e-llm-inference-service] [ [e2e-llm-inference-service] pytest.param( [e2e-llm-inference-service] TestCase( [e2e-llm-inference-service] base_refs=[ [e2e-llm-inference-service] "router-with-gateway-ref", [e2e-llm-inference-service] "router-with-managed-route", [e2e-llm-inference-service] "model-fb-opt-125m", [e2e-llm-inference-service] "workload-llmd-simulator", [e2e-llm-inference-service] ], [e2e-llm-inference-service] endpoint="/v1/completions", [e2e-llm-inference-service] prompt="KServe is a", [e2e-llm-inference-service] payload_formatter=completions_payload, [e2e-llm-inference-service] response_assertion=create_response_assertion(with_field="choices"), [e2e-llm-inference-service] expected_gateway=ROUTER_GATEWAYS[0], [e2e-llm-inference-service] before_test=[ [e2e-llm-inference-service] lambda: create_router_resources( [e2e-llm-inference-service] gateways=[ROUTER_GATEWAYS[0]], [e2e-llm-inference-service] ) [e2e-llm-inference-service] ], [e2e-llm-inference-service] ), [e2e-llm-inference-service] marks=[ [e2e-llm-inference-service] pytest.mark.cluster_cpu, [e2e-llm-inference-service] pytest.mark.cluster_single_node, [e2e-llm-inference-service] pytest.mark.llmd_simulator, [e2e-llm-inference-service] pytest.mark.custom_gateway, [e2e-llm-inference-service] ], [e2e-llm-inference-service] ), [e2e-llm-inference-service] pytest.param( [e2e-llm-inference-service] TestCase( [e2e-llm-inference-service] base_refs=[ [e2e-llm-inference-service] "router-managed", [e2e-llm-inference-service] "workload-single-cpu", [e2e-llm-inference-service] "model-fb-opt-125m", [e2e-llm-inference-service] ], [e2e-llm-inference-service] prompt="KServe is a", [e2e-llm-inference-service] payload_formatter=completions_payload, [e2e-llm-inference-service] response_assertion=assert_200_with_choices, [e2e-llm-inference-service] ), [e2e-llm-inference-service] marks=[pytest.mark.cluster_cpu, pytest.mark.cluster_single_node], [e2e-llm-inference-service] ), [e2e-llm-inference-service] pytest.param( [e2e-llm-inference-service] TestCase( [e2e-llm-inference-service] base_refs=[ [e2e-llm-inference-service] "router-custom-route-timeout", [e2e-llm-inference-service] "scheduler-managed", [e2e-llm-inference-service] "workload-single-cpu", [e2e-llm-inference-service] "model-fb-opt-125m", [e2e-llm-inference-service] ], [e2e-llm-inference-service] prompt="KServe is a", [e2e-llm-inference-service] service_name="custom-route-timeout-test", [e2e-llm-inference-service] ), [e2e-llm-inference-service] marks=[pytest.mark.cluster_cpu, pytest.mark.cluster_single_node], [e2e-llm-inference-service] ), [e2e-llm-inference-service] pytest.param( [e2e-llm-inference-service] TestCase( [e2e-llm-inference-service] base_refs=[ [e2e-llm-inference-service] "router-with-refs", [e2e-llm-inference-service] "scheduler-managed", [e2e-llm-inference-service] "workload-single-cpu", [e2e-llm-inference-service] "model-fb-opt-125m", [e2e-llm-inference-service] ], [e2e-llm-inference-service] prompt="KServe is a", [e2e-llm-inference-service] service_name="router-with-refs-test", [e2e-llm-inference-service] expected_gateway=ROUTER_GATEWAYS[0], [e2e-llm-inference-service] before_test=[ [e2e-llm-inference-service] lambda: create_router_resources( [e2e-llm-inference-service] gateways=[ROUTER_GATEWAYS[0]], [e2e-llm-inference-service] routes=[ROUTER_ROUTES[0], ROUTER_ROUTES[1]], [e2e-llm-inference-service] ) [e2e-llm-inference-service] ], [e2e-llm-inference-service] ), [e2e-llm-inference-service] marks=[ [e2e-llm-inference-service] pytest.mark.cluster_cpu, [e2e-llm-inference-service] pytest.mark.cluster_single_node, [e2e-llm-inference-service] pytest.mark.custom_gateway, [e2e-llm-inference-service] ], [e2e-llm-inference-service] ), [e2e-llm-inference-service] pytest.param( [e2e-llm-inference-service] TestCase( [e2e-llm-inference-service] base_refs=["router-managed", "workload-pd-cpu", "model-fb-opt-125m"], [e2e-llm-inference-service] prompt="You are an expert in Kubernetes-native machine learning serving platforms, with deep knowledge of the KServe project. " [e2e-llm-inference-service] "Explain the challenges of serving large-scale models, GPU scheduling, and how KServe integrates with capabilities like multi-model serving. " [e2e-llm-inference-service] "Provide a detailed comparison with open source alternatives, focusing on operational trade-offs.", [e2e-llm-inference-service] response_assertion=assert_200_with_choices, [e2e-llm-inference-service] ), [e2e-llm-inference-service] marks=[pytest.mark.cluster_cpu, pytest.mark.cluster_single_node], [e2e-llm-inference-service] ), [e2e-llm-inference-service] pytest.param( [e2e-llm-inference-service] TestCase( [e2e-llm-inference-service] base_refs=[ [e2e-llm-inference-service] "router-custom-route-timeout-pd", [e2e-llm-inference-service] "scheduler-managed", [e2e-llm-inference-service] "workload-pd-cpu", [e2e-llm-inference-service] "model-fb-opt-125m", [e2e-llm-inference-service] ], [e2e-llm-inference-service] prompt="You are an expert in Kubernetes-native machine learning serving platforms, with deep knowledge of the KServe project. " [e2e-llm-inference-service] "Explain the challenges of serving large-scale models, GPU scheduling, and how KServe integrates with capabilities like multi-model serving. " [e2e-llm-inference-service] "Provide a detailed comparison with open source alternatives, focusing on operational trade-offs.", [e2e-llm-inference-service] service_name="custom-route-timeout-pd-test", [e2e-llm-inference-service] response_assertion=assert_200_with_choices, [e2e-llm-inference-service] ), [e2e-llm-inference-service] marks=[pytest.mark.cluster_cpu, pytest.mark.cluster_single_node], [e2e-llm-inference-service] ), [e2e-llm-inference-service] pytest.param( [e2e-llm-inference-service] TestCase( [e2e-llm-inference-service] base_refs=[ [e2e-llm-inference-service] "router-with-refs-pd", [e2e-llm-inference-service] "scheduler-managed", [e2e-llm-inference-service] "workload-pd-cpu", [e2e-llm-inference-service] "model-fb-opt-125m", [e2e-llm-inference-service] ], [e2e-llm-inference-service] prompt="You are an expert in Kubernetes-native machine learning serving platforms, with deep knowledge of the KServe project. " [e2e-llm-inference-service] "Explain the challenges of serving large-scale models, GPU scheduling, and how KServe integrates with capabilities like multi-model serving. " [e2e-llm-inference-service] "Provide a detailed comparison with open source alternatives, focusing on operational trade-offs.", [e2e-llm-inference-service] service_name="router-with-refs-pd-test", [e2e-llm-inference-service] response_assertion=assert_200_with_choices, [e2e-llm-inference-service] expected_gateway=ROUTER_GATEWAYS[1], [e2e-llm-inference-service] before_test=[ [e2e-llm-inference-service] lambda: create_router_resources( [e2e-llm-inference-service] gateways=[ROUTER_GATEWAYS[1]], [e2e-llm-inference-service] routes=[ROUTER_ROUTES[2], ROUTER_ROUTES[3]], [e2e-llm-inference-service] ) [e2e-llm-inference-service] ], [e2e-llm-inference-service] ), [e2e-llm-inference-service] marks=[ [e2e-llm-inference-service] pytest.mark.cluster_cpu, [e2e-llm-inference-service] pytest.mark.cluster_single_node, [e2e-llm-inference-service] pytest.mark.custom_gateway, [e2e-llm-inference-service] ], [e2e-llm-inference-service] ), [e2e-llm-inference-service] pytest.param( [e2e-llm-inference-service] TestCase( [e2e-llm-inference-service] base_refs=[ [e2e-llm-inference-service] "router-managed", [e2e-llm-inference-service] "workload-dp-ep-gpu", [e2e-llm-inference-service] "workload-dp-ep-prefill-gpu", [e2e-llm-inference-service] "model-deepseek-v2-lite", [e2e-llm-inference-service] ], [e2e-llm-inference-service] prompt="Delve into the multifaceted implications of a fully disaggregated cloud architecture, specifically " [e2e-llm-inference-service] "where the compute plane (P) and the data plane (D) are independently deployed and managed for a " [e2e-llm-inference-service] "geographically distributed, high-throughput, low-latency microservices ecosystem. Beyond the " [e2e-llm-inference-service] "fundamental challenges of network latency and data consistency, elaborate on the advanced " [e2e-llm-inference-service] "considerations and trade-offs inherent in such a setup: 1. Network Architecture and Protocols: " [e2e-llm-inference-service] "How would the network fabric and underlying protocols (e.g., RDMA, custom transport layers) need to " [e2e-llm-inference-service] "evolve to support optimal performance and minimize inter-plane communication overhead, especially for " [e2e-llm-inference-service] "synchronous operations? Discuss the role of network programmability (e.g., SDN, P4) in dynamically " [e2e-llm-inference-service] "optimizing routing and traffic flow between P and D. 2. Advanced Data Consistency and Durability: " [e2e-llm-inference-service] "Explore sophisticated data consistency models (e.g., causal consistency, strong eventual consistency) " [e2e-llm-inference-service] "and their applicability in balancing performance and data integrity across a globally distributed data plane. " [e2e-llm-inference-service] "Detail strategies for ensuring data durability and fault tolerance, including multi-region replication, " [e2e-llm-inference-service] "intelligent partitioning, and recovery mechanisms in the event of partial or full plane failures. " [e2e-llm-inference-service] "3. Dynamic Resource Orchestration and Cost Optimization: Analyze how an orchestration layer would intelligently " [e2e-llm-inference-service] "manage the independent scaling of compute (P) and data (D) resources, considering fluctuating workloads, " [e2e-llm-inference-service] "cost efficiency, and performance targets (e.g., using predictive analytics for resource provisioning). " [e2e-llm-inference-service] "Discuss mechanisms for dynamically reallocating compute nodes to different data partitions based on " [e2e-llm-inference-service] "workload patterns and data locality, potentially involving live migration strategies. " [e2e-llm-inference-service] "4. Security and Compliance in a Distributed Landscape: Address the enhanced security perimeter " [e2e-llm-inference-service] "challenges, including securing communication channels between P and D (encryption in transit, mutual TLS), " [e2e-llm-inference-service] "fine-grained access control to data at rest and in motion, and identity management across disaggregated " [e2e-llm-inference-service] "components. Discuss how such an architecture impacts compliance with regulatory frameworks (e.g., GDPR, HIPAA) " [e2e-llm-inference-service] "concerning data sovereignty, privacy, and auditability. 5. Operational Complexity and Observability: " [e2e-llm-inference-service] "Examine the increased complexity in monitoring, logging, and tracing across highly decoupled compute and " [e2e-llm-inference-service] "data planes. What specialized tooling and practices (e.g., distributed tracing with OpenTelemetry, advanced AIOps) " [e2e-llm-inference-service] "would be essential? How would incident response and troubleshooting differ in this disaggregated environment " [e2e-llm-inference-service] "compared to traditional integrated systems? Consider the challenges of pinpointing root causes across " [e2e-llm-inference-service] "independent failures. 6. Real-world Applicability and Future Trends: Identify specific industries " [e2e-llm-inference-service] "or use cases (e.g., high-frequency trading, IoT edge processing, large language model inference) " [e2e-llm-inference-service] "where the benefits of P/D disaggregation would strongly outweigh its complexities. " [e2e-llm-inference-service] "Conclude by speculating on emerging technologies or paradigms (e.g., serverless compute functions " [e2e-llm-inference-service] "directly interacting with object storage, in-memory disaggregation) that could further drive or " [e2e-llm-inference-service] "transform P/D disaggregation in cloud computing.", [e2e-llm-inference-service] max_tokens=2000, [e2e-llm-inference-service] ), [e2e-llm-inference-service] marks=[ [e2e-llm-inference-service] pytest.mark.cluster_gpu, [e2e-llm-inference-service] pytest.mark.cluster_nvidia, [e2e-llm-inference-service] pytest.mark.cluster_nvidia_roce, [e2e-llm-inference-service] ], [e2e-llm-inference-service] ), [e2e-llm-inference-service] pytest.param( [e2e-llm-inference-service] TestCase( [e2e-llm-inference-service] base_refs=[ [e2e-llm-inference-service] "router-no-scheduler", [e2e-llm-inference-service] "workload-single-cpu", [e2e-llm-inference-service] "model-fb-opt-125m", [e2e-llm-inference-service] ], [e2e-llm-inference-service] prompt="What is KServe?", [e2e-llm-inference-service] ), [e2e-llm-inference-service] marks=[ [e2e-llm-inference-service] pytest.mark.cluster_cpu, [e2e-llm-inference-service] pytest.mark.cluster_single_node, [e2e-llm-inference-service] pytest.mark.no_scheduler, [e2e-llm-inference-service] ], [e2e-llm-inference-service] ), [e2e-llm-inference-service] pytest.param( [e2e-llm-inference-service] TestCase( [e2e-llm-inference-service] base_refs=[ [e2e-llm-inference-service] "router-managed", [e2e-llm-inference-service] "workload-simulated-dp-ep-cpu", [e2e-llm-inference-service] "model-fb-opt-125m", [e2e-llm-inference-service] ], [e2e-llm-inference-service] prompt="This test simulates DP+EP that can run on CPU, the idea is to test the LWS-based deployment, " [e2e-llm-inference-service] "but without the resources requirements for DP+EP (GPUs and ROCe/IB).", [e2e-llm-inference-service] ), [e2e-llm-inference-service] marks=[pytest.mark.cluster_cpu, pytest.mark.cluster_multi_node], [e2e-llm-inference-service] ), [e2e-llm-inference-service] # Scheduler config tests [e2e-llm-inference-service] pytest.param( [e2e-llm-inference-service] TestCase( [e2e-llm-inference-service] base_refs=[ [e2e-llm-inference-service] "router-managed", [e2e-llm-inference-service] "scheduler-with-inline-config", [e2e-llm-inference-service] "workload-llmd-simulator", [e2e-llm-inference-service] ], [e2e-llm-inference-service] prompt="KServe is a", [e2e-llm-inference-service] service_name="scheduler-inline-config-test", [e2e-llm-inference-service] ), [e2e-llm-inference-service] marks=[pytest.mark.cluster_cpu, pytest.mark.cluster_single_node], [e2e-llm-inference-service] ), [e2e-llm-inference-service] # Chat completions endpoint coverage [e2e-llm-inference-service] pytest.param( [e2e-llm-inference-service] TestCase( [e2e-llm-inference-service] base_refs=[ [e2e-llm-inference-service] "router-managed", [e2e-llm-inference-service] "workload-llmd-simulator", [e2e-llm-inference-service] "model-qwen2.5-0.5b", [e2e-llm-inference-service] ], [e2e-llm-inference-service] model_name="Qwen/Qwen2.5-0.5B-Instruct", [e2e-llm-inference-service] endpoint="/v1/chat/completions", [e2e-llm-inference-service] prompt="What is KServe?", [e2e-llm-inference-service] payload_formatter=chat_completions_payload, [e2e-llm-inference-service] response_assertion=create_response_assertion(with_field="choices"), [e2e-llm-inference-service] ), [e2e-llm-inference-service] marks=[ [e2e-llm-inference-service] pytest.mark.cluster_cpu, [e2e-llm-inference-service] pytest.mark.cluster_single_node, [e2e-llm-inference-service] pytest.mark.llmd_simulator, [e2e-llm-inference-service] ], [e2e-llm-inference-service] ), [e2e-llm-inference-service] pytest.param( [e2e-llm-inference-service] TestCase( [e2e-llm-inference-service] base_refs=[ [e2e-llm-inference-service] "router-managed", [e2e-llm-inference-service] "scheduler-with-configmap-ref", [e2e-llm-inference-service] "workload-llmd-simulator", [e2e-llm-inference-service] ], [e2e-llm-inference-service] prompt="KServe is a", [e2e-llm-inference-service] service_name="scheduler-configmap-ref-test", [e2e-llm-inference-service] before_test=[create_scheduler_configmap], [e2e-llm-inference-service] after_test=[delete_scheduler_configmap], [e2e-llm-inference-service] ), [e2e-llm-inference-service] marks=[pytest.mark.cluster_cpu, pytest.mark.cluster_single_node], [e2e-llm-inference-service] ), [e2e-llm-inference-service] pytest.param( [e2e-llm-inference-service] TestCase( [e2e-llm-inference-service] base_refs=[ [e2e-llm-inference-service] "router-managed", [e2e-llm-inference-service] "scheduler-with-replicas", [e2e-llm-inference-service] "workload-llmd-simulator", [e2e-llm-inference-service] ], [e2e-llm-inference-service] prompt="KServe is a", [e2e-llm-inference-service] service_name="scheduler-ha-replicas-test", [e2e-llm-inference-service] ), [e2e-llm-inference-service] marks=[pytest.mark.cluster_cpu, pytest.mark.cluster_single_node], [e2e-llm-inference-service] ), [e2e-llm-inference-service] pytest.param( [e2e-llm-inference-service] TestCase( [e2e-llm-inference-service] base_refs=[ [e2e-llm-inference-service] "router-managed", [e2e-llm-inference-service] "scheduler-with-custom-template", [e2e-llm-inference-service] "workload-llmd-simulator", [e2e-llm-inference-service] ], [e2e-llm-inference-service] prompt="KServe is a", [e2e-llm-inference-service] service_name="scheduler-custom-template-test", [e2e-llm-inference-service] ), [e2e-llm-inference-service] marks=[pytest.mark.cluster_cpu, pytest.mark.cluster_single_node], [e2e-llm-inference-service] ), [e2e-llm-inference-service] # Scheduler v0.6 → v0.7 migration tests. [e2e-llm-inference-service] # Deploy v0.6-style configs and verify the controller migrates them [e2e-llm-inference-service] # so the v0.7 scheduler boots successfully. [e2e-llm-inference-service] pytest.param( [e2e-llm-inference-service] TestCase( [e2e-llm-inference-service] base_refs=[ [e2e-llm-inference-service] "router-managed", [e2e-llm-inference-service] "scheduler-v06-pd-config-migration", [e2e-llm-inference-service] "workload-llmd-simulator-pd", [e2e-llm-inference-service] ], [e2e-llm-inference-service] prompt="KServe is a", [e2e-llm-inference-service] service_name="scheduler-v06-pd-migration-test", [e2e-llm-inference-service] response_assertion=assert_200_with_choices, [e2e-llm-inference-service] ), [e2e-llm-inference-service] marks=[ [e2e-llm-inference-service] pytest.mark.cluster_cpu, [e2e-llm-inference-service] pytest.mark.cluster_single_node, [e2e-llm-inference-service] pytest.mark.llmd_simulator, [e2e-llm-inference-service] ], [e2e-llm-inference-service] ), [e2e-llm-inference-service] pytest.param( [e2e-llm-inference-service] TestCase( [e2e-llm-inference-service] base_refs=[ [e2e-llm-inference-service] "router-managed", [e2e-llm-inference-service] "scheduler-v06-nonzero-threshold-migration", [e2e-llm-inference-service] "workload-llmd-simulator-pd", [e2e-llm-inference-service] ], [e2e-llm-inference-service] prompt="KServe is a", [e2e-llm-inference-service] service_name="scheduler-v06-threshold-migration-test", [e2e-llm-inference-service] response_assertion=assert_200_with_choices, [e2e-llm-inference-service] ), [e2e-llm-inference-service] marks=[ [e2e-llm-inference-service] pytest.mark.cluster_cpu, [e2e-llm-inference-service] pytest.mark.cluster_single_node, [e2e-llm-inference-service] pytest.mark.llmd_simulator, [e2e-llm-inference-service] ], [e2e-llm-inference-service] ), [e2e-llm-inference-service] # Precise prefix KV cache routing test [e2e-llm-inference-service] pytest.param( [e2e-llm-inference-service] TestCase( [e2e-llm-inference-service] base_refs=[ [e2e-llm-inference-service] "router-managed", [e2e-llm-inference-service] "scheduler-with-precise-prefix-cache-inline-config", [e2e-llm-inference-service] "workload-llmd-simulator-kvcache", [e2e-llm-inference-service] ], [e2e-llm-inference-service] prompt="KServe is a", [e2e-llm-inference-service] service_name="precise-prefix-cache-test", [e2e-llm-inference-service] ), [e2e-llm-inference-service] marks=[ [e2e-llm-inference-service] pytest.mark.cluster_cpu, [e2e-llm-inference-service] pytest.mark.cluster_single_node, [e2e-llm-inference-service] pytest.mark.llmd_simulator, [e2e-llm-inference-service] ], [e2e-llm-inference-service] ), [e2e-llm-inference-service] # Models endpoint coverage [e2e-llm-inference-service] pytest.param( [e2e-llm-inference-service] TestCase( [e2e-llm-inference-service] base_refs=[ [e2e-llm-inference-service] "router-managed", [e2e-llm-inference-service] "workload-llmd-simulator", [e2e-llm-inference-service] ], [e2e-llm-inference-service] endpoint="/v1/models", [e2e-llm-inference-service] response_assertion=create_response_assertion(with_field="data"), [e2e-llm-inference-service] ), [e2e-llm-inference-service] marks=[ [e2e-llm-inference-service] pytest.mark.cluster_cpu, [e2e-llm-inference-service] pytest.mark.cluster_single_node, [e2e-llm-inference-service] pytest.mark.llmd_simulator, [e2e-llm-inference-service] ], [e2e-llm-inference-service] ), [e2e-llm-inference-service] # Model-based routing via X-Gateway-Model-Name header — /v1/completions [e2e-llm-inference-service] pytest.param( [e2e-llm-inference-service] TestCase( [e2e-llm-inference-service] base_refs=[ [e2e-llm-inference-service] "router-managed", [e2e-llm-inference-service] "workload-llmd-simulator", [e2e-llm-inference-service] ], [e2e-llm-inference-service] endpoint="/v1/completions", [e2e-llm-inference-service] prompt="KServe is a", [e2e-llm-inference-service] payload_formatter=completions_payload, [e2e-llm-inference-service] response_assertion=assert_model_field_matches("facebook/opt-125m"), [e2e-llm-inference-service] url_getter=get_model_routing_url, [e2e-llm-inference-service] extra_headers={ [e2e-llm-inference-service] MODEL_ROUTING_HEADER: f"publishers/{KSERVE_TEST_NAMESPACE}/models/facebook/opt-125m", [e2e-llm-inference-service] }, [e2e-llm-inference-service] peers=[ [e2e-llm-inference-service] TestCase( [e2e-llm-inference-service] base_refs=[ [e2e-llm-inference-service] "router-managed", [e2e-llm-inference-service] "workload-llmd-simulator", [e2e-llm-inference-service] "model-qwen2.5-0.5b", [e2e-llm-inference-service] ], [e2e-llm-inference-service] endpoint="/v1/completions", [e2e-llm-inference-service] prompt="KServe is a", [e2e-llm-inference-service] payload_formatter=completions_payload, [e2e-llm-inference-service] response_assertion=assert_model_field_matches( [e2e-llm-inference-service] "Qwen/Qwen2.5-0.5B-Instruct" [e2e-llm-inference-service] ), [e2e-llm-inference-service] url_getter=get_model_routing_url, [e2e-llm-inference-service] extra_headers={ [e2e-llm-inference-service] MODEL_ROUTING_HEADER: f"publishers/{KSERVE_TEST_NAMESPACE}/models/Qwen/Qwen2.5-0.5B-Instruct", [e2e-llm-inference-service] }, [e2e-llm-inference-service] ), [e2e-llm-inference-service] ], [e2e-llm-inference-service] ), [e2e-llm-inference-service] marks=[ [e2e-llm-inference-service] pytest.mark.cluster_cpu, [e2e-llm-inference-service] pytest.mark.cluster_single_node, [e2e-llm-inference-service] pytest.mark.llmd_simulator, [e2e-llm-inference-service] pytest.mark.model_routing, [e2e-llm-inference-service] ], [e2e-llm-inference-service] ), [e2e-llm-inference-service] # Model-based routing via X-Gateway-Model-Name header — /v1/chat/completions [e2e-llm-inference-service] pytest.param( [e2e-llm-inference-service] TestCase( [e2e-llm-inference-service] base_refs=[ [e2e-llm-inference-service] "router-managed", [e2e-llm-inference-service] "workload-llmd-simulator", [e2e-llm-inference-service] ], [e2e-llm-inference-service] endpoint="/v1/chat/completions", [e2e-llm-inference-service] prompt="What is KServe?", [e2e-llm-inference-service] payload_formatter=chat_completions_payload, [e2e-llm-inference-service] response_assertion=assert_model_field_matches("facebook/opt-125m"), [e2e-llm-inference-service] url_getter=get_model_routing_url, [e2e-llm-inference-service] extra_headers={ [e2e-llm-inference-service] MODEL_ROUTING_HEADER: f"publishers/{KSERVE_TEST_NAMESPACE}/models/facebook/opt-125m", [e2e-llm-inference-service] }, [e2e-llm-inference-service] peers=[ [e2e-llm-inference-service] TestCase( [e2e-llm-inference-service] base_refs=[ [e2e-llm-inference-service] "router-managed", [e2e-llm-inference-service] "workload-llmd-simulator", [e2e-llm-inference-service] "model-qwen2.5-0.5b", [e2e-llm-inference-service] ], [e2e-llm-inference-service] endpoint="/v1/chat/completions", [e2e-llm-inference-service] prompt="What is KServe?", [e2e-llm-inference-service] payload_formatter=chat_completions_payload, [e2e-llm-inference-service] response_assertion=assert_model_field_matches( [e2e-llm-inference-service] "Qwen/Qwen2.5-0.5B-Instruct" [e2e-llm-inference-service] ), [e2e-llm-inference-service] url_getter=get_model_routing_url, [e2e-llm-inference-service] extra_headers={ [e2e-llm-inference-service] MODEL_ROUTING_HEADER: f"publishers/{KSERVE_TEST_NAMESPACE}/models/Qwen/Qwen2.5-0.5B-Instruct", [e2e-llm-inference-service] }, [e2e-llm-inference-service] ), [e2e-llm-inference-service] ], [e2e-llm-inference-service] ), [e2e-llm-inference-service] marks=[ [e2e-llm-inference-service] pytest.mark.cluster_cpu, [e2e-llm-inference-service] pytest.mark.cluster_single_node, [e2e-llm-inference-service] pytest.mark.llmd_simulator, [e2e-llm-inference-service] pytest.mark.model_routing, [e2e-llm-inference-service] ], [e2e-llm-inference-service] ), [e2e-llm-inference-service] # Model-based routing via X-Gateway-Model-Name header — LoRA adapter [e2e-llm-inference-service] pytest.param( [e2e-llm-inference-service] TestCase( [e2e-llm-inference-service] base_refs=[ [e2e-llm-inference-service] "router-managed", [e2e-llm-inference-service] "workload-single-cpu", [e2e-llm-inference-service] "model-fb-opt-125m-with-lora-hf", [e2e-llm-inference-service] ], [e2e-llm-inference-service] endpoint="/v1/completions", [e2e-llm-inference-service] prompt="KServe is a", [e2e-llm-inference-service] model_name=f"publishers/{KSERVE_TEST_NAMESPACE}/models/lora-adapter-1", [e2e-llm-inference-service] payload_formatter=completions_payload, [e2e-llm-inference-service] response_assertion=assert_model_field_matches( [e2e-llm-inference-service] f"publishers/{KSERVE_TEST_NAMESPACE}/models/lora-adapter-1" [e2e-llm-inference-service] ), [e2e-llm-inference-service] url_getter=get_model_routing_url, [e2e-llm-inference-service] extra_headers={ [e2e-llm-inference-service] MODEL_ROUTING_HEADER: f"publishers/{KSERVE_TEST_NAMESPACE}/models/lora-adapter-1", [e2e-llm-inference-service] }, [e2e-llm-inference-service] ), [e2e-llm-inference-service] marks=[ [e2e-llm-inference-service] pytest.mark.cluster_cpu, [e2e-llm-inference-service] pytest.mark.cluster_single_node, [e2e-llm-inference-service] pytest.mark.model_routing, [e2e-llm-inference-service] pytest.mark.lora, [e2e-llm-inference-service] ], [e2e-llm-inference-service] ), [e2e-llm-inference-service] # Model-based routing via X-Gateway-Model-Name header — /v1/models (base + LoRA) [e2e-llm-inference-service] pytest.param( [e2e-llm-inference-service] TestCase( [e2e-llm-inference-service] base_refs=[ [e2e-llm-inference-service] "router-managed", [e2e-llm-inference-service] "workload-single-cpu", [e2e-llm-inference-service] "model-fb-opt-125m-with-lora-hf", [e2e-llm-inference-service] ], [e2e-llm-inference-service] endpoint="/v1/models", [e2e-llm-inference-service] response_assertion=assert_models_contains( [e2e-llm-inference-service] "facebook/opt-125m", [e2e-llm-inference-service] f"publishers/{KSERVE_TEST_NAMESPACE}/models/facebook/opt-125m", [e2e-llm-inference-service] "lora-adapter-1", [e2e-llm-inference-service] f"publishers/{KSERVE_TEST_NAMESPACE}/models/lora-adapter-1", [e2e-llm-inference-service] ), [e2e-llm-inference-service] url_getter=get_model_routing_url, [e2e-llm-inference-service] extra_headers={ [e2e-llm-inference-service] MODEL_ROUTING_HEADER: f"publishers/{KSERVE_TEST_NAMESPACE}/models/facebook/opt-125m", [e2e-llm-inference-service] }, [e2e-llm-inference-service] ), [e2e-llm-inference-service] marks=[ [e2e-llm-inference-service] pytest.mark.cluster_cpu, [e2e-llm-inference-service] pytest.mark.cluster_single_node, [e2e-llm-inference-service] pytest.mark.model_routing, [e2e-llm-inference-service] pytest.mark.lora, [e2e-llm-inference-service] ], [e2e-llm-inference-service] ), [e2e-llm-inference-service] # PVC storage tests -- validate direct PVC volume mount with real vLLM serving [e2e-llm-inference-service] pytest.param( [e2e-llm-inference-service] TestCase( [e2e-llm-inference-service] base_refs=[ [e2e-llm-inference-service] "router-managed", [e2e-llm-inference-service] "workload-single-cpu", [e2e-llm-inference-service] "model-pvc", [e2e-llm-inference-service] ], [e2e-llm-inference-service] prompt="KServe is a", [e2e-llm-inference-service] response_assertion=assert_200_with_choices, [e2e-llm-inference-service] before_test=[ensure_pvc_with_model], [e2e-llm-inference-service] ), [e2e-llm-inference-service] marks=[ [e2e-llm-inference-service] pytest.mark.cluster_cpu, [e2e-llm-inference-service] pytest.mark.cluster_single_node, [e2e-llm-inference-service] pytest.mark.pvc_storage, [e2e-llm-inference-service] ], [e2e-llm-inference-service] ), [e2e-llm-inference-service] pytest.param( [e2e-llm-inference-service] TestCase( [e2e-llm-inference-service] base_refs=[ [e2e-llm-inference-service] "router-managed", [e2e-llm-inference-service] "workload-pd-cpu", [e2e-llm-inference-service] "model-pvc", [e2e-llm-inference-service] ], [e2e-llm-inference-service] prompt="KServe is a", [e2e-llm-inference-service] response_assertion=assert_200_with_choices, [e2e-llm-inference-service] before_test=[ensure_pvc_with_model], [e2e-llm-inference-service] ), [e2e-llm-inference-service] marks=[ [e2e-llm-inference-service] pytest.mark.cluster_cpu, [e2e-llm-inference-service] pytest.mark.cluster_single_node, [e2e-llm-inference-service] pytest.mark.pvc_storage, [e2e-llm-inference-service] ], [e2e-llm-inference-service] ), [e2e-llm-inference-service] pytest.param( [e2e-llm-inference-service] TestCase( [e2e-llm-inference-service] base_refs=[ [e2e-llm-inference-service] "router-managed", [e2e-llm-inference-service] "workload-simulated-dp-ep-cpu", [e2e-llm-inference-service] "model-pvc", [e2e-llm-inference-service] ], [e2e-llm-inference-service] prompt="KServe is a", [e2e-llm-inference-service] before_test=[ensure_pvc_with_model], [e2e-llm-inference-service] ), [e2e-llm-inference-service] marks=[ [e2e-llm-inference-service] pytest.mark.cluster_cpu, [e2e-llm-inference-service] pytest.mark.cluster_multi_node, [e2e-llm-inference-service] pytest.mark.pvc_storage, [e2e-llm-inference-service] ], [e2e-llm-inference-service] ), [e2e-llm-inference-service] ], [e2e-llm-inference-service] indirect=["test_case"], [e2e-llm-inference-service] ids=generate_test_id, [e2e-llm-inference-service] ) [e2e-llm-inference-service] @log_execution [e2e-llm-inference-service] def test_llm_inference_service(test_case: TestCase): # noqa: F811 [e2e-llm-inference-service] inject_k8s_proxy() [e2e-llm-inference-service] [e2e-llm-inference-service] kserve_client = KServeClient( [e2e-llm-inference-service] config_file=os.environ.get("KUBECONFIG", "~/.kube/config"), [e2e-llm-inference-service] client_configuration=client.Configuration(), [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] service_name = test_case.llm_service.metadata.name [e2e-llm-inference-service] if not test_case.llm_service.metadata.annotations: [e2e-llm-inference-service] test_case.llm_service.metadata.annotations = {} [e2e-llm-inference-service] [e2e-llm-inference-service] test_case.llm_service.metadata.annotations[ [e2e-llm-inference-service] "security.opendatahub.io/enable-auth" [e2e-llm-inference-service] ] = "false" [e2e-llm-inference-service] prefix = test_case.log_prefix [e2e-llm-inference-service] [e2e-llm-inference-service] test_failed = False [e2e-llm-inference-service] try: [e2e-llm-inference-service] print(f"{prefix} Creating LLMInferenceService {service_name}") [e2e-llm-inference-service] create_llmisvc(kserve_client, test_case.llm_service) [e2e-llm-inference-service] print(f"{prefix} Waiting for LLMInferenceService {service_name} to be ready") [e2e-llm-inference-service] wait_for_llm_isvc_ready( [e2e-llm-inference-service] kserve_client, test_case.llm_service, test_case.wait_timeout [e2e-llm-inference-service] ) [e2e-llm-inference-service] print(f"{prefix} Waiting for model response from {service_name}") [e2e-llm-inference-service] > wait_for_model_response( [e2e-llm-inference-service] kserve_client, [e2e-llm-inference-service] test_case, [e2e-llm-inference-service] test_case.wait_timeout, [e2e-llm-inference-service] extra_headers=test_case.extra_headers, [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] llmisvc/test_llm_inference_service.py:816: [e2e-llm-inference-service] _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ [e2e-llm-inference-service] [e2e-llm-inference-service] args = (, TestCase(base_refs=['router-managed', 'workload-sin... {'name': 'model-fb-opt-125m-llmisvc-model-2a331009'}]}, [e2e-llm-inference-service] 'status': None}, model_name='facebook/opt-125m'), 900) [e2e-llm-inference-service] kwargs = {'extra_headers': None}, func_name = 'wait_for_model_response' [e2e-llm-inference-service] timestamp_start = '2026-07-07T14:33:41.745372', start_time = 1783434821.7457728 [e2e-llm-inference-service] duration = 904.5519132614136, timestamp_end = '2026-07-07T14:48:46.297689' [e2e-llm-inference-service] [e2e-llm-inference-service] @functools.wraps(func) [e2e-llm-inference-service] def wrapper(*args, **kwargs): [e2e-llm-inference-service] func_name = func.__name__ [e2e-llm-inference-service] [e2e-llm-inference-service] timestamp_start = datetime.now().isoformat() [e2e-llm-inference-service] logger.info( [e2e-llm-inference-service] f"[{func_name}] [{timestamp_start}] start - args={args}, kwargs={kwargs}" [e2e-llm-inference-service] ) [e2e-llm-inference-service] start_time = time.time() [e2e-llm-inference-service] [e2e-llm-inference-service] try: [e2e-llm-inference-service] > result = func(*args, **kwargs) [e2e-llm-inference-service] [e2e-llm-inference-service] llmisvc/logging.py:40: [e2e-llm-inference-service] _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ [e2e-llm-inference-service] [e2e-llm-inference-service] kserve_client = [e2e-llm-inference-service] test_case = TestCase(base_refs=['router-managed', 'workload-single-cpu', 'model-fb-opt-125m'], prompt='KServe is a', service_name=... {'name': 'model-fb-opt-125m-llmisvc-model-2a331009'}]}, [e2e-llm-inference-service] 'status': None}, model_name='facebook/opt-125m') [e2e-llm-inference-service] timeout_seconds = 900, extra_headers = None [e2e-llm-inference-service] [e2e-llm-inference-service] @log_execution [e2e-llm-inference-service] def wait_for_model_response( [e2e-llm-inference-service] kserve_client: KServeClient, [e2e-llm-inference-service] test_case: TestCase, # noqa: F811 [e2e-llm-inference-service] timeout_seconds: int = 900, [e2e-llm-inference-service] extra_headers: Optional[Dict[str, str]] = None, [e2e-llm-inference-service] ) -> str: [e2e-llm-inference-service] def get_successful_response(): [e2e-llm-inference-service] try: [e2e-llm-inference-service] if test_case.url_getter: [e2e-llm-inference-service] service_url = test_case.url_getter(kserve_client, test_case.llm_service) [e2e-llm-inference-service] else: [e2e-llm-inference-service] service_url = get_llm_service_url(kserve_client, test_case.llm_service) [e2e-llm-inference-service] except Exception as e: [e2e-llm-inference-service] raise AssertionError(f"❌ Failed to get service URL: {e}") from e [e2e-llm-inference-service] [e2e-llm-inference-service] model_url = service_url + test_case.endpoint [e2e-llm-inference-service] [e2e-llm-inference-service] headers = {"Content-Type": "application/json"} [e2e-llm-inference-service] if extra_headers: [e2e-llm-inference-service] headers.update(extra_headers) [e2e-llm-inference-service] [e2e-llm-inference-service] if test_case.payload_formatter is not None: [e2e-llm-inference-service] test_payload = test_case.payload_formatter(test_case) [e2e-llm-inference-service] elif test_case.prompt is not None: [e2e-llm-inference-service] test_payload = { [e2e-llm-inference-service] "model": test_case.model_name [e2e-llm-inference-service] if not extra_headers or MODEL_ROUTING_HEADER not in extra_headers [e2e-llm-inference-service] else extra_headers[MODEL_ROUTING_HEADER], [e2e-llm-inference-service] "prompt": test_case.prompt, [e2e-llm-inference-service] "max_tokens": test_case.max_tokens, [e2e-llm-inference-service] } [e2e-llm-inference-service] else: [e2e-llm-inference-service] test_payload = None [e2e-llm-inference-service] [e2e-llm-inference-service] logger.info(f"Calling LLM service at {model_url} with payload {test_payload}") [e2e-llm-inference-service] try: [e2e-llm-inference-service] if test_payload is not None: [e2e-llm-inference-service] response = post_with_retry( [e2e-llm-inference-service] model_url, [e2e-llm-inference-service] headers=headers, [e2e-llm-inference-service] json_data=test_payload, [e2e-llm-inference-service] timeout=test_case.response_timeout, [e2e-llm-inference-service] ) [e2e-llm-inference-service] else: [e2e-llm-inference-service] response = get_with_retry( [e2e-llm-inference-service] model_url, [e2e-llm-inference-service] headers=headers, [e2e-llm-inference-service] timeout=test_case.response_timeout, [e2e-llm-inference-service] ) [e2e-llm-inference-service] except Exception as e: [e2e-llm-inference-service] logger.error(f"❌ Failed to call model: {e}") [e2e-llm-inference-service] raise AssertionError(f"❌ Failed to call model: {e}") from e [e2e-llm-inference-service] [e2e-llm-inference-service] logger.info(f"Model response is {response.status_code}: {response.text[:500]}") [e2e-llm-inference-service] [e2e-llm-inference-service] if 200 <= response.status_code < 300: [e2e-llm-inference-service] return response [e2e-llm-inference-service] raise AssertionError( [e2e-llm-inference-service] f"Service returned {response.status_code}: {response.text}" [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] > response = wait_for(get_successful_response, timeout=timeout_seconds, interval=5.0) [e2e-llm-inference-service] [e2e-llm-inference-service] llmisvc/test_llm_inference_service.py:1119: [e2e-llm-inference-service] _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ [e2e-llm-inference-service] [e2e-llm-inference-service] assertion_fn = .get_successful_response at 0x7f4f0dde3e20> [e2e-llm-inference-service] timeout = 900, interval = 5.0 [e2e-llm-inference-service] [e2e-llm-inference-service] def wait_for( [e2e-llm-inference-service] assertion_fn: Callable[[], Any], timeout: float = 5.0, interval: float = 0.1 [e2e-llm-inference-service] ) -> Any: [e2e-llm-inference-service] """Wait for the assertion to succeed within timeout.""" [e2e-llm-inference-service] deadline = time.time() + timeout [e2e-llm-inference-service] last_msg = None [e2e-llm-inference-service] while True: [e2e-llm-inference-service] try: [e2e-llm-inference-service] > return assertion_fn() [e2e-llm-inference-service] [e2e-llm-inference-service] llmisvc/test_llm_inference_service.py:1215: [e2e-llm-inference-service] _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ [e2e-llm-inference-service] [e2e-llm-inference-service] def get_successful_response(): [e2e-llm-inference-service] try: [e2e-llm-inference-service] if test_case.url_getter: [e2e-llm-inference-service] service_url = test_case.url_getter(kserve_client, test_case.llm_service) [e2e-llm-inference-service] else: [e2e-llm-inference-service] service_url = get_llm_service_url(kserve_client, test_case.llm_service) [e2e-llm-inference-service] except Exception as e: [e2e-llm-inference-service] raise AssertionError(f"❌ Failed to get service URL: {e}") from e [e2e-llm-inference-service] [e2e-llm-inference-service] model_url = service_url + test_case.endpoint [e2e-llm-inference-service] [e2e-llm-inference-service] headers = {"Content-Type": "application/json"} [e2e-llm-inference-service] if extra_headers: [e2e-llm-inference-service] headers.update(extra_headers) [e2e-llm-inference-service] [e2e-llm-inference-service] if test_case.payload_formatter is not None: [e2e-llm-inference-service] test_payload = test_case.payload_formatter(test_case) [e2e-llm-inference-service] elif test_case.prompt is not None: [e2e-llm-inference-service] test_payload = { [e2e-llm-inference-service] "model": test_case.model_name [e2e-llm-inference-service] if not extra_headers or MODEL_ROUTING_HEADER not in extra_headers [e2e-llm-inference-service] else extra_headers[MODEL_ROUTING_HEADER], [e2e-llm-inference-service] "prompt": test_case.prompt, [e2e-llm-inference-service] "max_tokens": test_case.max_tokens, [e2e-llm-inference-service] } [e2e-llm-inference-service] else: [e2e-llm-inference-service] test_payload = None [e2e-llm-inference-service] [e2e-llm-inference-service] logger.info(f"Calling LLM service at {model_url} with payload {test_payload}") [e2e-llm-inference-service] try: [e2e-llm-inference-service] if test_payload is not None: [e2e-llm-inference-service] response = post_with_retry( [e2e-llm-inference-service] model_url, [e2e-llm-inference-service] headers=headers, [e2e-llm-inference-service] json_data=test_payload, [e2e-llm-inference-service] timeout=test_case.response_timeout, [e2e-llm-inference-service] ) [e2e-llm-inference-service] else: [e2e-llm-inference-service] response = get_with_retry( [e2e-llm-inference-service] model_url, [e2e-llm-inference-service] headers=headers, [e2e-llm-inference-service] timeout=test_case.response_timeout, [e2e-llm-inference-service] ) [e2e-llm-inference-service] except Exception as e: [e2e-llm-inference-service] logger.error(f"❌ Failed to call model: {e}") [e2e-llm-inference-service] > raise AssertionError(f"❌ Failed to call model: {e}") from e [e2e-llm-inference-service] E AssertionError: ❌ Failed to call model: HTTPConnectionPool(host='ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com', port=80): Max retries exceeded with url: /kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-route-87882a8e/v1/completions (Caused by ReadTimeoutError("HTTPConnectionPool(host='ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com', port=80): Read timed out. (read timeout=60)")) [e2e-llm-inference-service] [e2e-llm-inference-service] llmisvc/test_llm_inference_service.py:1109: AssertionError [e2e-llm-inference-service] ------------------------------ Captured log setup ------------------------------ [e2e-llm-inference-service] INFO e2e.llmisvc.logging:fixtures.py:1637 No HTTP proxy configured for k8s client [e2e-llm-inference-service] INFO e2e.llmisvc.logging:fixtures.py:1586 Checking LLMInferenceServiceConfig router-managed-llmisvc-model-fb-aee408e0 in namespace kserve-ci-e2e-test [e2e-llm-inference-service] INFO e2e.llmisvc.logging:fixtures.py:1612 Resource not found, creating LLMInferenceServiceConfig router-managed-llmisvc-model-fb-aee408e0 [e2e-llm-inference-service] INFO e2e.llmisvc.logging:fixtures.py:1622 ✓ Successfully created LLMInferenceServiceConfig router-managed-llmisvc-model-fb-aee408e0 [e2e-llm-inference-service] INFO e2e.llmisvc.logging:fixtures.py:1586 Checking LLMInferenceServiceConfig workload-single-cpu-llmisvc-mod-c7f60866 in namespace kserve-ci-e2e-test [e2e-llm-inference-service] INFO e2e.llmisvc.logging:fixtures.py:1612 Resource not found, creating LLMInferenceServiceConfig workload-single-cpu-llmisvc-mod-c7f60866 [e2e-llm-inference-service] INFO e2e.llmisvc.logging:fixtures.py:1622 ✓ Successfully created LLMInferenceServiceConfig workload-single-cpu-llmisvc-mod-c7f60866 [e2e-llm-inference-service] INFO e2e.llmisvc.logging:fixtures.py:1586 Checking LLMInferenceServiceConfig model-fb-opt-125m-llmisvc-model-2a331009 in namespace kserve-ci-e2e-test [e2e-llm-inference-service] INFO e2e.llmisvc.logging:fixtures.py:1612 Resource not found, creating LLMInferenceServiceConfig model-fb-opt-125m-llmisvc-model-2a331009 [e2e-llm-inference-service] INFO e2e.llmisvc.logging:fixtures.py:1622 ✓ Successfully created LLMInferenceServiceConfig model-fb-opt-125m-llmisvc-model-2a331009 [e2e-llm-inference-service] ------------------------------ Captured log call ------------------------------- [e2e-llm-inference-service] INFO e2e.llmisvc.logging:logging.py:34 [test_llm_inference_service] [2026-07-07T14:31:35.834557] start - args=(), kwargs={'test_case': TestCase(base_refs=['router-managed', 'workload-single-cpu', 'model-fb-opt-125m'], prompt='KServe is a', service_name='llmisvc-model-fb-opt-125m-route-87882a8e', endpoint='/v1/completions', max_tokens=20, payload_formatter=, response_assertion=, wait_timeout=900, response_timeout=60, extra_headers=None, url_getter=None, expected_gateway=None, before_test=[], after_test=[], peers=[], llm_service={'api_version': 'serving.kserve.io/v1alpha1', [e2e-llm-inference-service] 'kind': 'LLMInferenceService', [e2e-llm-inference-service] 'metadata': {'annotations': None, [e2e-llm-inference-service] 'creation_timestamp': None, [e2e-llm-inference-service] 'deletion_grace_period_seconds': None, [e2e-llm-inference-service] 'deletion_timestamp': None, [e2e-llm-inference-service] 'finalizers': None, [e2e-llm-inference-service] 'generate_name': None, [e2e-llm-inference-service] 'generation': None, [e2e-llm-inference-service] 'labels': None, [e2e-llm-inference-service] 'managed_fields': None, [e2e-llm-inference-service] 'name': 'llmisvc-model-fb-opt-125m-route-87882a8e', [e2e-llm-inference-service] 'namespace': 'kserve-ci-e2e-test', [e2e-llm-inference-service] 'owner_references': None, [e2e-llm-inference-service] 'resource_version': None, [e2e-llm-inference-service] 'self_link': None, [e2e-llm-inference-service] 'uid': None}, [e2e-llm-inference-service] 'spec': {'baseRefs': [{'name': 'router-managed-llmisvc-model-fb-aee408e0'}, [e2e-llm-inference-service] {'name': 'workload-single-cpu-llmisvc-mod-c7f60866'}, [e2e-llm-inference-service] {'name': 'model-fb-opt-125m-llmisvc-model-2a331009'}]}, [e2e-llm-inference-service] 'status': None}, model_name='facebook/opt-125m')} [e2e-llm-inference-service] INFO e2e.llmisvc.logging:fixtures.py:1637 No HTTP proxy configured for k8s client [e2e-llm-inference-service] INFO e2e.llmisvc.logging:logging.py:34 [create_llmisvc] [2026-07-07T14:31:35.847212] start - args=(, {'api_version': 'serving.kserve.io/v1alpha1', [e2e-llm-inference-service] 'kind': 'LLMInferenceService', [e2e-llm-inference-service] 'metadata': {'annotations': {'security.opendatahub.io/enable-auth': 'false'}, [e2e-llm-inference-service] 'creation_timestamp': None, [e2e-llm-inference-service] 'deletion_grace_period_seconds': None, [e2e-llm-inference-service] 'deletion_timestamp': None, [e2e-llm-inference-service] 'finalizers': None, [e2e-llm-inference-service] 'generate_name': None, [e2e-llm-inference-service] 'generation': None, [e2e-llm-inference-service] 'labels': None, [e2e-llm-inference-service] 'managed_fields': None, [e2e-llm-inference-service] 'name': 'llmisvc-model-fb-opt-125m-route-87882a8e', [e2e-llm-inference-service] 'namespace': 'kserve-ci-e2e-test', [e2e-llm-inference-service] 'owner_references': None, [e2e-llm-inference-service] 'resource_version': None, [e2e-llm-inference-service] 'self_link': None, [e2e-llm-inference-service] 'uid': None}, [e2e-llm-inference-service] 'spec': {'baseRefs': [{'name': 'router-managed-llmisvc-model-fb-aee408e0'}, [e2e-llm-inference-service] {'name': 'workload-single-cpu-llmisvc-mod-c7f60866'}, [e2e-llm-inference-service] {'name': 'model-fb-opt-125m-llmisvc-model-2a331009'}]}, [e2e-llm-inference-service] 'status': None}), kwargs={} [e2e-llm-inference-service] INFO e2e.llmisvc.logging:logging.py:43 [create_llmisvc] [2026-07-07T14:31:35.905915] end - ✅ in 0.058s [e2e-llm-inference-service] INFO e2e.llmisvc.logging:logging.py:34 [wait_for_llm_isvc_ready] [2026-07-07T14:31:35.906001] start - args=(, {'api_version': 'serving.kserve.io/v1alpha1', [e2e-llm-inference-service] 'kind': 'LLMInferenceService', [e2e-llm-inference-service] 'metadata': {'annotations': {'security.opendatahub.io/enable-auth': 'false'}, [e2e-llm-inference-service] 'creation_timestamp': None, [e2e-llm-inference-service] 'deletion_grace_period_seconds': None, [e2e-llm-inference-service] 'deletion_timestamp': None, [e2e-llm-inference-service] 'finalizers': None, [e2e-llm-inference-service] 'generate_name': None, [e2e-llm-inference-service] 'generation': None, [e2e-llm-inference-service] 'labels': None, [e2e-llm-inference-service] 'managed_fields': None, [e2e-llm-inference-service] 'name': 'llmisvc-model-fb-opt-125m-route-87882a8e', [e2e-llm-inference-service] 'namespace': 'kserve-ci-e2e-test', [e2e-llm-inference-service] 'owner_references': None, [e2e-llm-inference-service] 'resource_version': None, [e2e-llm-inference-service] 'self_link': None, [e2e-llm-inference-service] 'uid': None}, [e2e-llm-inference-service] 'spec': {'baseRefs': [{'name': 'router-managed-llmisvc-model-fb-aee408e0'}, [e2e-llm-inference-service] {'name': 'workload-single-cpu-llmisvc-mod-c7f60866'}, [e2e-llm-inference-service] {'name': 'model-fb-opt-125m-llmisvc-model-2a331009'}]}, [e2e-llm-inference-service] 'status': None}, 900), kwargs={} [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1222 Waiting: No conditions found in status [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1222 Waiting: Missing true conditions: {'Ready', 'RouterReady', 'WorkloadsReady'}, expected {'Ready', 'RouterReady', 'WorkloadsReady'}, got [{'lastTransitionTime': '2026-07-07T14:31:42Z', 'message': 'The following HTTPRoutes are not ready: [kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-route-87882a8e-kserve-route: "False" (reason "InvalidKind", message "referencing unsupported backendRef: group \\"inference.networking.x-k8s.io\\" kind \\"InferencePool\\"")]', 'reason': 'HTTPRoutesNotReady', 'severity': 'Info', 'status': 'False', 'type': 'HTTPRoutesReady'}, {'lastTransitionTime': '2026-07-07T14:31:42Z', 'message': 'Inference Pool kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-route-87882a8e-inference-pool exists but no Gateway controller has accepted it yet', 'reason': 'WaitingForGateway', 'severity': 'Info', 'status': 'False', 'type': 'InferencePoolReady'}, {'lastTransitionTime': '2026-07-07T14:31:42Z', 'reason': 'Progressing', 'severity': 'Info', 'status': 'False', 'type': 'MainWorkloadReady'}, {'lastTransitionTime': '2026-07-07T14:31:42Z', 'severity': 'Info', 'status': 'True', 'type': 'PresetsCombined'}, {'lastTransitionTime': '2026-07-07T14:31:42Z', 'message': 'The following HTTPRoutes are not ready: [kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-route-87882a8e-kserve-route: "False" (reason "InvalidKind", message "referencing unsupported backendRef: group \\"inference.networking.x-k8s.io\\" kind \\"InferencePool\\"")]', 'reason': 'HTTPRoutesNotReady', 'status': 'False', 'type': 'Ready'}, {'lastTransitionTime': '2026-07-07T14:31:42Z', 'message': 'The following HTTPRoutes are not ready: [kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-route-87882a8e-kserve-route: "False" (reason "InvalidKind", message "referencing unsupported backendRef: group \\"inference.networking.x-k8s.io\\" kind \\"InferencePool\\"")]', 'reason': 'HTTPRoutesNotReady', 'status': 'False', 'type': 'RouterReady'}, {'lastTransitionTime': '2026-07-07T14:31:42Z', 'message': 'Deployment rollout in progress', 'reason': 'Progressing', 'severity': 'Info', 'status': 'False', 'type': 'SchedulerWorkloadReady'}, {'lastTransitionTime': '2026-07-07T14:31:42Z', 'reason': 'Progressing', 'status': 'False', 'type': 'WorkloadsReady'}] [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1222 Waiting: Missing true conditions: {'Ready', 'RouterReady', 'WorkloadsReady'}, expected {'Ready', 'RouterReady', 'WorkloadsReady'}, got [{'lastTransitionTime': '2026-07-07T14:31:47Z', 'severity': 'Info', 'status': 'True', 'type': 'HTTPRoutesReady'}, {'lastTransitionTime': '2026-07-07T14:31:47Z', 'severity': 'Info', 'status': 'True', 'type': 'InferencePoolReady'}, {'lastTransitionTime': '2026-07-07T14:31:47Z', 'message': 'Deployment does not have minimum availability.', 'reason': 'MinimumReplicasUnavailable', 'severity': 'Info', 'status': 'False', 'type': 'MainWorkloadReady'}, {'lastTransitionTime': '2026-07-07T14:31:42Z', 'severity': 'Info', 'status': 'True', 'type': 'PresetsCombined'}, {'lastTransitionTime': '2026-07-07T14:31:47Z', 'message': 'Deployment does not have minimum availability.', 'reason': 'MinimumReplicasUnavailable', 'status': 'False', 'type': 'Ready'}, {'lastTransitionTime': '2026-07-07T14:31:47Z', 'message': 'Deployment does not have minimum availability.', 'reason': 'MinimumReplicasUnavailable', 'status': 'False', 'type': 'RouterReady'}, {'lastTransitionTime': '2026-07-07T14:31:47Z', 'message': 'Deployment does not have minimum availability.', 'reason': 'MinimumReplicasUnavailable', 'severity': 'Info', 'status': 'False', 'type': 'SchedulerWorkloadReady'}, {'lastTransitionTime': '2026-07-07T14:31:47Z', 'message': 'Deployment does not have minimum availability.', 'reason': 'MinimumReplicasUnavailable', 'status': 'False', 'type': 'WorkloadsReady'}] [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1222 Waiting: Missing true conditions: {'Ready', 'WorkloadsReady'}, expected {'Ready', 'RouterReady', 'WorkloadsReady'}, got [{'lastTransitionTime': '2026-07-07T14:31:47Z', 'severity': 'Info', 'status': 'True', 'type': 'HTTPRoutesReady'}, {'lastTransitionTime': '2026-07-07T14:31:47Z', 'severity': 'Info', 'status': 'True', 'type': 'InferencePoolReady'}, {'lastTransitionTime': '2026-07-07T14:31:47Z', 'message': 'Deployment does not have minimum availability.', 'reason': 'MinimumReplicasUnavailable', 'severity': 'Info', 'status': 'False', 'type': 'MainWorkloadReady'}, {'lastTransitionTime': '2026-07-07T14:31:42Z', 'severity': 'Info', 'status': 'True', 'type': 'PresetsCombined'}, {'lastTransitionTime': '2026-07-07T14:31:47Z', 'message': 'Deployment does not have minimum availability.', 'reason': 'MinimumReplicasUnavailable', 'status': 'False', 'type': 'Ready'}, {'lastTransitionTime': '2026-07-07T14:32:15Z', 'status': 'True', 'type': 'RouterReady'}, {'lastTransitionTime': '2026-07-07T14:32:15Z', 'severity': 'Info', 'status': 'True', 'type': 'SchedulerWorkloadReady'}, {'lastTransitionTime': '2026-07-07T14:31:47Z', 'message': 'Deployment does not have minimum availability.', 'reason': 'MinimumReplicasUnavailable', 'status': 'False', 'type': 'WorkloadsReady'}] [e2e-llm-inference-service] INFO e2e.llmisvc.logging:logging.py:43 [wait_for_llm_isvc_ready] [2026-07-07T14:33:41.745112] end - ✅ in 125.839s [e2e-llm-inference-service] INFO e2e.llmisvc.logging:logging.py:34 [wait_for_model_response] [2026-07-07T14:33:41.745372] start - args=(, TestCase(base_refs=['router-managed', 'workload-single-cpu', 'model-fb-opt-125m'], prompt='KServe is a', service_name='llmisvc-model-fb-opt-125m-route-87882a8e', endpoint='/v1/completions', max_tokens=20, payload_formatter=, response_assertion=, wait_timeout=900, response_timeout=60, extra_headers=None, url_getter=None, expected_gateway=None, before_test=[], after_test=[], peers=[], llm_service={'api_version': 'serving.kserve.io/v1alpha1', [e2e-llm-inference-service] 'kind': 'LLMInferenceService', [e2e-llm-inference-service] 'metadata': {'annotations': {'security.opendatahub.io/enable-auth': 'false'}, [e2e-llm-inference-service] 'creation_timestamp': None, [e2e-llm-inference-service] 'deletion_grace_period_seconds': None, [e2e-llm-inference-service] 'deletion_timestamp': None, [e2e-llm-inference-service] 'finalizers': None, [e2e-llm-inference-service] 'generate_name': None, [e2e-llm-inference-service] 'generation': None, [e2e-llm-inference-service] 'labels': None, [e2e-llm-inference-service] 'managed_fields': None, [e2e-llm-inference-service] 'name': 'llmisvc-model-fb-opt-125m-route-87882a8e', [e2e-llm-inference-service] 'namespace': 'kserve-ci-e2e-test', [e2e-llm-inference-service] 'owner_references': None, [e2e-llm-inference-service] 'resource_version': None, [e2e-llm-inference-service] 'self_link': None, [e2e-llm-inference-service] 'uid': None}, [e2e-llm-inference-service] 'spec': {'baseRefs': [{'name': 'router-managed-llmisvc-model-fb-aee408e0'}, [e2e-llm-inference-service] {'name': 'workload-single-cpu-llmisvc-mod-c7f60866'}, [e2e-llm-inference-service] {'name': 'model-fb-opt-125m-llmisvc-model-2a331009'}]}, [e2e-llm-inference-service] 'status': None}, model_name='facebook/opt-125m'), 900), kwargs={'extra_headers': None} [e2e-llm-inference-service] INFO e2e.llmisvc.logging:logging.py:34 [get_llm_service_url] [2026-07-07T14:33:41.745778] start - args=(, {'api_version': 'serving.kserve.io/v1alpha1', [e2e-llm-inference-service] 'kind': 'LLMInferenceService', [e2e-llm-inference-service] 'metadata': {'annotations': {'security.opendatahub.io/enable-auth': 'false'}, [e2e-llm-inference-service] 'creation_timestamp': None, [e2e-llm-inference-service] 'deletion_grace_period_seconds': None, [e2e-llm-inference-service] 'deletion_timestamp': None, [e2e-llm-inference-service] 'finalizers': None, [e2e-llm-inference-service] 'generate_name': None, [e2e-llm-inference-service] 'generation': None, [e2e-llm-inference-service] 'labels': None, [e2e-llm-inference-service] 'managed_fields': None, [e2e-llm-inference-service] 'name': 'llmisvc-model-fb-opt-125m-route-87882a8e', [e2e-llm-inference-service] 'namespace': 'kserve-ci-e2e-test', [e2e-llm-inference-service] 'owner_references': None, [e2e-llm-inference-service] 'resource_version': None, [e2e-llm-inference-service] 'self_link': None, [e2e-llm-inference-service] 'uid': None}, [e2e-llm-inference-service] 'spec': {'baseRefs': [{'name': 'router-managed-llmisvc-model-fb-aee408e0'}, [e2e-llm-inference-service] {'name': 'workload-single-cpu-llmisvc-mod-c7f60866'}, [e2e-llm-inference-service] {'name': 'model-fb-opt-125m-llmisvc-model-2a331009'}]}, [e2e-llm-inference-service] 'status': None}), kwargs={} [e2e-llm-inference-service] INFO e2e.llmisvc.logging:logging.py:43 [get_llm_service_url] [2026-07-07T14:33:41.758367] end - ✅ in 0.012s [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1092 Calling LLM service at http://ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com/kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-route-87882a8e/v1/completions with payload {'model': 'facebook/opt-125m', 'prompt': 'KServe is a', 'max_tokens': 20} [e2e-llm-inference-service] WARNING urllib3.connectionpool:connectionpool.py:868 Retrying (Retry(total=7, connect=None, read=None, redirect=None, status=None)) after connection broken by 'ReadTimeoutError("HTTPConnectionPool(host='ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com', port=80): Read timed out. (read timeout=60)")': /kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-route-87882a8e/v1/completions [e2e-llm-inference-service] WARNING urllib3.connectionpool:connectionpool.py:868 Retrying (Retry(total=6, connect=None, read=None, redirect=None, status=None)) after connection broken by 'ReadTimeoutError("HTTPConnectionPool(host='ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com', port=80): Read timed out. (read timeout=60)")': /kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-route-87882a8e/v1/completions [e2e-llm-inference-service] WARNING urllib3.connectionpool:connectionpool.py:868 Retrying (Retry(total=5, connect=None, read=None, redirect=None, status=None)) after connection broken by 'ReadTimeoutError("HTTPConnectionPool(host='ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com', port=80): Read timed out. (read timeout=60)")': /kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-route-87882a8e/v1/completions [e2e-llm-inference-service] WARNING urllib3.connectionpool:connectionpool.py:868 Retrying (Retry(total=4, connect=None, read=None, redirect=None, status=None)) after connection broken by 'ReadTimeoutError("HTTPConnectionPool(host='ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com', port=80): Read timed out. (read timeout=60)")': /kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-route-87882a8e/v1/completions [e2e-llm-inference-service] WARNING urllib3.connectionpool:connectionpool.py:868 Retrying (Retry(total=3, connect=None, read=None, redirect=None, status=None)) after connection broken by 'ReadTimeoutError("HTTPConnectionPool(host='ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com', port=80): Read timed out. (read timeout=60)")': /kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-route-87882a8e/v1/completions [e2e-llm-inference-service] WARNING urllib3.connectionpool:connectionpool.py:868 Retrying (Retry(total=2, connect=None, read=None, redirect=None, status=None)) after connection broken by 'RemoteDisconnected('Remote end closed connection without response')': /kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-route-87882a8e/v1/completions [e2e-llm-inference-service] WARNING urllib3.connectionpool:connectionpool.py:868 Retrying (Retry(total=1, connect=None, read=None, redirect=None, status=None)) after connection broken by 'ReadTimeoutError("HTTPConnectionPool(host='ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com', port=80): Read timed out. (read timeout=60)")': /kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-route-87882a8e/v1/completions [e2e-llm-inference-service] WARNING urllib3.connectionpool:connectionpool.py:868 Retrying (Retry(total=0, connect=None, read=None, redirect=None, status=None)) after connection broken by 'ReadTimeoutError("HTTPConnectionPool(host='ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com', port=80): Read timed out. (read timeout=60)")': /kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-route-87882a8e/v1/completions [e2e-llm-inference-service] ERROR e2e.llmisvc.logging:test_llm_inference_service.py:1108 ❌ Failed to call model: HTTPConnectionPool(host='ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com', port=80): Max retries exceeded with url: /kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-route-87882a8e/v1/completions (Caused by ReadTimeoutError("HTTPConnectionPool(host='ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com', port=80): Read timed out. (read timeout=60)")) [e2e-llm-inference-service] ERROR e2e.llmisvc.logging:test_llm_inference_service.py:1219 Timed out waiting: ❌ Failed to call model: HTTPConnectionPool(host='ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com', port=80): Max retries exceeded with url: /kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-route-87882a8e/v1/completions (Caused by ReadTimeoutError("HTTPConnectionPool(host='ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com', port=80): Read timed out. (read timeout=60)")) [e2e-llm-inference-service] ERROR e2e.llmisvc.logging:logging.py:48 [wait_for_model_response] [2026-07-07T14:48:46.297689] end - ❌ 904.552s: ❌ Failed to call model: HTTPConnectionPool(host='ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com', port=80): Max retries exceeded with url: /kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-route-87882a8e/v1/completions (Caused by ReadTimeoutError("HTTPConnectionPool(host='ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com', port=80): Read timed out. (read timeout=60)")) [e2e-llm-inference-service] ERROR e2e.llmisvc.logging:test_llm_inference_service.py:831 [router-managed-workload-single-cpu-model-fb-opt-125m] ❌ ERROR: Failed to call llm inference service llmisvc-model-fb-opt-125m-route-87882a8e: ❌ Failed to call model: HTTPConnectionPool(host='ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com', port=80): Max retries exceeded with url: /kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-route-87882a8e/v1/completions (Caused by ReadTimeoutError("HTTPConnectionPool(host='ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com', port=80): Read timed out. (read timeout=60)")) [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1240 🔍 # Diagnostics for 'llmisvc-model-fb-opt-125m-route-87882a8e' in 'kserve-ci-e2e-test' [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1241 --- [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1242 # LLMInferenceService llmisvc-model-fb-opt-125m-route-87882a8e [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1245 apiVersion: serving.kserve.io/v1alpha1 [e2e-llm-inference-service] kind: LLMInferenceService [e2e-llm-inference-service] metadata: [e2e-llm-inference-service] annotations: [e2e-llm-inference-service] security.opendatahub.io/enable-auth: 'false' [e2e-llm-inference-service] creationTimestamp: '2026-07-07T14:31:35Z' [e2e-llm-inference-service] finalizers: [e2e-llm-inference-service] - serving.kserve.io/llmisvc-finalizer [e2e-llm-inference-service] generation: 1 [e2e-llm-inference-service] managedFields: [e2e-llm-inference-service] - apiVersion: serving.kserve.io/v1alpha1 [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:metadata: [e2e-llm-inference-service] f:annotations: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:security.opendatahub.io/enable-auth: {} [e2e-llm-inference-service] f:spec: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:baseRefs: {} [e2e-llm-inference-service] manager: OpenAPI-Generator [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] time: '2026-07-07T14:31:35Z' [e2e-llm-inference-service] - apiVersion: serving.kserve.io/v1alpha2 [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:metadata: [e2e-llm-inference-service] f:finalizers: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] v:"serving.kserve.io/llmisvc-finalizer": {} [e2e-llm-inference-service] manager: manager [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] time: '2026-07-07T14:31:35Z' [e2e-llm-inference-service] - apiVersion: serving.kserve.io/v1alpha2 [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:status: [e2e-llm-inference-service] f:addresses: {} [e2e-llm-inference-service] f:annotations: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:serving.kserve.io/config-llm-decode-template: {} [e2e-llm-inference-service] f:serving.kserve.io/config-llm-decode-worker-data-parallel: {} [e2e-llm-inference-service] f:serving.kserve.io/config-llm-prefill-template: {} [e2e-llm-inference-service] f:serving.kserve.io/config-llm-prefill-worker-data-parallel: {} [e2e-llm-inference-service] f:serving.kserve.io/config-llm-router-route: {} [e2e-llm-inference-service] f:serving.kserve.io/config-llm-scheduler: {} [e2e-llm-inference-service] f:serving.kserve.io/config-llm-template: {} [e2e-llm-inference-service] f:serving.kserve.io/config-llm-tracing: {} [e2e-llm-inference-service] f:serving.kserve.io/config-llm-worker-data-parallel: {} [e2e-llm-inference-service] f:appliedConfigs: {} [e2e-llm-inference-service] f:conditions: {} [e2e-llm-inference-service] f:observedGeneration: {} [e2e-llm-inference-service] f:router: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:gateways: {} [e2e-llm-inference-service] f:scheduler: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:inferencePool: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:group: {} [e2e-llm-inference-service] f:kind: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:service: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:group: {} [e2e-llm-inference-service] f:kind: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:url: {} [e2e-llm-inference-service] f:workloads: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:primary: {} [e2e-llm-inference-service] f:scheduler: {} [e2e-llm-inference-service] f:service: {} [e2e-llm-inference-service] manager: manager [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] subresource: status [e2e-llm-inference-service] time: '2026-07-07T14:33:40Z' [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-route-87882a8e [e2e-llm-inference-service] namespace: kserve-ci-e2e-test [e2e-llm-inference-service] resourceVersion: '33116' [e2e-llm-inference-service] uid: 88c60d0e-95bd-401f-ae1c-2c4c0b1e7c9d [e2e-llm-inference-service] spec: [e2e-llm-inference-service] baseRefs: [e2e-llm-inference-service] - name: router-managed-llmisvc-model-fb-aee408e0 [e2e-llm-inference-service] - name: workload-single-cpu-llmisvc-mod-c7f60866 [e2e-llm-inference-service] - name: model-fb-opt-125m-llmisvc-model-2a331009 [e2e-llm-inference-service] model: [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-route-87882a8e [e2e-llm-inference-service] uri: '' [e2e-llm-inference-service] status: [e2e-llm-inference-service] addresses: [e2e-llm-inference-service] - name: gateway-external-model-routing [e2e-llm-inference-service] url: http://ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com/ [e2e-llm-inference-service] - name: gateway-external [e2e-llm-inference-service] url: http://ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com/kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-route-87882a8e [e2e-llm-inference-service] - name: gateway-internal-model-routing [e2e-llm-inference-service] url: http://openshift-ai-inference-openshift-default.openshift-ingress.svc.cluster.local/ [e2e-llm-inference-service] - name: gateway-internal [e2e-llm-inference-service] url: http://openshift-ai-inference-openshift-default.openshift-ingress.svc.cluster.local/kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-route-87882a8e [e2e-llm-inference-service] annotations: [e2e-llm-inference-service] serving.kserve.io/config-llm-decode-template: kserve-config-llm-decode-template [e2e-llm-inference-service] serving.kserve.io/config-llm-decode-worker-data-parallel: kserve-config-llm-decode-worker-data-parallel [e2e-llm-inference-service] serving.kserve.io/config-llm-prefill-template: kserve-config-llm-prefill-template [e2e-llm-inference-service] serving.kserve.io/config-llm-prefill-worker-data-parallel: kserve-config-llm-prefill-worker-data-parallel [e2e-llm-inference-service] serving.kserve.io/config-llm-router-route: kserve-config-llm-router-route [e2e-llm-inference-service] serving.kserve.io/config-llm-scheduler: kserve-config-llm-scheduler [e2e-llm-inference-service] serving.kserve.io/config-llm-template: kserve-config-llm-template [e2e-llm-inference-service] serving.kserve.io/config-llm-tracing: kserve-config-llm-tracing [e2e-llm-inference-service] serving.kserve.io/config-llm-worker-data-parallel: kserve-config-llm-worker-data-parallel [e2e-llm-inference-service] conditions: [e2e-llm-inference-service] - lastTransitionTime: '2026-07-07T14:31:47Z' [e2e-llm-inference-service] severity: Info [e2e-llm-inference-service] status: 'True' [e2e-llm-inference-service] type: HTTPRoutesReady [e2e-llm-inference-service] - lastTransitionTime: '2026-07-07T14:31:47Z' [e2e-llm-inference-service] severity: Info [e2e-llm-inference-service] status: 'True' [e2e-llm-inference-service] type: InferencePoolReady [e2e-llm-inference-service] - lastTransitionTime: '2026-07-07T14:33:40Z' [e2e-llm-inference-service] severity: Info [e2e-llm-inference-service] status: 'True' [e2e-llm-inference-service] type: MainWorkloadReady [e2e-llm-inference-service] - lastTransitionTime: '2026-07-07T14:31:42Z' [e2e-llm-inference-service] severity: Info [e2e-llm-inference-service] status: 'True' [e2e-llm-inference-service] type: PresetsCombined [e2e-llm-inference-service] - lastTransitionTime: '2026-07-07T14:33:40Z' [e2e-llm-inference-service] status: 'True' [e2e-llm-inference-service] type: Ready [e2e-llm-inference-service] - lastTransitionTime: '2026-07-07T14:32:15Z' [e2e-llm-inference-service] status: 'True' [e2e-llm-inference-service] type: RouterReady [e2e-llm-inference-service] - lastTransitionTime: '2026-07-07T14:32:15Z' [e2e-llm-inference-service] severity: Info [e2e-llm-inference-service] status: 'True' [e2e-llm-inference-service] type: SchedulerWorkloadReady [e2e-llm-inference-service] - lastTransitionTime: '2026-07-07T14:33:40Z' [e2e-llm-inference-service] status: 'True' [e2e-llm-inference-service] type: WorkloadsReady [e2e-llm-inference-service] observedGeneration: 1 [e2e-llm-inference-service] url: http://ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com/kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-route-87882a8e [e2e-llm-inference-service] [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:44 TIME NAMESPACE SOURCE TYPE REASON MESSAGE [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:45 -------------------------------------------------------------------------------------------------- [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 None kserve-ci-e2e-test Normal Scheduled Successfully assigned kserve-ci-e2e-test/auth-disabled-test-kserve-658fdbccfc-7k7fm to ip-10-0-139-47.ec2.internal [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:28:27 kserve-ci-e2e-test multus Normal AddedInterface Add eth0 [10.134.0.29/23] from ovn-kubernetes [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:28:28 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Pulled Container image "quay.io/opendatahub/kserve-storage-initializer@sha256:13ad765c0f323200e1db7b45c99661287e65c5158d6b54b16139829636d20be4" already present on machine [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:28:28 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Created Created container: storage-initializer [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:28:28 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Started Started container storage-initializer [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:28:31 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Pulled Container image "public.ecr.aws/q9t5s3a7/vllm-cpu-release-repo:v0.19.0" already present on machine [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:28:32 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Created Created container: main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:28:32 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Started Started container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:30:17 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Warning Unhealthy Startup probe failed: Get "https://10.134.0.29:8000/health": dial tcp 10.134.0.29:8000: connect: connection refused [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:30:32 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Killing Stopping container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:30:47 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Warning Unhealthy Readiness probe failed: Get "https://10.134.0.29:8000/health": dial tcp 10.134.0.29:8000: connect: connection refused [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:28:27 kserve-ci-e2e-test replicaset-controller Normal SuccessfulCreate Created pod: auth-disabled-test-kserve-658fdbccfc-7k7fm [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 None kserve-ci-e2e-test Normal Scheduled Successfully assigned kserve-ci-e2e-test/auth-disabled-test-kserve-router-scheduler-58c676b848-mfsnv to ip-10-0-133-25.ec2.internal [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:28:28 kserve-ci-e2e-test multus Normal AddedInterface Add eth0 [10.132.0.51/23] from ovn-kubernetes [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:28:28 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Pulled Container image "quay.io/opendatahub/kserve-storage-initializer@sha256:13ad765c0f323200e1db7b45c99661287e65c5158d6b54b16139829636d20be4" already present on machine [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:28:28 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Created Created container: storage-initializer [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:28:28 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Started Started container storage-initializer [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:28:29 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Pulled Container image "ghcr.io/llm-d/llm-d-router-endpoint-picker:v0.9.0-rc.2" already present on machine [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:28:29 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Created Created container: main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:28:29 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Started Started container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:30:32 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Killing Stopping container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:30:32 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Warning FailedPreStopHook PreStopHook failed [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:28:28 kserve-ci-e2e-test replicaset-controller Normal SuccessfulCreate Created pod: auth-disabled-test-kserve-router-scheduler-58c676b848-mfsnv [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:28:27 kserve-ci-e2e-test deployment-controller Normal ScalingReplicaSet Scaled up replica set auth-disabled-test-kserve-router-scheduler-58c676b848 from 0 to 1 [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:28:27 kserve-ci-e2e-test deployment-controller Normal ScalingReplicaSet Scaled up replica set auth-disabled-test-kserve-658fdbccfc from 0 to 1 [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:28:27 kserve-ci-e2e-test OpenDataHubModelController Warning ReconcileError Failed to reconcile LLMInferenceService: 1 error occurred: * failed to get HTTPRoute for AuthPolicy auth-disabled-test-kserve-route-authn: failed to get HTTPRoute kserve-ci-e2e-test/auth-disabled-test-kserve-route: HTTPRoute.gateway.networking.k8s.io "auth-disabled-test-kserve-route" not found [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:28:27 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Secret kserve-ci-e2e-test/auth-disabled-test-kserve-self-signed-certs [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:28:27 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Deployment kserve-ci-e2e-test/auth-disabled-test-kserve [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:28:27 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Service kserve-ci-e2e-test/auth-disabled-test-kserve-workload-svc [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:28:27 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.ServiceAccount kserve-ci-e2e-test/auth-disabled-test-epp-sa [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:28:27 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.ClusterRoleBinding /kserve-ci-e2e-test-auth-disabled-test-epp-auth-rb [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:28:27 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Role kserve-ci-e2e-test/auth-disabled-test-epp-role [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:28:27 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.RoleBinding kserve-ci-e2e-test/auth-disabled-test-epp-rb [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:28:27 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Deployment kserve-ci-e2e-test/auth-disabled-test-kserve-router-scheduler [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:28:27 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Service kserve-ci-e2e-test/auth-disabled-test-epp-service [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:28:38 kserve-ci-e2e-test LLMInferenceServiceController Normal Created (combined from similar events): Created v1.DestinationRule kserve-ci-e2e-test/auth-disabled-test-kserve-shadow-svc [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:28:38 kserve-ci-e2e-test LLMInferenceServiceController Normal Updated Updated v1.Secret kserve-ci-e2e-test/auth-disabled-test-kserve-self-signed-certs [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:28:38 kserve-ci-e2e-test LLMInferenceServiceController Normal Updated Updated v1.HTTPRoute kserve-ci-e2e-test/auth-disabled-test-kserve-route [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:30:28 kserve-ci-e2e-test LLMInferenceServiceController Normal LLMInferenceServiceReady LLMInferenceService [auth-disabled-test] is Ready [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:30:32 kserve-ci-e2e-test LLMInferenceServiceController Normal Deleted Deleted v1.ClusterRoleBinding /kserve-ci-e2e-test-auth-disabled-test-epp-auth-rb [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 None kserve-ci-e2e-test Normal Scheduled Successfully assigned kserve-ci-e2e-test/auth-enabled-test-kserve-87bd45bf5-zxvsp to ip-10-0-139-47.ec2.internal [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:33 kserve-ci-e2e-test multus Normal AddedInterface Add eth0 [10.134.0.24/23] from ovn-kubernetes [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:33 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Pulled Container image "quay.io/opendatahub/kserve-storage-initializer@sha256:13ad765c0f323200e1db7b45c99661287e65c5158d6b54b16139829636d20be4" already present on machine [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:33 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Created Created container: storage-initializer [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:33 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Started Started container storage-initializer [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:39 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Pulling Pulling image "public.ecr.aws/q9t5s3a7/vllm-cpu-release-repo:v0.19.0" [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:23:08 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Pulled Successfully pulled image "public.ecr.aws/q9t5s3a7/vllm-cpu-release-repo:v0.19.0" in 28.615s (28.615s including waiting). Image size: 3531177328 bytes. [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:23:08 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Created Created container: main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:23:08 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Started Started container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:24:53 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Warning Unhealthy Startup probe failed: Get "https://10.134.0.24:8000/health": dial tcp 10.134.0.24:8000: connect: connection refused [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:10 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Killing Stopping container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:25 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Warning Unhealthy Readiness probe failed: Get "https://10.134.0.24:8000/health": dial tcp 10.134.0.24:8000: connect: connection refused [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:32 kserve-ci-e2e-test replicaset-controller Normal SuccessfulCreate Created pod: auth-enabled-test-kserve-87bd45bf5-zxvsp [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 None kserve-ci-e2e-test Normal Scheduled Successfully assigned kserve-ci-e2e-test/auth-enabled-test-kserve-router-scheduler-6bbc585b87-nf5lr to ip-10-0-133-25.ec2.internal [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:33 kserve-ci-e2e-test multus Normal AddedInterface Add eth0 [10.132.0.47/23] from ovn-kubernetes [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:33 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Pulled Container image "quay.io/opendatahub/kserve-storage-initializer@sha256:13ad765c0f323200e1db7b45c99661287e65c5158d6b54b16139829636d20be4" already present on machine [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:33 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Created Created container: storage-initializer [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:33 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Started Started container storage-initializer [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:34 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Pulling Pulling image "ghcr.io/llm-d/llm-d-router-endpoint-picker:v0.9.0-rc.2" [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:35 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Pulled Successfully pulled image "ghcr.io/llm-d/llm-d-router-endpoint-picker:v0.9.0-rc.2" in 1.131s (1.131s including waiting). Image size: 75073927 bytes. [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:35 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Created Created container: main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:35 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Started Started container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:10 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Killing Stopping container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:10 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Warning FailedPreStopHook PreStopHook failed [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:33 kserve-ci-e2e-test replicaset-controller Normal SuccessfulCreate Created pod: auth-enabled-test-kserve-router-scheduler-6bbc585b87-nf5lr [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:33 kserve-ci-e2e-test deployment-controller Normal ScalingReplicaSet Scaled up replica set auth-enabled-test-kserve-router-scheduler-6bbc585b87 from 0 to 1 [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:32 kserve-ci-e2e-test deployment-controller Normal ScalingReplicaSet Scaled up replica set auth-enabled-test-kserve-87bd45bf5 from 0 to 1 [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:32 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Secret kserve-ci-e2e-test/auth-enabled-test-kserve-self-signed-certs [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:32 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Deployment kserve-ci-e2e-test/auth-enabled-test-kserve [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:32 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Service kserve-ci-e2e-test/auth-enabled-test-kserve-workload-svc [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:32 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.ServiceAccount kserve-ci-e2e-test/auth-enabled-test-epp-sa [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:33 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.ClusterRoleBinding /kserve-ci-e2e-test-auth-enabled-test-epp-auth-rb [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:33 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Role kserve-ci-e2e-test/auth-enabled-test-epp-role [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:33 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.RoleBinding kserve-ci-e2e-test/auth-enabled-test-epp-rb [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:33 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Deployment kserve-ci-e2e-test/auth-enabled-test-kserve-router-scheduler [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:33 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Service kserve-ci-e2e-test/auth-enabled-test-epp-service [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:43 kserve-ci-e2e-test LLMInferenceServiceController Normal Created (combined from similar events): Created v1.DestinationRule kserve-ci-e2e-test/auth-enabled-test-kserve-shadow-svc [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:42 kserve-ci-e2e-test LLMInferenceServiceController Normal Updated Updated v1.Secret kserve-ci-e2e-test/auth-enabled-test-kserve-self-signed-certs [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:43 kserve-ci-e2e-test LLMInferenceServiceController Normal Updated Updated v1.HTTPRoute kserve-ci-e2e-test/auth-enabled-test-kserve-route [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:05 kserve-ci-e2e-test LLMInferenceServiceController Normal LLMInferenceServiceReady LLMInferenceService [auth-enabled-test] is Ready [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:10 kserve-ci-e2e-test LLMInferenceServiceController Normal Deleted Deleted v1.ClusterRoleBinding /kserve-ci-e2e-test-auth-enabled-test-epp-auth-rb [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 None kserve-ci-e2e-test Normal Scheduled Successfully assigned kserve-ci-e2e-test/auth-invalid-token-test-kserve-6467b8fbf9-jhlvl to ip-10-0-139-47.ec2.internal [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:44 kserve-ci-e2e-test multus Normal AddedInterface Add eth0 [10.134.0.27/23] from ovn-kubernetes [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:44 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Pulled Container image "quay.io/opendatahub/kserve-storage-initializer@sha256:13ad765c0f323200e1db7b45c99661287e65c5158d6b54b16139829636d20be4" already present on machine [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:44 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Created Created container: storage-initializer [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:44 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Started Started container storage-initializer [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:49 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Pulled Container image "public.ecr.aws/q9t5s3a7/vllm-cpu-release-repo:v0.19.0" already present on machine [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:49 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Created Created container: main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:49 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Started Started container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:34 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Warning Unhealthy Startup probe failed: Get "https://10.134.0.27:8000/health": dial tcp 10.134.0.27:8000: connect: connection refused [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:46 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Killing Stopping container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:28:03 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Warning Unhealthy Readiness probe failed: Get "https://10.134.0.27:8000/health": context deadline exceeded [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:43 kserve-ci-e2e-test replicaset-controller Normal SuccessfulCreate Created pod: auth-invalid-token-test-kserve-6467b8fbf9-jhlvl [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:43 kserve-ci-e2e-test replicaset-controller Normal SuccessfulCreate Created pod: auth-invalid-token-test-kserve-router-scheduler-7c68b57cd5h6v2p [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 None kserve-ci-e2e-test Normal Scheduled Successfully assigned kserve-ci-e2e-test/auth-invalid-token-test-kserve-router-scheduler-7c68b57cd5h6v2p to ip-10-0-133-25.ec2.internal [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:44 kserve-ci-e2e-test multus Normal AddedInterface Add eth0 [10.132.0.49/23] from ovn-kubernetes [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:44 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Pulled Container image "quay.io/opendatahub/kserve-storage-initializer@sha256:13ad765c0f323200e1db7b45c99661287e65c5158d6b54b16139829636d20be4" already present on machine [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:44 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Created Created container: storage-initializer [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:44 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Started Started container storage-initializer [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:45 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Pulled Container image "ghcr.io/llm-d/llm-d-router-endpoint-picker:v0.9.0-rc.2" already present on machine [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:45 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Created Created container: main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:45 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Started Started container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:46 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Killing Stopping container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:46 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Warning FailedPreStopHook PreStopHook failed [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:43 kserve-ci-e2e-test deployment-controller Normal ScalingReplicaSet Scaled up replica set auth-invalid-token-test-kserve-router-scheduler-7c68b57cd5 from 0 to 1 [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:43 kserve-ci-e2e-test deployment-controller Normal ScalingReplicaSet Scaled up replica set auth-invalid-token-test-kserve-6467b8fbf9 from 0 to 1 [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:43 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Secret kserve-ci-e2e-test/auth-invalid-token-test-kserve-self-signed-certs [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:43 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Deployment kserve-ci-e2e-test/auth-invalid-token-test-kserve [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:43 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Service kserve-ci-e2e-test/auth-invalid-token-test-kserve-workload-svc [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:43 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.ServiceAccount kserve-ci-e2e-test/auth-invalid-token-test-epp-sa [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:43 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.ClusterRoleBinding /kserve-ci-e2e-test-auth-invalid-token-test-epp-auth-rb [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:43 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Role kserve-ci-e2e-test/auth-invalid-token-test-epp-role [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:43 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.RoleBinding kserve-ci-e2e-test/auth-invalid-token-test-epp-rb [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:43 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Deployment kserve-ci-e2e-test/auth-invalid-token-test-kserve-router-scheduler [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:43 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Service kserve-ci-e2e-test/auth-invalid-token-test-epp-service [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:26:10 kserve-ci-e2e-test LLMInferenceServiceController Normal Created (combined from similar events): Created v1.DestinationRule kserve-ci-e2e-test/auth-invalid-token-test-kserve-shadow-svc [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:26:09 kserve-ci-e2e-test LLMInferenceServiceController Normal Updated Updated v1.Secret kserve-ci-e2e-test/auth-invalid-token-test-kserve-self-signed-certs [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:26:09 kserve-ci-e2e-test LLMInferenceServiceController Normal Updated Updated v1.HTTPRoute kserve-ci-e2e-test/auth-invalid-token-test-kserve-route [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:45 kserve-ci-e2e-test LLMInferenceServiceController Normal LLMInferenceServiceReady LLMInferenceService [auth-invalid-token-test] is Ready [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:46 kserve-ci-e2e-test LLMInferenceServiceController Normal Deleted Deleted v1.ClusterRoleBinding /kserve-ci-e2e-test-auth-invalid-token-test-epp-auth-rb [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:45 kserve-ci-e2e-test replicaset-controller Normal SuccessfulCreate Created pod: gw-section-name-router-with-gat-2f0a622e-kserve-7c9c8cffc558295 [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 None kserve-ci-e2e-test Normal Scheduled Successfully assigned kserve-ci-e2e-test/gw-section-name-router-with-gat-2f0a622e-kserve-7c9c8cffc558295 to ip-10-0-139-47.ec2.internal [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:46 kserve-ci-e2e-test multus Normal AddedInterface Add eth0 [10.134.0.21/23] from ovn-kubernetes [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:46 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Pulling Pulling image "ghcr.io/llm-d/llm-d-inference-sim:v0.8.2" [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:47 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Pulled Successfully pulled image "ghcr.io/llm-d/llm-d-inference-sim:v0.8.2" in 1.195s (1.195s including waiting). Image size: 98346788 bytes. [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:47 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Created Created container: main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:47 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Started Started container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:58 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Killing Stopping container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:45 kserve-ci-e2e-test deployment-controller Normal ScalingReplicaSet Scaled up replica set gw-section-name-router-with-gat-2f0a622e-kserve-7c9c8cffc5 from 0 to 1 [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:45 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Secret kserve-ci-e2e-test/gw-sec0c69dceeb48768325d1a53a749e65786-kserve-self-signed-certs [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:45 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Deployment kserve-ci-e2e-test/gw-section-name-router-with-gat-2f0a622e-kserve [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:45 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Service kserve-ci-e2e-test/gw-section-name-router-with-gat-2f0a622e-kserve-workload-svc [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:45 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.HTTPRoute kserve-ci-e2e-test/gw-section-name-router-with-gat-2f0a622e-kserve-route [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:46 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.DestinationRule kserve-ci-e2e-test/gw-section-name-router-with-gat-2f0a622e-kserve-workload-svc [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:10 kserve-ci-e2e-test replicaset-controller Normal SuccessfulCreate Created pod: gw-section-name-router-with-gat-f1d92d0f-kserve-7bc8dd6c5bcr4rf [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 None kserve-ci-e2e-test Normal Scheduled Successfully assigned kserve-ci-e2e-test/gw-section-name-router-with-gat-f1d92d0f-kserve-7bc8dd6c5bcr4rf to ip-10-0-133-25.ec2.internal [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:10 kserve-ci-e2e-test multus Normal AddedInterface Add eth0 [10.132.0.46/23] from ovn-kubernetes [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:10 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Pulled Container image "ghcr.io/llm-d/llm-d-inference-sim:v0.8.2" already present on machine [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:10 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Created Created container: main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:10 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Started Started container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:27 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Killing Stopping container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:10 kserve-ci-e2e-test deployment-controller Normal ScalingReplicaSet Scaled up replica set gw-section-name-router-with-gat-f1d92d0f-kserve-7bc8dd6c5b from 0 to 1 [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:10 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Secret kserve-ci-e2e-test/gw-sec2774c263d49959f50d9eebc552e13bf9-kserve-self-signed-certs [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:10 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Deployment kserve-ci-e2e-test/gw-section-name-router-with-gat-f1d92d0f-kserve [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:10 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Service kserve-ci-e2e-test/gw-section-name-router-with-gat-f1d92d0f-kserve-workload-svc [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:10 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.HTTPRoute kserve-ci-e2e-test/gw-section-name-router-with-gat-f1d92d0f-kserve-route [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:10 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.DestinationRule kserve-ci-e2e-test/gw-section-name-router-with-gat-f1d92d0f-kserve-workload-svc [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:31:40 kserve-ci-e2e-test replicaset-controller Normal SuccessfulCreate Created pod: llmisvc-model-fb-opt-125m-route-87882a8e-kserve-5bd7864c6ctnql9 [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 None kserve-ci-e2e-test Normal Scheduled Successfully assigned kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-route-87882a8e-kserve-5bd7864c6ctnql9 to ip-10-0-139-47.ec2.internal [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:31:40 kserve-ci-e2e-test multus Normal AddedInterface Add eth0 [10.134.0.31/23] from ovn-kubernetes [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:31:40 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Pulled Container image "quay.io/opendatahub/kserve-storage-initializer@sha256:13ad765c0f323200e1db7b45c99661287e65c5158d6b54b16139829636d20be4" already present on machine [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:31:40 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Created Created container: storage-initializer [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:31:40 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Started Started container storage-initializer [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:31:45 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Pulled Container image "public.ecr.aws/q9t5s3a7/vllm-cpu-release-repo:v0.19.0" already present on machine [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:31:45 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Created Created container: main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:31:45 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Started Started container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:33:30 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Warning Unhealthy Startup probe failed: Get "https://10.134.0.31:8000/health": dial tcp 10.134.0.31:8000: connect: connection refused [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:31:40 kserve-ci-e2e-test deployment-controller Normal ScalingReplicaSet Scaled up replica set llmisvc-model-fb-opt-125m-route-87882a8e-kserve-5bd7864c6c from 0 to 1 [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:31:38 kserve-ci-e2e-test OpenDataHubModelController Warning ReconcileError Failed to reconcile LLMInferenceService: 1 error occurred: * failed to get HTTPRoute for AuthPolicy llmisvc-model-fb-opt-125m-route-87882a8e-kserve-route-authn: failed to get HTTPRoute kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-route-87882a8e-kserve-route: HTTPRoute.gateway.networking.k8s.io "llmisvc-model-fb-opt-125m-route-87882a8e-kserve-route" not found [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:31:40 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Secret kserve-ci-e2e-test/llmisva690bbc929faec8bc98c767f16c003c1-kserve-self-signed-certs [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:31:40 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Deployment kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-route-87882a8e-kserve [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:31:40 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Service kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-route-87882a8e-kserve-workload-svc [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:31:40 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.ServiceAccount kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-route-87882a8e-epp-sa [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:31:40 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.ClusterRoleBinding /kserve-ci-e2e-test21fe6730fe484f3a92b1a16afe1bac8f-llmisvc-mode [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:31:40 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Role kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-route-87882a8e-epp-role [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:31:40 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.RoleBinding kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-route-87882a8e-epp-rb [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:31:40 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Deployment kserve-ci-e2e-test/llmisvca690bbc929faec8bc98c767f16c003c1-kserve-router-scheduler [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:31:40 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Service kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-route-87882a8e-epp-service [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:31:47 kserve-ci-e2e-test LLMInferenceServiceController Normal Created (combined from similar events): Created v1.DestinationRule kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-route-87882a8e-kserve-shadow-svc [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:31:45 kserve-ci-e2e-test LLMInferenceServiceController Normal Updated Updated v1.Secret kserve-ci-e2e-test/llmisva690bbc929faec8bc98c767f16c003c1-kserve-self-signed-certs [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:31:46 kserve-ci-e2e-test LLMInferenceServiceController Normal Updated Updated v1.HTTPRoute kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-route-87882a8e-kserve-route [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:33:40 kserve-ci-e2e-test LLMInferenceServiceController Normal LLMInferenceServiceReady LLMInferenceService [llmisvc-model-fb-opt-125m-route-87882a8e] is Ready [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:31:02 kserve-ci-e2e-test replicaset-controller Normal SuccessfulCreate Created pod: llmisvc-model-fb-opt-125m-route-e95b1dc1-kserve-66874c76d46pqbc [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 None kserve-ci-e2e-test Normal Scheduled Successfully assigned kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-route-e95b1dc1-kserve-66874c76d46pqbc to ip-10-0-135-220.ec2.internal [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:31:02 kserve-ci-e2e-test multus Normal AddedInterface Add eth0 [10.133.0.37/23] from ovn-kubernetes [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:31:02 kserve-ci-e2e-test kubelet/ip-10-0-135-220.ec2.internal Normal Pulled Container image "ghcr.io/llm-d/llm-d-inference-sim:v0.8.2" already present on machine [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:31:02 kserve-ci-e2e-test kubelet/ip-10-0-135-220.ec2.internal Normal Created Created container: main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:31:02 kserve-ci-e2e-test kubelet/ip-10-0-135-220.ec2.internal Normal Started Started container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:31:15 kserve-ci-e2e-test kubelet/ip-10-0-135-220.ec2.internal Normal Killing Stopping container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:31:02 kserve-ci-e2e-test deployment-controller Normal ScalingReplicaSet Scaled up replica set llmisvc-model-fb-opt-125m-route-e95b1dc1-kserve-66874c76d4 from 0 to 1 [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:30:58 kserve-ci-e2e-test OpenDataHubModelController Warning ReconcileError Failed to reconcile LLMInferenceService: 1 error occurred: * failed to get HTTPRoute for AuthPolicy llmisvc-model-fb-opt-125m-route-e95b1dc1-kserve-route-authn: failed to get HTTPRoute kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-route-e95b1dc1-kserve-route: HTTPRoute.gateway.networking.k8s.io "llmisvc-model-fb-opt-125m-route-e95b1dc1-kserve-route" not found [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:31:02 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Secret kserve-ci-e2e-test/llmisv122f03714c5bdf915a2917fdf1262b98-kserve-self-signed-certs [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:31:02 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Deployment kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-route-e95b1dc1-kserve [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:31:02 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Service kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-route-e95b1dc1-kserve-workload-svc [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:31:02 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.HTTPRoute kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-route-e95b1dc1-kserve-route [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:31:02 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.DestinationRule kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-route-e95b1dc1-kserve-workload-svc [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:31:14 kserve-ci-e2e-test LLMInferenceServiceController Normal Updated Updated v1.Secret kserve-ci-e2e-test/llmisv122f03714c5bdf915a2917fdf1262b98-kserve-self-signed-certs [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:31:14 kserve-ci-e2e-test LLMInferenceServiceController Normal LLMInferenceServiceReady LLMInferenceService [llmisvc-model-fb-opt-125m-route-e95b1dc1] is Ready [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 None kserve-ci-e2e-test Normal Scheduled Successfully assigned kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-with-7ca60146-kserve-7cfdf49d74-pd9px to ip-10-0-139-47.ec2.internal [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:29:25 kserve-ci-e2e-test multus Normal AddedInterface Add eth0 [10.134.0.30/23] from ovn-kubernetes [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:29:25 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Pulled Container image "quay.io/opendatahub/kserve-storage-initializer@sha256:13ad765c0f323200e1db7b45c99661287e65c5158d6b54b16139829636d20be4" already present on machine [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:29:25 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Created Created container: storage-initializer [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:29:25 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Started Started container storage-initializer [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:29:30 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Pulled Container image "public.ecr.aws/q9t5s3a7/vllm-cpu-release-repo:v0.19.0" already present on machine [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:29:30 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Created Created container: main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:29:30 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Started Started container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:30:34 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Warning Unhealthy Startup probe failed: Get "https://10.134.0.30:8000/health": dial tcp 10.134.0.30:8000: connect: connection refused [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:29:24 kserve-ci-e2e-test replicaset-controller Normal SuccessfulCreate Created pod: llmisvc-model-fb-opt-125m-with-7ca60146-kserve-7cfdf49d74-pd9px [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:29:24 kserve-ci-e2e-test replicaset-controller Normal SuccessfulCreate Created pod: llmisvc-model-fb-opt-125m-with-7ca60146-kserve-router-schenp5mk [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:29:24 kserve-ci-e2e-test deployment-controller Normal ScalingReplicaSet Scaled up replica set llmisvc-model-fb-opt-125m-with-7ca60146-kserve-router-scheduler-778547fbc from 0 to 1 [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 None kserve-ci-e2e-test Normal Scheduled Successfully assigned kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-with-7ca60146-kserve-router-schenp5mk to ip-10-0-133-25.ec2.internal [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:29:25 kserve-ci-e2e-test multus Normal AddedInterface Add eth0 [10.132.0.52/23] from ovn-kubernetes [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:29:25 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Pulled Container image "quay.io/opendatahub/kserve-storage-initializer@sha256:13ad765c0f323200e1db7b45c99661287e65c5158d6b54b16139829636d20be4" already present on machine [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:29:25 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Created Created container: storage-initializer [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:29:25 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Started Started container storage-initializer [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:29:26 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Pulled Container image "ghcr.io/llm-d/llm-d-router-endpoint-picker:v0.9.0-rc.2" already present on machine [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:29:27 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Created Created container: main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:29:27 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Started Started container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:29:24 kserve-ci-e2e-test deployment-controller Normal ScalingReplicaSet Scaled up replica set llmisvc-model-fb-opt-125m-with-7ca60146-kserve-7cfdf49d74 from 0 to 1 [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:29:23 kserve-ci-e2e-test OpenDataHubModelController Warning ReconcileError Failed to reconcile LLMInferenceService: 1 error occurred: * failed to get HTTPRoute for AuthPolicy llmisvc-model-fb-opt-125m-with-7ca60146-kserve-route-authn: failed to get HTTPRoute kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-with-7ca60146-kserve-route: HTTPRoute.gateway.networking.k8s.io "llmisvc-model-fb-opt-125m-with-7ca60146-kserve-route" not found [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:29:24 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Secret kserve-ci-e2e-test/llmisv3e414c2ba058a022dfd694dbcbac5b51-kserve-self-signed-certs [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:29:24 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Deployment kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-with-7ca60146-kserve [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:29:24 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Service kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-with-7ca60146-kserve-workload-svc [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:29:24 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.ServiceAccount kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-with-7ca60146-epp-sa [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:29:24 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.ClusterRoleBinding /kserve-ci-e2e-test5216bfd716f919dc046bc693ceb22e41-llmisvc-mode [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:29:24 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Role kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-with-7ca60146-epp-role [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:29:24 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.RoleBinding kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-with-7ca60146-epp-rb [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:29:24 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Deployment kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-with-7ca60146-kserve-router-scheduler [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:29:24 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Service kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-with-7ca60146-epp-service [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:29:35 kserve-ci-e2e-test LLMInferenceServiceController Normal Created (combined from similar events): Created v1.DestinationRule kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-with-7ca60146-kserve-shadow-svc [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:29:33 kserve-ci-e2e-test LLMInferenceServiceController Normal Updated Updated v1.Secret kserve-ci-e2e-test/llmisv3e414c2ba058a022dfd694dbcbac5b51-kserve-self-signed-certs [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:29:34 kserve-ci-e2e-test LLMInferenceServiceController Normal Updated Updated v1.HTTPRoute kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-with-7ca60146-kserve-route [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:30:45 kserve-ci-e2e-test LLMInferenceServiceController Normal LLMInferenceServiceReady LLMInferenceService [llmisvc-model-fb-opt-125m-with-7ca60146] is Ready [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 None kserve-ci-e2e-test Normal Scheduled Successfully assigned kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-with-ba4d693a-kserve-5b998d5789-t6wf4 to ip-10-0-139-47.ec2.internal [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:45:56 kserve-ci-e2e-test multus Normal AddedInterface Add eth0 [10.134.0.32/23] from ovn-kubernetes [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:45:56 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Pulled Container image "quay.io/opendatahub/kserve-storage-initializer@sha256:13ad765c0f323200e1db7b45c99661287e65c5158d6b54b16139829636d20be4" already present on machine [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:45:56 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Created Created container: storage-initializer [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:45:56 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Started Started container storage-initializer [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:46:01 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Pulled Container image "public.ecr.aws/q9t5s3a7/vllm-cpu-release-repo:v0.19.0" already present on machine [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:46:01 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Created Created container: main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:46:01 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Started Started container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:47:06 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Warning Unhealthy Startup probe failed: Get "https://10.134.0.32:8000/health": dial tcp 10.134.0.32:8000: connect: connection refused [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:45:56 kserve-ci-e2e-test replicaset-controller Normal SuccessfulCreate Created pod: llmisvc-model-fb-opt-125m-with-ba4d693a-kserve-5b998d5789-t6wf4 [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 None kserve-ci-e2e-test Normal Scheduled Successfully assigned kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-with-ba4d693a-kserve-router-sche4ghr9 to ip-10-0-133-25.ec2.internal [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:45:56 kserve-ci-e2e-test multus Normal AddedInterface Add eth0 [10.132.0.54/23] from ovn-kubernetes [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:45:56 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Pulled Container image "quay.io/opendatahub/kserve-storage-initializer@sha256:13ad765c0f323200e1db7b45c99661287e65c5158d6b54b16139829636d20be4" already present on machine [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:45:57 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Created Created container: storage-initializer [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:45:57 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Started Started container storage-initializer [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:45:57 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Pulled Container image "ghcr.io/llm-d/llm-d-router-endpoint-picker:v0.9.0-rc.2" already present on machine [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:45:57 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Created Created container: main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:45:57 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Started Started container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:45:56 kserve-ci-e2e-test replicaset-controller Normal SuccessfulCreate Created pod: llmisvc-model-fb-opt-125m-with-ba4d693a-kserve-router-sche4ghr9 [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:45:56 kserve-ci-e2e-test deployment-controller Normal ScalingReplicaSet Scaled up replica set llmisvc-model-fb-opt-125m-with-ba4d693a-kserve-router-scheduler-79f5455bb from 0 to 1 [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:45:56 kserve-ci-e2e-test deployment-controller Normal ScalingReplicaSet Scaled up replica set llmisvc-model-fb-opt-125m-with-ba4d693a-kserve-5b998d5789 from 0 to 1 [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:45:56 kserve-ci-e2e-test OpenDataHubModelController Warning ReconcileError Failed to reconcile LLMInferenceService: 1 error occurred: * failed to get HTTPRoute for AuthPolicy llmisvc-model-fb-opt-125m-with-ba4d693a-kserve-route-authn: failed to get HTTPRoute kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-with-ba4d693a-kserve-route: HTTPRoute.gateway.networking.k8s.io "llmisvc-model-fb-opt-125m-with-ba4d693a-kserve-route" not found [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:45:56 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Secret kserve-ci-e2e-test/llmisv77ff2528d3e9b4972cd9335229fce9f0-kserve-self-signed-certs [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:45:56 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Deployment kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-with-ba4d693a-kserve [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:45:56 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Service kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-with-ba4d693a-kserve-workload-svc [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:45:56 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.ServiceAccount kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-with-ba4d693a-epp-sa [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:45:56 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.ClusterRoleBinding /kserve-ci-e2e-test05addb65ba05195619f26ef266e8fc04-llmisvc-mode [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:45:56 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Role kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-with-ba4d693a-epp-role [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:45:56 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.RoleBinding kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-with-ba4d693a-epp-rb [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:45:56 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Deployment kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-with-ba4d693a-kserve-router-scheduler [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:45:56 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Service kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-with-ba4d693a-epp-service [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:46:17 kserve-ci-e2e-test LLMInferenceServiceController Normal Created (combined from similar events): Created v1.DestinationRule kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-with-ba4d693a-kserve-shadow-svc [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:46:16 kserve-ci-e2e-test LLMInferenceServiceController Normal Updated Updated v1.Secret kserve-ci-e2e-test/llmisv77ff2528d3e9b4972cd9335229fce9f0-kserve-self-signed-certs [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:46:16 kserve-ci-e2e-test LLMInferenceServiceController Normal Updated Updated v1.HTTPRoute kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-with-ba4d693a-kserve-route [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:47:17 kserve-ci-e2e-test LLMInferenceServiceController Normal LLMInferenceServiceReady LLMInferenceService [llmisvc-model-fb-opt-125m-with-ba4d693a] is Ready [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:54 kserve-ci-e2e-test replicaset-controller Normal SuccessfulCreate Created pod: llmisvc-model-qwen2-5-0-5b-rout-4f8c0978-kserve-5c54ddb98flf892 [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 None kserve-ci-e2e-test Normal Scheduled Successfully assigned kserve-ci-e2e-test/llmisvc-model-qwen2-5-0-5b-rout-4f8c0978-kserve-5c54ddb98flf892 to ip-10-0-135-220.ec2.internal [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:54 kserve-ci-e2e-test multus Normal AddedInterface Add eth0 [10.133.0.36/23] from ovn-kubernetes [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:54 kserve-ci-e2e-test kubelet/ip-10-0-135-220.ec2.internal Normal Pulled Container image "ghcr.io/llm-d/llm-d-inference-sim:v0.8.2" already present on machine [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:54 kserve-ci-e2e-test kubelet/ip-10-0-135-220.ec2.internal Normal Created Created container: main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:54 kserve-ci-e2e-test kubelet/ip-10-0-135-220.ec2.internal Normal Started Started container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:28:40 kserve-ci-e2e-test kubelet/ip-10-0-135-220.ec2.internal Normal Killing Stopping container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:28:56 kserve-ci-e2e-test kubelet/ip-10-0-135-220.ec2.internal Warning Unhealthy Readiness probe failed: Get "https://10.133.0.36:8000/health": context deadline exceeded [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:54 kserve-ci-e2e-test deployment-controller Normal ScalingReplicaSet Scaled up replica set llmisvc-model-qwen2-5-0-5b-rout-4f8c0978-kserve-5c54ddb98f from 0 to 1 [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:53 kserve-ci-e2e-test OpenDataHubModelController Warning ReconcileError Failed to reconcile LLMInferenceService: 1 error occurred: * failed to get HTTPRoute for AuthPolicy llmisvc-model-qwen2-5-0-5b-rout-4f8c0978-kserve-route-authn: failed to get HTTPRoute kserve-ci-e2e-test/llmisvc-model-qwen2-5-0-5b-rout-4f8c0978-kserve-route: HTTPRoute.gateway.networking.k8s.io "llmisvc-model-qwen2-5-0-5b-rout-4f8c0978-kserve-route" not found [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:54 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Secret kserve-ci-e2e-test/llmisv08544b88a8d963ffd553cc1f3ed82d16-kserve-self-signed-certs [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:54 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Deployment kserve-ci-e2e-test/llmisvc-model-qwen2-5-0-5b-rout-4f8c0978-kserve [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:54 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Service kserve-ci-e2e-test/llmisvc-model-qwen2-5-0-5b-rout-4f8c0978-kserve-workload-svc [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:54 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.ServiceAccount kserve-ci-e2e-test/llmisvc-model-qwen2-5-0-5b-rout-4f8c0978-epp-sa [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:54 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.ClusterRoleBinding /kserve-ci-e2e-test76d7f532acb694e4a7bcef75d32cd8a1-llmisvc-mode [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:54 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Role kserve-ci-e2e-test/llmisvc-model-qwen2-5-0-5b-rout-4f8c0978-epp-role [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:54 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.RoleBinding kserve-ci-e2e-test/llmisvc-model-qwen2-5-0-5b-rout-4f8c0978-epp-rb [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:54 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Deployment kserve-ci-e2e-test/llmisvc08544b88a8d963ffd553cc1f3ed82d16-kserve-router-scheduler [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:54 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Service kserve-ci-e2e-test/llmisvc-model-qwen2-5-0-5b-rout-4f8c0978-epp-service [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:28:06 kserve-ci-e2e-test LLMInferenceServiceController Normal Created (combined from similar events): Created v1.DestinationRule kserve-ci-e2e-test/llmisvc-model-qwen2-5-0-5b-rout-4f8c0978-kserve-shadow-svc [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:28:05 kserve-ci-e2e-test LLMInferenceServiceController Normal Updated Updated v1.Secret kserve-ci-e2e-test/llmisv08544b88a8d963ffd553cc1f3ed82d16-kserve-self-signed-certs [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:28:06 kserve-ci-e2e-test LLMInferenceServiceController Normal Updated Updated v1.HTTPRoute kserve-ci-e2e-test/llmisvc-model-qwen2-5-0-5b-rout-4f8c0978-kserve-route [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:28:31 kserve-ci-e2e-test LLMInferenceServiceController Normal LLMInferenceServiceReady LLMInferenceService [llmisvc-model-qwen2-5-0-5b-rout-4f8c0978] is Ready [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:28:40 kserve-ci-e2e-test LLMInferenceServiceController Normal Deleted Deleted v1.ClusterRoleBinding /kserve-ci-e2e-test76d7f532acb694e4a7bcef75d32cd8a1-llmisvc-mode [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:23 kserve-ci-e2e-test replicaset-controller Normal SuccessfulCreate Created pod: llmisvc-model-qwen2-5-0-5b-rout-a50492e9-kserve-5d8ffd58ddqnd8m [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 None kserve-ci-e2e-test Normal Scheduled Successfully assigned kserve-ci-e2e-test/llmisvc-model-qwen2-5-0-5b-rout-a50492e9-kserve-5d8ffd58ddqnd8m to ip-10-0-135-220.ec2.internal [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:23 kserve-ci-e2e-test multus Normal AddedInterface Add eth0 [10.133.0.34/23] from ovn-kubernetes [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:23 kserve-ci-e2e-test kubelet/ip-10-0-135-220.ec2.internal Normal Pulled Container image "ghcr.io/llm-d/llm-d-inference-sim:v0.8.2" already present on machine [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:23 kserve-ci-e2e-test kubelet/ip-10-0-135-220.ec2.internal Normal Created Created container: main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:23 kserve-ci-e2e-test kubelet/ip-10-0-135-220.ec2.internal Normal Started Started container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:26:13 kserve-ci-e2e-test kubelet/ip-10-0-135-220.ec2.internal Normal Killing Stopping container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:23 kserve-ci-e2e-test deployment-controller Normal ScalingReplicaSet Scaled up replica set llmisvc-model-qwen2-5-0-5b-rout-a50492e9-kserve-5d8ffd58dd from 0 to 1 [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:21 kserve-ci-e2e-test OpenDataHubModelController Warning ReconcileError Failed to reconcile LLMInferenceService: 1 error occurred: * failed to get HTTPRoute for AuthPolicy llmisvc-model-qwen2-5-0-5b-rout-a50492e9-kserve-route-authn: failed to get HTTPRoute kserve-ci-e2e-test/llmisvc-model-qwen2-5-0-5b-rout-a50492e9-kserve-route: HTTPRoute.gateway.networking.k8s.io "llmisvc-model-qwen2-5-0-5b-rout-a50492e9-kserve-route" not found [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:23 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Secret kserve-ci-e2e-test/llmisvb11a9c9832b99b016bc8f8e0ea095712-kserve-self-signed-certs [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:23 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Deployment kserve-ci-e2e-test/llmisvc-model-qwen2-5-0-5b-rout-a50492e9-kserve [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:23 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Service kserve-ci-e2e-test/llmisvc-model-qwen2-5-0-5b-rout-a50492e9-kserve-workload-svc [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:23 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.ServiceAccount kserve-ci-e2e-test/llmisvc-model-qwen2-5-0-5b-rout-a50492e9-epp-sa [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:23 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.ClusterRoleBinding /kserve-ci-e2e-testb7025ca4d8a6f8f5b2fd08b5581d2678-llmisvc-mode [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:23 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Role kserve-ci-e2e-test/llmisvc-model-qwen2-5-0-5b-rout-a50492e9-epp-role [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:23 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.RoleBinding kserve-ci-e2e-test/llmisvc-model-qwen2-5-0-5b-rout-a50492e9-epp-rb [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:23 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Deployment kserve-ci-e2e-test/llmisvcb11a9c9832b99b016bc8f8e0ea095712-kserve-router-scheduler [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:23 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Service kserve-ci-e2e-test/llmisvc-model-qwen2-5-0-5b-rout-a50492e9-epp-service [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:34 kserve-ci-e2e-test LLMInferenceServiceController Normal Created (combined from similar events): Created v1.DestinationRule kserve-ci-e2e-test/llmisvc-model-qwen2-5-0-5b-rout-a50492e9-kserve-shadow-svc [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:33 kserve-ci-e2e-test LLMInferenceServiceController Normal Updated Updated v1.Secret kserve-ci-e2e-test/llmisvb11a9c9832b99b016bc8f8e0ea095712-kserve-self-signed-certs [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:33 kserve-ci-e2e-test LLMInferenceServiceController Normal Updated Updated v1.HTTPRoute kserve-ci-e2e-test/llmisvc-model-qwen2-5-0-5b-rout-a50492e9-kserve-route [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:26:12 kserve-ci-e2e-test LLMInferenceServiceController Normal LLMInferenceServiceReady LLMInferenceService [llmisvc-model-qwen2-5-0-5b-rout-a50492e9] is Ready [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:26:13 kserve-ci-e2e-test LLMInferenceServiceController Normal Deleted Deleted v1.ClusterRoleBinding /kserve-ci-e2e-testb7025ca4d8a6f8f5b2fd08b5581d2678-llmisvc-mode [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 None kserve-ci-e2e-test Normal Scheduled Successfully assigned kserve-ci-e2e-test/llmisvc-router-managed-test-llm-4b931143-kserve-bd545d56-jcs4n to ip-10-0-135-220.ec2.internal [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:18 kserve-ci-e2e-test multus Normal AddedInterface Add eth0 [10.133.0.35/23] from ovn-kubernetes [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:18 kserve-ci-e2e-test kubelet/ip-10-0-135-220.ec2.internal Normal Pulled Container image "ghcr.io/llm-d/llm-d-inference-sim:v0.8.2" already present on machine [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:18 kserve-ci-e2e-test kubelet/ip-10-0-135-220.ec2.internal Normal Created Created container: main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:18 kserve-ci-e2e-test kubelet/ip-10-0-135-220.ec2.internal Normal Started Started container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:28:58 kserve-ci-e2e-test kubelet/ip-10-0-135-220.ec2.internal Normal Killing Stopping container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:17 kserve-ci-e2e-test replicaset-controller Normal SuccessfulCreate Created pod: llmisvc-router-managed-test-llm-4b931143-kserve-bd545d56-jcs4n [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:17 kserve-ci-e2e-test deployment-controller Normal ScalingReplicaSet Scaled up replica set llmisvc-router-managed-test-llm-4b931143-kserve-bd545d56 from 0 to 1 [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:14 kserve-ci-e2e-test OpenDataHubModelController Warning ReconcileError Failed to reconcile LLMInferenceService: 1 error occurred: * failed to get HTTPRoute for AuthPolicy llmisvc-router-managed-test-llm-4b931143-kserve-route-authn: failed to get HTTPRoute kserve-ci-e2e-test/llmisvc-router-managed-test-llm-4b931143-kserve-route: HTTPRoute.gateway.networking.k8s.io "llmisvc-router-managed-test-llm-4b931143-kserve-route" not found [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:17 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Secret kserve-ci-e2e-test/llmisvca2d2d7d499abb359505529ebe02c136-kserve-self-signed-certs [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:17 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Deployment kserve-ci-e2e-test/llmisvc-router-managed-test-llm-4b931143-kserve [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:17 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Service kserve-ci-e2e-test/llmisvc-router-managed-test-llm-4b931143-kserve-workload-svc [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:17 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.ServiceAccount kserve-ci-e2e-test/llmisvc-router-managed-test-llm-4b931143-epp-sa [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:17 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.ClusterRoleBinding /kserve-ci-e2e-test8ac8e3d2264ccb939eb021b0b835847c-llmisvc-rout [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:17 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Role kserve-ci-e2e-test/llmisvc-router-managed-test-llm-4b931143-epp-role [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:17 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.RoleBinding kserve-ci-e2e-test/llmisvc-router-managed-test-llm-4b931143-epp-rb [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:17 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Deployment kserve-ci-e2e-test/llmisvcca2d2d7d499abb359505529ebe02c136-kserve-router-scheduler [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:17 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Service kserve-ci-e2e-test/llmisvc-router-managed-test-llm-4b931143-epp-service [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:43 kserve-ci-e2e-test LLMInferenceServiceController Normal Created (combined from similar events): Created v1.DestinationRule kserve-ci-e2e-test/llmisvc-router-managed-test-llm-4b931143-kserve-shadow-svc [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:42 kserve-ci-e2e-test LLMInferenceServiceController Normal Updated Updated v1.Secret kserve-ci-e2e-test/llmisvca2d2d7d499abb359505529ebe02c136-kserve-self-signed-certs [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:43 kserve-ci-e2e-test LLMInferenceServiceController Normal Updated Updated v1.HTTPRoute kserve-ci-e2e-test/llmisvc-router-managed-test-llm-4b931143-kserve-route [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:50 kserve-ci-e2e-test LLMInferenceServiceController Normal LLMInferenceServiceReady LLMInferenceService [llmisvc-router-managed-test-llm-4b931143] is Ready [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:28:58 kserve-ci-e2e-test LLMInferenceServiceController Normal Deleted Deleted v1.ClusterRoleBinding /kserve-ci-e2e-test8ac8e3d2264ccb939eb021b0b835847c-llmisvc-rout [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:24:42 kserve-ci-e2e-test replicaset-controller Normal SuccessfulCreate Created pod: llmisvc-router-managed-test-llm-5b1e8f15-kserve-64df7bddb4xtf2m [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 None kserve-ci-e2e-test Normal Scheduled Successfully assigned kserve-ci-e2e-test/llmisvc-router-managed-test-llm-5b1e8f15-kserve-64df7bddb4xtf2m to ip-10-0-135-220.ec2.internal [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:24:43 kserve-ci-e2e-test multus Normal AddedInterface Add eth0 [10.133.0.33/23] from ovn-kubernetes [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:24:43 kserve-ci-e2e-test kubelet/ip-10-0-135-220.ec2.internal Normal Pulled Container image "ghcr.io/llm-d/llm-d-inference-sim:v0.8.2" already present on machine [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:24:43 kserve-ci-e2e-test kubelet/ip-10-0-135-220.ec2.internal Normal Created Created container: main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:24:43 kserve-ci-e2e-test kubelet/ip-10-0-135-220.ec2.internal Normal Started Started container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:26:33 kserve-ci-e2e-test kubelet/ip-10-0-135-220.ec2.internal Normal Killing Stopping container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:24:42 kserve-ci-e2e-test deployment-controller Normal ScalingReplicaSet Scaled up replica set llmisvc-router-managed-test-llm-5b1e8f15-kserve-64df7bddb4 from 0 to 1 [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:24:40 kserve-ci-e2e-test OpenDataHubModelController Warning ReconcileError Failed to reconcile LLMInferenceService: 1 error occurred: * failed to get HTTPRoute for AuthPolicy llmisvc-router-managed-test-llm-5b1e8f15-kserve-route-authn: failed to get HTTPRoute kserve-ci-e2e-test/llmisvc-router-managed-test-llm-5b1e8f15-kserve-route: HTTPRoute.gateway.networking.k8s.io "llmisvc-router-managed-test-llm-5b1e8f15-kserve-route" not found [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:24:42 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Secret kserve-ci-e2e-test/llmisve55ae740357a3a31a27cdb8b66ffe20f-kserve-self-signed-certs [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:24:42 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Deployment kserve-ci-e2e-test/llmisvc-router-managed-test-llm-5b1e8f15-kserve [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:24:42 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Service kserve-ci-e2e-test/llmisvc-router-managed-test-llm-5b1e8f15-kserve-workload-svc [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:24:42 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.ServiceAccount kserve-ci-e2e-test/llmisvc-router-managed-test-llm-5b1e8f15-epp-sa [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:24:42 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.ClusterRoleBinding /kserve-ci-e2e-test7f54e84970003a6e7372bdbcb574f7ed-llmisvc-rout [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:24:42 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Role kserve-ci-e2e-test/llmisvc-router-managed-test-llm-5b1e8f15-epp-role [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:24:43 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.RoleBinding kserve-ci-e2e-test/llmisvc-router-managed-test-llm-5b1e8f15-epp-rb [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:24:43 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Deployment kserve-ci-e2e-test/llmisvce55ae740357a3a31a27cdb8b66ffe20f-kserve-router-scheduler [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:24:43 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Service kserve-ci-e2e-test/llmisvc-router-managed-test-llm-5b1e8f15-epp-service [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:03 kserve-ci-e2e-test LLMInferenceServiceController Normal Created (combined from similar events): Created v1.DestinationRule kserve-ci-e2e-test/llmisvc-router-managed-test-llm-5b1e8f15-kserve-shadow-svc [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:03 kserve-ci-e2e-test LLMInferenceServiceController Normal Updated Updated v1.Secret kserve-ci-e2e-test/llmisve55ae740357a3a31a27cdb8b66ffe20f-kserve-self-signed-certs [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:03 kserve-ci-e2e-test LLMInferenceServiceController Normal Updated Updated v1.HTTPRoute kserve-ci-e2e-test/llmisvc-router-managed-test-llm-5b1e8f15-kserve-route [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:15 kserve-ci-e2e-test LLMInferenceServiceController Normal LLMInferenceServiceReady LLMInferenceService [llmisvc-router-managed-test-llm-5b1e8f15] is Ready [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:26:33 kserve-ci-e2e-test LLMInferenceServiceController Normal Deleted Deleted v1.ClusterRoleBinding /kserve-ci-e2e-test7f54e84970003a6e7372bdbcb574f7ed-llmisvc-rout [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:23:41 kserve-ci-e2e-test replicaset-controller Normal SuccessfulCreate Created pod: llmisvc-router-managed-test-llm-e45d1f79-kserve-7fdbbd4879ffz96 [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 None kserve-ci-e2e-test Normal Scheduled Successfully assigned kserve-ci-e2e-test/llmisvc-router-managed-test-llm-e45d1f79-kserve-7fdbbd4879ffz96 to ip-10-0-135-220.ec2.internal [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:23:41 kserve-ci-e2e-test multus Normal AddedInterface Add eth0 [10.133.0.32/23] from ovn-kubernetes [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:23:41 kserve-ci-e2e-test kubelet/ip-10-0-135-220.ec2.internal Normal Pulling Pulling image "ghcr.io/llm-d/llm-d-inference-sim:v0.8.2" [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:23:43 kserve-ci-e2e-test kubelet/ip-10-0-135-220.ec2.internal Normal Pulled Successfully pulled image "ghcr.io/llm-d/llm-d-inference-sim:v0.8.2" in 1.407s (1.407s including waiting). Image size: 98346788 bytes. [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:23:43 kserve-ci-e2e-test kubelet/ip-10-0-135-220.ec2.internal Normal Created Created container: main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:23:43 kserve-ci-e2e-test kubelet/ip-10-0-135-220.ec2.internal Normal Started Started container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:24:15 kserve-ci-e2e-test kubelet/ip-10-0-135-220.ec2.internal Normal Killing Stopping container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:23:41 kserve-ci-e2e-test deployment-controller Normal ScalingReplicaSet Scaled up replica set llmisvc-router-managed-test-llm-e45d1f79-kserve-7fdbbd4879 from 0 to 1 [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:23:36 kserve-ci-e2e-test OpenDataHubModelController Warning ReconcileError Failed to reconcile LLMInferenceService: 1 error occurred: * failed to get HTTPRoute for AuthPolicy llmisvc-router-managed-test-llm-e45d1f79-kserve-route-authn: failed to get HTTPRoute kserve-ci-e2e-test/llmisvc-router-managed-test-llm-e45d1f79-kserve-route: HTTPRoute.gateway.networking.k8s.io "llmisvc-router-managed-test-llm-e45d1f79-kserve-route" not found [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:23:41 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Secret kserve-ci-e2e-test/llmisv5c7e67b6c51568d1d6d13829a9337f2a-kserve-self-signed-certs [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:23:41 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Deployment kserve-ci-e2e-test/llmisvc-router-managed-test-llm-e45d1f79-kserve [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:23:41 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Service kserve-ci-e2e-test/llmisvc-router-managed-test-llm-e45d1f79-kserve-workload-svc [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:23:41 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.ServiceAccount kserve-ci-e2e-test/llmisvc-router-managed-test-llm-e45d1f79-epp-sa [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:23:41 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.ClusterRoleBinding /kserve-ci-e2e-testef4d2875be14b30dc1561ed84d0d4bde-llmisvc-rout [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:23:41 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Role kserve-ci-e2e-test/llmisvc-router-managed-test-llm-e45d1f79-epp-role [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:23:41 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.RoleBinding kserve-ci-e2e-test/llmisvc-router-managed-test-llm-e45d1f79-epp-rb [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:23:41 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Deployment kserve-ci-e2e-test/llmisvc5c7e67b6c51568d1d6d13829a9337f2a-kserve-router-scheduler [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:23:41 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Service kserve-ci-e2e-test/llmisvc-router-managed-test-llm-e45d1f79-epp-service [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:23:53 kserve-ci-e2e-test LLMInferenceServiceController Normal Created (combined from similar events): Created v1.DestinationRule kserve-ci-e2e-test/llmisvc-router-managed-test-llm-e45d1f79-kserve-shadow-svc [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:23:52 kserve-ci-e2e-test LLMInferenceServiceController Normal Updated Updated v1.Secret kserve-ci-e2e-test/llmisv5c7e67b6c51568d1d6d13829a9337f2a-kserve-self-signed-certs [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:23:52 kserve-ci-e2e-test LLMInferenceServiceController Normal Updated Updated v1.HTTPRoute kserve-ci-e2e-test/llmisvc-router-managed-test-llm-e45d1f79-kserve-route [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:24:14 kserve-ci-e2e-test LLMInferenceServiceController Normal LLMInferenceServiceReady LLMInferenceService [llmisvc-router-managed-test-llm-e45d1f79] is Ready [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:24:15 kserve-ci-e2e-test LLMInferenceServiceController Normal Deleted Deleted v1.ClusterRoleBinding /kserve-ci-e2e-testef4d2875be14b30dc1561ed84d0d4bde-llmisvc-rout [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:54 kserve-ci-e2e-test replicaset-controller Normal SuccessfulCreate Created pod: llmisvc08544b88a8d963ffd553cc1f3ed82d16-kserve-router-schefvtgh [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:54 kserve-ci-e2e-test deployment-controller Normal ScalingReplicaSet Scaled up replica set llmisvc08544b88a8d963ffd553cc1f3ed82d16-kserve-router-scheduler-6fcb489785 from 0 to 1 [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 None kserve-ci-e2e-test Normal Scheduled Successfully assigned kserve-ci-e2e-test/llmisvc08544b88a8d963ffd553cc1f3ed82d16-kserve-router-schefvtgh to ip-10-0-133-25.ec2.internal [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:55 kserve-ci-e2e-test multus Normal AddedInterface Add eth0 [10.132.0.50/23] from ovn-kubernetes [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:55 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Pulled Container image "ghcr.io/llm-d/llm-d-router-endpoint-picker:v0.9.0-rc.2" already present on machine [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:55 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Created Created container: main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:55 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Started Started container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:28:40 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Killing Stopping container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:28:40 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Warning FailedPreStopHook PreStopHook failed [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:23:41 kserve-ci-e2e-test replicaset-controller Normal SuccessfulCreate Created pod: llmisvc5c7e67b6c51568d1d6d13829a9337f2a-kserve-router-schexv898 [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:23:41 kserve-ci-e2e-test deployment-controller Normal ScalingReplicaSet Scaled up replica set llmisvc5c7e67b6c51568d1d6d13829a9337f2a-kserve-router-scheduler-5dd88bfbb7 from 0 to 1 [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 None kserve-ci-e2e-test Normal Scheduled Successfully assigned kserve-ci-e2e-test/llmisvc5c7e67b6c51568d1d6d13829a9337f2a-kserve-router-schexv898 to ip-10-0-139-47.ec2.internal [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:23:42 kserve-ci-e2e-test multus Normal AddedInterface Add eth0 [10.134.0.25/23] from ovn-kubernetes [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:23:42 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Pulled Container image "ghcr.io/llm-d/llm-d-router-endpoint-picker:v0.9.0-rc.2" already present on machine [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:23:42 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Created Created container: main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:23:42 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Started Started container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:24:15 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Killing Stopping container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:24:15 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Warning FailedPreStopHook PreStopHook failed [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:31:40 kserve-ci-e2e-test replicaset-controller Normal SuccessfulCreate Created pod: llmisvca690bbc929faec8bc98c767f16c003c1-kserve-router-scheq44kv [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:31:40 kserve-ci-e2e-test deployment-controller Normal ScalingReplicaSet Scaled up replica set llmisvca690bbc929faec8bc98c767f16c003c1-kserve-router-scheduler-5d6998448f from 0 to 1 [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 None kserve-ci-e2e-test Normal Scheduled Successfully assigned kserve-ci-e2e-test/llmisvca690bbc929faec8bc98c767f16c003c1-kserve-router-scheq44kv to ip-10-0-133-25.ec2.internal [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:31:40 kserve-ci-e2e-test multus Normal AddedInterface Add eth0 [10.132.0.53/23] from ovn-kubernetes [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:31:41 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Pulled Container image "quay.io/opendatahub/kserve-storage-initializer@sha256:13ad765c0f323200e1db7b45c99661287e65c5158d6b54b16139829636d20be4" already present on machine [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:31:41 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Created Created container: storage-initializer [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:31:41 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Started Started container storage-initializer [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:31:42 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Pulled Container image "ghcr.io/llm-d/llm-d-router-endpoint-picker:v0.9.0-rc.2" already present on machine [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:31:42 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Created Created container: main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:31:42 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Started Started container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:23 kserve-ci-e2e-test replicaset-controller Normal SuccessfulCreate Created pod: llmisvcb11a9c9832b99b016bc8f8e0ea095712-kserve-router-scheknzg7 [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:23 kserve-ci-e2e-test deployment-controller Normal ScalingReplicaSet Scaled up replica set llmisvcb11a9c9832b99b016bc8f8e0ea095712-kserve-router-scheduler-6d86bd4d9d from 0 to 1 [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 None kserve-ci-e2e-test Normal Scheduled Successfully assigned kserve-ci-e2e-test/llmisvcb11a9c9832b99b016bc8f8e0ea095712-kserve-router-scheknzg7 to ip-10-0-133-25.ec2.internal [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:24 kserve-ci-e2e-test multus Normal AddedInterface Add eth0 [10.132.0.48/23] from ovn-kubernetes [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:24 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Pulled Container image "ghcr.io/llm-d/llm-d-router-endpoint-picker:v0.9.0-rc.2" already present on machine [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:24 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Created Created container: main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:24 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Started Started container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:26:13 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Killing Stopping container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:26:13 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Warning FailedPreStopHook PreStopHook failed [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 None kserve-ci-e2e-test Normal Scheduled Successfully assigned kserve-ci-e2e-test/llmisvcca2d2d7d499abb359505529ebe02c136-kserve-router-scheb2glb to ip-10-0-139-47.ec2.internal [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:18 kserve-ci-e2e-test multus Normal AddedInterface Add eth0 [10.134.0.28/23] from ovn-kubernetes [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:18 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Pulled Container image "ghcr.io/llm-d/llm-d-router-endpoint-picker:v0.9.0-rc.2" already present on machine [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:18 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Created Created container: main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:18 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Started Started container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:28:58 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Killing Stopping container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:28:58 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Warning FailedPreStopHook PreStopHook failed [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:17 kserve-ci-e2e-test replicaset-controller Normal SuccessfulCreate Created pod: llmisvcca2d2d7d499abb359505529ebe02c136-kserve-router-scheb2glb [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:17 kserve-ci-e2e-test deployment-controller Normal ScalingReplicaSet Scaled up replica set llmisvcca2d2d7d499abb359505529ebe02c136-kserve-router-scheduler-68cc9685d6 from 0 to 1 [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:24:43 kserve-ci-e2e-test replicaset-controller Normal SuccessfulCreate Created pod: llmisvce55ae740357a3a31a27cdb8b66ffe20f-kserve-router-schex4lqs [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:24:43 kserve-ci-e2e-test deployment-controller Normal ScalingReplicaSet Scaled up replica set llmisvce55ae740357a3a31a27cdb8b66ffe20f-kserve-router-scheduler-749449dbc8 from 0 to 1 [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 None kserve-ci-e2e-test Normal Scheduled Successfully assigned kserve-ci-e2e-test/llmisvce55ae740357a3a31a27cdb8b66ffe20f-kserve-router-schex4lqs to ip-10-0-139-47.ec2.internal [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:24:43 kserve-ci-e2e-test multus Normal AddedInterface Add eth0 [10.134.0.26/23] from ovn-kubernetes [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:24:43 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Pulled Container image "ghcr.io/llm-d/llm-d-router-endpoint-picker:v0.9.0-rc.2" already present on machine [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:24:43 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Created Created container: main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:24:43 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Started Started container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:26:33 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Killing Stopping container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:26:33 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Warning FailedPreStopHook PreStopHook failed [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 None kserve-ci-e2e-test Normal Scheduled Successfully assigned kserve-ci-e2e-test/precise-prefix-cache-test-kserve-5755bc98d9-f88w6 to ip-10-0-139-47.ec2.internal [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:55 kserve-ci-e2e-test multus Normal AddedInterface Add eth0 [10.134.0.22/23] from ovn-kubernetes [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:55 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Pulled Container image "quay.io/opendatahub/kserve-storage-initializer@sha256:13ad765c0f323200e1db7b45c99661287e65c5158d6b54b16139829636d20be4" already present on machine [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:55 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Created Created container: storage-initializer [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:55 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Started Started container storage-initializer [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:00 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Pulled Container image "ghcr.io/llm-d/llm-d-inference-sim:v0.8.2" already present on machine [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:01 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Created Created container: main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:01 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Started Started container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:23:10 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Killing Stopping container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 None kserve-ci-e2e-test Normal Scheduled Successfully assigned kserve-ci-e2e-test/precise-prefix-cache-test-kserve-5755bc98d9-hdpjr to ip-10-0-133-25.ec2.internal [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:55 kserve-ci-e2e-test multus Normal AddedInterface Add eth0 [10.132.0.45/23] from ovn-kubernetes [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:55 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Pulling Pulling image "quay.io/opendatahub/kserve-storage-initializer@sha256:13ad765c0f323200e1db7b45c99661287e65c5158d6b54b16139829636d20be4" [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:00 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Pulled Successfully pulled image "quay.io/opendatahub/kserve-storage-initializer@sha256:13ad765c0f323200e1db7b45c99661287e65c5158d6b54b16139829636d20be4" in 5.546s (5.546s including waiting). Image size: 301879255 bytes. [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:00 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Created Created container: storage-initializer [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:00 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Started Started container storage-initializer [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:07 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Pulling Pulling image "ghcr.io/llm-d/llm-d-inference-sim:v0.8.2" [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:08 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Pulled Successfully pulled image "ghcr.io/llm-d/llm-d-inference-sim:v0.8.2" in 1.434s (1.434s including waiting). Image size: 98346788 bytes. [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:09 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Created Created container: main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:09 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Started Started container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:23:10 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Killing Stopping container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:54 kserve-ci-e2e-test replicaset-controller Normal SuccessfulCreate Created pod: precise-prefix-cache-test-kserve-5755bc98d9-f88w6 [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:54 kserve-ci-e2e-test replicaset-controller Normal SuccessfulCreate Created pod: precise-prefix-cache-test-kserve-5755bc98d9-hdpjr [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:55 kserve-ci-e2e-test replicaset-controller Normal SuccessfulCreate Created pod: precise-prefix-cache-test-kserve-router-scheduler-647bfdf7jdncq [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 None kserve-ci-e2e-test Normal Scheduled Successfully assigned kserve-ci-e2e-test/precise-prefix-cache-test-kserve-router-scheduler-647bfdf7jdncq to ip-10-0-139-47.ec2.internal [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:55 kserve-ci-e2e-test multus Normal AddedInterface Add eth0 [10.134.0.23/23] from ovn-kubernetes [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:55 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Pulled Container image "quay.io/opendatahub/kserve-storage-initializer@sha256:13ad765c0f323200e1db7b45c99661287e65c5158d6b54b16139829636d20be4" already present on machine [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:55 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Created Created container: storage-initializer [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:55 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Started Started container storage-initializer [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:57 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Pulling Pulling image "ghcr.io/llm-d/llm-d-router-endpoint-picker:v0.9.0-rc.2" [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:58 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Pulled Successfully pulled image "ghcr.io/llm-d/llm-d-router-endpoint-picker:v0.9.0-rc.2" in 1.604s (1.604s including waiting). Image size: 75073927 bytes. [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:30 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Created Created container: main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:30 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Started Started container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:58 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Pulling Pulling image "ghcr.io/llm-d/llm-d-uds-tokenizer:vllm-v0.19.1" [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:29 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Pulled Successfully pulled image "ghcr.io/llm-d/llm-d-uds-tokenizer:vllm-v0.19.1" in 30.722s (30.722s including waiting). Image size: 2989890188 bytes. [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:29 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Created Created container: tokenizer [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:29 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Started Started container tokenizer [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:30 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Pulled Container image "ghcr.io/llm-d/llm-d-router-endpoint-picker:v0.9.0-rc.2" already present on machine [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:35 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Warning Unhealthy Startup probe failed: Get "http://10.134.0.23:8082/healthz": dial tcp 10.134.0.23:8082: connect: connection refused [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:46 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Warning Unhealthy Liveness probe failed: timeout: failed to connect service "10.134.0.23:9003" within 1s: context deadline exceeded [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:23:10 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Killing Stopping container tokenizer [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:23:10 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Killing Stopping container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:23:10 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Warning FailedPreStopHook PreStopHook failed [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:55 kserve-ci-e2e-test deployment-controller Normal ScalingReplicaSet Scaled up replica set precise-prefix-cache-test-kserve-router-scheduler-647bfdf79c from 0 to 1 [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:54 kserve-ci-e2e-test deployment-controller Normal ScalingReplicaSet Scaled up replica set precise-prefix-cache-test-kserve-5755bc98d9 from 0 to 2 [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:47 kserve-ci-e2e-test OpenDataHubModelController Warning ReconcileError Failed to reconcile LLMInferenceService: 1 error occurred: * failed to get HTTPRoute for AuthPolicy precise-prefix-cache-test-kserve-route-authn: failed to get HTTPRoute kserve-ci-e2e-test/precise-prefix-cache-test-kserve-route: HTTPRoute.gateway.networking.k8s.io "precise-prefix-cache-test-kserve-route" not found [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:54 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Secret kserve-ci-e2e-test/precise-prefix-cache-test-kserve-self-signed-certs [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:54 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Deployment kserve-ci-e2e-test/precise-prefix-cache-test-kserve [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:54 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Service kserve-ci-e2e-test/precise-prefix-cache-test-kserve-workload-svc [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:54 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.ServiceAccount kserve-ci-e2e-test/precise-prefix-cache-test-epp-sa [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:54 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.ClusterRoleBinding /kserve-ci-e2e-test-precise-prefix-cache-test-epp-auth-rb [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:54 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Role kserve-ci-e2e-test/precise-prefix-cache-test-epp-role [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:55 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.RoleBinding kserve-ci-e2e-test/precise-prefix-cache-test-epp-rb [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:55 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Deployment kserve-ci-e2e-test/precise-prefix-cache-test-kserve-router-scheduler [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:55 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Service kserve-ci-e2e-test/precise-prefix-cache-test-epp-service [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:26 kserve-ci-e2e-test LLMInferenceServiceController Normal Created (combined from similar events): Created v1.DestinationRule kserve-ci-e2e-test/precise-prefix-cache-test-kserve-shadow-svc [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:25 kserve-ci-e2e-test LLMInferenceServiceController Normal Updated Updated v1.Secret kserve-ci-e2e-test/precise-prefix-cache-test-kserve-self-signed-certs [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:26 kserve-ci-e2e-test LLMInferenceServiceController Normal Updated Updated v1.HTTPRoute kserve-ci-e2e-test/precise-prefix-cache-test-kserve-route [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:23:09 kserve-ci-e2e-test LLMInferenceServiceController Normal LLMInferenceServiceReady LLMInferenceService [precise-prefix-cache-test] is Ready [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:23:10 kserve-ci-e2e-test LLMInferenceServiceController Normal Deleted Deleted v1.ClusterRoleBinding /kserve-ci-e2e-test-precise-prefix-cache-test-epp-auth-rb [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 None kserve-ci-e2e-test Normal Scheduled Successfully assigned kserve-ci-e2e-test/router-gateway-1-openshift-default-799f46c59b-vt5f9 to ip-10-0-133-25.ec2.internal [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:36 kserve-ci-e2e-test multus Normal AddedInterface Add eth0 [10.132.0.44/23] from ovn-kubernetes [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:56 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Pulled Container image "registry.redhat.io/openshift-service-mesh/istio-proxyv2-rhel9@sha256:7d15cebf9b62f3f235c0eab5158ac8ff2fda86a1d193490dc94c301402c99da8" already present on machine [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:56 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Created Created container: istio-proxy [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:56 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Started Started container istio-proxy [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:54 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Warning Unhealthy Readiness probe failed: Get "http://10.132.0.44:15021/healthz/ready": context deadline exceeded (Client.Timeout exceeded while awaiting headers) [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:02 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Warning Unhealthy Startup probe failed: Get "http://10.132.0.44:15021/healthz/ready": dial tcp 10.132.0.44:15021: connect: connection refused [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:35 kserve-ci-e2e-test replicaset-controller Normal SuccessfulCreate Created pod: router-gateway-1-openshift-default-799f46c59b-vt5f9 [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:35 kserve-ci-e2e-test deployment-controller Normal ScalingReplicaSet Scaled up replica set router-gateway-1-openshift-default-799f46c59b from 0 to 1 [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:35 kserve-ci-e2e-test service-controller Normal EnsuringLoadBalancer Ensuring load balancer [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:40 kserve-ci-e2e-test service-controller Normal EnsuredLoadBalancer Ensured load balancer [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:30:52 kserve-ci-e2e-test gateway_labeler_controller Normal AddedLabel Added label istio.io/rev=openshift-gateway to gateway router-gateway-1 [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:148 ### Pod llmisvc-model-fb-opt-125m-route-87882a8e-kserve-5bd7864c6ctnql9 (phase=Running) [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:188 #### init-container 'storage-initializer' (restarts=0) [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:201 # -- logs (current) -- [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:202 2026-07-07 14:31:40.902 1 storage.initializer INFO [initializer-entrypoint:():17] Initializing, args: (src_uri, dest_path): [('hf://facebook/opt-125m', '/mnt/models')] [e2e-llm-inference-service] 2026-07-07 14:31:40.903 1 storage.initializer INFO [kserve_storage.py:download():166] Copying contents of hf://facebook/opt-125m to local [e2e-llm-inference-service] Could not set the permissions on the file '/mnt/models/.cache/huggingface/download/wPaCkH-WbT7GsmxMKKrNZTV4nSM=.ac481c8eb05e4d2496fbe076a38a7b4835dd733d.incomplete'. Error: [Errno 13] Permission denied: '/mnt/tmp_1eef8f63-8973-4aa5-8ca1-2f362240fe19'. [e2e-llm-inference-service] Continuing without setting permissions. [e2e-llm-inference-service] Could not set the permissions on the file '/mnt/models/.cache/huggingface/download/5HHJ6px3_ZRDOG3OxNZMhuycwOk=.a591333512516f58bf2002045dece909a0ccdb8b.incomplete'. Error: [Errno 13] Permission denied: '/mnt/tmp_3ca1efc3-8e4f-4d15-a745-7445c46f1858'. [e2e-llm-inference-service] Continuing without setting permissions. [e2e-llm-inference-service] Could not set the permissions on the file '/mnt/models/.cache/huggingface/download/Xn7B-BWUGOee2Y6hCZtEhtFu4BE=.38c05904caf6e5b9f04ecda5c973d77e6c1da151.incomplete'. Error: [Errno 13] Permission denied: '/mnt/tmp_c4677999-a51a-46c8-8bb8-fe4be7f5250c'. [e2e-llm-inference-service] Continuing without setting permissions. [e2e-llm-inference-service] Could not set the permissions on the file '/mnt/models/.cache/huggingface/download/8_PA_wEVGiVa2goH2H4KQOQpvVY=.b3fb716a3024261980becb2382e31a3780985130.incomplete'. Error: [Errno 13] Permission denied: '/mnt/tmp_a9eecabd-b9c1-4fc2-95b2-d76dceb07d90'. [e2e-llm-inference-service] Continuing without setting permissions. [e2e-llm-inference-service] Could not set the permissions on the file '/mnt/models/.cache/huggingface/download/gPcsVCQDYDHk-_n0G9uADl7PXIM=.61c60ec52ed43038fff0fbbd68b080c94b0d94b4c8458dbd65965f9b17631c89.incomplete'. Error: [Errno 13] Permission denied: '/mnt/tmp_39047282-ade9-4dd2-9f1c-3b37a42ee226'. [e2e-llm-inference-service] Continuing without setting permissions. [e2e-llm-inference-service] Could not set the permissions on the file '/mnt/models/.cache/huggingface/download/3EVKVggOldJcKSsGjSdoUCN1AyQ=.cf739e3ba86db7791ebab2828cc34b8a5acd3a86.incomplete'. Error: [Errno 13] Permission denied: '/mnt/tmp_604e9c3e-5a6e-4481-a96a-dcbecd51d28e'. [e2e-llm-inference-service] Continuing without setting permissions. [e2e-llm-inference-service] Could not set the permissions on the file '/mnt/models/.cache/huggingface/download/PtHk0z_I45atnj23IIRhTExwT3w=.226b0752cac7789c48f0cb3ec53eda48b7be36cc.incomplete'. Error: [Errno 13] Permission denied: '/mnt/tmp_d529176f-28aa-4cb9-bba3-64a0bd9161ff'. [e2e-llm-inference-service] Continuing without setting permissions. [e2e-llm-inference-service] Could not set the permissions on the file '/mnt/models/.cache/huggingface/download/Q1p2l2BzM1m6P5jKvr8WTq1TUio=.2d74da6615135c58cf3cf9ad4cb11e7c613ff9e55fe658a47ab83b6c8d1174a9.incomplete'. Error: [Errno 13] Permission denied: '/mnt/tmp_46f131c6-0092-4e99-b0b5-0f8a54b22fe4'. [e2e-llm-inference-service] Continuing without setting permissions. [e2e-llm-inference-service] Could not set the permissions on the file '/mnt/models/.cache/huggingface/download/ahkChHUJFxEmOdq5GDFEmerRzCY=.5dfa36546b8eddce0e04df3133c30df43fcc3828.incomplete'. Error: [Errno 13] Permission denied: '/mnt/tmp_3b437f89-858a-421e-95e0-57d30eff73c4'. [e2e-llm-inference-service] Continuing without setting permissions. [e2e-llm-inference-service] Could not set the permissions on the file '/mnt/models/.cache/huggingface/download/a7eHxRFT3OeMBIFg52k2nfj5m7w=.db7090b0c8b34dd957a7e0656c718f978f9203cc874018f37dda44108be5970a.incomplete'. Error: [Errno 13] Permission denied: '/mnt/tmp_37620263-7d52-4ef4-ba76-023a4ef5dd8e'. [e2e-llm-inference-service] Continuing without setting permissions. [e2e-llm-inference-service] Could not set the permissions on the file '/mnt/models/.cache/huggingface/download/vzaExXFZNBay89bvlQv-ZcI6BTg=.27c24ca9d908d0b678b20c698aeb9e950c44d865.incomplete'. Error: [Errno 13] Permission denied: '/mnt/tmp_8ed72b42-e15b-4015-8155-9fa36dc8b3e3'. [e2e-llm-inference-service] Continuing without setting permissions. [e2e-llm-inference-service] Could not set the permissions on the file '/mnt/models/.cache/huggingface/download/j3m-Hy6QvBddw8RXA1uSWl1AJ0c=.0a39732b2d8be8e493cab3da68b68cc3e28221de.incomplete'. Error: [Errno 13] Permission denied: '/mnt/tmp_5efb9074-4189-45d9-a8e4-681e57dcd813'. [e2e-llm-inference-service] Continuing without setting permissions. [e2e-llm-inference-service] 2026-07-07 14:31:45.290 1 storage.initializer INFO [kserve_storage.py:download():234] Successfully copied hf://facebook/opt-125m to /mnt/models [e2e-llm-inference-service] 2026-07-07 14:31:45.290 1 storage.initializer INFO [kserve_storage.py:download():235] Model downloaded in 4.38715659200011 seconds. [e2e-llm-inference-service] [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:188 #### container 'main' (restarts=0) [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:201 # -- logs (current) -- [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:202 (APIServer pid=1) DEBUG 07-07 14:46:00 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:46:01 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:46:02 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:46:03 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:46:04 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:46:05 [v1/metrics/loggers.py:259] Engine 000: Avg prompt throughput: 0.0 tokens/s, Avg generation throughput: 0.0 tokens/s, Running: 0 reqs, Waiting: 0 reqs, GPU KV cache usage: 0.0%, Prefix cache hit rate: 0.0% [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:46:05 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:46:06 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:46:07 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:46:08 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:46:09 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:46:10 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:46:10 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:46:11 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:46:12 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:46:13 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:46:14 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:46:15 [v1/metrics/loggers.py:259] Engine 000: Avg prompt throughput: 0.0 tokens/s, Avg generation throughput: 0.0 tokens/s, Running: 0 reqs, Waiting: 0 reqs, GPU KV cache usage: 0.0%, Prefix cache hit rate: 0.0% [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:46:15 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:46:16 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:46:17 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:46:18 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:46:19 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:46:20 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:46:20 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:46:21 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:46:22 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:46:23 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:46:24 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:46:25 [v1/metrics/loggers.py:259] Engine 000: Avg prompt throughput: 0.0 tokens/s, Avg generation throughput: 0.0 tokens/s, Running: 0 reqs, Waiting: 0 reqs, GPU KV cache usage: 0.0%, Prefix cache hit rate: 0.0% [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:46:25 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:46:26 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:46:27 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:46:28 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:46:29 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:46:30 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:46:30 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:46:31 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:46:32 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:46:33 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:46:34 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:46:35 [v1/metrics/loggers.py:259] Engine 000: Avg prompt throughput: 0.0 tokens/s, Avg generation throughput: 0.0 tokens/s, Running: 0 reqs, Waiting: 0 reqs, GPU KV cache usage: 0.0%, Prefix cache hit rate: 0.0% [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:46:35 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:46:36 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:46:37 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:46:38 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:46:39 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:46:40 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:46:40 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:46:41 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:46:42 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:46:43 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:46:44 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:46:45 [v1/metrics/loggers.py:259] Engine 000: Avg prompt throughput: 0.0 tokens/s, Avg generation throughput: 0.0 tokens/s, Running: 0 reqs, Waiting: 0 reqs, GPU KV cache usage: 0.0%, Prefix cache hit rate: 0.0% [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:46:45 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:46:46 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:46:47 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:46:48 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:46:49 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:46:50 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:46:50 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:46:51 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:46:52 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:46:53 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:46:54 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:46:55 [v1/metrics/loggers.py:259] Engine 000: Avg prompt throughput: 0.0 tokens/s, Avg generation throughput: 0.0 tokens/s, Running: 0 reqs, Waiting: 0 reqs, GPU KV cache usage: 0.0%, Prefix cache hit rate: 0.0% [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:46:55 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:46:56 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:46:57 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:46:58 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:46:59 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:47:00 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:47:00 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:47:01 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:47:02 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:47:03 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:47:04 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:47:05 [v1/metrics/loggers.py:259] Engine 000: Avg prompt throughput: 0.0 tokens/s, Avg generation throughput: 0.0 tokens/s, Running: 0 reqs, Waiting: 0 reqs, GPU KV cache usage: 0.0%, Prefix cache hit rate: 0.0% [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:47:05 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:47:06 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:47:07 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:47:08 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:47:09 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:47:10 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:47:10 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:47:11 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:47:12 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:47:13 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:47:14 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:47:15 [v1/metrics/loggers.py:259] Engine 000: Avg prompt throughput: 0.0 tokens/s, Avg generation throughput: 0.0 tokens/s, Running: 0 reqs, Waiting: 0 reqs, GPU KV cache usage: 0.0%, Prefix cache hit rate: 0.0% [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:47:15 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:47:16 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:47:17 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:47:18 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:47:19 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:47:20 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:47:20 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:47:21 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:47:22 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:47:23 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:47:24 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:47:25 [v1/metrics/loggers.py:259] Engine 000: Avg prompt throughput: 0.0 tokens/s, Avg generation throughput: 0.0 tokens/s, Running: 0 reqs, Waiting: 0 reqs, GPU KV cache usage: 0.0%, Prefix cache hit rate: 0.0% [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:47:25 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:47:26 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:47:27 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:47:28 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:47:29 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:47:30 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:47:30 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:47:31 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:47:32 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:47:33 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:47:34 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:47:35 [v1/metrics/loggers.py:259] Engine 000: Avg prompt throughput: 0.0 tokens/s, Avg generation throughput: 0.0 tokens/s, Running: 0 reqs, Waiting: 0 reqs, GPU KV cache usage: 0.0%, Prefix cache hit rate: 0.0% [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:47:35 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:47:36 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:47:37 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:47:38 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:47:39 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:47:40 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:47:40 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:47:41 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:47:42 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:47:43 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:47:44 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:47:45 [v1/metrics/loggers.py:259] Engine 000: Avg prompt throughput: 0.0 tokens/s, Avg generation throughput: 0.0 tokens/s, Running: 0 reqs, Waiting: 0 reqs, GPU KV cache usage: 0.0%, Prefix cache hit rate: 0.0% [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:47:45 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:47:46 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:47:47 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:47:48 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:47:49 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:47:50 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:47:50 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:47:51 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:47:52 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:47:53 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:47:54 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:47:55 [v1/metrics/loggers.py:259] Engine 000: Avg prompt throughput: 0.0 tokens/s, Avg generation throughput: 0.0 tokens/s, Running: 0 reqs, Waiting: 0 reqs, GPU KV cache usage: 0.0%, Prefix cache hit rate: 0.0% [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:47:55 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:47:56 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:47:57 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:47:58 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:47:59 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:48:00 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:48:00 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:48:01 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:48:02 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:48:03 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:48:04 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:48:05 [v1/metrics/loggers.py:259] Engine 000: Avg prompt throughput: 0.0 tokens/s, Avg generation throughput: 0.0 tokens/s, Running: 0 reqs, Waiting: 0 reqs, GPU KV cache usage: 0.0%, Prefix cache hit rate: 0.0% [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:48:05 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:48:06 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:48:07 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:48:08 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:48:09 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:48:10 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:48:10 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:48:11 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:48:12 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:48:13 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:48:14 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:48:15 [v1/metrics/loggers.py:259] Engine 000: Avg prompt throughput: 0.0 tokens/s, Avg generation throughput: 0.0 tokens/s, Running: 0 reqs, Waiting: 0 reqs, GPU KV cache usage: 0.0%, Prefix cache hit rate: 0.0% [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:48:15 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:48:16 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:48:17 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:48:18 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:48:19 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:48:20 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:48:20 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:48:21 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:48:22 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:48:23 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:48:24 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:48:25 [v1/metrics/loggers.py:259] Engine 000: Avg prompt throughput: 0.0 tokens/s, Avg generation throughput: 0.0 tokens/s, Running: 0 reqs, Waiting: 0 reqs, GPU KV cache usage: 0.0%, Prefix cache hit rate: 0.0% [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:48:25 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:48:26 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:48:27 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:48:28 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:48:29 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:48:30 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:48:30 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:48:31 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:48:32 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:48:33 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:48:34 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:48:35 [v1/metrics/loggers.py:259] Engine 000: Avg prompt throughput: 0.0 tokens/s, Avg generation throughput: 0.0 tokens/s, Running: 0 reqs, Waiting: 0 reqs, GPU KV cache usage: 0.0%, Prefix cache hit rate: 0.0% [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:48:35 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:48:36 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:48:37 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:48:38 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:48:39 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:48:40 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:48:40 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:48:41 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:48:42 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:48:43 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:48:44 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:48:45 [v1/metrics/loggers.py:259] Engine 000: Avg prompt throughput: 0.0 tokens/s, Avg generation throughput: 0.0 tokens/s, Running: 0 reqs, Waiting: 0 reqs, GPU KV cache usage: 0.0%, Prefix cache hit rate: 0.0% [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:48:45 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:48:46 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:148 ### Pod llmisvca690bbc929faec8bc98c767f16c003c1-kserve-router-scheq44kv (phase=Running) [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:188 #### init-container 'storage-initializer' (restarts=0) [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:201 # -- logs (current) -- [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:202 2026-07-07 14:31:41.330 1 storage.initializer INFO [initializer-entrypoint:():17] Initializing, args: (src_uri, dest_path): [('hf://facebook/opt-125m', '/mnt/models')] [e2e-llm-inference-service] 2026-07-07 14:31:41.330 1 storage.initializer INFO [kserve_storage.py:download():166] Copying contents of hf://facebook/opt-125m to local [e2e-llm-inference-service] 2026-07-07 14:31:41.330 1 storage.initializer INFO [kserve_storage.py:download():169] Allow patterns: ['tokenizer.json', 'tokenizer_config.json', 'special_tokens_map.json', 'vocab.json', 'merges.txt', 'config.json', 'generation_config.json'] [e2e-llm-inference-service] Could not set the permissions on the file '/mnt/models/.cache/huggingface/download/8_PA_wEVGiVa2goH2H4KQOQpvVY=.b3fb716a3024261980becb2382e31a3780985130.incomplete'. Error: [Errno 13] Permission denied: '/mnt/tmp_0e02c7ba-1c50-4034-957a-0e68033773d2'. [e2e-llm-inference-service] Continuing without setting permissions. [e2e-llm-inference-service] Could not set the permissions on the file '/mnt/models/.cache/huggingface/download/3EVKVggOldJcKSsGjSdoUCN1AyQ=.cf739e3ba86db7791ebab2828cc34b8a5acd3a86.incomplete'. Error: [Errno 13] Permission denied: '/mnt/tmp_2c93c76f-a57b-45cd-bb68-d819dc0be348'. [e2e-llm-inference-service] Continuing without setting permissions. [e2e-llm-inference-service] Could not set the permissions on the file '/mnt/models/.cache/huggingface/download/PtHk0z_I45atnj23IIRhTExwT3w=.226b0752cac7789c48f0cb3ec53eda48b7be36cc.incomplete'. Error: [Errno 13] Permission denied: '/mnt/tmp_1a7b97ca-36cd-4735-9d8c-b332d99c41f2'. [e2e-llm-inference-service] Continuing without setting permissions. [e2e-llm-inference-service] Could not set the permissions on the file '/mnt/models/.cache/huggingface/download/ahkChHUJFxEmOdq5GDFEmerRzCY=.5dfa36546b8eddce0e04df3133c30df43fcc3828.incomplete'. Error: [Errno 13] Permission denied: '/mnt/tmp_df0d43a4-bb35-4605-a958-1829ce660075'. [e2e-llm-inference-service] Continuing without setting permissions. [e2e-llm-inference-service] Could not set the permissions on the file '/mnt/models/.cache/huggingface/download/vzaExXFZNBay89bvlQv-ZcI6BTg=.27c24ca9d908d0b678b20c698aeb9e950c44d865.incomplete'. Error: [Errno 13] Permission denied: '/mnt/tmp_4c26d301-0301-49e9-93d0-0eda160eac88'. [e2e-llm-inference-service] Continuing without setting permissions. [e2e-llm-inference-service] Could not set the permissions on the file '/mnt/models/.cache/huggingface/download/j3m-Hy6QvBddw8RXA1uSWl1AJ0c=.0a39732b2d8be8e493cab3da68b68cc3e28221de.incomplete'. Error: [Errno 13] Permission denied: '/mnt/tmp_dcfeb0c7-a984-46a1-9596-ec0ed17dc098'. [e2e-llm-inference-service] Continuing without setting permissions. [e2e-llm-inference-service] 2026-07-07 14:31:41.757 1 storage.initializer INFO [kserve_storage.py:download():234] Successfully copied hf://facebook/opt-125m to /mnt/models [e2e-llm-inference-service] 2026-07-07 14:31:41.757 1 storage.initializer INFO [kserve_storage.py:download():235] Model downloaded in 0.4266467590000502 seconds. [e2e-llm-inference-service] [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:188 #### container 'main' (restarts=0) [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:201 # -- logs (current) -- [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:202 {"level":"info","ts":1783434702.9090648,"logger":"setup","caller":"runner/runner.go:196","msg":"GIE build","commit-sha":"181aa8358916e19b8844ccc752b2d6153d4b2ad6","build-ref":"v0.9.0-rc.2"} [e2e-llm-inference-service] Flag --model-server-metrics-scheme has been deprecated, This flag is deprecated. Configure via EndpointPickerConfig data layer plugin parameters instead. [e2e-llm-inference-service] {"level":"info","ts":1783434702.9091983,"logger":"setup","caller":"runner/runner.go:217","msg":"Flags processed","flags":{"cert-path":"/var/run/kserve/tls","config-file":"","config-text":"apiVersion: inference.networking.x-k8s.io/v1alpha1\nkind: EndpointPickerConfig\nplugins:\n- type: single-profile-handler\n- type: queue-scorer\n- type: prefix-cache-scorer\n- type: max-score-picker\nschedulingProfiles:\n- name: default\n plugins:\n - pluginRef: queue-scorer\n weight: 2\n - pluginRef: prefix-cache-scorer\n weight: 3\n - pluginRef: max-score-picker\n","disable-endpoint-subset-filter":false,"enable-cert-reload":true,"enable-grpc-stream-metrics":false,"enable-pprof":true,"endpoint-selector":"","endpoint-target-ports":{},"grpc-health-port":9003,"grpc-max-recv-msg-size":"","grpc-max-send-msg-size":"","grpc-port":9002,"ha-enable-leader-election":false,"health-checking":false,"metrics-endpoint-auth":true,"metrics-port":9090,"metrics-staleness-threshold":2000000000,"model-server-metrics-https-insecure-skip-verify":true,"model-server-metrics-path":"/metrics","model-server-metrics-port":0,"model-server-metrics-scheme":"https","pool-group":"inference.networking.k8s.io","pool-name":"llmisvc-model-fb-opt-125m-route-87882a8e-inference-pool","pool-namespace":"kserve-ci-e2e-test","refresh-metrics-interval":50000000,"refresh-prometheus-metrics-interval":5000000000,"secure-serving":true,"tracing":true,"v":2,"zap-devel":{},"zap-encoder":{},"zap-log-level":{},"zap-stacktrace-level":{},"zap-time-encoding":{}}} [e2e-llm-inference-service] {"level":"info","ts":1783434702.9095712,"logger":"setup.trace","caller":"tracing/telemetry.go:123","msg":"init OTel trace exporter","type":"console"} [e2e-llm-inference-service] {"level":"info","ts":1783434702.9112048,"caller":"loader/configloader.go:89","msg":"DEPRECATION: apiVersion inference.networking.x-k8s.io/v1alpha1/EndpointPickerConfig is deprecated","replacement":"llm-d.ai/v1alpha1/EndpointPickerConfig"} [e2e-llm-inference-service] {"level":"info","ts":1783434702.9112492,"caller":"loader/configloader.go:121","msg":"Loaded raw configuration","config":"{Plugins: [{Type: single-profile-handler} {Type: queue-scorer} {Type: prefix-cache-scorer} {Type: max-score-picker}], SchedulingProfiles: [{Name: default, Plugins: [{PluginRef: queue-scorer, Weight: 2.00} {PluginRef: prefix-cache-scorer, Weight: 3.00} {PluginRef: max-score-picker}]}]}"} [e2e-llm-inference-service] {"level":"info","ts":1783434702.9112594,"logger":"setup","caller":"runner/runner.go:622","msg":"Data layer: ENABLED"} [e2e-llm-inference-service] {"level":"info","ts":1783434702.9115467,"logger":"setup","caller":"runner/runner.go:281","msg":"Raw config after phase one","config":{"apiVersion":"inference.networking.x-k8s.io/v1alpha1","dataLayer":null,"kind":"EndpointPickerConfig","plugins":[{"name":"single-profile-handler","parameters":null,"type":"single-profile-handler"},{"name":"queue-scorer","parameters":null,"type":"queue-scorer"},{"name":"prefix-cache-scorer","parameters":null,"type":"prefix-cache-scorer"},{"name":"max-score-picker","parameters":null,"type":"max-score-picker"}],"schedulingProfiles":[{"name":"default","plugins":[{"pluginRef":"queue-scorer","weight":2},{"pluginRef":"prefix-cache-scorer","weight":3},{"pluginRef":"max-score-picker","weight":null}]}]}} [e2e-llm-inference-service] {"level":"info","ts":1783434702.9362562,"logger":"utilization-detector/utilization-detector","caller":"utilization/detector.go:83","msg":"Creating new UtilizationDetector","queueDepthThreshold":5,"kvCacheUtilThreshold":0.8,"metricsStalenessThreshold":"200ms","headroom":0} [e2e-llm-inference-service] {"level":"info","ts":1783434702.9363484,"caller":"metrics/factories.go:230","msg":"Registered engine mapping","engine":"vllm","mapping":"Mapping{all specs enabled}"} [e2e-llm-inference-service] {"level":"info","ts":1783434702.9363823,"caller":"metrics/factories.go:230","msg":"Registered engine mapping","engine":"sglang","mapping":"Mapping{disabled: [lora]}"} [e2e-llm-inference-service] {"level":"info","ts":1783434702.9364195,"caller":"metrics/factories.go:230","msg":"Registered engine mapping","engine":"trtllm-serve","mapping":"Mapping{disabled: [lora, cacheInfo]}"} [e2e-llm-inference-service] {"level":"info","ts":1783434702.9364827,"caller":"metrics/factories.go:230","msg":"Registered engine mapping","engine":"triton-tensorrt-llm","mapping":"Mapping{disabled: [lora, cacheInfo]}"} [e2e-llm-inference-service] {"level":"info","ts":1783434702.9365075,"caller":"metrics/factories.go:230","msg":"Registered engine mapping","engine":"triton","mapping":"Mapping{disabled: [kv, lora, cacheInfo]}"} [e2e-llm-inference-service] {"level":"info","ts":1783434702.9365592,"caller":"loader/configloader.go:154","msg":"Instantiated all plugins and applied system defaults. Effective raw configuration","config":"{Plugins: [{Name: single-profile-handler, Type: single-profile-handler} {Name: queue-scorer, Type: queue-scorer} {Name: prefix-cache-scorer, Type: prefix-cache-scorer} {Name: max-score-picker, Type: max-score-picker} {Name: fcfs-ordering-policy, Type: fcfs-ordering-policy} {Name: global-strict-fairness-policy, Type: global-strict-fairness-policy} {Name: static-usage-limit-policy, Type: static-usage-limit-policy} {Name: openai-parser, Type: openai-parser} {Name: anthropic-parser, Type: anthropic-parser} {Name: vllmhttp-parser, Type: vllmhttp-parser} {Name: utilization-detector, Type: utilization-detector} {Name: metrics-data-source, Type: metrics-data-source} {Name: core-metrics-extractor, Type: core-metrics-extractor}], SchedulingProfiles: [{Name: default, Plugins: [{PluginRef: queue-scorer, Weight: 2.00} {PluginRef: prefix-cache-scorer, Weight: 3.00} {PluginRef: max-score-picker}]}], DataLayer: {Sources: [{PluginRef: metrics-data-source, Extractors: [{PluginRef: core-metrics-extractor}]}], Discovery: }, FlowControl: {MaxBytes: unlimited, MaxRequests: unlimited, SaturationDetector: {PluginRef: utilization-detector}}, RequestHandler: {Parsers: [{PluginRef: openai-parser}, {PluginRef: anthropic-parser}, {PluginRef: vllmhttp-parser}]}}"} [e2e-llm-inference-service] {"level":"info","ts":1783434702.936613,"caller":"approximateprefix/plugin.go:88","msg":"Prefix DataProducer initialized","config":{"autoTune":true,"blockSizeTokens":16,"blockSize":0,"maxPrefixBlocksToMatch":2048,"maxPrefixTokensToMatch":131072,"lruCapacityPerServer":31250}} [e2e-llm-inference-service] {"level":"info","ts":1783434702.936699,"caller":"approximateprefix/plugin.go:111","msg":"WARNING: configured blockSizeTokens is below the recommended minimum, overriding it.","blockSizeTokens":16,"minimum":64,"issue":"https://github.com/llm-d/llm-d-router/issues/1158"} [e2e-llm-inference-service] {"level":"info","ts":1783434702.9367456,"caller":"datalayer/data_graph.go:116","msg":"auto-created default producer","producer":"approx-prefix-cache-producer/approx-prefix-cache-producer","dataKey":"PrefixCacheMatchInfoDataKey/approx-prefix-cache-producer","consumer":"prefix-cache-scorer"} [e2e-llm-inference-service] {"level":"info","ts":1783434702.9367864,"caller":"datalayer/data_graph.go:116","msg":"auto-created default producer","producer":"token-producer/token-producer","dataKey":"TokenizedPrompt/token-producer","consumer":"approx-prefix-cache-producer"} [e2e-llm-inference-service] {"level":"info","ts":1783434702.9369216,"caller":"runner/runner.go:685","msg":"loaded configuration from file/text successfully"} [e2e-llm-inference-service] {"level":"info","ts":1783434702.9369345,"logger":"setup","caller":"runner/runner.go:308","msg":"EPP config after phase two","config":"{SchedulerConfig:{ProfileHandler: single-profile-handler/single-profile-handler, Profiles: map[default:{Filters: [], Scorers: [queue-scorer/queue-scorer: 2.000000, prefix-cache-scorer/prefix-cache-scorer: 3.000000], Picker: max-score-picker/max-score-picker}]} SaturationDetector:0xc000815440 DataConfig:{Sources:[{Plugin:0xc0004bce10 Extractors:[0xc000815640]}]} FlowControlConfig: ParserRegistry:0xc000815ac0}"} [e2e-llm-inference-service] {"level":"info","ts":1783434702.961596,"logger":"setup","caller":"runner/runner.go:352","msg":"Setting pprof handlers"} [e2e-llm-inference-service] {"level":"info","ts":1783434702.9616292,"caller":"manager/internal.go:201","msg":"Registering metrics http server extra handler","path":"/debug/pprof/goroutine"} [e2e-llm-inference-service] {"level":"info","ts":1783434702.9616442,"caller":"manager/internal.go:201","msg":"Registering metrics http server extra handler","path":"/debug/pprof/mutex"} [e2e-llm-inference-service] {"level":"info","ts":1783434702.9616494,"caller":"manager/internal.go:201","msg":"Registering metrics http server extra handler","path":"/debug/pprof/"} [e2e-llm-inference-service] {"level":"info","ts":1783434702.9616542,"caller":"manager/internal.go:201","msg":"Registering metrics http server extra handler","path":"/debug/pprof/profile"} [e2e-llm-inference-service] {"level":"info","ts":1783434702.961659,"caller":"manager/internal.go:201","msg":"Registering metrics http server extra handler","path":"/debug/pprof/symbol"} [e2e-llm-inference-service] {"level":"info","ts":1783434702.9616632,"caller":"manager/internal.go:201","msg":"Registering metrics http server extra handler","path":"/debug/pprof/trace"} [e2e-llm-inference-service] {"level":"info","ts":1783434702.9616678,"caller":"manager/internal.go:201","msg":"Registering metrics http server extra handler","path":"/debug/pprof/allocs"} [e2e-llm-inference-service] {"level":"info","ts":1783434702.9616723,"caller":"manager/internal.go:201","msg":"Registering metrics http server extra handler","path":"/debug/pprof/threadcreate"} [e2e-llm-inference-service] {"level":"info","ts":1783434702.961677,"caller":"manager/internal.go:201","msg":"Registering metrics http server extra handler","path":"/debug/pprof/block"} [e2e-llm-inference-service] {"level":"info","ts":1783434702.9616816,"caller":"manager/internal.go:201","msg":"Registering metrics http server extra handler","path":"/debug/pprof/cmdline"} [e2e-llm-inference-service] {"level":"info","ts":1783434702.9616864,"caller":"manager/internal.go:201","msg":"Registering metrics http server extra handler","path":"/debug/pprof/heap"} [e2e-llm-inference-service] {"level":"info","ts":1783434702.9616983,"caller":"manager/internal.go:201","msg":"Registering metrics http server extra handler","path":"/debug/plugins/state"} [e2e-llm-inference-service] {"level":"info","ts":1783434702.9617028,"logger":"setup","caller":"runner/runner.go:373","msg":"parsed config","scheduler-config":"{ProfileHandler: single-profile-handler/single-profile-handler, Profiles: map[default:{Filters: [], Scorers: [queue-scorer/queue-scorer: 2.000000, prefix-cache-scorer/prefix-cache-scorer: 3.000000], Picker: max-score-picker/max-score-picker}]}"} [e2e-llm-inference-service] {"level":"info","ts":1783434702.9617293,"logger":"setup","caller":"datalayer/runtime.go:99","msg":"Configuring datalayer runtime","numSources":1} [e2e-llm-inference-service] {"level":"info","ts":1783434702.9617376,"logger":"setup","caller":"datalayer/runtime.go:118","msg":"Processing source","source":"metrics-data-source","numExtractors":1} [e2e-llm-inference-service] {"level":"info","ts":1783434702.9617527,"logger":"setup","caller":"datalayer/runtime.go:147","msg":"Source configured","source":"metrics-data-source","extractors":["core-metrics-extractor/core-metrics-extractor"]} [e2e-llm-inference-service] {"level":"info","ts":1783434702.9617655,"logger":"setup","caller":"datalayer/runtime.go:206","msg":"Datalayer runtime configured","pollers":1,"notifiers":0,"endpointSources":0} [e2e-llm-inference-service] {"level":"info","ts":1783434702.9617746,"logger":"setup","caller":"runner/runner.go:833","msg":"Experimental Flow Control layer is disabled, using legacy admission control"} [e2e-llm-inference-service] {"level":"info","ts":1783434702.9618921,"logger":"setup","caller":"runner/runner.go:721","msg":"ExtProc server runner added to manager."} [e2e-llm-inference-service] {"level":"info","ts":1783434702.961903,"logger":"setup","caller":"runner/runner.go:260","msg":"Controller manager starting"} [e2e-llm-inference-service] {"level":"info","ts":1783434702.9619296,"logger":"controller-runtime.metrics","caller":"server/server.go:208","msg":"Starting metrics server"} [e2e-llm-inference-service] {"level":"info","ts":1783434702.9621432,"caller":"runnable/grpc.go:35","msg":"gRPC server starting","name":"health"} [e2e-llm-inference-service] {"level":"info","ts":1783434702.9622252,"caller":"runnable/grpc.go:43","msg":"gRPC server listening","name":"health","port":9003} [e2e-llm-inference-service] {"level":"info","ts":1783434702.9622834,"logger":"controller-runtime.metrics","caller":"server/server.go:247","msg":"Serving metrics server","bindAddress":":9090","secure":false} [e2e-llm-inference-service] {"level":"info","ts":1783434702.9624577,"caller":"controller/controller.go:370","msg":"Starting EventSource","controller":"inferencepool","controllerGroup":"inference.networking.k8s.io","controllerKind":"InferencePool","source":"kind source: *v1.InferencePool"} [e2e-llm-inference-service] {"level":"info","ts":1783434702.9624846,"caller":"controller/controller.go:370","msg":"Starting EventSource","controller":"inferencemodelrewrite","controllerGroup":"inference.networking.x-k8s.io","controllerKind":"InferenceModelRewrite","source":"kind source: *v1alpha2.InferenceModelRewrite"} [e2e-llm-inference-service] {"level":"info","ts":1783434702.9628294,"caller":"controller/controller.go:370","msg":"Starting EventSource","controller":"inferenceobjective","controllerGroup":"inference.networking.x-k8s.io","controllerKind":"InferenceObjective","source":"kind source: *v1alpha2.InferenceObjective"} [e2e-llm-inference-service] {"level":"info","ts":1783434702.9629414,"caller":"controller/controller.go:370","msg":"Starting EventSource","controller":"pod","controllerGroup":"","controllerKind":"Pod","source":"kind source: *v1.Pod"} [e2e-llm-inference-service] {"level":"info","ts":1783434702.963434,"caller":"runnable/grpc.go:35","msg":"gRPC server starting","name":"ext-proc"} [e2e-llm-inference-service] {"level":"info","ts":1783434702.9635067,"caller":"runnable/grpc.go:43","msg":"gRPC server listening","name":"ext-proc","port":9002} [e2e-llm-inference-service] {"level":"info","ts":1783434702.9689121,"logger":"controller-runtime.cache","caller":"cache/reflector.go:446","msg":"Caches populated","type":"*v1alpha2.InferenceObjective","reflector":"pkg/mod/k8s.io/client-go@v0.35.6/tools/cache/reflector.go:289"} [e2e-llm-inference-service] {"level":"info","ts":1783434702.9689252,"logger":"controller-runtime.cache","caller":"cache/reflector.go:446","msg":"Caches populated","type":"*v1alpha2.InferenceModelRewrite","reflector":"pkg/mod/k8s.io/client-go@v0.35.6/tools/cache/reflector.go:289"} [e2e-llm-inference-service] {"level":"info","ts":1783434702.9701436,"logger":"controller-runtime.cache","caller":"cache/reflector.go:446","msg":"Caches populated","type":"*v1.Pod","reflector":"pkg/mod/k8s.io/client-go@v0.35.6/tools/cache/reflector.go:289"} [e2e-llm-inference-service] {"level":"info","ts":1783434702.9702542,"logger":"controller-runtime.cache","caller":"cache/reflector.go:446","msg":"Caches populated","type":"*v1.InferencePool","reflector":"pkg/mod/k8s.io/client-go@v0.35.6/tools/cache/reflector.go:289"} [e2e-llm-inference-service] {"level":"info","ts":1783434703.0631862,"caller":"controller/controller.go:303","msg":"Starting Controller","controller":"inferenceobjective","controllerGroup":"inference.networking.x-k8s.io","controllerKind":"InferenceObjective"} [e2e-llm-inference-service] {"level":"info","ts":1783434703.0632215,"caller":"controller/controller.go:306","msg":"Starting workers","controller":"inferenceobjective","controllerGroup":"inference.networking.x-k8s.io","controllerKind":"InferenceObjective","worker count":1} [e2e-llm-inference-service] {"level":"info","ts":1783434703.0631862,"caller":"controller/controller.go:303","msg":"Starting Controller","controller":"inferencemodelrewrite","controllerGroup":"inference.networking.x-k8s.io","controllerKind":"InferenceModelRewrite"} [e2e-llm-inference-service] {"level":"info","ts":1783434703.0632522,"caller":"controller/controller.go:306","msg":"Starting workers","controller":"inferencemodelrewrite","controllerGroup":"inference.networking.x-k8s.io","controllerKind":"InferenceModelRewrite","worker count":1} [e2e-llm-inference-service] {"level":"info","ts":1783434703.163454,"caller":"controller/controller.go:303","msg":"Starting Controller","controller":"inferencepool","controllerGroup":"inference.networking.k8s.io","controllerKind":"InferencePool"} [e2e-llm-inference-service] {"level":"info","ts":1783434703.1634703,"caller":"controller/controller.go:303","msg":"Starting Controller","controller":"pod","controllerGroup":"","controllerKind":"Pod"} [e2e-llm-inference-service] {"level":"info","ts":1783434703.1634834,"caller":"controller/controller.go:306","msg":"Starting workers","controller":"inferencepool","controllerGroup":"inference.networking.k8s.io","controllerKind":"InferencePool","worker count":1} [e2e-llm-inference-service] {"level":"info","ts":1783434703.1634912,"caller":"controller/controller.go:306","msg":"Starting workers","controller":"pod","controllerGroup":"","controllerKind":"Pod","worker count":1} [e2e-llm-inference-service] {"level":"info","ts":1783434703.163628,"caller":"controller/inferencepool_reconciler.go:46","msg":"Reconciling InferencePool","controller":"inferencepool","controllerGroup":"inference.networking.k8s.io","controllerKind":"InferencePool","InferencePool":{"name":"llmisvc-model-fb-opt-125m-route-87882a8e-inference-pool","namespace":"kserve-ci-e2e-test"},"namespace":"kserve-ci-e2e-test","name":"llmisvc-model-fb-opt-125m-route-87882a8e-inference-pool","reconcileID":"49d47884-00e2-4697-87b1-56fbf49b7cd2"} [e2e-llm-inference-service] {"level":"info","ts":1783434706.3161318,"caller":"controller/inferencepool_reconciler.go:46","msg":"Reconciling InferencePool","controller":"inferencepool","controllerGroup":"inference.networking.k8s.io","controllerKind":"InferencePool","InferencePool":{"name":"llmisvc-model-fb-opt-125m-route-87882a8e-inference-pool","namespace":"kserve-ci-e2e-test"},"namespace":"kserve-ci-e2e-test","name":"llmisvc-model-fb-opt-125m-route-87882a8e-inference-pool","reconcileID":"788b3afb-1555-4121-ba6d-4340bc9bf76a"} [e2e-llm-inference-service] {"level":"info","ts":1783434820.51026,"caller":"controller/pod_reconciler.go:99","msg":"Pod already exists","controller":"pod","controllerGroup":"","controllerKind":"Pod","Pod":{"name":"llmisvc-model-fb-opt-125m-route-87882a8e-kserve-5bd7864c6ctnql9","namespace":"kserve-ci-e2e-test"},"namespace":"kserve-ci-e2e-test","name":"llmisvc-model-fb-opt-125m-route-87882a8e-kserve-5bd7864c6ctnql9","reconcileID":"3665ef00-9093-455a-86f8-bea10b7c7674"} [e2e-llm-inference-service] [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1254 --- [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1255 metadata: [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-route-87882a8e-epp-service [e2e-llm-inference-service] namespace: kserve-ci-e2e-test [e2e-llm-inference-service] uid: abb4f84f-2bba-481f-90bd-479198fd420f [e2e-llm-inference-service] resourceVersion: '31666' [e2e-llm-inference-service] creationTimestamp: '2026-07-07T14:31:40Z' [e2e-llm-inference-service] labels: [e2e-llm-inference-service] app.kubernetes.io/component: llminferenceservice-router-scheduler [e2e-llm-inference-service] app.kubernetes.io/name: llmisvc-model-fb-opt-125m-route-87882a8e [e2e-llm-inference-service] app.kubernetes.io/part-of: llminferenceservice [e2e-llm-inference-service] endpoints.kubernetes.io/managed-by: endpoint-controller [e2e-llm-inference-service] annotations: [e2e-llm-inference-service] endpoints.kubernetes.io/last-change-trigger-time: '2026-07-07T14:32:14Z' [e2e-llm-inference-service] managedFields: [e2e-llm-inference-service] - manager: kube-controller-manager [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] apiVersion: v1 [e2e-llm-inference-service] time: '2026-07-07T14:32:14Z' [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:metadata: [e2e-llm-inference-service] f:annotations: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:endpoints.kubernetes.io/last-change-trigger-time: {} [e2e-llm-inference-service] f:labels: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:app.kubernetes.io/component: {} [e2e-llm-inference-service] f:app.kubernetes.io/name: {} [e2e-llm-inference-service] f:app.kubernetes.io/part-of: {} [e2e-llm-inference-service] f:endpoints.kubernetes.io/managed-by: {} [e2e-llm-inference-service] f:subsets: {} [e2e-llm-inference-service] subsets: [e2e-llm-inference-service] - addresses: [e2e-llm-inference-service] - ip: 10.132.0.53 [e2e-llm-inference-service] nodeName: ip-10-0-133-25.ec2.internal [e2e-llm-inference-service] targetRef: [e2e-llm-inference-service] kind: Pod [e2e-llm-inference-service] namespace: kserve-ci-e2e-test [e2e-llm-inference-service] name: llmisvca690bbc929faec8bc98c767f16c003c1-kserve-router-scheq44kv [e2e-llm-inference-service] uid: aa1b020a-833a-4de1-9ebe-0c421b524055 [e2e-llm-inference-service] ports: [e2e-llm-inference-service] - name: grpc-health [e2e-llm-inference-service] port: 9003 [e2e-llm-inference-service] protocol: TCP [e2e-llm-inference-service] - name: metrics [e2e-llm-inference-service] port: 9090 [e2e-llm-inference-service] protocol: TCP [e2e-llm-inference-service] - name: zmq [e2e-llm-inference-service] port: 5557 [e2e-llm-inference-service] protocol: TCP [e2e-llm-inference-service] - name: grpc [e2e-llm-inference-service] port: 9002 [e2e-llm-inference-service] protocol: TCP [e2e-llm-inference-service] apiVersion: v1 [e2e-llm-inference-service] kind: Endpoints [e2e-llm-inference-service] [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1254 --- [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1255 metadata: [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-route-87882a8e-kserve-workload-svc [e2e-llm-inference-service] namespace: kserve-ci-e2e-test [e2e-llm-inference-service] uid: 19e7ef83-5dbc-45b1-bedc-3d67bd54314f [e2e-llm-inference-service] resourceVersion: '33109' [e2e-llm-inference-service] creationTimestamp: '2026-07-07T14:31:40Z' [e2e-llm-inference-service] labels: [e2e-llm-inference-service] app.kubernetes.io/component: llminferenceservice-workload [e2e-llm-inference-service] app.kubernetes.io/name: llmisvc-model-fb-opt-125m-route-87882a8e [e2e-llm-inference-service] app.kubernetes.io/part-of: llminferenceservice [e2e-llm-inference-service] endpoints.kubernetes.io/managed-by: endpoint-controller [e2e-llm-inference-service] annotations: [e2e-llm-inference-service] endpoints.kubernetes.io/last-change-trigger-time: '2026-07-07T14:33:40Z' [e2e-llm-inference-service] managedFields: [e2e-llm-inference-service] - manager: kube-controller-manager [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] apiVersion: v1 [e2e-llm-inference-service] time: '2026-07-07T14:33:40Z' [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:metadata: [e2e-llm-inference-service] f:annotations: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:endpoints.kubernetes.io/last-change-trigger-time: {} [e2e-llm-inference-service] f:labels: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:app.kubernetes.io/component: {} [e2e-llm-inference-service] f:app.kubernetes.io/name: {} [e2e-llm-inference-service] f:app.kubernetes.io/part-of: {} [e2e-llm-inference-service] f:endpoints.kubernetes.io/managed-by: {} [e2e-llm-inference-service] f:subsets: {} [e2e-llm-inference-service] subsets: [e2e-llm-inference-service] - addresses: [e2e-llm-inference-service] - ip: 10.134.0.31 [e2e-llm-inference-service] nodeName: ip-10-0-139-47.ec2.internal [e2e-llm-inference-service] targetRef: [e2e-llm-inference-service] kind: Pod [e2e-llm-inference-service] namespace: kserve-ci-e2e-test [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-route-87882a8e-kserve-5bd7864c6ctnql9 [e2e-llm-inference-service] uid: 8abe33d5-7242-4f9f-a98e-f1814d77afd9 [e2e-llm-inference-service] ports: [e2e-llm-inference-service] - name: https [e2e-llm-inference-service] port: 8000 [e2e-llm-inference-service] protocol: TCP [e2e-llm-inference-service] appProtocol: https [e2e-llm-inference-service] apiVersion: v1 [e2e-llm-inference-service] kind: Endpoints [e2e-llm-inference-service] [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1254 --- [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1255 metadata: [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-route-87882a8e-kserve-5bd7864c6ctnql9 [e2e-llm-inference-service] generateName: llmisvc-model-fb-opt-125m-route-87882a8e-kserve-5bd7864c6c- [e2e-llm-inference-service] namespace: kserve-ci-e2e-test [e2e-llm-inference-service] uid: 8abe33d5-7242-4f9f-a98e-f1814d77afd9 [e2e-llm-inference-service] resourceVersion: '33105' [e2e-llm-inference-service] generation: 1 [e2e-llm-inference-service] creationTimestamp: '2026-07-07T14:31:40Z' [e2e-llm-inference-service] labels: [e2e-llm-inference-service] app.kubernetes.io/component: llminferenceservice-workload [e2e-llm-inference-service] app.kubernetes.io/name: llmisvc-model-fb-opt-125m-route-87882a8e [e2e-llm-inference-service] app.kubernetes.io/part-of: llminferenceservice [e2e-llm-inference-service] kserve.io/component: workload [e2e-llm-inference-service] llm-d.ai/role: both [e2e-llm-inference-service] pod-template-hash: 5bd7864c6c [e2e-llm-inference-service] annotations: [e2e-llm-inference-service] k8s.ovn.org/pod-networks: '{"default":{"ip_addresses":["10.134.0.31/23"],"mac_address":"0a:58:0a:86:00:1f","gateway_ips":["10.134.0.1"],"routes":[{"dest":"10.132.0.0/14","nextHop":"10.134.0.1"},{"dest":"172.31.0.0/16","nextHop":"10.134.0.1"},{"dest":"169.254.0.5/32","nextHop":"10.134.0.1"},{"dest":"100.64.0.0/16","nextHop":"10.134.0.1"}],"ip_address":"10.134.0.31/23","gateway_ip":"10.134.0.1","role":"primary"}}' [e2e-llm-inference-service] k8s.v1.cni.cncf.io/network-status: "[{\n \"name\": \"ovn-kubernetes\",\n \ [e2e-llm-inference-service] \ \"interface\": \"eth0\",\n \"ips\": [\n \"10.134.0.31\"\n ],\n\ [e2e-llm-inference-service] \ \"mac\": \"0a:58:0a:86:00:1f\",\n \"default\": true,\n \"dns\": {}\n\ [e2e-llm-inference-service] }]" [e2e-llm-inference-service] openshift.io/scc: restricted-v2 [e2e-llm-inference-service] seccomp.security.alpha.kubernetes.io/pod: runtime/default [e2e-llm-inference-service] security.openshift.io/validated-scc-subject-type: user [e2e-llm-inference-service] ownerReferences: [e2e-llm-inference-service] - apiVersion: apps/v1 [e2e-llm-inference-service] kind: ReplicaSet [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-route-87882a8e-kserve-5bd7864c6c [e2e-llm-inference-service] uid: 6072d6ea-6830-4187-b75d-75a1ebdb2b52 [e2e-llm-inference-service] controller: true [e2e-llm-inference-service] blockOwnerDeletion: true [e2e-llm-inference-service] managedFields: [e2e-llm-inference-service] - manager: ip-10-0-139-47 [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] apiVersion: v1 [e2e-llm-inference-service] time: '2026-07-07T14:31:40Z' [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:metadata: [e2e-llm-inference-service] f:annotations: [e2e-llm-inference-service] f:k8s.ovn.org/pod-networks: {} [e2e-llm-inference-service] subresource: status [e2e-llm-inference-service] - manager: kube-controller-manager [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] apiVersion: v1 [e2e-llm-inference-service] time: '2026-07-07T14:31:40Z' [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:metadata: [e2e-llm-inference-service] f:generateName: {} [e2e-llm-inference-service] f:labels: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:app.kubernetes.io/component: {} [e2e-llm-inference-service] f:app.kubernetes.io/name: {} [e2e-llm-inference-service] f:app.kubernetes.io/part-of: {} [e2e-llm-inference-service] f:kserve.io/component: {} [e2e-llm-inference-service] f:llm-d.ai/role: {} [e2e-llm-inference-service] f:pod-template-hash: {} [e2e-llm-inference-service] f:ownerReferences: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"uid":"6072d6ea-6830-4187-b75d-75a1ebdb2b52"}: {} [e2e-llm-inference-service] f:spec: [e2e-llm-inference-service] f:containers: [e2e-llm-inference-service] k:{"name":"main"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:command: {} [e2e-llm-inference-service] f:env: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"name":"HF_HUB_CACHE"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"HOME"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"TORCHINDUCTOR_CACHE_DIR"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"USER"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"VLLM_CPU_KVCACHE_SPACE"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"VLLM_ENABLE_V1_MULTIPROCESSING"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"VLLM_LOGGING_LEVEL"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] f:image: {} [e2e-llm-inference-service] f:imagePullPolicy: {} [e2e-llm-inference-service] f:lifecycle: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:preStop: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:exec: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:command: {} [e2e-llm-inference-service] f:livenessProbe: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:failureThreshold: {} [e2e-llm-inference-service] f:httpGet: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:path: {} [e2e-llm-inference-service] f:port: {} [e2e-llm-inference-service] f:scheme: {} [e2e-llm-inference-service] f:periodSeconds: {} [e2e-llm-inference-service] f:successThreshold: {} [e2e-llm-inference-service] f:timeoutSeconds: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:ports: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"containerPort":8000,"protocol":"TCP"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:containerPort: {} [e2e-llm-inference-service] f:protocol: {} [e2e-llm-inference-service] f:readinessProbe: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:failureThreshold: {} [e2e-llm-inference-service] f:httpGet: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:path: {} [e2e-llm-inference-service] f:port: {} [e2e-llm-inference-service] f:scheme: {} [e2e-llm-inference-service] f:periodSeconds: {} [e2e-llm-inference-service] f:successThreshold: {} [e2e-llm-inference-service] f:timeoutSeconds: {} [e2e-llm-inference-service] f:resources: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:limits: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:cpu: {} [e2e-llm-inference-service] f:memory: {} [e2e-llm-inference-service] f:requests: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:cpu: {} [e2e-llm-inference-service] f:memory: {} [e2e-llm-inference-service] f:securityContext: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:allowPrivilegeEscalation: {} [e2e-llm-inference-service] f:capabilities: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:drop: {} [e2e-llm-inference-service] f:readOnlyRootFilesystem: {} [e2e-llm-inference-service] f:runAsNonRoot: {} [e2e-llm-inference-service] f:seccompProfile: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:type: {} [e2e-llm-inference-service] f:startupProbe: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:failureThreshold: {} [e2e-llm-inference-service] f:httpGet: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:path: {} [e2e-llm-inference-service] f:port: {} [e2e-llm-inference-service] f:scheme: {} [e2e-llm-inference-service] f:periodSeconds: {} [e2e-llm-inference-service] f:successThreshold: {} [e2e-llm-inference-service] f:timeoutSeconds: {} [e2e-llm-inference-service] f:terminationMessagePath: {} [e2e-llm-inference-service] f:terminationMessagePolicy: {} [e2e-llm-inference-service] f:volumeMounts: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"mountPath":"/dev/shm"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:mountPath: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] k:{"mountPath":"/home"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:mountPath: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] k:{"mountPath":"/mnt/models"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:mountPath: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:readOnly: {} [e2e-llm-inference-service] k:{"mountPath":"/models"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:mountPath: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] k:{"mountPath":"/tmp"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:mountPath: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] k:{"mountPath":"/var/run/kserve/tls"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:mountPath: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:readOnly: {} [e2e-llm-inference-service] f:dnsPolicy: {} [e2e-llm-inference-service] f:enableServiceLinks: {} [e2e-llm-inference-service] f:initContainers: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"name":"storage-initializer"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:args: {} [e2e-llm-inference-service] f:env: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"name":"AWS_ACCESS_KEY_ID"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:valueFrom: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:secretKeyRef: {} [e2e-llm-inference-service] k:{"name":"AWS_CA_BUNDLE"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"AWS_CA_BUNDLE_CONFIGMAP"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"AWS_ENDPOINT_URL"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"AWS_SECRET_ACCESS_KEY"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:valueFrom: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:secretKeyRef: {} [e2e-llm-inference-service] k:{"name":"HF_HUB_ENABLE_HF_TRANSFER"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"HF_XET_HIGH_PERFORMANCE"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"HF_XET_NUM_CONCURRENT_RANGE_GETS"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"S3_ENDPOINT"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"S3_USE_HTTPS"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"S3_VERIFY_SSL"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] f:image: {} [e2e-llm-inference-service] f:imagePullPolicy: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:resources: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:limits: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:cpu: {} [e2e-llm-inference-service] f:memory: {} [e2e-llm-inference-service] f:requests: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:cpu: {} [e2e-llm-inference-service] f:memory: {} [e2e-llm-inference-service] f:terminationMessagePath: {} [e2e-llm-inference-service] f:terminationMessagePolicy: {} [e2e-llm-inference-service] f:volumeMounts: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"mountPath":"/mnt/models"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:mountPath: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:restartPolicy: {} [e2e-llm-inference-service] f:schedulerName: {} [e2e-llm-inference-service] f:securityContext: {} [e2e-llm-inference-service] f:terminationGracePeriodSeconds: {} [e2e-llm-inference-service] f:volumes: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"name":"dshm"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:emptyDir: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:medium: {} [e2e-llm-inference-service] f:sizeLimit: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] k:{"name":"home"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:emptyDir: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] k:{"name":"kserve-provision-location"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:emptyDir: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] k:{"name":"model-cache"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:emptyDir: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] k:{"name":"tls-certs"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:secret: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:defaultMode: {} [e2e-llm-inference-service] f:secretName: {} [e2e-llm-inference-service] k:{"name":"tmp-dir"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:emptyDir: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] - manager: multus-daemon [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] apiVersion: v1 [e2e-llm-inference-service] time: '2026-07-07T14:31:40Z' [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:metadata: [e2e-llm-inference-service] f:annotations: [e2e-llm-inference-service] f:k8s.v1.cni.cncf.io/network-status: {} [e2e-llm-inference-service] subresource: status [e2e-llm-inference-service] - manager: kubelet [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] apiVersion: v1 [e2e-llm-inference-service] time: '2026-07-07T14:33:40Z' [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:status: [e2e-llm-inference-service] f:conditions: [e2e-llm-inference-service] k:{"type":"ContainersReady"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:lastProbeTime: {} [e2e-llm-inference-service] f:lastTransitionTime: {} [e2e-llm-inference-service] f:observedGeneration: {} [e2e-llm-inference-service] f:status: {} [e2e-llm-inference-service] f:type: {} [e2e-llm-inference-service] k:{"type":"Initialized"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:lastProbeTime: {} [e2e-llm-inference-service] f:lastTransitionTime: {} [e2e-llm-inference-service] f:observedGeneration: {} [e2e-llm-inference-service] f:status: {} [e2e-llm-inference-service] f:type: {} [e2e-llm-inference-service] k:{"type":"PodReadyToStartContainers"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:lastProbeTime: {} [e2e-llm-inference-service] f:lastTransitionTime: {} [e2e-llm-inference-service] f:observedGeneration: {} [e2e-llm-inference-service] f:status: {} [e2e-llm-inference-service] f:type: {} [e2e-llm-inference-service] k:{"type":"PodScheduled"}: [e2e-llm-inference-service] f:observedGeneration: {} [e2e-llm-inference-service] k:{"type":"Ready"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:lastProbeTime: {} [e2e-llm-inference-service] f:lastTransitionTime: {} [e2e-llm-inference-service] f:observedGeneration: {} [e2e-llm-inference-service] f:status: {} [e2e-llm-inference-service] f:type: {} [e2e-llm-inference-service] f:containerStatuses: {} [e2e-llm-inference-service] f:hostIP: {} [e2e-llm-inference-service] f:hostIPs: {} [e2e-llm-inference-service] f:initContainerStatuses: {} [e2e-llm-inference-service] f:observedGeneration: {} [e2e-llm-inference-service] f:phase: {} [e2e-llm-inference-service] f:podIP: {} [e2e-llm-inference-service] f:podIPs: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"ip":"10.134.0.31"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:ip: {} [e2e-llm-inference-service] f:startTime: {} [e2e-llm-inference-service] subresource: status [e2e-llm-inference-service] spec: [e2e-llm-inference-service] volumes: [e2e-llm-inference-service] - name: home [e2e-llm-inference-service] emptyDir: {} [e2e-llm-inference-service] - name: dshm [e2e-llm-inference-service] emptyDir: [e2e-llm-inference-service] medium: Memory [e2e-llm-inference-service] sizeLimit: 1Gi [e2e-llm-inference-service] - name: model-cache [e2e-llm-inference-service] emptyDir: {} [e2e-llm-inference-service] - name: tmp-dir [e2e-llm-inference-service] emptyDir: {} [e2e-llm-inference-service] - name: tls-certs [e2e-llm-inference-service] secret: [e2e-llm-inference-service] secretName: llmisva690bbc929faec8bc98c767f16c003c1-kserve-self-signed-certs [e2e-llm-inference-service] defaultMode: 420 [e2e-llm-inference-service] - name: kserve-provision-location [e2e-llm-inference-service] emptyDir: {} [e2e-llm-inference-service] - name: kube-api-access-w65vh [e2e-llm-inference-service] projected: [e2e-llm-inference-service] sources: [e2e-llm-inference-service] - serviceAccountToken: [e2e-llm-inference-service] expirationSeconds: 3607 [e2e-llm-inference-service] path: token [e2e-llm-inference-service] - configMap: [e2e-llm-inference-service] name: kube-root-ca.crt [e2e-llm-inference-service] items: [e2e-llm-inference-service] - key: ca.crt [e2e-llm-inference-service] path: ca.crt [e2e-llm-inference-service] - downwardAPI: [e2e-llm-inference-service] items: [e2e-llm-inference-service] - path: namespace [e2e-llm-inference-service] fieldRef: [e2e-llm-inference-service] apiVersion: v1 [e2e-llm-inference-service] fieldPath: metadata.namespace [e2e-llm-inference-service] - configMap: [e2e-llm-inference-service] name: openshift-service-ca.crt [e2e-llm-inference-service] items: [e2e-llm-inference-service] - key: service-ca.crt [e2e-llm-inference-service] path: service-ca.crt [e2e-llm-inference-service] defaultMode: 420 [e2e-llm-inference-service] initContainers: [e2e-llm-inference-service] - name: storage-initializer [e2e-llm-inference-service] image: quay.io/opendatahub/kserve-storage-initializer@sha256:13ad765c0f323200e1db7b45c99661287e65c5158d6b54b16139829636d20be4 [e2e-llm-inference-service] args: [e2e-llm-inference-service] - hf://facebook/opt-125m [e2e-llm-inference-service] - /mnt/models [e2e-llm-inference-service] env: [e2e-llm-inference-service] - name: AWS_ACCESS_KEY_ID [e2e-llm-inference-service] valueFrom: [e2e-llm-inference-service] secretKeyRef: [e2e-llm-inference-service] name: seaweedfs-s3-creds [e2e-llm-inference-service] key: AWS_ACCESS_KEY_ID [e2e-llm-inference-service] - name: AWS_SECRET_ACCESS_KEY [e2e-llm-inference-service] valueFrom: [e2e-llm-inference-service] secretKeyRef: [e2e-llm-inference-service] name: seaweedfs-s3-creds [e2e-llm-inference-service] key: AWS_SECRET_ACCESS_KEY [e2e-llm-inference-service] - name: S3_USE_HTTPS [e2e-llm-inference-service] value: '0' [e2e-llm-inference-service] - name: S3_ENDPOINT [e2e-llm-inference-service] value: s3-service.kserve:8333 [e2e-llm-inference-service] - name: AWS_ENDPOINT_URL [e2e-llm-inference-service] value: http://s3-service.kserve:8333 [e2e-llm-inference-service] - name: S3_VERIFY_SSL [e2e-llm-inference-service] value: '0' [e2e-llm-inference-service] - name: AWS_CA_BUNDLE [e2e-llm-inference-service] value: /etc/ssl/custom-certs/cabundle.crt [e2e-llm-inference-service] - name: AWS_CA_BUNDLE_CONFIGMAP [e2e-llm-inference-service] value: odh-kserve-custom-ca-bundle [e2e-llm-inference-service] - name: HF_HUB_ENABLE_HF_TRANSFER [e2e-llm-inference-service] value: '1' [e2e-llm-inference-service] - name: HF_XET_HIGH_PERFORMANCE [e2e-llm-inference-service] value: '1' [e2e-llm-inference-service] - name: HF_XET_NUM_CONCURRENT_RANGE_GETS [e2e-llm-inference-service] value: '8' [e2e-llm-inference-service] resources: [e2e-llm-inference-service] limits: [e2e-llm-inference-service] cpu: '1' [e2e-llm-inference-service] memory: 24Gi [e2e-llm-inference-service] requests: [e2e-llm-inference-service] cpu: 100m [e2e-llm-inference-service] memory: 100Mi [e2e-llm-inference-service] volumeMounts: [e2e-llm-inference-service] - name: kserve-provision-location [e2e-llm-inference-service] mountPath: /mnt/models [e2e-llm-inference-service] - name: kube-api-access-w65vh [e2e-llm-inference-service] readOnly: true [e2e-llm-inference-service] mountPath: /var/run/secrets/kubernetes.io/serviceaccount [e2e-llm-inference-service] terminationMessagePath: /dev/termination-log [e2e-llm-inference-service] terminationMessagePolicy: FallbackToLogsOnError [e2e-llm-inference-service] imagePullPolicy: IfNotPresent [e2e-llm-inference-service] securityContext: [e2e-llm-inference-service] capabilities: [e2e-llm-inference-service] drop: [e2e-llm-inference-service] - ALL [e2e-llm-inference-service] runAsUser: 1000690000 [e2e-llm-inference-service] runAsNonRoot: true [e2e-llm-inference-service] allowPrivilegeEscalation: false [e2e-llm-inference-service] containers: [e2e-llm-inference-service] - name: main [e2e-llm-inference-service] image: public.ecr.aws/q9t5s3a7/vllm-cpu-release-repo:v0.19.0 [e2e-llm-inference-service] command: [e2e-llm-inference-service] - /bin/bash [e2e-llm-inference-service] - -c [e2e-llm-inference-service] - "if [ -f /etc/profile.d/ibm-aiu-setup.sh ]; then\n source /etc/profile.d/ibm-aiu-setup.sh\n\ [e2e-llm-inference-service] fi\n\nif [ \"$KSERVE_INFER_ROCE\" = \"true\" ]; then\n echo \"Trying to infer\ [e2e-llm-inference-service] \ RoCE configs ... \"\n grep -H . /sys/class/infiniband/*/ports/*/gids/* 2>/dev/null\n\ [e2e-llm-inference-service] \ grep -H . /sys/class/infiniband/*/ports/*/gid_attrs/types/* 2>/dev/null\n\ [e2e-llm-inference-service] \n cat /proc/driver/nvidia/params\n\n KSERVE_INFER_IB_GID_INDEX_GREP=${KSERVE_INFER_IB_GID_INDEX_GREP:-\"\ [e2e-llm-inference-service] RoCE v2\"}\n\n echo \"[Infer RoCE] Discovering active HCAs ...\"\n active_hcas=()\n\ [e2e-llm-inference-service] \ # Loop through all mlx5 devices found in sysfs\n for hca_dir in /sys/class/infiniband/mlx5_*;\ [e2e-llm-inference-service] \ do\n # Ensure it's a directory before proceeding\n if [ -d \"$hca_dir\"\ [e2e-llm-inference-service] \ ]; then\n hca_name=$(basename \"$hca_dir\")\n port_state_file=\"\ [e2e-llm-inference-service] $hca_dir/ports/1/state\" # Assume port 1\n type_file=\"$hca_dir/ports/1/gid_attrs/types/*\"\ [e2e-llm-inference-service] \n\n echo \"[Infer RoCE] Check if the port state file ${port_state_file}\ [e2e-llm-inference-service] \ exists and contains 'ACTIVE'\"\n if [ -f \"$port_state_file\" ] &&\ [e2e-llm-inference-service] \ grep -q \"ACTIVE\" \"$port_state_file\" && grep -q \"${KSERVE_INFER_IB_GID_INDEX_GREP}\"\ [e2e-llm-inference-service] \ ${type_file} 2>/dev/null; then\n echo \"[Infer RoCE] Found active\ [e2e-llm-inference-service] \ HCA: $hca_name\"\n active_hcas+=(\"$hca_name\")\n else\n\ [e2e-llm-inference-service] \ echo \"[Infer RoCE] Skipping inactive or down HCA: $hca_name\"\ [e2e-llm-inference-service] \n fi\n fi\n done\n\n # Check if we found any active HCAs\n\ [e2e-llm-inference-service] \ if [ ${#active_hcas[@]} -gt 0 ]; then\n # Join the array elements with\ [e2e-llm-inference-service] \ a comma\n hca_port_pairs=()\n for hca in \"${active_hcas[@]}\";\ [e2e-llm-inference-service] \ do\n hca_port_pairs+=(\"${hca}:1\")\n done\n\n active_hca_list=$(IFS=,;\ [e2e-llm-inference-service] \ echo \"${active_hcas[*]}\")\n hca_port_pairs_list=$(IFS=,; echo \"${hca_port_pairs[*]}\"\ [e2e-llm-inference-service] )\n echo \"[Infer RoCE] Setting active HCAs: ${active_hca_list}\"\n \ [e2e-llm-inference-service] \ export NCCL_IB_HCA=${NCCL_IB_HCA:-${active_hca_list}}\n export NVSHMEM_HCA_LIST=${NVSHMEM_HCA_LIST:-${hca_port_pairs_list}}\n\ [e2e-llm-inference-service] \ export UCX_NET_DEVICES=${UCX_NET_DEVICES:-${hca_port_pairs_list}}\n\n\ [e2e-llm-inference-service] \ echo \"[Infer RoCE] NCCL_IB_HCA=${NCCL_IB_HCA}\"\n echo \"[Infer\ [e2e-llm-inference-service] \ RoCE] NVSHMEM_HCA_LIST=${NVSHMEM_HCA_LIST}\"\n echo \"[Infer RoCE] UCX_NET_DEVICES=${UCX_NET_DEVICES}\"\ [e2e-llm-inference-service] \n else\n echo \"[Infer RoCE] WARNING: No active RoCE HCAs found. NCCL_IB_HCA\ [e2e-llm-inference-service] \ will not be set.\"\n fi\n\n if [ ${#active_hcas[@]} -gt 0 ]; then\n \ [e2e-llm-inference-service] \ echo \"[Infer RoCE] Finding GID_INDEX for each active HCA (SR-IOV compatible)...\"\ [e2e-llm-inference-service] \n\n # For SR-IOV environments, find the most common IPv4 RoCE v2 GID index\ [e2e-llm-inference-service] \ across all HCAs\n declare -A gid_index_count\n declare -A hca_gid_index\n\ [e2e-llm-inference-service] \n for hca_name in \"${active_hcas[@]}\"; do\n echo \"[Infer RoCE]\ [e2e-llm-inference-service] \ Processing HCA: ${hca_name}\"\n\n # Find all RoCE v2 IPv4 GIDs for\ [e2e-llm-inference-service] \ this HCA and count by index\n for tpath in /sys/class/infiniband/${hca_name}/ports/1/gid_attrs/types/*;\ [e2e-llm-inference-service] \ do\n if grep -q \"${KSERVE_INFER_IB_GID_INDEX_GREP}\" \"$tpath\"\ [e2e-llm-inference-service] \ 2>/dev/null; then\n idx=$(basename \"$tpath\")\n \ [e2e-llm-inference-service] \ gid_file=\"/sys/class/infiniband/${hca_name}/ports/1/gids/${idx}\"\ [e2e-llm-inference-service] \n # Check for IPv4 GID (contains ffff:)\n \ [e2e-llm-inference-service] \ if [ -f \"$gid_file\" ] && grep -q \"ffff:\" \"$gid_file\"; then\n \ [e2e-llm-inference-service] \ gid_value=$(cat \"$gid_file\" 2>/dev/null || echo \"\")\n \ [e2e-llm-inference-service] \ echo \"[Infer RoCE] Found IPv4 RoCE v2 GID for ${hca_name}:\ [e2e-llm-inference-service] \ index=${idx}, gid=${gid_value}\"\n hca_gid_index[\"${hca_name}\"\ [e2e-llm-inference-service] ]=\"${idx}\"\n gid_index_count[\"${idx}\"]=$((${gid_index_count[\"\ [e2e-llm-inference-service] ${idx}\"]} + 1))\n break # Use first found IPv4 GID per\ [e2e-llm-inference-service] \ HCA\n fi\n fi\n done\n done\n\n\ [e2e-llm-inference-service] \ # Find the most common GID index (most likely to be consistent across\ [e2e-llm-inference-service] \ nodes)\n best_gid_index=\"\"\n max_count=0\n for idx in \"\ [e2e-llm-inference-service] ${!gid_index_count[@]}\"; do\n count=${gid_index_count[\"${idx}\"]}\n\ [e2e-llm-inference-service] \ echo \"[Infer RoCE] GID_INDEX ${idx} found on ${count} HCAs\"\n \ [e2e-llm-inference-service] \ if [ $count -gt $max_count ]; then\n max_count=$count\n\ [e2e-llm-inference-service] \ best_gid_index=\"$idx\"\n fi\n done\n\n #\ [e2e-llm-inference-service] \ Use deterministic fallback if tied - prefer index 3 (SR-IOV standard)\n \ [e2e-llm-inference-service] \ if [ ${#gid_index_count[@]} -gt 1 ]; then\n echo \"[Infer RoCE]\ [e2e-llm-inference-service] \ Multiple GID indices found, selecting most common: ${best_gid_index}\"\n \ [e2e-llm-inference-service] \ # If there's a tie, prefer index 3 as it's most common in SR-IOV setups\n\ [e2e-llm-inference-service] \ if [ -n \"${gid_index_count['3']}\" ] && [ \"${gid_index_count['3']}\"\ [e2e-llm-inference-service] \ -eq \"$max_count\" ]; then\n best_gid_index=\"3\"\n \ [e2e-llm-inference-service] \ echo \"[Infer RoCE] Using deterministic fallback: GID_INDEX=3 (SR-IOV\ [e2e-llm-inference-service] \ standard)\"\n fi\n fi\n\n # Check if GID_INDEX is already\ [e2e-llm-inference-service] \ set via environment variables\n if [ -n \"${NCCL_IB_GID_INDEX}\" ]; then\n\ [e2e-llm-inference-service] \ echo \"[Infer RoCE] Using pre-configured NCCL_IB_GID_INDEX=${NCCL_IB_GID_INDEX}\ [e2e-llm-inference-service] \ from environment\"\n export NVSHMEM_IB_GID_INDEX=${NVSHMEM_IB_GID_INDEX:-$NCCL_IB_GID_INDEX}\n\ [e2e-llm-inference-service] \ export UCX_IB_GID_INDEX=${UCX_IB_GID_INDEX:-$NCCL_IB_GID_INDEX}\n\ [e2e-llm-inference-service] \ echo \"[Infer RoCE] Using pre-configured GID_INDEX=${NCCL_IB_GID_INDEX}\ [e2e-llm-inference-service] \ for NCCL, NVSHMEM, and UCX\"\n elif [ -n \"$best_gid_index\" ]; then\n\ [e2e-llm-inference-service] \ echo \"[Infer RoCE] Selected GID_INDEX: ${best_gid_index} (found\ [e2e-llm-inference-service] \ on ${max_count} HCAs)\"\n\n export NCCL_IB_GID_INDEX=${NCCL_IB_GID_INDEX:-$best_gid_index}\n\ [e2e-llm-inference-service] \ export NVSHMEM_IB_GID_INDEX=${NVSHMEM_IB_GID_INDEX:-$best_gid_index}\n\ [e2e-llm-inference-service] \ export UCX_IB_GID_INDEX=${UCX_IB_GID_INDEX:-$best_gid_index}\n\n\ [e2e-llm-inference-service] \ echo \"[Infer RoCE] Exported GID_INDEX=${best_gid_index} for NCCL,\ [e2e-llm-inference-service] \ NVSHMEM, and UCX\"\n else\n echo \"[Infer RoCE] ERROR: No valid\ [e2e-llm-inference-service] \ IPv4 ${KSERVE_INFER_IB_GID_INDEX_GREP} GID_INDEX found on any HCA.\"\n \ [e2e-llm-inference-service] \ fi\n else\n echo \"[Infer RoCE] No active HCAs found, skipping GID_INDEX\ [e2e-llm-inference-service] \ inference.\"\n fi\nfi\n\n# --disable-access-log-for-endpoints landed in vLLM\ [e2e-llm-inference-service] \ 0.16.0 (vllm-project/vllm#30011).\n# Older versions still need the blanket\ [e2e-llm-inference-service] \ --disable-uvicorn-access-log.\nACCESS_LOG_ARGS=\"--disable-uvicorn-access-log\"\ [e2e-llm-inference-service] \nVLLM_VERSION=$(vllm --version 2>/dev/null | tail -1 | awk '{print $NF}')\n\ [e2e-llm-inference-service] echo \"[access-log-detect] vllm version='${VLLM_VERSION}'\"\nif [[ \"$VLLM_VERSION\"\ [e2e-llm-inference-service] \ =~ ^[0-9]+\\.[0-9]+ ]] && [ \"$(printf '%s\\n%s\\n' \"0.16.0\" \"${VLLM_VERSION}\"\ [e2e-llm-inference-service] \ | sort -V | head -1)\" = \"0.16.0\" ]; then\n ACCESS_LOG_ARGS=\"--disable-access-log-for-endpoints\ [e2e-llm-inference-service] \ /health,/metrics,/ping\"\nfi\necho \"[access-log-detect] selected ACCESS_LOG_ARGS='${ACCESS_LOG_ARGS}'\"\ [e2e-llm-inference-service] \n\n# --shutdown-timeout landed in vLLM 0.18.0 (vllm-project/vllm#36666).\n\ [e2e-llm-inference-service] SHUTDOWN_TIMEOUT_ARGS=\"\"\nif [[ \"$VLLM_VERSION\" =~ ^[0-9]+\\.[0-9]+ ]] &&\ [e2e-llm-inference-service] \ [ \"$(printf '%s\\n%s\\n' \"0.18.0\" \"${VLLM_VERSION}\" | sort -V | head\ [e2e-llm-inference-service] \ -1)\" = \"0.18.0\" ]; then\n SHUTDOWN_TIMEOUT_ARGS=\"--shutdown-timeout 40\"\ [e2e-llm-inference-service] \nfi\n\neval \"exec vllm serve /mnt/models \\\n --served-model-name \"facebook/opt-125m\"\ [e2e-llm-inference-service] \ \"publishers/kserve-ci-e2e-test/models/facebook/opt-125m\" \\\n --port 8000\ [e2e-llm-inference-service] \ \\\n ${ACCESS_LOG_ARGS} \\\n ${SHUTDOWN_TIMEOUT_ARGS} \\\n --enable-ssl-refresh\ [e2e-llm-inference-service] \ \\\n --ssl-certfile /var/run/kserve/tls/tls.crt \\\n --ssl-keyfile /var/run/kserve/tls/tls.key\ [e2e-llm-inference-service] \ \\\n ${VLLM_ADDITIONAL_ARGS} \\\n $@\"" [e2e-llm-inference-service] - -- [e2e-llm-inference-service] ports: [e2e-llm-inference-service] - containerPort: 8000 [e2e-llm-inference-service] protocol: TCP [e2e-llm-inference-service] env: [e2e-llm-inference-service] - name: HOME [e2e-llm-inference-service] value: /home [e2e-llm-inference-service] - name: VLLM_LOGGING_LEVEL [e2e-llm-inference-service] value: DEBUG [e2e-llm-inference-service] - name: VLLM_CPU_KVCACHE_SPACE [e2e-llm-inference-service] value: '1' [e2e-llm-inference-service] - name: VLLM_ENABLE_V1_MULTIPROCESSING [e2e-llm-inference-service] value: '0' [e2e-llm-inference-service] - name: USER [e2e-llm-inference-service] value: nonroot [e2e-llm-inference-service] - name: TORCHINDUCTOR_CACHE_DIR [e2e-llm-inference-service] value: /tmp/torchinductor-cache [e2e-llm-inference-service] - name: HF_HUB_CACHE [e2e-llm-inference-service] value: /models [e2e-llm-inference-service] resources: [e2e-llm-inference-service] limits: [e2e-llm-inference-service] cpu: '2' [e2e-llm-inference-service] memory: 7Gi [e2e-llm-inference-service] requests: [e2e-llm-inference-service] cpu: 200m [e2e-llm-inference-service] memory: 2Gi [e2e-llm-inference-service] volumeMounts: [e2e-llm-inference-service] - name: home [e2e-llm-inference-service] mountPath: /home [e2e-llm-inference-service] - name: tmp-dir [e2e-llm-inference-service] mountPath: /tmp [e2e-llm-inference-service] - name: dshm [e2e-llm-inference-service] mountPath: /dev/shm [e2e-llm-inference-service] - name: model-cache [e2e-llm-inference-service] mountPath: /models [e2e-llm-inference-service] - name: tls-certs [e2e-llm-inference-service] readOnly: true [e2e-llm-inference-service] mountPath: /var/run/kserve/tls [e2e-llm-inference-service] - name: kserve-provision-location [e2e-llm-inference-service] readOnly: true [e2e-llm-inference-service] mountPath: /mnt/models [e2e-llm-inference-service] - name: kube-api-access-w65vh [e2e-llm-inference-service] readOnly: true [e2e-llm-inference-service] mountPath: /var/run/secrets/kubernetes.io/serviceaccount [e2e-llm-inference-service] livenessProbe: [e2e-llm-inference-service] httpGet: [e2e-llm-inference-service] path: /health [e2e-llm-inference-service] port: 8000 [e2e-llm-inference-service] scheme: HTTPS [e2e-llm-inference-service] timeoutSeconds: 1 [e2e-llm-inference-service] periodSeconds: 10 [e2e-llm-inference-service] successThreshold: 1 [e2e-llm-inference-service] failureThreshold: 10 [e2e-llm-inference-service] readinessProbe: [e2e-llm-inference-service] httpGet: [e2e-llm-inference-service] path: /health [e2e-llm-inference-service] port: 8000 [e2e-llm-inference-service] scheme: HTTPS [e2e-llm-inference-service] timeoutSeconds: 1 [e2e-llm-inference-service] periodSeconds: 1 [e2e-llm-inference-service] successThreshold: 1 [e2e-llm-inference-service] failureThreshold: 2 [e2e-llm-inference-service] startupProbe: [e2e-llm-inference-service] httpGet: [e2e-llm-inference-service] path: /health [e2e-llm-inference-service] port: 8000 [e2e-llm-inference-service] scheme: HTTPS [e2e-llm-inference-service] timeoutSeconds: 1 [e2e-llm-inference-service] periodSeconds: 10 [e2e-llm-inference-service] successThreshold: 1 [e2e-llm-inference-service] failureThreshold: 60 [e2e-llm-inference-service] lifecycle: [e2e-llm-inference-service] preStop: [e2e-llm-inference-service] exec: [e2e-llm-inference-service] command: [e2e-llm-inference-service] - /bin/sleep [e2e-llm-inference-service] - '15' [e2e-llm-inference-service] terminationMessagePath: /dev/termination-log [e2e-llm-inference-service] terminationMessagePolicy: FallbackToLogsOnError [e2e-llm-inference-service] imagePullPolicy: IfNotPresent [e2e-llm-inference-service] securityContext: [e2e-llm-inference-service] capabilities: [e2e-llm-inference-service] drop: [e2e-llm-inference-service] - ALL [e2e-llm-inference-service] runAsUser: 1000690000 [e2e-llm-inference-service] runAsNonRoot: true [e2e-llm-inference-service] readOnlyRootFilesystem: true [e2e-llm-inference-service] allowPrivilegeEscalation: false [e2e-llm-inference-service] seccompProfile: [e2e-llm-inference-service] type: RuntimeDefault [e2e-llm-inference-service] restartPolicy: Always [e2e-llm-inference-service] terminationGracePeriodSeconds: 60 [e2e-llm-inference-service] dnsPolicy: ClusterFirst [e2e-llm-inference-service] serviceAccountName: default [e2e-llm-inference-service] serviceAccount: default [e2e-llm-inference-service] nodeName: ip-10-0-139-47.ec2.internal [e2e-llm-inference-service] securityContext: [e2e-llm-inference-service] seLinuxOptions: [e2e-llm-inference-service] level: s0:c26,c20 [e2e-llm-inference-service] fsGroup: 1000690000 [e2e-llm-inference-service] seccompProfile: [e2e-llm-inference-service] type: RuntimeDefault [e2e-llm-inference-service] imagePullSecrets: [e2e-llm-inference-service] - name: default-dockercfg-tbbgr [e2e-llm-inference-service] schedulerName: default-scheduler [e2e-llm-inference-service] tolerations: [e2e-llm-inference-service] - key: node.kubernetes.io/not-ready [e2e-llm-inference-service] operator: Exists [e2e-llm-inference-service] effect: NoExecute [e2e-llm-inference-service] tolerationSeconds: 300 [e2e-llm-inference-service] - key: node.kubernetes.io/unreachable [e2e-llm-inference-service] operator: Exists [e2e-llm-inference-service] effect: NoExecute [e2e-llm-inference-service] tolerationSeconds: 300 [e2e-llm-inference-service] - key: node.kubernetes.io/memory-pressure [e2e-llm-inference-service] operator: Exists [e2e-llm-inference-service] effect: NoSchedule [e2e-llm-inference-service] priority: 0 [e2e-llm-inference-service] enableServiceLinks: true [e2e-llm-inference-service] preemptionPolicy: PreemptLowerPriority [e2e-llm-inference-service] status: [e2e-llm-inference-service] observedGeneration: 1 [e2e-llm-inference-service] phase: Running [e2e-llm-inference-service] conditions: [e2e-llm-inference-service] - type: PodReadyToStartContainers [e2e-llm-inference-service] observedGeneration: 1 [e2e-llm-inference-service] status: 'True' [e2e-llm-inference-service] lastProbeTime: null [e2e-llm-inference-service] lastTransitionTime: '2026-07-07T14:31:40Z' [e2e-llm-inference-service] - type: Initialized [e2e-llm-inference-service] observedGeneration: 1 [e2e-llm-inference-service] status: 'True' [e2e-llm-inference-service] lastProbeTime: null [e2e-llm-inference-service] lastTransitionTime: '2026-07-07T14:31:45Z' [e2e-llm-inference-service] - type: Ready [e2e-llm-inference-service] observedGeneration: 1 [e2e-llm-inference-service] status: 'True' [e2e-llm-inference-service] lastProbeTime: null [e2e-llm-inference-service] lastTransitionTime: '2026-07-07T14:33:40Z' [e2e-llm-inference-service] - type: ContainersReady [e2e-llm-inference-service] observedGeneration: 1 [e2e-llm-inference-service] status: 'True' [e2e-llm-inference-service] lastProbeTime: null [e2e-llm-inference-service] lastTransitionTime: '2026-07-07T14:33:40Z' [e2e-llm-inference-service] - type: PodScheduled [e2e-llm-inference-service] observedGeneration: 1 [e2e-llm-inference-service] status: 'True' [e2e-llm-inference-service] lastProbeTime: null [e2e-llm-inference-service] lastTransitionTime: '2026-07-07T14:31:40Z' [e2e-llm-inference-service] hostIP: 10.0.139.47 [e2e-llm-inference-service] hostIPs: [e2e-llm-inference-service] - ip: 10.0.139.47 [e2e-llm-inference-service] podIP: 10.134.0.31 [e2e-llm-inference-service] podIPs: [e2e-llm-inference-service] - ip: 10.134.0.31 [e2e-llm-inference-service] startTime: '2026-07-07T14:31:40Z' [e2e-llm-inference-service] initContainerStatuses: [e2e-llm-inference-service] - name: storage-initializer [e2e-llm-inference-service] state: [e2e-llm-inference-service] terminated: [e2e-llm-inference-service] exitCode: 0 [e2e-llm-inference-service] reason: Completed [e2e-llm-inference-service] startedAt: '2026-07-07T14:31:40Z' [e2e-llm-inference-service] finishedAt: '2026-07-07T14:31:45Z' [e2e-llm-inference-service] containerID: cri-o://80b747a3d03d170b45367c4a318c25a77fbd724260c5be81258335308e76277e [e2e-llm-inference-service] lastState: {} [e2e-llm-inference-service] ready: true [e2e-llm-inference-service] restartCount: 0 [e2e-llm-inference-service] image: quay.io/opendatahub/kserve-storage-initializer@sha256:13ad765c0f323200e1db7b45c99661287e65c5158d6b54b16139829636d20be4 [e2e-llm-inference-service] imageID: quay.io/opendatahub/kserve-storage-initializer@sha256:13ad765c0f323200e1db7b45c99661287e65c5158d6b54b16139829636d20be4 [e2e-llm-inference-service] containerID: cri-o://80b747a3d03d170b45367c4a318c25a77fbd724260c5be81258335308e76277e [e2e-llm-inference-service] started: false [e2e-llm-inference-service] allocatedResources: [e2e-llm-inference-service] cpu: 100m [e2e-llm-inference-service] memory: 100Mi [e2e-llm-inference-service] resources: [e2e-llm-inference-service] limits: [e2e-llm-inference-service] cpu: '1' [e2e-llm-inference-service] memory: 24Gi [e2e-llm-inference-service] requests: [e2e-llm-inference-service] cpu: 100m [e2e-llm-inference-service] memory: 100Mi [e2e-llm-inference-service] volumeMounts: [e2e-llm-inference-service] - name: kserve-provision-location [e2e-llm-inference-service] mountPath: /mnt/models [e2e-llm-inference-service] - name: kube-api-access-w65vh [e2e-llm-inference-service] mountPath: /var/run/secrets/kubernetes.io/serviceaccount [e2e-llm-inference-service] readOnly: true [e2e-llm-inference-service] recursiveReadOnly: Disabled [e2e-llm-inference-service] user: [e2e-llm-inference-service] linux: [e2e-llm-inference-service] uid: 1000690000 [e2e-llm-inference-service] gid: 0 [e2e-llm-inference-service] supplementalGroups: [e2e-llm-inference-service] - 0 [e2e-llm-inference-service] - 1000690000 [e2e-llm-inference-service] containerStatuses: [e2e-llm-inference-service] - name: main [e2e-llm-inference-service] state: [e2e-llm-inference-service] running: [e2e-llm-inference-service] startedAt: '2026-07-07T14:31:45Z' [e2e-llm-inference-service] lastState: {} [e2e-llm-inference-service] ready: true [e2e-llm-inference-service] restartCount: 0 [e2e-llm-inference-service] image: public.ecr.aws/q9t5s3a7/vllm-cpu-release-repo:v0.19.0 [e2e-llm-inference-service] imageID: public.ecr.aws/q9t5s3a7/vllm-cpu-release-repo@sha256:afb39fca138b51d019d986229d546531b45a2a3deb73bcf59bd42406e13fbba0 [e2e-llm-inference-service] containerID: cri-o://7365f30974f1b746fc81e7dc06b259d6ee3d74c6fae2bb99a9eefebacbe3f0b0 [e2e-llm-inference-service] started: true [e2e-llm-inference-service] allocatedResources: [e2e-llm-inference-service] cpu: 200m [e2e-llm-inference-service] memory: 2Gi [e2e-llm-inference-service] resources: [e2e-llm-inference-service] limits: [e2e-llm-inference-service] cpu: '2' [e2e-llm-inference-service] memory: 7Gi [e2e-llm-inference-service] requests: [e2e-llm-inference-service] cpu: 200m [e2e-llm-inference-service] memory: 2Gi [e2e-llm-inference-service] volumeMounts: [e2e-llm-inference-service] - name: home [e2e-llm-inference-service] mountPath: /home [e2e-llm-inference-service] - name: tmp-dir [e2e-llm-inference-service] mountPath: /tmp [e2e-llm-inference-service] - name: dshm [e2e-llm-inference-service] mountPath: /dev/shm [e2e-llm-inference-service] - name: model-cache [e2e-llm-inference-service] mountPath: /models [e2e-llm-inference-service] - name: tls-certs [e2e-llm-inference-service] mountPath: /var/run/kserve/tls [e2e-llm-inference-service] readOnly: true [e2e-llm-inference-service] recursiveReadOnly: Disabled [e2e-llm-inference-service] - name: kserve-provision-location [e2e-llm-inference-service] mountPath: /mnt/models [e2e-llm-inference-service] readOnly: true [e2e-llm-inference-service] recursiveReadOnly: Disabled [e2e-llm-inference-service] - name: kube-api-access-w65vh [e2e-llm-inference-service] mountPath: /var/run/secrets/kubernetes.io/serviceaccount [e2e-llm-inference-service] readOnly: true [e2e-llm-inference-service] recursiveReadOnly: Disabled [e2e-llm-inference-service] user: [e2e-llm-inference-service] linux: [e2e-llm-inference-service] uid: 1000690000 [e2e-llm-inference-service] gid: 0 [e2e-llm-inference-service] supplementalGroups: [e2e-llm-inference-service] - 0 [e2e-llm-inference-service] - 1000690000 [e2e-llm-inference-service] qosClass: Burstable [e2e-llm-inference-service] apiVersion: v1 [e2e-llm-inference-service] kind: Pod [e2e-llm-inference-service] [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1254 --- [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1255 metadata: [e2e-llm-inference-service] name: llmisvca690bbc929faec8bc98c767f16c003c1-kserve-router-scheq44kv [e2e-llm-inference-service] generateName: llmisvca690bbc929faec8bc98c767f16c003c1-kserve-router-scheduler-5d6998448f- [e2e-llm-inference-service] namespace: kserve-ci-e2e-test [e2e-llm-inference-service] uid: aa1b020a-833a-4de1-9ebe-0c421b524055 [e2e-llm-inference-service] resourceVersion: '31664' [e2e-llm-inference-service] generation: 1 [e2e-llm-inference-service] creationTimestamp: '2026-07-07T14:31:40Z' [e2e-llm-inference-service] labels: [e2e-llm-inference-service] app.kubernetes.io/component: llminferenceservice-router-scheduler [e2e-llm-inference-service] app.kubernetes.io/name: llmisvc-model-fb-opt-125m-route-87882a8e [e2e-llm-inference-service] app.kubernetes.io/part-of: llminferenceservice [e2e-llm-inference-service] pod-template-hash: 5d6998448f [e2e-llm-inference-service] annotations: [e2e-llm-inference-service] app.kubernetes.io/version: 0.9.0 [e2e-llm-inference-service] certificates.kserve.io/expiration-v2: 'true' [e2e-llm-inference-service] k8s.ovn.org/pod-networks: '{"default":{"ip_addresses":["10.132.0.53/23"],"mac_address":"0a:58:0a:84:00:35","gateway_ips":["10.132.0.1"],"routes":[{"dest":"10.132.0.0/14","nextHop":"10.132.0.1"},{"dest":"172.31.0.0/16","nextHop":"10.132.0.1"},{"dest":"169.254.0.5/32","nextHop":"10.132.0.1"},{"dest":"100.64.0.0/16","nextHop":"10.132.0.1"}],"ip_address":"10.132.0.53/23","gateway_ip":"10.132.0.1","role":"primary"}}' [e2e-llm-inference-service] k8s.v1.cni.cncf.io/network-status: "[{\n \"name\": \"ovn-kubernetes\",\n \ [e2e-llm-inference-service] \ \"interface\": \"eth0\",\n \"ips\": [\n \"10.132.0.53\"\n ],\n\ [e2e-llm-inference-service] \ \"mac\": \"0a:58:0a:84:00:35\",\n \"default\": true,\n \"dns\": {}\n\ [e2e-llm-inference-service] }]" [e2e-llm-inference-service] openshift.io/scc: restricted-v2 [e2e-llm-inference-service] seccomp.security.alpha.kubernetes.io/pod: runtime/default [e2e-llm-inference-service] security.openshift.io/validated-scc-subject-type: user [e2e-llm-inference-service] ownerReferences: [e2e-llm-inference-service] - apiVersion: apps/v1 [e2e-llm-inference-service] kind: ReplicaSet [e2e-llm-inference-service] name: llmisvca690bbc929faec8bc98c767f16c003c1-kserve-router-scheduler-5d6998448f [e2e-llm-inference-service] uid: b08110a1-5c9b-4623-b862-05a01c144c6a [e2e-llm-inference-service] controller: true [e2e-llm-inference-service] blockOwnerDeletion: true [e2e-llm-inference-service] managedFields: [e2e-llm-inference-service] - manager: ip-10-0-133-25 [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] apiVersion: v1 [e2e-llm-inference-service] time: '2026-07-07T14:31:40Z' [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:metadata: [e2e-llm-inference-service] f:annotations: [e2e-llm-inference-service] f:k8s.ovn.org/pod-networks: {} [e2e-llm-inference-service] subresource: status [e2e-llm-inference-service] - manager: kube-controller-manager [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] apiVersion: v1 [e2e-llm-inference-service] time: '2026-07-07T14:31:40Z' [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:metadata: [e2e-llm-inference-service] f:annotations: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:app.kubernetes.io/version: {} [e2e-llm-inference-service] f:certificates.kserve.io/expiration-v2: {} [e2e-llm-inference-service] f:generateName: {} [e2e-llm-inference-service] f:labels: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:app.kubernetes.io/component: {} [e2e-llm-inference-service] f:app.kubernetes.io/name: {} [e2e-llm-inference-service] f:app.kubernetes.io/part-of: {} [e2e-llm-inference-service] f:pod-template-hash: {} [e2e-llm-inference-service] f:ownerReferences: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"uid":"b08110a1-5c9b-4623-b862-05a01c144c6a"}: {} [e2e-llm-inference-service] f:spec: [e2e-llm-inference-service] f:containers: [e2e-llm-inference-service] k:{"name":"main"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:args: {} [e2e-llm-inference-service] f:command: {} [e2e-llm-inference-service] f:env: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"name":"SSL_CERT_DIR"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] f:image: {} [e2e-llm-inference-service] f:imagePullPolicy: {} [e2e-llm-inference-service] f:lifecycle: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:preStop: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:exec: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:command: {} [e2e-llm-inference-service] f:livenessProbe: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:failureThreshold: {} [e2e-llm-inference-service] f:grpc: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:port: {} [e2e-llm-inference-service] f:service: {} [e2e-llm-inference-service] f:initialDelaySeconds: {} [e2e-llm-inference-service] f:periodSeconds: {} [e2e-llm-inference-service] f:successThreshold: {} [e2e-llm-inference-service] f:timeoutSeconds: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:ports: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"containerPort":5557,"protocol":"TCP"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:containerPort: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:protocol: {} [e2e-llm-inference-service] k:{"containerPort":9002,"protocol":"TCP"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:containerPort: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:protocol: {} [e2e-llm-inference-service] k:{"containerPort":9003,"protocol":"TCP"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:containerPort: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:protocol: {} [e2e-llm-inference-service] k:{"containerPort":9090,"protocol":"TCP"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:containerPort: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:protocol: {} [e2e-llm-inference-service] f:readinessProbe: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:failureThreshold: {} [e2e-llm-inference-service] f:grpc: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:port: {} [e2e-llm-inference-service] f:service: {} [e2e-llm-inference-service] f:initialDelaySeconds: {} [e2e-llm-inference-service] f:periodSeconds: {} [e2e-llm-inference-service] f:successThreshold: {} [e2e-llm-inference-service] f:timeoutSeconds: {} [e2e-llm-inference-service] f:resources: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:requests: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:cpu: {} [e2e-llm-inference-service] f:memory: {} [e2e-llm-inference-service] f:securityContext: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:allowPrivilegeEscalation: {} [e2e-llm-inference-service] f:capabilities: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:drop: {} [e2e-llm-inference-service] f:readOnlyRootFilesystem: {} [e2e-llm-inference-service] f:runAsNonRoot: {} [e2e-llm-inference-service] f:seccompProfile: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:type: {} [e2e-llm-inference-service] f:terminationMessagePath: {} [e2e-llm-inference-service] f:terminationMessagePolicy: {} [e2e-llm-inference-service] f:volumeMounts: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"mountPath":"/tmp/tokenizer"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:mountPath: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] k:{"mountPath":"/var/run/kserve/tls"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:mountPath: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:readOnly: {} [e2e-llm-inference-service] f:dnsPolicy: {} [e2e-llm-inference-service] f:enableServiceLinks: {} [e2e-llm-inference-service] f:initContainers: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"name":"storage-initializer"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:args: {} [e2e-llm-inference-service] f:env: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"name":"AWS_ACCESS_KEY_ID"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:valueFrom: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:secretKeyRef: {} [e2e-llm-inference-service] k:{"name":"AWS_CA_BUNDLE"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"AWS_CA_BUNDLE_CONFIGMAP"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"AWS_ENDPOINT_URL"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"AWS_SECRET_ACCESS_KEY"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:valueFrom: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:secretKeyRef: {} [e2e-llm-inference-service] k:{"name":"HF_HUB_ENABLE_HF_TRANSFER"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"HF_XET_HIGH_PERFORMANCE"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"HF_XET_NUM_CONCURRENT_RANGE_GETS"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"S3_ENDPOINT"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"S3_USE_HTTPS"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"S3_VERIFY_SSL"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"STORAGE_ALLOW_PATTERNS"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] f:image: {} [e2e-llm-inference-service] f:imagePullPolicy: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:resources: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:limits: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:cpu: {} [e2e-llm-inference-service] f:memory: {} [e2e-llm-inference-service] f:requests: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:cpu: {} [e2e-llm-inference-service] f:memory: {} [e2e-llm-inference-service] f:terminationMessagePath: {} [e2e-llm-inference-service] f:terminationMessagePolicy: {} [e2e-llm-inference-service] f:volumeMounts: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"mountPath":"/mnt/models"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:mountPath: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:restartPolicy: {} [e2e-llm-inference-service] f:schedulerName: {} [e2e-llm-inference-service] f:securityContext: {} [e2e-llm-inference-service] f:serviceAccount: {} [e2e-llm-inference-service] f:serviceAccountName: {} [e2e-llm-inference-service] f:terminationGracePeriodSeconds: {} [e2e-llm-inference-service] f:volumes: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"name":"kserve-provision-location"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:emptyDir: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] k:{"name":"tls-certs"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:secret: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:defaultMode: {} [e2e-llm-inference-service] f:secretName: {} [e2e-llm-inference-service] k:{"name":"tokenizer-cache"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:emptyDir: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] k:{"name":"tokenizer-tmp"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:emptyDir: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] k:{"name":"tokenizer-uds"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:emptyDir: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] - manager: multus-daemon [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] apiVersion: v1 [e2e-llm-inference-service] time: '2026-07-07T14:31:41Z' [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:metadata: [e2e-llm-inference-service] f:annotations: [e2e-llm-inference-service] f:k8s.v1.cni.cncf.io/network-status: {} [e2e-llm-inference-service] subresource: status [e2e-llm-inference-service] - manager: kubelet [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] apiVersion: v1 [e2e-llm-inference-service] time: '2026-07-07T14:32:14Z' [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:status: [e2e-llm-inference-service] f:conditions: [e2e-llm-inference-service] k:{"type":"ContainersReady"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:lastProbeTime: {} [e2e-llm-inference-service] f:lastTransitionTime: {} [e2e-llm-inference-service] f:observedGeneration: {} [e2e-llm-inference-service] f:status: {} [e2e-llm-inference-service] f:type: {} [e2e-llm-inference-service] k:{"type":"Initialized"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:lastProbeTime: {} [e2e-llm-inference-service] f:lastTransitionTime: {} [e2e-llm-inference-service] f:observedGeneration: {} [e2e-llm-inference-service] f:status: {} [e2e-llm-inference-service] f:type: {} [e2e-llm-inference-service] k:{"type":"PodReadyToStartContainers"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:lastProbeTime: {} [e2e-llm-inference-service] f:lastTransitionTime: {} [e2e-llm-inference-service] f:observedGeneration: {} [e2e-llm-inference-service] f:status: {} [e2e-llm-inference-service] f:type: {} [e2e-llm-inference-service] k:{"type":"PodScheduled"}: [e2e-llm-inference-service] f:observedGeneration: {} [e2e-llm-inference-service] k:{"type":"Ready"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:lastProbeTime: {} [e2e-llm-inference-service] f:lastTransitionTime: {} [e2e-llm-inference-service] f:observedGeneration: {} [e2e-llm-inference-service] f:status: {} [e2e-llm-inference-service] f:type: {} [e2e-llm-inference-service] f:containerStatuses: {} [e2e-llm-inference-service] f:hostIP: {} [e2e-llm-inference-service] f:hostIPs: {} [e2e-llm-inference-service] f:initContainerStatuses: {} [e2e-llm-inference-service] f:observedGeneration: {} [e2e-llm-inference-service] f:phase: {} [e2e-llm-inference-service] f:podIP: {} [e2e-llm-inference-service] f:podIPs: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"ip":"10.132.0.53"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:ip: {} [e2e-llm-inference-service] f:startTime: {} [e2e-llm-inference-service] subresource: status [e2e-llm-inference-service] spec: [e2e-llm-inference-service] volumes: [e2e-llm-inference-service] - name: tls-certs [e2e-llm-inference-service] secret: [e2e-llm-inference-service] secretName: llmisva690bbc929faec8bc98c767f16c003c1-kserve-self-signed-certs [e2e-llm-inference-service] defaultMode: 420 [e2e-llm-inference-service] - name: tokenizer-uds [e2e-llm-inference-service] emptyDir: {} [e2e-llm-inference-service] - name: tokenizer-tmp [e2e-llm-inference-service] emptyDir: {} [e2e-llm-inference-service] - name: tokenizer-cache [e2e-llm-inference-service] emptyDir: {} [e2e-llm-inference-service] - name: kserve-provision-location [e2e-llm-inference-service] emptyDir: {} [e2e-llm-inference-service] - name: kube-api-access-lsb6d [e2e-llm-inference-service] projected: [e2e-llm-inference-service] sources: [e2e-llm-inference-service] - serviceAccountToken: [e2e-llm-inference-service] expirationSeconds: 3607 [e2e-llm-inference-service] path: token [e2e-llm-inference-service] - configMap: [e2e-llm-inference-service] name: kube-root-ca.crt [e2e-llm-inference-service] items: [e2e-llm-inference-service] - key: ca.crt [e2e-llm-inference-service] path: ca.crt [e2e-llm-inference-service] - downwardAPI: [e2e-llm-inference-service] items: [e2e-llm-inference-service] - path: namespace [e2e-llm-inference-service] fieldRef: [e2e-llm-inference-service] apiVersion: v1 [e2e-llm-inference-service] fieldPath: metadata.namespace [e2e-llm-inference-service] - configMap: [e2e-llm-inference-service] name: openshift-service-ca.crt [e2e-llm-inference-service] items: [e2e-llm-inference-service] - key: service-ca.crt [e2e-llm-inference-service] path: service-ca.crt [e2e-llm-inference-service] defaultMode: 420 [e2e-llm-inference-service] initContainers: [e2e-llm-inference-service] - name: storage-initializer [e2e-llm-inference-service] image: quay.io/opendatahub/kserve-storage-initializer@sha256:13ad765c0f323200e1db7b45c99661287e65c5158d6b54b16139829636d20be4 [e2e-llm-inference-service] args: [e2e-llm-inference-service] - hf://facebook/opt-125m [e2e-llm-inference-service] - /mnt/models [e2e-llm-inference-service] env: [e2e-llm-inference-service] - name: AWS_ACCESS_KEY_ID [e2e-llm-inference-service] valueFrom: [e2e-llm-inference-service] secretKeyRef: [e2e-llm-inference-service] name: seaweedfs-s3-creds [e2e-llm-inference-service] key: AWS_ACCESS_KEY_ID [e2e-llm-inference-service] - name: AWS_SECRET_ACCESS_KEY [e2e-llm-inference-service] valueFrom: [e2e-llm-inference-service] secretKeyRef: [e2e-llm-inference-service] name: seaweedfs-s3-creds [e2e-llm-inference-service] key: AWS_SECRET_ACCESS_KEY [e2e-llm-inference-service] - name: S3_USE_HTTPS [e2e-llm-inference-service] value: '0' [e2e-llm-inference-service] - name: S3_ENDPOINT [e2e-llm-inference-service] value: s3-service.kserve:8333 [e2e-llm-inference-service] - name: AWS_ENDPOINT_URL [e2e-llm-inference-service] value: http://s3-service.kserve:8333 [e2e-llm-inference-service] - name: S3_VERIFY_SSL [e2e-llm-inference-service] value: '0' [e2e-llm-inference-service] - name: AWS_CA_BUNDLE [e2e-llm-inference-service] value: /etc/ssl/custom-certs/cabundle.crt [e2e-llm-inference-service] - name: AWS_CA_BUNDLE_CONFIGMAP [e2e-llm-inference-service] value: odh-kserve-custom-ca-bundle [e2e-llm-inference-service] - name: HF_HUB_ENABLE_HF_TRANSFER [e2e-llm-inference-service] value: '1' [e2e-llm-inference-service] - name: HF_XET_HIGH_PERFORMANCE [e2e-llm-inference-service] value: '1' [e2e-llm-inference-service] - name: HF_XET_NUM_CONCURRENT_RANGE_GETS [e2e-llm-inference-service] value: '8' [e2e-llm-inference-service] - name: STORAGE_ALLOW_PATTERNS [e2e-llm-inference-service] value: '["tokenizer.json", "tokenizer_config.json", "special_tokens_map.json", [e2e-llm-inference-service] "vocab.json", "merges.txt", "config.json", "generation_config.json"]' [e2e-llm-inference-service] resources: [e2e-llm-inference-service] limits: [e2e-llm-inference-service] cpu: '1' [e2e-llm-inference-service] memory: 24Gi [e2e-llm-inference-service] requests: [e2e-llm-inference-service] cpu: 100m [e2e-llm-inference-service] memory: 100Mi [e2e-llm-inference-service] volumeMounts: [e2e-llm-inference-service] - name: kserve-provision-location [e2e-llm-inference-service] mountPath: /mnt/models [e2e-llm-inference-service] - name: kube-api-access-lsb6d [e2e-llm-inference-service] readOnly: true [e2e-llm-inference-service] mountPath: /var/run/secrets/kubernetes.io/serviceaccount [e2e-llm-inference-service] terminationMessagePath: /dev/termination-log [e2e-llm-inference-service] terminationMessagePolicy: FallbackToLogsOnError [e2e-llm-inference-service] imagePullPolicy: IfNotPresent [e2e-llm-inference-service] securityContext: [e2e-llm-inference-service] capabilities: [e2e-llm-inference-service] drop: [e2e-llm-inference-service] - ALL [e2e-llm-inference-service] runAsUser: 1000690000 [e2e-llm-inference-service] runAsNonRoot: true [e2e-llm-inference-service] allowPrivilegeEscalation: false [e2e-llm-inference-service] containers: [e2e-llm-inference-service] - name: main [e2e-llm-inference-service] image: ghcr.io/llm-d/llm-d-router-endpoint-picker:v0.9.0-rc.2 [e2e-llm-inference-service] command: [e2e-llm-inference-service] - /app/epp [e2e-llm-inference-service] - --pool-name [e2e-llm-inference-service] - llmisvc-model-fb-opt-125m-route-87882a8e-inference-pool [e2e-llm-inference-service] - --pool-namespace [e2e-llm-inference-service] - kserve-ci-e2e-test [e2e-llm-inference-service] - --zap-encoder [e2e-llm-inference-service] - json [e2e-llm-inference-service] - --grpc-port [e2e-llm-inference-service] - '9002' [e2e-llm-inference-service] - --grpc-health-port [e2e-llm-inference-service] - '9003' [e2e-llm-inference-service] - --enable-cert-reload=true [e2e-llm-inference-service] - --secure-serving=true [e2e-llm-inference-service] - --model-server-metrics-scheme=https [e2e-llm-inference-service] - --cert-path=/var/run/kserve/tls [e2e-llm-inference-service] args: [e2e-llm-inference-service] - --config-text [e2e-llm-inference-service] - "apiVersion: inference.networking.x-k8s.io/v1alpha1\nkind: EndpointPickerConfig\n\ [e2e-llm-inference-service] plugins:\n- type: single-profile-handler\n- type: queue-scorer\n- type: prefix-cache-scorer\n\ [e2e-llm-inference-service] - type: max-score-picker\nschedulingProfiles:\n- name: default\n plugins:\n\ [e2e-llm-inference-service] \ - pluginRef: queue-scorer\n weight: 2\n - pluginRef: prefix-cache-scorer\n\ [e2e-llm-inference-service] \ weight: 3\n - pluginRef: max-score-picker\n" [e2e-llm-inference-service] ports: [e2e-llm-inference-service] - name: grpc [e2e-llm-inference-service] containerPort: 9002 [e2e-llm-inference-service] protocol: TCP [e2e-llm-inference-service] - name: grpc-health [e2e-llm-inference-service] containerPort: 9003 [e2e-llm-inference-service] protocol: TCP [e2e-llm-inference-service] - name: metrics [e2e-llm-inference-service] containerPort: 9090 [e2e-llm-inference-service] protocol: TCP [e2e-llm-inference-service] - name: zmq [e2e-llm-inference-service] containerPort: 5557 [e2e-llm-inference-service] protocol: TCP [e2e-llm-inference-service] env: [e2e-llm-inference-service] - name: SSL_CERT_DIR [e2e-llm-inference-service] value: /var/run/kserve/tls:/var/run/secrets/kubernetes.io/serviceaccount:/etc/pki/tls/certs [e2e-llm-inference-service] resources: [e2e-llm-inference-service] requests: [e2e-llm-inference-service] cpu: 256m [e2e-llm-inference-service] memory: 500Mi [e2e-llm-inference-service] volumeMounts: [e2e-llm-inference-service] - name: tls-certs [e2e-llm-inference-service] readOnly: true [e2e-llm-inference-service] mountPath: /var/run/kserve/tls [e2e-llm-inference-service] - name: tokenizer-uds [e2e-llm-inference-service] mountPath: /tmp/tokenizer [e2e-llm-inference-service] - name: kube-api-access-lsb6d [e2e-llm-inference-service] readOnly: true [e2e-llm-inference-service] mountPath: /var/run/secrets/kubernetes.io/serviceaccount [e2e-llm-inference-service] livenessProbe: [e2e-llm-inference-service] grpc: [e2e-llm-inference-service] port: 9003 [e2e-llm-inference-service] service: liveness [e2e-llm-inference-service] initialDelaySeconds: 5 [e2e-llm-inference-service] timeoutSeconds: 1 [e2e-llm-inference-service] periodSeconds: 10 [e2e-llm-inference-service] successThreshold: 1 [e2e-llm-inference-service] failureThreshold: 3 [e2e-llm-inference-service] readinessProbe: [e2e-llm-inference-service] grpc: [e2e-llm-inference-service] port: 9003 [e2e-llm-inference-service] service: readiness [e2e-llm-inference-service] initialDelaySeconds: 30 [e2e-llm-inference-service] timeoutSeconds: 1 [e2e-llm-inference-service] periodSeconds: 10 [e2e-llm-inference-service] successThreshold: 1 [e2e-llm-inference-service] failureThreshold: 3 [e2e-llm-inference-service] lifecycle: [e2e-llm-inference-service] preStop: [e2e-llm-inference-service] exec: [e2e-llm-inference-service] command: [e2e-llm-inference-service] - /bin/sleep [e2e-llm-inference-service] - '15' [e2e-llm-inference-service] terminationMessagePath: /dev/termination-log [e2e-llm-inference-service] terminationMessagePolicy: FallbackToLogsOnError [e2e-llm-inference-service] imagePullPolicy: IfNotPresent [e2e-llm-inference-service] securityContext: [e2e-llm-inference-service] capabilities: [e2e-llm-inference-service] drop: [e2e-llm-inference-service] - ALL [e2e-llm-inference-service] runAsUser: 1000690000 [e2e-llm-inference-service] runAsNonRoot: true [e2e-llm-inference-service] readOnlyRootFilesystem: true [e2e-llm-inference-service] allowPrivilegeEscalation: false [e2e-llm-inference-service] seccompProfile: [e2e-llm-inference-service] type: RuntimeDefault [e2e-llm-inference-service] restartPolicy: Always [e2e-llm-inference-service] terminationGracePeriodSeconds: 60 [e2e-llm-inference-service] dnsPolicy: ClusterFirst [e2e-llm-inference-service] serviceAccountName: llmisvc-model-fb-opt-125m-route-87882a8e-epp-sa [e2e-llm-inference-service] serviceAccount: llmisvc-model-fb-opt-125m-route-87882a8e-epp-sa [e2e-llm-inference-service] nodeName: ip-10-0-133-25.ec2.internal [e2e-llm-inference-service] securityContext: [e2e-llm-inference-service] seLinuxOptions: [e2e-llm-inference-service] level: s0:c26,c20 [e2e-llm-inference-service] fsGroup: 1000690000 [e2e-llm-inference-service] seccompProfile: [e2e-llm-inference-service] type: RuntimeDefault [e2e-llm-inference-service] imagePullSecrets: [e2e-llm-inference-service] - name: default-dockercfg-tbbgr [e2e-llm-inference-service] - name: llmisvc-model-fb-opt-125m-route-87882a8e-epp-sa-dockercfg-r9q2l [e2e-llm-inference-service] schedulerName: default-scheduler [e2e-llm-inference-service] tolerations: [e2e-llm-inference-service] - key: node.kubernetes.io/not-ready [e2e-llm-inference-service] operator: Exists [e2e-llm-inference-service] effect: NoExecute [e2e-llm-inference-service] tolerationSeconds: 300 [e2e-llm-inference-service] - key: node.kubernetes.io/unreachable [e2e-llm-inference-service] operator: Exists [e2e-llm-inference-service] effect: NoExecute [e2e-llm-inference-service] tolerationSeconds: 300 [e2e-llm-inference-service] - key: node.kubernetes.io/memory-pressure [e2e-llm-inference-service] operator: Exists [e2e-llm-inference-service] effect: NoSchedule [e2e-llm-inference-service] priority: 0 [e2e-llm-inference-service] enableServiceLinks: true [e2e-llm-inference-service] preemptionPolicy: PreemptLowerPriority [e2e-llm-inference-service] status: [e2e-llm-inference-service] observedGeneration: 1 [e2e-llm-inference-service] phase: Running [e2e-llm-inference-service] conditions: [e2e-llm-inference-service] - type: PodReadyToStartContainers [e2e-llm-inference-service] observedGeneration: 1 [e2e-llm-inference-service] status: 'True' [e2e-llm-inference-service] lastProbeTime: null [e2e-llm-inference-service] lastTransitionTime: '2026-07-07T14:31:41Z' [e2e-llm-inference-service] - type: Initialized [e2e-llm-inference-service] observedGeneration: 1 [e2e-llm-inference-service] status: 'True' [e2e-llm-inference-service] lastProbeTime: null [e2e-llm-inference-service] lastTransitionTime: '2026-07-07T14:31:42Z' [e2e-llm-inference-service] - type: Ready [e2e-llm-inference-service] observedGeneration: 1 [e2e-llm-inference-service] status: 'True' [e2e-llm-inference-service] lastProbeTime: null [e2e-llm-inference-service] lastTransitionTime: '2026-07-07T14:32:14Z' [e2e-llm-inference-service] - type: ContainersReady [e2e-llm-inference-service] observedGeneration: 1 [e2e-llm-inference-service] status: 'True' [e2e-llm-inference-service] lastProbeTime: null [e2e-llm-inference-service] lastTransitionTime: '2026-07-07T14:32:14Z' [e2e-llm-inference-service] - type: PodScheduled [e2e-llm-inference-service] observedGeneration: 1 [e2e-llm-inference-service] status: 'True' [e2e-llm-inference-service] lastProbeTime: null [e2e-llm-inference-service] lastTransitionTime: '2026-07-07T14:31:40Z' [e2e-llm-inference-service] hostIP: 10.0.133.25 [e2e-llm-inference-service] hostIPs: [e2e-llm-inference-service] - ip: 10.0.133.25 [e2e-llm-inference-service] podIP: 10.132.0.53 [e2e-llm-inference-service] podIPs: [e2e-llm-inference-service] - ip: 10.132.0.53 [e2e-llm-inference-service] startTime: '2026-07-07T14:31:40Z' [e2e-llm-inference-service] initContainerStatuses: [e2e-llm-inference-service] - name: storage-initializer [e2e-llm-inference-service] state: [e2e-llm-inference-service] terminated: [e2e-llm-inference-service] exitCode: 0 [e2e-llm-inference-service] reason: Completed [e2e-llm-inference-service] startedAt: '2026-07-07T14:31:41Z' [e2e-llm-inference-service] finishedAt: '2026-07-07T14:31:41Z' [e2e-llm-inference-service] containerID: cri-o://0ff49e83832733ee3b55ad70f94acc8b4667a99f94f767f87831407b64593977 [e2e-llm-inference-service] lastState: {} [e2e-llm-inference-service] ready: true [e2e-llm-inference-service] restartCount: 0 [e2e-llm-inference-service] image: quay.io/opendatahub/kserve-storage-initializer@sha256:13ad765c0f323200e1db7b45c99661287e65c5158d6b54b16139829636d20be4 [e2e-llm-inference-service] imageID: quay.io/opendatahub/kserve-storage-initializer@sha256:13ad765c0f323200e1db7b45c99661287e65c5158d6b54b16139829636d20be4 [e2e-llm-inference-service] containerID: cri-o://0ff49e83832733ee3b55ad70f94acc8b4667a99f94f767f87831407b64593977 [e2e-llm-inference-service] started: false [e2e-llm-inference-service] allocatedResources: [e2e-llm-inference-service] cpu: 100m [e2e-llm-inference-service] memory: 100Mi [e2e-llm-inference-service] resources: [e2e-llm-inference-service] limits: [e2e-llm-inference-service] cpu: '1' [e2e-llm-inference-service] memory: 24Gi [e2e-llm-inference-service] requests: [e2e-llm-inference-service] cpu: 100m [e2e-llm-inference-service] memory: 100Mi [e2e-llm-inference-service] volumeMounts: [e2e-llm-inference-service] - name: kserve-provision-location [e2e-llm-inference-service] mountPath: /mnt/models [e2e-llm-inference-service] - name: kube-api-access-lsb6d [e2e-llm-inference-service] mountPath: /var/run/secrets/kubernetes.io/serviceaccount [e2e-llm-inference-service] readOnly: true [e2e-llm-inference-service] recursiveReadOnly: Disabled [e2e-llm-inference-service] user: [e2e-llm-inference-service] linux: [e2e-llm-inference-service] uid: 1000690000 [e2e-llm-inference-service] gid: 0 [e2e-llm-inference-service] supplementalGroups: [e2e-llm-inference-service] - 0 [e2e-llm-inference-service] - 1000690000 [e2e-llm-inference-service] containerStatuses: [e2e-llm-inference-service] - name: main [e2e-llm-inference-service] state: [e2e-llm-inference-service] running: [e2e-llm-inference-service] startedAt: '2026-07-07T14:31:42Z' [e2e-llm-inference-service] lastState: {} [e2e-llm-inference-service] ready: true [e2e-llm-inference-service] restartCount: 0 [e2e-llm-inference-service] image: ghcr.io/llm-d/llm-d-router-endpoint-picker:v0.9.0-rc.2 [e2e-llm-inference-service] imageID: ghcr.io/llm-d/llm-d-router-endpoint-picker@sha256:06b6c75d77afd0e07053402752a9736c2dfbc12a306d0d37d963aac4c1d4e6a6 [e2e-llm-inference-service] containerID: cri-o://12029edf3a006b500eb8405d9ef175adada92634ec7f11e0a399d1aab4a55711 [e2e-llm-inference-service] started: true [e2e-llm-inference-service] allocatedResources: [e2e-llm-inference-service] cpu: 256m [e2e-llm-inference-service] memory: 500Mi [e2e-llm-inference-service] resources: [e2e-llm-inference-service] requests: [e2e-llm-inference-service] cpu: 256m [e2e-llm-inference-service] memory: 500Mi [e2e-llm-inference-service] volumeMounts: [e2e-llm-inference-service] - name: tls-certs [e2e-llm-inference-service] mountPath: /var/run/kserve/tls [e2e-llm-inference-service] readOnly: true [e2e-llm-inference-service] recursiveReadOnly: Disabled [e2e-llm-inference-service] - name: tokenizer-uds [e2e-llm-inference-service] mountPath: /tmp/tokenizer [e2e-llm-inference-service] - name: kube-api-access-lsb6d [e2e-llm-inference-service] mountPath: /var/run/secrets/kubernetes.io/serviceaccount [e2e-llm-inference-service] readOnly: true [e2e-llm-inference-service] recursiveReadOnly: Disabled [e2e-llm-inference-service] user: [e2e-llm-inference-service] linux: [e2e-llm-inference-service] uid: 1000690000 [e2e-llm-inference-service] gid: 0 [e2e-llm-inference-service] supplementalGroups: [e2e-llm-inference-service] - 0 [e2e-llm-inference-service] - 1000690000 [e2e-llm-inference-service] qosClass: Burstable [e2e-llm-inference-service] apiVersion: v1 [e2e-llm-inference-service] kind: Pod [e2e-llm-inference-service] [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1254 --- [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1255 metadata: [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-route-87882a8e-epp-sa [e2e-llm-inference-service] namespace: kserve-ci-e2e-test [e2e-llm-inference-service] uid: 6daef5dc-9a7d-4cc5-ba36-94a7dc430239 [e2e-llm-inference-service] resourceVersion: '30977' [e2e-llm-inference-service] creationTimestamp: '2026-07-07T14:31:40Z' [e2e-llm-inference-service] labels: [e2e-llm-inference-service] app.kubernetes.io/component: llminferenceservice-router-scheduler [e2e-llm-inference-service] app.kubernetes.io/name: llmisvc-model-fb-opt-125m-route-87882a8e [e2e-llm-inference-service] app.kubernetes.io/part-of: llminferenceservice [e2e-llm-inference-service] annotations: [e2e-llm-inference-service] openshift.io/internal-registry-pull-secret-ref: llmisvc-model-fb-opt-125m-route-87882a8e-epp-sa-dockercfg-r9q2l [e2e-llm-inference-service] ownerReferences: [e2e-llm-inference-service] - apiVersion: serving.kserve.io/v1alpha2 [e2e-llm-inference-service] kind: LLMInferenceService [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-route-87882a8e [e2e-llm-inference-service] uid: 88c60d0e-95bd-401f-ae1c-2c4c0b1e7c9d [e2e-llm-inference-service] controller: true [e2e-llm-inference-service] blockOwnerDeletion: true [e2e-llm-inference-service] managedFields: [e2e-llm-inference-service] - manager: openshift.io/image-registry-pull-secrets_service-account-controller [e2e-llm-inference-service] operation: Apply [e2e-llm-inference-service] apiVersion: v1 [e2e-llm-inference-service] time: '2026-07-07T14:31:40Z' [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:imagePullSecrets: {} [e2e-llm-inference-service] f:metadata: [e2e-llm-inference-service] f:annotations: [e2e-llm-inference-service] f:openshift.io/internal-registry-pull-secret-ref: {} [e2e-llm-inference-service] f:secrets: [e2e-llm-inference-service] k:{"name":"llmisvc-model-fb-opt-125m-route-87882a8e-epp-sa-dockercfg-r9q2l"}: {} [e2e-llm-inference-service] - manager: manager [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] apiVersion: v1 [e2e-llm-inference-service] time: '2026-07-07T14:31:40Z' [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:metadata: [e2e-llm-inference-service] f:labels: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:app.kubernetes.io/component: {} [e2e-llm-inference-service] f:app.kubernetes.io/name: {} [e2e-llm-inference-service] f:app.kubernetes.io/part-of: {} [e2e-llm-inference-service] f:ownerReferences: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"uid":"88c60d0e-95bd-401f-ae1c-2c4c0b1e7c9d"}: {} [e2e-llm-inference-service] f:secrets: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"name":"default-dockercfg-tbbgr"}: {} [e2e-llm-inference-service] k:{"name":"seaweedfs-s3-creds"}: {} [e2e-llm-inference-service] secrets: [e2e-llm-inference-service] - name: default-dockercfg-tbbgr [e2e-llm-inference-service] - name: seaweedfs-s3-creds [e2e-llm-inference-service] - name: llmisvc-model-fb-opt-125m-route-87882a8e-epp-sa-dockercfg-r9q2l [e2e-llm-inference-service] imagePullSecrets: [e2e-llm-inference-service] - name: default-dockercfg-tbbgr [e2e-llm-inference-service] - name: llmisvc-model-fb-opt-125m-route-87882a8e-epp-sa-dockercfg-r9q2l [e2e-llm-inference-service] apiVersion: v1 [e2e-llm-inference-service] kind: ServiceAccount [e2e-llm-inference-service] [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1254 --- [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1255 metadata: [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-route-87882a8e-epp-service [e2e-llm-inference-service] namespace: kserve-ci-e2e-test [e2e-llm-inference-service] uid: c8576b5b-94c8-4bab-b493-e867f77d1f8d [e2e-llm-inference-service] resourceVersion: '30994' [e2e-llm-inference-service] creationTimestamp: '2026-07-07T14:31:40Z' [e2e-llm-inference-service] labels: [e2e-llm-inference-service] app.kubernetes.io/component: llminferenceservice-router-scheduler [e2e-llm-inference-service] app.kubernetes.io/name: llmisvc-model-fb-opt-125m-route-87882a8e [e2e-llm-inference-service] app.kubernetes.io/part-of: llminferenceservice [e2e-llm-inference-service] ownerReferences: [e2e-llm-inference-service] - apiVersion: serving.kserve.io/v1alpha2 [e2e-llm-inference-service] kind: LLMInferenceService [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-route-87882a8e [e2e-llm-inference-service] uid: 88c60d0e-95bd-401f-ae1c-2c4c0b1e7c9d [e2e-llm-inference-service] controller: true [e2e-llm-inference-service] blockOwnerDeletion: true [e2e-llm-inference-service] managedFields: [e2e-llm-inference-service] - manager: manager [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] apiVersion: v1 [e2e-llm-inference-service] time: '2026-07-07T14:31:40Z' [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:metadata: [e2e-llm-inference-service] f:labels: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:app.kubernetes.io/component: {} [e2e-llm-inference-service] f:app.kubernetes.io/name: {} [e2e-llm-inference-service] f:app.kubernetes.io/part-of: {} [e2e-llm-inference-service] f:ownerReferences: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"uid":"88c60d0e-95bd-401f-ae1c-2c4c0b1e7c9d"}: {} [e2e-llm-inference-service] f:spec: [e2e-llm-inference-service] f:internalTrafficPolicy: {} [e2e-llm-inference-service] f:ports: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"port":5557,"protocol":"TCP"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:port: {} [e2e-llm-inference-service] f:protocol: {} [e2e-llm-inference-service] f:targetPort: {} [e2e-llm-inference-service] k:{"port":9002,"protocol":"TCP"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:port: {} [e2e-llm-inference-service] f:protocol: {} [e2e-llm-inference-service] f:targetPort: {} [e2e-llm-inference-service] k:{"port":9003,"protocol":"TCP"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:port: {} [e2e-llm-inference-service] f:protocol: {} [e2e-llm-inference-service] f:targetPort: {} [e2e-llm-inference-service] k:{"port":9090,"protocol":"TCP"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:port: {} [e2e-llm-inference-service] f:protocol: {} [e2e-llm-inference-service] f:targetPort: {} [e2e-llm-inference-service] f:selector: {} [e2e-llm-inference-service] f:sessionAffinity: {} [e2e-llm-inference-service] f:type: {} [e2e-llm-inference-service] spec: [e2e-llm-inference-service] ports: [e2e-llm-inference-service] - name: grpc [e2e-llm-inference-service] protocol: TCP [e2e-llm-inference-service] port: 9002 [e2e-llm-inference-service] targetPort: grpc [e2e-llm-inference-service] - name: grpc-health [e2e-llm-inference-service] protocol: TCP [e2e-llm-inference-service] port: 9003 [e2e-llm-inference-service] targetPort: grpc-health [e2e-llm-inference-service] - name: metrics [e2e-llm-inference-service] protocol: TCP [e2e-llm-inference-service] port: 9090 [e2e-llm-inference-service] targetPort: metrics [e2e-llm-inference-service] - name: zmq [e2e-llm-inference-service] protocol: TCP [e2e-llm-inference-service] port: 5557 [e2e-llm-inference-service] targetPort: zmq [e2e-llm-inference-service] selector: [e2e-llm-inference-service] app.kubernetes.io/component: llminferenceservice-router-scheduler [e2e-llm-inference-service] app.kubernetes.io/name: llmisvc-model-fb-opt-125m-route-87882a8e [e2e-llm-inference-service] app.kubernetes.io/part-of: llminferenceservice [e2e-llm-inference-service] clusterIP: 172.31.105.164 [e2e-llm-inference-service] clusterIPs: [e2e-llm-inference-service] - 172.31.105.164 [e2e-llm-inference-service] type: ClusterIP [e2e-llm-inference-service] sessionAffinity: None [e2e-llm-inference-service] ipFamilies: [e2e-llm-inference-service] - IPv4 [e2e-llm-inference-service] ipFamilyPolicy: SingleStack [e2e-llm-inference-service] internalTrafficPolicy: Cluster [e2e-llm-inference-service] status: [e2e-llm-inference-service] loadBalancer: {} [e2e-llm-inference-service] apiVersion: v1 [e2e-llm-inference-service] kind: Service [e2e-llm-inference-service] [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1254 --- [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1255 metadata: [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-route-87882a8e-kserve-workload-svc [e2e-llm-inference-service] namespace: kserve-ci-e2e-test [e2e-llm-inference-service] uid: f49fc6fa-967a-4564-b13b-50c3a33b3c52 [e2e-llm-inference-service] resourceVersion: '30957' [e2e-llm-inference-service] creationTimestamp: '2026-07-07T14:31:40Z' [e2e-llm-inference-service] labels: [e2e-llm-inference-service] app.kubernetes.io/component: llminferenceservice-workload [e2e-llm-inference-service] app.kubernetes.io/name: llmisvc-model-fb-opt-125m-route-87882a8e [e2e-llm-inference-service] app.kubernetes.io/part-of: llminferenceservice [e2e-llm-inference-service] ownerReferences: [e2e-llm-inference-service] - apiVersion: serving.kserve.io/v1alpha2 [e2e-llm-inference-service] kind: LLMInferenceService [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-route-87882a8e [e2e-llm-inference-service] uid: 88c60d0e-95bd-401f-ae1c-2c4c0b1e7c9d [e2e-llm-inference-service] controller: true [e2e-llm-inference-service] blockOwnerDeletion: true [e2e-llm-inference-service] managedFields: [e2e-llm-inference-service] - manager: manager [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] apiVersion: v1 [e2e-llm-inference-service] time: '2026-07-07T14:31:40Z' [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:metadata: [e2e-llm-inference-service] f:labels: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:app.kubernetes.io/component: {} [e2e-llm-inference-service] f:app.kubernetes.io/name: {} [e2e-llm-inference-service] f:app.kubernetes.io/part-of: {} [e2e-llm-inference-service] f:ownerReferences: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"uid":"88c60d0e-95bd-401f-ae1c-2c4c0b1e7c9d"}: {} [e2e-llm-inference-service] f:spec: [e2e-llm-inference-service] f:internalTrafficPolicy: {} [e2e-llm-inference-service] f:ports: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"port":8000,"protocol":"TCP"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:appProtocol: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:port: {} [e2e-llm-inference-service] f:protocol: {} [e2e-llm-inference-service] f:targetPort: {} [e2e-llm-inference-service] f:selector: {} [e2e-llm-inference-service] f:sessionAffinity: {} [e2e-llm-inference-service] f:type: {} [e2e-llm-inference-service] spec: [e2e-llm-inference-service] ports: [e2e-llm-inference-service] - name: https [e2e-llm-inference-service] protocol: TCP [e2e-llm-inference-service] appProtocol: https [e2e-llm-inference-service] port: 8000 [e2e-llm-inference-service] targetPort: 8000 [e2e-llm-inference-service] selector: [e2e-llm-inference-service] app.kubernetes.io/name: llmisvc-model-fb-opt-125m-route-87882a8e [e2e-llm-inference-service] app.kubernetes.io/part-of: llminferenceservice [e2e-llm-inference-service] kserve.io/component: workload [e2e-llm-inference-service] clusterIP: 172.31.184.206 [e2e-llm-inference-service] clusterIPs: [e2e-llm-inference-service] - 172.31.184.206 [e2e-llm-inference-service] type: ClusterIP [e2e-llm-inference-service] sessionAffinity: None [e2e-llm-inference-service] ipFamilies: [e2e-llm-inference-service] - IPv4 [e2e-llm-inference-service] ipFamilyPolicy: SingleStack [e2e-llm-inference-service] internalTrafficPolicy: Cluster [e2e-llm-inference-service] status: [e2e-llm-inference-service] loadBalancer: {} [e2e-llm-inference-service] apiVersion: v1 [e2e-llm-inference-service] kind: Service [e2e-llm-inference-service] [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1254 --- [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1255 metadata: [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-route-87882a8e-kserve [e2e-llm-inference-service] namespace: kserve-ci-e2e-test [e2e-llm-inference-service] uid: 6006b5ee-4a45-4a5d-9465-bae07199d5c5 [e2e-llm-inference-service] resourceVersion: '33111' [e2e-llm-inference-service] generation: 1 [e2e-llm-inference-service] creationTimestamp: '2026-07-07T14:31:40Z' [e2e-llm-inference-service] labels: [e2e-llm-inference-service] app.kubernetes.io/component: llminferenceservice-workload [e2e-llm-inference-service] app.kubernetes.io/name: llmisvc-model-fb-opt-125m-route-87882a8e [e2e-llm-inference-service] app.kubernetes.io/part-of: llminferenceservice [e2e-llm-inference-service] kserve.io/component: workload [e2e-llm-inference-service] llm-d.ai/role: both [e2e-llm-inference-service] annotations: [e2e-llm-inference-service] deployment.kubernetes.io/revision: '1' [e2e-llm-inference-service] ownerReferences: [e2e-llm-inference-service] - apiVersion: serving.kserve.io/v1alpha2 [e2e-llm-inference-service] kind: LLMInferenceService [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-route-87882a8e [e2e-llm-inference-service] uid: 88c60d0e-95bd-401f-ae1c-2c4c0b1e7c9d [e2e-llm-inference-service] controller: true [e2e-llm-inference-service] blockOwnerDeletion: true [e2e-llm-inference-service] managedFields: [e2e-llm-inference-service] - manager: manager [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] apiVersion: apps/v1 [e2e-llm-inference-service] time: '2026-07-07T14:31:40Z' [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:metadata: [e2e-llm-inference-service] f:labels: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:app.kubernetes.io/component: {} [e2e-llm-inference-service] f:app.kubernetes.io/name: {} [e2e-llm-inference-service] f:app.kubernetes.io/part-of: {} [e2e-llm-inference-service] f:kserve.io/component: {} [e2e-llm-inference-service] f:llm-d.ai/role: {} [e2e-llm-inference-service] f:ownerReferences: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"uid":"88c60d0e-95bd-401f-ae1c-2c4c0b1e7c9d"}: {} [e2e-llm-inference-service] f:spec: [e2e-llm-inference-service] f:progressDeadlineSeconds: {} [e2e-llm-inference-service] f:replicas: {} [e2e-llm-inference-service] f:revisionHistoryLimit: {} [e2e-llm-inference-service] f:selector: {} [e2e-llm-inference-service] f:strategy: [e2e-llm-inference-service] f:rollingUpdate: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:maxSurge: {} [e2e-llm-inference-service] f:maxUnavailable: {} [e2e-llm-inference-service] f:type: {} [e2e-llm-inference-service] f:template: [e2e-llm-inference-service] f:metadata: [e2e-llm-inference-service] f:labels: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:app.kubernetes.io/component: {} [e2e-llm-inference-service] f:app.kubernetes.io/name: {} [e2e-llm-inference-service] f:app.kubernetes.io/part-of: {} [e2e-llm-inference-service] f:kserve.io/component: {} [e2e-llm-inference-service] f:llm-d.ai/role: {} [e2e-llm-inference-service] f:spec: [e2e-llm-inference-service] f:containers: [e2e-llm-inference-service] k:{"name":"main"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:command: {} [e2e-llm-inference-service] f:env: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"name":"HF_HUB_CACHE"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"HOME"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"TORCHINDUCTOR_CACHE_DIR"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"USER"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"VLLM_CPU_KVCACHE_SPACE"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"VLLM_ENABLE_V1_MULTIPROCESSING"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"VLLM_LOGGING_LEVEL"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] f:image: {} [e2e-llm-inference-service] f:imagePullPolicy: {} [e2e-llm-inference-service] f:lifecycle: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:preStop: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:exec: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:command: {} [e2e-llm-inference-service] f:livenessProbe: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:failureThreshold: {} [e2e-llm-inference-service] f:httpGet: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:path: {} [e2e-llm-inference-service] f:port: {} [e2e-llm-inference-service] f:scheme: {} [e2e-llm-inference-service] f:periodSeconds: {} [e2e-llm-inference-service] f:successThreshold: {} [e2e-llm-inference-service] f:timeoutSeconds: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:ports: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"containerPort":8000,"protocol":"TCP"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:containerPort: {} [e2e-llm-inference-service] f:protocol: {} [e2e-llm-inference-service] f:readinessProbe: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:failureThreshold: {} [e2e-llm-inference-service] f:httpGet: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:path: {} [e2e-llm-inference-service] f:port: {} [e2e-llm-inference-service] f:scheme: {} [e2e-llm-inference-service] f:periodSeconds: {} [e2e-llm-inference-service] f:successThreshold: {} [e2e-llm-inference-service] f:timeoutSeconds: {} [e2e-llm-inference-service] f:resources: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:limits: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:cpu: {} [e2e-llm-inference-service] f:memory: {} [e2e-llm-inference-service] f:requests: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:cpu: {} [e2e-llm-inference-service] f:memory: {} [e2e-llm-inference-service] f:securityContext: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:allowPrivilegeEscalation: {} [e2e-llm-inference-service] f:capabilities: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:drop: {} [e2e-llm-inference-service] f:readOnlyRootFilesystem: {} [e2e-llm-inference-service] f:runAsNonRoot: {} [e2e-llm-inference-service] f:seccompProfile: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:type: {} [e2e-llm-inference-service] f:startupProbe: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:failureThreshold: {} [e2e-llm-inference-service] f:httpGet: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:path: {} [e2e-llm-inference-service] f:port: {} [e2e-llm-inference-service] f:scheme: {} [e2e-llm-inference-service] f:periodSeconds: {} [e2e-llm-inference-service] f:successThreshold: {} [e2e-llm-inference-service] f:timeoutSeconds: {} [e2e-llm-inference-service] f:terminationMessagePath: {} [e2e-llm-inference-service] f:terminationMessagePolicy: {} [e2e-llm-inference-service] f:volumeMounts: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"mountPath":"/dev/shm"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:mountPath: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] k:{"mountPath":"/home"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:mountPath: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] k:{"mountPath":"/mnt/models"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:mountPath: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:readOnly: {} [e2e-llm-inference-service] k:{"mountPath":"/models"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:mountPath: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] k:{"mountPath":"/tmp"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:mountPath: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] k:{"mountPath":"/var/run/kserve/tls"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:mountPath: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:readOnly: {} [e2e-llm-inference-service] f:dnsPolicy: {} [e2e-llm-inference-service] f:initContainers: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"name":"storage-initializer"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:args: {} [e2e-llm-inference-service] f:env: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"name":"AWS_ACCESS_KEY_ID"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:valueFrom: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:secretKeyRef: {} [e2e-llm-inference-service] k:{"name":"AWS_CA_BUNDLE"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"AWS_CA_BUNDLE_CONFIGMAP"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"AWS_ENDPOINT_URL"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"AWS_SECRET_ACCESS_KEY"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:valueFrom: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:secretKeyRef: {} [e2e-llm-inference-service] k:{"name":"HF_HUB_ENABLE_HF_TRANSFER"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"HF_XET_HIGH_PERFORMANCE"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"HF_XET_NUM_CONCURRENT_RANGE_GETS"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"S3_ENDPOINT"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"S3_USE_HTTPS"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"S3_VERIFY_SSL"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] f:image: {} [e2e-llm-inference-service] f:imagePullPolicy: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:resources: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:limits: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:cpu: {} [e2e-llm-inference-service] f:memory: {} [e2e-llm-inference-service] f:requests: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:cpu: {} [e2e-llm-inference-service] f:memory: {} [e2e-llm-inference-service] f:terminationMessagePath: {} [e2e-llm-inference-service] f:terminationMessagePolicy: {} [e2e-llm-inference-service] f:volumeMounts: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"mountPath":"/mnt/models"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:mountPath: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:restartPolicy: {} [e2e-llm-inference-service] f:schedulerName: {} [e2e-llm-inference-service] f:securityContext: {} [e2e-llm-inference-service] f:terminationGracePeriodSeconds: {} [e2e-llm-inference-service] f:volumes: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"name":"dshm"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:emptyDir: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:medium: {} [e2e-llm-inference-service] f:sizeLimit: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] k:{"name":"home"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:emptyDir: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] k:{"name":"kserve-provision-location"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:emptyDir: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] k:{"name":"model-cache"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:emptyDir: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] k:{"name":"tls-certs"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:secret: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:defaultMode: {} [e2e-llm-inference-service] f:secretName: {} [e2e-llm-inference-service] k:{"name":"tmp-dir"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:emptyDir: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] - manager: kube-controller-manager [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] apiVersion: apps/v1 [e2e-llm-inference-service] time: '2026-07-07T14:33:40Z' [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:metadata: [e2e-llm-inference-service] f:annotations: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:deployment.kubernetes.io/revision: {} [e2e-llm-inference-service] f:status: [e2e-llm-inference-service] f:availableReplicas: {} [e2e-llm-inference-service] f:conditions: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"type":"Available"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:lastTransitionTime: {} [e2e-llm-inference-service] f:lastUpdateTime: {} [e2e-llm-inference-service] f:message: {} [e2e-llm-inference-service] f:reason: {} [e2e-llm-inference-service] f:status: {} [e2e-llm-inference-service] f:type: {} [e2e-llm-inference-service] k:{"type":"Progressing"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:lastTransitionTime: {} [e2e-llm-inference-service] f:lastUpdateTime: {} [e2e-llm-inference-service] f:message: {} [e2e-llm-inference-service] f:reason: {} [e2e-llm-inference-service] f:status: {} [e2e-llm-inference-service] f:type: {} [e2e-llm-inference-service] f:observedGeneration: {} [e2e-llm-inference-service] f:readyReplicas: {} [e2e-llm-inference-service] f:replicas: {} [e2e-llm-inference-service] f:updatedReplicas: {} [e2e-llm-inference-service] subresource: status [e2e-llm-inference-service] spec: [e2e-llm-inference-service] replicas: 1 [e2e-llm-inference-service] selector: [e2e-llm-inference-service] matchLabels: [e2e-llm-inference-service] app.kubernetes.io/component: llminferenceservice-workload [e2e-llm-inference-service] app.kubernetes.io/name: llmisvc-model-fb-opt-125m-route-87882a8e [e2e-llm-inference-service] app.kubernetes.io/part-of: llminferenceservice [e2e-llm-inference-service] kserve.io/component: workload [e2e-llm-inference-service] llm-d.ai/role: both [e2e-llm-inference-service] template: [e2e-llm-inference-service] metadata: [e2e-llm-inference-service] labels: [e2e-llm-inference-service] app.kubernetes.io/component: llminferenceservice-workload [e2e-llm-inference-service] app.kubernetes.io/name: llmisvc-model-fb-opt-125m-route-87882a8e [e2e-llm-inference-service] app.kubernetes.io/part-of: llminferenceservice [e2e-llm-inference-service] kserve.io/component: workload [e2e-llm-inference-service] llm-d.ai/role: both [e2e-llm-inference-service] spec: [e2e-llm-inference-service] volumes: [e2e-llm-inference-service] - name: home [e2e-llm-inference-service] emptyDir: {} [e2e-llm-inference-service] - name: dshm [e2e-llm-inference-service] emptyDir: [e2e-llm-inference-service] medium: Memory [e2e-llm-inference-service] sizeLimit: 1Gi [e2e-llm-inference-service] - name: model-cache [e2e-llm-inference-service] emptyDir: {} [e2e-llm-inference-service] - name: tmp-dir [e2e-llm-inference-service] emptyDir: {} [e2e-llm-inference-service] - name: tls-certs [e2e-llm-inference-service] secret: [e2e-llm-inference-service] secretName: llmisva690bbc929faec8bc98c767f16c003c1-kserve-self-signed-certs [e2e-llm-inference-service] defaultMode: 420 [e2e-llm-inference-service] - name: kserve-provision-location [e2e-llm-inference-service] emptyDir: {} [e2e-llm-inference-service] initContainers: [e2e-llm-inference-service] - name: storage-initializer [e2e-llm-inference-service] image: quay.io/opendatahub/kserve-storage-initializer@sha256:13ad765c0f323200e1db7b45c99661287e65c5158d6b54b16139829636d20be4 [e2e-llm-inference-service] args: [e2e-llm-inference-service] - hf://facebook/opt-125m [e2e-llm-inference-service] - /mnt/models [e2e-llm-inference-service] env: [e2e-llm-inference-service] - name: AWS_ACCESS_KEY_ID [e2e-llm-inference-service] valueFrom: [e2e-llm-inference-service] secretKeyRef: [e2e-llm-inference-service] name: seaweedfs-s3-creds [e2e-llm-inference-service] key: AWS_ACCESS_KEY_ID [e2e-llm-inference-service] - name: AWS_SECRET_ACCESS_KEY [e2e-llm-inference-service] valueFrom: [e2e-llm-inference-service] secretKeyRef: [e2e-llm-inference-service] name: seaweedfs-s3-creds [e2e-llm-inference-service] key: AWS_SECRET_ACCESS_KEY [e2e-llm-inference-service] - name: S3_USE_HTTPS [e2e-llm-inference-service] value: '0' [e2e-llm-inference-service] - name: S3_ENDPOINT [e2e-llm-inference-service] value: s3-service.kserve:8333 [e2e-llm-inference-service] - name: AWS_ENDPOINT_URL [e2e-llm-inference-service] value: http://s3-service.kserve:8333 [e2e-llm-inference-service] - name: S3_VERIFY_SSL [e2e-llm-inference-service] value: '0' [e2e-llm-inference-service] - name: AWS_CA_BUNDLE [e2e-llm-inference-service] value: /etc/ssl/custom-certs/cabundle.crt [e2e-llm-inference-service] - name: AWS_CA_BUNDLE_CONFIGMAP [e2e-llm-inference-service] value: odh-kserve-custom-ca-bundle [e2e-llm-inference-service] - name: HF_HUB_ENABLE_HF_TRANSFER [e2e-llm-inference-service] value: '1' [e2e-llm-inference-service] - name: HF_XET_HIGH_PERFORMANCE [e2e-llm-inference-service] value: '1' [e2e-llm-inference-service] - name: HF_XET_NUM_CONCURRENT_RANGE_GETS [e2e-llm-inference-service] value: '8' [e2e-llm-inference-service] resources: [e2e-llm-inference-service] limits: [e2e-llm-inference-service] cpu: '1' [e2e-llm-inference-service] memory: 24Gi [e2e-llm-inference-service] requests: [e2e-llm-inference-service] cpu: 100m [e2e-llm-inference-service] memory: 100Mi [e2e-llm-inference-service] volumeMounts: [e2e-llm-inference-service] - name: kserve-provision-location [e2e-llm-inference-service] mountPath: /mnt/models [e2e-llm-inference-service] terminationMessagePath: /dev/termination-log [e2e-llm-inference-service] terminationMessagePolicy: FallbackToLogsOnError [e2e-llm-inference-service] imagePullPolicy: IfNotPresent [e2e-llm-inference-service] containers: [e2e-llm-inference-service] - name: main [e2e-llm-inference-service] image: public.ecr.aws/q9t5s3a7/vllm-cpu-release-repo:v0.19.0 [e2e-llm-inference-service] command: [e2e-llm-inference-service] - /bin/bash [e2e-llm-inference-service] - -c [e2e-llm-inference-service] - "if [ -f /etc/profile.d/ibm-aiu-setup.sh ]; then\n source /etc/profile.d/ibm-aiu-setup.sh\n\ [e2e-llm-inference-service] fi\n\nif [ \"$KSERVE_INFER_ROCE\" = \"true\" ]; then\n echo \"Trying to\ [e2e-llm-inference-service] \ infer RoCE configs ... \"\n grep -H . /sys/class/infiniband/*/ports/*/gids/*\ [e2e-llm-inference-service] \ 2>/dev/null\n grep -H . /sys/class/infiniband/*/ports/*/gid_attrs/types/*\ [e2e-llm-inference-service] \ 2>/dev/null\n\n cat /proc/driver/nvidia/params\n\n KSERVE_INFER_IB_GID_INDEX_GREP=${KSERVE_INFER_IB_GID_INDEX_GREP:-\"\ [e2e-llm-inference-service] RoCE v2\"}\n\n echo \"[Infer RoCE] Discovering active HCAs ...\"\n active_hcas=()\n\ [e2e-llm-inference-service] \ # Loop through all mlx5 devices found in sysfs\n for hca_dir in /sys/class/infiniband/mlx5_*;\ [e2e-llm-inference-service] \ do\n # Ensure it's a directory before proceeding\n if [ -d \"\ [e2e-llm-inference-service] $hca_dir\" ]; then\n hca_name=$(basename \"$hca_dir\")\n \ [e2e-llm-inference-service] \ port_state_file=\"$hca_dir/ports/1/state\" # Assume port 1\n \ [e2e-llm-inference-service] \ type_file=\"$hca_dir/ports/1/gid_attrs/types/*\"\n\n echo\ [e2e-llm-inference-service] \ \"[Infer RoCE] Check if the port state file ${port_state_file} exists\ [e2e-llm-inference-service] \ and contains 'ACTIVE'\"\n if [ -f \"$port_state_file\" ] && grep\ [e2e-llm-inference-service] \ -q \"ACTIVE\" \"$port_state_file\" && grep -q \"${KSERVE_INFER_IB_GID_INDEX_GREP}\"\ [e2e-llm-inference-service] \ ${type_file} 2>/dev/null; then\n echo \"[Infer RoCE] Found\ [e2e-llm-inference-service] \ active HCA: $hca_name\"\n active_hcas+=(\"$hca_name\")\n\ [e2e-llm-inference-service] \ else\n echo \"[Infer RoCE] Skipping inactive or\ [e2e-llm-inference-service] \ down HCA: $hca_name\"\n fi\n fi\n done\n\n # Check if\ [e2e-llm-inference-service] \ we found any active HCAs\n if [ ${#active_hcas[@]} -gt 0 ]; then\n \ [e2e-llm-inference-service] \ # Join the array elements with a comma\n hca_port_pairs=()\n \ [e2e-llm-inference-service] \ for hca in \"${active_hcas[@]}\"; do\n hca_port_pairs+=(\"\ [e2e-llm-inference-service] ${hca}:1\")\n done\n\n active_hca_list=$(IFS=,; echo \"${active_hcas[*]}\"\ [e2e-llm-inference-service] )\n hca_port_pairs_list=$(IFS=,; echo \"${hca_port_pairs[*]}\")\n \ [e2e-llm-inference-service] \ echo \"[Infer RoCE] Setting active HCAs: ${active_hca_list}\"\n \ [e2e-llm-inference-service] \ export NCCL_IB_HCA=${NCCL_IB_HCA:-${active_hca_list}}\n export\ [e2e-llm-inference-service] \ NVSHMEM_HCA_LIST=${NVSHMEM_HCA_LIST:-${hca_port_pairs_list}}\n export\ [e2e-llm-inference-service] \ UCX_NET_DEVICES=${UCX_NET_DEVICES:-${hca_port_pairs_list}}\n\n echo\ [e2e-llm-inference-service] \ \"[Infer RoCE] NCCL_IB_HCA=${NCCL_IB_HCA}\"\n echo \"[Infer RoCE]\ [e2e-llm-inference-service] \ NVSHMEM_HCA_LIST=${NVSHMEM_HCA_LIST}\"\n echo \"[Infer RoCE] UCX_NET_DEVICES=${UCX_NET_DEVICES}\"\ [e2e-llm-inference-service] \n else\n echo \"[Infer RoCE] WARNING: No active RoCE HCAs found.\ [e2e-llm-inference-service] \ NCCL_IB_HCA will not be set.\"\n fi\n\n if [ ${#active_hcas[@]} -gt\ [e2e-llm-inference-service] \ 0 ]; then\n echo \"[Infer RoCE] Finding GID_INDEX for each active\ [e2e-llm-inference-service] \ HCA (SR-IOV compatible)...\"\n\n # For SR-IOV environments, find\ [e2e-llm-inference-service] \ the most common IPv4 RoCE v2 GID index across all HCAs\n declare\ [e2e-llm-inference-service] \ -A gid_index_count\n declare -A hca_gid_index\n\n for hca_name\ [e2e-llm-inference-service] \ in \"${active_hcas[@]}\"; do\n echo \"[Infer RoCE] Processing\ [e2e-llm-inference-service] \ HCA: ${hca_name}\"\n\n # Find all RoCE v2 IPv4 GIDs for this\ [e2e-llm-inference-service] \ HCA and count by index\n for tpath in /sys/class/infiniband/${hca_name}/ports/1/gid_attrs/types/*;\ [e2e-llm-inference-service] \ do\n if grep -q \"${KSERVE_INFER_IB_GID_INDEX_GREP}\" \"\ [e2e-llm-inference-service] $tpath\" 2>/dev/null; then\n idx=$(basename \"$tpath\"\ [e2e-llm-inference-service] )\n gid_file=\"/sys/class/infiniband/${hca_name}/ports/1/gids/${idx}\"\ [e2e-llm-inference-service] \n # Check for IPv4 GID (contains ffff:)\n \ [e2e-llm-inference-service] \ if [ -f \"$gid_file\" ] && grep -q \"ffff:\" \"$gid_file\"; then\n\ [e2e-llm-inference-service] \ gid_value=$(cat \"$gid_file\" 2>/dev/null || echo\ [e2e-llm-inference-service] \ \"\")\n echo \"[Infer RoCE] Found IPv4 RoCE v2 GID\ [e2e-llm-inference-service] \ for ${hca_name}: index=${idx}, gid=${gid_value}\"\n \ [e2e-llm-inference-service] \ hca_gid_index[\"${hca_name}\"]=\"${idx}\"\n gid_index_count[\"\ [e2e-llm-inference-service] ${idx}\"]=$((${gid_index_count[\"${idx}\"]} + 1))\n \ [e2e-llm-inference-service] \ break # Use first found IPv4 GID per HCA\n fi\n \ [e2e-llm-inference-service] \ fi\n done\n done\n\n # Find the most common\ [e2e-llm-inference-service] \ GID index (most likely to be consistent across nodes)\n best_gid_index=\"\ [e2e-llm-inference-service] \"\n max_count=0\n for idx in \"${!gid_index_count[@]}\"; do\n\ [e2e-llm-inference-service] \ count=${gid_index_count[\"${idx}\"]}\n echo \"[Infer\ [e2e-llm-inference-service] \ RoCE] GID_INDEX ${idx} found on ${count} HCAs\"\n if [ $count\ [e2e-llm-inference-service] \ -gt $max_count ]; then\n max_count=$count\n \ [e2e-llm-inference-service] \ best_gid_index=\"$idx\"\n fi\n done\n\n # Use deterministic\ [e2e-llm-inference-service] \ fallback if tied - prefer index 3 (SR-IOV standard)\n if [ ${#gid_index_count[@]}\ [e2e-llm-inference-service] \ -gt 1 ]; then\n echo \"[Infer RoCE] Multiple GID indices found,\ [e2e-llm-inference-service] \ selecting most common: ${best_gid_index}\"\n # If there's a tie,\ [e2e-llm-inference-service] \ prefer index 3 as it's most common in SR-IOV setups\n if [ -n\ [e2e-llm-inference-service] \ \"${gid_index_count['3']}\" ] && [ \"${gid_index_count['3']}\" -eq \"\ [e2e-llm-inference-service] $max_count\" ]; then\n best_gid_index=\"3\"\n \ [e2e-llm-inference-service] \ echo \"[Infer RoCE] Using deterministic fallback: GID_INDEX=3 (SR-IOV\ [e2e-llm-inference-service] \ standard)\"\n fi\n fi\n\n # Check if GID_INDEX is already\ [e2e-llm-inference-service] \ set via environment variables\n if [ -n \"${NCCL_IB_GID_INDEX}\"\ [e2e-llm-inference-service] \ ]; then\n echo \"[Infer RoCE] Using pre-configured NCCL_IB_GID_INDEX=${NCCL_IB_GID_INDEX}\ [e2e-llm-inference-service] \ from environment\"\n export NVSHMEM_IB_GID_INDEX=${NVSHMEM_IB_GID_INDEX:-$NCCL_IB_GID_INDEX}\n\ [e2e-llm-inference-service] \ export UCX_IB_GID_INDEX=${UCX_IB_GID_INDEX:-$NCCL_IB_GID_INDEX}\n\ [e2e-llm-inference-service] \ echo \"[Infer RoCE] Using pre-configured GID_INDEX=${NCCL_IB_GID_INDEX}\ [e2e-llm-inference-service] \ for NCCL, NVSHMEM, and UCX\"\n elif [ -n \"$best_gid_index\" ]; then\n\ [e2e-llm-inference-service] \ echo \"[Infer RoCE] Selected GID_INDEX: ${best_gid_index} (found\ [e2e-llm-inference-service] \ on ${max_count} HCAs)\"\n\n export NCCL_IB_GID_INDEX=${NCCL_IB_GID_INDEX:-$best_gid_index}\n\ [e2e-llm-inference-service] \ export NVSHMEM_IB_GID_INDEX=${NVSHMEM_IB_GID_INDEX:-$best_gid_index}\n\ [e2e-llm-inference-service] \ export UCX_IB_GID_INDEX=${UCX_IB_GID_INDEX:-$best_gid_index}\n\ [e2e-llm-inference-service] \n echo \"[Infer RoCE] Exported GID_INDEX=${best_gid_index} for\ [e2e-llm-inference-service] \ NCCL, NVSHMEM, and UCX\"\n else\n echo \"[Infer RoCE] ERROR:\ [e2e-llm-inference-service] \ No valid IPv4 ${KSERVE_INFER_IB_GID_INDEX_GREP} GID_INDEX found on any\ [e2e-llm-inference-service] \ HCA.\"\n fi\n else\n echo \"[Infer RoCE] No active HCAs found,\ [e2e-llm-inference-service] \ skipping GID_INDEX inference.\"\n fi\nfi\n\n# --disable-access-log-for-endpoints\ [e2e-llm-inference-service] \ landed in vLLM 0.16.0 (vllm-project/vllm#30011).\n# Older versions still\ [e2e-llm-inference-service] \ need the blanket --disable-uvicorn-access-log.\nACCESS_LOG_ARGS=\"--disable-uvicorn-access-log\"\ [e2e-llm-inference-service] \nVLLM_VERSION=$(vllm --version 2>/dev/null | tail -1 | awk '{print $NF}')\n\ [e2e-llm-inference-service] echo \"[access-log-detect] vllm version='${VLLM_VERSION}'\"\nif [[ \"$VLLM_VERSION\"\ [e2e-llm-inference-service] \ =~ ^[0-9]+\\.[0-9]+ ]] && [ \"$(printf '%s\\n%s\\n' \"0.16.0\" \"${VLLM_VERSION}\"\ [e2e-llm-inference-service] \ | sort -V | head -1)\" = \"0.16.0\" ]; then\n ACCESS_LOG_ARGS=\"--disable-access-log-for-endpoints\ [e2e-llm-inference-service] \ /health,/metrics,/ping\"\nfi\necho \"[access-log-detect] selected ACCESS_LOG_ARGS='${ACCESS_LOG_ARGS}'\"\ [e2e-llm-inference-service] \n\n# --shutdown-timeout landed in vLLM 0.18.0 (vllm-project/vllm#36666).\n\ [e2e-llm-inference-service] SHUTDOWN_TIMEOUT_ARGS=\"\"\nif [[ \"$VLLM_VERSION\" =~ ^[0-9]+\\.[0-9]+\ [e2e-llm-inference-service] \ ]] && [ \"$(printf '%s\\n%s\\n' \"0.18.0\" \"${VLLM_VERSION}\" | sort\ [e2e-llm-inference-service] \ -V | head -1)\" = \"0.18.0\" ]; then\n SHUTDOWN_TIMEOUT_ARGS=\"--shutdown-timeout\ [e2e-llm-inference-service] \ 40\"\nfi\n\neval \"exec vllm serve /mnt/models \\\n --served-model-name\ [e2e-llm-inference-service] \ \"facebook/opt-125m\" \"publishers/kserve-ci-e2e-test/models/facebook/opt-125m\"\ [e2e-llm-inference-service] \ \\\n --port 8000 \\\n ${ACCESS_LOG_ARGS} \\\n ${SHUTDOWN_TIMEOUT_ARGS}\ [e2e-llm-inference-service] \ \\\n --enable-ssl-refresh \\\n --ssl-certfile /var/run/kserve/tls/tls.crt\ [e2e-llm-inference-service] \ \\\n --ssl-keyfile /var/run/kserve/tls/tls.key \\\n ${VLLM_ADDITIONAL_ARGS}\ [e2e-llm-inference-service] \ \\\n $@\"" [e2e-llm-inference-service] - -- [e2e-llm-inference-service] ports: [e2e-llm-inference-service] - containerPort: 8000 [e2e-llm-inference-service] protocol: TCP [e2e-llm-inference-service] env: [e2e-llm-inference-service] - name: HOME [e2e-llm-inference-service] value: /home [e2e-llm-inference-service] - name: VLLM_LOGGING_LEVEL [e2e-llm-inference-service] value: DEBUG [e2e-llm-inference-service] - name: VLLM_CPU_KVCACHE_SPACE [e2e-llm-inference-service] value: '1' [e2e-llm-inference-service] - name: VLLM_ENABLE_V1_MULTIPROCESSING [e2e-llm-inference-service] value: '0' [e2e-llm-inference-service] - name: USER [e2e-llm-inference-service] value: nonroot [e2e-llm-inference-service] - name: TORCHINDUCTOR_CACHE_DIR [e2e-llm-inference-service] value: /tmp/torchinductor-cache [e2e-llm-inference-service] - name: HF_HUB_CACHE [e2e-llm-inference-service] value: /models [e2e-llm-inference-service] resources: [e2e-llm-inference-service] limits: [e2e-llm-inference-service] cpu: '2' [e2e-llm-inference-service] memory: 7Gi [e2e-llm-inference-service] requests: [e2e-llm-inference-service] cpu: 200m [e2e-llm-inference-service] memory: 2Gi [e2e-llm-inference-service] volumeMounts: [e2e-llm-inference-service] - name: home [e2e-llm-inference-service] mountPath: /home [e2e-llm-inference-service] - name: tmp-dir [e2e-llm-inference-service] mountPath: /tmp [e2e-llm-inference-service] - name: dshm [e2e-llm-inference-service] mountPath: /dev/shm [e2e-llm-inference-service] - name: model-cache [e2e-llm-inference-service] mountPath: /models [e2e-llm-inference-service] - name: tls-certs [e2e-llm-inference-service] readOnly: true [e2e-llm-inference-service] mountPath: /var/run/kserve/tls [e2e-llm-inference-service] - name: kserve-provision-location [e2e-llm-inference-service] readOnly: true [e2e-llm-inference-service] mountPath: /mnt/models [e2e-llm-inference-service] livenessProbe: [e2e-llm-inference-service] httpGet: [e2e-llm-inference-service] path: /health [e2e-llm-inference-service] port: 8000 [e2e-llm-inference-service] scheme: HTTPS [e2e-llm-inference-service] timeoutSeconds: 1 [e2e-llm-inference-service] periodSeconds: 10 [e2e-llm-inference-service] successThreshold: 1 [e2e-llm-inference-service] failureThreshold: 10 [e2e-llm-inference-service] readinessProbe: [e2e-llm-inference-service] httpGet: [e2e-llm-inference-service] path: /health [e2e-llm-inference-service] port: 8000 [e2e-llm-inference-service] scheme: HTTPS [e2e-llm-inference-service] timeoutSeconds: 1 [e2e-llm-inference-service] periodSeconds: 1 [e2e-llm-inference-service] successThreshold: 1 [e2e-llm-inference-service] failureThreshold: 2 [e2e-llm-inference-service] startupProbe: [e2e-llm-inference-service] httpGet: [e2e-llm-inference-service] path: /health [e2e-llm-inference-service] port: 8000 [e2e-llm-inference-service] scheme: HTTPS [e2e-llm-inference-service] timeoutSeconds: 1 [e2e-llm-inference-service] periodSeconds: 10 [e2e-llm-inference-service] successThreshold: 1 [e2e-llm-inference-service] failureThreshold: 60 [e2e-llm-inference-service] lifecycle: [e2e-llm-inference-service] preStop: [e2e-llm-inference-service] exec: [e2e-llm-inference-service] command: [e2e-llm-inference-service] - /bin/sleep [e2e-llm-inference-service] - '15' [e2e-llm-inference-service] terminationMessagePath: /dev/termination-log [e2e-llm-inference-service] terminationMessagePolicy: FallbackToLogsOnError [e2e-llm-inference-service] imagePullPolicy: IfNotPresent [e2e-llm-inference-service] securityContext: [e2e-llm-inference-service] capabilities: [e2e-llm-inference-service] drop: [e2e-llm-inference-service] - ALL [e2e-llm-inference-service] runAsNonRoot: true [e2e-llm-inference-service] readOnlyRootFilesystem: true [e2e-llm-inference-service] allowPrivilegeEscalation: false [e2e-llm-inference-service] seccompProfile: [e2e-llm-inference-service] type: RuntimeDefault [e2e-llm-inference-service] restartPolicy: Always [e2e-llm-inference-service] terminationGracePeriodSeconds: 60 [e2e-llm-inference-service] dnsPolicy: ClusterFirst [e2e-llm-inference-service] securityContext: {} [e2e-llm-inference-service] schedulerName: default-scheduler [e2e-llm-inference-service] strategy: [e2e-llm-inference-service] type: RollingUpdate [e2e-llm-inference-service] rollingUpdate: [e2e-llm-inference-service] maxUnavailable: 25% [e2e-llm-inference-service] maxSurge: 25% [e2e-llm-inference-service] revisionHistoryLimit: 10 [e2e-llm-inference-service] progressDeadlineSeconds: 600 [e2e-llm-inference-service] status: [e2e-llm-inference-service] observedGeneration: 1 [e2e-llm-inference-service] replicas: 1 [e2e-llm-inference-service] updatedReplicas: 1 [e2e-llm-inference-service] readyReplicas: 1 [e2e-llm-inference-service] availableReplicas: 1 [e2e-llm-inference-service] conditions: [e2e-llm-inference-service] - type: Available [e2e-llm-inference-service] status: 'True' [e2e-llm-inference-service] lastUpdateTime: '2026-07-07T14:33:40Z' [e2e-llm-inference-service] lastTransitionTime: '2026-07-07T14:33:40Z' [e2e-llm-inference-service] reason: MinimumReplicasAvailable [e2e-llm-inference-service] message: Deployment has minimum availability. [e2e-llm-inference-service] - type: Progressing [e2e-llm-inference-service] status: 'True' [e2e-llm-inference-service] lastUpdateTime: '2026-07-07T14:33:40Z' [e2e-llm-inference-service] lastTransitionTime: '2026-07-07T14:31:40Z' [e2e-llm-inference-service] reason: NewReplicaSetAvailable [e2e-llm-inference-service] message: ReplicaSet "llmisvc-model-fb-opt-125m-route-87882a8e-kserve-5bd7864c6c" [e2e-llm-inference-service] has successfully progressed. [e2e-llm-inference-service] apiVersion: apps/v1 [e2e-llm-inference-service] kind: Deployment [e2e-llm-inference-service] [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1254 --- [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1255 metadata: [e2e-llm-inference-service] name: llmisvca690bbc929faec8bc98c767f16c003c1-kserve-router-scheduler [e2e-llm-inference-service] namespace: kserve-ci-e2e-test [e2e-llm-inference-service] uid: 3c03b0dd-9720-4679-b828-aaa0b86d2325 [e2e-llm-inference-service] resourceVersion: '31669' [e2e-llm-inference-service] generation: 1 [e2e-llm-inference-service] creationTimestamp: '2026-07-07T14:31:40Z' [e2e-llm-inference-service] labels: [e2e-llm-inference-service] app.kubernetes.io/component: llminferenceservice-router-scheduler [e2e-llm-inference-service] app.kubernetes.io/name: llmisvc-model-fb-opt-125m-route-87882a8e [e2e-llm-inference-service] app.kubernetes.io/part-of: llminferenceservice [e2e-llm-inference-service] annotations: [e2e-llm-inference-service] deployment.kubernetes.io/revision: '1' [e2e-llm-inference-service] ownerReferences: [e2e-llm-inference-service] - apiVersion: serving.kserve.io/v1alpha2 [e2e-llm-inference-service] kind: LLMInferenceService [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-route-87882a8e [e2e-llm-inference-service] uid: 88c60d0e-95bd-401f-ae1c-2c4c0b1e7c9d [e2e-llm-inference-service] controller: true [e2e-llm-inference-service] blockOwnerDeletion: true [e2e-llm-inference-service] managedFields: [e2e-llm-inference-service] - manager: manager [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] apiVersion: apps/v1 [e2e-llm-inference-service] time: '2026-07-07T14:31:40Z' [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:metadata: [e2e-llm-inference-service] f:labels: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:app.kubernetes.io/component: {} [e2e-llm-inference-service] f:app.kubernetes.io/name: {} [e2e-llm-inference-service] f:app.kubernetes.io/part-of: {} [e2e-llm-inference-service] f:ownerReferences: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"uid":"88c60d0e-95bd-401f-ae1c-2c4c0b1e7c9d"}: {} [e2e-llm-inference-service] f:spec: [e2e-llm-inference-service] f:progressDeadlineSeconds: {} [e2e-llm-inference-service] f:replicas: {} [e2e-llm-inference-service] f:revisionHistoryLimit: {} [e2e-llm-inference-service] f:selector: {} [e2e-llm-inference-service] f:strategy: [e2e-llm-inference-service] f:type: {} [e2e-llm-inference-service] f:template: [e2e-llm-inference-service] f:metadata: [e2e-llm-inference-service] f:annotations: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:app.kubernetes.io/version: {} [e2e-llm-inference-service] f:certificates.kserve.io/expiration-v2: {} [e2e-llm-inference-service] f:labels: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:app.kubernetes.io/component: {} [e2e-llm-inference-service] f:app.kubernetes.io/name: {} [e2e-llm-inference-service] f:app.kubernetes.io/part-of: {} [e2e-llm-inference-service] f:spec: [e2e-llm-inference-service] f:containers: [e2e-llm-inference-service] k:{"name":"main"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:args: {} [e2e-llm-inference-service] f:command: {} [e2e-llm-inference-service] f:env: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"name":"SSL_CERT_DIR"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] f:image: {} [e2e-llm-inference-service] f:imagePullPolicy: {} [e2e-llm-inference-service] f:lifecycle: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:preStop: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:exec: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:command: {} [e2e-llm-inference-service] f:livenessProbe: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:failureThreshold: {} [e2e-llm-inference-service] f:grpc: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:port: {} [e2e-llm-inference-service] f:service: {} [e2e-llm-inference-service] f:initialDelaySeconds: {} [e2e-llm-inference-service] f:periodSeconds: {} [e2e-llm-inference-service] f:successThreshold: {} [e2e-llm-inference-service] f:timeoutSeconds: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:ports: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"containerPort":5557,"protocol":"TCP"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:containerPort: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:protocol: {} [e2e-llm-inference-service] k:{"containerPort":9002,"protocol":"TCP"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:containerPort: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:protocol: {} [e2e-llm-inference-service] k:{"containerPort":9003,"protocol":"TCP"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:containerPort: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:protocol: {} [e2e-llm-inference-service] k:{"containerPort":9090,"protocol":"TCP"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:containerPort: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:protocol: {} [e2e-llm-inference-service] f:readinessProbe: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:failureThreshold: {} [e2e-llm-inference-service] f:grpc: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:port: {} [e2e-llm-inference-service] f:service: {} [e2e-llm-inference-service] f:initialDelaySeconds: {} [e2e-llm-inference-service] f:periodSeconds: {} [e2e-llm-inference-service] f:successThreshold: {} [e2e-llm-inference-service] f:timeoutSeconds: {} [e2e-llm-inference-service] f:resources: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:requests: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:cpu: {} [e2e-llm-inference-service] f:memory: {} [e2e-llm-inference-service] f:securityContext: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:allowPrivilegeEscalation: {} [e2e-llm-inference-service] f:capabilities: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:drop: {} [e2e-llm-inference-service] f:readOnlyRootFilesystem: {} [e2e-llm-inference-service] f:runAsNonRoot: {} [e2e-llm-inference-service] f:seccompProfile: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:type: {} [e2e-llm-inference-service] f:terminationMessagePath: {} [e2e-llm-inference-service] f:terminationMessagePolicy: {} [e2e-llm-inference-service] f:volumeMounts: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"mountPath":"/tmp/tokenizer"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:mountPath: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] k:{"mountPath":"/var/run/kserve/tls"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:mountPath: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:readOnly: {} [e2e-llm-inference-service] f:dnsPolicy: {} [e2e-llm-inference-service] f:initContainers: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"name":"storage-initializer"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:args: {} [e2e-llm-inference-service] f:env: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"name":"AWS_ACCESS_KEY_ID"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:valueFrom: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:secretKeyRef: {} [e2e-llm-inference-service] k:{"name":"AWS_CA_BUNDLE"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"AWS_CA_BUNDLE_CONFIGMAP"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"AWS_ENDPOINT_URL"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"AWS_SECRET_ACCESS_KEY"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:valueFrom: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:secretKeyRef: {} [e2e-llm-inference-service] k:{"name":"HF_HUB_ENABLE_HF_TRANSFER"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"HF_XET_HIGH_PERFORMANCE"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"HF_XET_NUM_CONCURRENT_RANGE_GETS"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"S3_ENDPOINT"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"S3_USE_HTTPS"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"S3_VERIFY_SSL"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"STORAGE_ALLOW_PATTERNS"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] f:image: {} [e2e-llm-inference-service] f:imagePullPolicy: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:resources: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:limits: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:cpu: {} [e2e-llm-inference-service] f:memory: {} [e2e-llm-inference-service] f:requests: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:cpu: {} [e2e-llm-inference-service] f:memory: {} [e2e-llm-inference-service] f:terminationMessagePath: {} [e2e-llm-inference-service] f:terminationMessagePolicy: {} [e2e-llm-inference-service] f:volumeMounts: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"mountPath":"/mnt/models"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:mountPath: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:restartPolicy: {} [e2e-llm-inference-service] f:schedulerName: {} [e2e-llm-inference-service] f:securityContext: {} [e2e-llm-inference-service] f:serviceAccount: {} [e2e-llm-inference-service] f:serviceAccountName: {} [e2e-llm-inference-service] f:terminationGracePeriodSeconds: {} [e2e-llm-inference-service] f:volumes: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"name":"kserve-provision-location"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:emptyDir: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] k:{"name":"tls-certs"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:secret: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:defaultMode: {} [e2e-llm-inference-service] f:secretName: {} [e2e-llm-inference-service] k:{"name":"tokenizer-cache"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:emptyDir: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] k:{"name":"tokenizer-tmp"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:emptyDir: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] k:{"name":"tokenizer-uds"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:emptyDir: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] - manager: kube-controller-manager [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] apiVersion: apps/v1 [e2e-llm-inference-service] time: '2026-07-07T14:32:14Z' [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:metadata: [e2e-llm-inference-service] f:annotations: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:deployment.kubernetes.io/revision: {} [e2e-llm-inference-service] f:status: [e2e-llm-inference-service] f:availableReplicas: {} [e2e-llm-inference-service] f:conditions: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"type":"Available"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:lastTransitionTime: {} [e2e-llm-inference-service] f:lastUpdateTime: {} [e2e-llm-inference-service] f:message: {} [e2e-llm-inference-service] f:reason: {} [e2e-llm-inference-service] f:status: {} [e2e-llm-inference-service] f:type: {} [e2e-llm-inference-service] k:{"type":"Progressing"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:lastTransitionTime: {} [e2e-llm-inference-service] f:lastUpdateTime: {} [e2e-llm-inference-service] f:message: {} [e2e-llm-inference-service] f:reason: {} [e2e-llm-inference-service] f:status: {} [e2e-llm-inference-service] f:type: {} [e2e-llm-inference-service] f:observedGeneration: {} [e2e-llm-inference-service] f:readyReplicas: {} [e2e-llm-inference-service] f:replicas: {} [e2e-llm-inference-service] f:updatedReplicas: {} [e2e-llm-inference-service] subresource: status [e2e-llm-inference-service] spec: [e2e-llm-inference-service] replicas: 1 [e2e-llm-inference-service] selector: [e2e-llm-inference-service] matchLabels: [e2e-llm-inference-service] app.kubernetes.io/component: llminferenceservice-router-scheduler [e2e-llm-inference-service] app.kubernetes.io/name: llmisvc-model-fb-opt-125m-route-87882a8e [e2e-llm-inference-service] app.kubernetes.io/part-of: llminferenceservice [e2e-llm-inference-service] template: [e2e-llm-inference-service] metadata: [e2e-llm-inference-service] labels: [e2e-llm-inference-service] app.kubernetes.io/component: llminferenceservice-router-scheduler [e2e-llm-inference-service] app.kubernetes.io/name: llmisvc-model-fb-opt-125m-route-87882a8e [e2e-llm-inference-service] app.kubernetes.io/part-of: llminferenceservice [e2e-llm-inference-service] annotations: [e2e-llm-inference-service] app.kubernetes.io/version: 0.9.0 [e2e-llm-inference-service] certificates.kserve.io/expiration-v2: 'true' [e2e-llm-inference-service] spec: [e2e-llm-inference-service] volumes: [e2e-llm-inference-service] - name: tls-certs [e2e-llm-inference-service] secret: [e2e-llm-inference-service] secretName: llmisva690bbc929faec8bc98c767f16c003c1-kserve-self-signed-certs [e2e-llm-inference-service] defaultMode: 420 [e2e-llm-inference-service] - name: tokenizer-uds [e2e-llm-inference-service] emptyDir: {} [e2e-llm-inference-service] - name: tokenizer-tmp [e2e-llm-inference-service] emptyDir: {} [e2e-llm-inference-service] - name: tokenizer-cache [e2e-llm-inference-service] emptyDir: {} [e2e-llm-inference-service] - name: kserve-provision-location [e2e-llm-inference-service] emptyDir: {} [e2e-llm-inference-service] initContainers: [e2e-llm-inference-service] - name: storage-initializer [e2e-llm-inference-service] image: quay.io/opendatahub/kserve-storage-initializer@sha256:13ad765c0f323200e1db7b45c99661287e65c5158d6b54b16139829636d20be4 [e2e-llm-inference-service] args: [e2e-llm-inference-service] - hf://facebook/opt-125m [e2e-llm-inference-service] - /mnt/models [e2e-llm-inference-service] env: [e2e-llm-inference-service] - name: AWS_ACCESS_KEY_ID [e2e-llm-inference-service] valueFrom: [e2e-llm-inference-service] secretKeyRef: [e2e-llm-inference-service] name: seaweedfs-s3-creds [e2e-llm-inference-service] key: AWS_ACCESS_KEY_ID [e2e-llm-inference-service] - name: AWS_SECRET_ACCESS_KEY [e2e-llm-inference-service] valueFrom: [e2e-llm-inference-service] secretKeyRef: [e2e-llm-inference-service] name: seaweedfs-s3-creds [e2e-llm-inference-service] key: AWS_SECRET_ACCESS_KEY [e2e-llm-inference-service] - name: S3_USE_HTTPS [e2e-llm-inference-service] value: '0' [e2e-llm-inference-service] - name: S3_ENDPOINT [e2e-llm-inference-service] value: s3-service.kserve:8333 [e2e-llm-inference-service] - name: AWS_ENDPOINT_URL [e2e-llm-inference-service] value: http://s3-service.kserve:8333 [e2e-llm-inference-service] - name: S3_VERIFY_SSL [e2e-llm-inference-service] value: '0' [e2e-llm-inference-service] - name: AWS_CA_BUNDLE [e2e-llm-inference-service] value: /etc/ssl/custom-certs/cabundle.crt [e2e-llm-inference-service] - name: AWS_CA_BUNDLE_CONFIGMAP [e2e-llm-inference-service] value: odh-kserve-custom-ca-bundle [e2e-llm-inference-service] - name: HF_HUB_ENABLE_HF_TRANSFER [e2e-llm-inference-service] value: '1' [e2e-llm-inference-service] - name: HF_XET_HIGH_PERFORMANCE [e2e-llm-inference-service] value: '1' [e2e-llm-inference-service] - name: HF_XET_NUM_CONCURRENT_RANGE_GETS [e2e-llm-inference-service] value: '8' [e2e-llm-inference-service] - name: STORAGE_ALLOW_PATTERNS [e2e-llm-inference-service] value: '["tokenizer.json", "tokenizer_config.json", "special_tokens_map.json", [e2e-llm-inference-service] "vocab.json", "merges.txt", "config.json", "generation_config.json"]' [e2e-llm-inference-service] resources: [e2e-llm-inference-service] limits: [e2e-llm-inference-service] cpu: '1' [e2e-llm-inference-service] memory: 24Gi [e2e-llm-inference-service] requests: [e2e-llm-inference-service] cpu: 100m [e2e-llm-inference-service] memory: 100Mi [e2e-llm-inference-service] volumeMounts: [e2e-llm-inference-service] - name: kserve-provision-location [e2e-llm-inference-service] mountPath: /mnt/models [e2e-llm-inference-service] terminationMessagePath: /dev/termination-log [e2e-llm-inference-service] terminationMessagePolicy: FallbackToLogsOnError [e2e-llm-inference-service] imagePullPolicy: IfNotPresent [e2e-llm-inference-service] containers: [e2e-llm-inference-service] - name: main [e2e-llm-inference-service] image: ghcr.io/llm-d/llm-d-router-endpoint-picker:v0.9.0-rc.2 [e2e-llm-inference-service] command: [e2e-llm-inference-service] - /app/epp [e2e-llm-inference-service] - --pool-name [e2e-llm-inference-service] - llmisvc-model-fb-opt-125m-route-87882a8e-inference-pool [e2e-llm-inference-service] - --pool-namespace [e2e-llm-inference-service] - kserve-ci-e2e-test [e2e-llm-inference-service] - --zap-encoder [e2e-llm-inference-service] - json [e2e-llm-inference-service] - --grpc-port [e2e-llm-inference-service] - '9002' [e2e-llm-inference-service] - --grpc-health-port [e2e-llm-inference-service] - '9003' [e2e-llm-inference-service] - --enable-cert-reload=true [e2e-llm-inference-service] - --secure-serving=true [e2e-llm-inference-service] - --model-server-metrics-scheme=https [e2e-llm-inference-service] - --cert-path=/var/run/kserve/tls [e2e-llm-inference-service] args: [e2e-llm-inference-service] - --config-text [e2e-llm-inference-service] - "apiVersion: inference.networking.x-k8s.io/v1alpha1\nkind: EndpointPickerConfig\n\ [e2e-llm-inference-service] plugins:\n- type: single-profile-handler\n- type: queue-scorer\n- type:\ [e2e-llm-inference-service] \ prefix-cache-scorer\n- type: max-score-picker\nschedulingProfiles:\n-\ [e2e-llm-inference-service] \ name: default\n plugins:\n - pluginRef: queue-scorer\n weight: 2\n\ [e2e-llm-inference-service] \ - pluginRef: prefix-cache-scorer\n weight: 3\n - pluginRef: max-score-picker\n" [e2e-llm-inference-service] ports: [e2e-llm-inference-service] - name: grpc [e2e-llm-inference-service] containerPort: 9002 [e2e-llm-inference-service] protocol: TCP [e2e-llm-inference-service] - name: grpc-health [e2e-llm-inference-service] containerPort: 9003 [e2e-llm-inference-service] protocol: TCP [e2e-llm-inference-service] - name: metrics [e2e-llm-inference-service] containerPort: 9090 [e2e-llm-inference-service] protocol: TCP [e2e-llm-inference-service] - name: zmq [e2e-llm-inference-service] containerPort: 5557 [e2e-llm-inference-service] protocol: TCP [e2e-llm-inference-service] env: [e2e-llm-inference-service] - name: SSL_CERT_DIR [e2e-llm-inference-service] value: /var/run/kserve/tls:/var/run/secrets/kubernetes.io/serviceaccount:/etc/pki/tls/certs [e2e-llm-inference-service] resources: [e2e-llm-inference-service] requests: [e2e-llm-inference-service] cpu: 256m [e2e-llm-inference-service] memory: 500Mi [e2e-llm-inference-service] volumeMounts: [e2e-llm-inference-service] - name: tls-certs [e2e-llm-inference-service] readOnly: true [e2e-llm-inference-service] mountPath: /var/run/kserve/tls [e2e-llm-inference-service] - name: tokenizer-uds [e2e-llm-inference-service] mountPath: /tmp/tokenizer [e2e-llm-inference-service] livenessProbe: [e2e-llm-inference-service] grpc: [e2e-llm-inference-service] port: 9003 [e2e-llm-inference-service] service: liveness [e2e-llm-inference-service] initialDelaySeconds: 5 [e2e-llm-inference-service] timeoutSeconds: 1 [e2e-llm-inference-service] periodSeconds: 10 [e2e-llm-inference-service] successThreshold: 1 [e2e-llm-inference-service] failureThreshold: 3 [e2e-llm-inference-service] readinessProbe: [e2e-llm-inference-service] grpc: [e2e-llm-inference-service] port: 9003 [e2e-llm-inference-service] service: readiness [e2e-llm-inference-service] initialDelaySeconds: 30 [e2e-llm-inference-service] timeoutSeconds: 1 [e2e-llm-inference-service] periodSeconds: 10 [e2e-llm-inference-service] successThreshold: 1 [e2e-llm-inference-service] failureThreshold: 3 [e2e-llm-inference-service] lifecycle: [e2e-llm-inference-service] preStop: [e2e-llm-inference-service] exec: [e2e-llm-inference-service] command: [e2e-llm-inference-service] - /bin/sleep [e2e-llm-inference-service] - '15' [e2e-llm-inference-service] terminationMessagePath: /dev/termination-log [e2e-llm-inference-service] terminationMessagePolicy: FallbackToLogsOnError [e2e-llm-inference-service] imagePullPolicy: IfNotPresent [e2e-llm-inference-service] securityContext: [e2e-llm-inference-service] capabilities: [e2e-llm-inference-service] drop: [e2e-llm-inference-service] - ALL [e2e-llm-inference-service] runAsNonRoot: true [e2e-llm-inference-service] readOnlyRootFilesystem: true [e2e-llm-inference-service] allowPrivilegeEscalation: false [e2e-llm-inference-service] seccompProfile: [e2e-llm-inference-service] type: RuntimeDefault [e2e-llm-inference-service] restartPolicy: Always [e2e-llm-inference-service] terminationGracePeriodSeconds: 60 [e2e-llm-inference-service] dnsPolicy: ClusterFirst [e2e-llm-inference-service] serviceAccountName: llmisvc-model-fb-opt-125m-route-87882a8e-epp-sa [e2e-llm-inference-service] serviceAccount: llmisvc-model-fb-opt-125m-route-87882a8e-epp-sa [e2e-llm-inference-service] securityContext: {} [e2e-llm-inference-service] schedulerName: default-scheduler [e2e-llm-inference-service] strategy: [e2e-llm-inference-service] type: Recreate [e2e-llm-inference-service] revisionHistoryLimit: 10 [e2e-llm-inference-service] progressDeadlineSeconds: 600 [e2e-llm-inference-service] status: [e2e-llm-inference-service] observedGeneration: 1 [e2e-llm-inference-service] replicas: 1 [e2e-llm-inference-service] updatedReplicas: 1 [e2e-llm-inference-service] readyReplicas: 1 [e2e-llm-inference-service] availableReplicas: 1 [e2e-llm-inference-service] conditions: [e2e-llm-inference-service] - type: Available [e2e-llm-inference-service] status: 'True' [e2e-llm-inference-service] lastUpdateTime: '2026-07-07T14:32:14Z' [e2e-llm-inference-service] lastTransitionTime: '2026-07-07T14:32:14Z' [e2e-llm-inference-service] reason: MinimumReplicasAvailable [e2e-llm-inference-service] message: Deployment has minimum availability. [e2e-llm-inference-service] - type: Progressing [e2e-llm-inference-service] status: 'True' [e2e-llm-inference-service] lastUpdateTime: '2026-07-07T14:32:14Z' [e2e-llm-inference-service] lastTransitionTime: '2026-07-07T14:31:40Z' [e2e-llm-inference-service] reason: NewReplicaSetAvailable [e2e-llm-inference-service] message: ReplicaSet "llmisvca690bbc929faec8bc98c767f16c003c1-kserve-router-scheduler-5d6998448f" [e2e-llm-inference-service] has successfully progressed. [e2e-llm-inference-service] apiVersion: apps/v1 [e2e-llm-inference-service] kind: Deployment [e2e-llm-inference-service] [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1254 --- [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1255 metadata: [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-route-87882a8e-kserve-5bd7864c6c [e2e-llm-inference-service] namespace: kserve-ci-e2e-test [e2e-llm-inference-service] uid: 6072d6ea-6830-4187-b75d-75a1ebdb2b52 [e2e-llm-inference-service] resourceVersion: '33110' [e2e-llm-inference-service] generation: 1 [e2e-llm-inference-service] creationTimestamp: '2026-07-07T14:31:40Z' [e2e-llm-inference-service] labels: [e2e-llm-inference-service] app.kubernetes.io/component: llminferenceservice-workload [e2e-llm-inference-service] app.kubernetes.io/name: llmisvc-model-fb-opt-125m-route-87882a8e [e2e-llm-inference-service] app.kubernetes.io/part-of: llminferenceservice [e2e-llm-inference-service] kserve.io/component: workload [e2e-llm-inference-service] llm-d.ai/role: both [e2e-llm-inference-service] pod-template-hash: 5bd7864c6c [e2e-llm-inference-service] annotations: [e2e-llm-inference-service] deployment.kubernetes.io/desired-replicas: '1' [e2e-llm-inference-service] deployment.kubernetes.io/max-replicas: '2' [e2e-llm-inference-service] deployment.kubernetes.io/revision: '1' [e2e-llm-inference-service] ownerReferences: [e2e-llm-inference-service] - apiVersion: apps/v1 [e2e-llm-inference-service] kind: Deployment [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-route-87882a8e-kserve [e2e-llm-inference-service] uid: 6006b5ee-4a45-4a5d-9465-bae07199d5c5 [e2e-llm-inference-service] controller: true [e2e-llm-inference-service] blockOwnerDeletion: true [e2e-llm-inference-service] managedFields: [e2e-llm-inference-service] - manager: kube-controller-manager [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] apiVersion: apps/v1 [e2e-llm-inference-service] time: '2026-07-07T14:31:40Z' [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:metadata: [e2e-llm-inference-service] f:annotations: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:deployment.kubernetes.io/desired-replicas: {} [e2e-llm-inference-service] f:deployment.kubernetes.io/max-replicas: {} [e2e-llm-inference-service] f:deployment.kubernetes.io/revision: {} [e2e-llm-inference-service] f:labels: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:app.kubernetes.io/component: {} [e2e-llm-inference-service] f:app.kubernetes.io/name: {} [e2e-llm-inference-service] f:app.kubernetes.io/part-of: {} [e2e-llm-inference-service] f:kserve.io/component: {} [e2e-llm-inference-service] f:llm-d.ai/role: {} [e2e-llm-inference-service] f:pod-template-hash: {} [e2e-llm-inference-service] f:ownerReferences: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"uid":"6006b5ee-4a45-4a5d-9465-bae07199d5c5"}: {} [e2e-llm-inference-service] f:spec: [e2e-llm-inference-service] f:replicas: {} [e2e-llm-inference-service] f:selector: {} [e2e-llm-inference-service] f:template: [e2e-llm-inference-service] f:metadata: [e2e-llm-inference-service] f:labels: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:app.kubernetes.io/component: {} [e2e-llm-inference-service] f:app.kubernetes.io/name: {} [e2e-llm-inference-service] f:app.kubernetes.io/part-of: {} [e2e-llm-inference-service] f:kserve.io/component: {} [e2e-llm-inference-service] f:llm-d.ai/role: {} [e2e-llm-inference-service] f:pod-template-hash: {} [e2e-llm-inference-service] f:spec: [e2e-llm-inference-service] f:containers: [e2e-llm-inference-service] k:{"name":"main"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:command: {} [e2e-llm-inference-service] f:env: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"name":"HF_HUB_CACHE"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"HOME"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"TORCHINDUCTOR_CACHE_DIR"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"USER"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"VLLM_CPU_KVCACHE_SPACE"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"VLLM_ENABLE_V1_MULTIPROCESSING"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"VLLM_LOGGING_LEVEL"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] f:image: {} [e2e-llm-inference-service] f:imagePullPolicy: {} [e2e-llm-inference-service] f:lifecycle: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:preStop: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:exec: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:command: {} [e2e-llm-inference-service] f:livenessProbe: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:failureThreshold: {} [e2e-llm-inference-service] f:httpGet: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:path: {} [e2e-llm-inference-service] f:port: {} [e2e-llm-inference-service] f:scheme: {} [e2e-llm-inference-service] f:periodSeconds: {} [e2e-llm-inference-service] f:successThreshold: {} [e2e-llm-inference-service] f:timeoutSeconds: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:ports: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"containerPort":8000,"protocol":"TCP"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:containerPort: {} [e2e-llm-inference-service] f:protocol: {} [e2e-llm-inference-service] f:readinessProbe: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:failureThreshold: {} [e2e-llm-inference-service] f:httpGet: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:path: {} [e2e-llm-inference-service] f:port: {} [e2e-llm-inference-service] f:scheme: {} [e2e-llm-inference-service] f:periodSeconds: {} [e2e-llm-inference-service] f:successThreshold: {} [e2e-llm-inference-service] f:timeoutSeconds: {} [e2e-llm-inference-service] f:resources: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:limits: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:cpu: {} [e2e-llm-inference-service] f:memory: {} [e2e-llm-inference-service] f:requests: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:cpu: {} [e2e-llm-inference-service] f:memory: {} [e2e-llm-inference-service] f:securityContext: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:allowPrivilegeEscalation: {} [e2e-llm-inference-service] f:capabilities: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:drop: {} [e2e-llm-inference-service] f:readOnlyRootFilesystem: {} [e2e-llm-inference-service] f:runAsNonRoot: {} [e2e-llm-inference-service] f:seccompProfile: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:type: {} [e2e-llm-inference-service] f:startupProbe: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:failureThreshold: {} [e2e-llm-inference-service] f:httpGet: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:path: {} [e2e-llm-inference-service] f:port: {} [e2e-llm-inference-service] f:scheme: {} [e2e-llm-inference-service] f:periodSeconds: {} [e2e-llm-inference-service] f:successThreshold: {} [e2e-llm-inference-service] f:timeoutSeconds: {} [e2e-llm-inference-service] f:terminationMessagePath: {} [e2e-llm-inference-service] f:terminationMessagePolicy: {} [e2e-llm-inference-service] f:volumeMounts: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"mountPath":"/dev/shm"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:mountPath: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] k:{"mountPath":"/home"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:mountPath: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] k:{"mountPath":"/mnt/models"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:mountPath: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:readOnly: {} [e2e-llm-inference-service] k:{"mountPath":"/models"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:mountPath: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] k:{"mountPath":"/tmp"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:mountPath: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] k:{"mountPath":"/var/run/kserve/tls"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:mountPath: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:readOnly: {} [e2e-llm-inference-service] f:dnsPolicy: {} [e2e-llm-inference-service] f:initContainers: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"name":"storage-initializer"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:args: {} [e2e-llm-inference-service] f:env: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"name":"AWS_ACCESS_KEY_ID"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:valueFrom: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:secretKeyRef: {} [e2e-llm-inference-service] k:{"name":"AWS_CA_BUNDLE"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"AWS_CA_BUNDLE_CONFIGMAP"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"AWS_ENDPOINT_URL"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"AWS_SECRET_ACCESS_KEY"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:valueFrom: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:secretKeyRef: {} [e2e-llm-inference-service] k:{"name":"HF_HUB_ENABLE_HF_TRANSFER"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"HF_XET_HIGH_PERFORMANCE"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"HF_XET_NUM_CONCURRENT_RANGE_GETS"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"S3_ENDPOINT"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"S3_USE_HTTPS"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"S3_VERIFY_SSL"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] f:image: {} [e2e-llm-inference-service] f:imagePullPolicy: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:resources: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:limits: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:cpu: {} [e2e-llm-inference-service] f:memory: {} [e2e-llm-inference-service] f:requests: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:cpu: {} [e2e-llm-inference-service] f:memory: {} [e2e-llm-inference-service] f:terminationMessagePath: {} [e2e-llm-inference-service] f:terminationMessagePolicy: {} [e2e-llm-inference-service] f:volumeMounts: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"mountPath":"/mnt/models"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:mountPath: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:restartPolicy: {} [e2e-llm-inference-service] f:schedulerName: {} [e2e-llm-inference-service] f:securityContext: {} [e2e-llm-inference-service] f:terminationGracePeriodSeconds: {} [e2e-llm-inference-service] f:volumes: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"name":"dshm"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:emptyDir: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:medium: {} [e2e-llm-inference-service] f:sizeLimit: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] k:{"name":"home"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:emptyDir: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] k:{"name":"kserve-provision-location"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:emptyDir: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] k:{"name":"model-cache"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:emptyDir: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] k:{"name":"tls-certs"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:secret: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:defaultMode: {} [e2e-llm-inference-service] f:secretName: {} [e2e-llm-inference-service] k:{"name":"tmp-dir"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:emptyDir: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] - manager: kube-controller-manager [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] apiVersion: apps/v1 [e2e-llm-inference-service] time: '2026-07-07T14:33:40Z' [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:status: [e2e-llm-inference-service] f:availableReplicas: {} [e2e-llm-inference-service] f:fullyLabeledReplicas: {} [e2e-llm-inference-service] f:observedGeneration: {} [e2e-llm-inference-service] f:readyReplicas: {} [e2e-llm-inference-service] f:replicas: {} [e2e-llm-inference-service] subresource: status [e2e-llm-inference-service] spec: [e2e-llm-inference-service] replicas: 1 [e2e-llm-inference-service] selector: [e2e-llm-inference-service] matchLabels: [e2e-llm-inference-service] app.kubernetes.io/component: llminferenceservice-workload [e2e-llm-inference-service] app.kubernetes.io/name: llmisvc-model-fb-opt-125m-route-87882a8e [e2e-llm-inference-service] app.kubernetes.io/part-of: llminferenceservice [e2e-llm-inference-service] kserve.io/component: workload [e2e-llm-inference-service] llm-d.ai/role: both [e2e-llm-inference-service] pod-template-hash: 5bd7864c6c [e2e-llm-inference-service] template: [e2e-llm-inference-service] metadata: [e2e-llm-inference-service] labels: [e2e-llm-inference-service] app.kubernetes.io/component: llminferenceservice-workload [e2e-llm-inference-service] app.kubernetes.io/name: llmisvc-model-fb-opt-125m-route-87882a8e [e2e-llm-inference-service] app.kubernetes.io/part-of: llminferenceservice [e2e-llm-inference-service] kserve.io/component: workload [e2e-llm-inference-service] llm-d.ai/role: both [e2e-llm-inference-service] pod-template-hash: 5bd7864c6c [e2e-llm-inference-service] spec: [e2e-llm-inference-service] volumes: [e2e-llm-inference-service] - name: home [e2e-llm-inference-service] emptyDir: {} [e2e-llm-inference-service] - name: dshm [e2e-llm-inference-service] emptyDir: [e2e-llm-inference-service] medium: Memory [e2e-llm-inference-service] sizeLimit: 1Gi [e2e-llm-inference-service] - name: model-cache [e2e-llm-inference-service] emptyDir: {} [e2e-llm-inference-service] - name: tmp-dir [e2e-llm-inference-service] emptyDir: {} [e2e-llm-inference-service] - name: tls-certs [e2e-llm-inference-service] secret: [e2e-llm-inference-service] secretName: llmisva690bbc929faec8bc98c767f16c003c1-kserve-self-signed-certs [e2e-llm-inference-service] defaultMode: 420 [e2e-llm-inference-service] - name: kserve-provision-location [e2e-llm-inference-service] emptyDir: {} [e2e-llm-inference-service] initContainers: [e2e-llm-inference-service] - name: storage-initializer [e2e-llm-inference-service] image: quay.io/opendatahub/kserve-storage-initializer@sha256:13ad765c0f323200e1db7b45c99661287e65c5158d6b54b16139829636d20be4 [e2e-llm-inference-service] args: [e2e-llm-inference-service] - hf://facebook/opt-125m [e2e-llm-inference-service] - /mnt/models [e2e-llm-inference-service] env: [e2e-llm-inference-service] - name: AWS_ACCESS_KEY_ID [e2e-llm-inference-service] valueFrom: [e2e-llm-inference-service] secretKeyRef: [e2e-llm-inference-service] name: seaweedfs-s3-creds [e2e-llm-inference-service] key: AWS_ACCESS_KEY_ID [e2e-llm-inference-service] - name: AWS_SECRET_ACCESS_KEY [e2e-llm-inference-service] valueFrom: [e2e-llm-inference-service] secretKeyRef: [e2e-llm-inference-service] name: seaweedfs-s3-creds [e2e-llm-inference-service] key: AWS_SECRET_ACCESS_KEY [e2e-llm-inference-service] - name: S3_USE_HTTPS [e2e-llm-inference-service] value: '0' [e2e-llm-inference-service] - name: S3_ENDPOINT [e2e-llm-inference-service] value: s3-service.kserve:8333 [e2e-llm-inference-service] - name: AWS_ENDPOINT_URL [e2e-llm-inference-service] value: http://s3-service.kserve:8333 [e2e-llm-inference-service] - name: S3_VERIFY_SSL [e2e-llm-inference-service] value: '0' [e2e-llm-inference-service] - name: AWS_CA_BUNDLE [e2e-llm-inference-service] value: /etc/ssl/custom-certs/cabundle.crt [e2e-llm-inference-service] - name: AWS_CA_BUNDLE_CONFIGMAP [e2e-llm-inference-service] value: odh-kserve-custom-ca-bundle [e2e-llm-inference-service] - name: HF_HUB_ENABLE_HF_TRANSFER [e2e-llm-inference-service] value: '1' [e2e-llm-inference-service] - name: HF_XET_HIGH_PERFORMANCE [e2e-llm-inference-service] value: '1' [e2e-llm-inference-service] - name: HF_XET_NUM_CONCURRENT_RANGE_GETS [e2e-llm-inference-service] value: '8' [e2e-llm-inference-service] resources: [e2e-llm-inference-service] limits: [e2e-llm-inference-service] cpu: '1' [e2e-llm-inference-service] memory: 24Gi [e2e-llm-inference-service] requests: [e2e-llm-inference-service] cpu: 100m [e2e-llm-inference-service] memory: 100Mi [e2e-llm-inference-service] volumeMounts: [e2e-llm-inference-service] - name: kserve-provision-location [e2e-llm-inference-service] mountPath: /mnt/models [e2e-llm-inference-service] terminationMessagePath: /dev/termination-log [e2e-llm-inference-service] terminationMessagePolicy: FallbackToLogsOnError [e2e-llm-inference-service] imagePullPolicy: IfNotPresent [e2e-llm-inference-service] containers: [e2e-llm-inference-service] - name: main [e2e-llm-inference-service] image: public.ecr.aws/q9t5s3a7/vllm-cpu-release-repo:v0.19.0 [e2e-llm-inference-service] command: [e2e-llm-inference-service] - /bin/bash [e2e-llm-inference-service] - -c [e2e-llm-inference-service] - "if [ -f /etc/profile.d/ibm-aiu-setup.sh ]; then\n source /etc/profile.d/ibm-aiu-setup.sh\n\ [e2e-llm-inference-service] fi\n\nif [ \"$KSERVE_INFER_ROCE\" = \"true\" ]; then\n echo \"Trying to\ [e2e-llm-inference-service] \ infer RoCE configs ... \"\n grep -H . /sys/class/infiniband/*/ports/*/gids/*\ [e2e-llm-inference-service] \ 2>/dev/null\n grep -H . /sys/class/infiniband/*/ports/*/gid_attrs/types/*\ [e2e-llm-inference-service] \ 2>/dev/null\n\n cat /proc/driver/nvidia/params\n\n KSERVE_INFER_IB_GID_INDEX_GREP=${KSERVE_INFER_IB_GID_INDEX_GREP:-\"\ [e2e-llm-inference-service] RoCE v2\"}\n\n echo \"[Infer RoCE] Discovering active HCAs ...\"\n active_hcas=()\n\ [e2e-llm-inference-service] \ # Loop through all mlx5 devices found in sysfs\n for hca_dir in /sys/class/infiniband/mlx5_*;\ [e2e-llm-inference-service] \ do\n # Ensure it's a directory before proceeding\n if [ -d \"\ [e2e-llm-inference-service] $hca_dir\" ]; then\n hca_name=$(basename \"$hca_dir\")\n \ [e2e-llm-inference-service] \ port_state_file=\"$hca_dir/ports/1/state\" # Assume port 1\n \ [e2e-llm-inference-service] \ type_file=\"$hca_dir/ports/1/gid_attrs/types/*\"\n\n echo\ [e2e-llm-inference-service] \ \"[Infer RoCE] Check if the port state file ${port_state_file} exists\ [e2e-llm-inference-service] \ and contains 'ACTIVE'\"\n if [ -f \"$port_state_file\" ] && grep\ [e2e-llm-inference-service] \ -q \"ACTIVE\" \"$port_state_file\" && grep -q \"${KSERVE_INFER_IB_GID_INDEX_GREP}\"\ [e2e-llm-inference-service] \ ${type_file} 2>/dev/null; then\n echo \"[Infer RoCE] Found\ [e2e-llm-inference-service] \ active HCA: $hca_name\"\n active_hcas+=(\"$hca_name\")\n\ [e2e-llm-inference-service] \ else\n echo \"[Infer RoCE] Skipping inactive or\ [e2e-llm-inference-service] \ down HCA: $hca_name\"\n fi\n fi\n done\n\n # Check if\ [e2e-llm-inference-service] \ we found any active HCAs\n if [ ${#active_hcas[@]} -gt 0 ]; then\n \ [e2e-llm-inference-service] \ # Join the array elements with a comma\n hca_port_pairs=()\n \ [e2e-llm-inference-service] \ for hca in \"${active_hcas[@]}\"; do\n hca_port_pairs+=(\"\ [e2e-llm-inference-service] ${hca}:1\")\n done\n\n active_hca_list=$(IFS=,; echo \"${active_hcas[*]}\"\ [e2e-llm-inference-service] )\n hca_port_pairs_list=$(IFS=,; echo \"${hca_port_pairs[*]}\")\n \ [e2e-llm-inference-service] \ echo \"[Infer RoCE] Setting active HCAs: ${active_hca_list}\"\n \ [e2e-llm-inference-service] \ export NCCL_IB_HCA=${NCCL_IB_HCA:-${active_hca_list}}\n export\ [e2e-llm-inference-service] \ NVSHMEM_HCA_LIST=${NVSHMEM_HCA_LIST:-${hca_port_pairs_list}}\n export\ [e2e-llm-inference-service] \ UCX_NET_DEVICES=${UCX_NET_DEVICES:-${hca_port_pairs_list}}\n\n echo\ [e2e-llm-inference-service] \ \"[Infer RoCE] NCCL_IB_HCA=${NCCL_IB_HCA}\"\n echo \"[Infer RoCE]\ [e2e-llm-inference-service] \ NVSHMEM_HCA_LIST=${NVSHMEM_HCA_LIST}\"\n echo \"[Infer RoCE] UCX_NET_DEVICES=${UCX_NET_DEVICES}\"\ [e2e-llm-inference-service] \n else\n echo \"[Infer RoCE] WARNING: No active RoCE HCAs found.\ [e2e-llm-inference-service] \ NCCL_IB_HCA will not be set.\"\n fi\n\n if [ ${#active_hcas[@]} -gt\ [e2e-llm-inference-service] \ 0 ]; then\n echo \"[Infer RoCE] Finding GID_INDEX for each active\ [e2e-llm-inference-service] \ HCA (SR-IOV compatible)...\"\n\n # For SR-IOV environments, find\ [e2e-llm-inference-service] \ the most common IPv4 RoCE v2 GID index across all HCAs\n declare\ [e2e-llm-inference-service] \ -A gid_index_count\n declare -A hca_gid_index\n\n for hca_name\ [e2e-llm-inference-service] \ in \"${active_hcas[@]}\"; do\n echo \"[Infer RoCE] Processing\ [e2e-llm-inference-service] \ HCA: ${hca_name}\"\n\n # Find all RoCE v2 IPv4 GIDs for this\ [e2e-llm-inference-service] \ HCA and count by index\n for tpath in /sys/class/infiniband/${hca_name}/ports/1/gid_attrs/types/*;\ [e2e-llm-inference-service] \ do\n if grep -q \"${KSERVE_INFER_IB_GID_INDEX_GREP}\" \"\ [e2e-llm-inference-service] $tpath\" 2>/dev/null; then\n idx=$(basename \"$tpath\"\ [e2e-llm-inference-service] )\n gid_file=\"/sys/class/infiniband/${hca_name}/ports/1/gids/${idx}\"\ [e2e-llm-inference-service] \n # Check for IPv4 GID (contains ffff:)\n \ [e2e-llm-inference-service] \ if [ -f \"$gid_file\" ] && grep -q \"ffff:\" \"$gid_file\"; then\n\ [e2e-llm-inference-service] \ gid_value=$(cat \"$gid_file\" 2>/dev/null || echo\ [e2e-llm-inference-service] \ \"\")\n echo \"[Infer RoCE] Found IPv4 RoCE v2 GID\ [e2e-llm-inference-service] \ for ${hca_name}: index=${idx}, gid=${gid_value}\"\n \ [e2e-llm-inference-service] \ hca_gid_index[\"${hca_name}\"]=\"${idx}\"\n gid_index_count[\"\ [e2e-llm-inference-service] ${idx}\"]=$((${gid_index_count[\"${idx}\"]} + 1))\n \ [e2e-llm-inference-service] \ break # Use first found IPv4 GID per HCA\n fi\n \ [e2e-llm-inference-service] \ fi\n done\n done\n\n # Find the most common\ [e2e-llm-inference-service] \ GID index (most likely to be consistent across nodes)\n best_gid_index=\"\ [e2e-llm-inference-service] \"\n max_count=0\n for idx in \"${!gid_index_count[@]}\"; do\n\ [e2e-llm-inference-service] \ count=${gid_index_count[\"${idx}\"]}\n echo \"[Infer\ [e2e-llm-inference-service] \ RoCE] GID_INDEX ${idx} found on ${count} HCAs\"\n if [ $count\ [e2e-llm-inference-service] \ -gt $max_count ]; then\n max_count=$count\n \ [e2e-llm-inference-service] \ best_gid_index=\"$idx\"\n fi\n done\n\n # Use deterministic\ [e2e-llm-inference-service] \ fallback if tied - prefer index 3 (SR-IOV standard)\n if [ ${#gid_index_count[@]}\ [e2e-llm-inference-service] \ -gt 1 ]; then\n echo \"[Infer RoCE] Multiple GID indices found,\ [e2e-llm-inference-service] \ selecting most common: ${best_gid_index}\"\n # If there's a tie,\ [e2e-llm-inference-service] \ prefer index 3 as it's most common in SR-IOV setups\n if [ -n\ [e2e-llm-inference-service] \ \"${gid_index_count['3']}\" ] && [ \"${gid_index_count['3']}\" -eq \"\ [e2e-llm-inference-service] $max_count\" ]; then\n best_gid_index=\"3\"\n \ [e2e-llm-inference-service] \ echo \"[Infer RoCE] Using deterministic fallback: GID_INDEX=3 (SR-IOV\ [e2e-llm-inference-service] \ standard)\"\n fi\n fi\n\n # Check if GID_INDEX is already\ [e2e-llm-inference-service] \ set via environment variables\n if [ -n \"${NCCL_IB_GID_INDEX}\"\ [e2e-llm-inference-service] \ ]; then\n echo \"[Infer RoCE] Using pre-configured NCCL_IB_GID_INDEX=${NCCL_IB_GID_INDEX}\ [e2e-llm-inference-service] \ from environment\"\n export NVSHMEM_IB_GID_INDEX=${NVSHMEM_IB_GID_INDEX:-$NCCL_IB_GID_INDEX}\n\ [e2e-llm-inference-service] \ export UCX_IB_GID_INDEX=${UCX_IB_GID_INDEX:-$NCCL_IB_GID_INDEX}\n\ [e2e-llm-inference-service] \ echo \"[Infer RoCE] Using pre-configured GID_INDEX=${NCCL_IB_GID_INDEX}\ [e2e-llm-inference-service] \ for NCCL, NVSHMEM, and UCX\"\n elif [ -n \"$best_gid_index\" ]; then\n\ [e2e-llm-inference-service] \ echo \"[Infer RoCE] Selected GID_INDEX: ${best_gid_index} (found\ [e2e-llm-inference-service] \ on ${max_count} HCAs)\"\n\n export NCCL_IB_GID_INDEX=${NCCL_IB_GID_INDEX:-$best_gid_index}\n\ [e2e-llm-inference-service] \ export NVSHMEM_IB_GID_INDEX=${NVSHMEM_IB_GID_INDEX:-$best_gid_index}\n\ [e2e-llm-inference-service] \ export UCX_IB_GID_INDEX=${UCX_IB_GID_INDEX:-$best_gid_index}\n\ [e2e-llm-inference-service] \n echo \"[Infer RoCE] Exported GID_INDEX=${best_gid_index} for\ [e2e-llm-inference-service] \ NCCL, NVSHMEM, and UCX\"\n else\n echo \"[Infer RoCE] ERROR:\ [e2e-llm-inference-service] \ No valid IPv4 ${KSERVE_INFER_IB_GID_INDEX_GREP} GID_INDEX found on any\ [e2e-llm-inference-service] \ HCA.\"\n fi\n else\n echo \"[Infer RoCE] No active HCAs found,\ [e2e-llm-inference-service] \ skipping GID_INDEX inference.\"\n fi\nfi\n\n# --disable-access-log-for-endpoints\ [e2e-llm-inference-service] \ landed in vLLM 0.16.0 (vllm-project/vllm#30011).\n# Older versions still\ [e2e-llm-inference-service] \ need the blanket --disable-uvicorn-access-log.\nACCESS_LOG_ARGS=\"--disable-uvicorn-access-log\"\ [e2e-llm-inference-service] \nVLLM_VERSION=$(vllm --version 2>/dev/null | tail -1 | awk '{print $NF}')\n\ [e2e-llm-inference-service] echo \"[access-log-detect] vllm version='${VLLM_VERSION}'\"\nif [[ \"$VLLM_VERSION\"\ [e2e-llm-inference-service] \ =~ ^[0-9]+\\.[0-9]+ ]] && [ \"$(printf '%s\\n%s\\n' \"0.16.0\" \"${VLLM_VERSION}\"\ [e2e-llm-inference-service] \ | sort -V | head -1)\" = \"0.16.0\" ]; then\n ACCESS_LOG_ARGS=\"--disable-access-log-for-endpoints\ [e2e-llm-inference-service] \ /health,/metrics,/ping\"\nfi\necho \"[access-log-detect] selected ACCESS_LOG_ARGS='${ACCESS_LOG_ARGS}'\"\ [e2e-llm-inference-service] \n\n# --shutdown-timeout landed in vLLM 0.18.0 (vllm-project/vllm#36666).\n\ [e2e-llm-inference-service] SHUTDOWN_TIMEOUT_ARGS=\"\"\nif [[ \"$VLLM_VERSION\" =~ ^[0-9]+\\.[0-9]+\ [e2e-llm-inference-service] \ ]] && [ \"$(printf '%s\\n%s\\n' \"0.18.0\" \"${VLLM_VERSION}\" | sort\ [e2e-llm-inference-service] \ -V | head -1)\" = \"0.18.0\" ]; then\n SHUTDOWN_TIMEOUT_ARGS=\"--shutdown-timeout\ [e2e-llm-inference-service] \ 40\"\nfi\n\neval \"exec vllm serve /mnt/models \\\n --served-model-name\ [e2e-llm-inference-service] \ \"facebook/opt-125m\" \"publishers/kserve-ci-e2e-test/models/facebook/opt-125m\"\ [e2e-llm-inference-service] \ \\\n --port 8000 \\\n ${ACCESS_LOG_ARGS} \\\n ${SHUTDOWN_TIMEOUT_ARGS}\ [e2e-llm-inference-service] \ \\\n --enable-ssl-refresh \\\n --ssl-certfile /var/run/kserve/tls/tls.crt\ [e2e-llm-inference-service] \ \\\n --ssl-keyfile /var/run/kserve/tls/tls.key \\\n ${VLLM_ADDITIONAL_ARGS}\ [e2e-llm-inference-service] \ \\\n $@\"" [e2e-llm-inference-service] - -- [e2e-llm-inference-service] ports: [e2e-llm-inference-service] - containerPort: 8000 [e2e-llm-inference-service] protocol: TCP [e2e-llm-inference-service] env: [e2e-llm-inference-service] - name: HOME [e2e-llm-inference-service] value: /home [e2e-llm-inference-service] - name: VLLM_LOGGING_LEVEL [e2e-llm-inference-service] value: DEBUG [e2e-llm-inference-service] - name: VLLM_CPU_KVCACHE_SPACE [e2e-llm-inference-service] value: '1' [e2e-llm-inference-service] - name: VLLM_ENABLE_V1_MULTIPROCESSING [e2e-llm-inference-service] value: '0' [e2e-llm-inference-service] - name: USER [e2e-llm-inference-service] value: nonroot [e2e-llm-inference-service] - name: TORCHINDUCTOR_CACHE_DIR [e2e-llm-inference-service] value: /tmp/torchinductor-cache [e2e-llm-inference-service] - name: HF_HUB_CACHE [e2e-llm-inference-service] value: /models [e2e-llm-inference-service] resources: [e2e-llm-inference-service] limits: [e2e-llm-inference-service] cpu: '2' [e2e-llm-inference-service] memory: 7Gi [e2e-llm-inference-service] requests: [e2e-llm-inference-service] cpu: 200m [e2e-llm-inference-service] memory: 2Gi [e2e-llm-inference-service] volumeMounts: [e2e-llm-inference-service] - name: home [e2e-llm-inference-service] mountPath: /home [e2e-llm-inference-service] - name: tmp-dir [e2e-llm-inference-service] mountPath: /tmp [e2e-llm-inference-service] - name: dshm [e2e-llm-inference-service] mountPath: /dev/shm [e2e-llm-inference-service] - name: model-cache [e2e-llm-inference-service] mountPath: /models [e2e-llm-inference-service] - name: tls-certs [e2e-llm-inference-service] readOnly: true [e2e-llm-inference-service] mountPath: /var/run/kserve/tls [e2e-llm-inference-service] - name: kserve-provision-location [e2e-llm-inference-service] readOnly: true [e2e-llm-inference-service] mountPath: /mnt/models [e2e-llm-inference-service] livenessProbe: [e2e-llm-inference-service] httpGet: [e2e-llm-inference-service] path: /health [e2e-llm-inference-service] port: 8000 [e2e-llm-inference-service] scheme: HTTPS [e2e-llm-inference-service] timeoutSeconds: 1 [e2e-llm-inference-service] periodSeconds: 10 [e2e-llm-inference-service] successThreshold: 1 [e2e-llm-inference-service] failureThreshold: 10 [e2e-llm-inference-service] readinessProbe: [e2e-llm-inference-service] httpGet: [e2e-llm-inference-service] path: /health [e2e-llm-inference-service] port: 8000 [e2e-llm-inference-service] scheme: HTTPS [e2e-llm-inference-service] timeoutSeconds: 1 [e2e-llm-inference-service] periodSeconds: 1 [e2e-llm-inference-service] successThreshold: 1 [e2e-llm-inference-service] failureThreshold: 2 [e2e-llm-inference-service] startupProbe: [e2e-llm-inference-service] httpGet: [e2e-llm-inference-service] path: /health [e2e-llm-inference-service] port: 8000 [e2e-llm-inference-service] scheme: HTTPS [e2e-llm-inference-service] timeoutSeconds: 1 [e2e-llm-inference-service] periodSeconds: 10 [e2e-llm-inference-service] successThreshold: 1 [e2e-llm-inference-service] failureThreshold: 60 [e2e-llm-inference-service] lifecycle: [e2e-llm-inference-service] preStop: [e2e-llm-inference-service] exec: [e2e-llm-inference-service] command: [e2e-llm-inference-service] - /bin/sleep [e2e-llm-inference-service] - '15' [e2e-llm-inference-service] terminationMessagePath: /dev/termination-log [e2e-llm-inference-service] terminationMessagePolicy: FallbackToLogsOnError [e2e-llm-inference-service] imagePullPolicy: IfNotPresent [e2e-llm-inference-service] securityContext: [e2e-llm-inference-service] capabilities: [e2e-llm-inference-service] drop: [e2e-llm-inference-service] - ALL [e2e-llm-inference-service] runAsNonRoot: true [e2e-llm-inference-service] readOnlyRootFilesystem: true [e2e-llm-inference-service] allowPrivilegeEscalation: false [e2e-llm-inference-service] seccompProfile: [e2e-llm-inference-service] type: RuntimeDefault [e2e-llm-inference-service] restartPolicy: Always [e2e-llm-inference-service] terminationGracePeriodSeconds: 60 [e2e-llm-inference-service] dnsPolicy: ClusterFirst [e2e-llm-inference-service] securityContext: {} [e2e-llm-inference-service] schedulerName: default-scheduler [e2e-llm-inference-service] status: [e2e-llm-inference-service] replicas: 1 [e2e-llm-inference-service] fullyLabeledReplicas: 1 [e2e-llm-inference-service] readyReplicas: 1 [e2e-llm-inference-service] availableReplicas: 1 [e2e-llm-inference-service] observedGeneration: 1 [e2e-llm-inference-service] apiVersion: apps/v1 [e2e-llm-inference-service] kind: ReplicaSet [e2e-llm-inference-service] [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1254 --- [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1255 metadata: [e2e-llm-inference-service] name: llmisvca690bbc929faec8bc98c767f16c003c1-kserve-router-scheduler-5d6998448f [e2e-llm-inference-service] namespace: kserve-ci-e2e-test [e2e-llm-inference-service] uid: b08110a1-5c9b-4623-b862-05a01c144c6a [e2e-llm-inference-service] resourceVersion: '31668' [e2e-llm-inference-service] generation: 1 [e2e-llm-inference-service] creationTimestamp: '2026-07-07T14:31:40Z' [e2e-llm-inference-service] labels: [e2e-llm-inference-service] app.kubernetes.io/component: llminferenceservice-router-scheduler [e2e-llm-inference-service] app.kubernetes.io/name: llmisvc-model-fb-opt-125m-route-87882a8e [e2e-llm-inference-service] app.kubernetes.io/part-of: llminferenceservice [e2e-llm-inference-service] pod-template-hash: 5d6998448f [e2e-llm-inference-service] annotations: [e2e-llm-inference-service] deployment.kubernetes.io/desired-replicas: '1' [e2e-llm-inference-service] deployment.kubernetes.io/max-replicas: '1' [e2e-llm-inference-service] deployment.kubernetes.io/revision: '1' [e2e-llm-inference-service] ownerReferences: [e2e-llm-inference-service] - apiVersion: apps/v1 [e2e-llm-inference-service] kind: Deployment [e2e-llm-inference-service] name: llmisvca690bbc929faec8bc98c767f16c003c1-kserve-router-scheduler [e2e-llm-inference-service] uid: 3c03b0dd-9720-4679-b828-aaa0b86d2325 [e2e-llm-inference-service] controller: true [e2e-llm-inference-service] blockOwnerDeletion: true [e2e-llm-inference-service] managedFields: [e2e-llm-inference-service] - manager: kube-controller-manager [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] apiVersion: apps/v1 [e2e-llm-inference-service] time: '2026-07-07T14:31:40Z' [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:metadata: [e2e-llm-inference-service] f:annotations: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:deployment.kubernetes.io/desired-replicas: {} [e2e-llm-inference-service] f:deployment.kubernetes.io/max-replicas: {} [e2e-llm-inference-service] f:deployment.kubernetes.io/revision: {} [e2e-llm-inference-service] f:labels: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:app.kubernetes.io/component: {} [e2e-llm-inference-service] f:app.kubernetes.io/name: {} [e2e-llm-inference-service] f:app.kubernetes.io/part-of: {} [e2e-llm-inference-service] f:pod-template-hash: {} [e2e-llm-inference-service] f:ownerReferences: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"uid":"3c03b0dd-9720-4679-b828-aaa0b86d2325"}: {} [e2e-llm-inference-service] f:spec: [e2e-llm-inference-service] f:replicas: {} [e2e-llm-inference-service] f:selector: {} [e2e-llm-inference-service] f:template: [e2e-llm-inference-service] f:metadata: [e2e-llm-inference-service] f:annotations: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:app.kubernetes.io/version: {} [e2e-llm-inference-service] f:certificates.kserve.io/expiration-v2: {} [e2e-llm-inference-service] f:labels: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:app.kubernetes.io/component: {} [e2e-llm-inference-service] f:app.kubernetes.io/name: {} [e2e-llm-inference-service] f:app.kubernetes.io/part-of: {} [e2e-llm-inference-service] f:pod-template-hash: {} [e2e-llm-inference-service] f:spec: [e2e-llm-inference-service] f:containers: [e2e-llm-inference-service] k:{"name":"main"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:args: {} [e2e-llm-inference-service] f:command: {} [e2e-llm-inference-service] f:env: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"name":"SSL_CERT_DIR"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] f:image: {} [e2e-llm-inference-service] f:imagePullPolicy: {} [e2e-llm-inference-service] f:lifecycle: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:preStop: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:exec: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:command: {} [e2e-llm-inference-service] f:livenessProbe: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:failureThreshold: {} [e2e-llm-inference-service] f:grpc: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:port: {} [e2e-llm-inference-service] f:service: {} [e2e-llm-inference-service] f:initialDelaySeconds: {} [e2e-llm-inference-service] f:periodSeconds: {} [e2e-llm-inference-service] f:successThreshold: {} [e2e-llm-inference-service] f:timeoutSeconds: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:ports: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"containerPort":5557,"protocol":"TCP"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:containerPort: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:protocol: {} [e2e-llm-inference-service] k:{"containerPort":9002,"protocol":"TCP"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:containerPort: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:protocol: {} [e2e-llm-inference-service] k:{"containerPort":9003,"protocol":"TCP"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:containerPort: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:protocol: {} [e2e-llm-inference-service] k:{"containerPort":9090,"protocol":"TCP"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:containerPort: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:protocol: {} [e2e-llm-inference-service] f:readinessProbe: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:failureThreshold: {} [e2e-llm-inference-service] f:grpc: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:port: {} [e2e-llm-inference-service] f:service: {} [e2e-llm-inference-service] f:initialDelaySeconds: {} [e2e-llm-inference-service] f:periodSeconds: {} [e2e-llm-inference-service] f:successThreshold: {} [e2e-llm-inference-service] f:timeoutSeconds: {} [e2e-llm-inference-service] f:resources: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:requests: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:cpu: {} [e2e-llm-inference-service] f:memory: {} [e2e-llm-inference-service] f:securityContext: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:allowPrivilegeEscalation: {} [e2e-llm-inference-service] f:capabilities: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:drop: {} [e2e-llm-inference-service] f:readOnlyRootFilesystem: {} [e2e-llm-inference-service] f:runAsNonRoot: {} [e2e-llm-inference-service] f:seccompProfile: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:type: {} [e2e-llm-inference-service] f:terminationMessagePath: {} [e2e-llm-inference-service] f:terminationMessagePolicy: {} [e2e-llm-inference-service] f:volumeMounts: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"mountPath":"/tmp/tokenizer"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:mountPath: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] k:{"mountPath":"/var/run/kserve/tls"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:mountPath: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:readOnly: {} [e2e-llm-inference-service] f:dnsPolicy: {} [e2e-llm-inference-service] f:initContainers: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"name":"storage-initializer"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:args: {} [e2e-llm-inference-service] f:env: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"name":"AWS_ACCESS_KEY_ID"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:valueFrom: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:secretKeyRef: {} [e2e-llm-inference-service] k:{"name":"AWS_CA_BUNDLE"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"AWS_CA_BUNDLE_CONFIGMAP"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"AWS_ENDPOINT_URL"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"AWS_SECRET_ACCESS_KEY"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:valueFrom: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:secretKeyRef: {} [e2e-llm-inference-service] k:{"name":"HF_HUB_ENABLE_HF_TRANSFER"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"HF_XET_HIGH_PERFORMANCE"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"HF_XET_NUM_CONCURRENT_RANGE_GETS"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"S3_ENDPOINT"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"S3_USE_HTTPS"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"S3_VERIFY_SSL"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"STORAGE_ALLOW_PATTERNS"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] f:image: {} [e2e-llm-inference-service] f:imagePullPolicy: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:resources: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:limits: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:cpu: {} [e2e-llm-inference-service] f:memory: {} [e2e-llm-inference-service] f:requests: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:cpu: {} [e2e-llm-inference-service] f:memory: {} [e2e-llm-inference-service] f:terminationMessagePath: {} [e2e-llm-inference-service] f:terminationMessagePolicy: {} [e2e-llm-inference-service] f:volumeMounts: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"mountPath":"/mnt/models"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:mountPath: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:restartPolicy: {} [e2e-llm-inference-service] f:schedulerName: {} [e2e-llm-inference-service] f:securityContext: {} [e2e-llm-inference-service] f:serviceAccount: {} [e2e-llm-inference-service] f:serviceAccountName: {} [e2e-llm-inference-service] f:terminationGracePeriodSeconds: {} [e2e-llm-inference-service] f:volumes: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"name":"kserve-provision-location"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:emptyDir: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] k:{"name":"tls-certs"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:secret: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:defaultMode: {} [e2e-llm-inference-service] f:secretName: {} [e2e-llm-inference-service] k:{"name":"tokenizer-cache"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:emptyDir: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] k:{"name":"tokenizer-tmp"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:emptyDir: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] k:{"name":"tokenizer-uds"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:emptyDir: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] - manager: kube-controller-manager [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] apiVersion: apps/v1 [e2e-llm-inference-service] time: '2026-07-07T14:32:14Z' [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:status: [e2e-llm-inference-service] f:availableReplicas: {} [e2e-llm-inference-service] f:fullyLabeledReplicas: {} [e2e-llm-inference-service] f:observedGeneration: {} [e2e-llm-inference-service] f:readyReplicas: {} [e2e-llm-inference-service] f:replicas: {} [e2e-llm-inference-service] subresource: status [e2e-llm-inference-service] spec: [e2e-llm-inference-service] replicas: 1 [e2e-llm-inference-service] selector: [e2e-llm-inference-service] matchLabels: [e2e-llm-inference-service] app.kubernetes.io/component: llminferenceservice-router-scheduler [e2e-llm-inference-service] app.kubernetes.io/name: llmisvc-model-fb-opt-125m-route-87882a8e [e2e-llm-inference-service] app.kubernetes.io/part-of: llminferenceservice [e2e-llm-inference-service] pod-template-hash: 5d6998448f [e2e-llm-inference-service] template: [e2e-llm-inference-service] metadata: [e2e-llm-inference-service] labels: [e2e-llm-inference-service] app.kubernetes.io/component: llminferenceservice-router-scheduler [e2e-llm-inference-service] app.kubernetes.io/name: llmisvc-model-fb-opt-125m-route-87882a8e [e2e-llm-inference-service] app.kubernetes.io/part-of: llminferenceservice [e2e-llm-inference-service] pod-template-hash: 5d6998448f [e2e-llm-inference-service] annotations: [e2e-llm-inference-service] app.kubernetes.io/version: 0.9.0 [e2e-llm-inference-service] certificates.kserve.io/expiration-v2: 'true' [e2e-llm-inference-service] spec: [e2e-llm-inference-service] volumes: [e2e-llm-inference-service] - name: tls-certs [e2e-llm-inference-service] secret: [e2e-llm-inference-service] secretName: llmisva690bbc929faec8bc98c767f16c003c1-kserve-self-signed-certs [e2e-llm-inference-service] defaultMode: 420 [e2e-llm-inference-service] - name: tokenizer-uds [e2e-llm-inference-service] emptyDir: {} [e2e-llm-inference-service] - name: tokenizer-tmp [e2e-llm-inference-service] emptyDir: {} [e2e-llm-inference-service] - name: tokenizer-cache [e2e-llm-inference-service] emptyDir: {} [e2e-llm-inference-service] - name: kserve-provision-location [e2e-llm-inference-service] emptyDir: {} [e2e-llm-inference-service] initContainers: [e2e-llm-inference-service] - name: storage-initializer [e2e-llm-inference-service] image: quay.io/opendatahub/kserve-storage-initializer@sha256:13ad765c0f323200e1db7b45c99661287e65c5158d6b54b16139829636d20be4 [e2e-llm-inference-service] args: [e2e-llm-inference-service] - hf://facebook/opt-125m [e2e-llm-inference-service] - /mnt/models [e2e-llm-inference-service] env: [e2e-llm-inference-service] - name: AWS_ACCESS_KEY_ID [e2e-llm-inference-service] valueFrom: [e2e-llm-inference-service] secretKeyRef: [e2e-llm-inference-service] name: seaweedfs-s3-creds [e2e-llm-inference-service] key: AWS_ACCESS_KEY_ID [e2e-llm-inference-service] - name: AWS_SECRET_ACCESS_KEY [e2e-llm-inference-service] valueFrom: [e2e-llm-inference-service] secretKeyRef: [e2e-llm-inference-service] name: seaweedfs-s3-creds [e2e-llm-inference-service] key: AWS_SECRET_ACCESS_KEY [e2e-llm-inference-service] - name: S3_USE_HTTPS [e2e-llm-inference-service] value: '0' [e2e-llm-inference-service] - name: S3_ENDPOINT [e2e-llm-inference-service] value: s3-service.kserve:8333 [e2e-llm-inference-service] - name: AWS_ENDPOINT_URL [e2e-llm-inference-service] value: http://s3-service.kserve:8333 [e2e-llm-inference-service] - name: S3_VERIFY_SSL [e2e-llm-inference-service] value: '0' [e2e-llm-inference-service] - name: AWS_CA_BUNDLE [e2e-llm-inference-service] value: /etc/ssl/custom-certs/cabundle.crt [e2e-llm-inference-service] - name: AWS_CA_BUNDLE_CONFIGMAP [e2e-llm-inference-service] value: odh-kserve-custom-ca-bundle [e2e-llm-inference-service] - name: HF_HUB_ENABLE_HF_TRANSFER [e2e-llm-inference-service] value: '1' [e2e-llm-inference-service] - name: HF_XET_HIGH_PERFORMANCE [e2e-llm-inference-service] value: '1' [e2e-llm-inference-service] - name: HF_XET_NUM_CONCURRENT_RANGE_GETS [e2e-llm-inference-service] value: '8' [e2e-llm-inference-service] - name: STORAGE_ALLOW_PATTERNS [e2e-llm-inference-service] value: '["tokenizer.json", "tokenizer_config.json", "special_tokens_map.json", [e2e-llm-inference-service] "vocab.json", "merges.txt", "config.json", "generation_config.json"]' [e2e-llm-inference-service] resources: [e2e-llm-inference-service] limits: [e2e-llm-inference-service] cpu: '1' [e2e-llm-inference-service] memory: 24Gi [e2e-llm-inference-service] requests: [e2e-llm-inference-service] cpu: 100m [e2e-llm-inference-service] memory: 100Mi [e2e-llm-inference-service] volumeMounts: [e2e-llm-inference-service] - name: kserve-provision-location [e2e-llm-inference-service] mountPath: /mnt/models [e2e-llm-inference-service] terminationMessagePath: /dev/termination-log [e2e-llm-inference-service] terminationMessagePolicy: FallbackToLogsOnError [e2e-llm-inference-service] imagePullPolicy: IfNotPresent [e2e-llm-inference-service] containers: [e2e-llm-inference-service] - name: main [e2e-llm-inference-service] image: ghcr.io/llm-d/llm-d-router-endpoint-picker:v0.9.0-rc.2 [e2e-llm-inference-service] command: [e2e-llm-inference-service] - /app/epp [e2e-llm-inference-service] - --pool-name [e2e-llm-inference-service] - llmisvc-model-fb-opt-125m-route-87882a8e-inference-pool [e2e-llm-inference-service] - --pool-namespace [e2e-llm-inference-service] - kserve-ci-e2e-test [e2e-llm-inference-service] - --zap-encoder [e2e-llm-inference-service] - json [e2e-llm-inference-service] - --grpc-port [e2e-llm-inference-service] - '9002' [e2e-llm-inference-service] - --grpc-health-port [e2e-llm-inference-service] - '9003' [e2e-llm-inference-service] - --enable-cert-reload=true [e2e-llm-inference-service] - --secure-serving=true [e2e-llm-inference-service] - --model-server-metrics-scheme=https [e2e-llm-inference-service] - --cert-path=/var/run/kserve/tls [e2e-llm-inference-service] args: [e2e-llm-inference-service] - --config-text [e2e-llm-inference-service] - "apiVersion: inference.networking.x-k8s.io/v1alpha1\nkind: EndpointPickerConfig\n\ [e2e-llm-inference-service] plugins:\n- type: single-profile-handler\n- type: queue-scorer\n- type:\ [e2e-llm-inference-service] \ prefix-cache-scorer\n- type: max-score-picker\nschedulingProfiles:\n-\ [e2e-llm-inference-service] \ name: default\n plugins:\n - pluginRef: queue-scorer\n weight: 2\n\ [e2e-llm-inference-service] \ - pluginRef: prefix-cache-scorer\n weight: 3\n - pluginRef: max-score-picker\n" [e2e-llm-inference-service] ports: [e2e-llm-inference-service] - name: grpc [e2e-llm-inference-service] containerPort: 9002 [e2e-llm-inference-service] protocol: TCP [e2e-llm-inference-service] - name: grpc-health [e2e-llm-inference-service] containerPort: 9003 [e2e-llm-inference-service] protocol: TCP [e2e-llm-inference-service] - name: metrics [e2e-llm-inference-service] containerPort: 9090 [e2e-llm-inference-service] protocol: TCP [e2e-llm-inference-service] - name: zmq [e2e-llm-inference-service] containerPort: 5557 [e2e-llm-inference-service] protocol: TCP [e2e-llm-inference-service] env: [e2e-llm-inference-service] - name: SSL_CERT_DIR [e2e-llm-inference-service] value: /var/run/kserve/tls:/var/run/secrets/kubernetes.io/serviceaccount:/etc/pki/tls/certs [e2e-llm-inference-service] resources: [e2e-llm-inference-service] requests: [e2e-llm-inference-service] cpu: 256m [e2e-llm-inference-service] memory: 500Mi [e2e-llm-inference-service] volumeMounts: [e2e-llm-inference-service] - name: tls-certs [e2e-llm-inference-service] readOnly: true [e2e-llm-inference-service] mountPath: /var/run/kserve/tls [e2e-llm-inference-service] - name: tokenizer-uds [e2e-llm-inference-service] mountPath: /tmp/tokenizer [e2e-llm-inference-service] livenessProbe: [e2e-llm-inference-service] grpc: [e2e-llm-inference-service] port: 9003 [e2e-llm-inference-service] service: liveness [e2e-llm-inference-service] initialDelaySeconds: 5 [e2e-llm-inference-service] timeoutSeconds: 1 [e2e-llm-inference-service] periodSeconds: 10 [e2e-llm-inference-service] successThreshold: 1 [e2e-llm-inference-service] failureThreshold: 3 [e2e-llm-inference-service] readinessProbe: [e2e-llm-inference-service] grpc: [e2e-llm-inference-service] port: 9003 [e2e-llm-inference-service] service: readiness [e2e-llm-inference-service] initialDelaySeconds: 30 [e2e-llm-inference-service] timeoutSeconds: 1 [e2e-llm-inference-service] periodSeconds: 10 [e2e-llm-inference-service] successThreshold: 1 [e2e-llm-inference-service] failureThreshold: 3 [e2e-llm-inference-service] lifecycle: [e2e-llm-inference-service] preStop: [e2e-llm-inference-service] exec: [e2e-llm-inference-service] command: [e2e-llm-inference-service] - /bin/sleep [e2e-llm-inference-service] - '15' [e2e-llm-inference-service] terminationMessagePath: /dev/termination-log [e2e-llm-inference-service] terminationMessagePolicy: FallbackToLogsOnError [e2e-llm-inference-service] imagePullPolicy: IfNotPresent [e2e-llm-inference-service] securityContext: [e2e-llm-inference-service] capabilities: [e2e-llm-inference-service] drop: [e2e-llm-inference-service] - ALL [e2e-llm-inference-service] runAsNonRoot: true [e2e-llm-inference-service] readOnlyRootFilesystem: true [e2e-llm-inference-service] allowPrivilegeEscalation: false [e2e-llm-inference-service] seccompProfile: [e2e-llm-inference-service] type: RuntimeDefault [e2e-llm-inference-service] restartPolicy: Always [e2e-llm-inference-service] terminationGracePeriodSeconds: 60 [e2e-llm-inference-service] dnsPolicy: ClusterFirst [e2e-llm-inference-service] serviceAccountName: llmisvc-model-fb-opt-125m-route-87882a8e-epp-sa [e2e-llm-inference-service] serviceAccount: llmisvc-model-fb-opt-125m-route-87882a8e-epp-sa [e2e-llm-inference-service] securityContext: {} [e2e-llm-inference-service] schedulerName: default-scheduler [e2e-llm-inference-service] status: [e2e-llm-inference-service] replicas: 1 [e2e-llm-inference-service] fullyLabeledReplicas: 1 [e2e-llm-inference-service] readyReplicas: 1 [e2e-llm-inference-service] availableReplicas: 1 [e2e-llm-inference-service] observedGeneration: 1 [e2e-llm-inference-service] apiVersion: apps/v1 [e2e-llm-inference-service] kind: ReplicaSet [e2e-llm-inference-service] [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1254 --- [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1255 metadata: [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-route-87882a8e-epp-rb [e2e-llm-inference-service] namespace: kserve-ci-e2e-test [e2e-llm-inference-service] uid: 654c00c6-cdf3-4f63-9614-3ab3b01c2cd6 [e2e-llm-inference-service] resourceVersion: '30988' [e2e-llm-inference-service] creationTimestamp: '2026-07-07T14:31:40Z' [e2e-llm-inference-service] labels: [e2e-llm-inference-service] app.kubernetes.io/component: llminferenceservice-router-scheduler [e2e-llm-inference-service] app.kubernetes.io/name: llmisvc-model-fb-opt-125m-route-87882a8e [e2e-llm-inference-service] app.kubernetes.io/part-of: llminferenceservice [e2e-llm-inference-service] ownerReferences: [e2e-llm-inference-service] - apiVersion: serving.kserve.io/v1alpha2 [e2e-llm-inference-service] kind: LLMInferenceService [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-route-87882a8e [e2e-llm-inference-service] uid: 88c60d0e-95bd-401f-ae1c-2c4c0b1e7c9d [e2e-llm-inference-service] controller: true [e2e-llm-inference-service] blockOwnerDeletion: true [e2e-llm-inference-service] managedFields: [e2e-llm-inference-service] - manager: manager [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] apiVersion: rbac.authorization.k8s.io/v1 [e2e-llm-inference-service] time: '2026-07-07T14:31:40Z' [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:metadata: [e2e-llm-inference-service] f:labels: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:app.kubernetes.io/component: {} [e2e-llm-inference-service] f:app.kubernetes.io/name: {} [e2e-llm-inference-service] f:app.kubernetes.io/part-of: {} [e2e-llm-inference-service] f:ownerReferences: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"uid":"88c60d0e-95bd-401f-ae1c-2c4c0b1e7c9d"}: {} [e2e-llm-inference-service] f:roleRef: {} [e2e-llm-inference-service] f:subjects: {} [e2e-llm-inference-service] subjects: [e2e-llm-inference-service] - kind: ServiceAccount [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-route-87882a8e-epp-sa [e2e-llm-inference-service] namespace: kserve-ci-e2e-test [e2e-llm-inference-service] roleRef: [e2e-llm-inference-service] apiGroup: rbac.authorization.k8s.io [e2e-llm-inference-service] kind: Role [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-route-87882a8e-epp-role [e2e-llm-inference-service] apiVersion: rbac.authorization.k8s.io/v1 [e2e-llm-inference-service] kind: RoleBinding [e2e-llm-inference-service] [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1254 --- [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1255 metadata: [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-route-87882a8e-epp-role [e2e-llm-inference-service] namespace: kserve-ci-e2e-test [e2e-llm-inference-service] uid: dd4dffa9-097a-4938-a613-8cfc6e17c1f1 [e2e-llm-inference-service] resourceVersion: '30984' [e2e-llm-inference-service] creationTimestamp: '2026-07-07T14:31:40Z' [e2e-llm-inference-service] labels: [e2e-llm-inference-service] app.kubernetes.io/component: llminferenceservice-router-scheduler [e2e-llm-inference-service] app.kubernetes.io/name: llmisvc-model-fb-opt-125m-route-87882a8e [e2e-llm-inference-service] app.kubernetes.io/part-of: llminferenceservice [e2e-llm-inference-service] ownerReferences: [e2e-llm-inference-service] - apiVersion: serving.kserve.io/v1alpha2 [e2e-llm-inference-service] kind: LLMInferenceService [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-route-87882a8e [e2e-llm-inference-service] uid: 88c60d0e-95bd-401f-ae1c-2c4c0b1e7c9d [e2e-llm-inference-service] controller: true [e2e-llm-inference-service] blockOwnerDeletion: true [e2e-llm-inference-service] managedFields: [e2e-llm-inference-service] - manager: manager [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] apiVersion: rbac.authorization.k8s.io/v1 [e2e-llm-inference-service] time: '2026-07-07T14:31:40Z' [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:metadata: [e2e-llm-inference-service] f:labels: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:app.kubernetes.io/component: {} [e2e-llm-inference-service] f:app.kubernetes.io/name: {} [e2e-llm-inference-service] f:app.kubernetes.io/part-of: {} [e2e-llm-inference-service] f:ownerReferences: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"uid":"88c60d0e-95bd-401f-ae1c-2c4c0b1e7c9d"}: {} [e2e-llm-inference-service] f:rules: {} [e2e-llm-inference-service] rules: [e2e-llm-inference-service] - verbs: [e2e-llm-inference-service] - get [e2e-llm-inference-service] - list [e2e-llm-inference-service] - watch [e2e-llm-inference-service] apiGroups: [e2e-llm-inference-service] - '' [e2e-llm-inference-service] resources: [e2e-llm-inference-service] - pods [e2e-llm-inference-service] - verbs: [e2e-llm-inference-service] - get [e2e-llm-inference-service] - list [e2e-llm-inference-service] - watch [e2e-llm-inference-service] apiGroups: [e2e-llm-inference-service] - inference.networking.k8s.io [e2e-llm-inference-service] - inference.networking.x-k8s.io [e2e-llm-inference-service] resources: [e2e-llm-inference-service] - inferencepools [e2e-llm-inference-service] - inferenceobjectives [e2e-llm-inference-service] - inferencemodels [e2e-llm-inference-service] - verbs: [e2e-llm-inference-service] - get [e2e-llm-inference-service] - list [e2e-llm-inference-service] - watch [e2e-llm-inference-service] apiGroups: [e2e-llm-inference-service] - inference.networking.x-k8s.io [e2e-llm-inference-service] resources: [e2e-llm-inference-service] - inferencemodelrewrites [e2e-llm-inference-service] - inferencepoolimports [e2e-llm-inference-service] - verbs: [e2e-llm-inference-service] - get [e2e-llm-inference-service] - list [e2e-llm-inference-service] - watch [e2e-llm-inference-service] apiGroups: [e2e-llm-inference-service] - discovery.k8s.io [e2e-llm-inference-service] resources: [e2e-llm-inference-service] - endpointslices [e2e-llm-inference-service] - verbs: [e2e-llm-inference-service] - get [e2e-llm-inference-service] - list [e2e-llm-inference-service] - watch [e2e-llm-inference-service] - create [e2e-llm-inference-service] - update [e2e-llm-inference-service] - patch [e2e-llm-inference-service] - delete [e2e-llm-inference-service] apiGroups: [e2e-llm-inference-service] - coordination.k8s.io [e2e-llm-inference-service] resources: [e2e-llm-inference-service] - leases [e2e-llm-inference-service] apiVersion: rbac.authorization.k8s.io/v1 [e2e-llm-inference-service] kind: Role [e2e-llm-inference-service] [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1254 --- [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1255 metadata: [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-route-87882a8e-epp-service-t5h4g [e2e-llm-inference-service] generateName: llmisvc-model-fb-opt-125m-route-87882a8e-epp-service- [e2e-llm-inference-service] namespace: kserve-ci-e2e-test [e2e-llm-inference-service] uid: 1bc89184-0294-47b6-a366-434686d89869 [e2e-llm-inference-service] resourceVersion: '31667' [e2e-llm-inference-service] generation: 3 [e2e-llm-inference-service] creationTimestamp: '2026-07-07T14:31:40Z' [e2e-llm-inference-service] labels: [e2e-llm-inference-service] app.kubernetes.io/component: llminferenceservice-router-scheduler [e2e-llm-inference-service] app.kubernetes.io/name: llmisvc-model-fb-opt-125m-route-87882a8e [e2e-llm-inference-service] app.kubernetes.io/part-of: llminferenceservice [e2e-llm-inference-service] endpointslice.kubernetes.io/managed-by: endpointslice-controller.k8s.io [e2e-llm-inference-service] kubernetes.io/service-name: llmisvc-model-fb-opt-125m-route-87882a8e-epp-service [e2e-llm-inference-service] annotations: [e2e-llm-inference-service] endpoints.kubernetes.io/last-change-trigger-time: '2026-07-07T14:32:14Z' [e2e-llm-inference-service] ownerReferences: [e2e-llm-inference-service] - apiVersion: v1 [e2e-llm-inference-service] kind: Service [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-route-87882a8e-epp-service [e2e-llm-inference-service] uid: c8576b5b-94c8-4bab-b493-e867f77d1f8d [e2e-llm-inference-service] controller: true [e2e-llm-inference-service] blockOwnerDeletion: true [e2e-llm-inference-service] managedFields: [e2e-llm-inference-service] - manager: kube-controller-manager [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] apiVersion: discovery.k8s.io/v1 [e2e-llm-inference-service] time: '2026-07-07T14:32:14Z' [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:addressType: {} [e2e-llm-inference-service] f:endpoints: {} [e2e-llm-inference-service] f:metadata: [e2e-llm-inference-service] f:annotations: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:endpoints.kubernetes.io/last-change-trigger-time: {} [e2e-llm-inference-service] f:generateName: {} [e2e-llm-inference-service] f:labels: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:app.kubernetes.io/component: {} [e2e-llm-inference-service] f:app.kubernetes.io/name: {} [e2e-llm-inference-service] f:app.kubernetes.io/part-of: {} [e2e-llm-inference-service] f:endpointslice.kubernetes.io/managed-by: {} [e2e-llm-inference-service] f:kubernetes.io/service-name: {} [e2e-llm-inference-service] f:ownerReferences: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"uid":"c8576b5b-94c8-4bab-b493-e867f77d1f8d"}: {} [e2e-llm-inference-service] f:ports: {} [e2e-llm-inference-service] addressType: IPv4 [e2e-llm-inference-service] endpoints: [e2e-llm-inference-service] - addresses: [e2e-llm-inference-service] - 10.132.0.53 [e2e-llm-inference-service] conditions: [e2e-llm-inference-service] ready: true [e2e-llm-inference-service] serving: true [e2e-llm-inference-service] terminating: false [e2e-llm-inference-service] targetRef: [e2e-llm-inference-service] kind: Pod [e2e-llm-inference-service] namespace: kserve-ci-e2e-test [e2e-llm-inference-service] name: llmisvca690bbc929faec8bc98c767f16c003c1-kserve-router-scheq44kv [e2e-llm-inference-service] uid: aa1b020a-833a-4de1-9ebe-0c421b524055 [e2e-llm-inference-service] nodeName: ip-10-0-133-25.ec2.internal [e2e-llm-inference-service] zone: us-east-1a [e2e-llm-inference-service] ports: [e2e-llm-inference-service] - name: grpc [e2e-llm-inference-service] protocol: TCP [e2e-llm-inference-service] port: 9002 [e2e-llm-inference-service] - name: grpc-health [e2e-llm-inference-service] protocol: TCP [e2e-llm-inference-service] port: 9003 [e2e-llm-inference-service] - name: metrics [e2e-llm-inference-service] protocol: TCP [e2e-llm-inference-service] port: 9090 [e2e-llm-inference-service] - name: zmq [e2e-llm-inference-service] protocol: TCP [e2e-llm-inference-service] port: 5557 [e2e-llm-inference-service] apiVersion: discovery.k8s.io/v1 [e2e-llm-inference-service] kind: EndpointSlice [e2e-llm-inference-service] [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1254 --- [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1255 metadata: [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-route-87882a8e-kserve-workload-stg2jv [e2e-llm-inference-service] generateName: llmisvc-model-fb-opt-125m-route-87882a8e-kserve-workload-svc- [e2e-llm-inference-service] namespace: kserve-ci-e2e-test [e2e-llm-inference-service] uid: 7de5e0f7-54fe-4077-8dac-4020dc9dc03c [e2e-llm-inference-service] resourceVersion: '33107' [e2e-llm-inference-service] generation: 3 [e2e-llm-inference-service] creationTimestamp: '2026-07-07T14:31:40Z' [e2e-llm-inference-service] labels: [e2e-llm-inference-service] app.kubernetes.io/component: llminferenceservice-workload [e2e-llm-inference-service] app.kubernetes.io/name: llmisvc-model-fb-opt-125m-route-87882a8e [e2e-llm-inference-service] app.kubernetes.io/part-of: llminferenceservice [e2e-llm-inference-service] endpointslice.kubernetes.io/managed-by: endpointslice-controller.k8s.io [e2e-llm-inference-service] kubernetes.io/service-name: llmisvc-model-fb-opt-125m-route-87882a8e-kserve-workload-svc [e2e-llm-inference-service] annotations: [e2e-llm-inference-service] endpoints.kubernetes.io/last-change-trigger-time: '2026-07-07T14:33:40Z' [e2e-llm-inference-service] ownerReferences: [e2e-llm-inference-service] - apiVersion: v1 [e2e-llm-inference-service] kind: Service [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-route-87882a8e-kserve-workload-svc [e2e-llm-inference-service] uid: f49fc6fa-967a-4564-b13b-50c3a33b3c52 [e2e-llm-inference-service] controller: true [e2e-llm-inference-service] blockOwnerDeletion: true [e2e-llm-inference-service] managedFields: [e2e-llm-inference-service] - manager: kube-controller-manager [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] apiVersion: discovery.k8s.io/v1 [e2e-llm-inference-service] time: '2026-07-07T14:33:40Z' [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:addressType: {} [e2e-llm-inference-service] f:endpoints: {} [e2e-llm-inference-service] f:metadata: [e2e-llm-inference-service] f:annotations: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:endpoints.kubernetes.io/last-change-trigger-time: {} [e2e-llm-inference-service] f:generateName: {} [e2e-llm-inference-service] f:labels: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:app.kubernetes.io/component: {} [e2e-llm-inference-service] f:app.kubernetes.io/name: {} [e2e-llm-inference-service] f:app.kubernetes.io/part-of: {} [e2e-llm-inference-service] f:endpointslice.kubernetes.io/managed-by: {} [e2e-llm-inference-service] f:kubernetes.io/service-name: {} [e2e-llm-inference-service] f:ownerReferences: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"uid":"f49fc6fa-967a-4564-b13b-50c3a33b3c52"}: {} [e2e-llm-inference-service] f:ports: {} [e2e-llm-inference-service] addressType: IPv4 [e2e-llm-inference-service] endpoints: [e2e-llm-inference-service] - addresses: [e2e-llm-inference-service] - 10.134.0.31 [e2e-llm-inference-service] conditions: [e2e-llm-inference-service] ready: true [e2e-llm-inference-service] serving: true [e2e-llm-inference-service] terminating: false [e2e-llm-inference-service] targetRef: [e2e-llm-inference-service] kind: Pod [e2e-llm-inference-service] namespace: kserve-ci-e2e-test [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-route-87882a8e-kserve-5bd7864c6ctnql9 [e2e-llm-inference-service] uid: 8abe33d5-7242-4f9f-a98e-f1814d77afd9 [e2e-llm-inference-service] nodeName: ip-10-0-139-47.ec2.internal [e2e-llm-inference-service] zone: us-east-1a [e2e-llm-inference-service] ports: [e2e-llm-inference-service] - name: https [e2e-llm-inference-service] protocol: TCP [e2e-llm-inference-service] port: 8000 [e2e-llm-inference-service] appProtocol: https [e2e-llm-inference-service] apiVersion: discovery.k8s.io/v1 [e2e-llm-inference-service] kind: EndpointSlice [e2e-llm-inference-service] [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1254 --- [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1255 metadata: [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-route-87882a8e-epp-rb [e2e-llm-inference-service] namespace: kserve-ci-e2e-test [e2e-llm-inference-service] uid: 654c00c6-cdf3-4f63-9614-3ab3b01c2cd6 [e2e-llm-inference-service] resourceVersion: '30988' [e2e-llm-inference-service] creationTimestamp: '2026-07-07T14:31:40Z' [e2e-llm-inference-service] labels: [e2e-llm-inference-service] app.kubernetes.io/component: llminferenceservice-router-scheduler [e2e-llm-inference-service] app.kubernetes.io/name: llmisvc-model-fb-opt-125m-route-87882a8e [e2e-llm-inference-service] app.kubernetes.io/part-of: llminferenceservice [e2e-llm-inference-service] ownerReferences: [e2e-llm-inference-service] - apiVersion: serving.kserve.io/v1alpha2 [e2e-llm-inference-service] kind: LLMInferenceService [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-route-87882a8e [e2e-llm-inference-service] uid: 88c60d0e-95bd-401f-ae1c-2c4c0b1e7c9d [e2e-llm-inference-service] controller: true [e2e-llm-inference-service] blockOwnerDeletion: true [e2e-llm-inference-service] managedFields: [e2e-llm-inference-service] - manager: manager [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] apiVersion: rbac.authorization.k8s.io/v1 [e2e-llm-inference-service] time: '2026-07-07T14:31:40Z' [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:metadata: [e2e-llm-inference-service] f:labels: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:app.kubernetes.io/component: {} [e2e-llm-inference-service] f:app.kubernetes.io/name: {} [e2e-llm-inference-service] f:app.kubernetes.io/part-of: {} [e2e-llm-inference-service] f:ownerReferences: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"uid":"88c60d0e-95bd-401f-ae1c-2c4c0b1e7c9d"}: {} [e2e-llm-inference-service] f:roleRef: {} [e2e-llm-inference-service] f:subjects: {} [e2e-llm-inference-service] userNames: [e2e-llm-inference-service] - system:serviceaccount:kserve-ci-e2e-test:llmisvc-model-fb-opt-125m-route-87882a8e-epp-sa [e2e-llm-inference-service] groupNames: null [e2e-llm-inference-service] subjects: [e2e-llm-inference-service] - kind: ServiceAccount [e2e-llm-inference-service] namespace: kserve-ci-e2e-test [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-route-87882a8e-epp-sa [e2e-llm-inference-service] roleRef: [e2e-llm-inference-service] namespace: kserve-ci-e2e-test [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-route-87882a8e-epp-role [e2e-llm-inference-service] apiVersion: authorization.openshift.io/v1 [e2e-llm-inference-service] kind: RoleBinding [e2e-llm-inference-service] [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1254 --- [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1255 metadata: [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-route-87882a8e-epp-role [e2e-llm-inference-service] namespace: kserve-ci-e2e-test [e2e-llm-inference-service] uid: dd4dffa9-097a-4938-a613-8cfc6e17c1f1 [e2e-llm-inference-service] resourceVersion: '30984' [e2e-llm-inference-service] creationTimestamp: '2026-07-07T14:31:40Z' [e2e-llm-inference-service] labels: [e2e-llm-inference-service] app.kubernetes.io/component: llminferenceservice-router-scheduler [e2e-llm-inference-service] app.kubernetes.io/name: llmisvc-model-fb-opt-125m-route-87882a8e [e2e-llm-inference-service] app.kubernetes.io/part-of: llminferenceservice [e2e-llm-inference-service] ownerReferences: [e2e-llm-inference-service] - apiVersion: serving.kserve.io/v1alpha2 [e2e-llm-inference-service] kind: LLMInferenceService [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-route-87882a8e [e2e-llm-inference-service] uid: 88c60d0e-95bd-401f-ae1c-2c4c0b1e7c9d [e2e-llm-inference-service] controller: true [e2e-llm-inference-service] blockOwnerDeletion: true [e2e-llm-inference-service] managedFields: [e2e-llm-inference-service] - manager: manager [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] apiVersion: rbac.authorization.k8s.io/v1 [e2e-llm-inference-service] time: '2026-07-07T14:31:40Z' [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:metadata: [e2e-llm-inference-service] f:labels: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:app.kubernetes.io/component: {} [e2e-llm-inference-service] f:app.kubernetes.io/name: {} [e2e-llm-inference-service] f:app.kubernetes.io/part-of: {} [e2e-llm-inference-service] f:ownerReferences: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"uid":"88c60d0e-95bd-401f-ae1c-2c4c0b1e7c9d"}: {} [e2e-llm-inference-service] f:rules: {} [e2e-llm-inference-service] rules: [e2e-llm-inference-service] - verbs: [e2e-llm-inference-service] - get [e2e-llm-inference-service] - list [e2e-llm-inference-service] - watch [e2e-llm-inference-service] attributeRestrictions: null [e2e-llm-inference-service] apiGroups: [e2e-llm-inference-service] - '' [e2e-llm-inference-service] resources: [e2e-llm-inference-service] - pods [e2e-llm-inference-service] - verbs: [e2e-llm-inference-service] - get [e2e-llm-inference-service] - list [e2e-llm-inference-service] - watch [e2e-llm-inference-service] attributeRestrictions: null [e2e-llm-inference-service] apiGroups: [e2e-llm-inference-service] - inference.networking.k8s.io [e2e-llm-inference-service] - inference.networking.x-k8s.io [e2e-llm-inference-service] resources: [e2e-llm-inference-service] - inferencemodels [e2e-llm-inference-service] - inferenceobjectives [e2e-llm-inference-service] - inferencepools [e2e-llm-inference-service] - verbs: [e2e-llm-inference-service] - get [e2e-llm-inference-service] - list [e2e-llm-inference-service] - watch [e2e-llm-inference-service] attributeRestrictions: null [e2e-llm-inference-service] apiGroups: [e2e-llm-inference-service] - inference.networking.x-k8s.io [e2e-llm-inference-service] resources: [e2e-llm-inference-service] - inferencemodelrewrites [e2e-llm-inference-service] - inferencepoolimports [e2e-llm-inference-service] - verbs: [e2e-llm-inference-service] - get [e2e-llm-inference-service] - list [e2e-llm-inference-service] - watch [e2e-llm-inference-service] attributeRestrictions: null [e2e-llm-inference-service] apiGroups: [e2e-llm-inference-service] - discovery.k8s.io [e2e-llm-inference-service] resources: [e2e-llm-inference-service] - endpointslices [e2e-llm-inference-service] - verbs: [e2e-llm-inference-service] - create [e2e-llm-inference-service] - delete [e2e-llm-inference-service] - get [e2e-llm-inference-service] - list [e2e-llm-inference-service] - patch [e2e-llm-inference-service] - update [e2e-llm-inference-service] - watch [e2e-llm-inference-service] attributeRestrictions: null [e2e-llm-inference-service] apiGroups: [e2e-llm-inference-service] - coordination.k8s.io [e2e-llm-inference-service] resources: [e2e-llm-inference-service] - leases [e2e-llm-inference-service] apiVersion: authorization.openshift.io/v1 [e2e-llm-inference-service] kind: Role [e2e-llm-inference-service] [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1254 --- [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1255 apiVersion: gateway.networking.k8s.io/v1 [e2e-llm-inference-service] kind: HTTPRoute [e2e-llm-inference-service] metadata: [e2e-llm-inference-service] annotations: [e2e-llm-inference-service] serving.kserve.io/inference-pool-migrated: v1 [e2e-llm-inference-service] creationTimestamp: '2026-07-07T14:31:40Z' [e2e-llm-inference-service] generation: 2 [e2e-llm-inference-service] labels: [e2e-llm-inference-service] app.kubernetes.io/component: llminferenceservice-router [e2e-llm-inference-service] app.kubernetes.io/name: llmisvc-model-fb-opt-125m-route-87882a8e [e2e-llm-inference-service] app.kubernetes.io/part-of: llminferenceservice [e2e-llm-inference-service] managedFields: [e2e-llm-inference-service] - apiVersion: gateway.networking.k8s.io/v1 [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:metadata: [e2e-llm-inference-service] f:annotations: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:serving.kserve.io/inference-pool-migrated: {} [e2e-llm-inference-service] f:labels: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:app.kubernetes.io/component: {} [e2e-llm-inference-service] f:app.kubernetes.io/name: {} [e2e-llm-inference-service] f:app.kubernetes.io/part-of: {} [e2e-llm-inference-service] f:ownerReferences: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"uid":"88c60d0e-95bd-401f-ae1c-2c4c0b1e7c9d"}: {} [e2e-llm-inference-service] f:spec: {} [e2e-llm-inference-service] manager: manager [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] time: '2026-07-07T14:31:46Z' [e2e-llm-inference-service] - apiVersion: gateway.networking.k8s.io/v1 [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:status: [e2e-llm-inference-service] f:parents: {} [e2e-llm-inference-service] manager: manager [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] subresource: status [e2e-llm-inference-service] time: '2026-07-07T14:31:46Z' [e2e-llm-inference-service] - apiVersion: gateway.networking.k8s.io/v1beta1 [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:status: {} [e2e-llm-inference-service] manager: pilot-discovery [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] subresource: status [e2e-llm-inference-service] time: '2026-07-07T14:31:46Z' [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-route-87882a8e-kserve-route [e2e-llm-inference-service] namespace: kserve-ci-e2e-test [e2e-llm-inference-service] ownerReferences: [e2e-llm-inference-service] - apiVersion: serving.kserve.io/v1alpha2 [e2e-llm-inference-service] blockOwnerDeletion: true [e2e-llm-inference-service] controller: true [e2e-llm-inference-service] kind: LLMInferenceService [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-route-87882a8e [e2e-llm-inference-service] uid: 88c60d0e-95bd-401f-ae1c-2c4c0b1e7c9d [e2e-llm-inference-service] resourceVersion: '31188' [e2e-llm-inference-service] uid: 30cc6b70-2948-4232-b987-d2b1a439b44f [e2e-llm-inference-service] spec: [e2e-llm-inference-service] parentRefs: [e2e-llm-inference-service] - group: gateway.networking.k8s.io [e2e-llm-inference-service] kind: Gateway [e2e-llm-inference-service] name: openshift-ai-inference [e2e-llm-inference-service] namespace: openshift-ingress [e2e-llm-inference-service] rules: [e2e-llm-inference-service] - backendRefs: [e2e-llm-inference-service] - group: inference.networking.k8s.io [e2e-llm-inference-service] kind: InferencePool [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-route-87882a8e-inference-pool [e2e-llm-inference-service] port: 8000 [e2e-llm-inference-service] weight: 1 [e2e-llm-inference-service] filters: [e2e-llm-inference-service] - type: URLRewrite [e2e-llm-inference-service] urlRewrite: [e2e-llm-inference-service] path: [e2e-llm-inference-service] replacePrefixMatch: /v1/completions [e2e-llm-inference-service] type: ReplacePrefixMatch [e2e-llm-inference-service] matches: [e2e-llm-inference-service] - path: [e2e-llm-inference-service] type: PathPrefix [e2e-llm-inference-service] value: /kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-route-87882a8e/v1/completions [e2e-llm-inference-service] timeouts: [e2e-llm-inference-service] backendRequest: 0s [e2e-llm-inference-service] request: 0s [e2e-llm-inference-service] - backendRefs: [e2e-llm-inference-service] - group: inference.networking.k8s.io [e2e-llm-inference-service] kind: InferencePool [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-route-87882a8e-inference-pool [e2e-llm-inference-service] port: 8000 [e2e-llm-inference-service] weight: 1 [e2e-llm-inference-service] matches: [e2e-llm-inference-service] - headers: [e2e-llm-inference-service] - name: X-Gateway-Model-Name [e2e-llm-inference-service] type: Exact [e2e-llm-inference-service] value: publishers/kserve-ci-e2e-test/models/facebook/opt-125m [e2e-llm-inference-service] path: [e2e-llm-inference-service] type: Exact [e2e-llm-inference-service] value: /v1/completions [e2e-llm-inference-service] - headers: [e2e-llm-inference-service] - name: X-Gateway-Model-Name [e2e-llm-inference-service] type: Exact [e2e-llm-inference-service] value: publishers/kserve-ci-e2e-test/models/facebook/opt-125m [e2e-llm-inference-service] path: [e2e-llm-inference-service] type: Exact [e2e-llm-inference-service] value: /v1/completions/ [e2e-llm-inference-service] timeouts: [e2e-llm-inference-service] backendRequest: 0s [e2e-llm-inference-service] request: 0s [e2e-llm-inference-service] - backendRefs: [e2e-llm-inference-service] - group: inference.networking.k8s.io [e2e-llm-inference-service] kind: InferencePool [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-route-87882a8e-inference-pool [e2e-llm-inference-service] port: 8000 [e2e-llm-inference-service] weight: 1 [e2e-llm-inference-service] filters: [e2e-llm-inference-service] - type: URLRewrite [e2e-llm-inference-service] urlRewrite: [e2e-llm-inference-service] path: [e2e-llm-inference-service] replacePrefixMatch: /v1/chat/completions [e2e-llm-inference-service] type: ReplacePrefixMatch [e2e-llm-inference-service] matches: [e2e-llm-inference-service] - path: [e2e-llm-inference-service] type: PathPrefix [e2e-llm-inference-service] value: /kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-route-87882a8e/v1/chat/completions [e2e-llm-inference-service] timeouts: [e2e-llm-inference-service] backendRequest: 0s [e2e-llm-inference-service] request: 0s [e2e-llm-inference-service] - backendRefs: [e2e-llm-inference-service] - group: inference.networking.k8s.io [e2e-llm-inference-service] kind: InferencePool [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-route-87882a8e-inference-pool [e2e-llm-inference-service] port: 8000 [e2e-llm-inference-service] weight: 1 [e2e-llm-inference-service] matches: [e2e-llm-inference-service] - headers: [e2e-llm-inference-service] - name: X-Gateway-Model-Name [e2e-llm-inference-service] type: Exact [e2e-llm-inference-service] value: publishers/kserve-ci-e2e-test/models/facebook/opt-125m [e2e-llm-inference-service] path: [e2e-llm-inference-service] type: Exact [e2e-llm-inference-service] value: /v1/chat/completions [e2e-llm-inference-service] - headers: [e2e-llm-inference-service] - name: X-Gateway-Model-Name [e2e-llm-inference-service] type: Exact [e2e-llm-inference-service] value: publishers/kserve-ci-e2e-test/models/facebook/opt-125m [e2e-llm-inference-service] path: [e2e-llm-inference-service] type: Exact [e2e-llm-inference-service] value: /v1/chat/completions/ [e2e-llm-inference-service] timeouts: [e2e-llm-inference-service] backendRequest: 0s [e2e-llm-inference-service] request: 0s [e2e-llm-inference-service] - backendRefs: [e2e-llm-inference-service] - group: inference.networking.k8s.io [e2e-llm-inference-service] kind: InferencePool [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-route-87882a8e-inference-pool [e2e-llm-inference-service] port: 8000 [e2e-llm-inference-service] weight: 1 [e2e-llm-inference-service] filters: [e2e-llm-inference-service] - type: URLRewrite [e2e-llm-inference-service] urlRewrite: [e2e-llm-inference-service] path: [e2e-llm-inference-service] replacePrefixMatch: /v1/responses [e2e-llm-inference-service] type: ReplacePrefixMatch [e2e-llm-inference-service] matches: [e2e-llm-inference-service] - path: [e2e-llm-inference-service] type: PathPrefix [e2e-llm-inference-service] value: /kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-route-87882a8e/v1/responses [e2e-llm-inference-service] timeouts: [e2e-llm-inference-service] backendRequest: 0s [e2e-llm-inference-service] request: 0s [e2e-llm-inference-service] - backendRefs: [e2e-llm-inference-service] - group: inference.networking.k8s.io [e2e-llm-inference-service] kind: InferencePool [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-route-87882a8e-inference-pool [e2e-llm-inference-service] port: 8000 [e2e-llm-inference-service] weight: 1 [e2e-llm-inference-service] matches: [e2e-llm-inference-service] - headers: [e2e-llm-inference-service] - name: X-Gateway-Model-Name [e2e-llm-inference-service] type: Exact [e2e-llm-inference-service] value: publishers/kserve-ci-e2e-test/models/facebook/opt-125m [e2e-llm-inference-service] path: [e2e-llm-inference-service] type: Exact [e2e-llm-inference-service] value: /v1/responses [e2e-llm-inference-service] - headers: [e2e-llm-inference-service] - name: X-Gateway-Model-Name [e2e-llm-inference-service] type: Exact [e2e-llm-inference-service] value: publishers/kserve-ci-e2e-test/models/facebook/opt-125m [e2e-llm-inference-service] path: [e2e-llm-inference-service] type: Exact [e2e-llm-inference-service] value: /v1/responses/ [e2e-llm-inference-service] timeouts: [e2e-llm-inference-service] backendRequest: 0s [e2e-llm-inference-service] request: 0s [e2e-llm-inference-service] - backendRefs: [e2e-llm-inference-service] - group: inference.networking.k8s.io [e2e-llm-inference-service] kind: InferencePool [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-route-87882a8e-inference-pool [e2e-llm-inference-service] port: 8000 [e2e-llm-inference-service] weight: 1 [e2e-llm-inference-service] filters: [e2e-llm-inference-service] - type: URLRewrite [e2e-llm-inference-service] urlRewrite: [e2e-llm-inference-service] path: [e2e-llm-inference-service] replacePrefixMatch: /v1/messages [e2e-llm-inference-service] type: ReplacePrefixMatch [e2e-llm-inference-service] matches: [e2e-llm-inference-service] - path: [e2e-llm-inference-service] type: PathPrefix [e2e-llm-inference-service] value: /kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-route-87882a8e/v1/messages [e2e-llm-inference-service] timeouts: [e2e-llm-inference-service] backendRequest: 0s [e2e-llm-inference-service] request: 0s [e2e-llm-inference-service] - backendRefs: [e2e-llm-inference-service] - group: inference.networking.k8s.io [e2e-llm-inference-service] kind: InferencePool [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-route-87882a8e-inference-pool [e2e-llm-inference-service] port: 8000 [e2e-llm-inference-service] weight: 1 [e2e-llm-inference-service] matches: [e2e-llm-inference-service] - headers: [e2e-llm-inference-service] - name: X-Gateway-Model-Name [e2e-llm-inference-service] type: Exact [e2e-llm-inference-service] value: publishers/kserve-ci-e2e-test/models/facebook/opt-125m [e2e-llm-inference-service] path: [e2e-llm-inference-service] type: Exact [e2e-llm-inference-service] value: /v1/messages [e2e-llm-inference-service] - headers: [e2e-llm-inference-service] - name: X-Gateway-Model-Name [e2e-llm-inference-service] type: Exact [e2e-llm-inference-service] value: publishers/kserve-ci-e2e-test/models/facebook/opt-125m [e2e-llm-inference-service] path: [e2e-llm-inference-service] type: Exact [e2e-llm-inference-service] value: /v1/messages/ [e2e-llm-inference-service] timeouts: [e2e-llm-inference-service] backendRequest: 0s [e2e-llm-inference-service] request: 0s [e2e-llm-inference-service] - backendRefs: [e2e-llm-inference-service] - group: '' [e2e-llm-inference-service] kind: Service [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-route-87882a8e-kserve-workload-svc [e2e-llm-inference-service] port: 8000 [e2e-llm-inference-service] weight: 1 [e2e-llm-inference-service] filters: [e2e-llm-inference-service] - type: URLRewrite [e2e-llm-inference-service] urlRewrite: [e2e-llm-inference-service] path: [e2e-llm-inference-service] replacePrefixMatch: / [e2e-llm-inference-service] type: ReplacePrefixMatch [e2e-llm-inference-service] matches: [e2e-llm-inference-service] - path: [e2e-llm-inference-service] type: PathPrefix [e2e-llm-inference-service] value: /kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-route-87882a8e [e2e-llm-inference-service] timeouts: [e2e-llm-inference-service] backendRequest: 0s [e2e-llm-inference-service] request: 0s [e2e-llm-inference-service] - backendRefs: [e2e-llm-inference-service] - group: '' [e2e-llm-inference-service] kind: Service [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-route-87882a8e-kserve-workload-svc [e2e-llm-inference-service] port: 8000 [e2e-llm-inference-service] weight: 1 [e2e-llm-inference-service] matches: [e2e-llm-inference-service] - headers: [e2e-llm-inference-service] - name: X-Gateway-Model-Name [e2e-llm-inference-service] type: Exact [e2e-llm-inference-service] value: publishers/kserve-ci-e2e-test/models/facebook/opt-125m [e2e-llm-inference-service] path: [e2e-llm-inference-service] type: PathPrefix [e2e-llm-inference-service] value: / [e2e-llm-inference-service] timeouts: [e2e-llm-inference-service] backendRequest: 0s [e2e-llm-inference-service] request: 0s [e2e-llm-inference-service] status: [e2e-llm-inference-service] parents: [e2e-llm-inference-service] - conditions: [e2e-llm-inference-service] - lastTransitionTime: '2026-07-07T14:31:40Z' [e2e-llm-inference-service] message: Route was valid [e2e-llm-inference-service] observedGeneration: 2 [e2e-llm-inference-service] reason: Accepted [e2e-llm-inference-service] status: 'True' [e2e-llm-inference-service] type: Accepted [e2e-llm-inference-service] - lastTransitionTime: '2026-07-07T14:31:46Z' [e2e-llm-inference-service] message: All references resolved [e2e-llm-inference-service] observedGeneration: 2 [e2e-llm-inference-service] reason: ResolvedRefs [e2e-llm-inference-service] status: 'True' [e2e-llm-inference-service] type: ResolvedRefs [e2e-llm-inference-service] controllerName: openshift.io/gateway-controller/v1 [e2e-llm-inference-service] parentRef: [e2e-llm-inference-service] group: gateway.networking.k8s.io [e2e-llm-inference-service] kind: Gateway [e2e-llm-inference-service] name: openshift-ai-inference [e2e-llm-inference-service] namespace: openshift-ingress [e2e-llm-inference-service] - conditions: [e2e-llm-inference-service] - lastTransitionTime: '2026-07-07T14:31:41Z' [e2e-llm-inference-service] message: Object affected by AuthPolicy [kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-route-87882a8e-kserve-route-authn [e2e-llm-inference-service] openshift-ingress/openshift-ai-inference-authn] [e2e-llm-inference-service] observedGeneration: 2 [e2e-llm-inference-service] reason: Accepted [e2e-llm-inference-service] status: 'True' [e2e-llm-inference-service] type: kuadrant.io/AuthPolicyAffected [e2e-llm-inference-service] controllerName: kuadrant.io/policy-controller [e2e-llm-inference-service] parentRef: [e2e-llm-inference-service] group: gateway.networking.k8s.io [e2e-llm-inference-service] kind: Gateway [e2e-llm-inference-service] name: openshift-ai-inference [e2e-llm-inference-service] namespace: openshift-ingress [e2e-llm-inference-service] [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1254 --- [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1255 apiVersion: gateway.networking.k8s.io/v1beta1 [e2e-llm-inference-service] kind: HTTPRoute [e2e-llm-inference-service] metadata: [e2e-llm-inference-service] annotations: [e2e-llm-inference-service] serving.kserve.io/inference-pool-migrated: v1 [e2e-llm-inference-service] creationTimestamp: '2026-07-07T14:31:40Z' [e2e-llm-inference-service] generation: 2 [e2e-llm-inference-service] labels: [e2e-llm-inference-service] app.kubernetes.io/component: llminferenceservice-router [e2e-llm-inference-service] app.kubernetes.io/name: llmisvc-model-fb-opt-125m-route-87882a8e [e2e-llm-inference-service] app.kubernetes.io/part-of: llminferenceservice [e2e-llm-inference-service] managedFields: [e2e-llm-inference-service] - apiVersion: gateway.networking.k8s.io/v1 [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:metadata: [e2e-llm-inference-service] f:annotations: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:serving.kserve.io/inference-pool-migrated: {} [e2e-llm-inference-service] f:labels: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:app.kubernetes.io/component: {} [e2e-llm-inference-service] f:app.kubernetes.io/name: {} [e2e-llm-inference-service] f:app.kubernetes.io/part-of: {} [e2e-llm-inference-service] f:ownerReferences: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"uid":"88c60d0e-95bd-401f-ae1c-2c4c0b1e7c9d"}: {} [e2e-llm-inference-service] f:spec: {} [e2e-llm-inference-service] manager: manager [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] time: '2026-07-07T14:31:46Z' [e2e-llm-inference-service] - apiVersion: gateway.networking.k8s.io/v1 [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:status: [e2e-llm-inference-service] f:parents: {} [e2e-llm-inference-service] manager: manager [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] subresource: status [e2e-llm-inference-service] time: '2026-07-07T14:31:46Z' [e2e-llm-inference-service] - apiVersion: gateway.networking.k8s.io/v1beta1 [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:status: {} [e2e-llm-inference-service] manager: pilot-discovery [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] subresource: status [e2e-llm-inference-service] time: '2026-07-07T14:31:46Z' [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-route-87882a8e-kserve-route [e2e-llm-inference-service] namespace: kserve-ci-e2e-test [e2e-llm-inference-service] ownerReferences: [e2e-llm-inference-service] - apiVersion: serving.kserve.io/v1alpha2 [e2e-llm-inference-service] blockOwnerDeletion: true [e2e-llm-inference-service] controller: true [e2e-llm-inference-service] kind: LLMInferenceService [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-route-87882a8e [e2e-llm-inference-service] uid: 88c60d0e-95bd-401f-ae1c-2c4c0b1e7c9d [e2e-llm-inference-service] resourceVersion: '31188' [e2e-llm-inference-service] uid: 30cc6b70-2948-4232-b987-d2b1a439b44f [e2e-llm-inference-service] spec: [e2e-llm-inference-service] parentRefs: [e2e-llm-inference-service] - group: gateway.networking.k8s.io [e2e-llm-inference-service] kind: Gateway [e2e-llm-inference-service] name: openshift-ai-inference [e2e-llm-inference-service] namespace: openshift-ingress [e2e-llm-inference-service] rules: [e2e-llm-inference-service] - backendRefs: [e2e-llm-inference-service] - group: inference.networking.k8s.io [e2e-llm-inference-service] kind: InferencePool [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-route-87882a8e-inference-pool [e2e-llm-inference-service] port: 8000 [e2e-llm-inference-service] weight: 1 [e2e-llm-inference-service] filters: [e2e-llm-inference-service] - type: URLRewrite [e2e-llm-inference-service] urlRewrite: [e2e-llm-inference-service] path: [e2e-llm-inference-service] replacePrefixMatch: /v1/completions [e2e-llm-inference-service] type: ReplacePrefixMatch [e2e-llm-inference-service] matches: [e2e-llm-inference-service] - path: [e2e-llm-inference-service] type: PathPrefix [e2e-llm-inference-service] value: /kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-route-87882a8e/v1/completions [e2e-llm-inference-service] timeouts: [e2e-llm-inference-service] backendRequest: 0s [e2e-llm-inference-service] request: 0s [e2e-llm-inference-service] - backendRefs: [e2e-llm-inference-service] - group: inference.networking.k8s.io [e2e-llm-inference-service] kind: InferencePool [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-route-87882a8e-inference-pool [e2e-llm-inference-service] port: 8000 [e2e-llm-inference-service] weight: 1 [e2e-llm-inference-service] matches: [e2e-llm-inference-service] - headers: [e2e-llm-inference-service] - name: X-Gateway-Model-Name [e2e-llm-inference-service] type: Exact [e2e-llm-inference-service] value: publishers/kserve-ci-e2e-test/models/facebook/opt-125m [e2e-llm-inference-service] path: [e2e-llm-inference-service] type: Exact [e2e-llm-inference-service] value: /v1/completions [e2e-llm-inference-service] - headers: [e2e-llm-inference-service] - name: X-Gateway-Model-Name [e2e-llm-inference-service] type: Exact [e2e-llm-inference-service] value: publishers/kserve-ci-e2e-test/models/facebook/opt-125m [e2e-llm-inference-service] path: [e2e-llm-inference-service] type: Exact [e2e-llm-inference-service] value: /v1/completions/ [e2e-llm-inference-service] timeouts: [e2e-llm-inference-service] backendRequest: 0s [e2e-llm-inference-service] request: 0s [e2e-llm-inference-service] - backendRefs: [e2e-llm-inference-service] - group: inference.networking.k8s.io [e2e-llm-inference-service] kind: InferencePool [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-route-87882a8e-inference-pool [e2e-llm-inference-service] port: 8000 [e2e-llm-inference-service] weight: 1 [e2e-llm-inference-service] filters: [e2e-llm-inference-service] - type: URLRewrite [e2e-llm-inference-service] urlRewrite: [e2e-llm-inference-service] path: [e2e-llm-inference-service] replacePrefixMatch: /v1/chat/completions [e2e-llm-inference-service] type: ReplacePrefixMatch [e2e-llm-inference-service] matches: [e2e-llm-inference-service] - path: [e2e-llm-inference-service] type: PathPrefix [e2e-llm-inference-service] value: /kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-route-87882a8e/v1/chat/completions [e2e-llm-inference-service] timeouts: [e2e-llm-inference-service] backendRequest: 0s [e2e-llm-inference-service] request: 0s [e2e-llm-inference-service] - backendRefs: [e2e-llm-inference-service] - group: inference.networking.k8s.io [e2e-llm-inference-service] kind: InferencePool [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-route-87882a8e-inference-pool [e2e-llm-inference-service] port: 8000 [e2e-llm-inference-service] weight: 1 [e2e-llm-inference-service] matches: [e2e-llm-inference-service] - headers: [e2e-llm-inference-service] - name: X-Gateway-Model-Name [e2e-llm-inference-service] type: Exact [e2e-llm-inference-service] value: publishers/kserve-ci-e2e-test/models/facebook/opt-125m [e2e-llm-inference-service] path: [e2e-llm-inference-service] type: Exact [e2e-llm-inference-service] value: /v1/chat/completions [e2e-llm-inference-service] - headers: [e2e-llm-inference-service] - name: X-Gateway-Model-Name [e2e-llm-inference-service] type: Exact [e2e-llm-inference-service] value: publishers/kserve-ci-e2e-test/models/facebook/opt-125m [e2e-llm-inference-service] path: [e2e-llm-inference-service] type: Exact [e2e-llm-inference-service] value: /v1/chat/completions/ [e2e-llm-inference-service] timeouts: [e2e-llm-inference-service] backendRequest: 0s [e2e-llm-inference-service] request: 0s [e2e-llm-inference-service] - backendRefs: [e2e-llm-inference-service] - group: inference.networking.k8s.io [e2e-llm-inference-service] kind: InferencePool [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-route-87882a8e-inference-pool [e2e-llm-inference-service] port: 8000 [e2e-llm-inference-service] weight: 1 [e2e-llm-inference-service] filters: [e2e-llm-inference-service] - type: URLRewrite [e2e-llm-inference-service] urlRewrite: [e2e-llm-inference-service] path: [e2e-llm-inference-service] replacePrefixMatch: /v1/responses [e2e-llm-inference-service] type: ReplacePrefixMatch [e2e-llm-inference-service] matches: [e2e-llm-inference-service] - path: [e2e-llm-inference-service] type: PathPrefix [e2e-llm-inference-service] value: /kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-route-87882a8e/v1/responses [e2e-llm-inference-service] timeouts: [e2e-llm-inference-service] backendRequest: 0s [e2e-llm-inference-service] request: 0s [e2e-llm-inference-service] - backendRefs: [e2e-llm-inference-service] - group: inference.networking.k8s.io [e2e-llm-inference-service] kind: InferencePool [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-route-87882a8e-inference-pool [e2e-llm-inference-service] port: 8000 [e2e-llm-inference-service] weight: 1 [e2e-llm-inference-service] matches: [e2e-llm-inference-service] - headers: [e2e-llm-inference-service] - name: X-Gateway-Model-Name [e2e-llm-inference-service] type: Exact [e2e-llm-inference-service] value: publishers/kserve-ci-e2e-test/models/facebook/opt-125m [e2e-llm-inference-service] path: [e2e-llm-inference-service] type: Exact [e2e-llm-inference-service] value: /v1/responses [e2e-llm-inference-service] - headers: [e2e-llm-inference-service] - name: X-Gateway-Model-Name [e2e-llm-inference-service] type: Exact [e2e-llm-inference-service] value: publishers/kserve-ci-e2e-test/models/facebook/opt-125m [e2e-llm-inference-service] path: [e2e-llm-inference-service] type: Exact [e2e-llm-inference-service] value: /v1/responses/ [e2e-llm-inference-service] timeouts: [e2e-llm-inference-service] backendRequest: 0s [e2e-llm-inference-service] request: 0s [e2e-llm-inference-service] - backendRefs: [e2e-llm-inference-service] - group: inference.networking.k8s.io [e2e-llm-inference-service] kind: InferencePool [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-route-87882a8e-inference-pool [e2e-llm-inference-service] port: 8000 [e2e-llm-inference-service] weight: 1 [e2e-llm-inference-service] filters: [e2e-llm-inference-service] - type: URLRewrite [e2e-llm-inference-service] urlRewrite: [e2e-llm-inference-service] path: [e2e-llm-inference-service] replacePrefixMatch: /v1/messages [e2e-llm-inference-service] type: ReplacePrefixMatch [e2e-llm-inference-service] matches: [e2e-llm-inference-service] - path: [e2e-llm-inference-service] type: PathPrefix [e2e-llm-inference-service] value: /kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-route-87882a8e/v1/messages [e2e-llm-inference-service] timeouts: [e2e-llm-inference-service] backendRequest: 0s [e2e-llm-inference-service] request: 0s [e2e-llm-inference-service] - backendRefs: [e2e-llm-inference-service] - group: inference.networking.k8s.io [e2e-llm-inference-service] kind: InferencePool [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-route-87882a8e-inference-pool [e2e-llm-inference-service] port: 8000 [e2e-llm-inference-service] weight: 1 [e2e-llm-inference-service] matches: [e2e-llm-inference-service] - headers: [e2e-llm-inference-service] - name: X-Gateway-Model-Name [e2e-llm-inference-service] type: Exact [e2e-llm-inference-service] value: publishers/kserve-ci-e2e-test/models/facebook/opt-125m [e2e-llm-inference-service] path: [e2e-llm-inference-service] type: Exact [e2e-llm-inference-service] value: /v1/messages [e2e-llm-inference-service] - headers: [e2e-llm-inference-service] - name: X-Gateway-Model-Name [e2e-llm-inference-service] type: Exact [e2e-llm-inference-service] value: publishers/kserve-ci-e2e-test/models/facebook/opt-125m [e2e-llm-inference-service] path: [e2e-llm-inference-service] type: Exact [e2e-llm-inference-service] value: /v1/messages/ [e2e-llm-inference-service] timeouts: [e2e-llm-inference-service] backendRequest: 0s [e2e-llm-inference-service] request: 0s [e2e-llm-inference-service] - backendRefs: [e2e-llm-inference-service] - group: '' [e2e-llm-inference-service] kind: Service [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-route-87882a8e-kserve-workload-svc [e2e-llm-inference-service] port: 8000 [e2e-llm-inference-service] weight: 1 [e2e-llm-inference-service] filters: [e2e-llm-inference-service] - type: URLRewrite [e2e-llm-inference-service] urlRewrite: [e2e-llm-inference-service] path: [e2e-llm-inference-service] replacePrefixMatch: / [e2e-llm-inference-service] type: ReplacePrefixMatch [e2e-llm-inference-service] matches: [e2e-llm-inference-service] - path: [e2e-llm-inference-service] type: PathPrefix [e2e-llm-inference-service] value: /kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-route-87882a8e [e2e-llm-inference-service] timeouts: [e2e-llm-inference-service] backendRequest: 0s [e2e-llm-inference-service] request: 0s [e2e-llm-inference-service] - backendRefs: [e2e-llm-inference-service] - group: '' [e2e-llm-inference-service] kind: Service [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-route-87882a8e-kserve-workload-svc [e2e-llm-inference-service] port: 8000 [e2e-llm-inference-service] weight: 1 [e2e-llm-inference-service] matches: [e2e-llm-inference-service] - headers: [e2e-llm-inference-service] - name: X-Gateway-Model-Name [e2e-llm-inference-service] type: Exact [e2e-llm-inference-service] value: publishers/kserve-ci-e2e-test/models/facebook/opt-125m [e2e-llm-inference-service] path: [e2e-llm-inference-service] type: PathPrefix [e2e-llm-inference-service] value: / [e2e-llm-inference-service] timeouts: [e2e-llm-inference-service] backendRequest: 0s [e2e-llm-inference-service] request: 0s [e2e-llm-inference-service] status: [e2e-llm-inference-service] parents: [e2e-llm-inference-service] - conditions: [e2e-llm-inference-service] - lastTransitionTime: '2026-07-07T14:31:40Z' [e2e-llm-inference-service] message: Route was valid [e2e-llm-inference-service] observedGeneration: 2 [e2e-llm-inference-service] reason: Accepted [e2e-llm-inference-service] status: 'True' [e2e-llm-inference-service] type: Accepted [e2e-llm-inference-service] - lastTransitionTime: '2026-07-07T14:31:46Z' [e2e-llm-inference-service] message: All references resolved [e2e-llm-inference-service] observedGeneration: 2 [e2e-llm-inference-service] reason: ResolvedRefs [e2e-llm-inference-service] status: 'True' [e2e-llm-inference-service] type: ResolvedRefs [e2e-llm-inference-service] controllerName: openshift.io/gateway-controller/v1 [e2e-llm-inference-service] parentRef: [e2e-llm-inference-service] group: gateway.networking.k8s.io [e2e-llm-inference-service] kind: Gateway [e2e-llm-inference-service] name: openshift-ai-inference [e2e-llm-inference-service] namespace: openshift-ingress [e2e-llm-inference-service] - conditions: [e2e-llm-inference-service] - lastTransitionTime: '2026-07-07T14:31:41Z' [e2e-llm-inference-service] message: Object affected by AuthPolicy [kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-route-87882a8e-kserve-route-authn [e2e-llm-inference-service] openshift-ingress/openshift-ai-inference-authn] [e2e-llm-inference-service] observedGeneration: 2 [e2e-llm-inference-service] reason: Accepted [e2e-llm-inference-service] status: 'True' [e2e-llm-inference-service] type: kuadrant.io/AuthPolicyAffected [e2e-llm-inference-service] controllerName: kuadrant.io/policy-controller [e2e-llm-inference-service] parentRef: [e2e-llm-inference-service] group: gateway.networking.k8s.io [e2e-llm-inference-service] kind: Gateway [e2e-llm-inference-service] name: openshift-ai-inference [e2e-llm-inference-service] namespace: openshift-ingress [e2e-llm-inference-service] [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1254 --- [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1255 apiVersion: inference.networking.k8s.io/v1 [e2e-llm-inference-service] kind: InferencePool [e2e-llm-inference-service] metadata: [e2e-llm-inference-service] creationTimestamp: '2026-07-07T14:31:40Z' [e2e-llm-inference-service] generation: 1 [e2e-llm-inference-service] labels: [e2e-llm-inference-service] app.kubernetes.io/component: llminferenceservice-router-scheduler [e2e-llm-inference-service] app.kubernetes.io/name: llmisvc-model-fb-opt-125m-route-87882a8e [e2e-llm-inference-service] app.kubernetes.io/part-of: llminferenceservice [e2e-llm-inference-service] managedFields: [e2e-llm-inference-service] - apiVersion: inference.networking.k8s.io/v1 [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:metadata: [e2e-llm-inference-service] f:labels: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:app.kubernetes.io/component: {} [e2e-llm-inference-service] f:app.kubernetes.io/name: {} [e2e-llm-inference-service] f:app.kubernetes.io/part-of: {} [e2e-llm-inference-service] f:ownerReferences: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"uid":"88c60d0e-95bd-401f-ae1c-2c4c0b1e7c9d"}: {} [e2e-llm-inference-service] f:spec: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:appProtocol: {} [e2e-llm-inference-service] f:endpointPickerRef: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:failureMode: {} [e2e-llm-inference-service] f:group: {} [e2e-llm-inference-service] f:kind: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:port: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:number: {} [e2e-llm-inference-service] f:selector: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:matchLabels: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:app.kubernetes.io/name: {} [e2e-llm-inference-service] f:app.kubernetes.io/part-of: {} [e2e-llm-inference-service] f:kserve.io/component: {} [e2e-llm-inference-service] f:targetPorts: {} [e2e-llm-inference-service] manager: manager [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] time: '2026-07-07T14:31:40Z' [e2e-llm-inference-service] - apiVersion: inference.networking.k8s.io/v1 [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:status: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:parents: {} [e2e-llm-inference-service] manager: pilot-discovery [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] subresource: status [e2e-llm-inference-service] time: '2026-07-07T14:31:46Z' [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-route-87882a8e-inference-pool [e2e-llm-inference-service] namespace: kserve-ci-e2e-test [e2e-llm-inference-service] ownerReferences: [e2e-llm-inference-service] - apiVersion: serving.kserve.io/v1alpha2 [e2e-llm-inference-service] blockOwnerDeletion: true [e2e-llm-inference-service] controller: true [e2e-llm-inference-service] kind: LLMInferenceService [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-route-87882a8e [e2e-llm-inference-service] uid: 88c60d0e-95bd-401f-ae1c-2c4c0b1e7c9d [e2e-llm-inference-service] resourceVersion: '31170' [e2e-llm-inference-service] uid: 8fbd0a5d-18e4-40c4-8f18-639a2f9db911 [e2e-llm-inference-service] spec: [e2e-llm-inference-service] appProtocol: http [e2e-llm-inference-service] endpointPickerRef: [e2e-llm-inference-service] failureMode: FailOpen [e2e-llm-inference-service] group: '' [e2e-llm-inference-service] kind: Service [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-route-87882a8e-epp-service [e2e-llm-inference-service] port: [e2e-llm-inference-service] number: 9002 [e2e-llm-inference-service] selector: [e2e-llm-inference-service] matchLabels: [e2e-llm-inference-service] app.kubernetes.io/name: llmisvc-model-fb-opt-125m-route-87882a8e [e2e-llm-inference-service] app.kubernetes.io/part-of: llminferenceservice [e2e-llm-inference-service] kserve.io/component: workload [e2e-llm-inference-service] targetPorts: [e2e-llm-inference-service] - number: 8000 [e2e-llm-inference-service] status: [e2e-llm-inference-service] parents: [e2e-llm-inference-service] - conditions: [e2e-llm-inference-service] - lastTransitionTime: '2026-07-07T14:31:46Z' [e2e-llm-inference-service] message: Referenced by an HTTPRoute accepted by the parentRef Gateway [e2e-llm-inference-service] observedGeneration: 1 [e2e-llm-inference-service] reason: Accepted [e2e-llm-inference-service] status: 'True' [e2e-llm-inference-service] type: Accepted [e2e-llm-inference-service] - lastTransitionTime: '2026-07-07T14:31:46Z' [e2e-llm-inference-service] message: Referenced ExtensionRef resolved successfully [e2e-llm-inference-service] observedGeneration: 1 [e2e-llm-inference-service] reason: ResolvedRefs [e2e-llm-inference-service] status: 'True' [e2e-llm-inference-service] type: ResolvedRefs [e2e-llm-inference-service] parentRef: [e2e-llm-inference-service] group: networking.istio.io [e2e-llm-inference-service] kind: Gateway [e2e-llm-inference-service] name: openshift-ai-inference [e2e-llm-inference-service] namespace: openshift-ingress [e2e-llm-inference-service] [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1254 --- [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1255 apiVersion: kuadrant.io/v1 [e2e-llm-inference-service] kind: AuthPolicy [e2e-llm-inference-service] metadata: [e2e-llm-inference-service] creationTimestamp: '2026-07-07T14:31:40Z' [e2e-llm-inference-service] generation: 1 [e2e-llm-inference-service] labels: [e2e-llm-inference-service] app.kubernetes.io/component: llminferenceservice-policies [e2e-llm-inference-service] app.kubernetes.io/managed-by: odh-model-controller [e2e-llm-inference-service] app.kubernetes.io/name: llmisvc-model-fb-opt-125m-route-87882a8e [e2e-llm-inference-service] app.kubernetes.io/part-of: llminferenceservice [e2e-llm-inference-service] managedFields: [e2e-llm-inference-service] - apiVersion: kuadrant.io/v1 [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:metadata: [e2e-llm-inference-service] f:labels: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:app.kubernetes.io/component: {} [e2e-llm-inference-service] f:app.kubernetes.io/managed-by: {} [e2e-llm-inference-service] f:app.kubernetes.io/name: {} [e2e-llm-inference-service] f:app.kubernetes.io/part-of: {} [e2e-llm-inference-service] f:ownerReferences: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"uid":"88c60d0e-95bd-401f-ae1c-2c4c0b1e7c9d"}: {} [e2e-llm-inference-service] f:spec: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:rules: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:authentication: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:public: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:anonymous: {} [e2e-llm-inference-service] f:credentials: {} [e2e-llm-inference-service] f:metrics: {} [e2e-llm-inference-service] f:overrides: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:fairness: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] f:objective: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] f:priority: {} [e2e-llm-inference-service] f:response: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:success: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:headers: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:x-gateway-inference-fairness-id: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:metrics: {} [e2e-llm-inference-service] f:plain: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:expression: {} [e2e-llm-inference-service] f:priority: {} [e2e-llm-inference-service] f:x-gateway-inference-objective: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:metrics: {} [e2e-llm-inference-service] f:plain: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:expression: {} [e2e-llm-inference-service] f:priority: {} [e2e-llm-inference-service] f:targetRef: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:group: {} [e2e-llm-inference-service] f:kind: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] manager: manager [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] time: '2026-07-07T14:31:40Z' [e2e-llm-inference-service] - apiVersion: kuadrant.io/v1 [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:status: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:conditions: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"type":"Accepted"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:lastTransitionTime: {} [e2e-llm-inference-service] f:message: {} [e2e-llm-inference-service] f:reason: {} [e2e-llm-inference-service] f:status: {} [e2e-llm-inference-service] f:type: {} [e2e-llm-inference-service] k:{"type":"Enforced"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:lastTransitionTime: {} [e2e-llm-inference-service] f:message: {} [e2e-llm-inference-service] f:reason: {} [e2e-llm-inference-service] f:status: {} [e2e-llm-inference-service] f:type: {} [e2e-llm-inference-service] f:observedGeneration: {} [e2e-llm-inference-service] manager: manager [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] subresource: status [e2e-llm-inference-service] time: '2026-07-07T14:31:42Z' [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-route-87882a8e-kserve-route-authn [e2e-llm-inference-service] namespace: kserve-ci-e2e-test [e2e-llm-inference-service] ownerReferences: [e2e-llm-inference-service] - apiVersion: serving.kserve.io/v1alpha2 [e2e-llm-inference-service] blockOwnerDeletion: true [e2e-llm-inference-service] controller: true [e2e-llm-inference-service] kind: LLMInferenceService [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-route-87882a8e [e2e-llm-inference-service] uid: 88c60d0e-95bd-401f-ae1c-2c4c0b1e7c9d [e2e-llm-inference-service] resourceVersion: '31097' [e2e-llm-inference-service] uid: 37f23904-b644-4e44-920e-6ea461f5230d [e2e-llm-inference-service] spec: [e2e-llm-inference-service] rules: [e2e-llm-inference-service] authentication: [e2e-llm-inference-service] public: [e2e-llm-inference-service] anonymous: {} [e2e-llm-inference-service] credentials: {} [e2e-llm-inference-service] metrics: false [e2e-llm-inference-service] overrides: [e2e-llm-inference-service] fairness: [e2e-llm-inference-service] value: unauthenticated [e2e-llm-inference-service] objective: [e2e-llm-inference-service] value: unauthenticated [e2e-llm-inference-service] priority: 0 [e2e-llm-inference-service] response: [e2e-llm-inference-service] success: [e2e-llm-inference-service] headers: [e2e-llm-inference-service] x-gateway-inference-fairness-id: [e2e-llm-inference-service] metrics: false [e2e-llm-inference-service] plain: [e2e-llm-inference-service] expression: auth.identity.fairness [e2e-llm-inference-service] priority: 0 [e2e-llm-inference-service] x-gateway-inference-objective: [e2e-llm-inference-service] metrics: false [e2e-llm-inference-service] plain: [e2e-llm-inference-service] expression: auth.identity.objective [e2e-llm-inference-service] priority: 0 [e2e-llm-inference-service] targetRef: [e2e-llm-inference-service] group: gateway.networking.k8s.io [e2e-llm-inference-service] kind: HTTPRoute [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-route-87882a8e-kserve-route [e2e-llm-inference-service] status: [e2e-llm-inference-service] conditions: [e2e-llm-inference-service] - lastTransitionTime: '2026-07-07T14:31:41Z' [e2e-llm-inference-service] message: AuthPolicy has been accepted [e2e-llm-inference-service] reason: Accepted [e2e-llm-inference-service] status: 'True' [e2e-llm-inference-service] type: Accepted [e2e-llm-inference-service] - lastTransitionTime: '2026-07-07T14:31:42Z' [e2e-llm-inference-service] message: AuthPolicy has been successfully enforced [e2e-llm-inference-service] reason: Enforced [e2e-llm-inference-service] status: 'True' [e2e-llm-inference-service] type: Enforced [e2e-llm-inference-service] observedGeneration: 1 [e2e-llm-inference-service] [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1254 --- [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1255 apiVersion: networking.istio.io/v1 [e2e-llm-inference-service] kind: DestinationRule [e2e-llm-inference-service] metadata: [e2e-llm-inference-service] creationTimestamp: '2026-07-07T14:31:40Z' [e2e-llm-inference-service] generation: 1 [e2e-llm-inference-service] labels: [e2e-llm-inference-service] app.kubernetes.io/component: llminferenceservice-router-scheduler [e2e-llm-inference-service] app.kubernetes.io/name: llmisvc-model-fb-opt-125m-route-87882a8e [e2e-llm-inference-service] app.kubernetes.io/part-of: llminferenceservice [e2e-llm-inference-service] llm-d.ai/managed: 'true' [e2e-llm-inference-service] managedFields: [e2e-llm-inference-service] - apiVersion: networking.istio.io/v1 [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:metadata: [e2e-llm-inference-service] f:labels: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:app.kubernetes.io/component: {} [e2e-llm-inference-service] f:app.kubernetes.io/name: {} [e2e-llm-inference-service] f:app.kubernetes.io/part-of: {} [e2e-llm-inference-service] f:llm-d.ai/managed: {} [e2e-llm-inference-service] f:ownerReferences: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"uid":"88c60d0e-95bd-401f-ae1c-2c4c0b1e7c9d"}: {} [e2e-llm-inference-service] f:spec: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:exportTo: {} [e2e-llm-inference-service] f:host: {} [e2e-llm-inference-service] f:trafficPolicy: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:tls: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:insecureSkipVerify: {} [e2e-llm-inference-service] f:mode: {} [e2e-llm-inference-service] f:sni: {} [e2e-llm-inference-service] manager: manager [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] time: '2026-07-07T14:31:40Z' [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-route-87882a8e-kserve-scheduler [e2e-llm-inference-service] namespace: kserve-ci-e2e-test [e2e-llm-inference-service] ownerReferences: [e2e-llm-inference-service] - apiVersion: serving.kserve.io/v1alpha2 [e2e-llm-inference-service] blockOwnerDeletion: true [e2e-llm-inference-service] controller: true [e2e-llm-inference-service] kind: LLMInferenceService [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-route-87882a8e [e2e-llm-inference-service] uid: 88c60d0e-95bd-401f-ae1c-2c4c0b1e7c9d [e2e-llm-inference-service] resourceVersion: '31037' [e2e-llm-inference-service] uid: 36ac8ef0-fac2-4479-a42e-21b95a707ea2 [e2e-llm-inference-service] spec: [e2e-llm-inference-service] exportTo: [e2e-llm-inference-service] - '*' [e2e-llm-inference-service] host: llmisvc-model-fb-opt-125m-route-87882a8e-epp-service.kserve-ci-e2e-test.svc.cluster.local [e2e-llm-inference-service] trafficPolicy: [e2e-llm-inference-service] tls: [e2e-llm-inference-service] insecureSkipVerify: true [e2e-llm-inference-service] mode: SIMPLE [e2e-llm-inference-service] sni: llmisvc-model-fb-opt-125m-route-87882a8e-epp-service.kserve-ci-e2e-test.svc.cluster.local [e2e-llm-inference-service] [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1254 --- [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1255 apiVersion: networking.istio.io/v1 [e2e-llm-inference-service] kind: DestinationRule [e2e-llm-inference-service] metadata: [e2e-llm-inference-service] creationTimestamp: '2026-07-07T14:31:46Z' [e2e-llm-inference-service] generation: 1 [e2e-llm-inference-service] labels: [e2e-llm-inference-service] app.kubernetes.io/component: llminferenceservice-shadow-service [e2e-llm-inference-service] app.kubernetes.io/name: llmisvc-model-fb-opt-125m-route-87882a8e [e2e-llm-inference-service] app.kubernetes.io/part-of: llminferenceservice [e2e-llm-inference-service] llm-d.ai/managed: 'true' [e2e-llm-inference-service] managedFields: [e2e-llm-inference-service] - apiVersion: networking.istio.io/v1 [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:metadata: [e2e-llm-inference-service] f:labels: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:app.kubernetes.io/component: {} [e2e-llm-inference-service] f:app.kubernetes.io/name: {} [e2e-llm-inference-service] f:app.kubernetes.io/part-of: {} [e2e-llm-inference-service] f:llm-d.ai/managed: {} [e2e-llm-inference-service] f:ownerReferences: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"uid":"88c60d0e-95bd-401f-ae1c-2c4c0b1e7c9d"}: {} [e2e-llm-inference-service] f:spec: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:exportTo: {} [e2e-llm-inference-service] f:host: {} [e2e-llm-inference-service] f:trafficPolicy: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:tls: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:insecureSkipVerify: {} [e2e-llm-inference-service] f:mode: {} [e2e-llm-inference-service] f:sni: {} [e2e-llm-inference-service] manager: manager [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] time: '2026-07-07T14:31:46Z' [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-route-87882a8e-kserve-shadow-svc [e2e-llm-inference-service] namespace: kserve-ci-e2e-test [e2e-llm-inference-service] ownerReferences: [e2e-llm-inference-service] - apiVersion: serving.kserve.io/v1alpha2 [e2e-llm-inference-service] blockOwnerDeletion: true [e2e-llm-inference-service] controller: true [e2e-llm-inference-service] kind: LLMInferenceService [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-route-87882a8e [e2e-llm-inference-service] uid: 88c60d0e-95bd-401f-ae1c-2c4c0b1e7c9d [e2e-llm-inference-service] resourceVersion: '31187' [e2e-llm-inference-service] uid: 42e3e102-30d1-4377-b348-f59f61d9fc50 [e2e-llm-inference-service] spec: [e2e-llm-inference-service] exportTo: [e2e-llm-inference-service] - '*' [e2e-llm-inference-service] host: llmisvc-model-fb-opt-125m-route-87882a8e-inference--ip-aaa905cb.kserve-ci-e2e-test.svc.cluster.local [e2e-llm-inference-service] trafficPolicy: [e2e-llm-inference-service] tls: [e2e-llm-inference-service] insecureSkipVerify: true [e2e-llm-inference-service] mode: SIMPLE [e2e-llm-inference-service] sni: llmisvc-model-fb-opt-125m-route-87882a8e-kserve-workload-svc.kserve-ci-e2e-test.svc.cluster.local [e2e-llm-inference-service] [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1254 --- [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1255 apiVersion: networking.istio.io/v1 [e2e-llm-inference-service] kind: DestinationRule [e2e-llm-inference-service] metadata: [e2e-llm-inference-service] creationTimestamp: '2026-07-07T14:31:41Z' [e2e-llm-inference-service] generation: 1 [e2e-llm-inference-service] labels: [e2e-llm-inference-service] app.kubernetes.io/component: llminferenceservice-workload [e2e-llm-inference-service] app.kubernetes.io/name: llmisvc-model-fb-opt-125m-route-87882a8e [e2e-llm-inference-service] app.kubernetes.io/part-of: llminferenceservice [e2e-llm-inference-service] llm-d.ai/managed: 'true' [e2e-llm-inference-service] managedFields: [e2e-llm-inference-service] - apiVersion: networking.istio.io/v1 [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:metadata: [e2e-llm-inference-service] f:labels: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:app.kubernetes.io/component: {} [e2e-llm-inference-service] f:app.kubernetes.io/name: {} [e2e-llm-inference-service] f:app.kubernetes.io/part-of: {} [e2e-llm-inference-service] f:llm-d.ai/managed: {} [e2e-llm-inference-service] f:ownerReferences: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"uid":"88c60d0e-95bd-401f-ae1c-2c4c0b1e7c9d"}: {} [e2e-llm-inference-service] f:spec: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:exportTo: {} [e2e-llm-inference-service] f:host: {} [e2e-llm-inference-service] f:trafficPolicy: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:tls: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:caCertificates: {} [e2e-llm-inference-service] f:insecureSkipVerify: {} [e2e-llm-inference-service] f:mode: {} [e2e-llm-inference-service] f:sni: {} [e2e-llm-inference-service] manager: manager [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] time: '2026-07-07T14:31:41Z' [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-route-87882a8e-kserve-workload-svc [e2e-llm-inference-service] namespace: kserve-ci-e2e-test [e2e-llm-inference-service] ownerReferences: [e2e-llm-inference-service] - apiVersion: serving.kserve.io/v1alpha2 [e2e-llm-inference-service] blockOwnerDeletion: true [e2e-llm-inference-service] controller: true [e2e-llm-inference-service] kind: LLMInferenceService [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-route-87882a8e [e2e-llm-inference-service] uid: 88c60d0e-95bd-401f-ae1c-2c4c0b1e7c9d [e2e-llm-inference-service] resourceVersion: '31046' [e2e-llm-inference-service] uid: dd2603eb-b21f-4a15-b5a5-45dd42d3e4bf [e2e-llm-inference-service] spec: [e2e-llm-inference-service] exportTo: [e2e-llm-inference-service] - '*' [e2e-llm-inference-service] host: llmisvc-model-fb-opt-125m-route-87882a8e-kserve-workload-svc.kserve-ci-e2e-test.svc.cluster.local [e2e-llm-inference-service] trafficPolicy: [e2e-llm-inference-service] tls: [e2e-llm-inference-service] caCertificates: /var/run/secrets/kubernetes.io/serviceaccount/service-ca.crt [e2e-llm-inference-service] insecureSkipVerify: false [e2e-llm-inference-service] mode: SIMPLE [e2e-llm-inference-service] sni: llmisvc-model-fb-opt-125m-route-87882a8e-kserve-workload-svc.kserve-ci-e2e-test.svc.cluster.local [e2e-llm-inference-service] [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1254 --- [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1255 apiVersion: networking.istio.io/v1beta1 [e2e-llm-inference-service] kind: DestinationRule [e2e-llm-inference-service] metadata: [e2e-llm-inference-service] creationTimestamp: '2026-07-07T14:31:40Z' [e2e-llm-inference-service] generation: 1 [e2e-llm-inference-service] labels: [e2e-llm-inference-service] app.kubernetes.io/component: llminferenceservice-router-scheduler [e2e-llm-inference-service] app.kubernetes.io/name: llmisvc-model-fb-opt-125m-route-87882a8e [e2e-llm-inference-service] app.kubernetes.io/part-of: llminferenceservice [e2e-llm-inference-service] llm-d.ai/managed: 'true' [e2e-llm-inference-service] managedFields: [e2e-llm-inference-service] - apiVersion: networking.istio.io/v1 [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:metadata: [e2e-llm-inference-service] f:labels: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:app.kubernetes.io/component: {} [e2e-llm-inference-service] f:app.kubernetes.io/name: {} [e2e-llm-inference-service] f:app.kubernetes.io/part-of: {} [e2e-llm-inference-service] f:llm-d.ai/managed: {} [e2e-llm-inference-service] f:ownerReferences: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"uid":"88c60d0e-95bd-401f-ae1c-2c4c0b1e7c9d"}: {} [e2e-llm-inference-service] f:spec: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:exportTo: {} [e2e-llm-inference-service] f:host: {} [e2e-llm-inference-service] f:trafficPolicy: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:tls: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:insecureSkipVerify: {} [e2e-llm-inference-service] f:mode: {} [e2e-llm-inference-service] f:sni: {} [e2e-llm-inference-service] manager: manager [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] time: '2026-07-07T14:31:40Z' [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-route-87882a8e-kserve-scheduler [e2e-llm-inference-service] namespace: kserve-ci-e2e-test [e2e-llm-inference-service] ownerReferences: [e2e-llm-inference-service] - apiVersion: serving.kserve.io/v1alpha2 [e2e-llm-inference-service] blockOwnerDeletion: true [e2e-llm-inference-service] controller: true [e2e-llm-inference-service] kind: LLMInferenceService [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-route-87882a8e [e2e-llm-inference-service] uid: 88c60d0e-95bd-401f-ae1c-2c4c0b1e7c9d [e2e-llm-inference-service] resourceVersion: '31037' [e2e-llm-inference-service] uid: 36ac8ef0-fac2-4479-a42e-21b95a707ea2 [e2e-llm-inference-service] spec: [e2e-llm-inference-service] exportTo: [e2e-llm-inference-service] - '*' [e2e-llm-inference-service] host: llmisvc-model-fb-opt-125m-route-87882a8e-epp-service.kserve-ci-e2e-test.svc.cluster.local [e2e-llm-inference-service] trafficPolicy: [e2e-llm-inference-service] tls: [e2e-llm-inference-service] insecureSkipVerify: true [e2e-llm-inference-service] mode: SIMPLE [e2e-llm-inference-service] sni: llmisvc-model-fb-opt-125m-route-87882a8e-epp-service.kserve-ci-e2e-test.svc.cluster.local [e2e-llm-inference-service] [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1254 --- [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1255 apiVersion: networking.istio.io/v1beta1 [e2e-llm-inference-service] kind: DestinationRule [e2e-llm-inference-service] metadata: [e2e-llm-inference-service] creationTimestamp: '2026-07-07T14:31:46Z' [e2e-llm-inference-service] generation: 1 [e2e-llm-inference-service] labels: [e2e-llm-inference-service] app.kubernetes.io/component: llminferenceservice-shadow-service [e2e-llm-inference-service] app.kubernetes.io/name: llmisvc-model-fb-opt-125m-route-87882a8e [e2e-llm-inference-service] app.kubernetes.io/part-of: llminferenceservice [e2e-llm-inference-service] llm-d.ai/managed: 'true' [e2e-llm-inference-service] managedFields: [e2e-llm-inference-service] - apiVersion: networking.istio.io/v1 [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:metadata: [e2e-llm-inference-service] f:labels: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:app.kubernetes.io/component: {} [e2e-llm-inference-service] f:app.kubernetes.io/name: {} [e2e-llm-inference-service] f:app.kubernetes.io/part-of: {} [e2e-llm-inference-service] f:llm-d.ai/managed: {} [e2e-llm-inference-service] f:ownerReferences: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"uid":"88c60d0e-95bd-401f-ae1c-2c4c0b1e7c9d"}: {} [e2e-llm-inference-service] f:spec: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:exportTo: {} [e2e-llm-inference-service] f:host: {} [e2e-llm-inference-service] f:trafficPolicy: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:tls: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:insecureSkipVerify: {} [e2e-llm-inference-service] f:mode: {} [e2e-llm-inference-service] f:sni: {} [e2e-llm-inference-service] manager: manager [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] time: '2026-07-07T14:31:46Z' [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-route-87882a8e-kserve-shadow-svc [e2e-llm-inference-service] namespace: kserve-ci-e2e-test [e2e-llm-inference-service] ownerReferences: [e2e-llm-inference-service] - apiVersion: serving.kserve.io/v1alpha2 [e2e-llm-inference-service] blockOwnerDeletion: true [e2e-llm-inference-service] controller: true [e2e-llm-inference-service] kind: LLMInferenceService [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-route-87882a8e [e2e-llm-inference-service] uid: 88c60d0e-95bd-401f-ae1c-2c4c0b1e7c9d [e2e-llm-inference-service] resourceVersion: '31187' [e2e-llm-inference-service] uid: 42e3e102-30d1-4377-b348-f59f61d9fc50 [e2e-llm-inference-service] spec: [e2e-llm-inference-service] exportTo: [e2e-llm-inference-service] - '*' [e2e-llm-inference-service] host: llmisvc-model-fb-opt-125m-route-87882a8e-inference--ip-aaa905cb.kserve-ci-e2e-test.svc.cluster.local [e2e-llm-inference-service] trafficPolicy: [e2e-llm-inference-service] tls: [e2e-llm-inference-service] insecureSkipVerify: true [e2e-llm-inference-service] mode: SIMPLE [e2e-llm-inference-service] sni: llmisvc-model-fb-opt-125m-route-87882a8e-kserve-workload-svc.kserve-ci-e2e-test.svc.cluster.local [e2e-llm-inference-service] [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1254 --- [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1255 apiVersion: networking.istio.io/v1beta1 [e2e-llm-inference-service] kind: DestinationRule [e2e-llm-inference-service] metadata: [e2e-llm-inference-service] creationTimestamp: '2026-07-07T14:31:41Z' [e2e-llm-inference-service] generation: 1 [e2e-llm-inference-service] labels: [e2e-llm-inference-service] app.kubernetes.io/component: llminferenceservice-workload [e2e-llm-inference-service] app.kubernetes.io/name: llmisvc-model-fb-opt-125m-route-87882a8e [e2e-llm-inference-service] app.kubernetes.io/part-of: llminferenceservice [e2e-llm-inference-service] llm-d.ai/managed: 'true' [e2e-llm-inference-service] managedFields: [e2e-llm-inference-service] - apiVersion: networking.istio.io/v1 [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:metadata: [e2e-llm-inference-service] f:labels: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:app.kubernetes.io/component: {} [e2e-llm-inference-service] f:app.kubernetes.io/name: {} [e2e-llm-inference-service] f:app.kubernetes.io/part-of: {} [e2e-llm-inference-service] f:llm-d.ai/managed: {} [e2e-llm-inference-service] f:ownerReferences: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"uid":"88c60d0e-95bd-401f-ae1c-2c4c0b1e7c9d"}: {} [e2e-llm-inference-service] f:spec: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:exportTo: {} [e2e-llm-inference-service] f:host: {} [e2e-llm-inference-service] f:trafficPolicy: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:tls: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:caCertificates: {} [e2e-llm-inference-service] f:insecureSkipVerify: {} [e2e-llm-inference-service] f:mode: {} [e2e-llm-inference-service] f:sni: {} [e2e-llm-inference-service] manager: manager [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] time: '2026-07-07T14:31:41Z' [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-route-87882a8e-kserve-workload-svc [e2e-llm-inference-service] namespace: kserve-ci-e2e-test [e2e-llm-inference-service] ownerReferences: [e2e-llm-inference-service] - apiVersion: serving.kserve.io/v1alpha2 [e2e-llm-inference-service] blockOwnerDeletion: true [e2e-llm-inference-service] controller: true [e2e-llm-inference-service] kind: LLMInferenceService [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-route-87882a8e [e2e-llm-inference-service] uid: 88c60d0e-95bd-401f-ae1c-2c4c0b1e7c9d [e2e-llm-inference-service] resourceVersion: '31046' [e2e-llm-inference-service] uid: dd2603eb-b21f-4a15-b5a5-45dd42d3e4bf [e2e-llm-inference-service] spec: [e2e-llm-inference-service] exportTo: [e2e-llm-inference-service] - '*' [e2e-llm-inference-service] host: llmisvc-model-fb-opt-125m-route-87882a8e-kserve-workload-svc.kserve-ci-e2e-test.svc.cluster.local [e2e-llm-inference-service] trafficPolicy: [e2e-llm-inference-service] tls: [e2e-llm-inference-service] caCertificates: /var/run/secrets/kubernetes.io/serviceaccount/service-ca.crt [e2e-llm-inference-service] insecureSkipVerify: false [e2e-llm-inference-service] mode: SIMPLE [e2e-llm-inference-service] sni: llmisvc-model-fb-opt-125m-route-87882a8e-kserve-workload-svc.kserve-ci-e2e-test.svc.cluster.local [e2e-llm-inference-service] [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1254 --- [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1255 apiVersion: networking.istio.io/v1alpha3 [e2e-llm-inference-service] kind: DestinationRule [e2e-llm-inference-service] metadata: [e2e-llm-inference-service] creationTimestamp: '2026-07-07T14:31:40Z' [e2e-llm-inference-service] generation: 1 [e2e-llm-inference-service] labels: [e2e-llm-inference-service] app.kubernetes.io/component: llminferenceservice-router-scheduler [e2e-llm-inference-service] app.kubernetes.io/name: llmisvc-model-fb-opt-125m-route-87882a8e [e2e-llm-inference-service] app.kubernetes.io/part-of: llminferenceservice [e2e-llm-inference-service] llm-d.ai/managed: 'true' [e2e-llm-inference-service] managedFields: [e2e-llm-inference-service] - apiVersion: networking.istio.io/v1 [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:metadata: [e2e-llm-inference-service] f:labels: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:app.kubernetes.io/component: {} [e2e-llm-inference-service] f:app.kubernetes.io/name: {} [e2e-llm-inference-service] f:app.kubernetes.io/part-of: {} [e2e-llm-inference-service] f:llm-d.ai/managed: {} [e2e-llm-inference-service] f:ownerReferences: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"uid":"88c60d0e-95bd-401f-ae1c-2c4c0b1e7c9d"}: {} [e2e-llm-inference-service] f:spec: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:exportTo: {} [e2e-llm-inference-service] f:host: {} [e2e-llm-inference-service] f:trafficPolicy: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:tls: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:insecureSkipVerify: {} [e2e-llm-inference-service] f:mode: {} [e2e-llm-inference-service] f:sni: {} [e2e-llm-inference-service] manager: manager [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] time: '2026-07-07T14:31:40Z' [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-route-87882a8e-kserve-scheduler [e2e-llm-inference-service] namespace: kserve-ci-e2e-test [e2e-llm-inference-service] ownerReferences: [e2e-llm-inference-service] - apiVersion: serving.kserve.io/v1alpha2 [e2e-llm-inference-service] blockOwnerDeletion: true [e2e-llm-inference-service] controller: true [e2e-llm-inference-service] kind: LLMInferenceService [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-route-87882a8e [e2e-llm-inference-service] uid: 88c60d0e-95bd-401f-ae1c-2c4c0b1e7c9d [e2e-llm-inference-service] resourceVersion: '31037' [e2e-llm-inference-service] uid: 36ac8ef0-fac2-4479-a42e-21b95a707ea2 [e2e-llm-inference-service] spec: [e2e-llm-inference-service] exportTo: [e2e-llm-inference-service] - '*' [e2e-llm-inference-service] host: llmisvc-model-fb-opt-125m-route-87882a8e-epp-service.kserve-ci-e2e-test.svc.cluster.local [e2e-llm-inference-service] trafficPolicy: [e2e-llm-inference-service] tls: [e2e-llm-inference-service] insecureSkipVerify: true [e2e-llm-inference-service] mode: SIMPLE [e2e-llm-inference-service] sni: llmisvc-model-fb-opt-125m-route-87882a8e-epp-service.kserve-ci-e2e-test.svc.cluster.local [e2e-llm-inference-service] [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1254 --- [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1255 apiVersion: networking.istio.io/v1alpha3 [e2e-llm-inference-service] kind: DestinationRule [e2e-llm-inference-service] metadata: [e2e-llm-inference-service] creationTimestamp: '2026-07-07T14:31:46Z' [e2e-llm-inference-service] generation: 1 [e2e-llm-inference-service] labels: [e2e-llm-inference-service] app.kubernetes.io/component: llminferenceservice-shadow-service [e2e-llm-inference-service] app.kubernetes.io/name: llmisvc-model-fb-opt-125m-route-87882a8e [e2e-llm-inference-service] app.kubernetes.io/part-of: llminferenceservice [e2e-llm-inference-service] llm-d.ai/managed: 'true' [e2e-llm-inference-service] managedFields: [e2e-llm-inference-service] - apiVersion: networking.istio.io/v1 [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:metadata: [e2e-llm-inference-service] f:labels: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:app.kubernetes.io/component: {} [e2e-llm-inference-service] f:app.kubernetes.io/name: {} [e2e-llm-inference-service] f:app.kubernetes.io/part-of: {} [e2e-llm-inference-service] f:llm-d.ai/managed: {} [e2e-llm-inference-service] f:ownerReferences: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"uid":"88c60d0e-95bd-401f-ae1c-2c4c0b1e7c9d"}: {} [e2e-llm-inference-service] f:spec: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:exportTo: {} [e2e-llm-inference-service] f:host: {} [e2e-llm-inference-service] f:trafficPolicy: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:tls: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:insecureSkipVerify: {} [e2e-llm-inference-service] f:mode: {} [e2e-llm-inference-service] f:sni: {} [e2e-llm-inference-service] manager: manager [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] time: '2026-07-07T14:31:46Z' [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-route-87882a8e-kserve-shadow-svc [e2e-llm-inference-service] namespace: kserve-ci-e2e-test [e2e-llm-inference-service] ownerReferences: [e2e-llm-inference-service] - apiVersion: serving.kserve.io/v1alpha2 [e2e-llm-inference-service] blockOwnerDeletion: true [e2e-llm-inference-service] controller: true [e2e-llm-inference-service] kind: LLMInferenceService [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-route-87882a8e [e2e-llm-inference-service] uid: 88c60d0e-95bd-401f-ae1c-2c4c0b1e7c9d [e2e-llm-inference-service] resourceVersion: '31187' [e2e-llm-inference-service] uid: 42e3e102-30d1-4377-b348-f59f61d9fc50 [e2e-llm-inference-service] spec: [e2e-llm-inference-service] exportTo: [e2e-llm-inference-service] - '*' [e2e-llm-inference-service] host: llmisvc-model-fb-opt-125m-route-87882a8e-inference--ip-aaa905cb.kserve-ci-e2e-test.svc.cluster.local [e2e-llm-inference-service] trafficPolicy: [e2e-llm-inference-service] tls: [e2e-llm-inference-service] insecureSkipVerify: true [e2e-llm-inference-service] mode: SIMPLE [e2e-llm-inference-service] sni: llmisvc-model-fb-opt-125m-route-87882a8e-kserve-workload-svc.kserve-ci-e2e-test.svc.cluster.local [e2e-llm-inference-service] [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1254 --- [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1255 apiVersion: networking.istio.io/v1alpha3 [e2e-llm-inference-service] kind: DestinationRule [e2e-llm-inference-service] metadata: [e2e-llm-inference-service] creationTimestamp: '2026-07-07T14:31:41Z' [e2e-llm-inference-service] generation: 1 [e2e-llm-inference-service] labels: [e2e-llm-inference-service] app.kubernetes.io/component: llminferenceservice-workload [e2e-llm-inference-service] app.kubernetes.io/name: llmisvc-model-fb-opt-125m-route-87882a8e [e2e-llm-inference-service] app.kubernetes.io/part-of: llminferenceservice [e2e-llm-inference-service] llm-d.ai/managed: 'true' [e2e-llm-inference-service] managedFields: [e2e-llm-inference-service] - apiVersion: networking.istio.io/v1 [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:metadata: [e2e-llm-inference-service] f:labels: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:app.kubernetes.io/component: {} [e2e-llm-inference-service] f:app.kubernetes.io/name: {} [e2e-llm-inference-service] f:app.kubernetes.io/part-of: {} [e2e-llm-inference-service] f:llm-d.ai/managed: {} [e2e-llm-inference-service] f:ownerReferences: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"uid":"88c60d0e-95bd-401f-ae1c-2c4c0b1e7c9d"}: {} [e2e-llm-inference-service] f:spec: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:exportTo: {} [e2e-llm-inference-service] f:host: {} [e2e-llm-inference-service] f:trafficPolicy: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:tls: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:caCertificates: {} [e2e-llm-inference-service] f:insecureSkipVerify: {} [e2e-llm-inference-service] f:mode: {} [e2e-llm-inference-service] f:sni: {} [e2e-llm-inference-service] manager: manager [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] time: '2026-07-07T14:31:41Z' [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-route-87882a8e-kserve-workload-svc [e2e-llm-inference-service] namespace: kserve-ci-e2e-test [e2e-llm-inference-service] ownerReferences: [e2e-llm-inference-service] - apiVersion: serving.kserve.io/v1alpha2 [e2e-llm-inference-service] blockOwnerDeletion: true [e2e-llm-inference-service] controller: true [e2e-llm-inference-service] kind: LLMInferenceService [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-route-87882a8e [e2e-llm-inference-service] uid: 88c60d0e-95bd-401f-ae1c-2c4c0b1e7c9d [e2e-llm-inference-service] resourceVersion: '31046' [e2e-llm-inference-service] uid: dd2603eb-b21f-4a15-b5a5-45dd42d3e4bf [e2e-llm-inference-service] spec: [e2e-llm-inference-service] exportTo: [e2e-llm-inference-service] - '*' [e2e-llm-inference-service] host: llmisvc-model-fb-opt-125m-route-87882a8e-kserve-workload-svc.kserve-ci-e2e-test.svc.cluster.local [e2e-llm-inference-service] trafficPolicy: [e2e-llm-inference-service] tls: [e2e-llm-inference-service] caCertificates: /var/run/secrets/kubernetes.io/serviceaccount/service-ca.crt [e2e-llm-inference-service] insecureSkipVerify: false [e2e-llm-inference-service] mode: SIMPLE [e2e-llm-inference-service] sni: llmisvc-model-fb-opt-125m-route-87882a8e-kserve-workload-svc.kserve-ci-e2e-test.svc.cluster.local [e2e-llm-inference-service] [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1254 --- [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1255 apiVersion: inference.networking.x-k8s.io/v1alpha2 [e2e-llm-inference-service] kind: InferencePool [e2e-llm-inference-service] metadata: [e2e-llm-inference-service] creationTimestamp: '2026-07-07T14:31:40Z' [e2e-llm-inference-service] generation: 1 [e2e-llm-inference-service] labels: [e2e-llm-inference-service] app.kubernetes.io/component: llminferenceservice-router-scheduler [e2e-llm-inference-service] app.kubernetes.io/name: llmisvc-model-fb-opt-125m-route-87882a8e [e2e-llm-inference-service] app.kubernetes.io/part-of: llminferenceservice [e2e-llm-inference-service] managedFields: [e2e-llm-inference-service] - apiVersion: inference.networking.x-k8s.io/v1alpha2 [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:metadata: [e2e-llm-inference-service] f:labels: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:app.kubernetes.io/component: {} [e2e-llm-inference-service] f:app.kubernetes.io/name: {} [e2e-llm-inference-service] f:app.kubernetes.io/part-of: {} [e2e-llm-inference-service] f:ownerReferences: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"uid":"88c60d0e-95bd-401f-ae1c-2c4c0b1e7c9d"}: {} [e2e-llm-inference-service] f:spec: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:extensionRef: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:failureMode: {} [e2e-llm-inference-service] f:group: {} [e2e-llm-inference-service] f:kind: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:portNumber: {} [e2e-llm-inference-service] f:selector: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:app.kubernetes.io/name: {} [e2e-llm-inference-service] f:app.kubernetes.io/part-of: {} [e2e-llm-inference-service] f:kserve.io/component: {} [e2e-llm-inference-service] f:targetPortNumber: {} [e2e-llm-inference-service] manager: manager [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] time: '2026-07-07T14:31:40Z' [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-route-87882a8e-inference-pool [e2e-llm-inference-service] namespace: kserve-ci-e2e-test [e2e-llm-inference-service] ownerReferences: [e2e-llm-inference-service] - apiVersion: serving.kserve.io/v1alpha2 [e2e-llm-inference-service] blockOwnerDeletion: true [e2e-llm-inference-service] controller: true [e2e-llm-inference-service] kind: LLMInferenceService [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-route-87882a8e [e2e-llm-inference-service] uid: 88c60d0e-95bd-401f-ae1c-2c4c0b1e7c9d [e2e-llm-inference-service] resourceVersion: '31003' [e2e-llm-inference-service] uid: 1c77c5b4-240e-49c7-8a8c-81b1bad17253 [e2e-llm-inference-service] spec: [e2e-llm-inference-service] extensionRef: [e2e-llm-inference-service] failureMode: FailOpen [e2e-llm-inference-service] group: '' [e2e-llm-inference-service] kind: Service [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-route-87882a8e-epp-service [e2e-llm-inference-service] portNumber: 9002 [e2e-llm-inference-service] selector: [e2e-llm-inference-service] app.kubernetes.io/name: llmisvc-model-fb-opt-125m-route-87882a8e [e2e-llm-inference-service] app.kubernetes.io/part-of: llminferenceservice [e2e-llm-inference-service] kserve.io/component: workload [e2e-llm-inference-service] targetPortNumber: 8000 [e2e-llm-inference-service] status: [e2e-llm-inference-service] parent: [e2e-llm-inference-service] - conditions: [e2e-llm-inference-service] - lastTransitionTime: '1970-01-01T00:00:00Z' [e2e-llm-inference-service] message: Waiting for controller [e2e-llm-inference-service] reason: Pending [e2e-llm-inference-service] status: Unknown [e2e-llm-inference-service] type: Accepted [e2e-llm-inference-service] parentRef: [e2e-llm-inference-service] group: gateway.networking.k8s.io [e2e-llm-inference-service] kind: Status [e2e-llm-inference-service] name: default [e2e-llm-inference-service] [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1254 --- [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1255 metadata: [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-route-87882a8e-kserve-5bd7864c6ctnql9 [e2e-llm-inference-service] namespace: kserve-ci-e2e-test [e2e-llm-inference-service] creationTimestamp: '2026-07-07T14:48:47Z' [e2e-llm-inference-service] labels: [e2e-llm-inference-service] app.kubernetes.io/component: llminferenceservice-workload [e2e-llm-inference-service] app.kubernetes.io/name: llmisvc-model-fb-opt-125m-route-87882a8e [e2e-llm-inference-service] app.kubernetes.io/part-of: llminferenceservice [e2e-llm-inference-service] kserve.io/component: workload [e2e-llm-inference-service] llm-d.ai/role: both [e2e-llm-inference-service] pod-template-hash: 5bd7864c6c [e2e-llm-inference-service] timestamp: '2026-07-07T14:48:35Z' [e2e-llm-inference-service] window: 10.068s [e2e-llm-inference-service] containers: [e2e-llm-inference-service] - name: main [e2e-llm-inference-service] usage: [e2e-llm-inference-service] cpu: 100342173n [e2e-llm-inference-service] memory: 2400744Ki [e2e-llm-inference-service] apiVersion: metrics.k8s.io/v1beta1 [e2e-llm-inference-service] kind: PodMetrics [e2e-llm-inference-service] [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1254 --- [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1255 metadata: [e2e-llm-inference-service] name: llmisvca690bbc929faec8bc98c767f16c003c1-kserve-router-scheq44kv [e2e-llm-inference-service] namespace: kserve-ci-e2e-test [e2e-llm-inference-service] creationTimestamp: '2026-07-07T14:48:47Z' [e2e-llm-inference-service] labels: [e2e-llm-inference-service] app.kubernetes.io/component: llminferenceservice-router-scheduler [e2e-llm-inference-service] app.kubernetes.io/name: llmisvc-model-fb-opt-125m-route-87882a8e [e2e-llm-inference-service] app.kubernetes.io/part-of: llminferenceservice [e2e-llm-inference-service] pod-template-hash: 5d6998448f [e2e-llm-inference-service] timestamp: '2026-07-07T14:48:29Z' [e2e-llm-inference-service] window: 11.227s [e2e-llm-inference-service] containers: [e2e-llm-inference-service] - name: main [e2e-llm-inference-service] usage: [e2e-llm-inference-service] cpu: 58855615n [e2e-llm-inference-service] memory: 31180Ki [e2e-llm-inference-service] apiVersion: metrics.k8s.io/v1beta1 [e2e-llm-inference-service] kind: PodMetrics [e2e-llm-inference-service] [e2e-llm-inference-service] ERROR e2e.llmisvc.logging:logging.py:48 [test_llm_inference_service] [2026-07-07T14:48:48.082266] end - ❌ 1032.247s: ❌ Failed to call model: HTTPConnectionPool(host='ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com', port=80): Max retries exceeded with url: /kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-route-87882a8e/v1/completions (Caused by ReadTimeoutError("HTTPConnectionPool(host='ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com', port=80): Read timed out. (read timeout=60)")) [e2e-llm-inference-service] _ test_llm_inference_service[router-managed-workload-single-cpu-model-fb-opt-125m-with-lora-hf1] _ [e2e-llm-inference-service] [gw1] linux -- Python 3.11.13 /workspace/source/python/kserve/.venv/bin/python [e2e-llm-inference-service] [e2e-llm-inference-service] self = [e2e-llm-inference-service] conn = [e2e-llm-inference-service] method = 'GET', url = '/v1/models', body = None [e2e-llm-inference-service] headers = {'User-Agent': 'python-requests/2.32.3', 'Accept-Encoding': 'gzip, deflate', 'Accept': '*/*', 'Connection': 'keep-alive', 'Content-Type': 'application/json', 'X-Gateway-Model-Name': 'publishers/kserve-ci-e2e-test/models/facebook/opt-125m'} [e2e-llm-inference-service] retries = Retry(total=0, connect=None, read=None, redirect=None, status=None) [e2e-llm-inference-service] timeout = Timeout(connect=60, read=60, total=None), chunked = False [e2e-llm-inference-service] response_conn = [e2e-llm-inference-service] preload_content = False, decode_content = False, enforce_content_length = True [e2e-llm-inference-service] [e2e-llm-inference-service] def _make_request( [e2e-llm-inference-service] self, [e2e-llm-inference-service] conn: BaseHTTPConnection, [e2e-llm-inference-service] method: str, [e2e-llm-inference-service] url: str, [e2e-llm-inference-service] body: _TYPE_BODY | None = None, [e2e-llm-inference-service] headers: typing.Mapping[str, str] | None = None, [e2e-llm-inference-service] retries: Retry | None = None, [e2e-llm-inference-service] timeout: _TYPE_TIMEOUT = _DEFAULT_TIMEOUT, [e2e-llm-inference-service] chunked: bool = False, [e2e-llm-inference-service] response_conn: BaseHTTPConnection | None = None, [e2e-llm-inference-service] preload_content: bool = True, [e2e-llm-inference-service] decode_content: bool = True, [e2e-llm-inference-service] enforce_content_length: bool = True, [e2e-llm-inference-service] ) -> BaseHTTPResponse: [e2e-llm-inference-service] """ [e2e-llm-inference-service] Perform a request on a given urllib connection object taken from our [e2e-llm-inference-service] pool. [e2e-llm-inference-service] [e2e-llm-inference-service] :param conn: [e2e-llm-inference-service] a connection from one of our connection pools [e2e-llm-inference-service] [e2e-llm-inference-service] :param method: [e2e-llm-inference-service] HTTP request method (such as GET, POST, PUT, etc.) [e2e-llm-inference-service] [e2e-llm-inference-service] :param url: [e2e-llm-inference-service] The URL to perform the request on. [e2e-llm-inference-service] [e2e-llm-inference-service] :param body: [e2e-llm-inference-service] Data to send in the request body, either :class:`str`, :class:`bytes`, [e2e-llm-inference-service] an iterable of :class:`str`/:class:`bytes`, or a file-like object. [e2e-llm-inference-service] [e2e-llm-inference-service] :param headers: [e2e-llm-inference-service] Dictionary of custom headers to send, such as User-Agent, [e2e-llm-inference-service] If-None-Match, etc. If None, pool headers are used. If provided, [e2e-llm-inference-service] these headers completely replace any pool-specific headers. [e2e-llm-inference-service] [e2e-llm-inference-service] :param retries: [e2e-llm-inference-service] Configure the number of retries to allow before raising a [e2e-llm-inference-service] :class:`~urllib3.exceptions.MaxRetryError` exception. [e2e-llm-inference-service] [e2e-llm-inference-service] Pass ``None`` to retry until you receive a response. Pass a [e2e-llm-inference-service] :class:`~urllib3.util.retry.Retry` object for fine-grained control [e2e-llm-inference-service] over different types of retries. [e2e-llm-inference-service] Pass an integer number to retry connection errors that many times, [e2e-llm-inference-service] but no other types of errors. Pass zero to never retry. [e2e-llm-inference-service] [e2e-llm-inference-service] If ``False``, then retries are disabled and any exception is raised [e2e-llm-inference-service] immediately. Also, instead of raising a MaxRetryError on redirects, [e2e-llm-inference-service] the redirect response will be returned. [e2e-llm-inference-service] [e2e-llm-inference-service] :type retries: :class:`~urllib3.util.retry.Retry`, False, or an int. [e2e-llm-inference-service] [e2e-llm-inference-service] :param timeout: [e2e-llm-inference-service] If specified, overrides the default timeout for this one [e2e-llm-inference-service] request. It may be a float (in seconds) or an instance of [e2e-llm-inference-service] :class:`urllib3.util.Timeout`. [e2e-llm-inference-service] [e2e-llm-inference-service] :param chunked: [e2e-llm-inference-service] If True, urllib3 will send the body using chunked transfer [e2e-llm-inference-service] encoding. Otherwise, urllib3 will send the body using the standard [e2e-llm-inference-service] content-length form. Defaults to False. [e2e-llm-inference-service] [e2e-llm-inference-service] :param response_conn: [e2e-llm-inference-service] Set this to ``None`` if you will handle releasing the connection or [e2e-llm-inference-service] set the connection to have the response release it. [e2e-llm-inference-service] [e2e-llm-inference-service] :param preload_content: [e2e-llm-inference-service] If True, the response's body will be preloaded during construction. [e2e-llm-inference-service] [e2e-llm-inference-service] :param decode_content: [e2e-llm-inference-service] If True, will attempt to decode the body based on the [e2e-llm-inference-service] 'content-encoding' header. [e2e-llm-inference-service] [e2e-llm-inference-service] :param enforce_content_length: [e2e-llm-inference-service] Enforce content length checking. Body returned by server must match [e2e-llm-inference-service] value of Content-Length header, if present. Otherwise, raise error. [e2e-llm-inference-service] """ [e2e-llm-inference-service] self.num_requests += 1 [e2e-llm-inference-service] [e2e-llm-inference-service] timeout_obj = self._get_timeout(timeout) [e2e-llm-inference-service] timeout_obj.start_connect() [e2e-llm-inference-service] conn.timeout = Timeout.resolve_default_timeout(timeout_obj.connect_timeout) [e2e-llm-inference-service] [e2e-llm-inference-service] try: [e2e-llm-inference-service] # Trigger any extra validation we need to do. [e2e-llm-inference-service] try: [e2e-llm-inference-service] self._validate_conn(conn) [e2e-llm-inference-service] except (SocketTimeout, BaseSSLError) as e: [e2e-llm-inference-service] self._raise_timeout(err=e, url=url, timeout_value=conn.timeout) [e2e-llm-inference-service] raise [e2e-llm-inference-service] [e2e-llm-inference-service] # _validate_conn() starts the connection to an HTTPS proxy [e2e-llm-inference-service] # so we need to wrap errors with 'ProxyError' here too. [e2e-llm-inference-service] except ( [e2e-llm-inference-service] OSError, [e2e-llm-inference-service] NewConnectionError, [e2e-llm-inference-service] TimeoutError, [e2e-llm-inference-service] BaseSSLError, [e2e-llm-inference-service] CertificateError, [e2e-llm-inference-service] SSLError, [e2e-llm-inference-service] ) as e: [e2e-llm-inference-service] new_e: Exception = e [e2e-llm-inference-service] if isinstance(e, (BaseSSLError, CertificateError)): [e2e-llm-inference-service] new_e = SSLError(e) [e2e-llm-inference-service] # If the connection didn't successfully connect to it's proxy [e2e-llm-inference-service] # then there [e2e-llm-inference-service] if isinstance( [e2e-llm-inference-service] new_e, (OSError, NewConnectionError, TimeoutError, SSLError) [e2e-llm-inference-service] ) and (conn and conn.proxy and not conn.has_connected_to_proxy): [e2e-llm-inference-service] new_e = _wrap_proxy_error(new_e, conn.proxy.scheme) [e2e-llm-inference-service] raise new_e [e2e-llm-inference-service] [e2e-llm-inference-service] # conn.request() calls http.client.*.request, not the method in [e2e-llm-inference-service] # urllib3.request. It also calls makefile (recv) on the socket. [e2e-llm-inference-service] try: [e2e-llm-inference-service] conn.request( [e2e-llm-inference-service] method, [e2e-llm-inference-service] url, [e2e-llm-inference-service] body=body, [e2e-llm-inference-service] headers=headers, [e2e-llm-inference-service] chunked=chunked, [e2e-llm-inference-service] preload_content=preload_content, [e2e-llm-inference-service] decode_content=decode_content, [e2e-llm-inference-service] enforce_content_length=enforce_content_length, [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] # We are swallowing BrokenPipeError (errno.EPIPE) since the server is [e2e-llm-inference-service] # legitimately able to close the connection after sending a valid response. [e2e-llm-inference-service] # With this behaviour, the received response is still readable. [e2e-llm-inference-service] except BrokenPipeError: [e2e-llm-inference-service] pass [e2e-llm-inference-service] except OSError as e: [e2e-llm-inference-service] # MacOS/Linux [e2e-llm-inference-service] # EPROTOTYPE and ECONNRESET are needed on macOS [e2e-llm-inference-service] # https://erickt.github.io/blog/2014/11/19/adventures-in-debugging-a-potential-osx-kernel-bug/ [e2e-llm-inference-service] # Condition changed later to emit ECONNRESET instead of only EPROTOTYPE. [e2e-llm-inference-service] if e.errno != errno.EPROTOTYPE and e.errno != errno.ECONNRESET: [e2e-llm-inference-service] raise [e2e-llm-inference-service] [e2e-llm-inference-service] # Reset the timeout for the recv() on the socket [e2e-llm-inference-service] read_timeout = timeout_obj.read_timeout [e2e-llm-inference-service] [e2e-llm-inference-service] if not conn.is_closed: [e2e-llm-inference-service] # In Python 3 socket.py will catch EAGAIN and return None when you [e2e-llm-inference-service] # try and read into the file pointer created by http.client, which [e2e-llm-inference-service] # instead raises a BadStatusLine exception. Instead of catching [e2e-llm-inference-service] # the exception and assuming all BadStatusLine exceptions are read [e2e-llm-inference-service] # timeouts, check for a zero timeout before making the request. [e2e-llm-inference-service] if read_timeout == 0: [e2e-llm-inference-service] raise ReadTimeoutError( [e2e-llm-inference-service] self, url, f"Read timed out. (read timeout={read_timeout})" [e2e-llm-inference-service] ) [e2e-llm-inference-service] conn.timeout = read_timeout [e2e-llm-inference-service] [e2e-llm-inference-service] # Receive the response from the server [e2e-llm-inference-service] try: [e2e-llm-inference-service] > response = conn.getresponse() [e2e-llm-inference-service] [e2e-llm-inference-service] ../../python/kserve/.venv/lib64/python3.11/site-packages/urllib3/connectionpool.py:534: [e2e-llm-inference-service] _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ [e2e-llm-inference-service] [e2e-llm-inference-service] self = [e2e-llm-inference-service] [e2e-llm-inference-service] def getresponse( # type: ignore[override] [e2e-llm-inference-service] self, [e2e-llm-inference-service] ) -> HTTPResponse: [e2e-llm-inference-service] """ [e2e-llm-inference-service] Get the response from the server. [e2e-llm-inference-service] [e2e-llm-inference-service] If the HTTPConnection is in the correct state, returns an instance of HTTPResponse or of whatever object is returned by the response_class variable. [e2e-llm-inference-service] [e2e-llm-inference-service] If a request has not been sent or if a previous response has not be handled, ResponseNotReady is raised. If the HTTP response indicates that the connection should be closed, then it will be closed before the response is returned. When the connection is closed, the underlying socket is closed. [e2e-llm-inference-service] """ [e2e-llm-inference-service] # Raise the same error as http.client.HTTPConnection [e2e-llm-inference-service] if self._response_options is None: [e2e-llm-inference-service] raise ResponseNotReady() [e2e-llm-inference-service] [e2e-llm-inference-service] # Reset this attribute for being used again. [e2e-llm-inference-service] resp_options = self._response_options [e2e-llm-inference-service] self._response_options = None [e2e-llm-inference-service] [e2e-llm-inference-service] # Since the connection's timeout value may have been updated [e2e-llm-inference-service] # we need to set the timeout on the socket. [e2e-llm-inference-service] self.sock.settimeout(self.timeout) [e2e-llm-inference-service] [e2e-llm-inference-service] # This is needed here to avoid circular import errors [e2e-llm-inference-service] from .response import HTTPResponse [e2e-llm-inference-service] [e2e-llm-inference-service] # Save a reference to the shutdown function before ownership is passed [e2e-llm-inference-service] # to httplib_response [e2e-llm-inference-service] # TODO should we implement it everywhere? [e2e-llm-inference-service] _shutdown = getattr(self.sock, "shutdown", None) [e2e-llm-inference-service] [e2e-llm-inference-service] # Get the response from http.client.HTTPConnection [e2e-llm-inference-service] > httplib_response = super().getresponse() [e2e-llm-inference-service] [e2e-llm-inference-service] ../../python/kserve/.venv/lib64/python3.11/site-packages/urllib3/connection.py:571: [e2e-llm-inference-service] _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ [e2e-llm-inference-service] [e2e-llm-inference-service] self = [e2e-llm-inference-service] [e2e-llm-inference-service] def getresponse(self): [e2e-llm-inference-service] """Get the response from the server. [e2e-llm-inference-service] [e2e-llm-inference-service] If the HTTPConnection is in the correct state, returns an [e2e-llm-inference-service] instance of HTTPResponse or of whatever object is returned by [e2e-llm-inference-service] the response_class variable. [e2e-llm-inference-service] [e2e-llm-inference-service] If a request has not been sent or if a previous response has [e2e-llm-inference-service] not be handled, ResponseNotReady is raised. If the HTTP [e2e-llm-inference-service] response indicates that the connection should be closed, then [e2e-llm-inference-service] it will be closed before the response is returned. When the [e2e-llm-inference-service] connection is closed, the underlying socket is closed. [e2e-llm-inference-service] """ [e2e-llm-inference-service] [e2e-llm-inference-service] # if a prior response has been completed, then forget about it. [e2e-llm-inference-service] if self.__response and self.__response.isclosed(): [e2e-llm-inference-service] self.__response = None [e2e-llm-inference-service] [e2e-llm-inference-service] # if a prior response exists, then it must be completed (otherwise, we [e2e-llm-inference-service] # cannot read this response's header to determine the connection-close [e2e-llm-inference-service] # behavior) [e2e-llm-inference-service] # [e2e-llm-inference-service] # note: if a prior response existed, but was connection-close, then the [e2e-llm-inference-service] # socket and response were made independent of this HTTPConnection [e2e-llm-inference-service] # object since a new request requires that we open a whole new [e2e-llm-inference-service] # connection [e2e-llm-inference-service] # [e2e-llm-inference-service] # this means the prior response had one of two states: [e2e-llm-inference-service] # 1) will_close: this connection was reset and the prior socket and [e2e-llm-inference-service] # response operate independently [e2e-llm-inference-service] # 2) persistent: the response was retained and we await its [e2e-llm-inference-service] # isclosed() status to become true. [e2e-llm-inference-service] # [e2e-llm-inference-service] if self.__state != _CS_REQ_SENT or self.__response: [e2e-llm-inference-service] raise ResponseNotReady(self.__state) [e2e-llm-inference-service] [e2e-llm-inference-service] if self.debuglevel > 0: [e2e-llm-inference-service] response = self.response_class(self.sock, self.debuglevel, [e2e-llm-inference-service] method=self._method) [e2e-llm-inference-service] else: [e2e-llm-inference-service] response = self.response_class(self.sock, method=self._method) [e2e-llm-inference-service] [e2e-llm-inference-service] try: [e2e-llm-inference-service] try: [e2e-llm-inference-service] > response.begin() [e2e-llm-inference-service] [e2e-llm-inference-service] /usr/lib64/python3.11/http/client.py:1395: [e2e-llm-inference-service] _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ [e2e-llm-inference-service] [e2e-llm-inference-service] self = [e2e-llm-inference-service] [e2e-llm-inference-service] def begin(self): [e2e-llm-inference-service] if self.headers is not None: [e2e-llm-inference-service] # we've already started reading the response [e2e-llm-inference-service] return [e2e-llm-inference-service] [e2e-llm-inference-service] # read until we get a non-100 response [e2e-llm-inference-service] while True: [e2e-llm-inference-service] > version, status, reason = self._read_status() [e2e-llm-inference-service] [e2e-llm-inference-service] /usr/lib64/python3.11/http/client.py:325: [e2e-llm-inference-service] _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ [e2e-llm-inference-service] [e2e-llm-inference-service] self = [e2e-llm-inference-service] [e2e-llm-inference-service] def _read_status(self): [e2e-llm-inference-service] > line = str(self.fp.readline(_MAXLINE + 1), "iso-8859-1") [e2e-llm-inference-service] [e2e-llm-inference-service] /usr/lib64/python3.11/http/client.py:286: [e2e-llm-inference-service] _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ [e2e-llm-inference-service] [e2e-llm-inference-service] self = [e2e-llm-inference-service] b = [e2e-llm-inference-service] [e2e-llm-inference-service] def readinto(self, b): [e2e-llm-inference-service] """Read up to len(b) bytes into the writable buffer *b* and return [e2e-llm-inference-service] the number of bytes read. If the socket is non-blocking and no bytes [e2e-llm-inference-service] are available, None is returned. [e2e-llm-inference-service] [e2e-llm-inference-service] If *b* is non-empty, a 0 return value indicates that the connection [e2e-llm-inference-service] was shutdown at the other end. [e2e-llm-inference-service] """ [e2e-llm-inference-service] self._checkClosed() [e2e-llm-inference-service] self._checkReadable() [e2e-llm-inference-service] if self._timeout_occurred: [e2e-llm-inference-service] raise OSError("cannot read from timed out object") [e2e-llm-inference-service] while True: [e2e-llm-inference-service] try: [e2e-llm-inference-service] > return self._sock.recv_into(b) [e2e-llm-inference-service] E TimeoutError: timed out [e2e-llm-inference-service] [e2e-llm-inference-service] /usr/lib64/python3.11/socket.py:718: TimeoutError [e2e-llm-inference-service] [e2e-llm-inference-service] The above exception was the direct cause of the following exception: [e2e-llm-inference-service] [e2e-llm-inference-service] self = [e2e-llm-inference-service] method = 'GET', url = '/v1/models', body = None [e2e-llm-inference-service] headers = {'User-Agent': 'python-requests/2.32.3', 'Accept-Encoding': 'gzip, deflate', 'Accept': '*/*', 'Connection': 'keep-alive', 'Content-Type': 'application/json', 'X-Gateway-Model-Name': 'publishers/kserve-ci-e2e-test/models/facebook/opt-125m'} [e2e-llm-inference-service] retries = Retry(total=0, connect=None, read=None, redirect=None, status=None) [e2e-llm-inference-service] redirect = False, assert_same_host = False [e2e-llm-inference-service] timeout = Timeout(connect=60, read=60, total=None), pool_timeout = None [e2e-llm-inference-service] release_conn = False, chunked = False, body_pos = None, preload_content = False [e2e-llm-inference-service] decode_content = False, response_kw = {} [e2e-llm-inference-service] parsed_url = Url(scheme=None, auth=None, host=None, port=None, path='/v1/models', query=None, fragment=None) [e2e-llm-inference-service] destination_scheme = None, conn = None, release_this_conn = True [e2e-llm-inference-service] http_tunnel_required = False, err = None, clean_exit = False [e2e-llm-inference-service] [e2e-llm-inference-service] def urlopen( # type: ignore[override] [e2e-llm-inference-service] self, [e2e-llm-inference-service] method: str, [e2e-llm-inference-service] url: str, [e2e-llm-inference-service] body: _TYPE_BODY | None = None, [e2e-llm-inference-service] headers: typing.Mapping[str, str] | None = None, [e2e-llm-inference-service] retries: Retry | bool | int | None = None, [e2e-llm-inference-service] redirect: bool = True, [e2e-llm-inference-service] assert_same_host: bool = True, [e2e-llm-inference-service] timeout: _TYPE_TIMEOUT = _DEFAULT_TIMEOUT, [e2e-llm-inference-service] pool_timeout: int | None = None, [e2e-llm-inference-service] release_conn: bool | None = None, [e2e-llm-inference-service] chunked: bool = False, [e2e-llm-inference-service] body_pos: _TYPE_BODY_POSITION | None = None, [e2e-llm-inference-service] preload_content: bool = True, [e2e-llm-inference-service] decode_content: bool = True, [e2e-llm-inference-service] **response_kw: typing.Any, [e2e-llm-inference-service] ) -> BaseHTTPResponse: [e2e-llm-inference-service] """ [e2e-llm-inference-service] Get a connection from the pool and perform an HTTP request. This is the [e2e-llm-inference-service] lowest level call for making a request, so you'll need to specify all [e2e-llm-inference-service] the raw details. [e2e-llm-inference-service] [e2e-llm-inference-service] .. note:: [e2e-llm-inference-service] [e2e-llm-inference-service] More commonly, it's appropriate to use a convenience method [e2e-llm-inference-service] such as :meth:`request`. [e2e-llm-inference-service] [e2e-llm-inference-service] .. note:: [e2e-llm-inference-service] [e2e-llm-inference-service] `release_conn` will only behave as expected if [e2e-llm-inference-service] `preload_content=False` because we want to make [e2e-llm-inference-service] `preload_content=False` the default behaviour someday soon without [e2e-llm-inference-service] breaking backwards compatibility. [e2e-llm-inference-service] [e2e-llm-inference-service] :param method: [e2e-llm-inference-service] HTTP request method (such as GET, POST, PUT, etc.) [e2e-llm-inference-service] [e2e-llm-inference-service] :param url: [e2e-llm-inference-service] The URL to perform the request on. [e2e-llm-inference-service] [e2e-llm-inference-service] :param body: [e2e-llm-inference-service] Data to send in the request body, either :class:`str`, :class:`bytes`, [e2e-llm-inference-service] an iterable of :class:`str`/:class:`bytes`, or a file-like object. [e2e-llm-inference-service] [e2e-llm-inference-service] :param headers: [e2e-llm-inference-service] Dictionary of custom headers to send, such as User-Agent, [e2e-llm-inference-service] If-None-Match, etc. If None, pool headers are used. If provided, [e2e-llm-inference-service] these headers completely replace any pool-specific headers. [e2e-llm-inference-service] [e2e-llm-inference-service] :param retries: [e2e-llm-inference-service] Configure the number of retries to allow before raising a [e2e-llm-inference-service] :class:`~urllib3.exceptions.MaxRetryError` exception. [e2e-llm-inference-service] [e2e-llm-inference-service] If ``None`` (default) will retry 3 times, see ``Retry.DEFAULT``. Pass a [e2e-llm-inference-service] :class:`~urllib3.util.retry.Retry` object for fine-grained control [e2e-llm-inference-service] over different types of retries. [e2e-llm-inference-service] Pass an integer number to retry connection errors that many times, [e2e-llm-inference-service] but no other types of errors. Pass zero to never retry. [e2e-llm-inference-service] [e2e-llm-inference-service] If ``False``, then retries are disabled and any exception is raised [e2e-llm-inference-service] immediately. Also, instead of raising a MaxRetryError on redirects, [e2e-llm-inference-service] the redirect response will be returned. [e2e-llm-inference-service] [e2e-llm-inference-service] :type retries: :class:`~urllib3.util.retry.Retry`, False, or an int. [e2e-llm-inference-service] [e2e-llm-inference-service] :param redirect: [e2e-llm-inference-service] If True, automatically handle redirects (status codes 301, 302, [e2e-llm-inference-service] 303, 307, 308). Each redirect counts as a retry. Disabling retries [e2e-llm-inference-service] will disable redirect, too. [e2e-llm-inference-service] [e2e-llm-inference-service] :param assert_same_host: [e2e-llm-inference-service] If ``True``, will make sure that the host of the pool requests is [e2e-llm-inference-service] consistent else will raise HostChangedError. When ``False``, you can [e2e-llm-inference-service] use the pool on an HTTP proxy and request foreign hosts. [e2e-llm-inference-service] [e2e-llm-inference-service] :param timeout: [e2e-llm-inference-service] If specified, overrides the default timeout for this one [e2e-llm-inference-service] request. It may be a float (in seconds) or an instance of [e2e-llm-inference-service] :class:`urllib3.util.Timeout`. [e2e-llm-inference-service] [e2e-llm-inference-service] :param pool_timeout: [e2e-llm-inference-service] If set and the pool is set to block=True, then this method will [e2e-llm-inference-service] block for ``pool_timeout`` seconds and raise EmptyPoolError if no [e2e-llm-inference-service] connection is available within the time period. [e2e-llm-inference-service] [e2e-llm-inference-service] :param bool preload_content: [e2e-llm-inference-service] If True, the response's body will be preloaded into memory. [e2e-llm-inference-service] [e2e-llm-inference-service] :param bool decode_content: [e2e-llm-inference-service] If True, will attempt to decode the body based on the [e2e-llm-inference-service] 'content-encoding' header. [e2e-llm-inference-service] [e2e-llm-inference-service] :param release_conn: [e2e-llm-inference-service] If False, then the urlopen call will not release the connection [e2e-llm-inference-service] back into the pool once a response is received (but will release if [e2e-llm-inference-service] you read the entire contents of the response such as when [e2e-llm-inference-service] `preload_content=True`). This is useful if you're not preloading [e2e-llm-inference-service] the response's content immediately. You will need to call [e2e-llm-inference-service] ``r.release_conn()`` on the response ``r`` to return the connection [e2e-llm-inference-service] back into the pool. If None, it takes the value of ``preload_content`` [e2e-llm-inference-service] which defaults to ``True``. [e2e-llm-inference-service] [e2e-llm-inference-service] :param bool chunked: [e2e-llm-inference-service] If True, urllib3 will send the body using chunked transfer [e2e-llm-inference-service] encoding. Otherwise, urllib3 will send the body using the standard [e2e-llm-inference-service] content-length form. Defaults to False. [e2e-llm-inference-service] [e2e-llm-inference-service] :param int body_pos: [e2e-llm-inference-service] Position to seek to in file-like body in the event of a retry or [e2e-llm-inference-service] redirect. Typically this won't need to be set because urllib3 will [e2e-llm-inference-service] auto-populate the value when needed. [e2e-llm-inference-service] """ [e2e-llm-inference-service] parsed_url = parse_url(url) [e2e-llm-inference-service] destination_scheme = parsed_url.scheme [e2e-llm-inference-service] [e2e-llm-inference-service] if headers is None: [e2e-llm-inference-service] headers = self.headers [e2e-llm-inference-service] [e2e-llm-inference-service] if not isinstance(retries, Retry): [e2e-llm-inference-service] retries = Retry.from_int(retries, redirect=redirect, default=self.retries) [e2e-llm-inference-service] [e2e-llm-inference-service] if release_conn is None: [e2e-llm-inference-service] release_conn = preload_content [e2e-llm-inference-service] [e2e-llm-inference-service] # Check host [e2e-llm-inference-service] if assert_same_host and not self.is_same_host(url): [e2e-llm-inference-service] raise HostChangedError(self, url, retries) [e2e-llm-inference-service] [e2e-llm-inference-service] # Ensure that the URL we're connecting to is properly encoded [e2e-llm-inference-service] if url.startswith("/"): [e2e-llm-inference-service] url = to_str(_encode_target(url)) [e2e-llm-inference-service] else: [e2e-llm-inference-service] url = to_str(parsed_url.url) [e2e-llm-inference-service] [e2e-llm-inference-service] conn = None [e2e-llm-inference-service] [e2e-llm-inference-service] # Track whether `conn` needs to be released before [e2e-llm-inference-service] # returning/raising/recursing. Update this variable if necessary, and [e2e-llm-inference-service] # leave `release_conn` constant throughout the function. That way, if [e2e-llm-inference-service] # the function recurses, the original value of `release_conn` will be [e2e-llm-inference-service] # passed down into the recursive call, and its value will be respected. [e2e-llm-inference-service] # [e2e-llm-inference-service] # See issue #651 [1] for details. [e2e-llm-inference-service] # [e2e-llm-inference-service] # [1] [e2e-llm-inference-service] release_this_conn = release_conn [e2e-llm-inference-service] [e2e-llm-inference-service] http_tunnel_required = connection_requires_http_tunnel( [e2e-llm-inference-service] self.proxy, self.proxy_config, destination_scheme [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] # Merge the proxy headers. Only done when not using HTTP CONNECT. We [e2e-llm-inference-service] # have to copy the headers dict so we can safely change it without those [e2e-llm-inference-service] # changes being reflected in anyone else's copy. [e2e-llm-inference-service] if not http_tunnel_required: [e2e-llm-inference-service] headers = headers.copy() # type: ignore[attr-defined] [e2e-llm-inference-service] headers.update(self.proxy_headers) # type: ignore[union-attr] [e2e-llm-inference-service] [e2e-llm-inference-service] # Must keep the exception bound to a separate variable or else Python 3 [e2e-llm-inference-service] # complains about UnboundLocalError. [e2e-llm-inference-service] err = None [e2e-llm-inference-service] [e2e-llm-inference-service] # Keep track of whether we cleanly exited the except block. This [e2e-llm-inference-service] # ensures we do proper cleanup in finally. [e2e-llm-inference-service] clean_exit = False [e2e-llm-inference-service] [e2e-llm-inference-service] # Rewind body position, if needed. Record current position [e2e-llm-inference-service] # for future rewinds in the event of a redirect/retry. [e2e-llm-inference-service] body_pos = set_file_position(body, body_pos) [e2e-llm-inference-service] [e2e-llm-inference-service] try: [e2e-llm-inference-service] # Request a connection from the queue. [e2e-llm-inference-service] timeout_obj = self._get_timeout(timeout) [e2e-llm-inference-service] conn = self._get_conn(timeout=pool_timeout) [e2e-llm-inference-service] [e2e-llm-inference-service] conn.timeout = timeout_obj.connect_timeout # type: ignore[assignment] [e2e-llm-inference-service] [e2e-llm-inference-service] # Is this a closed/new connection that requires CONNECT tunnelling? [e2e-llm-inference-service] if self.proxy is not None and http_tunnel_required and conn.is_closed: [e2e-llm-inference-service] try: [e2e-llm-inference-service] self._prepare_proxy(conn) [e2e-llm-inference-service] except (BaseSSLError, OSError, SocketTimeout) as e: [e2e-llm-inference-service] self._raise_timeout( [e2e-llm-inference-service] err=e, url=self.proxy.url, timeout_value=conn.timeout [e2e-llm-inference-service] ) [e2e-llm-inference-service] raise [e2e-llm-inference-service] [e2e-llm-inference-service] # If we're going to release the connection in ``finally:``, then [e2e-llm-inference-service] # the response doesn't need to know about the connection. Otherwise [e2e-llm-inference-service] # it will also try to release it and we'll have a double-release [e2e-llm-inference-service] # mess. [e2e-llm-inference-service] response_conn = conn if not release_conn else None [e2e-llm-inference-service] [e2e-llm-inference-service] # Make the request on the HTTPConnection object [e2e-llm-inference-service] > response = self._make_request( [e2e-llm-inference-service] conn, [e2e-llm-inference-service] method, [e2e-llm-inference-service] url, [e2e-llm-inference-service] timeout=timeout_obj, [e2e-llm-inference-service] body=body, [e2e-llm-inference-service] headers=headers, [e2e-llm-inference-service] chunked=chunked, [e2e-llm-inference-service] retries=retries, [e2e-llm-inference-service] response_conn=response_conn, [e2e-llm-inference-service] preload_content=preload_content, [e2e-llm-inference-service] decode_content=decode_content, [e2e-llm-inference-service] **response_kw, [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] ../../python/kserve/.venv/lib64/python3.11/site-packages/urllib3/connectionpool.py:787: [e2e-llm-inference-service] _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ [e2e-llm-inference-service] [e2e-llm-inference-service] self = [e2e-llm-inference-service] conn = [e2e-llm-inference-service] method = 'GET', url = '/v1/models', body = None [e2e-llm-inference-service] headers = {'User-Agent': 'python-requests/2.32.3', 'Accept-Encoding': 'gzip, deflate', 'Accept': '*/*', 'Connection': 'keep-alive', 'Content-Type': 'application/json', 'X-Gateway-Model-Name': 'publishers/kserve-ci-e2e-test/models/facebook/opt-125m'} [e2e-llm-inference-service] retries = Retry(total=0, connect=None, read=None, redirect=None, status=None) [e2e-llm-inference-service] timeout = Timeout(connect=60, read=60, total=None), chunked = False [e2e-llm-inference-service] response_conn = [e2e-llm-inference-service] preload_content = False, decode_content = False, enforce_content_length = True [e2e-llm-inference-service] [e2e-llm-inference-service] def _make_request( [e2e-llm-inference-service] self, [e2e-llm-inference-service] conn: BaseHTTPConnection, [e2e-llm-inference-service] method: str, [e2e-llm-inference-service] url: str, [e2e-llm-inference-service] body: _TYPE_BODY | None = None, [e2e-llm-inference-service] headers: typing.Mapping[str, str] | None = None, [e2e-llm-inference-service] retries: Retry | None = None, [e2e-llm-inference-service] timeout: _TYPE_TIMEOUT = _DEFAULT_TIMEOUT, [e2e-llm-inference-service] chunked: bool = False, [e2e-llm-inference-service] response_conn: BaseHTTPConnection | None = None, [e2e-llm-inference-service] preload_content: bool = True, [e2e-llm-inference-service] decode_content: bool = True, [e2e-llm-inference-service] enforce_content_length: bool = True, [e2e-llm-inference-service] ) -> BaseHTTPResponse: [e2e-llm-inference-service] """ [e2e-llm-inference-service] Perform a request on a given urllib connection object taken from our [e2e-llm-inference-service] pool. [e2e-llm-inference-service] [e2e-llm-inference-service] :param conn: [e2e-llm-inference-service] a connection from one of our connection pools [e2e-llm-inference-service] [e2e-llm-inference-service] :param method: [e2e-llm-inference-service] HTTP request method (such as GET, POST, PUT, etc.) [e2e-llm-inference-service] [e2e-llm-inference-service] :param url: [e2e-llm-inference-service] The URL to perform the request on. [e2e-llm-inference-service] [e2e-llm-inference-service] :param body: [e2e-llm-inference-service] Data to send in the request body, either :class:`str`, :class:`bytes`, [e2e-llm-inference-service] an iterable of :class:`str`/:class:`bytes`, or a file-like object. [e2e-llm-inference-service] [e2e-llm-inference-service] :param headers: [e2e-llm-inference-service] Dictionary of custom headers to send, such as User-Agent, [e2e-llm-inference-service] If-None-Match, etc. If None, pool headers are used. If provided, [e2e-llm-inference-service] these headers completely replace any pool-specific headers. [e2e-llm-inference-service] [e2e-llm-inference-service] :param retries: [e2e-llm-inference-service] Configure the number of retries to allow before raising a [e2e-llm-inference-service] :class:`~urllib3.exceptions.MaxRetryError` exception. [e2e-llm-inference-service] [e2e-llm-inference-service] Pass ``None`` to retry until you receive a response. Pass a [e2e-llm-inference-service] :class:`~urllib3.util.retry.Retry` object for fine-grained control [e2e-llm-inference-service] over different types of retries. [e2e-llm-inference-service] Pass an integer number to retry connection errors that many times, [e2e-llm-inference-service] but no other types of errors. Pass zero to never retry. [e2e-llm-inference-service] [e2e-llm-inference-service] If ``False``, then retries are disabled and any exception is raised [e2e-llm-inference-service] immediately. Also, instead of raising a MaxRetryError on redirects, [e2e-llm-inference-service] the redirect response will be returned. [e2e-llm-inference-service] [e2e-llm-inference-service] :type retries: :class:`~urllib3.util.retry.Retry`, False, or an int. [e2e-llm-inference-service] [e2e-llm-inference-service] :param timeout: [e2e-llm-inference-service] If specified, overrides the default timeout for this one [e2e-llm-inference-service] request. It may be a float (in seconds) or an instance of [e2e-llm-inference-service] :class:`urllib3.util.Timeout`. [e2e-llm-inference-service] [e2e-llm-inference-service] :param chunked: [e2e-llm-inference-service] If True, urllib3 will send the body using chunked transfer [e2e-llm-inference-service] encoding. Otherwise, urllib3 will send the body using the standard [e2e-llm-inference-service] content-length form. Defaults to False. [e2e-llm-inference-service] [e2e-llm-inference-service] :param response_conn: [e2e-llm-inference-service] Set this to ``None`` if you will handle releasing the connection or [e2e-llm-inference-service] set the connection to have the response release it. [e2e-llm-inference-service] [e2e-llm-inference-service] :param preload_content: [e2e-llm-inference-service] If True, the response's body will be preloaded during construction. [e2e-llm-inference-service] [e2e-llm-inference-service] :param decode_content: [e2e-llm-inference-service] If True, will attempt to decode the body based on the [e2e-llm-inference-service] 'content-encoding' header. [e2e-llm-inference-service] [e2e-llm-inference-service] :param enforce_content_length: [e2e-llm-inference-service] Enforce content length checking. Body returned by server must match [e2e-llm-inference-service] value of Content-Length header, if present. Otherwise, raise error. [e2e-llm-inference-service] """ [e2e-llm-inference-service] self.num_requests += 1 [e2e-llm-inference-service] [e2e-llm-inference-service] timeout_obj = self._get_timeout(timeout) [e2e-llm-inference-service] timeout_obj.start_connect() [e2e-llm-inference-service] conn.timeout = Timeout.resolve_default_timeout(timeout_obj.connect_timeout) [e2e-llm-inference-service] [e2e-llm-inference-service] try: [e2e-llm-inference-service] # Trigger any extra validation we need to do. [e2e-llm-inference-service] try: [e2e-llm-inference-service] self._validate_conn(conn) [e2e-llm-inference-service] except (SocketTimeout, BaseSSLError) as e: [e2e-llm-inference-service] self._raise_timeout(err=e, url=url, timeout_value=conn.timeout) [e2e-llm-inference-service] raise [e2e-llm-inference-service] [e2e-llm-inference-service] # _validate_conn() starts the connection to an HTTPS proxy [e2e-llm-inference-service] # so we need to wrap errors with 'ProxyError' here too. [e2e-llm-inference-service] except ( [e2e-llm-inference-service] OSError, [e2e-llm-inference-service] NewConnectionError, [e2e-llm-inference-service] TimeoutError, [e2e-llm-inference-service] BaseSSLError, [e2e-llm-inference-service] CertificateError, [e2e-llm-inference-service] SSLError, [e2e-llm-inference-service] ) as e: [e2e-llm-inference-service] new_e: Exception = e [e2e-llm-inference-service] if isinstance(e, (BaseSSLError, CertificateError)): [e2e-llm-inference-service] new_e = SSLError(e) [e2e-llm-inference-service] # If the connection didn't successfully connect to it's proxy [e2e-llm-inference-service] # then there [e2e-llm-inference-service] if isinstance( [e2e-llm-inference-service] new_e, (OSError, NewConnectionError, TimeoutError, SSLError) [e2e-llm-inference-service] ) and (conn and conn.proxy and not conn.has_connected_to_proxy): [e2e-llm-inference-service] new_e = _wrap_proxy_error(new_e, conn.proxy.scheme) [e2e-llm-inference-service] raise new_e [e2e-llm-inference-service] [e2e-llm-inference-service] # conn.request() calls http.client.*.request, not the method in [e2e-llm-inference-service] # urllib3.request. It also calls makefile (recv) on the socket. [e2e-llm-inference-service] try: [e2e-llm-inference-service] conn.request( [e2e-llm-inference-service] method, [e2e-llm-inference-service] url, [e2e-llm-inference-service] body=body, [e2e-llm-inference-service] headers=headers, [e2e-llm-inference-service] chunked=chunked, [e2e-llm-inference-service] preload_content=preload_content, [e2e-llm-inference-service] decode_content=decode_content, [e2e-llm-inference-service] enforce_content_length=enforce_content_length, [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] # We are swallowing BrokenPipeError (errno.EPIPE) since the server is [e2e-llm-inference-service] # legitimately able to close the connection after sending a valid response. [e2e-llm-inference-service] # With this behaviour, the received response is still readable. [e2e-llm-inference-service] except BrokenPipeError: [e2e-llm-inference-service] pass [e2e-llm-inference-service] except OSError as e: [e2e-llm-inference-service] # MacOS/Linux [e2e-llm-inference-service] # EPROTOTYPE and ECONNRESET are needed on macOS [e2e-llm-inference-service] # https://erickt.github.io/blog/2014/11/19/adventures-in-debugging-a-potential-osx-kernel-bug/ [e2e-llm-inference-service] # Condition changed later to emit ECONNRESET instead of only EPROTOTYPE. [e2e-llm-inference-service] if e.errno != errno.EPROTOTYPE and e.errno != errno.ECONNRESET: [e2e-llm-inference-service] raise [e2e-llm-inference-service] [e2e-llm-inference-service] # Reset the timeout for the recv() on the socket [e2e-llm-inference-service] read_timeout = timeout_obj.read_timeout [e2e-llm-inference-service] [e2e-llm-inference-service] if not conn.is_closed: [e2e-llm-inference-service] # In Python 3 socket.py will catch EAGAIN and return None when you [e2e-llm-inference-service] # try and read into the file pointer created by http.client, which [e2e-llm-inference-service] # instead raises a BadStatusLine exception. Instead of catching [e2e-llm-inference-service] # the exception and assuming all BadStatusLine exceptions are read [e2e-llm-inference-service] # timeouts, check for a zero timeout before making the request. [e2e-llm-inference-service] if read_timeout == 0: [e2e-llm-inference-service] raise ReadTimeoutError( [e2e-llm-inference-service] self, url, f"Read timed out. (read timeout={read_timeout})" [e2e-llm-inference-service] ) [e2e-llm-inference-service] conn.timeout = read_timeout [e2e-llm-inference-service] [e2e-llm-inference-service] # Receive the response from the server [e2e-llm-inference-service] try: [e2e-llm-inference-service] response = conn.getresponse() [e2e-llm-inference-service] except (BaseSSLError, OSError) as e: [e2e-llm-inference-service] > self._raise_timeout(err=e, url=url, timeout_value=read_timeout) [e2e-llm-inference-service] [e2e-llm-inference-service] ../../python/kserve/.venv/lib64/python3.11/site-packages/urllib3/connectionpool.py:536: [e2e-llm-inference-service] _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ [e2e-llm-inference-service] [e2e-llm-inference-service] self = [e2e-llm-inference-service] err = TimeoutError('timed out'), url = '/v1/models', timeout_value = 60 [e2e-llm-inference-service] [e2e-llm-inference-service] def _raise_timeout( [e2e-llm-inference-service] self, [e2e-llm-inference-service] err: BaseSSLError | OSError | SocketTimeout, [e2e-llm-inference-service] url: str, [e2e-llm-inference-service] timeout_value: _TYPE_TIMEOUT | None, [e2e-llm-inference-service] ) -> None: [e2e-llm-inference-service] """Is the error actually a timeout? Will raise a ReadTimeout or pass""" [e2e-llm-inference-service] [e2e-llm-inference-service] if isinstance(err, SocketTimeout): [e2e-llm-inference-service] > raise ReadTimeoutError( [e2e-llm-inference-service] self, url, f"Read timed out. (read timeout={timeout_value})" [e2e-llm-inference-service] ) from err [e2e-llm-inference-service] E urllib3.exceptions.ReadTimeoutError: HTTPConnectionPool(host='ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com', port=80): Read timed out. (read timeout=60) [e2e-llm-inference-service] [e2e-llm-inference-service] ../../python/kserve/.venv/lib64/python3.11/site-packages/urllib3/connectionpool.py:367: ReadTimeoutError [e2e-llm-inference-service] [e2e-llm-inference-service] The above exception was the direct cause of the following exception: [e2e-llm-inference-service] [e2e-llm-inference-service] self = [e2e-llm-inference-service] request = , stream = False [e2e-llm-inference-service] timeout = Timeout(connect=60, read=60, total=None), verify = '/tmp/ca.crt' [e2e-llm-inference-service] cert = None, proxies = OrderedDict() [e2e-llm-inference-service] [e2e-llm-inference-service] def send( [e2e-llm-inference-service] self, request, stream=False, timeout=None, verify=True, cert=None, proxies=None [e2e-llm-inference-service] ): [e2e-llm-inference-service] """Sends PreparedRequest object. Returns Response object. [e2e-llm-inference-service] [e2e-llm-inference-service] :param request: The :class:`PreparedRequest ` being sent. [e2e-llm-inference-service] :param stream: (optional) Whether to stream the request content. [e2e-llm-inference-service] :param timeout: (optional) How long to wait for the server to send [e2e-llm-inference-service] data before giving up, as a float, or a :ref:`(connect timeout, [e2e-llm-inference-service] read timeout) ` tuple. [e2e-llm-inference-service] :type timeout: float or tuple or urllib3 Timeout object [e2e-llm-inference-service] :param verify: (optional) Either a boolean, in which case it controls whether [e2e-llm-inference-service] we verify the server's TLS certificate, or a string, in which case it [e2e-llm-inference-service] must be a path to a CA bundle to use [e2e-llm-inference-service] :param cert: (optional) Any user-provided SSL certificate to be trusted. [e2e-llm-inference-service] :param proxies: (optional) The proxies dictionary to apply to the request. [e2e-llm-inference-service] :rtype: requests.Response [e2e-llm-inference-service] """ [e2e-llm-inference-service] [e2e-llm-inference-service] try: [e2e-llm-inference-service] conn = self.get_connection_with_tls_context( [e2e-llm-inference-service] request, verify, proxies=proxies, cert=cert [e2e-llm-inference-service] ) [e2e-llm-inference-service] except LocationValueError as e: [e2e-llm-inference-service] raise InvalidURL(e, request=request) [e2e-llm-inference-service] [e2e-llm-inference-service] self.cert_verify(conn, request.url, verify, cert) [e2e-llm-inference-service] url = self.request_url(request, proxies) [e2e-llm-inference-service] self.add_headers( [e2e-llm-inference-service] request, [e2e-llm-inference-service] stream=stream, [e2e-llm-inference-service] timeout=timeout, [e2e-llm-inference-service] verify=verify, [e2e-llm-inference-service] cert=cert, [e2e-llm-inference-service] proxies=proxies, [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] chunked = not (request.body is None or "Content-Length" in request.headers) [e2e-llm-inference-service] [e2e-llm-inference-service] if isinstance(timeout, tuple): [e2e-llm-inference-service] try: [e2e-llm-inference-service] connect, read = timeout [e2e-llm-inference-service] timeout = TimeoutSauce(connect=connect, read=read) [e2e-llm-inference-service] except ValueError: [e2e-llm-inference-service] raise ValueError( [e2e-llm-inference-service] f"Invalid timeout {timeout}. Pass a (connect, read) timeout tuple, " [e2e-llm-inference-service] f"or a single float to set both timeouts to the same value." [e2e-llm-inference-service] ) [e2e-llm-inference-service] elif isinstance(timeout, TimeoutSauce): [e2e-llm-inference-service] pass [e2e-llm-inference-service] else: [e2e-llm-inference-service] timeout = TimeoutSauce(connect=timeout, read=timeout) [e2e-llm-inference-service] [e2e-llm-inference-service] try: [e2e-llm-inference-service] > resp = conn.urlopen( [e2e-llm-inference-service] method=request.method, [e2e-llm-inference-service] url=url, [e2e-llm-inference-service] body=request.body, [e2e-llm-inference-service] headers=request.headers, [e2e-llm-inference-service] redirect=False, [e2e-llm-inference-service] assert_same_host=False, [e2e-llm-inference-service] preload_content=False, [e2e-llm-inference-service] decode_content=False, [e2e-llm-inference-service] retries=self.max_retries, [e2e-llm-inference-service] timeout=timeout, [e2e-llm-inference-service] chunked=chunked, [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] ../../python/kserve/.venv/lib64/python3.11/site-packages/requests/adapters.py:667: [e2e-llm-inference-service] _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ [e2e-llm-inference-service] [e2e-llm-inference-service] self = [e2e-llm-inference-service] method = 'GET', url = '/v1/models', body = None [e2e-llm-inference-service] headers = {'User-Agent': 'python-requests/2.32.3', 'Accept-Encoding': 'gzip, deflate', 'Accept': '*/*', 'Connection': 'keep-alive', 'Content-Type': 'application/json', 'X-Gateway-Model-Name': 'publishers/kserve-ci-e2e-test/models/facebook/opt-125m'} [e2e-llm-inference-service] retries = Retry(total=7, connect=None, read=None, redirect=None, status=None) [e2e-llm-inference-service] redirect = False, assert_same_host = False [e2e-llm-inference-service] timeout = Timeout(connect=60, read=60, total=None), pool_timeout = None [e2e-llm-inference-service] release_conn = False, chunked = False, body_pos = None, preload_content = False [e2e-llm-inference-service] decode_content = False, response_kw = {} [e2e-llm-inference-service] parsed_url = Url(scheme=None, auth=None, host=None, port=None, path='/v1/models', query=None, fragment=None) [e2e-llm-inference-service] destination_scheme = None, conn = None, release_this_conn = True [e2e-llm-inference-service] http_tunnel_required = False [e2e-llm-inference-service] err = ReadTimeoutError("HTTPConnectionPool(host='ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com', port=80): Read timed out. (read timeout=60)") [e2e-llm-inference-service] clean_exit = False [e2e-llm-inference-service] [e2e-llm-inference-service] def urlopen( # type: ignore[override] [e2e-llm-inference-service] self, [e2e-llm-inference-service] method: str, [e2e-llm-inference-service] url: str, [e2e-llm-inference-service] body: _TYPE_BODY | None = None, [e2e-llm-inference-service] headers: typing.Mapping[str, str] | None = None, [e2e-llm-inference-service] retries: Retry | bool | int | None = None, [e2e-llm-inference-service] redirect: bool = True, [e2e-llm-inference-service] assert_same_host: bool = True, [e2e-llm-inference-service] timeout: _TYPE_TIMEOUT = _DEFAULT_TIMEOUT, [e2e-llm-inference-service] pool_timeout: int | None = None, [e2e-llm-inference-service] release_conn: bool | None = None, [e2e-llm-inference-service] chunked: bool = False, [e2e-llm-inference-service] body_pos: _TYPE_BODY_POSITION | None = None, [e2e-llm-inference-service] preload_content: bool = True, [e2e-llm-inference-service] decode_content: bool = True, [e2e-llm-inference-service] **response_kw: typing.Any, [e2e-llm-inference-service] ) -> BaseHTTPResponse: [e2e-llm-inference-service] """ [e2e-llm-inference-service] Get a connection from the pool and perform an HTTP request. This is the [e2e-llm-inference-service] lowest level call for making a request, so you'll need to specify all [e2e-llm-inference-service] the raw details. [e2e-llm-inference-service] [e2e-llm-inference-service] .. note:: [e2e-llm-inference-service] [e2e-llm-inference-service] More commonly, it's appropriate to use a convenience method [e2e-llm-inference-service] such as :meth:`request`. [e2e-llm-inference-service] [e2e-llm-inference-service] .. note:: [e2e-llm-inference-service] [e2e-llm-inference-service] `release_conn` will only behave as expected if [e2e-llm-inference-service] `preload_content=False` because we want to make [e2e-llm-inference-service] `preload_content=False` the default behaviour someday soon without [e2e-llm-inference-service] breaking backwards compatibility. [e2e-llm-inference-service] [e2e-llm-inference-service] :param method: [e2e-llm-inference-service] HTTP request method (such as GET, POST, PUT, etc.) [e2e-llm-inference-service] [e2e-llm-inference-service] :param url: [e2e-llm-inference-service] The URL to perform the request on. [e2e-llm-inference-service] [e2e-llm-inference-service] :param body: [e2e-llm-inference-service] Data to send in the request body, either :class:`str`, :class:`bytes`, [e2e-llm-inference-service] an iterable of :class:`str`/:class:`bytes`, or a file-like object. [e2e-llm-inference-service] [e2e-llm-inference-service] :param headers: [e2e-llm-inference-service] Dictionary of custom headers to send, such as User-Agent, [e2e-llm-inference-service] If-None-Match, etc. If None, pool headers are used. If provided, [e2e-llm-inference-service] these headers completely replace any pool-specific headers. [e2e-llm-inference-service] [e2e-llm-inference-service] :param retries: [e2e-llm-inference-service] Configure the number of retries to allow before raising a [e2e-llm-inference-service] :class:`~urllib3.exceptions.MaxRetryError` exception. [e2e-llm-inference-service] [e2e-llm-inference-service] If ``None`` (default) will retry 3 times, see ``Retry.DEFAULT``. Pass a [e2e-llm-inference-service] :class:`~urllib3.util.retry.Retry` object for fine-grained control [e2e-llm-inference-service] over different types of retries. [e2e-llm-inference-service] Pass an integer number to retry connection errors that many times, [e2e-llm-inference-service] but no other types of errors. Pass zero to never retry. [e2e-llm-inference-service] [e2e-llm-inference-service] If ``False``, then retries are disabled and any exception is raised [e2e-llm-inference-service] immediately. Also, instead of raising a MaxRetryError on redirects, [e2e-llm-inference-service] the redirect response will be returned. [e2e-llm-inference-service] [e2e-llm-inference-service] :type retries: :class:`~urllib3.util.retry.Retry`, False, or an int. [e2e-llm-inference-service] [e2e-llm-inference-service] :param redirect: [e2e-llm-inference-service] If True, automatically handle redirects (status codes 301, 302, [e2e-llm-inference-service] 303, 307, 308). Each redirect counts as a retry. Disabling retries [e2e-llm-inference-service] will disable redirect, too. [e2e-llm-inference-service] [e2e-llm-inference-service] :param assert_same_host: [e2e-llm-inference-service] If ``True``, will make sure that the host of the pool requests is [e2e-llm-inference-service] consistent else will raise HostChangedError. When ``False``, you can [e2e-llm-inference-service] use the pool on an HTTP proxy and request foreign hosts. [e2e-llm-inference-service] [e2e-llm-inference-service] :param timeout: [e2e-llm-inference-service] If specified, overrides the default timeout for this one [e2e-llm-inference-service] request. It may be a float (in seconds) or an instance of [e2e-llm-inference-service] :class:`urllib3.util.Timeout`. [e2e-llm-inference-service] [e2e-llm-inference-service] :param pool_timeout: [e2e-llm-inference-service] If set and the pool is set to block=True, then this method will [e2e-llm-inference-service] block for ``pool_timeout`` seconds and raise EmptyPoolError if no [e2e-llm-inference-service] connection is available within the time period. [e2e-llm-inference-service] [e2e-llm-inference-service] :param bool preload_content: [e2e-llm-inference-service] If True, the response's body will be preloaded into memory. [e2e-llm-inference-service] [e2e-llm-inference-service] :param bool decode_content: [e2e-llm-inference-service] If True, will attempt to decode the body based on the [e2e-llm-inference-service] 'content-encoding' header. [e2e-llm-inference-service] [e2e-llm-inference-service] :param release_conn: [e2e-llm-inference-service] If False, then the urlopen call will not release the connection [e2e-llm-inference-service] back into the pool once a response is received (but will release if [e2e-llm-inference-service] you read the entire contents of the response such as when [e2e-llm-inference-service] `preload_content=True`). This is useful if you're not preloading [e2e-llm-inference-service] the response's content immediately. You will need to call [e2e-llm-inference-service] ``r.release_conn()`` on the response ``r`` to return the connection [e2e-llm-inference-service] back into the pool. If None, it takes the value of ``preload_content`` [e2e-llm-inference-service] which defaults to ``True``. [e2e-llm-inference-service] [e2e-llm-inference-service] :param bool chunked: [e2e-llm-inference-service] If True, urllib3 will send the body using chunked transfer [e2e-llm-inference-service] encoding. Otherwise, urllib3 will send the body using the standard [e2e-llm-inference-service] content-length form. Defaults to False. [e2e-llm-inference-service] [e2e-llm-inference-service] :param int body_pos: [e2e-llm-inference-service] Position to seek to in file-like body in the event of a retry or [e2e-llm-inference-service] redirect. Typically this won't need to be set because urllib3 will [e2e-llm-inference-service] auto-populate the value when needed. [e2e-llm-inference-service] """ [e2e-llm-inference-service] parsed_url = parse_url(url) [e2e-llm-inference-service] destination_scheme = parsed_url.scheme [e2e-llm-inference-service] [e2e-llm-inference-service] if headers is None: [e2e-llm-inference-service] headers = self.headers [e2e-llm-inference-service] [e2e-llm-inference-service] if not isinstance(retries, Retry): [e2e-llm-inference-service] retries = Retry.from_int(retries, redirect=redirect, default=self.retries) [e2e-llm-inference-service] [e2e-llm-inference-service] if release_conn is None: [e2e-llm-inference-service] release_conn = preload_content [e2e-llm-inference-service] [e2e-llm-inference-service] # Check host [e2e-llm-inference-service] if assert_same_host and not self.is_same_host(url): [e2e-llm-inference-service] raise HostChangedError(self, url, retries) [e2e-llm-inference-service] [e2e-llm-inference-service] # Ensure that the URL we're connecting to is properly encoded [e2e-llm-inference-service] if url.startswith("/"): [e2e-llm-inference-service] url = to_str(_encode_target(url)) [e2e-llm-inference-service] else: [e2e-llm-inference-service] url = to_str(parsed_url.url) [e2e-llm-inference-service] [e2e-llm-inference-service] conn = None [e2e-llm-inference-service] [e2e-llm-inference-service] # Track whether `conn` needs to be released before [e2e-llm-inference-service] # returning/raising/recursing. Update this variable if necessary, and [e2e-llm-inference-service] # leave `release_conn` constant throughout the function. That way, if [e2e-llm-inference-service] # the function recurses, the original value of `release_conn` will be [e2e-llm-inference-service] # passed down into the recursive call, and its value will be respected. [e2e-llm-inference-service] # [e2e-llm-inference-service] # See issue #651 [1] for details. [e2e-llm-inference-service] # [e2e-llm-inference-service] # [1] [e2e-llm-inference-service] release_this_conn = release_conn [e2e-llm-inference-service] [e2e-llm-inference-service] http_tunnel_required = connection_requires_http_tunnel( [e2e-llm-inference-service] self.proxy, self.proxy_config, destination_scheme [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] # Merge the proxy headers. Only done when not using HTTP CONNECT. We [e2e-llm-inference-service] # have to copy the headers dict so we can safely change it without those [e2e-llm-inference-service] # changes being reflected in anyone else's copy. [e2e-llm-inference-service] if not http_tunnel_required: [e2e-llm-inference-service] headers = headers.copy() # type: ignore[attr-defined] [e2e-llm-inference-service] headers.update(self.proxy_headers) # type: ignore[union-attr] [e2e-llm-inference-service] [e2e-llm-inference-service] # Must keep the exception bound to a separate variable or else Python 3 [e2e-llm-inference-service] # complains about UnboundLocalError. [e2e-llm-inference-service] err = None [e2e-llm-inference-service] [e2e-llm-inference-service] # Keep track of whether we cleanly exited the except block. This [e2e-llm-inference-service] # ensures we do proper cleanup in finally. [e2e-llm-inference-service] clean_exit = False [e2e-llm-inference-service] [e2e-llm-inference-service] # Rewind body position, if needed. Record current position [e2e-llm-inference-service] # for future rewinds in the event of a redirect/retry. [e2e-llm-inference-service] body_pos = set_file_position(body, body_pos) [e2e-llm-inference-service] [e2e-llm-inference-service] try: [e2e-llm-inference-service] # Request a connection from the queue. [e2e-llm-inference-service] timeout_obj = self._get_timeout(timeout) [e2e-llm-inference-service] conn = self._get_conn(timeout=pool_timeout) [e2e-llm-inference-service] [e2e-llm-inference-service] conn.timeout = timeout_obj.connect_timeout # type: ignore[assignment] [e2e-llm-inference-service] [e2e-llm-inference-service] # Is this a closed/new connection that requires CONNECT tunnelling? [e2e-llm-inference-service] if self.proxy is not None and http_tunnel_required and conn.is_closed: [e2e-llm-inference-service] try: [e2e-llm-inference-service] self._prepare_proxy(conn) [e2e-llm-inference-service] except (BaseSSLError, OSError, SocketTimeout) as e: [e2e-llm-inference-service] self._raise_timeout( [e2e-llm-inference-service] err=e, url=self.proxy.url, timeout_value=conn.timeout [e2e-llm-inference-service] ) [e2e-llm-inference-service] raise [e2e-llm-inference-service] [e2e-llm-inference-service] # If we're going to release the connection in ``finally:``, then [e2e-llm-inference-service] # the response doesn't need to know about the connection. Otherwise [e2e-llm-inference-service] # it will also try to release it and we'll have a double-release [e2e-llm-inference-service] # mess. [e2e-llm-inference-service] response_conn = conn if not release_conn else None [e2e-llm-inference-service] [e2e-llm-inference-service] # Make the request on the HTTPConnection object [e2e-llm-inference-service] response = self._make_request( [e2e-llm-inference-service] conn, [e2e-llm-inference-service] method, [e2e-llm-inference-service] url, [e2e-llm-inference-service] timeout=timeout_obj, [e2e-llm-inference-service] body=body, [e2e-llm-inference-service] headers=headers, [e2e-llm-inference-service] chunked=chunked, [e2e-llm-inference-service] retries=retries, [e2e-llm-inference-service] response_conn=response_conn, [e2e-llm-inference-service] preload_content=preload_content, [e2e-llm-inference-service] decode_content=decode_content, [e2e-llm-inference-service] **response_kw, [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] # Everything went great! [e2e-llm-inference-service] clean_exit = True [e2e-llm-inference-service] [e2e-llm-inference-service] except EmptyPoolError: [e2e-llm-inference-service] # Didn't get a connection from the pool, no need to clean up [e2e-llm-inference-service] clean_exit = True [e2e-llm-inference-service] release_this_conn = False [e2e-llm-inference-service] raise [e2e-llm-inference-service] [e2e-llm-inference-service] except ( [e2e-llm-inference-service] TimeoutError, [e2e-llm-inference-service] HTTPException, [e2e-llm-inference-service] OSError, [e2e-llm-inference-service] ProtocolError, [e2e-llm-inference-service] BaseSSLError, [e2e-llm-inference-service] SSLError, [e2e-llm-inference-service] CertificateError, [e2e-llm-inference-service] ProxyError, [e2e-llm-inference-service] ) as e: [e2e-llm-inference-service] # Discard the connection for these exceptions. It will be [e2e-llm-inference-service] # replaced during the next _get_conn() call. [e2e-llm-inference-service] clean_exit = False [e2e-llm-inference-service] new_e: Exception = e [e2e-llm-inference-service] if isinstance(e, (BaseSSLError, CertificateError)): [e2e-llm-inference-service] new_e = SSLError(e) [e2e-llm-inference-service] if isinstance( [e2e-llm-inference-service] new_e, [e2e-llm-inference-service] ( [e2e-llm-inference-service] OSError, [e2e-llm-inference-service] NewConnectionError, [e2e-llm-inference-service] TimeoutError, [e2e-llm-inference-service] SSLError, [e2e-llm-inference-service] HTTPException, [e2e-llm-inference-service] ), [e2e-llm-inference-service] ) and (conn and conn.proxy and not conn.has_connected_to_proxy): [e2e-llm-inference-service] new_e = _wrap_proxy_error(new_e, conn.proxy.scheme) [e2e-llm-inference-service] elif isinstance(new_e, (OSError, HTTPException)): [e2e-llm-inference-service] new_e = ProtocolError("Connection aborted.", new_e) [e2e-llm-inference-service] [e2e-llm-inference-service] retries = retries.increment( [e2e-llm-inference-service] method, url, error=new_e, _pool=self, _stacktrace=sys.exc_info()[2] [e2e-llm-inference-service] ) [e2e-llm-inference-service] retries.sleep() [e2e-llm-inference-service] [e2e-llm-inference-service] # Keep track of the error for the retry warning. [e2e-llm-inference-service] err = e [e2e-llm-inference-service] [e2e-llm-inference-service] finally: [e2e-llm-inference-service] if not clean_exit: [e2e-llm-inference-service] # We hit some kind of exception, handled or otherwise. We need [e2e-llm-inference-service] # to throw the connection away unless explicitly told not to. [e2e-llm-inference-service] # Close the connection, set the variable to None, and make sure [e2e-llm-inference-service] # we put the None back in the pool to avoid leaking it. [e2e-llm-inference-service] if conn: [e2e-llm-inference-service] conn.close() [e2e-llm-inference-service] conn = None [e2e-llm-inference-service] release_this_conn = True [e2e-llm-inference-service] [e2e-llm-inference-service] if release_this_conn: [e2e-llm-inference-service] # Put the connection back to be reused. If the connection is [e2e-llm-inference-service] # expired then it will be None, which will get replaced with a [e2e-llm-inference-service] # fresh connection during _get_conn. [e2e-llm-inference-service] self._put_conn(conn) [e2e-llm-inference-service] [e2e-llm-inference-service] if not conn: [e2e-llm-inference-service] # Try again [e2e-llm-inference-service] log.warning( [e2e-llm-inference-service] "Retrying (%r) after connection broken by '%r': %s", retries, err, url [e2e-llm-inference-service] ) [e2e-llm-inference-service] > return self.urlopen( [e2e-llm-inference-service] method, [e2e-llm-inference-service] url, [e2e-llm-inference-service] body, [e2e-llm-inference-service] headers, [e2e-llm-inference-service] retries, [e2e-llm-inference-service] redirect, [e2e-llm-inference-service] assert_same_host, [e2e-llm-inference-service] timeout=timeout, [e2e-llm-inference-service] pool_timeout=pool_timeout, [e2e-llm-inference-service] release_conn=release_conn, [e2e-llm-inference-service] chunked=chunked, [e2e-llm-inference-service] body_pos=body_pos, [e2e-llm-inference-service] preload_content=preload_content, [e2e-llm-inference-service] decode_content=decode_content, [e2e-llm-inference-service] **response_kw, [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] ../../python/kserve/.venv/lib64/python3.11/site-packages/urllib3/connectionpool.py:871: [e2e-llm-inference-service] _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ [e2e-llm-inference-service] [e2e-llm-inference-service] self = [e2e-llm-inference-service] method = 'GET', url = '/v1/models', body = None [e2e-llm-inference-service] headers = {'User-Agent': 'python-requests/2.32.3', 'Accept-Encoding': 'gzip, deflate', 'Accept': '*/*', 'Connection': 'keep-alive', 'Content-Type': 'application/json', 'X-Gateway-Model-Name': 'publishers/kserve-ci-e2e-test/models/facebook/opt-125m'} [e2e-llm-inference-service] retries = Retry(total=6, connect=None, read=None, redirect=None, status=None) [e2e-llm-inference-service] redirect = False, assert_same_host = False [e2e-llm-inference-service] timeout = Timeout(connect=60, read=60, total=None), pool_timeout = None [e2e-llm-inference-service] release_conn = False, chunked = False, body_pos = None, preload_content = False [e2e-llm-inference-service] decode_content = False, response_kw = {} [e2e-llm-inference-service] parsed_url = Url(scheme=None, auth=None, host=None, port=None, path='/v1/models', query=None, fragment=None) [e2e-llm-inference-service] destination_scheme = None, conn = None, release_this_conn = True [e2e-llm-inference-service] http_tunnel_required = False [e2e-llm-inference-service] err = RemoteDisconnected('Remote end closed connection without response') [e2e-llm-inference-service] clean_exit = False [e2e-llm-inference-service] [e2e-llm-inference-service] def urlopen( # type: ignore[override] [e2e-llm-inference-service] self, [e2e-llm-inference-service] method: str, [e2e-llm-inference-service] url: str, [e2e-llm-inference-service] body: _TYPE_BODY | None = None, [e2e-llm-inference-service] headers: typing.Mapping[str, str] | None = None, [e2e-llm-inference-service] retries: Retry | bool | int | None = None, [e2e-llm-inference-service] redirect: bool = True, [e2e-llm-inference-service] assert_same_host: bool = True, [e2e-llm-inference-service] timeout: _TYPE_TIMEOUT = _DEFAULT_TIMEOUT, [e2e-llm-inference-service] pool_timeout: int | None = None, [e2e-llm-inference-service] release_conn: bool | None = None, [e2e-llm-inference-service] chunked: bool = False, [e2e-llm-inference-service] body_pos: _TYPE_BODY_POSITION | None = None, [e2e-llm-inference-service] preload_content: bool = True, [e2e-llm-inference-service] decode_content: bool = True, [e2e-llm-inference-service] **response_kw: typing.Any, [e2e-llm-inference-service] ) -> BaseHTTPResponse: [e2e-llm-inference-service] """ [e2e-llm-inference-service] Get a connection from the pool and perform an HTTP request. This is the [e2e-llm-inference-service] lowest level call for making a request, so you'll need to specify all [e2e-llm-inference-service] the raw details. [e2e-llm-inference-service] [e2e-llm-inference-service] .. note:: [e2e-llm-inference-service] [e2e-llm-inference-service] More commonly, it's appropriate to use a convenience method [e2e-llm-inference-service] such as :meth:`request`. [e2e-llm-inference-service] [e2e-llm-inference-service] .. note:: [e2e-llm-inference-service] [e2e-llm-inference-service] `release_conn` will only behave as expected if [e2e-llm-inference-service] `preload_content=False` because we want to make [e2e-llm-inference-service] `preload_content=False` the default behaviour someday soon without [e2e-llm-inference-service] breaking backwards compatibility. [e2e-llm-inference-service] [e2e-llm-inference-service] :param method: [e2e-llm-inference-service] HTTP request method (such as GET, POST, PUT, etc.) [e2e-llm-inference-service] [e2e-llm-inference-service] :param url: [e2e-llm-inference-service] The URL to perform the request on. [e2e-llm-inference-service] [e2e-llm-inference-service] :param body: [e2e-llm-inference-service] Data to send in the request body, either :class:`str`, :class:`bytes`, [e2e-llm-inference-service] an iterable of :class:`str`/:class:`bytes`, or a file-like object. [e2e-llm-inference-service] [e2e-llm-inference-service] :param headers: [e2e-llm-inference-service] Dictionary of custom headers to send, such as User-Agent, [e2e-llm-inference-service] If-None-Match, etc. If None, pool headers are used. If provided, [e2e-llm-inference-service] these headers completely replace any pool-specific headers. [e2e-llm-inference-service] [e2e-llm-inference-service] :param retries: [e2e-llm-inference-service] Configure the number of retries to allow before raising a [e2e-llm-inference-service] :class:`~urllib3.exceptions.MaxRetryError` exception. [e2e-llm-inference-service] [e2e-llm-inference-service] If ``None`` (default) will retry 3 times, see ``Retry.DEFAULT``. Pass a [e2e-llm-inference-service] :class:`~urllib3.util.retry.Retry` object for fine-grained control [e2e-llm-inference-service] over different types of retries. [e2e-llm-inference-service] Pass an integer number to retry connection errors that many times, [e2e-llm-inference-service] but no other types of errors. Pass zero to never retry. [e2e-llm-inference-service] [e2e-llm-inference-service] If ``False``, then retries are disabled and any exception is raised [e2e-llm-inference-service] immediately. Also, instead of raising a MaxRetryError on redirects, [e2e-llm-inference-service] the redirect response will be returned. [e2e-llm-inference-service] [e2e-llm-inference-service] :type retries: :class:`~urllib3.util.retry.Retry`, False, or an int. [e2e-llm-inference-service] [e2e-llm-inference-service] :param redirect: [e2e-llm-inference-service] If True, automatically handle redirects (status codes 301, 302, [e2e-llm-inference-service] 303, 307, 308). Each redirect counts as a retry. Disabling retries [e2e-llm-inference-service] will disable redirect, too. [e2e-llm-inference-service] [e2e-llm-inference-service] :param assert_same_host: [e2e-llm-inference-service] If ``True``, will make sure that the host of the pool requests is [e2e-llm-inference-service] consistent else will raise HostChangedError. When ``False``, you can [e2e-llm-inference-service] use the pool on an HTTP proxy and request foreign hosts. [e2e-llm-inference-service] [e2e-llm-inference-service] :param timeout: [e2e-llm-inference-service] If specified, overrides the default timeout for this one [e2e-llm-inference-service] request. It may be a float (in seconds) or an instance of [e2e-llm-inference-service] :class:`urllib3.util.Timeout`. [e2e-llm-inference-service] [e2e-llm-inference-service] :param pool_timeout: [e2e-llm-inference-service] If set and the pool is set to block=True, then this method will [e2e-llm-inference-service] block for ``pool_timeout`` seconds and raise EmptyPoolError if no [e2e-llm-inference-service] connection is available within the time period. [e2e-llm-inference-service] [e2e-llm-inference-service] :param bool preload_content: [e2e-llm-inference-service] If True, the response's body will be preloaded into memory. [e2e-llm-inference-service] [e2e-llm-inference-service] :param bool decode_content: [e2e-llm-inference-service] If True, will attempt to decode the body based on the [e2e-llm-inference-service] 'content-encoding' header. [e2e-llm-inference-service] [e2e-llm-inference-service] :param release_conn: [e2e-llm-inference-service] If False, then the urlopen call will not release the connection [e2e-llm-inference-service] back into the pool once a response is received (but will release if [e2e-llm-inference-service] you read the entire contents of the response such as when [e2e-llm-inference-service] `preload_content=True`). This is useful if you're not preloading [e2e-llm-inference-service] the response's content immediately. You will need to call [e2e-llm-inference-service] ``r.release_conn()`` on the response ``r`` to return the connection [e2e-llm-inference-service] back into the pool. If None, it takes the value of ``preload_content`` [e2e-llm-inference-service] which defaults to ``True``. [e2e-llm-inference-service] [e2e-llm-inference-service] :param bool chunked: [e2e-llm-inference-service] If True, urllib3 will send the body using chunked transfer [e2e-llm-inference-service] encoding. Otherwise, urllib3 will send the body using the standard [e2e-llm-inference-service] content-length form. Defaults to False. [e2e-llm-inference-service] [e2e-llm-inference-service] :param int body_pos: [e2e-llm-inference-service] Position to seek to in file-like body in the event of a retry or [e2e-llm-inference-service] redirect. Typically this won't need to be set because urllib3 will [e2e-llm-inference-service] auto-populate the value when needed. [e2e-llm-inference-service] """ [e2e-llm-inference-service] parsed_url = parse_url(url) [e2e-llm-inference-service] destination_scheme = parsed_url.scheme [e2e-llm-inference-service] [e2e-llm-inference-service] if headers is None: [e2e-llm-inference-service] headers = self.headers [e2e-llm-inference-service] [e2e-llm-inference-service] if not isinstance(retries, Retry): [e2e-llm-inference-service] retries = Retry.from_int(retries, redirect=redirect, default=self.retries) [e2e-llm-inference-service] [e2e-llm-inference-service] if release_conn is None: [e2e-llm-inference-service] release_conn = preload_content [e2e-llm-inference-service] [e2e-llm-inference-service] # Check host [e2e-llm-inference-service] if assert_same_host and not self.is_same_host(url): [e2e-llm-inference-service] raise HostChangedError(self, url, retries) [e2e-llm-inference-service] [e2e-llm-inference-service] # Ensure that the URL we're connecting to is properly encoded [e2e-llm-inference-service] if url.startswith("/"): [e2e-llm-inference-service] url = to_str(_encode_target(url)) [e2e-llm-inference-service] else: [e2e-llm-inference-service] url = to_str(parsed_url.url) [e2e-llm-inference-service] [e2e-llm-inference-service] conn = None [e2e-llm-inference-service] [e2e-llm-inference-service] # Track whether `conn` needs to be released before [e2e-llm-inference-service] # returning/raising/recursing. Update this variable if necessary, and [e2e-llm-inference-service] # leave `release_conn` constant throughout the function. That way, if [e2e-llm-inference-service] # the function recurses, the original value of `release_conn` will be [e2e-llm-inference-service] # passed down into the recursive call, and its value will be respected. [e2e-llm-inference-service] # [e2e-llm-inference-service] # See issue #651 [1] for details. [e2e-llm-inference-service] # [e2e-llm-inference-service] # [1] [e2e-llm-inference-service] release_this_conn = release_conn [e2e-llm-inference-service] [e2e-llm-inference-service] http_tunnel_required = connection_requires_http_tunnel( [e2e-llm-inference-service] self.proxy, self.proxy_config, destination_scheme [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] # Merge the proxy headers. Only done when not using HTTP CONNECT. We [e2e-llm-inference-service] # have to copy the headers dict so we can safely change it without those [e2e-llm-inference-service] # changes being reflected in anyone else's copy. [e2e-llm-inference-service] if not http_tunnel_required: [e2e-llm-inference-service] headers = headers.copy() # type: ignore[attr-defined] [e2e-llm-inference-service] headers.update(self.proxy_headers) # type: ignore[union-attr] [e2e-llm-inference-service] [e2e-llm-inference-service] # Must keep the exception bound to a separate variable or else Python 3 [e2e-llm-inference-service] # complains about UnboundLocalError. [e2e-llm-inference-service] err = None [e2e-llm-inference-service] [e2e-llm-inference-service] # Keep track of whether we cleanly exited the except block. This [e2e-llm-inference-service] # ensures we do proper cleanup in finally. [e2e-llm-inference-service] clean_exit = False [e2e-llm-inference-service] [e2e-llm-inference-service] # Rewind body position, if needed. Record current position [e2e-llm-inference-service] # for future rewinds in the event of a redirect/retry. [e2e-llm-inference-service] body_pos = set_file_position(body, body_pos) [e2e-llm-inference-service] [e2e-llm-inference-service] try: [e2e-llm-inference-service] # Request a connection from the queue. [e2e-llm-inference-service] timeout_obj = self._get_timeout(timeout) [e2e-llm-inference-service] conn = self._get_conn(timeout=pool_timeout) [e2e-llm-inference-service] [e2e-llm-inference-service] conn.timeout = timeout_obj.connect_timeout # type: ignore[assignment] [e2e-llm-inference-service] [e2e-llm-inference-service] # Is this a closed/new connection that requires CONNECT tunnelling? [e2e-llm-inference-service] if self.proxy is not None and http_tunnel_required and conn.is_closed: [e2e-llm-inference-service] try: [e2e-llm-inference-service] self._prepare_proxy(conn) [e2e-llm-inference-service] except (BaseSSLError, OSError, SocketTimeout) as e: [e2e-llm-inference-service] self._raise_timeout( [e2e-llm-inference-service] err=e, url=self.proxy.url, timeout_value=conn.timeout [e2e-llm-inference-service] ) [e2e-llm-inference-service] raise [e2e-llm-inference-service] [e2e-llm-inference-service] # If we're going to release the connection in ``finally:``, then [e2e-llm-inference-service] # the response doesn't need to know about the connection. Otherwise [e2e-llm-inference-service] # it will also try to release it and we'll have a double-release [e2e-llm-inference-service] # mess. [e2e-llm-inference-service] response_conn = conn if not release_conn else None [e2e-llm-inference-service] [e2e-llm-inference-service] # Make the request on the HTTPConnection object [e2e-llm-inference-service] response = self._make_request( [e2e-llm-inference-service] conn, [e2e-llm-inference-service] method, [e2e-llm-inference-service] url, [e2e-llm-inference-service] timeout=timeout_obj, [e2e-llm-inference-service] body=body, [e2e-llm-inference-service] headers=headers, [e2e-llm-inference-service] chunked=chunked, [e2e-llm-inference-service] retries=retries, [e2e-llm-inference-service] response_conn=response_conn, [e2e-llm-inference-service] preload_content=preload_content, [e2e-llm-inference-service] decode_content=decode_content, [e2e-llm-inference-service] **response_kw, [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] # Everything went great! [e2e-llm-inference-service] clean_exit = True [e2e-llm-inference-service] [e2e-llm-inference-service] except EmptyPoolError: [e2e-llm-inference-service] # Didn't get a connection from the pool, no need to clean up [e2e-llm-inference-service] clean_exit = True [e2e-llm-inference-service] release_this_conn = False [e2e-llm-inference-service] raise [e2e-llm-inference-service] [e2e-llm-inference-service] except ( [e2e-llm-inference-service] TimeoutError, [e2e-llm-inference-service] HTTPException, [e2e-llm-inference-service] OSError, [e2e-llm-inference-service] ProtocolError, [e2e-llm-inference-service] BaseSSLError, [e2e-llm-inference-service] SSLError, [e2e-llm-inference-service] CertificateError, [e2e-llm-inference-service] ProxyError, [e2e-llm-inference-service] ) as e: [e2e-llm-inference-service] # Discard the connection for these exceptions. It will be [e2e-llm-inference-service] # replaced during the next _get_conn() call. [e2e-llm-inference-service] clean_exit = False [e2e-llm-inference-service] new_e: Exception = e [e2e-llm-inference-service] if isinstance(e, (BaseSSLError, CertificateError)): [e2e-llm-inference-service] new_e = SSLError(e) [e2e-llm-inference-service] if isinstance( [e2e-llm-inference-service] new_e, [e2e-llm-inference-service] ( [e2e-llm-inference-service] OSError, [e2e-llm-inference-service] NewConnectionError, [e2e-llm-inference-service] TimeoutError, [e2e-llm-inference-service] SSLError, [e2e-llm-inference-service] HTTPException, [e2e-llm-inference-service] ), [e2e-llm-inference-service] ) and (conn and conn.proxy and not conn.has_connected_to_proxy): [e2e-llm-inference-service] new_e = _wrap_proxy_error(new_e, conn.proxy.scheme) [e2e-llm-inference-service] elif isinstance(new_e, (OSError, HTTPException)): [e2e-llm-inference-service] new_e = ProtocolError("Connection aborted.", new_e) [e2e-llm-inference-service] [e2e-llm-inference-service] retries = retries.increment( [e2e-llm-inference-service] method, url, error=new_e, _pool=self, _stacktrace=sys.exc_info()[2] [e2e-llm-inference-service] ) [e2e-llm-inference-service] retries.sleep() [e2e-llm-inference-service] [e2e-llm-inference-service] # Keep track of the error for the retry warning. [e2e-llm-inference-service] err = e [e2e-llm-inference-service] [e2e-llm-inference-service] finally: [e2e-llm-inference-service] if not clean_exit: [e2e-llm-inference-service] # We hit some kind of exception, handled or otherwise. We need [e2e-llm-inference-service] # to throw the connection away unless explicitly told not to. [e2e-llm-inference-service] # Close the connection, set the variable to None, and make sure [e2e-llm-inference-service] # we put the None back in the pool to avoid leaking it. [e2e-llm-inference-service] if conn: [e2e-llm-inference-service] conn.close() [e2e-llm-inference-service] conn = None [e2e-llm-inference-service] release_this_conn = True [e2e-llm-inference-service] [e2e-llm-inference-service] if release_this_conn: [e2e-llm-inference-service] # Put the connection back to be reused. If the connection is [e2e-llm-inference-service] # expired then it will be None, which will get replaced with a [e2e-llm-inference-service] # fresh connection during _get_conn. [e2e-llm-inference-service] self._put_conn(conn) [e2e-llm-inference-service] [e2e-llm-inference-service] if not conn: [e2e-llm-inference-service] # Try again [e2e-llm-inference-service] log.warning( [e2e-llm-inference-service] "Retrying (%r) after connection broken by '%r': %s", retries, err, url [e2e-llm-inference-service] ) [e2e-llm-inference-service] > return self.urlopen( [e2e-llm-inference-service] method, [e2e-llm-inference-service] url, [e2e-llm-inference-service] body, [e2e-llm-inference-service] headers, [e2e-llm-inference-service] retries, [e2e-llm-inference-service] redirect, [e2e-llm-inference-service] assert_same_host, [e2e-llm-inference-service] timeout=timeout, [e2e-llm-inference-service] pool_timeout=pool_timeout, [e2e-llm-inference-service] release_conn=release_conn, [e2e-llm-inference-service] chunked=chunked, [e2e-llm-inference-service] body_pos=body_pos, [e2e-llm-inference-service] preload_content=preload_content, [e2e-llm-inference-service] decode_content=decode_content, [e2e-llm-inference-service] **response_kw, [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] ../../python/kserve/.venv/lib64/python3.11/site-packages/urllib3/connectionpool.py:871: [e2e-llm-inference-service] _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ [e2e-llm-inference-service] [e2e-llm-inference-service] self = [e2e-llm-inference-service] method = 'GET', url = '/v1/models', body = None [e2e-llm-inference-service] headers = {'User-Agent': 'python-requests/2.32.3', 'Accept-Encoding': 'gzip, deflate', 'Accept': '*/*', 'Connection': 'keep-alive', 'Content-Type': 'application/json', 'X-Gateway-Model-Name': 'publishers/kserve-ci-e2e-test/models/facebook/opt-125m'} [e2e-llm-inference-service] retries = Retry(total=5, connect=None, read=None, redirect=None, status=None) [e2e-llm-inference-service] redirect = False, assert_same_host = False [e2e-llm-inference-service] timeout = Timeout(connect=60, read=60, total=None), pool_timeout = None [e2e-llm-inference-service] release_conn = False, chunked = False, body_pos = None, preload_content = False [e2e-llm-inference-service] decode_content = False, response_kw = {} [e2e-llm-inference-service] parsed_url = Url(scheme=None, auth=None, host=None, port=None, path='/v1/models', query=None, fragment=None) [e2e-llm-inference-service] destination_scheme = None, conn = None, release_this_conn = True [e2e-llm-inference-service] http_tunnel_required = False [e2e-llm-inference-service] err = ReadTimeoutError("HTTPConnectionPool(host='ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com', port=80): Read timed out. (read timeout=60)") [e2e-llm-inference-service] clean_exit = False [e2e-llm-inference-service] [e2e-llm-inference-service] def urlopen( # type: ignore[override] [e2e-llm-inference-service] self, [e2e-llm-inference-service] method: str, [e2e-llm-inference-service] url: str, [e2e-llm-inference-service] body: _TYPE_BODY | None = None, [e2e-llm-inference-service] headers: typing.Mapping[str, str] | None = None, [e2e-llm-inference-service] retries: Retry | bool | int | None = None, [e2e-llm-inference-service] redirect: bool = True, [e2e-llm-inference-service] assert_same_host: bool = True, [e2e-llm-inference-service] timeout: _TYPE_TIMEOUT = _DEFAULT_TIMEOUT, [e2e-llm-inference-service] pool_timeout: int | None = None, [e2e-llm-inference-service] release_conn: bool | None = None, [e2e-llm-inference-service] chunked: bool = False, [e2e-llm-inference-service] body_pos: _TYPE_BODY_POSITION | None = None, [e2e-llm-inference-service] preload_content: bool = True, [e2e-llm-inference-service] decode_content: bool = True, [e2e-llm-inference-service] **response_kw: typing.Any, [e2e-llm-inference-service] ) -> BaseHTTPResponse: [e2e-llm-inference-service] """ [e2e-llm-inference-service] Get a connection from the pool and perform an HTTP request. This is the [e2e-llm-inference-service] lowest level call for making a request, so you'll need to specify all [e2e-llm-inference-service] the raw details. [e2e-llm-inference-service] [e2e-llm-inference-service] .. note:: [e2e-llm-inference-service] [e2e-llm-inference-service] More commonly, it's appropriate to use a convenience method [e2e-llm-inference-service] such as :meth:`request`. [e2e-llm-inference-service] [e2e-llm-inference-service] .. note:: [e2e-llm-inference-service] [e2e-llm-inference-service] `release_conn` will only behave as expected if [e2e-llm-inference-service] `preload_content=False` because we want to make [e2e-llm-inference-service] `preload_content=False` the default behaviour someday soon without [e2e-llm-inference-service] breaking backwards compatibility. [e2e-llm-inference-service] [e2e-llm-inference-service] :param method: [e2e-llm-inference-service] HTTP request method (such as GET, POST, PUT, etc.) [e2e-llm-inference-service] [e2e-llm-inference-service] :param url: [e2e-llm-inference-service] The URL to perform the request on. [e2e-llm-inference-service] [e2e-llm-inference-service] :param body: [e2e-llm-inference-service] Data to send in the request body, either :class:`str`, :class:`bytes`, [e2e-llm-inference-service] an iterable of :class:`str`/:class:`bytes`, or a file-like object. [e2e-llm-inference-service] [e2e-llm-inference-service] :param headers: [e2e-llm-inference-service] Dictionary of custom headers to send, such as User-Agent, [e2e-llm-inference-service] If-None-Match, etc. If None, pool headers are used. If provided, [e2e-llm-inference-service] these headers completely replace any pool-specific headers. [e2e-llm-inference-service] [e2e-llm-inference-service] :param retries: [e2e-llm-inference-service] Configure the number of retries to allow before raising a [e2e-llm-inference-service] :class:`~urllib3.exceptions.MaxRetryError` exception. [e2e-llm-inference-service] [e2e-llm-inference-service] If ``None`` (default) will retry 3 times, see ``Retry.DEFAULT``. Pass a [e2e-llm-inference-service] :class:`~urllib3.util.retry.Retry` object for fine-grained control [e2e-llm-inference-service] over different types of retries. [e2e-llm-inference-service] Pass an integer number to retry connection errors that many times, [e2e-llm-inference-service] but no other types of errors. Pass zero to never retry. [e2e-llm-inference-service] [e2e-llm-inference-service] If ``False``, then retries are disabled and any exception is raised [e2e-llm-inference-service] immediately. Also, instead of raising a MaxRetryError on redirects, [e2e-llm-inference-service] the redirect response will be returned. [e2e-llm-inference-service] [e2e-llm-inference-service] :type retries: :class:`~urllib3.util.retry.Retry`, False, or an int. [e2e-llm-inference-service] [e2e-llm-inference-service] :param redirect: [e2e-llm-inference-service] If True, automatically handle redirects (status codes 301, 302, [e2e-llm-inference-service] 303, 307, 308). Each redirect counts as a retry. Disabling retries [e2e-llm-inference-service] will disable redirect, too. [e2e-llm-inference-service] [e2e-llm-inference-service] :param assert_same_host: [e2e-llm-inference-service] If ``True``, will make sure that the host of the pool requests is [e2e-llm-inference-service] consistent else will raise HostChangedError. When ``False``, you can [e2e-llm-inference-service] use the pool on an HTTP proxy and request foreign hosts. [e2e-llm-inference-service] [e2e-llm-inference-service] :param timeout: [e2e-llm-inference-service] If specified, overrides the default timeout for this one [e2e-llm-inference-service] request. It may be a float (in seconds) or an instance of [e2e-llm-inference-service] :class:`urllib3.util.Timeout`. [e2e-llm-inference-service] [e2e-llm-inference-service] :param pool_timeout: [e2e-llm-inference-service] If set and the pool is set to block=True, then this method will [e2e-llm-inference-service] block for ``pool_timeout`` seconds and raise EmptyPoolError if no [e2e-llm-inference-service] connection is available within the time period. [e2e-llm-inference-service] [e2e-llm-inference-service] :param bool preload_content: [e2e-llm-inference-service] If True, the response's body will be preloaded into memory. [e2e-llm-inference-service] [e2e-llm-inference-service] :param bool decode_content: [e2e-llm-inference-service] If True, will attempt to decode the body based on the [e2e-llm-inference-service] 'content-encoding' header. [e2e-llm-inference-service] [e2e-llm-inference-service] :param release_conn: [e2e-llm-inference-service] If False, then the urlopen call will not release the connection [e2e-llm-inference-service] back into the pool once a response is received (but will release if [e2e-llm-inference-service] you read the entire contents of the response such as when [e2e-llm-inference-service] `preload_content=True`). This is useful if you're not preloading [e2e-llm-inference-service] the response's content immediately. You will need to call [e2e-llm-inference-service] ``r.release_conn()`` on the response ``r`` to return the connection [e2e-llm-inference-service] back into the pool. If None, it takes the value of ``preload_content`` [e2e-llm-inference-service] which defaults to ``True``. [e2e-llm-inference-service] [e2e-llm-inference-service] :param bool chunked: [e2e-llm-inference-service] If True, urllib3 will send the body using chunked transfer [e2e-llm-inference-service] encoding. Otherwise, urllib3 will send the body using the standard [e2e-llm-inference-service] content-length form. Defaults to False. [e2e-llm-inference-service] [e2e-llm-inference-service] :param int body_pos: [e2e-llm-inference-service] Position to seek to in file-like body in the event of a retry or [e2e-llm-inference-service] redirect. Typically this won't need to be set because urllib3 will [e2e-llm-inference-service] auto-populate the value when needed. [e2e-llm-inference-service] """ [e2e-llm-inference-service] parsed_url = parse_url(url) [e2e-llm-inference-service] destination_scheme = parsed_url.scheme [e2e-llm-inference-service] [e2e-llm-inference-service] if headers is None: [e2e-llm-inference-service] headers = self.headers [e2e-llm-inference-service] [e2e-llm-inference-service] if not isinstance(retries, Retry): [e2e-llm-inference-service] retries = Retry.from_int(retries, redirect=redirect, default=self.retries) [e2e-llm-inference-service] [e2e-llm-inference-service] if release_conn is None: [e2e-llm-inference-service] release_conn = preload_content [e2e-llm-inference-service] [e2e-llm-inference-service] # Check host [e2e-llm-inference-service] if assert_same_host and not self.is_same_host(url): [e2e-llm-inference-service] raise HostChangedError(self, url, retries) [e2e-llm-inference-service] [e2e-llm-inference-service] # Ensure that the URL we're connecting to is properly encoded [e2e-llm-inference-service] if url.startswith("/"): [e2e-llm-inference-service] url = to_str(_encode_target(url)) [e2e-llm-inference-service] else: [e2e-llm-inference-service] url = to_str(parsed_url.url) [e2e-llm-inference-service] [e2e-llm-inference-service] conn = None [e2e-llm-inference-service] [e2e-llm-inference-service] # Track whether `conn` needs to be released before [e2e-llm-inference-service] # returning/raising/recursing. Update this variable if necessary, and [e2e-llm-inference-service] # leave `release_conn` constant throughout the function. That way, if [e2e-llm-inference-service] # the function recurses, the original value of `release_conn` will be [e2e-llm-inference-service] # passed down into the recursive call, and its value will be respected. [e2e-llm-inference-service] # [e2e-llm-inference-service] # See issue #651 [1] for details. [e2e-llm-inference-service] # [e2e-llm-inference-service] # [1] [e2e-llm-inference-service] release_this_conn = release_conn [e2e-llm-inference-service] [e2e-llm-inference-service] http_tunnel_required = connection_requires_http_tunnel( [e2e-llm-inference-service] self.proxy, self.proxy_config, destination_scheme [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] # Merge the proxy headers. Only done when not using HTTP CONNECT. We [e2e-llm-inference-service] # have to copy the headers dict so we can safely change it without those [e2e-llm-inference-service] # changes being reflected in anyone else's copy. [e2e-llm-inference-service] if not http_tunnel_required: [e2e-llm-inference-service] headers = headers.copy() # type: ignore[attr-defined] [e2e-llm-inference-service] headers.update(self.proxy_headers) # type: ignore[union-attr] [e2e-llm-inference-service] [e2e-llm-inference-service] # Must keep the exception bound to a separate variable or else Python 3 [e2e-llm-inference-service] # complains about UnboundLocalError. [e2e-llm-inference-service] err = None [e2e-llm-inference-service] [e2e-llm-inference-service] # Keep track of whether we cleanly exited the except block. This [e2e-llm-inference-service] # ensures we do proper cleanup in finally. [e2e-llm-inference-service] clean_exit = False [e2e-llm-inference-service] [e2e-llm-inference-service] # Rewind body position, if needed. Record current position [e2e-llm-inference-service] # for future rewinds in the event of a redirect/retry. [e2e-llm-inference-service] body_pos = set_file_position(body, body_pos) [e2e-llm-inference-service] [e2e-llm-inference-service] try: [e2e-llm-inference-service] # Request a connection from the queue. [e2e-llm-inference-service] timeout_obj = self._get_timeout(timeout) [e2e-llm-inference-service] conn = self._get_conn(timeout=pool_timeout) [e2e-llm-inference-service] [e2e-llm-inference-service] conn.timeout = timeout_obj.connect_timeout # type: ignore[assignment] [e2e-llm-inference-service] [e2e-llm-inference-service] # Is this a closed/new connection that requires CONNECT tunnelling? [e2e-llm-inference-service] if self.proxy is not None and http_tunnel_required and conn.is_closed: [e2e-llm-inference-service] try: [e2e-llm-inference-service] self._prepare_proxy(conn) [e2e-llm-inference-service] except (BaseSSLError, OSError, SocketTimeout) as e: [e2e-llm-inference-service] self._raise_timeout( [e2e-llm-inference-service] err=e, url=self.proxy.url, timeout_value=conn.timeout [e2e-llm-inference-service] ) [e2e-llm-inference-service] raise [e2e-llm-inference-service] [e2e-llm-inference-service] # If we're going to release the connection in ``finally:``, then [e2e-llm-inference-service] # the response doesn't need to know about the connection. Otherwise [e2e-llm-inference-service] # it will also try to release it and we'll have a double-release [e2e-llm-inference-service] # mess. [e2e-llm-inference-service] response_conn = conn if not release_conn else None [e2e-llm-inference-service] [e2e-llm-inference-service] # Make the request on the HTTPConnection object [e2e-llm-inference-service] response = self._make_request( [e2e-llm-inference-service] conn, [e2e-llm-inference-service] method, [e2e-llm-inference-service] url, [e2e-llm-inference-service] timeout=timeout_obj, [e2e-llm-inference-service] body=body, [e2e-llm-inference-service] headers=headers, [e2e-llm-inference-service] chunked=chunked, [e2e-llm-inference-service] retries=retries, [e2e-llm-inference-service] response_conn=response_conn, [e2e-llm-inference-service] preload_content=preload_content, [e2e-llm-inference-service] decode_content=decode_content, [e2e-llm-inference-service] **response_kw, [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] # Everything went great! [e2e-llm-inference-service] clean_exit = True [e2e-llm-inference-service] [e2e-llm-inference-service] except EmptyPoolError: [e2e-llm-inference-service] # Didn't get a connection from the pool, no need to clean up [e2e-llm-inference-service] clean_exit = True [e2e-llm-inference-service] release_this_conn = False [e2e-llm-inference-service] raise [e2e-llm-inference-service] [e2e-llm-inference-service] except ( [e2e-llm-inference-service] TimeoutError, [e2e-llm-inference-service] HTTPException, [e2e-llm-inference-service] OSError, [e2e-llm-inference-service] ProtocolError, [e2e-llm-inference-service] BaseSSLError, [e2e-llm-inference-service] SSLError, [e2e-llm-inference-service] CertificateError, [e2e-llm-inference-service] ProxyError, [e2e-llm-inference-service] ) as e: [e2e-llm-inference-service] # Discard the connection for these exceptions. It will be [e2e-llm-inference-service] # replaced during the next _get_conn() call. [e2e-llm-inference-service] clean_exit = False [e2e-llm-inference-service] new_e: Exception = e [e2e-llm-inference-service] if isinstance(e, (BaseSSLError, CertificateError)): [e2e-llm-inference-service] new_e = SSLError(e) [e2e-llm-inference-service] if isinstance( [e2e-llm-inference-service] new_e, [e2e-llm-inference-service] ( [e2e-llm-inference-service] OSError, [e2e-llm-inference-service] NewConnectionError, [e2e-llm-inference-service] TimeoutError, [e2e-llm-inference-service] SSLError, [e2e-llm-inference-service] HTTPException, [e2e-llm-inference-service] ), [e2e-llm-inference-service] ) and (conn and conn.proxy and not conn.has_connected_to_proxy): [e2e-llm-inference-service] new_e = _wrap_proxy_error(new_e, conn.proxy.scheme) [e2e-llm-inference-service] elif isinstance(new_e, (OSError, HTTPException)): [e2e-llm-inference-service] new_e = ProtocolError("Connection aborted.", new_e) [e2e-llm-inference-service] [e2e-llm-inference-service] retries = retries.increment( [e2e-llm-inference-service] method, url, error=new_e, _pool=self, _stacktrace=sys.exc_info()[2] [e2e-llm-inference-service] ) [e2e-llm-inference-service] retries.sleep() [e2e-llm-inference-service] [e2e-llm-inference-service] # Keep track of the error for the retry warning. [e2e-llm-inference-service] err = e [e2e-llm-inference-service] [e2e-llm-inference-service] finally: [e2e-llm-inference-service] if not clean_exit: [e2e-llm-inference-service] # We hit some kind of exception, handled or otherwise. We need [e2e-llm-inference-service] # to throw the connection away unless explicitly told not to. [e2e-llm-inference-service] # Close the connection, set the variable to None, and make sure [e2e-llm-inference-service] # we put the None back in the pool to avoid leaking it. [e2e-llm-inference-service] if conn: [e2e-llm-inference-service] conn.close() [e2e-llm-inference-service] conn = None [e2e-llm-inference-service] release_this_conn = True [e2e-llm-inference-service] [e2e-llm-inference-service] if release_this_conn: [e2e-llm-inference-service] # Put the connection back to be reused. If the connection is [e2e-llm-inference-service] # expired then it will be None, which will get replaced with a [e2e-llm-inference-service] # fresh connection during _get_conn. [e2e-llm-inference-service] self._put_conn(conn) [e2e-llm-inference-service] [e2e-llm-inference-service] if not conn: [e2e-llm-inference-service] # Try again [e2e-llm-inference-service] log.warning( [e2e-llm-inference-service] "Retrying (%r) after connection broken by '%r': %s", retries, err, url [e2e-llm-inference-service] ) [e2e-llm-inference-service] > return self.urlopen( [e2e-llm-inference-service] method, [e2e-llm-inference-service] url, [e2e-llm-inference-service] body, [e2e-llm-inference-service] headers, [e2e-llm-inference-service] retries, [e2e-llm-inference-service] redirect, [e2e-llm-inference-service] assert_same_host, [e2e-llm-inference-service] timeout=timeout, [e2e-llm-inference-service] pool_timeout=pool_timeout, [e2e-llm-inference-service] release_conn=release_conn, [e2e-llm-inference-service] chunked=chunked, [e2e-llm-inference-service] body_pos=body_pos, [e2e-llm-inference-service] preload_content=preload_content, [e2e-llm-inference-service] decode_content=decode_content, [e2e-llm-inference-service] **response_kw, [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] ../../python/kserve/.venv/lib64/python3.11/site-packages/urllib3/connectionpool.py:871: [e2e-llm-inference-service] _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ [e2e-llm-inference-service] [e2e-llm-inference-service] self = [e2e-llm-inference-service] method = 'GET', url = '/v1/models', body = None [e2e-llm-inference-service] headers = {'User-Agent': 'python-requests/2.32.3', 'Accept-Encoding': 'gzip, deflate', 'Accept': '*/*', 'Connection': 'keep-alive', 'Content-Type': 'application/json', 'X-Gateway-Model-Name': 'publishers/kserve-ci-e2e-test/models/facebook/opt-125m'} [e2e-llm-inference-service] retries = Retry(total=4, connect=None, read=None, redirect=None, status=None) [e2e-llm-inference-service] redirect = False, assert_same_host = False [e2e-llm-inference-service] timeout = Timeout(connect=60, read=60, total=None), pool_timeout = None [e2e-llm-inference-service] release_conn = False, chunked = False, body_pos = None, preload_content = False [e2e-llm-inference-service] decode_content = False, response_kw = {} [e2e-llm-inference-service] parsed_url = Url(scheme=None, auth=None, host=None, port=None, path='/v1/models', query=None, fragment=None) [e2e-llm-inference-service] destination_scheme = None, conn = None, release_this_conn = True [e2e-llm-inference-service] http_tunnel_required = False [e2e-llm-inference-service] err = ReadTimeoutError("HTTPConnectionPool(host='ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com', port=80): Read timed out. (read timeout=60)") [e2e-llm-inference-service] clean_exit = False [e2e-llm-inference-service] [e2e-llm-inference-service] def urlopen( # type: ignore[override] [e2e-llm-inference-service] self, [e2e-llm-inference-service] method: str, [e2e-llm-inference-service] url: str, [e2e-llm-inference-service] body: _TYPE_BODY | None = None, [e2e-llm-inference-service] headers: typing.Mapping[str, str] | None = None, [e2e-llm-inference-service] retries: Retry | bool | int | None = None, [e2e-llm-inference-service] redirect: bool = True, [e2e-llm-inference-service] assert_same_host: bool = True, [e2e-llm-inference-service] timeout: _TYPE_TIMEOUT = _DEFAULT_TIMEOUT, [e2e-llm-inference-service] pool_timeout: int | None = None, [e2e-llm-inference-service] release_conn: bool | None = None, [e2e-llm-inference-service] chunked: bool = False, [e2e-llm-inference-service] body_pos: _TYPE_BODY_POSITION | None = None, [e2e-llm-inference-service] preload_content: bool = True, [e2e-llm-inference-service] decode_content: bool = True, [e2e-llm-inference-service] **response_kw: typing.Any, [e2e-llm-inference-service] ) -> BaseHTTPResponse: [e2e-llm-inference-service] """ [e2e-llm-inference-service] Get a connection from the pool and perform an HTTP request. This is the [e2e-llm-inference-service] lowest level call for making a request, so you'll need to specify all [e2e-llm-inference-service] the raw details. [e2e-llm-inference-service] [e2e-llm-inference-service] .. note:: [e2e-llm-inference-service] [e2e-llm-inference-service] More commonly, it's appropriate to use a convenience method [e2e-llm-inference-service] such as :meth:`request`. [e2e-llm-inference-service] [e2e-llm-inference-service] .. note:: [e2e-llm-inference-service] [e2e-llm-inference-service] `release_conn` will only behave as expected if [e2e-llm-inference-service] `preload_content=False` because we want to make [e2e-llm-inference-service] `preload_content=False` the default behaviour someday soon without [e2e-llm-inference-service] breaking backwards compatibility. [e2e-llm-inference-service] [e2e-llm-inference-service] :param method: [e2e-llm-inference-service] HTTP request method (such as GET, POST, PUT, etc.) [e2e-llm-inference-service] [e2e-llm-inference-service] :param url: [e2e-llm-inference-service] The URL to perform the request on. [e2e-llm-inference-service] [e2e-llm-inference-service] :param body: [e2e-llm-inference-service] Data to send in the request body, either :class:`str`, :class:`bytes`, [e2e-llm-inference-service] an iterable of :class:`str`/:class:`bytes`, or a file-like object. [e2e-llm-inference-service] [e2e-llm-inference-service] :param headers: [e2e-llm-inference-service] Dictionary of custom headers to send, such as User-Agent, [e2e-llm-inference-service] If-None-Match, etc. If None, pool headers are used. If provided, [e2e-llm-inference-service] these headers completely replace any pool-specific headers. [e2e-llm-inference-service] [e2e-llm-inference-service] :param retries: [e2e-llm-inference-service] Configure the number of retries to allow before raising a [e2e-llm-inference-service] :class:`~urllib3.exceptions.MaxRetryError` exception. [e2e-llm-inference-service] [e2e-llm-inference-service] If ``None`` (default) will retry 3 times, see ``Retry.DEFAULT``. Pass a [e2e-llm-inference-service] :class:`~urllib3.util.retry.Retry` object for fine-grained control [e2e-llm-inference-service] over different types of retries. [e2e-llm-inference-service] Pass an integer number to retry connection errors that many times, [e2e-llm-inference-service] but no other types of errors. Pass zero to never retry. [e2e-llm-inference-service] [e2e-llm-inference-service] If ``False``, then retries are disabled and any exception is raised [e2e-llm-inference-service] immediately. Also, instead of raising a MaxRetryError on redirects, [e2e-llm-inference-service] the redirect response will be returned. [e2e-llm-inference-service] [e2e-llm-inference-service] :type retries: :class:`~urllib3.util.retry.Retry`, False, or an int. [e2e-llm-inference-service] [e2e-llm-inference-service] :param redirect: [e2e-llm-inference-service] If True, automatically handle redirects (status codes 301, 302, [e2e-llm-inference-service] 303, 307, 308). Each redirect counts as a retry. Disabling retries [e2e-llm-inference-service] will disable redirect, too. [e2e-llm-inference-service] [e2e-llm-inference-service] :param assert_same_host: [e2e-llm-inference-service] If ``True``, will make sure that the host of the pool requests is [e2e-llm-inference-service] consistent else will raise HostChangedError. When ``False``, you can [e2e-llm-inference-service] use the pool on an HTTP proxy and request foreign hosts. [e2e-llm-inference-service] [e2e-llm-inference-service] :param timeout: [e2e-llm-inference-service] If specified, overrides the default timeout for this one [e2e-llm-inference-service] request. It may be a float (in seconds) or an instance of [e2e-llm-inference-service] :class:`urllib3.util.Timeout`. [e2e-llm-inference-service] [e2e-llm-inference-service] :param pool_timeout: [e2e-llm-inference-service] If set and the pool is set to block=True, then this method will [e2e-llm-inference-service] block for ``pool_timeout`` seconds and raise EmptyPoolError if no [e2e-llm-inference-service] connection is available within the time period. [e2e-llm-inference-service] [e2e-llm-inference-service] :param bool preload_content: [e2e-llm-inference-service] If True, the response's body will be preloaded into memory. [e2e-llm-inference-service] [e2e-llm-inference-service] :param bool decode_content: [e2e-llm-inference-service] If True, will attempt to decode the body based on the [e2e-llm-inference-service] 'content-encoding' header. [e2e-llm-inference-service] [e2e-llm-inference-service] :param release_conn: [e2e-llm-inference-service] If False, then the urlopen call will not release the connection [e2e-llm-inference-service] back into the pool once a response is received (but will release if [e2e-llm-inference-service] you read the entire contents of the response such as when [e2e-llm-inference-service] `preload_content=True`). This is useful if you're not preloading [e2e-llm-inference-service] the response's content immediately. You will need to call [e2e-llm-inference-service] ``r.release_conn()`` on the response ``r`` to return the connection [e2e-llm-inference-service] back into the pool. If None, it takes the value of ``preload_content`` [e2e-llm-inference-service] which defaults to ``True``. [e2e-llm-inference-service] [e2e-llm-inference-service] :param bool chunked: [e2e-llm-inference-service] If True, urllib3 will send the body using chunked transfer [e2e-llm-inference-service] encoding. Otherwise, urllib3 will send the body using the standard [e2e-llm-inference-service] content-length form. Defaults to False. [e2e-llm-inference-service] [e2e-llm-inference-service] :param int body_pos: [e2e-llm-inference-service] Position to seek to in file-like body in the event of a retry or [e2e-llm-inference-service] redirect. Typically this won't need to be set because urllib3 will [e2e-llm-inference-service] auto-populate the value when needed. [e2e-llm-inference-service] """ [e2e-llm-inference-service] parsed_url = parse_url(url) [e2e-llm-inference-service] destination_scheme = parsed_url.scheme [e2e-llm-inference-service] [e2e-llm-inference-service] if headers is None: [e2e-llm-inference-service] headers = self.headers [e2e-llm-inference-service] [e2e-llm-inference-service] if not isinstance(retries, Retry): [e2e-llm-inference-service] retries = Retry.from_int(retries, redirect=redirect, default=self.retries) [e2e-llm-inference-service] [e2e-llm-inference-service] if release_conn is None: [e2e-llm-inference-service] release_conn = preload_content [e2e-llm-inference-service] [e2e-llm-inference-service] # Check host [e2e-llm-inference-service] if assert_same_host and not self.is_same_host(url): [e2e-llm-inference-service] raise HostChangedError(self, url, retries) [e2e-llm-inference-service] [e2e-llm-inference-service] # Ensure that the URL we're connecting to is properly encoded [e2e-llm-inference-service] if url.startswith("/"): [e2e-llm-inference-service] url = to_str(_encode_target(url)) [e2e-llm-inference-service] else: [e2e-llm-inference-service] url = to_str(parsed_url.url) [e2e-llm-inference-service] [e2e-llm-inference-service] conn = None [e2e-llm-inference-service] [e2e-llm-inference-service] # Track whether `conn` needs to be released before [e2e-llm-inference-service] # returning/raising/recursing. Update this variable if necessary, and [e2e-llm-inference-service] # leave `release_conn` constant throughout the function. That way, if [e2e-llm-inference-service] # the function recurses, the original value of `release_conn` will be [e2e-llm-inference-service] # passed down into the recursive call, and its value will be respected. [e2e-llm-inference-service] # [e2e-llm-inference-service] # See issue #651 [1] for details. [e2e-llm-inference-service] # [e2e-llm-inference-service] # [1] [e2e-llm-inference-service] release_this_conn = release_conn [e2e-llm-inference-service] [e2e-llm-inference-service] http_tunnel_required = connection_requires_http_tunnel( [e2e-llm-inference-service] self.proxy, self.proxy_config, destination_scheme [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] # Merge the proxy headers. Only done when not using HTTP CONNECT. We [e2e-llm-inference-service] # have to copy the headers dict so we can safely change it without those [e2e-llm-inference-service] # changes being reflected in anyone else's copy. [e2e-llm-inference-service] if not http_tunnel_required: [e2e-llm-inference-service] headers = headers.copy() # type: ignore[attr-defined] [e2e-llm-inference-service] headers.update(self.proxy_headers) # type: ignore[union-attr] [e2e-llm-inference-service] [e2e-llm-inference-service] # Must keep the exception bound to a separate variable or else Python 3 [e2e-llm-inference-service] # complains about UnboundLocalError. [e2e-llm-inference-service] err = None [e2e-llm-inference-service] [e2e-llm-inference-service] # Keep track of whether we cleanly exited the except block. This [e2e-llm-inference-service] # ensures we do proper cleanup in finally. [e2e-llm-inference-service] clean_exit = False [e2e-llm-inference-service] [e2e-llm-inference-service] # Rewind body position, if needed. Record current position [e2e-llm-inference-service] # for future rewinds in the event of a redirect/retry. [e2e-llm-inference-service] body_pos = set_file_position(body, body_pos) [e2e-llm-inference-service] [e2e-llm-inference-service] try: [e2e-llm-inference-service] # Request a connection from the queue. [e2e-llm-inference-service] timeout_obj = self._get_timeout(timeout) [e2e-llm-inference-service] conn = self._get_conn(timeout=pool_timeout) [e2e-llm-inference-service] [e2e-llm-inference-service] conn.timeout = timeout_obj.connect_timeout # type: ignore[assignment] [e2e-llm-inference-service] [e2e-llm-inference-service] # Is this a closed/new connection that requires CONNECT tunnelling? [e2e-llm-inference-service] if self.proxy is not None and http_tunnel_required and conn.is_closed: [e2e-llm-inference-service] try: [e2e-llm-inference-service] self._prepare_proxy(conn) [e2e-llm-inference-service] except (BaseSSLError, OSError, SocketTimeout) as e: [e2e-llm-inference-service] self._raise_timeout( [e2e-llm-inference-service] err=e, url=self.proxy.url, timeout_value=conn.timeout [e2e-llm-inference-service] ) [e2e-llm-inference-service] raise [e2e-llm-inference-service] [e2e-llm-inference-service] # If we're going to release the connection in ``finally:``, then [e2e-llm-inference-service] # the response doesn't need to know about the connection. Otherwise [e2e-llm-inference-service] # it will also try to release it and we'll have a double-release [e2e-llm-inference-service] # mess. [e2e-llm-inference-service] response_conn = conn if not release_conn else None [e2e-llm-inference-service] [e2e-llm-inference-service] # Make the request on the HTTPConnection object [e2e-llm-inference-service] response = self._make_request( [e2e-llm-inference-service] conn, [e2e-llm-inference-service] method, [e2e-llm-inference-service] url, [e2e-llm-inference-service] timeout=timeout_obj, [e2e-llm-inference-service] body=body, [e2e-llm-inference-service] headers=headers, [e2e-llm-inference-service] chunked=chunked, [e2e-llm-inference-service] retries=retries, [e2e-llm-inference-service] response_conn=response_conn, [e2e-llm-inference-service] preload_content=preload_content, [e2e-llm-inference-service] decode_content=decode_content, [e2e-llm-inference-service] **response_kw, [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] # Everything went great! [e2e-llm-inference-service] clean_exit = True [e2e-llm-inference-service] [e2e-llm-inference-service] except EmptyPoolError: [e2e-llm-inference-service] # Didn't get a connection from the pool, no need to clean up [e2e-llm-inference-service] clean_exit = True [e2e-llm-inference-service] release_this_conn = False [e2e-llm-inference-service] raise [e2e-llm-inference-service] [e2e-llm-inference-service] except ( [e2e-llm-inference-service] TimeoutError, [e2e-llm-inference-service] HTTPException, [e2e-llm-inference-service] OSError, [e2e-llm-inference-service] ProtocolError, [e2e-llm-inference-service] BaseSSLError, [e2e-llm-inference-service] SSLError, [e2e-llm-inference-service] CertificateError, [e2e-llm-inference-service] ProxyError, [e2e-llm-inference-service] ) as e: [e2e-llm-inference-service] # Discard the connection for these exceptions. It will be [e2e-llm-inference-service] # replaced during the next _get_conn() call. [e2e-llm-inference-service] clean_exit = False [e2e-llm-inference-service] new_e: Exception = e [e2e-llm-inference-service] if isinstance(e, (BaseSSLError, CertificateError)): [e2e-llm-inference-service] new_e = SSLError(e) [e2e-llm-inference-service] if isinstance( [e2e-llm-inference-service] new_e, [e2e-llm-inference-service] ( [e2e-llm-inference-service] OSError, [e2e-llm-inference-service] NewConnectionError, [e2e-llm-inference-service] TimeoutError, [e2e-llm-inference-service] SSLError, [e2e-llm-inference-service] HTTPException, [e2e-llm-inference-service] ), [e2e-llm-inference-service] ) and (conn and conn.proxy and not conn.has_connected_to_proxy): [e2e-llm-inference-service] new_e = _wrap_proxy_error(new_e, conn.proxy.scheme) [e2e-llm-inference-service] elif isinstance(new_e, (OSError, HTTPException)): [e2e-llm-inference-service] new_e = ProtocolError("Connection aborted.", new_e) [e2e-llm-inference-service] [e2e-llm-inference-service] retries = retries.increment( [e2e-llm-inference-service] method, url, error=new_e, _pool=self, _stacktrace=sys.exc_info()[2] [e2e-llm-inference-service] ) [e2e-llm-inference-service] retries.sleep() [e2e-llm-inference-service] [e2e-llm-inference-service] # Keep track of the error for the retry warning. [e2e-llm-inference-service] err = e [e2e-llm-inference-service] [e2e-llm-inference-service] finally: [e2e-llm-inference-service] if not clean_exit: [e2e-llm-inference-service] # We hit some kind of exception, handled or otherwise. We need [e2e-llm-inference-service] # to throw the connection away unless explicitly told not to. [e2e-llm-inference-service] # Close the connection, set the variable to None, and make sure [e2e-llm-inference-service] # we put the None back in the pool to avoid leaking it. [e2e-llm-inference-service] if conn: [e2e-llm-inference-service] conn.close() [e2e-llm-inference-service] conn = None [e2e-llm-inference-service] release_this_conn = True [e2e-llm-inference-service] [e2e-llm-inference-service] if release_this_conn: [e2e-llm-inference-service] # Put the connection back to be reused. If the connection is [e2e-llm-inference-service] # expired then it will be None, which will get replaced with a [e2e-llm-inference-service] # fresh connection during _get_conn. [e2e-llm-inference-service] self._put_conn(conn) [e2e-llm-inference-service] [e2e-llm-inference-service] if not conn: [e2e-llm-inference-service] # Try again [e2e-llm-inference-service] log.warning( [e2e-llm-inference-service] "Retrying (%r) after connection broken by '%r': %s", retries, err, url [e2e-llm-inference-service] ) [e2e-llm-inference-service] > return self.urlopen( [e2e-llm-inference-service] method, [e2e-llm-inference-service] url, [e2e-llm-inference-service] body, [e2e-llm-inference-service] headers, [e2e-llm-inference-service] retries, [e2e-llm-inference-service] redirect, [e2e-llm-inference-service] assert_same_host, [e2e-llm-inference-service] timeout=timeout, [e2e-llm-inference-service] pool_timeout=pool_timeout, [e2e-llm-inference-service] release_conn=release_conn, [e2e-llm-inference-service] chunked=chunked, [e2e-llm-inference-service] body_pos=body_pos, [e2e-llm-inference-service] preload_content=preload_content, [e2e-llm-inference-service] decode_content=decode_content, [e2e-llm-inference-service] **response_kw, [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] ../../python/kserve/.venv/lib64/python3.11/site-packages/urllib3/connectionpool.py:871: [e2e-llm-inference-service] _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ [e2e-llm-inference-service] [e2e-llm-inference-service] self = [e2e-llm-inference-service] method = 'GET', url = '/v1/models', body = None [e2e-llm-inference-service] headers = {'User-Agent': 'python-requests/2.32.3', 'Accept-Encoding': 'gzip, deflate', 'Accept': '*/*', 'Connection': 'keep-alive', 'Content-Type': 'application/json', 'X-Gateway-Model-Name': 'publishers/kserve-ci-e2e-test/models/facebook/opt-125m'} [e2e-llm-inference-service] retries = Retry(total=3, connect=None, read=None, redirect=None, status=None) [e2e-llm-inference-service] redirect = False, assert_same_host = False [e2e-llm-inference-service] timeout = Timeout(connect=60, read=60, total=None), pool_timeout = None [e2e-llm-inference-service] release_conn = False, chunked = False, body_pos = None, preload_content = False [e2e-llm-inference-service] decode_content = False, response_kw = {} [e2e-llm-inference-service] parsed_url = Url(scheme=None, auth=None, host=None, port=None, path='/v1/models', query=None, fragment=None) [e2e-llm-inference-service] destination_scheme = None, conn = None, release_this_conn = True [e2e-llm-inference-service] http_tunnel_required = False [e2e-llm-inference-service] err = ReadTimeoutError("HTTPConnectionPool(host='ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com', port=80): Read timed out. (read timeout=60)") [e2e-llm-inference-service] clean_exit = False [e2e-llm-inference-service] [e2e-llm-inference-service] def urlopen( # type: ignore[override] [e2e-llm-inference-service] self, [e2e-llm-inference-service] method: str, [e2e-llm-inference-service] url: str, [e2e-llm-inference-service] body: _TYPE_BODY | None = None, [e2e-llm-inference-service] headers: typing.Mapping[str, str] | None = None, [e2e-llm-inference-service] retries: Retry | bool | int | None = None, [e2e-llm-inference-service] redirect: bool = True, [e2e-llm-inference-service] assert_same_host: bool = True, [e2e-llm-inference-service] timeout: _TYPE_TIMEOUT = _DEFAULT_TIMEOUT, [e2e-llm-inference-service] pool_timeout: int | None = None, [e2e-llm-inference-service] release_conn: bool | None = None, [e2e-llm-inference-service] chunked: bool = False, [e2e-llm-inference-service] body_pos: _TYPE_BODY_POSITION | None = None, [e2e-llm-inference-service] preload_content: bool = True, [e2e-llm-inference-service] decode_content: bool = True, [e2e-llm-inference-service] **response_kw: typing.Any, [e2e-llm-inference-service] ) -> BaseHTTPResponse: [e2e-llm-inference-service] """ [e2e-llm-inference-service] Get a connection from the pool and perform an HTTP request. This is the [e2e-llm-inference-service] lowest level call for making a request, so you'll need to specify all [e2e-llm-inference-service] the raw details. [e2e-llm-inference-service] [e2e-llm-inference-service] .. note:: [e2e-llm-inference-service] [e2e-llm-inference-service] More commonly, it's appropriate to use a convenience method [e2e-llm-inference-service] such as :meth:`request`. [e2e-llm-inference-service] [e2e-llm-inference-service] .. note:: [e2e-llm-inference-service] [e2e-llm-inference-service] `release_conn` will only behave as expected if [e2e-llm-inference-service] `preload_content=False` because we want to make [e2e-llm-inference-service] `preload_content=False` the default behaviour someday soon without [e2e-llm-inference-service] breaking backwards compatibility. [e2e-llm-inference-service] [e2e-llm-inference-service] :param method: [e2e-llm-inference-service] HTTP request method (such as GET, POST, PUT, etc.) [e2e-llm-inference-service] [e2e-llm-inference-service] :param url: [e2e-llm-inference-service] The URL to perform the request on. [e2e-llm-inference-service] [e2e-llm-inference-service] :param body: [e2e-llm-inference-service] Data to send in the request body, either :class:`str`, :class:`bytes`, [e2e-llm-inference-service] an iterable of :class:`str`/:class:`bytes`, or a file-like object. [e2e-llm-inference-service] [e2e-llm-inference-service] :param headers: [e2e-llm-inference-service] Dictionary of custom headers to send, such as User-Agent, [e2e-llm-inference-service] If-None-Match, etc. If None, pool headers are used. If provided, [e2e-llm-inference-service] these headers completely replace any pool-specific headers. [e2e-llm-inference-service] [e2e-llm-inference-service] :param retries: [e2e-llm-inference-service] Configure the number of retries to allow before raising a [e2e-llm-inference-service] :class:`~urllib3.exceptions.MaxRetryError` exception. [e2e-llm-inference-service] [e2e-llm-inference-service] If ``None`` (default) will retry 3 times, see ``Retry.DEFAULT``. Pass a [e2e-llm-inference-service] :class:`~urllib3.util.retry.Retry` object for fine-grained control [e2e-llm-inference-service] over different types of retries. [e2e-llm-inference-service] Pass an integer number to retry connection errors that many times, [e2e-llm-inference-service] but no other types of errors. Pass zero to never retry. [e2e-llm-inference-service] [e2e-llm-inference-service] If ``False``, then retries are disabled and any exception is raised [e2e-llm-inference-service] immediately. Also, instead of raising a MaxRetryError on redirects, [e2e-llm-inference-service] the redirect response will be returned. [e2e-llm-inference-service] [e2e-llm-inference-service] :type retries: :class:`~urllib3.util.retry.Retry`, False, or an int. [e2e-llm-inference-service] [e2e-llm-inference-service] :param redirect: [e2e-llm-inference-service] If True, automatically handle redirects (status codes 301, 302, [e2e-llm-inference-service] 303, 307, 308). Each redirect counts as a retry. Disabling retries [e2e-llm-inference-service] will disable redirect, too. [e2e-llm-inference-service] [e2e-llm-inference-service] :param assert_same_host: [e2e-llm-inference-service] If ``True``, will make sure that the host of the pool requests is [e2e-llm-inference-service] consistent else will raise HostChangedError. When ``False``, you can [e2e-llm-inference-service] use the pool on an HTTP proxy and request foreign hosts. [e2e-llm-inference-service] [e2e-llm-inference-service] :param timeout: [e2e-llm-inference-service] If specified, overrides the default timeout for this one [e2e-llm-inference-service] request. It may be a float (in seconds) or an instance of [e2e-llm-inference-service] :class:`urllib3.util.Timeout`. [e2e-llm-inference-service] [e2e-llm-inference-service] :param pool_timeout: [e2e-llm-inference-service] If set and the pool is set to block=True, then this method will [e2e-llm-inference-service] block for ``pool_timeout`` seconds and raise EmptyPoolError if no [e2e-llm-inference-service] connection is available within the time period. [e2e-llm-inference-service] [e2e-llm-inference-service] :param bool preload_content: [e2e-llm-inference-service] If True, the response's body will be preloaded into memory. [e2e-llm-inference-service] [e2e-llm-inference-service] :param bool decode_content: [e2e-llm-inference-service] If True, will attempt to decode the body based on the [e2e-llm-inference-service] 'content-encoding' header. [e2e-llm-inference-service] [e2e-llm-inference-service] :param release_conn: [e2e-llm-inference-service] If False, then the urlopen call will not release the connection [e2e-llm-inference-service] back into the pool once a response is received (but will release if [e2e-llm-inference-service] you read the entire contents of the response such as when [e2e-llm-inference-service] `preload_content=True`). This is useful if you're not preloading [e2e-llm-inference-service] the response's content immediately. You will need to call [e2e-llm-inference-service] ``r.release_conn()`` on the response ``r`` to return the connection [e2e-llm-inference-service] back into the pool. If None, it takes the value of ``preload_content`` [e2e-llm-inference-service] which defaults to ``True``. [e2e-llm-inference-service] [e2e-llm-inference-service] :param bool chunked: [e2e-llm-inference-service] If True, urllib3 will send the body using chunked transfer [e2e-llm-inference-service] encoding. Otherwise, urllib3 will send the body using the standard [e2e-llm-inference-service] content-length form. Defaults to False. [e2e-llm-inference-service] [e2e-llm-inference-service] :param int body_pos: [e2e-llm-inference-service] Position to seek to in file-like body in the event of a retry or [e2e-llm-inference-service] redirect. Typically this won't need to be set because urllib3 will [e2e-llm-inference-service] auto-populate the value when needed. [e2e-llm-inference-service] """ [e2e-llm-inference-service] parsed_url = parse_url(url) [e2e-llm-inference-service] destination_scheme = parsed_url.scheme [e2e-llm-inference-service] [e2e-llm-inference-service] if headers is None: [e2e-llm-inference-service] headers = self.headers [e2e-llm-inference-service] [e2e-llm-inference-service] if not isinstance(retries, Retry): [e2e-llm-inference-service] retries = Retry.from_int(retries, redirect=redirect, default=self.retries) [e2e-llm-inference-service] [e2e-llm-inference-service] if release_conn is None: [e2e-llm-inference-service] release_conn = preload_content [e2e-llm-inference-service] [e2e-llm-inference-service] # Check host [e2e-llm-inference-service] if assert_same_host and not self.is_same_host(url): [e2e-llm-inference-service] raise HostChangedError(self, url, retries) [e2e-llm-inference-service] [e2e-llm-inference-service] # Ensure that the URL we're connecting to is properly encoded [e2e-llm-inference-service] if url.startswith("/"): [e2e-llm-inference-service] url = to_str(_encode_target(url)) [e2e-llm-inference-service] else: [e2e-llm-inference-service] url = to_str(parsed_url.url) [e2e-llm-inference-service] [e2e-llm-inference-service] conn = None [e2e-llm-inference-service] [e2e-llm-inference-service] # Track whether `conn` needs to be released before [e2e-llm-inference-service] # returning/raising/recursing. Update this variable if necessary, and [e2e-llm-inference-service] # leave `release_conn` constant throughout the function. That way, if [e2e-llm-inference-service] # the function recurses, the original value of `release_conn` will be [e2e-llm-inference-service] # passed down into the recursive call, and its value will be respected. [e2e-llm-inference-service] # [e2e-llm-inference-service] # See issue #651 [1] for details. [e2e-llm-inference-service] # [e2e-llm-inference-service] # [1] [e2e-llm-inference-service] release_this_conn = release_conn [e2e-llm-inference-service] [e2e-llm-inference-service] http_tunnel_required = connection_requires_http_tunnel( [e2e-llm-inference-service] self.proxy, self.proxy_config, destination_scheme [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] # Merge the proxy headers. Only done when not using HTTP CONNECT. We [e2e-llm-inference-service] # have to copy the headers dict so we can safely change it without those [e2e-llm-inference-service] # changes being reflected in anyone else's copy. [e2e-llm-inference-service] if not http_tunnel_required: [e2e-llm-inference-service] headers = headers.copy() # type: ignore[attr-defined] [e2e-llm-inference-service] headers.update(self.proxy_headers) # type: ignore[union-attr] [e2e-llm-inference-service] [e2e-llm-inference-service] # Must keep the exception bound to a separate variable or else Python 3 [e2e-llm-inference-service] # complains about UnboundLocalError. [e2e-llm-inference-service] err = None [e2e-llm-inference-service] [e2e-llm-inference-service] # Keep track of whether we cleanly exited the except block. This [e2e-llm-inference-service] # ensures we do proper cleanup in finally. [e2e-llm-inference-service] clean_exit = False [e2e-llm-inference-service] [e2e-llm-inference-service] # Rewind body position, if needed. Record current position [e2e-llm-inference-service] # for future rewinds in the event of a redirect/retry. [e2e-llm-inference-service] body_pos = set_file_position(body, body_pos) [e2e-llm-inference-service] [e2e-llm-inference-service] try: [e2e-llm-inference-service] # Request a connection from the queue. [e2e-llm-inference-service] timeout_obj = self._get_timeout(timeout) [e2e-llm-inference-service] conn = self._get_conn(timeout=pool_timeout) [e2e-llm-inference-service] [e2e-llm-inference-service] conn.timeout = timeout_obj.connect_timeout # type: ignore[assignment] [e2e-llm-inference-service] [e2e-llm-inference-service] # Is this a closed/new connection that requires CONNECT tunnelling? [e2e-llm-inference-service] if self.proxy is not None and http_tunnel_required and conn.is_closed: [e2e-llm-inference-service] try: [e2e-llm-inference-service] self._prepare_proxy(conn) [e2e-llm-inference-service] except (BaseSSLError, OSError, SocketTimeout) as e: [e2e-llm-inference-service] self._raise_timeout( [e2e-llm-inference-service] err=e, url=self.proxy.url, timeout_value=conn.timeout [e2e-llm-inference-service] ) [e2e-llm-inference-service] raise [e2e-llm-inference-service] [e2e-llm-inference-service] # If we're going to release the connection in ``finally:``, then [e2e-llm-inference-service] # the response doesn't need to know about the connection. Otherwise [e2e-llm-inference-service] # it will also try to release it and we'll have a double-release [e2e-llm-inference-service] # mess. [e2e-llm-inference-service] response_conn = conn if not release_conn else None [e2e-llm-inference-service] [e2e-llm-inference-service] # Make the request on the HTTPConnection object [e2e-llm-inference-service] response = self._make_request( [e2e-llm-inference-service] conn, [e2e-llm-inference-service] method, [e2e-llm-inference-service] url, [e2e-llm-inference-service] timeout=timeout_obj, [e2e-llm-inference-service] body=body, [e2e-llm-inference-service] headers=headers, [e2e-llm-inference-service] chunked=chunked, [e2e-llm-inference-service] retries=retries, [e2e-llm-inference-service] response_conn=response_conn, [e2e-llm-inference-service] preload_content=preload_content, [e2e-llm-inference-service] decode_content=decode_content, [e2e-llm-inference-service] **response_kw, [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] # Everything went great! [e2e-llm-inference-service] clean_exit = True [e2e-llm-inference-service] [e2e-llm-inference-service] except EmptyPoolError: [e2e-llm-inference-service] # Didn't get a connection from the pool, no need to clean up [e2e-llm-inference-service] clean_exit = True [e2e-llm-inference-service] release_this_conn = False [e2e-llm-inference-service] raise [e2e-llm-inference-service] [e2e-llm-inference-service] except ( [e2e-llm-inference-service] TimeoutError, [e2e-llm-inference-service] HTTPException, [e2e-llm-inference-service] OSError, [e2e-llm-inference-service] ProtocolError, [e2e-llm-inference-service] BaseSSLError, [e2e-llm-inference-service] SSLError, [e2e-llm-inference-service] CertificateError, [e2e-llm-inference-service] ProxyError, [e2e-llm-inference-service] ) as e: [e2e-llm-inference-service] # Discard the connection for these exceptions. It will be [e2e-llm-inference-service] # replaced during the next _get_conn() call. [e2e-llm-inference-service] clean_exit = False [e2e-llm-inference-service] new_e: Exception = e [e2e-llm-inference-service] if isinstance(e, (BaseSSLError, CertificateError)): [e2e-llm-inference-service] new_e = SSLError(e) [e2e-llm-inference-service] if isinstance( [e2e-llm-inference-service] new_e, [e2e-llm-inference-service] ( [e2e-llm-inference-service] OSError, [e2e-llm-inference-service] NewConnectionError, [e2e-llm-inference-service] TimeoutError, [e2e-llm-inference-service] SSLError, [e2e-llm-inference-service] HTTPException, [e2e-llm-inference-service] ), [e2e-llm-inference-service] ) and (conn and conn.proxy and not conn.has_connected_to_proxy): [e2e-llm-inference-service] new_e = _wrap_proxy_error(new_e, conn.proxy.scheme) [e2e-llm-inference-service] elif isinstance(new_e, (OSError, HTTPException)): [e2e-llm-inference-service] new_e = ProtocolError("Connection aborted.", new_e) [e2e-llm-inference-service] [e2e-llm-inference-service] retries = retries.increment( [e2e-llm-inference-service] method, url, error=new_e, _pool=self, _stacktrace=sys.exc_info()[2] [e2e-llm-inference-service] ) [e2e-llm-inference-service] retries.sleep() [e2e-llm-inference-service] [e2e-llm-inference-service] # Keep track of the error for the retry warning. [e2e-llm-inference-service] err = e [e2e-llm-inference-service] [e2e-llm-inference-service] finally: [e2e-llm-inference-service] if not clean_exit: [e2e-llm-inference-service] # We hit some kind of exception, handled or otherwise. We need [e2e-llm-inference-service] # to throw the connection away unless explicitly told not to. [e2e-llm-inference-service] # Close the connection, set the variable to None, and make sure [e2e-llm-inference-service] # we put the None back in the pool to avoid leaking it. [e2e-llm-inference-service] if conn: [e2e-llm-inference-service] conn.close() [e2e-llm-inference-service] conn = None [e2e-llm-inference-service] release_this_conn = True [e2e-llm-inference-service] [e2e-llm-inference-service] if release_this_conn: [e2e-llm-inference-service] # Put the connection back to be reused. If the connection is [e2e-llm-inference-service] # expired then it will be None, which will get replaced with a [e2e-llm-inference-service] # fresh connection during _get_conn. [e2e-llm-inference-service] self._put_conn(conn) [e2e-llm-inference-service] [e2e-llm-inference-service] if not conn: [e2e-llm-inference-service] # Try again [e2e-llm-inference-service] log.warning( [e2e-llm-inference-service] "Retrying (%r) after connection broken by '%r': %s", retries, err, url [e2e-llm-inference-service] ) [e2e-llm-inference-service] > return self.urlopen( [e2e-llm-inference-service] method, [e2e-llm-inference-service] url, [e2e-llm-inference-service] body, [e2e-llm-inference-service] headers, [e2e-llm-inference-service] retries, [e2e-llm-inference-service] redirect, [e2e-llm-inference-service] assert_same_host, [e2e-llm-inference-service] timeout=timeout, [e2e-llm-inference-service] pool_timeout=pool_timeout, [e2e-llm-inference-service] release_conn=release_conn, [e2e-llm-inference-service] chunked=chunked, [e2e-llm-inference-service] body_pos=body_pos, [e2e-llm-inference-service] preload_content=preload_content, [e2e-llm-inference-service] decode_content=decode_content, [e2e-llm-inference-service] **response_kw, [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] ../../python/kserve/.venv/lib64/python3.11/site-packages/urllib3/connectionpool.py:871: [e2e-llm-inference-service] _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ [e2e-llm-inference-service] [e2e-llm-inference-service] self = [e2e-llm-inference-service] method = 'GET', url = '/v1/models', body = None [e2e-llm-inference-service] headers = {'User-Agent': 'python-requests/2.32.3', 'Accept-Encoding': 'gzip, deflate', 'Accept': '*/*', 'Connection': 'keep-alive', 'Content-Type': 'application/json', 'X-Gateway-Model-Name': 'publishers/kserve-ci-e2e-test/models/facebook/opt-125m'} [e2e-llm-inference-service] retries = Retry(total=2, connect=None, read=None, redirect=None, status=None) [e2e-llm-inference-service] redirect = False, assert_same_host = False [e2e-llm-inference-service] timeout = Timeout(connect=60, read=60, total=None), pool_timeout = None [e2e-llm-inference-service] release_conn = False, chunked = False, body_pos = None, preload_content = False [e2e-llm-inference-service] decode_content = False, response_kw = {} [e2e-llm-inference-service] parsed_url = Url(scheme=None, auth=None, host=None, port=None, path='/v1/models', query=None, fragment=None) [e2e-llm-inference-service] destination_scheme = None, conn = None, release_this_conn = True [e2e-llm-inference-service] http_tunnel_required = False [e2e-llm-inference-service] err = ReadTimeoutError("HTTPConnectionPool(host='ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com', port=80): Read timed out. (read timeout=60)") [e2e-llm-inference-service] clean_exit = False [e2e-llm-inference-service] [e2e-llm-inference-service] def urlopen( # type: ignore[override] [e2e-llm-inference-service] self, [e2e-llm-inference-service] method: str, [e2e-llm-inference-service] url: str, [e2e-llm-inference-service] body: _TYPE_BODY | None = None, [e2e-llm-inference-service] headers: typing.Mapping[str, str] | None = None, [e2e-llm-inference-service] retries: Retry | bool | int | None = None, [e2e-llm-inference-service] redirect: bool = True, [e2e-llm-inference-service] assert_same_host: bool = True, [e2e-llm-inference-service] timeout: _TYPE_TIMEOUT = _DEFAULT_TIMEOUT, [e2e-llm-inference-service] pool_timeout: int | None = None, [e2e-llm-inference-service] release_conn: bool | None = None, [e2e-llm-inference-service] chunked: bool = False, [e2e-llm-inference-service] body_pos: _TYPE_BODY_POSITION | None = None, [e2e-llm-inference-service] preload_content: bool = True, [e2e-llm-inference-service] decode_content: bool = True, [e2e-llm-inference-service] **response_kw: typing.Any, [e2e-llm-inference-service] ) -> BaseHTTPResponse: [e2e-llm-inference-service] """ [e2e-llm-inference-service] Get a connection from the pool and perform an HTTP request. This is the [e2e-llm-inference-service] lowest level call for making a request, so you'll need to specify all [e2e-llm-inference-service] the raw details. [e2e-llm-inference-service] [e2e-llm-inference-service] .. note:: [e2e-llm-inference-service] [e2e-llm-inference-service] More commonly, it's appropriate to use a convenience method [e2e-llm-inference-service] such as :meth:`request`. [e2e-llm-inference-service] [e2e-llm-inference-service] .. note:: [e2e-llm-inference-service] [e2e-llm-inference-service] `release_conn` will only behave as expected if [e2e-llm-inference-service] `preload_content=False` because we want to make [e2e-llm-inference-service] `preload_content=False` the default behaviour someday soon without [e2e-llm-inference-service] breaking backwards compatibility. [e2e-llm-inference-service] [e2e-llm-inference-service] :param method: [e2e-llm-inference-service] HTTP request method (such as GET, POST, PUT, etc.) [e2e-llm-inference-service] [e2e-llm-inference-service] :param url: [e2e-llm-inference-service] The URL to perform the request on. [e2e-llm-inference-service] [e2e-llm-inference-service] :param body: [e2e-llm-inference-service] Data to send in the request body, either :class:`str`, :class:`bytes`, [e2e-llm-inference-service] an iterable of :class:`str`/:class:`bytes`, or a file-like object. [e2e-llm-inference-service] [e2e-llm-inference-service] :param headers: [e2e-llm-inference-service] Dictionary of custom headers to send, such as User-Agent, [e2e-llm-inference-service] If-None-Match, etc. If None, pool headers are used. If provided, [e2e-llm-inference-service] these headers completely replace any pool-specific headers. [e2e-llm-inference-service] [e2e-llm-inference-service] :param retries: [e2e-llm-inference-service] Configure the number of retries to allow before raising a [e2e-llm-inference-service] :class:`~urllib3.exceptions.MaxRetryError` exception. [e2e-llm-inference-service] [e2e-llm-inference-service] If ``None`` (default) will retry 3 times, see ``Retry.DEFAULT``. Pass a [e2e-llm-inference-service] :class:`~urllib3.util.retry.Retry` object for fine-grained control [e2e-llm-inference-service] over different types of retries. [e2e-llm-inference-service] Pass an integer number to retry connection errors that many times, [e2e-llm-inference-service] but no other types of errors. Pass zero to never retry. [e2e-llm-inference-service] [e2e-llm-inference-service] If ``False``, then retries are disabled and any exception is raised [e2e-llm-inference-service] immediately. Also, instead of raising a MaxRetryError on redirects, [e2e-llm-inference-service] the redirect response will be returned. [e2e-llm-inference-service] [e2e-llm-inference-service] :type retries: :class:`~urllib3.util.retry.Retry`, False, or an int. [e2e-llm-inference-service] [e2e-llm-inference-service] :param redirect: [e2e-llm-inference-service] If True, automatically handle redirects (status codes 301, 302, [e2e-llm-inference-service] 303, 307, 308). Each redirect counts as a retry. Disabling retries [e2e-llm-inference-service] will disable redirect, too. [e2e-llm-inference-service] [e2e-llm-inference-service] :param assert_same_host: [e2e-llm-inference-service] If ``True``, will make sure that the host of the pool requests is [e2e-llm-inference-service] consistent else will raise HostChangedError. When ``False``, you can [e2e-llm-inference-service] use the pool on an HTTP proxy and request foreign hosts. [e2e-llm-inference-service] [e2e-llm-inference-service] :param timeout: [e2e-llm-inference-service] If specified, overrides the default timeout for this one [e2e-llm-inference-service] request. It may be a float (in seconds) or an instance of [e2e-llm-inference-service] :class:`urllib3.util.Timeout`. [e2e-llm-inference-service] [e2e-llm-inference-service] :param pool_timeout: [e2e-llm-inference-service] If set and the pool is set to block=True, then this method will [e2e-llm-inference-service] block for ``pool_timeout`` seconds and raise EmptyPoolError if no [e2e-llm-inference-service] connection is available within the time period. [e2e-llm-inference-service] [e2e-llm-inference-service] :param bool preload_content: [e2e-llm-inference-service] If True, the response's body will be preloaded into memory. [e2e-llm-inference-service] [e2e-llm-inference-service] :param bool decode_content: [e2e-llm-inference-service] If True, will attempt to decode the body based on the [e2e-llm-inference-service] 'content-encoding' header. [e2e-llm-inference-service] [e2e-llm-inference-service] :param release_conn: [e2e-llm-inference-service] If False, then the urlopen call will not release the connection [e2e-llm-inference-service] back into the pool once a response is received (but will release if [e2e-llm-inference-service] you read the entire contents of the response such as when [e2e-llm-inference-service] `preload_content=True`). This is useful if you're not preloading [e2e-llm-inference-service] the response's content immediately. You will need to call [e2e-llm-inference-service] ``r.release_conn()`` on the response ``r`` to return the connection [e2e-llm-inference-service] back into the pool. If None, it takes the value of ``preload_content`` [e2e-llm-inference-service] which defaults to ``True``. [e2e-llm-inference-service] [e2e-llm-inference-service] :param bool chunked: [e2e-llm-inference-service] If True, urllib3 will send the body using chunked transfer [e2e-llm-inference-service] encoding. Otherwise, urllib3 will send the body using the standard [e2e-llm-inference-service] content-length form. Defaults to False. [e2e-llm-inference-service] [e2e-llm-inference-service] :param int body_pos: [e2e-llm-inference-service] Position to seek to in file-like body in the event of a retry or [e2e-llm-inference-service] redirect. Typically this won't need to be set because urllib3 will [e2e-llm-inference-service] auto-populate the value when needed. [e2e-llm-inference-service] """ [e2e-llm-inference-service] parsed_url = parse_url(url) [e2e-llm-inference-service] destination_scheme = parsed_url.scheme [e2e-llm-inference-service] [e2e-llm-inference-service] if headers is None: [e2e-llm-inference-service] headers = self.headers [e2e-llm-inference-service] [e2e-llm-inference-service] if not isinstance(retries, Retry): [e2e-llm-inference-service] retries = Retry.from_int(retries, redirect=redirect, default=self.retries) [e2e-llm-inference-service] [e2e-llm-inference-service] if release_conn is None: [e2e-llm-inference-service] release_conn = preload_content [e2e-llm-inference-service] [e2e-llm-inference-service] # Check host [e2e-llm-inference-service] if assert_same_host and not self.is_same_host(url): [e2e-llm-inference-service] raise HostChangedError(self, url, retries) [e2e-llm-inference-service] [e2e-llm-inference-service] # Ensure that the URL we're connecting to is properly encoded [e2e-llm-inference-service] if url.startswith("/"): [e2e-llm-inference-service] url = to_str(_encode_target(url)) [e2e-llm-inference-service] else: [e2e-llm-inference-service] url = to_str(parsed_url.url) [e2e-llm-inference-service] [e2e-llm-inference-service] conn = None [e2e-llm-inference-service] [e2e-llm-inference-service] # Track whether `conn` needs to be released before [e2e-llm-inference-service] # returning/raising/recursing. Update this variable if necessary, and [e2e-llm-inference-service] # leave `release_conn` constant throughout the function. That way, if [e2e-llm-inference-service] # the function recurses, the original value of `release_conn` will be [e2e-llm-inference-service] # passed down into the recursive call, and its value will be respected. [e2e-llm-inference-service] # [e2e-llm-inference-service] # See issue #651 [1] for details. [e2e-llm-inference-service] # [e2e-llm-inference-service] # [1] [e2e-llm-inference-service] release_this_conn = release_conn [e2e-llm-inference-service] [e2e-llm-inference-service] http_tunnel_required = connection_requires_http_tunnel( [e2e-llm-inference-service] self.proxy, self.proxy_config, destination_scheme [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] # Merge the proxy headers. Only done when not using HTTP CONNECT. We [e2e-llm-inference-service] # have to copy the headers dict so we can safely change it without those [e2e-llm-inference-service] # changes being reflected in anyone else's copy. [e2e-llm-inference-service] if not http_tunnel_required: [e2e-llm-inference-service] headers = headers.copy() # type: ignore[attr-defined] [e2e-llm-inference-service] headers.update(self.proxy_headers) # type: ignore[union-attr] [e2e-llm-inference-service] [e2e-llm-inference-service] # Must keep the exception bound to a separate variable or else Python 3 [e2e-llm-inference-service] # complains about UnboundLocalError. [e2e-llm-inference-service] err = None [e2e-llm-inference-service] [e2e-llm-inference-service] # Keep track of whether we cleanly exited the except block. This [e2e-llm-inference-service] # ensures we do proper cleanup in finally. [e2e-llm-inference-service] clean_exit = False [e2e-llm-inference-service] [e2e-llm-inference-service] # Rewind body position, if needed. Record current position [e2e-llm-inference-service] # for future rewinds in the event of a redirect/retry. [e2e-llm-inference-service] body_pos = set_file_position(body, body_pos) [e2e-llm-inference-service] [e2e-llm-inference-service] try: [e2e-llm-inference-service] # Request a connection from the queue. [e2e-llm-inference-service] timeout_obj = self._get_timeout(timeout) [e2e-llm-inference-service] conn = self._get_conn(timeout=pool_timeout) [e2e-llm-inference-service] [e2e-llm-inference-service] conn.timeout = timeout_obj.connect_timeout # type: ignore[assignment] [e2e-llm-inference-service] [e2e-llm-inference-service] # Is this a closed/new connection that requires CONNECT tunnelling? [e2e-llm-inference-service] if self.proxy is not None and http_tunnel_required and conn.is_closed: [e2e-llm-inference-service] try: [e2e-llm-inference-service] self._prepare_proxy(conn) [e2e-llm-inference-service] except (BaseSSLError, OSError, SocketTimeout) as e: [e2e-llm-inference-service] self._raise_timeout( [e2e-llm-inference-service] err=e, url=self.proxy.url, timeout_value=conn.timeout [e2e-llm-inference-service] ) [e2e-llm-inference-service] raise [e2e-llm-inference-service] [e2e-llm-inference-service] # If we're going to release the connection in ``finally:``, then [e2e-llm-inference-service] # the response doesn't need to know about the connection. Otherwise [e2e-llm-inference-service] # it will also try to release it and we'll have a double-release [e2e-llm-inference-service] # mess. [e2e-llm-inference-service] response_conn = conn if not release_conn else None [e2e-llm-inference-service] [e2e-llm-inference-service] # Make the request on the HTTPConnection object [e2e-llm-inference-service] response = self._make_request( [e2e-llm-inference-service] conn, [e2e-llm-inference-service] method, [e2e-llm-inference-service] url, [e2e-llm-inference-service] timeout=timeout_obj, [e2e-llm-inference-service] body=body, [e2e-llm-inference-service] headers=headers, [e2e-llm-inference-service] chunked=chunked, [e2e-llm-inference-service] retries=retries, [e2e-llm-inference-service] response_conn=response_conn, [e2e-llm-inference-service] preload_content=preload_content, [e2e-llm-inference-service] decode_content=decode_content, [e2e-llm-inference-service] **response_kw, [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] # Everything went great! [e2e-llm-inference-service] clean_exit = True [e2e-llm-inference-service] [e2e-llm-inference-service] except EmptyPoolError: [e2e-llm-inference-service] # Didn't get a connection from the pool, no need to clean up [e2e-llm-inference-service] clean_exit = True [e2e-llm-inference-service] release_this_conn = False [e2e-llm-inference-service] raise [e2e-llm-inference-service] [e2e-llm-inference-service] except ( [e2e-llm-inference-service] TimeoutError, [e2e-llm-inference-service] HTTPException, [e2e-llm-inference-service] OSError, [e2e-llm-inference-service] ProtocolError, [e2e-llm-inference-service] BaseSSLError, [e2e-llm-inference-service] SSLError, [e2e-llm-inference-service] CertificateError, [e2e-llm-inference-service] ProxyError, [e2e-llm-inference-service] ) as e: [e2e-llm-inference-service] # Discard the connection for these exceptions. It will be [e2e-llm-inference-service] # replaced during the next _get_conn() call. [e2e-llm-inference-service] clean_exit = False [e2e-llm-inference-service] new_e: Exception = e [e2e-llm-inference-service] if isinstance(e, (BaseSSLError, CertificateError)): [e2e-llm-inference-service] new_e = SSLError(e) [e2e-llm-inference-service] if isinstance( [e2e-llm-inference-service] new_e, [e2e-llm-inference-service] ( [e2e-llm-inference-service] OSError, [e2e-llm-inference-service] NewConnectionError, [e2e-llm-inference-service] TimeoutError, [e2e-llm-inference-service] SSLError, [e2e-llm-inference-service] HTTPException, [e2e-llm-inference-service] ), [e2e-llm-inference-service] ) and (conn and conn.proxy and not conn.has_connected_to_proxy): [e2e-llm-inference-service] new_e = _wrap_proxy_error(new_e, conn.proxy.scheme) [e2e-llm-inference-service] elif isinstance(new_e, (OSError, HTTPException)): [e2e-llm-inference-service] new_e = ProtocolError("Connection aborted.", new_e) [e2e-llm-inference-service] [e2e-llm-inference-service] retries = retries.increment( [e2e-llm-inference-service] method, url, error=new_e, _pool=self, _stacktrace=sys.exc_info()[2] [e2e-llm-inference-service] ) [e2e-llm-inference-service] retries.sleep() [e2e-llm-inference-service] [e2e-llm-inference-service] # Keep track of the error for the retry warning. [e2e-llm-inference-service] err = e [e2e-llm-inference-service] [e2e-llm-inference-service] finally: [e2e-llm-inference-service] if not clean_exit: [e2e-llm-inference-service] # We hit some kind of exception, handled or otherwise. We need [e2e-llm-inference-service] # to throw the connection away unless explicitly told not to. [e2e-llm-inference-service] # Close the connection, set the variable to None, and make sure [e2e-llm-inference-service] # we put the None back in the pool to avoid leaking it. [e2e-llm-inference-service] if conn: [e2e-llm-inference-service] conn.close() [e2e-llm-inference-service] conn = None [e2e-llm-inference-service] release_this_conn = True [e2e-llm-inference-service] [e2e-llm-inference-service] if release_this_conn: [e2e-llm-inference-service] # Put the connection back to be reused. If the connection is [e2e-llm-inference-service] # expired then it will be None, which will get replaced with a [e2e-llm-inference-service] # fresh connection during _get_conn. [e2e-llm-inference-service] self._put_conn(conn) [e2e-llm-inference-service] [e2e-llm-inference-service] if not conn: [e2e-llm-inference-service] # Try again [e2e-llm-inference-service] log.warning( [e2e-llm-inference-service] "Retrying (%r) after connection broken by '%r': %s", retries, err, url [e2e-llm-inference-service] ) [e2e-llm-inference-service] > return self.urlopen( [e2e-llm-inference-service] method, [e2e-llm-inference-service] url, [e2e-llm-inference-service] body, [e2e-llm-inference-service] headers, [e2e-llm-inference-service] retries, [e2e-llm-inference-service] redirect, [e2e-llm-inference-service] assert_same_host, [e2e-llm-inference-service] timeout=timeout, [e2e-llm-inference-service] pool_timeout=pool_timeout, [e2e-llm-inference-service] release_conn=release_conn, [e2e-llm-inference-service] chunked=chunked, [e2e-llm-inference-service] body_pos=body_pos, [e2e-llm-inference-service] preload_content=preload_content, [e2e-llm-inference-service] decode_content=decode_content, [e2e-llm-inference-service] **response_kw, [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] ../../python/kserve/.venv/lib64/python3.11/site-packages/urllib3/connectionpool.py:871: [e2e-llm-inference-service] _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ [e2e-llm-inference-service] [e2e-llm-inference-service] self = [e2e-llm-inference-service] method = 'GET', url = '/v1/models', body = None [e2e-llm-inference-service] headers = {'User-Agent': 'python-requests/2.32.3', 'Accept-Encoding': 'gzip, deflate', 'Accept': '*/*', 'Connection': 'keep-alive', 'Content-Type': 'application/json', 'X-Gateway-Model-Name': 'publishers/kserve-ci-e2e-test/models/facebook/opt-125m'} [e2e-llm-inference-service] retries = Retry(total=1, connect=None, read=None, redirect=None, status=None) [e2e-llm-inference-service] redirect = False, assert_same_host = False [e2e-llm-inference-service] timeout = Timeout(connect=60, read=60, total=None), pool_timeout = None [e2e-llm-inference-service] release_conn = False, chunked = False, body_pos = None, preload_content = False [e2e-llm-inference-service] decode_content = False, response_kw = {} [e2e-llm-inference-service] parsed_url = Url(scheme=None, auth=None, host=None, port=None, path='/v1/models', query=None, fragment=None) [e2e-llm-inference-service] destination_scheme = None, conn = None, release_this_conn = True [e2e-llm-inference-service] http_tunnel_required = False [e2e-llm-inference-service] err = ReadTimeoutError("HTTPConnectionPool(host='ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com', port=80): Read timed out. (read timeout=60)") [e2e-llm-inference-service] clean_exit = False [e2e-llm-inference-service] [e2e-llm-inference-service] def urlopen( # type: ignore[override] [e2e-llm-inference-service] self, [e2e-llm-inference-service] method: str, [e2e-llm-inference-service] url: str, [e2e-llm-inference-service] body: _TYPE_BODY | None = None, [e2e-llm-inference-service] headers: typing.Mapping[str, str] | None = None, [e2e-llm-inference-service] retries: Retry | bool | int | None = None, [e2e-llm-inference-service] redirect: bool = True, [e2e-llm-inference-service] assert_same_host: bool = True, [e2e-llm-inference-service] timeout: _TYPE_TIMEOUT = _DEFAULT_TIMEOUT, [e2e-llm-inference-service] pool_timeout: int | None = None, [e2e-llm-inference-service] release_conn: bool | None = None, [e2e-llm-inference-service] chunked: bool = False, [e2e-llm-inference-service] body_pos: _TYPE_BODY_POSITION | None = None, [e2e-llm-inference-service] preload_content: bool = True, [e2e-llm-inference-service] decode_content: bool = True, [e2e-llm-inference-service] **response_kw: typing.Any, [e2e-llm-inference-service] ) -> BaseHTTPResponse: [e2e-llm-inference-service] """ [e2e-llm-inference-service] Get a connection from the pool and perform an HTTP request. This is the [e2e-llm-inference-service] lowest level call for making a request, so you'll need to specify all [e2e-llm-inference-service] the raw details. [e2e-llm-inference-service] [e2e-llm-inference-service] .. note:: [e2e-llm-inference-service] [e2e-llm-inference-service] More commonly, it's appropriate to use a convenience method [e2e-llm-inference-service] such as :meth:`request`. [e2e-llm-inference-service] [e2e-llm-inference-service] .. note:: [e2e-llm-inference-service] [e2e-llm-inference-service] `release_conn` will only behave as expected if [e2e-llm-inference-service] `preload_content=False` because we want to make [e2e-llm-inference-service] `preload_content=False` the default behaviour someday soon without [e2e-llm-inference-service] breaking backwards compatibility. [e2e-llm-inference-service] [e2e-llm-inference-service] :param method: [e2e-llm-inference-service] HTTP request method (such as GET, POST, PUT, etc.) [e2e-llm-inference-service] [e2e-llm-inference-service] :param url: [e2e-llm-inference-service] The URL to perform the request on. [e2e-llm-inference-service] [e2e-llm-inference-service] :param body: [e2e-llm-inference-service] Data to send in the request body, either :class:`str`, :class:`bytes`, [e2e-llm-inference-service] an iterable of :class:`str`/:class:`bytes`, or a file-like object. [e2e-llm-inference-service] [e2e-llm-inference-service] :param headers: [e2e-llm-inference-service] Dictionary of custom headers to send, such as User-Agent, [e2e-llm-inference-service] If-None-Match, etc. If None, pool headers are used. If provided, [e2e-llm-inference-service] these headers completely replace any pool-specific headers. [e2e-llm-inference-service] [e2e-llm-inference-service] :param retries: [e2e-llm-inference-service] Configure the number of retries to allow before raising a [e2e-llm-inference-service] :class:`~urllib3.exceptions.MaxRetryError` exception. [e2e-llm-inference-service] [e2e-llm-inference-service] If ``None`` (default) will retry 3 times, see ``Retry.DEFAULT``. Pass a [e2e-llm-inference-service] :class:`~urllib3.util.retry.Retry` object for fine-grained control [e2e-llm-inference-service] over different types of retries. [e2e-llm-inference-service] Pass an integer number to retry connection errors that many times, [e2e-llm-inference-service] but no other types of errors. Pass zero to never retry. [e2e-llm-inference-service] [e2e-llm-inference-service] If ``False``, then retries are disabled and any exception is raised [e2e-llm-inference-service] immediately. Also, instead of raising a MaxRetryError on redirects, [e2e-llm-inference-service] the redirect response will be returned. [e2e-llm-inference-service] [e2e-llm-inference-service] :type retries: :class:`~urllib3.util.retry.Retry`, False, or an int. [e2e-llm-inference-service] [e2e-llm-inference-service] :param redirect: [e2e-llm-inference-service] If True, automatically handle redirects (status codes 301, 302, [e2e-llm-inference-service] 303, 307, 308). Each redirect counts as a retry. Disabling retries [e2e-llm-inference-service] will disable redirect, too. [e2e-llm-inference-service] [e2e-llm-inference-service] :param assert_same_host: [e2e-llm-inference-service] If ``True``, will make sure that the host of the pool requests is [e2e-llm-inference-service] consistent else will raise HostChangedError. When ``False``, you can [e2e-llm-inference-service] use the pool on an HTTP proxy and request foreign hosts. [e2e-llm-inference-service] [e2e-llm-inference-service] :param timeout: [e2e-llm-inference-service] If specified, overrides the default timeout for this one [e2e-llm-inference-service] request. It may be a float (in seconds) or an instance of [e2e-llm-inference-service] :class:`urllib3.util.Timeout`. [e2e-llm-inference-service] [e2e-llm-inference-service] :param pool_timeout: [e2e-llm-inference-service] If set and the pool is set to block=True, then this method will [e2e-llm-inference-service] block for ``pool_timeout`` seconds and raise EmptyPoolError if no [e2e-llm-inference-service] connection is available within the time period. [e2e-llm-inference-service] [e2e-llm-inference-service] :param bool preload_content: [e2e-llm-inference-service] If True, the response's body will be preloaded into memory. [e2e-llm-inference-service] [e2e-llm-inference-service] :param bool decode_content: [e2e-llm-inference-service] If True, will attempt to decode the body based on the [e2e-llm-inference-service] 'content-encoding' header. [e2e-llm-inference-service] [e2e-llm-inference-service] :param release_conn: [e2e-llm-inference-service] If False, then the urlopen call will not release the connection [e2e-llm-inference-service] back into the pool once a response is received (but will release if [e2e-llm-inference-service] you read the entire contents of the response such as when [e2e-llm-inference-service] `preload_content=True`). This is useful if you're not preloading [e2e-llm-inference-service] the response's content immediately. You will need to call [e2e-llm-inference-service] ``r.release_conn()`` on the response ``r`` to return the connection [e2e-llm-inference-service] back into the pool. If None, it takes the value of ``preload_content`` [e2e-llm-inference-service] which defaults to ``True``. [e2e-llm-inference-service] [e2e-llm-inference-service] :param bool chunked: [e2e-llm-inference-service] If True, urllib3 will send the body using chunked transfer [e2e-llm-inference-service] encoding. Otherwise, urllib3 will send the body using the standard [e2e-llm-inference-service] content-length form. Defaults to False. [e2e-llm-inference-service] [e2e-llm-inference-service] :param int body_pos: [e2e-llm-inference-service] Position to seek to in file-like body in the event of a retry or [e2e-llm-inference-service] redirect. Typically this won't need to be set because urllib3 will [e2e-llm-inference-service] auto-populate the value when needed. [e2e-llm-inference-service] """ [e2e-llm-inference-service] parsed_url = parse_url(url) [e2e-llm-inference-service] destination_scheme = parsed_url.scheme [e2e-llm-inference-service] [e2e-llm-inference-service] if headers is None: [e2e-llm-inference-service] headers = self.headers [e2e-llm-inference-service] [e2e-llm-inference-service] if not isinstance(retries, Retry): [e2e-llm-inference-service] retries = Retry.from_int(retries, redirect=redirect, default=self.retries) [e2e-llm-inference-service] [e2e-llm-inference-service] if release_conn is None: [e2e-llm-inference-service] release_conn = preload_content [e2e-llm-inference-service] [e2e-llm-inference-service] # Check host [e2e-llm-inference-service] if assert_same_host and not self.is_same_host(url): [e2e-llm-inference-service] raise HostChangedError(self, url, retries) [e2e-llm-inference-service] [e2e-llm-inference-service] # Ensure that the URL we're connecting to is properly encoded [e2e-llm-inference-service] if url.startswith("/"): [e2e-llm-inference-service] url = to_str(_encode_target(url)) [e2e-llm-inference-service] else: [e2e-llm-inference-service] url = to_str(parsed_url.url) [e2e-llm-inference-service] [e2e-llm-inference-service] conn = None [e2e-llm-inference-service] [e2e-llm-inference-service] # Track whether `conn` needs to be released before [e2e-llm-inference-service] # returning/raising/recursing. Update this variable if necessary, and [e2e-llm-inference-service] # leave `release_conn` constant throughout the function. That way, if [e2e-llm-inference-service] # the function recurses, the original value of `release_conn` will be [e2e-llm-inference-service] # passed down into the recursive call, and its value will be respected. [e2e-llm-inference-service] # [e2e-llm-inference-service] # See issue #651 [1] for details. [e2e-llm-inference-service] # [e2e-llm-inference-service] # [1] [e2e-llm-inference-service] release_this_conn = release_conn [e2e-llm-inference-service] [e2e-llm-inference-service] http_tunnel_required = connection_requires_http_tunnel( [e2e-llm-inference-service] self.proxy, self.proxy_config, destination_scheme [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] # Merge the proxy headers. Only done when not using HTTP CONNECT. We [e2e-llm-inference-service] # have to copy the headers dict so we can safely change it without those [e2e-llm-inference-service] # changes being reflected in anyone else's copy. [e2e-llm-inference-service] if not http_tunnel_required: [e2e-llm-inference-service] headers = headers.copy() # type: ignore[attr-defined] [e2e-llm-inference-service] headers.update(self.proxy_headers) # type: ignore[union-attr] [e2e-llm-inference-service] [e2e-llm-inference-service] # Must keep the exception bound to a separate variable or else Python 3 [e2e-llm-inference-service] # complains about UnboundLocalError. [e2e-llm-inference-service] err = None [e2e-llm-inference-service] [e2e-llm-inference-service] # Keep track of whether we cleanly exited the except block. This [e2e-llm-inference-service] # ensures we do proper cleanup in finally. [e2e-llm-inference-service] clean_exit = False [e2e-llm-inference-service] [e2e-llm-inference-service] # Rewind body position, if needed. Record current position [e2e-llm-inference-service] # for future rewinds in the event of a redirect/retry. [e2e-llm-inference-service] body_pos = set_file_position(body, body_pos) [e2e-llm-inference-service] [e2e-llm-inference-service] try: [e2e-llm-inference-service] # Request a connection from the queue. [e2e-llm-inference-service] timeout_obj = self._get_timeout(timeout) [e2e-llm-inference-service] conn = self._get_conn(timeout=pool_timeout) [e2e-llm-inference-service] [e2e-llm-inference-service] conn.timeout = timeout_obj.connect_timeout # type: ignore[assignment] [e2e-llm-inference-service] [e2e-llm-inference-service] # Is this a closed/new connection that requires CONNECT tunnelling? [e2e-llm-inference-service] if self.proxy is not None and http_tunnel_required and conn.is_closed: [e2e-llm-inference-service] try: [e2e-llm-inference-service] self._prepare_proxy(conn) [e2e-llm-inference-service] except (BaseSSLError, OSError, SocketTimeout) as e: [e2e-llm-inference-service] self._raise_timeout( [e2e-llm-inference-service] err=e, url=self.proxy.url, timeout_value=conn.timeout [e2e-llm-inference-service] ) [e2e-llm-inference-service] raise [e2e-llm-inference-service] [e2e-llm-inference-service] # If we're going to release the connection in ``finally:``, then [e2e-llm-inference-service] # the response doesn't need to know about the connection. Otherwise [e2e-llm-inference-service] # it will also try to release it and we'll have a double-release [e2e-llm-inference-service] # mess. [e2e-llm-inference-service] response_conn = conn if not release_conn else None [e2e-llm-inference-service] [e2e-llm-inference-service] # Make the request on the HTTPConnection object [e2e-llm-inference-service] response = self._make_request( [e2e-llm-inference-service] conn, [e2e-llm-inference-service] method, [e2e-llm-inference-service] url, [e2e-llm-inference-service] timeout=timeout_obj, [e2e-llm-inference-service] body=body, [e2e-llm-inference-service] headers=headers, [e2e-llm-inference-service] chunked=chunked, [e2e-llm-inference-service] retries=retries, [e2e-llm-inference-service] response_conn=response_conn, [e2e-llm-inference-service] preload_content=preload_content, [e2e-llm-inference-service] decode_content=decode_content, [e2e-llm-inference-service] **response_kw, [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] # Everything went great! [e2e-llm-inference-service] clean_exit = True [e2e-llm-inference-service] [e2e-llm-inference-service] except EmptyPoolError: [e2e-llm-inference-service] # Didn't get a connection from the pool, no need to clean up [e2e-llm-inference-service] clean_exit = True [e2e-llm-inference-service] release_this_conn = False [e2e-llm-inference-service] raise [e2e-llm-inference-service] [e2e-llm-inference-service] except ( [e2e-llm-inference-service] TimeoutError, [e2e-llm-inference-service] HTTPException, [e2e-llm-inference-service] OSError, [e2e-llm-inference-service] ProtocolError, [e2e-llm-inference-service] BaseSSLError, [e2e-llm-inference-service] SSLError, [e2e-llm-inference-service] CertificateError, [e2e-llm-inference-service] ProxyError, [e2e-llm-inference-service] ) as e: [e2e-llm-inference-service] # Discard the connection for these exceptions. It will be [e2e-llm-inference-service] # replaced during the next _get_conn() call. [e2e-llm-inference-service] clean_exit = False [e2e-llm-inference-service] new_e: Exception = e [e2e-llm-inference-service] if isinstance(e, (BaseSSLError, CertificateError)): [e2e-llm-inference-service] new_e = SSLError(e) [e2e-llm-inference-service] if isinstance( [e2e-llm-inference-service] new_e, [e2e-llm-inference-service] ( [e2e-llm-inference-service] OSError, [e2e-llm-inference-service] NewConnectionError, [e2e-llm-inference-service] TimeoutError, [e2e-llm-inference-service] SSLError, [e2e-llm-inference-service] HTTPException, [e2e-llm-inference-service] ), [e2e-llm-inference-service] ) and (conn and conn.proxy and not conn.has_connected_to_proxy): [e2e-llm-inference-service] new_e = _wrap_proxy_error(new_e, conn.proxy.scheme) [e2e-llm-inference-service] elif isinstance(new_e, (OSError, HTTPException)): [e2e-llm-inference-service] new_e = ProtocolError("Connection aborted.", new_e) [e2e-llm-inference-service] [e2e-llm-inference-service] retries = retries.increment( [e2e-llm-inference-service] method, url, error=new_e, _pool=self, _stacktrace=sys.exc_info()[2] [e2e-llm-inference-service] ) [e2e-llm-inference-service] retries.sleep() [e2e-llm-inference-service] [e2e-llm-inference-service] # Keep track of the error for the retry warning. [e2e-llm-inference-service] err = e [e2e-llm-inference-service] [e2e-llm-inference-service] finally: [e2e-llm-inference-service] if not clean_exit: [e2e-llm-inference-service] # We hit some kind of exception, handled or otherwise. We need [e2e-llm-inference-service] # to throw the connection away unless explicitly told not to. [e2e-llm-inference-service] # Close the connection, set the variable to None, and make sure [e2e-llm-inference-service] # we put the None back in the pool to avoid leaking it. [e2e-llm-inference-service] if conn: [e2e-llm-inference-service] conn.close() [e2e-llm-inference-service] conn = None [e2e-llm-inference-service] release_this_conn = True [e2e-llm-inference-service] [e2e-llm-inference-service] if release_this_conn: [e2e-llm-inference-service] # Put the connection back to be reused. If the connection is [e2e-llm-inference-service] # expired then it will be None, which will get replaced with a [e2e-llm-inference-service] # fresh connection during _get_conn. [e2e-llm-inference-service] self._put_conn(conn) [e2e-llm-inference-service] [e2e-llm-inference-service] if not conn: [e2e-llm-inference-service] # Try again [e2e-llm-inference-service] log.warning( [e2e-llm-inference-service] "Retrying (%r) after connection broken by '%r': %s", retries, err, url [e2e-llm-inference-service] ) [e2e-llm-inference-service] > return self.urlopen( [e2e-llm-inference-service] method, [e2e-llm-inference-service] url, [e2e-llm-inference-service] body, [e2e-llm-inference-service] headers, [e2e-llm-inference-service] retries, [e2e-llm-inference-service] redirect, [e2e-llm-inference-service] assert_same_host, [e2e-llm-inference-service] timeout=timeout, [e2e-llm-inference-service] pool_timeout=pool_timeout, [e2e-llm-inference-service] release_conn=release_conn, [e2e-llm-inference-service] chunked=chunked, [e2e-llm-inference-service] body_pos=body_pos, [e2e-llm-inference-service] preload_content=preload_content, [e2e-llm-inference-service] decode_content=decode_content, [e2e-llm-inference-service] **response_kw, [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] ../../python/kserve/.venv/lib64/python3.11/site-packages/urllib3/connectionpool.py:871: [e2e-llm-inference-service] _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ [e2e-llm-inference-service] [e2e-llm-inference-service] self = [e2e-llm-inference-service] method = 'GET', url = '/v1/models', body = None [e2e-llm-inference-service] headers = {'User-Agent': 'python-requests/2.32.3', 'Accept-Encoding': 'gzip, deflate', 'Accept': '*/*', 'Connection': 'keep-alive', 'Content-Type': 'application/json', 'X-Gateway-Model-Name': 'publishers/kserve-ci-e2e-test/models/facebook/opt-125m'} [e2e-llm-inference-service] retries = Retry(total=0, connect=None, read=None, redirect=None, status=None) [e2e-llm-inference-service] redirect = False, assert_same_host = False [e2e-llm-inference-service] timeout = Timeout(connect=60, read=60, total=None), pool_timeout = None [e2e-llm-inference-service] release_conn = False, chunked = False, body_pos = None, preload_content = False [e2e-llm-inference-service] decode_content = False, response_kw = {} [e2e-llm-inference-service] parsed_url = Url(scheme=None, auth=None, host=None, port=None, path='/v1/models', query=None, fragment=None) [e2e-llm-inference-service] destination_scheme = None, conn = None, release_this_conn = True [e2e-llm-inference-service] http_tunnel_required = False [e2e-llm-inference-service] err = ReadTimeoutError("HTTPConnectionPool(host='ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com', port=80): Read timed out. (read timeout=60)") [e2e-llm-inference-service] clean_exit = False [e2e-llm-inference-service] [e2e-llm-inference-service] def urlopen( # type: ignore[override] [e2e-llm-inference-service] self, [e2e-llm-inference-service] method: str, [e2e-llm-inference-service] url: str, [e2e-llm-inference-service] body: _TYPE_BODY | None = None, [e2e-llm-inference-service] headers: typing.Mapping[str, str] | None = None, [e2e-llm-inference-service] retries: Retry | bool | int | None = None, [e2e-llm-inference-service] redirect: bool = True, [e2e-llm-inference-service] assert_same_host: bool = True, [e2e-llm-inference-service] timeout: _TYPE_TIMEOUT = _DEFAULT_TIMEOUT, [e2e-llm-inference-service] pool_timeout: int | None = None, [e2e-llm-inference-service] release_conn: bool | None = None, [e2e-llm-inference-service] chunked: bool = False, [e2e-llm-inference-service] body_pos: _TYPE_BODY_POSITION | None = None, [e2e-llm-inference-service] preload_content: bool = True, [e2e-llm-inference-service] decode_content: bool = True, [e2e-llm-inference-service] **response_kw: typing.Any, [e2e-llm-inference-service] ) -> BaseHTTPResponse: [e2e-llm-inference-service] """ [e2e-llm-inference-service] Get a connection from the pool and perform an HTTP request. This is the [e2e-llm-inference-service] lowest level call for making a request, so you'll need to specify all [e2e-llm-inference-service] the raw details. [e2e-llm-inference-service] [e2e-llm-inference-service] .. note:: [e2e-llm-inference-service] [e2e-llm-inference-service] More commonly, it's appropriate to use a convenience method [e2e-llm-inference-service] such as :meth:`request`. [e2e-llm-inference-service] [e2e-llm-inference-service] .. note:: [e2e-llm-inference-service] [e2e-llm-inference-service] `release_conn` will only behave as expected if [e2e-llm-inference-service] `preload_content=False` because we want to make [e2e-llm-inference-service] `preload_content=False` the default behaviour someday soon without [e2e-llm-inference-service] breaking backwards compatibility. [e2e-llm-inference-service] [e2e-llm-inference-service] :param method: [e2e-llm-inference-service] HTTP request method (such as GET, POST, PUT, etc.) [e2e-llm-inference-service] [e2e-llm-inference-service] :param url: [e2e-llm-inference-service] The URL to perform the request on. [e2e-llm-inference-service] [e2e-llm-inference-service] :param body: [e2e-llm-inference-service] Data to send in the request body, either :class:`str`, :class:`bytes`, [e2e-llm-inference-service] an iterable of :class:`str`/:class:`bytes`, or a file-like object. [e2e-llm-inference-service] [e2e-llm-inference-service] :param headers: [e2e-llm-inference-service] Dictionary of custom headers to send, such as User-Agent, [e2e-llm-inference-service] If-None-Match, etc. If None, pool headers are used. If provided, [e2e-llm-inference-service] these headers completely replace any pool-specific headers. [e2e-llm-inference-service] [e2e-llm-inference-service] :param retries: [e2e-llm-inference-service] Configure the number of retries to allow before raising a [e2e-llm-inference-service] :class:`~urllib3.exceptions.MaxRetryError` exception. [e2e-llm-inference-service] [e2e-llm-inference-service] If ``None`` (default) will retry 3 times, see ``Retry.DEFAULT``. Pass a [e2e-llm-inference-service] :class:`~urllib3.util.retry.Retry` object for fine-grained control [e2e-llm-inference-service] over different types of retries. [e2e-llm-inference-service] Pass an integer number to retry connection errors that many times, [e2e-llm-inference-service] but no other types of errors. Pass zero to never retry. [e2e-llm-inference-service] [e2e-llm-inference-service] If ``False``, then retries are disabled and any exception is raised [e2e-llm-inference-service] immediately. Also, instead of raising a MaxRetryError on redirects, [e2e-llm-inference-service] the redirect response will be returned. [e2e-llm-inference-service] [e2e-llm-inference-service] :type retries: :class:`~urllib3.util.retry.Retry`, False, or an int. [e2e-llm-inference-service] [e2e-llm-inference-service] :param redirect: [e2e-llm-inference-service] If True, automatically handle redirects (status codes 301, 302, [e2e-llm-inference-service] 303, 307, 308). Each redirect counts as a retry. Disabling retries [e2e-llm-inference-service] will disable redirect, too. [e2e-llm-inference-service] [e2e-llm-inference-service] :param assert_same_host: [e2e-llm-inference-service] If ``True``, will make sure that the host of the pool requests is [e2e-llm-inference-service] consistent else will raise HostChangedError. When ``False``, you can [e2e-llm-inference-service] use the pool on an HTTP proxy and request foreign hosts. [e2e-llm-inference-service] [e2e-llm-inference-service] :param timeout: [e2e-llm-inference-service] If specified, overrides the default timeout for this one [e2e-llm-inference-service] request. It may be a float (in seconds) or an instance of [e2e-llm-inference-service] :class:`urllib3.util.Timeout`. [e2e-llm-inference-service] [e2e-llm-inference-service] :param pool_timeout: [e2e-llm-inference-service] If set and the pool is set to block=True, then this method will [e2e-llm-inference-service] block for ``pool_timeout`` seconds and raise EmptyPoolError if no [e2e-llm-inference-service] connection is available within the time period. [e2e-llm-inference-service] [e2e-llm-inference-service] :param bool preload_content: [e2e-llm-inference-service] If True, the response's body will be preloaded into memory. [e2e-llm-inference-service] [e2e-llm-inference-service] :param bool decode_content: [e2e-llm-inference-service] If True, will attempt to decode the body based on the [e2e-llm-inference-service] 'content-encoding' header. [e2e-llm-inference-service] [e2e-llm-inference-service] :param release_conn: [e2e-llm-inference-service] If False, then the urlopen call will not release the connection [e2e-llm-inference-service] back into the pool once a response is received (but will release if [e2e-llm-inference-service] you read the entire contents of the response such as when [e2e-llm-inference-service] `preload_content=True`). This is useful if you're not preloading [e2e-llm-inference-service] the response's content immediately. You will need to call [e2e-llm-inference-service] ``r.release_conn()`` on the response ``r`` to return the connection [e2e-llm-inference-service] back into the pool. If None, it takes the value of ``preload_content`` [e2e-llm-inference-service] which defaults to ``True``. [e2e-llm-inference-service] [e2e-llm-inference-service] :param bool chunked: [e2e-llm-inference-service] If True, urllib3 will send the body using chunked transfer [e2e-llm-inference-service] encoding. Otherwise, urllib3 will send the body using the standard [e2e-llm-inference-service] content-length form. Defaults to False. [e2e-llm-inference-service] [e2e-llm-inference-service] :param int body_pos: [e2e-llm-inference-service] Position to seek to in file-like body in the event of a retry or [e2e-llm-inference-service] redirect. Typically this won't need to be set because urllib3 will [e2e-llm-inference-service] auto-populate the value when needed. [e2e-llm-inference-service] """ [e2e-llm-inference-service] parsed_url = parse_url(url) [e2e-llm-inference-service] destination_scheme = parsed_url.scheme [e2e-llm-inference-service] [e2e-llm-inference-service] if headers is None: [e2e-llm-inference-service] headers = self.headers [e2e-llm-inference-service] [e2e-llm-inference-service] if not isinstance(retries, Retry): [e2e-llm-inference-service] retries = Retry.from_int(retries, redirect=redirect, default=self.retries) [e2e-llm-inference-service] [e2e-llm-inference-service] if release_conn is None: [e2e-llm-inference-service] release_conn = preload_content [e2e-llm-inference-service] [e2e-llm-inference-service] # Check host [e2e-llm-inference-service] if assert_same_host and not self.is_same_host(url): [e2e-llm-inference-service] raise HostChangedError(self, url, retries) [e2e-llm-inference-service] [e2e-llm-inference-service] # Ensure that the URL we're connecting to is properly encoded [e2e-llm-inference-service] if url.startswith("/"): [e2e-llm-inference-service] url = to_str(_encode_target(url)) [e2e-llm-inference-service] else: [e2e-llm-inference-service] url = to_str(parsed_url.url) [e2e-llm-inference-service] [e2e-llm-inference-service] conn = None [e2e-llm-inference-service] [e2e-llm-inference-service] # Track whether `conn` needs to be released before [e2e-llm-inference-service] # returning/raising/recursing. Update this variable if necessary, and [e2e-llm-inference-service] # leave `release_conn` constant throughout the function. That way, if [e2e-llm-inference-service] # the function recurses, the original value of `release_conn` will be [e2e-llm-inference-service] # passed down into the recursive call, and its value will be respected. [e2e-llm-inference-service] # [e2e-llm-inference-service] # See issue #651 [1] for details. [e2e-llm-inference-service] # [e2e-llm-inference-service] # [1] [e2e-llm-inference-service] release_this_conn = release_conn [e2e-llm-inference-service] [e2e-llm-inference-service] http_tunnel_required = connection_requires_http_tunnel( [e2e-llm-inference-service] self.proxy, self.proxy_config, destination_scheme [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] # Merge the proxy headers. Only done when not using HTTP CONNECT. We [e2e-llm-inference-service] # have to copy the headers dict so we can safely change it without those [e2e-llm-inference-service] # changes being reflected in anyone else's copy. [e2e-llm-inference-service] if not http_tunnel_required: [e2e-llm-inference-service] headers = headers.copy() # type: ignore[attr-defined] [e2e-llm-inference-service] headers.update(self.proxy_headers) # type: ignore[union-attr] [e2e-llm-inference-service] [e2e-llm-inference-service] # Must keep the exception bound to a separate variable or else Python 3 [e2e-llm-inference-service] # complains about UnboundLocalError. [e2e-llm-inference-service] err = None [e2e-llm-inference-service] [e2e-llm-inference-service] # Keep track of whether we cleanly exited the except block. This [e2e-llm-inference-service] # ensures we do proper cleanup in finally. [e2e-llm-inference-service] clean_exit = False [e2e-llm-inference-service] [e2e-llm-inference-service] # Rewind body position, if needed. Record current position [e2e-llm-inference-service] # for future rewinds in the event of a redirect/retry. [e2e-llm-inference-service] body_pos = set_file_position(body, body_pos) [e2e-llm-inference-service] [e2e-llm-inference-service] try: [e2e-llm-inference-service] # Request a connection from the queue. [e2e-llm-inference-service] timeout_obj = self._get_timeout(timeout) [e2e-llm-inference-service] conn = self._get_conn(timeout=pool_timeout) [e2e-llm-inference-service] [e2e-llm-inference-service] conn.timeout = timeout_obj.connect_timeout # type: ignore[assignment] [e2e-llm-inference-service] [e2e-llm-inference-service] # Is this a closed/new connection that requires CONNECT tunnelling? [e2e-llm-inference-service] if self.proxy is not None and http_tunnel_required and conn.is_closed: [e2e-llm-inference-service] try: [e2e-llm-inference-service] self._prepare_proxy(conn) [e2e-llm-inference-service] except (BaseSSLError, OSError, SocketTimeout) as e: [e2e-llm-inference-service] self._raise_timeout( [e2e-llm-inference-service] err=e, url=self.proxy.url, timeout_value=conn.timeout [e2e-llm-inference-service] ) [e2e-llm-inference-service] raise [e2e-llm-inference-service] [e2e-llm-inference-service] # If we're going to release the connection in ``finally:``, then [e2e-llm-inference-service] # the response doesn't need to know about the connection. Otherwise [e2e-llm-inference-service] # it will also try to release it and we'll have a double-release [e2e-llm-inference-service] # mess. [e2e-llm-inference-service] response_conn = conn if not release_conn else None [e2e-llm-inference-service] [e2e-llm-inference-service] # Make the request on the HTTPConnection object [e2e-llm-inference-service] response = self._make_request( [e2e-llm-inference-service] conn, [e2e-llm-inference-service] method, [e2e-llm-inference-service] url, [e2e-llm-inference-service] timeout=timeout_obj, [e2e-llm-inference-service] body=body, [e2e-llm-inference-service] headers=headers, [e2e-llm-inference-service] chunked=chunked, [e2e-llm-inference-service] retries=retries, [e2e-llm-inference-service] response_conn=response_conn, [e2e-llm-inference-service] preload_content=preload_content, [e2e-llm-inference-service] decode_content=decode_content, [e2e-llm-inference-service] **response_kw, [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] # Everything went great! [e2e-llm-inference-service] clean_exit = True [e2e-llm-inference-service] [e2e-llm-inference-service] except EmptyPoolError: [e2e-llm-inference-service] # Didn't get a connection from the pool, no need to clean up [e2e-llm-inference-service] clean_exit = True [e2e-llm-inference-service] release_this_conn = False [e2e-llm-inference-service] raise [e2e-llm-inference-service] [e2e-llm-inference-service] except ( [e2e-llm-inference-service] TimeoutError, [e2e-llm-inference-service] HTTPException, [e2e-llm-inference-service] OSError, [e2e-llm-inference-service] ProtocolError, [e2e-llm-inference-service] BaseSSLError, [e2e-llm-inference-service] SSLError, [e2e-llm-inference-service] CertificateError, [e2e-llm-inference-service] ProxyError, [e2e-llm-inference-service] ) as e: [e2e-llm-inference-service] # Discard the connection for these exceptions. It will be [e2e-llm-inference-service] # replaced during the next _get_conn() call. [e2e-llm-inference-service] clean_exit = False [e2e-llm-inference-service] new_e: Exception = e [e2e-llm-inference-service] if isinstance(e, (BaseSSLError, CertificateError)): [e2e-llm-inference-service] new_e = SSLError(e) [e2e-llm-inference-service] if isinstance( [e2e-llm-inference-service] new_e, [e2e-llm-inference-service] ( [e2e-llm-inference-service] OSError, [e2e-llm-inference-service] NewConnectionError, [e2e-llm-inference-service] TimeoutError, [e2e-llm-inference-service] SSLError, [e2e-llm-inference-service] HTTPException, [e2e-llm-inference-service] ), [e2e-llm-inference-service] ) and (conn and conn.proxy and not conn.has_connected_to_proxy): [e2e-llm-inference-service] new_e = _wrap_proxy_error(new_e, conn.proxy.scheme) [e2e-llm-inference-service] elif isinstance(new_e, (OSError, HTTPException)): [e2e-llm-inference-service] new_e = ProtocolError("Connection aborted.", new_e) [e2e-llm-inference-service] [e2e-llm-inference-service] retries = retries.increment( [e2e-llm-inference-service] method, url, error=new_e, _pool=self, _stacktrace=sys.exc_info()[2] [e2e-llm-inference-service] ) [e2e-llm-inference-service] retries.sleep() [e2e-llm-inference-service] [e2e-llm-inference-service] # Keep track of the error for the retry warning. [e2e-llm-inference-service] err = e [e2e-llm-inference-service] [e2e-llm-inference-service] finally: [e2e-llm-inference-service] if not clean_exit: [e2e-llm-inference-service] # We hit some kind of exception, handled or otherwise. We need [e2e-llm-inference-service] # to throw the connection away unless explicitly told not to. [e2e-llm-inference-service] # Close the connection, set the variable to None, and make sure [e2e-llm-inference-service] # we put the None back in the pool to avoid leaking it. [e2e-llm-inference-service] if conn: [e2e-llm-inference-service] conn.close() [e2e-llm-inference-service] conn = None [e2e-llm-inference-service] release_this_conn = True [e2e-llm-inference-service] [e2e-llm-inference-service] if release_this_conn: [e2e-llm-inference-service] # Put the connection back to be reused. If the connection is [e2e-llm-inference-service] # expired then it will be None, which will get replaced with a [e2e-llm-inference-service] # fresh connection during _get_conn. [e2e-llm-inference-service] self._put_conn(conn) [e2e-llm-inference-service] [e2e-llm-inference-service] if not conn: [e2e-llm-inference-service] # Try again [e2e-llm-inference-service] log.warning( [e2e-llm-inference-service] "Retrying (%r) after connection broken by '%r': %s", retries, err, url [e2e-llm-inference-service] ) [e2e-llm-inference-service] > return self.urlopen( [e2e-llm-inference-service] method, [e2e-llm-inference-service] url, [e2e-llm-inference-service] body, [e2e-llm-inference-service] headers, [e2e-llm-inference-service] retries, [e2e-llm-inference-service] redirect, [e2e-llm-inference-service] assert_same_host, [e2e-llm-inference-service] timeout=timeout, [e2e-llm-inference-service] pool_timeout=pool_timeout, [e2e-llm-inference-service] release_conn=release_conn, [e2e-llm-inference-service] chunked=chunked, [e2e-llm-inference-service] body_pos=body_pos, [e2e-llm-inference-service] preload_content=preload_content, [e2e-llm-inference-service] decode_content=decode_content, [e2e-llm-inference-service] **response_kw, [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] ../../python/kserve/.venv/lib64/python3.11/site-packages/urllib3/connectionpool.py:871: [e2e-llm-inference-service] _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ [e2e-llm-inference-service] [e2e-llm-inference-service] self = [e2e-llm-inference-service] method = 'GET', url = '/v1/models', body = None [e2e-llm-inference-service] headers = {'User-Agent': 'python-requests/2.32.3', 'Accept-Encoding': 'gzip, deflate', 'Accept': '*/*', 'Connection': 'keep-alive', 'Content-Type': 'application/json', 'X-Gateway-Model-Name': 'publishers/kserve-ci-e2e-test/models/facebook/opt-125m'} [e2e-llm-inference-service] retries = Retry(total=0, connect=None, read=None, redirect=None, status=None) [e2e-llm-inference-service] redirect = False, assert_same_host = False [e2e-llm-inference-service] timeout = Timeout(connect=60, read=60, total=None), pool_timeout = None [e2e-llm-inference-service] release_conn = False, chunked = False, body_pos = None, preload_content = False [e2e-llm-inference-service] decode_content = False, response_kw = {} [e2e-llm-inference-service] parsed_url = Url(scheme=None, auth=None, host=None, port=None, path='/v1/models', query=None, fragment=None) [e2e-llm-inference-service] destination_scheme = None, conn = None, release_this_conn = True [e2e-llm-inference-service] http_tunnel_required = False, err = None, clean_exit = False [e2e-llm-inference-service] [e2e-llm-inference-service] def urlopen( # type: ignore[override] [e2e-llm-inference-service] self, [e2e-llm-inference-service] method: str, [e2e-llm-inference-service] url: str, [e2e-llm-inference-service] body: _TYPE_BODY | None = None, [e2e-llm-inference-service] headers: typing.Mapping[str, str] | None = None, [e2e-llm-inference-service] retries: Retry | bool | int | None = None, [e2e-llm-inference-service] redirect: bool = True, [e2e-llm-inference-service] assert_same_host: bool = True, [e2e-llm-inference-service] timeout: _TYPE_TIMEOUT = _DEFAULT_TIMEOUT, [e2e-llm-inference-service] pool_timeout: int | None = None, [e2e-llm-inference-service] release_conn: bool | None = None, [e2e-llm-inference-service] chunked: bool = False, [e2e-llm-inference-service] body_pos: _TYPE_BODY_POSITION | None = None, [e2e-llm-inference-service] preload_content: bool = True, [e2e-llm-inference-service] decode_content: bool = True, [e2e-llm-inference-service] **response_kw: typing.Any, [e2e-llm-inference-service] ) -> BaseHTTPResponse: [e2e-llm-inference-service] """ [e2e-llm-inference-service] Get a connection from the pool and perform an HTTP request. This is the [e2e-llm-inference-service] lowest level call for making a request, so you'll need to specify all [e2e-llm-inference-service] the raw details. [e2e-llm-inference-service] [e2e-llm-inference-service] .. note:: [e2e-llm-inference-service] [e2e-llm-inference-service] More commonly, it's appropriate to use a convenience method [e2e-llm-inference-service] such as :meth:`request`. [e2e-llm-inference-service] [e2e-llm-inference-service] .. note:: [e2e-llm-inference-service] [e2e-llm-inference-service] `release_conn` will only behave as expected if [e2e-llm-inference-service] `preload_content=False` because we want to make [e2e-llm-inference-service] `preload_content=False` the default behaviour someday soon without [e2e-llm-inference-service] breaking backwards compatibility. [e2e-llm-inference-service] [e2e-llm-inference-service] :param method: [e2e-llm-inference-service] HTTP request method (such as GET, POST, PUT, etc.) [e2e-llm-inference-service] [e2e-llm-inference-service] :param url: [e2e-llm-inference-service] The URL to perform the request on. [e2e-llm-inference-service] [e2e-llm-inference-service] :param body: [e2e-llm-inference-service] Data to send in the request body, either :class:`str`, :class:`bytes`, [e2e-llm-inference-service] an iterable of :class:`str`/:class:`bytes`, or a file-like object. [e2e-llm-inference-service] [e2e-llm-inference-service] :param headers: [e2e-llm-inference-service] Dictionary of custom headers to send, such as User-Agent, [e2e-llm-inference-service] If-None-Match, etc. If None, pool headers are used. If provided, [e2e-llm-inference-service] these headers completely replace any pool-specific headers. [e2e-llm-inference-service] [e2e-llm-inference-service] :param retries: [e2e-llm-inference-service] Configure the number of retries to allow before raising a [e2e-llm-inference-service] :class:`~urllib3.exceptions.MaxRetryError` exception. [e2e-llm-inference-service] [e2e-llm-inference-service] If ``None`` (default) will retry 3 times, see ``Retry.DEFAULT``. Pass a [e2e-llm-inference-service] :class:`~urllib3.util.retry.Retry` object for fine-grained control [e2e-llm-inference-service] over different types of retries. [e2e-llm-inference-service] Pass an integer number to retry connection errors that many times, [e2e-llm-inference-service] but no other types of errors. Pass zero to never retry. [e2e-llm-inference-service] [e2e-llm-inference-service] If ``False``, then retries are disabled and any exception is raised [e2e-llm-inference-service] immediately. Also, instead of raising a MaxRetryError on redirects, [e2e-llm-inference-service] the redirect response will be returned. [e2e-llm-inference-service] [e2e-llm-inference-service] :type retries: :class:`~urllib3.util.retry.Retry`, False, or an int. [e2e-llm-inference-service] [e2e-llm-inference-service] :param redirect: [e2e-llm-inference-service] If True, automatically handle redirects (status codes 301, 302, [e2e-llm-inference-service] 303, 307, 308). Each redirect counts as a retry. Disabling retries [e2e-llm-inference-service] will disable redirect, too. [e2e-llm-inference-service] [e2e-llm-inference-service] :param assert_same_host: [e2e-llm-inference-service] If ``True``, will make sure that the host of the pool requests is [e2e-llm-inference-service] consistent else will raise HostChangedError. When ``False``, you can [e2e-llm-inference-service] use the pool on an HTTP proxy and request foreign hosts. [e2e-llm-inference-service] [e2e-llm-inference-service] :param timeout: [e2e-llm-inference-service] If specified, overrides the default timeout for this one [e2e-llm-inference-service] request. It may be a float (in seconds) or an instance of [e2e-llm-inference-service] :class:`urllib3.util.Timeout`. [e2e-llm-inference-service] [e2e-llm-inference-service] :param pool_timeout: [e2e-llm-inference-service] If set and the pool is set to block=True, then this method will [e2e-llm-inference-service] block for ``pool_timeout`` seconds and raise EmptyPoolError if no [e2e-llm-inference-service] connection is available within the time period. [e2e-llm-inference-service] [e2e-llm-inference-service] :param bool preload_content: [e2e-llm-inference-service] If True, the response's body will be preloaded into memory. [e2e-llm-inference-service] [e2e-llm-inference-service] :param bool decode_content: [e2e-llm-inference-service] If True, will attempt to decode the body based on the [e2e-llm-inference-service] 'content-encoding' header. [e2e-llm-inference-service] [e2e-llm-inference-service] :param release_conn: [e2e-llm-inference-service] If False, then the urlopen call will not release the connection [e2e-llm-inference-service] back into the pool once a response is received (but will release if [e2e-llm-inference-service] you read the entire contents of the response such as when [e2e-llm-inference-service] `preload_content=True`). This is useful if you're not preloading [e2e-llm-inference-service] the response's content immediately. You will need to call [e2e-llm-inference-service] ``r.release_conn()`` on the response ``r`` to return the connection [e2e-llm-inference-service] back into the pool. If None, it takes the value of ``preload_content`` [e2e-llm-inference-service] which defaults to ``True``. [e2e-llm-inference-service] [e2e-llm-inference-service] :param bool chunked: [e2e-llm-inference-service] If True, urllib3 will send the body using chunked transfer [e2e-llm-inference-service] encoding. Otherwise, urllib3 will send the body using the standard [e2e-llm-inference-service] content-length form. Defaults to False. [e2e-llm-inference-service] [e2e-llm-inference-service] :param int body_pos: [e2e-llm-inference-service] Position to seek to in file-like body in the event of a retry or [e2e-llm-inference-service] redirect. Typically this won't need to be set because urllib3 will [e2e-llm-inference-service] auto-populate the value when needed. [e2e-llm-inference-service] """ [e2e-llm-inference-service] parsed_url = parse_url(url) [e2e-llm-inference-service] destination_scheme = parsed_url.scheme [e2e-llm-inference-service] [e2e-llm-inference-service] if headers is None: [e2e-llm-inference-service] headers = self.headers [e2e-llm-inference-service] [e2e-llm-inference-service] if not isinstance(retries, Retry): [e2e-llm-inference-service] retries = Retry.from_int(retries, redirect=redirect, default=self.retries) [e2e-llm-inference-service] [e2e-llm-inference-service] if release_conn is None: [e2e-llm-inference-service] release_conn = preload_content [e2e-llm-inference-service] [e2e-llm-inference-service] # Check host [e2e-llm-inference-service] if assert_same_host and not self.is_same_host(url): [e2e-llm-inference-service] raise HostChangedError(self, url, retries) [e2e-llm-inference-service] [e2e-llm-inference-service] # Ensure that the URL we're connecting to is properly encoded [e2e-llm-inference-service] if url.startswith("/"): [e2e-llm-inference-service] url = to_str(_encode_target(url)) [e2e-llm-inference-service] else: [e2e-llm-inference-service] url = to_str(parsed_url.url) [e2e-llm-inference-service] [e2e-llm-inference-service] conn = None [e2e-llm-inference-service] [e2e-llm-inference-service] # Track whether `conn` needs to be released before [e2e-llm-inference-service] # returning/raising/recursing. Update this variable if necessary, and [e2e-llm-inference-service] # leave `release_conn` constant throughout the function. That way, if [e2e-llm-inference-service] # the function recurses, the original value of `release_conn` will be [e2e-llm-inference-service] # passed down into the recursive call, and its value will be respected. [e2e-llm-inference-service] # [e2e-llm-inference-service] # See issue #651 [1] for details. [e2e-llm-inference-service] # [e2e-llm-inference-service] # [1] [e2e-llm-inference-service] release_this_conn = release_conn [e2e-llm-inference-service] [e2e-llm-inference-service] http_tunnel_required = connection_requires_http_tunnel( [e2e-llm-inference-service] self.proxy, self.proxy_config, destination_scheme [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] # Merge the proxy headers. Only done when not using HTTP CONNECT. We [e2e-llm-inference-service] # have to copy the headers dict so we can safely change it without those [e2e-llm-inference-service] # changes being reflected in anyone else's copy. [e2e-llm-inference-service] if not http_tunnel_required: [e2e-llm-inference-service] headers = headers.copy() # type: ignore[attr-defined] [e2e-llm-inference-service] headers.update(self.proxy_headers) # type: ignore[union-attr] [e2e-llm-inference-service] [e2e-llm-inference-service] # Must keep the exception bound to a separate variable or else Python 3 [e2e-llm-inference-service] # complains about UnboundLocalError. [e2e-llm-inference-service] err = None [e2e-llm-inference-service] [e2e-llm-inference-service] # Keep track of whether we cleanly exited the except block. This [e2e-llm-inference-service] # ensures we do proper cleanup in finally. [e2e-llm-inference-service] clean_exit = False [e2e-llm-inference-service] [e2e-llm-inference-service] # Rewind body position, if needed. Record current position [e2e-llm-inference-service] # for future rewinds in the event of a redirect/retry. [e2e-llm-inference-service] body_pos = set_file_position(body, body_pos) [e2e-llm-inference-service] [e2e-llm-inference-service] try: [e2e-llm-inference-service] # Request a connection from the queue. [e2e-llm-inference-service] timeout_obj = self._get_timeout(timeout) [e2e-llm-inference-service] conn = self._get_conn(timeout=pool_timeout) [e2e-llm-inference-service] [e2e-llm-inference-service] conn.timeout = timeout_obj.connect_timeout # type: ignore[assignment] [e2e-llm-inference-service] [e2e-llm-inference-service] # Is this a closed/new connection that requires CONNECT tunnelling? [e2e-llm-inference-service] if self.proxy is not None and http_tunnel_required and conn.is_closed: [e2e-llm-inference-service] try: [e2e-llm-inference-service] self._prepare_proxy(conn) [e2e-llm-inference-service] except (BaseSSLError, OSError, SocketTimeout) as e: [e2e-llm-inference-service] self._raise_timeout( [e2e-llm-inference-service] err=e, url=self.proxy.url, timeout_value=conn.timeout [e2e-llm-inference-service] ) [e2e-llm-inference-service] raise [e2e-llm-inference-service] [e2e-llm-inference-service] # If we're going to release the connection in ``finally:``, then [e2e-llm-inference-service] # the response doesn't need to know about the connection. Otherwise [e2e-llm-inference-service] # it will also try to release it and we'll have a double-release [e2e-llm-inference-service] # mess. [e2e-llm-inference-service] response_conn = conn if not release_conn else None [e2e-llm-inference-service] [e2e-llm-inference-service] # Make the request on the HTTPConnection object [e2e-llm-inference-service] response = self._make_request( [e2e-llm-inference-service] conn, [e2e-llm-inference-service] method, [e2e-llm-inference-service] url, [e2e-llm-inference-service] timeout=timeout_obj, [e2e-llm-inference-service] body=body, [e2e-llm-inference-service] headers=headers, [e2e-llm-inference-service] chunked=chunked, [e2e-llm-inference-service] retries=retries, [e2e-llm-inference-service] response_conn=response_conn, [e2e-llm-inference-service] preload_content=preload_content, [e2e-llm-inference-service] decode_content=decode_content, [e2e-llm-inference-service] **response_kw, [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] # Everything went great! [e2e-llm-inference-service] clean_exit = True [e2e-llm-inference-service] [e2e-llm-inference-service] except EmptyPoolError: [e2e-llm-inference-service] # Didn't get a connection from the pool, no need to clean up [e2e-llm-inference-service] clean_exit = True [e2e-llm-inference-service] release_this_conn = False [e2e-llm-inference-service] raise [e2e-llm-inference-service] [e2e-llm-inference-service] except ( [e2e-llm-inference-service] TimeoutError, [e2e-llm-inference-service] HTTPException, [e2e-llm-inference-service] OSError, [e2e-llm-inference-service] ProtocolError, [e2e-llm-inference-service] BaseSSLError, [e2e-llm-inference-service] SSLError, [e2e-llm-inference-service] CertificateError, [e2e-llm-inference-service] ProxyError, [e2e-llm-inference-service] ) as e: [e2e-llm-inference-service] # Discard the connection for these exceptions. It will be [e2e-llm-inference-service] # replaced during the next _get_conn() call. [e2e-llm-inference-service] clean_exit = False [e2e-llm-inference-service] new_e: Exception = e [e2e-llm-inference-service] if isinstance(e, (BaseSSLError, CertificateError)): [e2e-llm-inference-service] new_e = SSLError(e) [e2e-llm-inference-service] if isinstance( [e2e-llm-inference-service] new_e, [e2e-llm-inference-service] ( [e2e-llm-inference-service] OSError, [e2e-llm-inference-service] NewConnectionError, [e2e-llm-inference-service] TimeoutError, [e2e-llm-inference-service] SSLError, [e2e-llm-inference-service] HTTPException, [e2e-llm-inference-service] ), [e2e-llm-inference-service] ) and (conn and conn.proxy and not conn.has_connected_to_proxy): [e2e-llm-inference-service] new_e = _wrap_proxy_error(new_e, conn.proxy.scheme) [e2e-llm-inference-service] elif isinstance(new_e, (OSError, HTTPException)): [e2e-llm-inference-service] new_e = ProtocolError("Connection aborted.", new_e) [e2e-llm-inference-service] [e2e-llm-inference-service] > retries = retries.increment( [e2e-llm-inference-service] method, url, error=new_e, _pool=self, _stacktrace=sys.exc_info()[2] [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] ../../python/kserve/.venv/lib64/python3.11/site-packages/urllib3/connectionpool.py:841: [e2e-llm-inference-service] _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ [e2e-llm-inference-service] [e2e-llm-inference-service] self = Retry(total=0, connect=None, read=None, redirect=None, status=None) [e2e-llm-inference-service] method = 'GET', url = '/v1/models', response = None [e2e-llm-inference-service] error = ReadTimeoutError("HTTPConnectionPool(host='ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com', port=80): Read timed out. (read timeout=60)") [e2e-llm-inference-service] _pool = [e2e-llm-inference-service] _stacktrace = [e2e-llm-inference-service] [e2e-llm-inference-service] def increment( [e2e-llm-inference-service] self, [e2e-llm-inference-service] method: str | None = None, [e2e-llm-inference-service] url: str | None = None, [e2e-llm-inference-service] response: BaseHTTPResponse | None = None, [e2e-llm-inference-service] error: Exception | None = None, [e2e-llm-inference-service] _pool: ConnectionPool | None = None, [e2e-llm-inference-service] _stacktrace: TracebackType | None = None, [e2e-llm-inference-service] ) -> Self: [e2e-llm-inference-service] """Return a new Retry object with incremented retry counters. [e2e-llm-inference-service] [e2e-llm-inference-service] :param response: A response object, or None, if the server did not [e2e-llm-inference-service] return a response. [e2e-llm-inference-service] :type response: :class:`~urllib3.response.BaseHTTPResponse` [e2e-llm-inference-service] :param Exception error: An error encountered during the request, or [e2e-llm-inference-service] None if the response was received successfully. [e2e-llm-inference-service] [e2e-llm-inference-service] :return: A new ``Retry`` object. [e2e-llm-inference-service] """ [e2e-llm-inference-service] if self.total is False and error: [e2e-llm-inference-service] # Disabled, indicate to re-raise the error. [e2e-llm-inference-service] raise reraise(type(error), error, _stacktrace) [e2e-llm-inference-service] [e2e-llm-inference-service] total = self.total [e2e-llm-inference-service] if total is not None: [e2e-llm-inference-service] total -= 1 [e2e-llm-inference-service] [e2e-llm-inference-service] connect = self.connect [e2e-llm-inference-service] read = self.read [e2e-llm-inference-service] redirect = self.redirect [e2e-llm-inference-service] status_count = self.status [e2e-llm-inference-service] other = self.other [e2e-llm-inference-service] cause = "unknown" [e2e-llm-inference-service] status = None [e2e-llm-inference-service] redirect_location = None [e2e-llm-inference-service] [e2e-llm-inference-service] if error and self._is_connection_error(error): [e2e-llm-inference-service] # Connect retry? [e2e-llm-inference-service] if connect is False: [e2e-llm-inference-service] raise reraise(type(error), error, _stacktrace) [e2e-llm-inference-service] elif connect is not None: [e2e-llm-inference-service] connect -= 1 [e2e-llm-inference-service] [e2e-llm-inference-service] elif error and self._is_read_error(error): [e2e-llm-inference-service] # Read retry? [e2e-llm-inference-service] if read is False or method is None or not self._is_method_retryable(method): [e2e-llm-inference-service] raise reraise(type(error), error, _stacktrace) [e2e-llm-inference-service] elif read is not None: [e2e-llm-inference-service] read -= 1 [e2e-llm-inference-service] [e2e-llm-inference-service] elif error: [e2e-llm-inference-service] # Other retry? [e2e-llm-inference-service] if other is not None: [e2e-llm-inference-service] other -= 1 [e2e-llm-inference-service] [e2e-llm-inference-service] elif response and response.get_redirect_location(): [e2e-llm-inference-service] # Redirect retry? [e2e-llm-inference-service] if redirect is not None: [e2e-llm-inference-service] redirect -= 1 [e2e-llm-inference-service] cause = "too many redirects" [e2e-llm-inference-service] response_redirect_location = response.get_redirect_location() [e2e-llm-inference-service] if response_redirect_location: [e2e-llm-inference-service] redirect_location = response_redirect_location [e2e-llm-inference-service] status = response.status [e2e-llm-inference-service] [e2e-llm-inference-service] else: [e2e-llm-inference-service] # Incrementing because of a server error like a 500 in [e2e-llm-inference-service] # status_forcelist and the given method is in the allowed_methods [e2e-llm-inference-service] cause = ResponseError.GENERIC_ERROR [e2e-llm-inference-service] if response and response.status: [e2e-llm-inference-service] if status_count is not None: [e2e-llm-inference-service] status_count -= 1 [e2e-llm-inference-service] cause = ResponseError.SPECIFIC_ERROR.format(status_code=response.status) [e2e-llm-inference-service] status = response.status [e2e-llm-inference-service] [e2e-llm-inference-service] history = self.history + ( [e2e-llm-inference-service] RequestHistory(method, url, error, status, redirect_location), [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] new_retry = self.new( [e2e-llm-inference-service] total=total, [e2e-llm-inference-service] connect=connect, [e2e-llm-inference-service] read=read, [e2e-llm-inference-service] redirect=redirect, [e2e-llm-inference-service] status=status_count, [e2e-llm-inference-service] other=other, [e2e-llm-inference-service] history=history, [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] if new_retry.is_exhausted(): [e2e-llm-inference-service] reason = error or ResponseError(cause) [e2e-llm-inference-service] > raise MaxRetryError(_pool, url, reason) from reason # type: ignore[arg-type] [e2e-llm-inference-service] E urllib3.exceptions.MaxRetryError: HTTPConnectionPool(host='ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com', port=80): Max retries exceeded with url: /v1/models (Caused by ReadTimeoutError("HTTPConnectionPool(host='ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com', port=80): Read timed out. (read timeout=60)")) [e2e-llm-inference-service] [e2e-llm-inference-service] ../../python/kserve/.venv/lib64/python3.11/site-packages/urllib3/util/retry.py:519: MaxRetryError [e2e-llm-inference-service] [e2e-llm-inference-service] During handling of the above exception, another exception occurred: [e2e-llm-inference-service] [e2e-llm-inference-service] def get_successful_response(): [e2e-llm-inference-service] try: [e2e-llm-inference-service] if test_case.url_getter: [e2e-llm-inference-service] service_url = test_case.url_getter(kserve_client, test_case.llm_service) [e2e-llm-inference-service] else: [e2e-llm-inference-service] service_url = get_llm_service_url(kserve_client, test_case.llm_service) [e2e-llm-inference-service] except Exception as e: [e2e-llm-inference-service] raise AssertionError(f"❌ Failed to get service URL: {e}") from e [e2e-llm-inference-service] [e2e-llm-inference-service] model_url = service_url + test_case.endpoint [e2e-llm-inference-service] [e2e-llm-inference-service] headers = {"Content-Type": "application/json"} [e2e-llm-inference-service] if extra_headers: [e2e-llm-inference-service] headers.update(extra_headers) [e2e-llm-inference-service] [e2e-llm-inference-service] if test_case.payload_formatter is not None: [e2e-llm-inference-service] test_payload = test_case.payload_formatter(test_case) [e2e-llm-inference-service] elif test_case.prompt is not None: [e2e-llm-inference-service] test_payload = { [e2e-llm-inference-service] "model": test_case.model_name [e2e-llm-inference-service] if not extra_headers or MODEL_ROUTING_HEADER not in extra_headers [e2e-llm-inference-service] else extra_headers[MODEL_ROUTING_HEADER], [e2e-llm-inference-service] "prompt": test_case.prompt, [e2e-llm-inference-service] "max_tokens": test_case.max_tokens, [e2e-llm-inference-service] } [e2e-llm-inference-service] else: [e2e-llm-inference-service] test_payload = None [e2e-llm-inference-service] [e2e-llm-inference-service] logger.info(f"Calling LLM service at {model_url} with payload {test_payload}") [e2e-llm-inference-service] try: [e2e-llm-inference-service] if test_payload is not None: [e2e-llm-inference-service] response = post_with_retry( [e2e-llm-inference-service] model_url, [e2e-llm-inference-service] headers=headers, [e2e-llm-inference-service] json_data=test_payload, [e2e-llm-inference-service] timeout=test_case.response_timeout, [e2e-llm-inference-service] ) [e2e-llm-inference-service] else: [e2e-llm-inference-service] > response = get_with_retry( [e2e-llm-inference-service] model_url, [e2e-llm-inference-service] headers=headers, [e2e-llm-inference-service] timeout=test_case.response_timeout, [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] llmisvc/test_llm_inference_service.py:1102: [e2e-llm-inference-service] _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ [e2e-llm-inference-service] [e2e-llm-inference-service] url = 'http://ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com/v1/models' [e2e-llm-inference-service] [e2e-llm-inference-service] def get_with_retry( [e2e-llm-inference-service] url: str, [e2e-llm-inference-service] *, [e2e-llm-inference-service] headers: Dict = None, [e2e-llm-inference-service] timeout: float = None, [e2e-llm-inference-service] total_retries: int = DEFAULT_RETRY_TOTAL, [e2e-llm-inference-service] backoff_factor: float = DEFAULT_RETRY_BACKOFF_FACTOR, [e2e-llm-inference-service] retry_status_codes=DEFAULT_RETRY_STATUS_CODES, [e2e-llm-inference-service] ) -> requests.Response: [e2e-llm-inference-service] """ [e2e-llm-inference-service] Send GET request with retries for transient HTTP and network failures. [e2e-llm-inference-service] """ [e2e-llm-inference-service] with _retry_session( [e2e-llm-inference-service] ["GET"], total_retries, backoff_factor, retry_status_codes [e2e-llm-inference-service] ) as session: [e2e-llm-inference-service] > return session.get(url, headers=headers, timeout=timeout) [e2e-llm-inference-service] [e2e-llm-inference-service] common/http_retry.py:46: [e2e-llm-inference-service] _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ [e2e-llm-inference-service] [e2e-llm-inference-service] self = [e2e-llm-inference-service] url = 'http://ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com/v1/models' [e2e-llm-inference-service] kwargs = {'allow_redirects': True, 'headers': {'Content-Type': 'application/json', 'X-Gateway-Model-Name': 'publishers/kserve-ci-e2e-test/models/facebook/opt-125m'}, 'timeout': 60} [e2e-llm-inference-service] [e2e-llm-inference-service] def get(self, url, **kwargs): [e2e-llm-inference-service] r"""Sends a GET request. Returns :class:`Response` object. [e2e-llm-inference-service] [e2e-llm-inference-service] :param url: URL for the new :class:`Request` object. [e2e-llm-inference-service] :param \*\*kwargs: Optional arguments that ``request`` takes. [e2e-llm-inference-service] :rtype: requests.Response [e2e-llm-inference-service] """ [e2e-llm-inference-service] [e2e-llm-inference-service] kwargs.setdefault("allow_redirects", True) [e2e-llm-inference-service] > return self.request("GET", url, **kwargs) [e2e-llm-inference-service] [e2e-llm-inference-service] ../../python/kserve/.venv/lib64/python3.11/site-packages/requests/sessions.py:602: [e2e-llm-inference-service] _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ [e2e-llm-inference-service] [e2e-llm-inference-service] self = , method = 'GET' [e2e-llm-inference-service] url = 'http://ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com/v1/models' [e2e-llm-inference-service] params = None, data = None [e2e-llm-inference-service] headers = {'Content-Type': 'application/json', 'X-Gateway-Model-Name': 'publishers/kserve-ci-e2e-test/models/facebook/opt-125m'} [e2e-llm-inference-service] cookies = None, files = None, auth = None, timeout = 60, allow_redirects = True [e2e-llm-inference-service] proxies = {}, hooks = None, stream = None, verify = None, cert = None [e2e-llm-inference-service] json = None [e2e-llm-inference-service] [e2e-llm-inference-service] def request( [e2e-llm-inference-service] self, [e2e-llm-inference-service] method, [e2e-llm-inference-service] url, [e2e-llm-inference-service] params=None, [e2e-llm-inference-service] data=None, [e2e-llm-inference-service] headers=None, [e2e-llm-inference-service] cookies=None, [e2e-llm-inference-service] files=None, [e2e-llm-inference-service] auth=None, [e2e-llm-inference-service] timeout=None, [e2e-llm-inference-service] allow_redirects=True, [e2e-llm-inference-service] proxies=None, [e2e-llm-inference-service] hooks=None, [e2e-llm-inference-service] stream=None, [e2e-llm-inference-service] verify=None, [e2e-llm-inference-service] cert=None, [e2e-llm-inference-service] json=None, [e2e-llm-inference-service] ): [e2e-llm-inference-service] """Constructs a :class:`Request `, prepares it and sends it. [e2e-llm-inference-service] Returns :class:`Response ` object. [e2e-llm-inference-service] [e2e-llm-inference-service] :param method: method for the new :class:`Request` object. [e2e-llm-inference-service] :param url: URL for the new :class:`Request` object. [e2e-llm-inference-service] :param params: (optional) Dictionary or bytes to be sent in the query [e2e-llm-inference-service] string for the :class:`Request`. [e2e-llm-inference-service] :param data: (optional) Dictionary, list of tuples, bytes, or file-like [e2e-llm-inference-service] object to send in the body of the :class:`Request`. [e2e-llm-inference-service] :param json: (optional) json to send in the body of the [e2e-llm-inference-service] :class:`Request`. [e2e-llm-inference-service] :param headers: (optional) Dictionary of HTTP Headers to send with the [e2e-llm-inference-service] :class:`Request`. [e2e-llm-inference-service] :param cookies: (optional) Dict or CookieJar object to send with the [e2e-llm-inference-service] :class:`Request`. [e2e-llm-inference-service] :param files: (optional) Dictionary of ``'filename': file-like-objects`` [e2e-llm-inference-service] for multipart encoding upload. [e2e-llm-inference-service] :param auth: (optional) Auth tuple or callable to enable [e2e-llm-inference-service] Basic/Digest/Custom HTTP Auth. [e2e-llm-inference-service] :param timeout: (optional) How long to wait for the server to send [e2e-llm-inference-service] data before giving up, as a float, or a :ref:`(connect timeout, [e2e-llm-inference-service] read timeout) ` tuple. [e2e-llm-inference-service] :type timeout: float or tuple [e2e-llm-inference-service] :param allow_redirects: (optional) Set to True by default. [e2e-llm-inference-service] :type allow_redirects: bool [e2e-llm-inference-service] :param proxies: (optional) Dictionary mapping protocol or protocol and [e2e-llm-inference-service] hostname to the URL of the proxy. [e2e-llm-inference-service] :param hooks: (optional) Dictionary mapping hook name to one event or [e2e-llm-inference-service] list of events, event must be callable. [e2e-llm-inference-service] :param stream: (optional) whether to immediately download the response [e2e-llm-inference-service] content. Defaults to ``False``. [e2e-llm-inference-service] :param verify: (optional) Either a boolean, in which case it controls whether we verify [e2e-llm-inference-service] the server's TLS certificate, or a string, in which case it must be a path [e2e-llm-inference-service] to a CA bundle to use. Defaults to ``True``. When set to [e2e-llm-inference-service] ``False``, requests will accept any TLS certificate presented by [e2e-llm-inference-service] the server, and will ignore hostname mismatches and/or expired [e2e-llm-inference-service] certificates, which will make your application vulnerable to [e2e-llm-inference-service] man-in-the-middle (MitM) attacks. Setting verify to ``False`` [e2e-llm-inference-service] may be useful during local development or testing. [e2e-llm-inference-service] :param cert: (optional) if String, path to ssl client cert file (.pem). [e2e-llm-inference-service] If Tuple, ('cert', 'key') pair. [e2e-llm-inference-service] :rtype: requests.Response [e2e-llm-inference-service] """ [e2e-llm-inference-service] # Create the Request. [e2e-llm-inference-service] req = Request( [e2e-llm-inference-service] method=method.upper(), [e2e-llm-inference-service] url=url, [e2e-llm-inference-service] headers=headers, [e2e-llm-inference-service] files=files, [e2e-llm-inference-service] data=data or {}, [e2e-llm-inference-service] json=json, [e2e-llm-inference-service] params=params or {}, [e2e-llm-inference-service] auth=auth, [e2e-llm-inference-service] cookies=cookies, [e2e-llm-inference-service] hooks=hooks, [e2e-llm-inference-service] ) [e2e-llm-inference-service] prep = self.prepare_request(req) [e2e-llm-inference-service] [e2e-llm-inference-service] proxies = proxies or {} [e2e-llm-inference-service] [e2e-llm-inference-service] settings = self.merge_environment_settings( [e2e-llm-inference-service] prep.url, proxies, stream, verify, cert [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] # Send the request. [e2e-llm-inference-service] send_kwargs = { [e2e-llm-inference-service] "timeout": timeout, [e2e-llm-inference-service] "allow_redirects": allow_redirects, [e2e-llm-inference-service] } [e2e-llm-inference-service] send_kwargs.update(settings) [e2e-llm-inference-service] > resp = self.send(prep, **send_kwargs) [e2e-llm-inference-service] [e2e-llm-inference-service] ../../python/kserve/.venv/lib64/python3.11/site-packages/requests/sessions.py:589: [e2e-llm-inference-service] _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ [e2e-llm-inference-service] [e2e-llm-inference-service] self = [e2e-llm-inference-service] request = [e2e-llm-inference-service] kwargs = {'cert': None, 'proxies': OrderedDict(), 'stream': False, 'timeout': 60, ...} [e2e-llm-inference-service] allow_redirects = True, stream = False, hooks = {'response': []} [e2e-llm-inference-service] adapter = [e2e-llm-inference-service] start = 1783435637.8514826 [e2e-llm-inference-service] [e2e-llm-inference-service] def send(self, request, **kwargs): [e2e-llm-inference-service] """Send a given PreparedRequest. [e2e-llm-inference-service] [e2e-llm-inference-service] :rtype: requests.Response [e2e-llm-inference-service] """ [e2e-llm-inference-service] # Set defaults that the hooks can utilize to ensure they always have [e2e-llm-inference-service] # the correct parameters to reproduce the previous request. [e2e-llm-inference-service] kwargs.setdefault("stream", self.stream) [e2e-llm-inference-service] kwargs.setdefault("verify", self.verify) [e2e-llm-inference-service] kwargs.setdefault("cert", self.cert) [e2e-llm-inference-service] if "proxies" not in kwargs: [e2e-llm-inference-service] kwargs["proxies"] = resolve_proxies(request, self.proxies, self.trust_env) [e2e-llm-inference-service] [e2e-llm-inference-service] # It's possible that users might accidentally send a Request object. [e2e-llm-inference-service] # Guard against that specific failure case. [e2e-llm-inference-service] if isinstance(request, Request): [e2e-llm-inference-service] raise ValueError("You can only send PreparedRequests.") [e2e-llm-inference-service] [e2e-llm-inference-service] # Set up variables needed for resolve_redirects and dispatching of hooks [e2e-llm-inference-service] allow_redirects = kwargs.pop("allow_redirects", True) [e2e-llm-inference-service] stream = kwargs.get("stream") [e2e-llm-inference-service] hooks = request.hooks [e2e-llm-inference-service] [e2e-llm-inference-service] # Get the appropriate adapter to use [e2e-llm-inference-service] adapter = self.get_adapter(url=request.url) [e2e-llm-inference-service] [e2e-llm-inference-service] # Start time (approximately) of the request [e2e-llm-inference-service] start = preferred_clock() [e2e-llm-inference-service] [e2e-llm-inference-service] # Send the request [e2e-llm-inference-service] > r = adapter.send(request, **kwargs) [e2e-llm-inference-service] [e2e-llm-inference-service] ../../python/kserve/.venv/lib64/python3.11/site-packages/requests/sessions.py:703: [e2e-llm-inference-service] _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ [e2e-llm-inference-service] [e2e-llm-inference-service] self = [e2e-llm-inference-service] request = , stream = False [e2e-llm-inference-service] timeout = Timeout(connect=60, read=60, total=None), verify = '/tmp/ca.crt' [e2e-llm-inference-service] cert = None, proxies = OrderedDict() [e2e-llm-inference-service] [e2e-llm-inference-service] def send( [e2e-llm-inference-service] self, request, stream=False, timeout=None, verify=True, cert=None, proxies=None [e2e-llm-inference-service] ): [e2e-llm-inference-service] """Sends PreparedRequest object. Returns Response object. [e2e-llm-inference-service] [e2e-llm-inference-service] :param request: The :class:`PreparedRequest ` being sent. [e2e-llm-inference-service] :param stream: (optional) Whether to stream the request content. [e2e-llm-inference-service] :param timeout: (optional) How long to wait for the server to send [e2e-llm-inference-service] data before giving up, as a float, or a :ref:`(connect timeout, [e2e-llm-inference-service] read timeout) ` tuple. [e2e-llm-inference-service] :type timeout: float or tuple or urllib3 Timeout object [e2e-llm-inference-service] :param verify: (optional) Either a boolean, in which case it controls whether [e2e-llm-inference-service] we verify the server's TLS certificate, or a string, in which case it [e2e-llm-inference-service] must be a path to a CA bundle to use [e2e-llm-inference-service] :param cert: (optional) Any user-provided SSL certificate to be trusted. [e2e-llm-inference-service] :param proxies: (optional) The proxies dictionary to apply to the request. [e2e-llm-inference-service] :rtype: requests.Response [e2e-llm-inference-service] """ [e2e-llm-inference-service] [e2e-llm-inference-service] try: [e2e-llm-inference-service] conn = self.get_connection_with_tls_context( [e2e-llm-inference-service] request, verify, proxies=proxies, cert=cert [e2e-llm-inference-service] ) [e2e-llm-inference-service] except LocationValueError as e: [e2e-llm-inference-service] raise InvalidURL(e, request=request) [e2e-llm-inference-service] [e2e-llm-inference-service] self.cert_verify(conn, request.url, verify, cert) [e2e-llm-inference-service] url = self.request_url(request, proxies) [e2e-llm-inference-service] self.add_headers( [e2e-llm-inference-service] request, [e2e-llm-inference-service] stream=stream, [e2e-llm-inference-service] timeout=timeout, [e2e-llm-inference-service] verify=verify, [e2e-llm-inference-service] cert=cert, [e2e-llm-inference-service] proxies=proxies, [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] chunked = not (request.body is None or "Content-Length" in request.headers) [e2e-llm-inference-service] [e2e-llm-inference-service] if isinstance(timeout, tuple): [e2e-llm-inference-service] try: [e2e-llm-inference-service] connect, read = timeout [e2e-llm-inference-service] timeout = TimeoutSauce(connect=connect, read=read) [e2e-llm-inference-service] except ValueError: [e2e-llm-inference-service] raise ValueError( [e2e-llm-inference-service] f"Invalid timeout {timeout}. Pass a (connect, read) timeout tuple, " [e2e-llm-inference-service] f"or a single float to set both timeouts to the same value." [e2e-llm-inference-service] ) [e2e-llm-inference-service] elif isinstance(timeout, TimeoutSauce): [e2e-llm-inference-service] pass [e2e-llm-inference-service] else: [e2e-llm-inference-service] timeout = TimeoutSauce(connect=timeout, read=timeout) [e2e-llm-inference-service] [e2e-llm-inference-service] try: [e2e-llm-inference-service] resp = conn.urlopen( [e2e-llm-inference-service] method=request.method, [e2e-llm-inference-service] url=url, [e2e-llm-inference-service] body=request.body, [e2e-llm-inference-service] headers=request.headers, [e2e-llm-inference-service] redirect=False, [e2e-llm-inference-service] assert_same_host=False, [e2e-llm-inference-service] preload_content=False, [e2e-llm-inference-service] decode_content=False, [e2e-llm-inference-service] retries=self.max_retries, [e2e-llm-inference-service] timeout=timeout, [e2e-llm-inference-service] chunked=chunked, [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] except (ProtocolError, OSError) as err: [e2e-llm-inference-service] raise ConnectionError(err, request=request) [e2e-llm-inference-service] [e2e-llm-inference-service] except MaxRetryError as e: [e2e-llm-inference-service] if isinstance(e.reason, ConnectTimeoutError): [e2e-llm-inference-service] # TODO: Remove this in 3.0.0: see #2811 [e2e-llm-inference-service] if not isinstance(e.reason, NewConnectionError): [e2e-llm-inference-service] raise ConnectTimeout(e, request=request) [e2e-llm-inference-service] [e2e-llm-inference-service] if isinstance(e.reason, ResponseError): [e2e-llm-inference-service] raise RetryError(e, request=request) [e2e-llm-inference-service] [e2e-llm-inference-service] if isinstance(e.reason, _ProxyError): [e2e-llm-inference-service] raise ProxyError(e, request=request) [e2e-llm-inference-service] [e2e-llm-inference-service] if isinstance(e.reason, _SSLError): [e2e-llm-inference-service] # This branch is for urllib3 v1.22 and later. [e2e-llm-inference-service] raise SSLError(e, request=request) [e2e-llm-inference-service] [e2e-llm-inference-service] > raise ConnectionError(e, request=request) [e2e-llm-inference-service] E requests.exceptions.ConnectionError: HTTPConnectionPool(host='ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com', port=80): Max retries exceeded with url: /v1/models (Caused by ReadTimeoutError("HTTPConnectionPool(host='ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com', port=80): Read timed out. (read timeout=60)")) [e2e-llm-inference-service] [e2e-llm-inference-service] ../../python/kserve/.venv/lib64/python3.11/site-packages/requests/adapters.py:700: ConnectionError [e2e-llm-inference-service] [e2e-llm-inference-service] The above exception was the direct cause of the following exception: [e2e-llm-inference-service] [e2e-llm-inference-service] test_case = TestCase(base_refs=['router-managed', 'workload-single-cpu', 'model-fb-opt-125m-with-lora-hf'], prompt=None, service_n... {'name': 'model-fb-opt-125m-with-lora-hf-c0d503b0'}]}, [e2e-llm-inference-service] 'status': None}, model_name='facebook/opt-125m') [e2e-llm-inference-service] [e2e-llm-inference-service] @pytest.mark.llminferenceservice [e2e-llm-inference-service] @pytest.mark.asyncio(loop_scope="session") [e2e-llm-inference-service] @pytest.mark.parametrize( [e2e-llm-inference-service] "test_case", [e2e-llm-inference-service] [ [e2e-llm-inference-service] pytest.param( [e2e-llm-inference-service] TestCase( [e2e-llm-inference-service] base_refs=[ [e2e-llm-inference-service] "router-with-gateway-ref", [e2e-llm-inference-service] "router-with-managed-route", [e2e-llm-inference-service] "model-fb-opt-125m", [e2e-llm-inference-service] "workload-llmd-simulator", [e2e-llm-inference-service] ], [e2e-llm-inference-service] endpoint="/v1/completions", [e2e-llm-inference-service] prompt="KServe is a", [e2e-llm-inference-service] payload_formatter=completions_payload, [e2e-llm-inference-service] response_assertion=create_response_assertion(with_field="choices"), [e2e-llm-inference-service] expected_gateway=ROUTER_GATEWAYS[0], [e2e-llm-inference-service] before_test=[ [e2e-llm-inference-service] lambda: create_router_resources( [e2e-llm-inference-service] gateways=[ROUTER_GATEWAYS[0]], [e2e-llm-inference-service] ) [e2e-llm-inference-service] ], [e2e-llm-inference-service] ), [e2e-llm-inference-service] marks=[ [e2e-llm-inference-service] pytest.mark.cluster_cpu, [e2e-llm-inference-service] pytest.mark.cluster_single_node, [e2e-llm-inference-service] pytest.mark.llmd_simulator, [e2e-llm-inference-service] pytest.mark.custom_gateway, [e2e-llm-inference-service] ], [e2e-llm-inference-service] ), [e2e-llm-inference-service] pytest.param( [e2e-llm-inference-service] TestCase( [e2e-llm-inference-service] base_refs=[ [e2e-llm-inference-service] "router-managed", [e2e-llm-inference-service] "workload-single-cpu", [e2e-llm-inference-service] "model-fb-opt-125m", [e2e-llm-inference-service] ], [e2e-llm-inference-service] prompt="KServe is a", [e2e-llm-inference-service] payload_formatter=completions_payload, [e2e-llm-inference-service] response_assertion=assert_200_with_choices, [e2e-llm-inference-service] ), [e2e-llm-inference-service] marks=[pytest.mark.cluster_cpu, pytest.mark.cluster_single_node], [e2e-llm-inference-service] ), [e2e-llm-inference-service] pytest.param( [e2e-llm-inference-service] TestCase( [e2e-llm-inference-service] base_refs=[ [e2e-llm-inference-service] "router-custom-route-timeout", [e2e-llm-inference-service] "scheduler-managed", [e2e-llm-inference-service] "workload-single-cpu", [e2e-llm-inference-service] "model-fb-opt-125m", [e2e-llm-inference-service] ], [e2e-llm-inference-service] prompt="KServe is a", [e2e-llm-inference-service] service_name="custom-route-timeout-test", [e2e-llm-inference-service] ), [e2e-llm-inference-service] marks=[pytest.mark.cluster_cpu, pytest.mark.cluster_single_node], [e2e-llm-inference-service] ), [e2e-llm-inference-service] pytest.param( [e2e-llm-inference-service] TestCase( [e2e-llm-inference-service] base_refs=[ [e2e-llm-inference-service] "router-with-refs", [e2e-llm-inference-service] "scheduler-managed", [e2e-llm-inference-service] "workload-single-cpu", [e2e-llm-inference-service] "model-fb-opt-125m", [e2e-llm-inference-service] ], [e2e-llm-inference-service] prompt="KServe is a", [e2e-llm-inference-service] service_name="router-with-refs-test", [e2e-llm-inference-service] expected_gateway=ROUTER_GATEWAYS[0], [e2e-llm-inference-service] before_test=[ [e2e-llm-inference-service] lambda: create_router_resources( [e2e-llm-inference-service] gateways=[ROUTER_GATEWAYS[0]], [e2e-llm-inference-service] routes=[ROUTER_ROUTES[0], ROUTER_ROUTES[1]], [e2e-llm-inference-service] ) [e2e-llm-inference-service] ], [e2e-llm-inference-service] ), [e2e-llm-inference-service] marks=[ [e2e-llm-inference-service] pytest.mark.cluster_cpu, [e2e-llm-inference-service] pytest.mark.cluster_single_node, [e2e-llm-inference-service] pytest.mark.custom_gateway, [e2e-llm-inference-service] ], [e2e-llm-inference-service] ), [e2e-llm-inference-service] pytest.param( [e2e-llm-inference-service] TestCase( [e2e-llm-inference-service] base_refs=["router-managed", "workload-pd-cpu", "model-fb-opt-125m"], [e2e-llm-inference-service] prompt="You are an expert in Kubernetes-native machine learning serving platforms, with deep knowledge of the KServe project. " [e2e-llm-inference-service] "Explain the challenges of serving large-scale models, GPU scheduling, and how KServe integrates with capabilities like multi-model serving. " [e2e-llm-inference-service] "Provide a detailed comparison with open source alternatives, focusing on operational trade-offs.", [e2e-llm-inference-service] response_assertion=assert_200_with_choices, [e2e-llm-inference-service] ), [e2e-llm-inference-service] marks=[pytest.mark.cluster_cpu, pytest.mark.cluster_single_node], [e2e-llm-inference-service] ), [e2e-llm-inference-service] pytest.param( [e2e-llm-inference-service] TestCase( [e2e-llm-inference-service] base_refs=[ [e2e-llm-inference-service] "router-custom-route-timeout-pd", [e2e-llm-inference-service] "scheduler-managed", [e2e-llm-inference-service] "workload-pd-cpu", [e2e-llm-inference-service] "model-fb-opt-125m", [e2e-llm-inference-service] ], [e2e-llm-inference-service] prompt="You are an expert in Kubernetes-native machine learning serving platforms, with deep knowledge of the KServe project. " [e2e-llm-inference-service] "Explain the challenges of serving large-scale models, GPU scheduling, and how KServe integrates with capabilities like multi-model serving. " [e2e-llm-inference-service] "Provide a detailed comparison with open source alternatives, focusing on operational trade-offs.", [e2e-llm-inference-service] service_name="custom-route-timeout-pd-test", [e2e-llm-inference-service] response_assertion=assert_200_with_choices, [e2e-llm-inference-service] ), [e2e-llm-inference-service] marks=[pytest.mark.cluster_cpu, pytest.mark.cluster_single_node], [e2e-llm-inference-service] ), [e2e-llm-inference-service] pytest.param( [e2e-llm-inference-service] TestCase( [e2e-llm-inference-service] base_refs=[ [e2e-llm-inference-service] "router-with-refs-pd", [e2e-llm-inference-service] "scheduler-managed", [e2e-llm-inference-service] "workload-pd-cpu", [e2e-llm-inference-service] "model-fb-opt-125m", [e2e-llm-inference-service] ], [e2e-llm-inference-service] prompt="You are an expert in Kubernetes-native machine learning serving platforms, with deep knowledge of the KServe project. " [e2e-llm-inference-service] "Explain the challenges of serving large-scale models, GPU scheduling, and how KServe integrates with capabilities like multi-model serving. " [e2e-llm-inference-service] "Provide a detailed comparison with open source alternatives, focusing on operational trade-offs.", [e2e-llm-inference-service] service_name="router-with-refs-pd-test", [e2e-llm-inference-service] response_assertion=assert_200_with_choices, [e2e-llm-inference-service] expected_gateway=ROUTER_GATEWAYS[1], [e2e-llm-inference-service] before_test=[ [e2e-llm-inference-service] lambda: create_router_resources( [e2e-llm-inference-service] gateways=[ROUTER_GATEWAYS[1]], [e2e-llm-inference-service] routes=[ROUTER_ROUTES[2], ROUTER_ROUTES[3]], [e2e-llm-inference-service] ) [e2e-llm-inference-service] ], [e2e-llm-inference-service] ), [e2e-llm-inference-service] marks=[ [e2e-llm-inference-service] pytest.mark.cluster_cpu, [e2e-llm-inference-service] pytest.mark.cluster_single_node, [e2e-llm-inference-service] pytest.mark.custom_gateway, [e2e-llm-inference-service] ], [e2e-llm-inference-service] ), [e2e-llm-inference-service] pytest.param( [e2e-llm-inference-service] TestCase( [e2e-llm-inference-service] base_refs=[ [e2e-llm-inference-service] "router-managed", [e2e-llm-inference-service] "workload-dp-ep-gpu", [e2e-llm-inference-service] "workload-dp-ep-prefill-gpu", [e2e-llm-inference-service] "model-deepseek-v2-lite", [e2e-llm-inference-service] ], [e2e-llm-inference-service] prompt="Delve into the multifaceted implications of a fully disaggregated cloud architecture, specifically " [e2e-llm-inference-service] "where the compute plane (P) and the data plane (D) are independently deployed and managed for a " [e2e-llm-inference-service] "geographically distributed, high-throughput, low-latency microservices ecosystem. Beyond the " [e2e-llm-inference-service] "fundamental challenges of network latency and data consistency, elaborate on the advanced " [e2e-llm-inference-service] "considerations and trade-offs inherent in such a setup: 1. Network Architecture and Protocols: " [e2e-llm-inference-service] "How would the network fabric and underlying protocols (e.g., RDMA, custom transport layers) need to " [e2e-llm-inference-service] "evolve to support optimal performance and minimize inter-plane communication overhead, especially for " [e2e-llm-inference-service] "synchronous operations? Discuss the role of network programmability (e.g., SDN, P4) in dynamically " [e2e-llm-inference-service] "optimizing routing and traffic flow between P and D. 2. Advanced Data Consistency and Durability: " [e2e-llm-inference-service] "Explore sophisticated data consistency models (e.g., causal consistency, strong eventual consistency) " [e2e-llm-inference-service] "and their applicability in balancing performance and data integrity across a globally distributed data plane. " [e2e-llm-inference-service] "Detail strategies for ensuring data durability and fault tolerance, including multi-region replication, " [e2e-llm-inference-service] "intelligent partitioning, and recovery mechanisms in the event of partial or full plane failures. " [e2e-llm-inference-service] "3. Dynamic Resource Orchestration and Cost Optimization: Analyze how an orchestration layer would intelligently " [e2e-llm-inference-service] "manage the independent scaling of compute (P) and data (D) resources, considering fluctuating workloads, " [e2e-llm-inference-service] "cost efficiency, and performance targets (e.g., using predictive analytics for resource provisioning). " [e2e-llm-inference-service] "Discuss mechanisms for dynamically reallocating compute nodes to different data partitions based on " [e2e-llm-inference-service] "workload patterns and data locality, potentially involving live migration strategies. " [e2e-llm-inference-service] "4. Security and Compliance in a Distributed Landscape: Address the enhanced security perimeter " [e2e-llm-inference-service] "challenges, including securing communication channels between P and D (encryption in transit, mutual TLS), " [e2e-llm-inference-service] "fine-grained access control to data at rest and in motion, and identity management across disaggregated " [e2e-llm-inference-service] "components. Discuss how such an architecture impacts compliance with regulatory frameworks (e.g., GDPR, HIPAA) " [e2e-llm-inference-service] "concerning data sovereignty, privacy, and auditability. 5. Operational Complexity and Observability: " [e2e-llm-inference-service] "Examine the increased complexity in monitoring, logging, and tracing across highly decoupled compute and " [e2e-llm-inference-service] "data planes. What specialized tooling and practices (e.g., distributed tracing with OpenTelemetry, advanced AIOps) " [e2e-llm-inference-service] "would be essential? How would incident response and troubleshooting differ in this disaggregated environment " [e2e-llm-inference-service] "compared to traditional integrated systems? Consider the challenges of pinpointing root causes across " [e2e-llm-inference-service] "independent failures. 6. Real-world Applicability and Future Trends: Identify specific industries " [e2e-llm-inference-service] "or use cases (e.g., high-frequency trading, IoT edge processing, large language model inference) " [e2e-llm-inference-service] "where the benefits of P/D disaggregation would strongly outweigh its complexities. " [e2e-llm-inference-service] "Conclude by speculating on emerging technologies or paradigms (e.g., serverless compute functions " [e2e-llm-inference-service] "directly interacting with object storage, in-memory disaggregation) that could further drive or " [e2e-llm-inference-service] "transform P/D disaggregation in cloud computing.", [e2e-llm-inference-service] max_tokens=2000, [e2e-llm-inference-service] ), [e2e-llm-inference-service] marks=[ [e2e-llm-inference-service] pytest.mark.cluster_gpu, [e2e-llm-inference-service] pytest.mark.cluster_nvidia, [e2e-llm-inference-service] pytest.mark.cluster_nvidia_roce, [e2e-llm-inference-service] ], [e2e-llm-inference-service] ), [e2e-llm-inference-service] pytest.param( [e2e-llm-inference-service] TestCase( [e2e-llm-inference-service] base_refs=[ [e2e-llm-inference-service] "router-no-scheduler", [e2e-llm-inference-service] "workload-single-cpu", [e2e-llm-inference-service] "model-fb-opt-125m", [e2e-llm-inference-service] ], [e2e-llm-inference-service] prompt="What is KServe?", [e2e-llm-inference-service] ), [e2e-llm-inference-service] marks=[ [e2e-llm-inference-service] pytest.mark.cluster_cpu, [e2e-llm-inference-service] pytest.mark.cluster_single_node, [e2e-llm-inference-service] pytest.mark.no_scheduler, [e2e-llm-inference-service] ], [e2e-llm-inference-service] ), [e2e-llm-inference-service] pytest.param( [e2e-llm-inference-service] TestCase( [e2e-llm-inference-service] base_refs=[ [e2e-llm-inference-service] "router-managed", [e2e-llm-inference-service] "workload-simulated-dp-ep-cpu", [e2e-llm-inference-service] "model-fb-opt-125m", [e2e-llm-inference-service] ], [e2e-llm-inference-service] prompt="This test simulates DP+EP that can run on CPU, the idea is to test the LWS-based deployment, " [e2e-llm-inference-service] "but without the resources requirements for DP+EP (GPUs and ROCe/IB).", [e2e-llm-inference-service] ), [e2e-llm-inference-service] marks=[pytest.mark.cluster_cpu, pytest.mark.cluster_multi_node], [e2e-llm-inference-service] ), [e2e-llm-inference-service] # Scheduler config tests [e2e-llm-inference-service] pytest.param( [e2e-llm-inference-service] TestCase( [e2e-llm-inference-service] base_refs=[ [e2e-llm-inference-service] "router-managed", [e2e-llm-inference-service] "scheduler-with-inline-config", [e2e-llm-inference-service] "workload-llmd-simulator", [e2e-llm-inference-service] ], [e2e-llm-inference-service] prompt="KServe is a", [e2e-llm-inference-service] service_name="scheduler-inline-config-test", [e2e-llm-inference-service] ), [e2e-llm-inference-service] marks=[pytest.mark.cluster_cpu, pytest.mark.cluster_single_node], [e2e-llm-inference-service] ), [e2e-llm-inference-service] # Chat completions endpoint coverage [e2e-llm-inference-service] pytest.param( [e2e-llm-inference-service] TestCase( [e2e-llm-inference-service] base_refs=[ [e2e-llm-inference-service] "router-managed", [e2e-llm-inference-service] "workload-llmd-simulator", [e2e-llm-inference-service] "model-qwen2.5-0.5b", [e2e-llm-inference-service] ], [e2e-llm-inference-service] model_name="Qwen/Qwen2.5-0.5B-Instruct", [e2e-llm-inference-service] endpoint="/v1/chat/completions", [e2e-llm-inference-service] prompt="What is KServe?", [e2e-llm-inference-service] payload_formatter=chat_completions_payload, [e2e-llm-inference-service] response_assertion=create_response_assertion(with_field="choices"), [e2e-llm-inference-service] ), [e2e-llm-inference-service] marks=[ [e2e-llm-inference-service] pytest.mark.cluster_cpu, [e2e-llm-inference-service] pytest.mark.cluster_single_node, [e2e-llm-inference-service] pytest.mark.llmd_simulator, [e2e-llm-inference-service] ], [e2e-llm-inference-service] ), [e2e-llm-inference-service] pytest.param( [e2e-llm-inference-service] TestCase( [e2e-llm-inference-service] base_refs=[ [e2e-llm-inference-service] "router-managed", [e2e-llm-inference-service] "scheduler-with-configmap-ref", [e2e-llm-inference-service] "workload-llmd-simulator", [e2e-llm-inference-service] ], [e2e-llm-inference-service] prompt="KServe is a", [e2e-llm-inference-service] service_name="scheduler-configmap-ref-test", [e2e-llm-inference-service] before_test=[create_scheduler_configmap], [e2e-llm-inference-service] after_test=[delete_scheduler_configmap], [e2e-llm-inference-service] ), [e2e-llm-inference-service] marks=[pytest.mark.cluster_cpu, pytest.mark.cluster_single_node], [e2e-llm-inference-service] ), [e2e-llm-inference-service] pytest.param( [e2e-llm-inference-service] TestCase( [e2e-llm-inference-service] base_refs=[ [e2e-llm-inference-service] "router-managed", [e2e-llm-inference-service] "scheduler-with-replicas", [e2e-llm-inference-service] "workload-llmd-simulator", [e2e-llm-inference-service] ], [e2e-llm-inference-service] prompt="KServe is a", [e2e-llm-inference-service] service_name="scheduler-ha-replicas-test", [e2e-llm-inference-service] ), [e2e-llm-inference-service] marks=[pytest.mark.cluster_cpu, pytest.mark.cluster_single_node], [e2e-llm-inference-service] ), [e2e-llm-inference-service] pytest.param( [e2e-llm-inference-service] TestCase( [e2e-llm-inference-service] base_refs=[ [e2e-llm-inference-service] "router-managed", [e2e-llm-inference-service] "scheduler-with-custom-template", [e2e-llm-inference-service] "workload-llmd-simulator", [e2e-llm-inference-service] ], [e2e-llm-inference-service] prompt="KServe is a", [e2e-llm-inference-service] service_name="scheduler-custom-template-test", [e2e-llm-inference-service] ), [e2e-llm-inference-service] marks=[pytest.mark.cluster_cpu, pytest.mark.cluster_single_node], [e2e-llm-inference-service] ), [e2e-llm-inference-service] # Scheduler v0.6 → v0.7 migration tests. [e2e-llm-inference-service] # Deploy v0.6-style configs and verify the controller migrates them [e2e-llm-inference-service] # so the v0.7 scheduler boots successfully. [e2e-llm-inference-service] pytest.param( [e2e-llm-inference-service] TestCase( [e2e-llm-inference-service] base_refs=[ [e2e-llm-inference-service] "router-managed", [e2e-llm-inference-service] "scheduler-v06-pd-config-migration", [e2e-llm-inference-service] "workload-llmd-simulator-pd", [e2e-llm-inference-service] ], [e2e-llm-inference-service] prompt="KServe is a", [e2e-llm-inference-service] service_name="scheduler-v06-pd-migration-test", [e2e-llm-inference-service] response_assertion=assert_200_with_choices, [e2e-llm-inference-service] ), [e2e-llm-inference-service] marks=[ [e2e-llm-inference-service] pytest.mark.cluster_cpu, [e2e-llm-inference-service] pytest.mark.cluster_single_node, [e2e-llm-inference-service] pytest.mark.llmd_simulator, [e2e-llm-inference-service] ], [e2e-llm-inference-service] ), [e2e-llm-inference-service] pytest.param( [e2e-llm-inference-service] TestCase( [e2e-llm-inference-service] base_refs=[ [e2e-llm-inference-service] "router-managed", [e2e-llm-inference-service] "scheduler-v06-nonzero-threshold-migration", [e2e-llm-inference-service] "workload-llmd-simulator-pd", [e2e-llm-inference-service] ], [e2e-llm-inference-service] prompt="KServe is a", [e2e-llm-inference-service] service_name="scheduler-v06-threshold-migration-test", [e2e-llm-inference-service] response_assertion=assert_200_with_choices, [e2e-llm-inference-service] ), [e2e-llm-inference-service] marks=[ [e2e-llm-inference-service] pytest.mark.cluster_cpu, [e2e-llm-inference-service] pytest.mark.cluster_single_node, [e2e-llm-inference-service] pytest.mark.llmd_simulator, [e2e-llm-inference-service] ], [e2e-llm-inference-service] ), [e2e-llm-inference-service] # Precise prefix KV cache routing test [e2e-llm-inference-service] pytest.param( [e2e-llm-inference-service] TestCase( [e2e-llm-inference-service] base_refs=[ [e2e-llm-inference-service] "router-managed", [e2e-llm-inference-service] "scheduler-with-precise-prefix-cache-inline-config", [e2e-llm-inference-service] "workload-llmd-simulator-kvcache", [e2e-llm-inference-service] ], [e2e-llm-inference-service] prompt="KServe is a", [e2e-llm-inference-service] service_name="precise-prefix-cache-test", [e2e-llm-inference-service] ), [e2e-llm-inference-service] marks=[ [e2e-llm-inference-service] pytest.mark.cluster_cpu, [e2e-llm-inference-service] pytest.mark.cluster_single_node, [e2e-llm-inference-service] pytest.mark.llmd_simulator, [e2e-llm-inference-service] ], [e2e-llm-inference-service] ), [e2e-llm-inference-service] # Models endpoint coverage [e2e-llm-inference-service] pytest.param( [e2e-llm-inference-service] TestCase( [e2e-llm-inference-service] base_refs=[ [e2e-llm-inference-service] "router-managed", [e2e-llm-inference-service] "workload-llmd-simulator", [e2e-llm-inference-service] ], [e2e-llm-inference-service] endpoint="/v1/models", [e2e-llm-inference-service] response_assertion=create_response_assertion(with_field="data"), [e2e-llm-inference-service] ), [e2e-llm-inference-service] marks=[ [e2e-llm-inference-service] pytest.mark.cluster_cpu, [e2e-llm-inference-service] pytest.mark.cluster_single_node, [e2e-llm-inference-service] pytest.mark.llmd_simulator, [e2e-llm-inference-service] ], [e2e-llm-inference-service] ), [e2e-llm-inference-service] # Model-based routing via X-Gateway-Model-Name header — /v1/completions [e2e-llm-inference-service] pytest.param( [e2e-llm-inference-service] TestCase( [e2e-llm-inference-service] base_refs=[ [e2e-llm-inference-service] "router-managed", [e2e-llm-inference-service] "workload-llmd-simulator", [e2e-llm-inference-service] ], [e2e-llm-inference-service] endpoint="/v1/completions", [e2e-llm-inference-service] prompt="KServe is a", [e2e-llm-inference-service] payload_formatter=completions_payload, [e2e-llm-inference-service] response_assertion=assert_model_field_matches("facebook/opt-125m"), [e2e-llm-inference-service] url_getter=get_model_routing_url, [e2e-llm-inference-service] extra_headers={ [e2e-llm-inference-service] MODEL_ROUTING_HEADER: f"publishers/{KSERVE_TEST_NAMESPACE}/models/facebook/opt-125m", [e2e-llm-inference-service] }, [e2e-llm-inference-service] peers=[ [e2e-llm-inference-service] TestCase( [e2e-llm-inference-service] base_refs=[ [e2e-llm-inference-service] "router-managed", [e2e-llm-inference-service] "workload-llmd-simulator", [e2e-llm-inference-service] "model-qwen2.5-0.5b", [e2e-llm-inference-service] ], [e2e-llm-inference-service] endpoint="/v1/completions", [e2e-llm-inference-service] prompt="KServe is a", [e2e-llm-inference-service] payload_formatter=completions_payload, [e2e-llm-inference-service] response_assertion=assert_model_field_matches( [e2e-llm-inference-service] "Qwen/Qwen2.5-0.5B-Instruct" [e2e-llm-inference-service] ), [e2e-llm-inference-service] url_getter=get_model_routing_url, [e2e-llm-inference-service] extra_headers={ [e2e-llm-inference-service] MODEL_ROUTING_HEADER: f"publishers/{KSERVE_TEST_NAMESPACE}/models/Qwen/Qwen2.5-0.5B-Instruct", [e2e-llm-inference-service] }, [e2e-llm-inference-service] ), [e2e-llm-inference-service] ], [e2e-llm-inference-service] ), [e2e-llm-inference-service] marks=[ [e2e-llm-inference-service] pytest.mark.cluster_cpu, [e2e-llm-inference-service] pytest.mark.cluster_single_node, [e2e-llm-inference-service] pytest.mark.llmd_simulator, [e2e-llm-inference-service] pytest.mark.model_routing, [e2e-llm-inference-service] ], [e2e-llm-inference-service] ), [e2e-llm-inference-service] # Model-based routing via X-Gateway-Model-Name header — /v1/chat/completions [e2e-llm-inference-service] pytest.param( [e2e-llm-inference-service] TestCase( [e2e-llm-inference-service] base_refs=[ [e2e-llm-inference-service] "router-managed", [e2e-llm-inference-service] "workload-llmd-simulator", [e2e-llm-inference-service] ], [e2e-llm-inference-service] endpoint="/v1/chat/completions", [e2e-llm-inference-service] prompt="What is KServe?", [e2e-llm-inference-service] payload_formatter=chat_completions_payload, [e2e-llm-inference-service] response_assertion=assert_model_field_matches("facebook/opt-125m"), [e2e-llm-inference-service] url_getter=get_model_routing_url, [e2e-llm-inference-service] extra_headers={ [e2e-llm-inference-service] MODEL_ROUTING_HEADER: f"publishers/{KSERVE_TEST_NAMESPACE}/models/facebook/opt-125m", [e2e-llm-inference-service] }, [e2e-llm-inference-service] peers=[ [e2e-llm-inference-service] TestCase( [e2e-llm-inference-service] base_refs=[ [e2e-llm-inference-service] "router-managed", [e2e-llm-inference-service] "workload-llmd-simulator", [e2e-llm-inference-service] "model-qwen2.5-0.5b", [e2e-llm-inference-service] ], [e2e-llm-inference-service] endpoint="/v1/chat/completions", [e2e-llm-inference-service] prompt="What is KServe?", [e2e-llm-inference-service] payload_formatter=chat_completions_payload, [e2e-llm-inference-service] response_assertion=assert_model_field_matches( [e2e-llm-inference-service] "Qwen/Qwen2.5-0.5B-Instruct" [e2e-llm-inference-service] ), [e2e-llm-inference-service] url_getter=get_model_routing_url, [e2e-llm-inference-service] extra_headers={ [e2e-llm-inference-service] MODEL_ROUTING_HEADER: f"publishers/{KSERVE_TEST_NAMESPACE}/models/Qwen/Qwen2.5-0.5B-Instruct", [e2e-llm-inference-service] }, [e2e-llm-inference-service] ), [e2e-llm-inference-service] ], [e2e-llm-inference-service] ), [e2e-llm-inference-service] marks=[ [e2e-llm-inference-service] pytest.mark.cluster_cpu, [e2e-llm-inference-service] pytest.mark.cluster_single_node, [e2e-llm-inference-service] pytest.mark.llmd_simulator, [e2e-llm-inference-service] pytest.mark.model_routing, [e2e-llm-inference-service] ], [e2e-llm-inference-service] ), [e2e-llm-inference-service] # Model-based routing via X-Gateway-Model-Name header — LoRA adapter [e2e-llm-inference-service] pytest.param( [e2e-llm-inference-service] TestCase( [e2e-llm-inference-service] base_refs=[ [e2e-llm-inference-service] "router-managed", [e2e-llm-inference-service] "workload-single-cpu", [e2e-llm-inference-service] "model-fb-opt-125m-with-lora-hf", [e2e-llm-inference-service] ], [e2e-llm-inference-service] endpoint="/v1/completions", [e2e-llm-inference-service] prompt="KServe is a", [e2e-llm-inference-service] model_name=f"publishers/{KSERVE_TEST_NAMESPACE}/models/lora-adapter-1", [e2e-llm-inference-service] payload_formatter=completions_payload, [e2e-llm-inference-service] response_assertion=assert_model_field_matches( [e2e-llm-inference-service] f"publishers/{KSERVE_TEST_NAMESPACE}/models/lora-adapter-1" [e2e-llm-inference-service] ), [e2e-llm-inference-service] url_getter=get_model_routing_url, [e2e-llm-inference-service] extra_headers={ [e2e-llm-inference-service] MODEL_ROUTING_HEADER: f"publishers/{KSERVE_TEST_NAMESPACE}/models/lora-adapter-1", [e2e-llm-inference-service] }, [e2e-llm-inference-service] ), [e2e-llm-inference-service] marks=[ [e2e-llm-inference-service] pytest.mark.cluster_cpu, [e2e-llm-inference-service] pytest.mark.cluster_single_node, [e2e-llm-inference-service] pytest.mark.model_routing, [e2e-llm-inference-service] pytest.mark.lora, [e2e-llm-inference-service] ], [e2e-llm-inference-service] ), [e2e-llm-inference-service] # Model-based routing via X-Gateway-Model-Name header — /v1/models (base + LoRA) [e2e-llm-inference-service] pytest.param( [e2e-llm-inference-service] TestCase( [e2e-llm-inference-service] base_refs=[ [e2e-llm-inference-service] "router-managed", [e2e-llm-inference-service] "workload-single-cpu", [e2e-llm-inference-service] "model-fb-opt-125m-with-lora-hf", [e2e-llm-inference-service] ], [e2e-llm-inference-service] endpoint="/v1/models", [e2e-llm-inference-service] response_assertion=assert_models_contains( [e2e-llm-inference-service] "facebook/opt-125m", [e2e-llm-inference-service] f"publishers/{KSERVE_TEST_NAMESPACE}/models/facebook/opt-125m", [e2e-llm-inference-service] "lora-adapter-1", [e2e-llm-inference-service] f"publishers/{KSERVE_TEST_NAMESPACE}/models/lora-adapter-1", [e2e-llm-inference-service] ), [e2e-llm-inference-service] url_getter=get_model_routing_url, [e2e-llm-inference-service] extra_headers={ [e2e-llm-inference-service] MODEL_ROUTING_HEADER: f"publishers/{KSERVE_TEST_NAMESPACE}/models/facebook/opt-125m", [e2e-llm-inference-service] }, [e2e-llm-inference-service] ), [e2e-llm-inference-service] marks=[ [e2e-llm-inference-service] pytest.mark.cluster_cpu, [e2e-llm-inference-service] pytest.mark.cluster_single_node, [e2e-llm-inference-service] pytest.mark.model_routing, [e2e-llm-inference-service] pytest.mark.lora, [e2e-llm-inference-service] ], [e2e-llm-inference-service] ), [e2e-llm-inference-service] # PVC storage tests -- validate direct PVC volume mount with real vLLM serving [e2e-llm-inference-service] pytest.param( [e2e-llm-inference-service] TestCase( [e2e-llm-inference-service] base_refs=[ [e2e-llm-inference-service] "router-managed", [e2e-llm-inference-service] "workload-single-cpu", [e2e-llm-inference-service] "model-pvc", [e2e-llm-inference-service] ], [e2e-llm-inference-service] prompt="KServe is a", [e2e-llm-inference-service] response_assertion=assert_200_with_choices, [e2e-llm-inference-service] before_test=[ensure_pvc_with_model], [e2e-llm-inference-service] ), [e2e-llm-inference-service] marks=[ [e2e-llm-inference-service] pytest.mark.cluster_cpu, [e2e-llm-inference-service] pytest.mark.cluster_single_node, [e2e-llm-inference-service] pytest.mark.pvc_storage, [e2e-llm-inference-service] ], [e2e-llm-inference-service] ), [e2e-llm-inference-service] pytest.param( [e2e-llm-inference-service] TestCase( [e2e-llm-inference-service] base_refs=[ [e2e-llm-inference-service] "router-managed", [e2e-llm-inference-service] "workload-pd-cpu", [e2e-llm-inference-service] "model-pvc", [e2e-llm-inference-service] ], [e2e-llm-inference-service] prompt="KServe is a", [e2e-llm-inference-service] response_assertion=assert_200_with_choices, [e2e-llm-inference-service] before_test=[ensure_pvc_with_model], [e2e-llm-inference-service] ), [e2e-llm-inference-service] marks=[ [e2e-llm-inference-service] pytest.mark.cluster_cpu, [e2e-llm-inference-service] pytest.mark.cluster_single_node, [e2e-llm-inference-service] pytest.mark.pvc_storage, [e2e-llm-inference-service] ], [e2e-llm-inference-service] ), [e2e-llm-inference-service] pytest.param( [e2e-llm-inference-service] TestCase( [e2e-llm-inference-service] base_refs=[ [e2e-llm-inference-service] "router-managed", [e2e-llm-inference-service] "workload-simulated-dp-ep-cpu", [e2e-llm-inference-service] "model-pvc", [e2e-llm-inference-service] ], [e2e-llm-inference-service] prompt="KServe is a", [e2e-llm-inference-service] before_test=[ensure_pvc_with_model], [e2e-llm-inference-service] ), [e2e-llm-inference-service] marks=[ [e2e-llm-inference-service] pytest.mark.cluster_cpu, [e2e-llm-inference-service] pytest.mark.cluster_multi_node, [e2e-llm-inference-service] pytest.mark.pvc_storage, [e2e-llm-inference-service] ], [e2e-llm-inference-service] ), [e2e-llm-inference-service] ], [e2e-llm-inference-service] indirect=["test_case"], [e2e-llm-inference-service] ids=generate_test_id, [e2e-llm-inference-service] ) [e2e-llm-inference-service] @log_execution [e2e-llm-inference-service] def test_llm_inference_service(test_case: TestCase): # noqa: F811 [e2e-llm-inference-service] inject_k8s_proxy() [e2e-llm-inference-service] [e2e-llm-inference-service] kserve_client = KServeClient( [e2e-llm-inference-service] config_file=os.environ.get("KUBECONFIG", "~/.kube/config"), [e2e-llm-inference-service] client_configuration=client.Configuration(), [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] service_name = test_case.llm_service.metadata.name [e2e-llm-inference-service] if not test_case.llm_service.metadata.annotations: [e2e-llm-inference-service] test_case.llm_service.metadata.annotations = {} [e2e-llm-inference-service] [e2e-llm-inference-service] test_case.llm_service.metadata.annotations[ [e2e-llm-inference-service] "security.opendatahub.io/enable-auth" [e2e-llm-inference-service] ] = "false" [e2e-llm-inference-service] prefix = test_case.log_prefix [e2e-llm-inference-service] [e2e-llm-inference-service] test_failed = False [e2e-llm-inference-service] try: [e2e-llm-inference-service] print(f"{prefix} Creating LLMInferenceService {service_name}") [e2e-llm-inference-service] create_llmisvc(kserve_client, test_case.llm_service) [e2e-llm-inference-service] print(f"{prefix} Waiting for LLMInferenceService {service_name} to be ready") [e2e-llm-inference-service] wait_for_llm_isvc_ready( [e2e-llm-inference-service] kserve_client, test_case.llm_service, test_case.wait_timeout [e2e-llm-inference-service] ) [e2e-llm-inference-service] print(f"{prefix} Waiting for model response from {service_name}") [e2e-llm-inference-service] > wait_for_model_response( [e2e-llm-inference-service] kserve_client, [e2e-llm-inference-service] test_case, [e2e-llm-inference-service] test_case.wait_timeout, [e2e-llm-inference-service] extra_headers=test_case.extra_headers, [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] llmisvc/test_llm_inference_service.py:816: [e2e-llm-inference-service] _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ [e2e-llm-inference-service] [e2e-llm-inference-service] args = (, TestCase(base_refs=['router-managed', 'workload-sin... {'name': 'model-fb-opt-125m-with-lora-hf-c0d503b0'}]}, [e2e-llm-inference-service] 'status': None}, model_name='facebook/opt-125m'), 900) [e2e-llm-inference-service] kwargs = {'extra_headers': {'X-Gateway-Model-Name': 'publishers/kserve-ci-e2e-test/models/facebook/opt-125m'}} [e2e-llm-inference-service] func_name = 'wait_for_model_response' [e2e-llm-inference-service] timestamp_start = '2026-07-07T14:47:17.842425', start_time = 1783435637.8427057 [e2e-llm-inference-service] duration = 904.6279535293579, timestamp_end = '2026-07-07T15:02:22.470662' [e2e-llm-inference-service] [e2e-llm-inference-service] @functools.wraps(func) [e2e-llm-inference-service] def wrapper(*args, **kwargs): [e2e-llm-inference-service] func_name = func.__name__ [e2e-llm-inference-service] [e2e-llm-inference-service] timestamp_start = datetime.now().isoformat() [e2e-llm-inference-service] logger.info( [e2e-llm-inference-service] f"[{func_name}] [{timestamp_start}] start - args={args}, kwargs={kwargs}" [e2e-llm-inference-service] ) [e2e-llm-inference-service] start_time = time.time() [e2e-llm-inference-service] [e2e-llm-inference-service] try: [e2e-llm-inference-service] > result = func(*args, **kwargs) [e2e-llm-inference-service] [e2e-llm-inference-service] llmisvc/logging.py:40: [e2e-llm-inference-service] _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ [e2e-llm-inference-service] [e2e-llm-inference-service] kserve_client = [e2e-llm-inference-service] test_case = TestCase(base_refs=['router-managed', 'workload-single-cpu', 'model-fb-opt-125m-with-lora-hf'], prompt=None, service_n... {'name': 'model-fb-opt-125m-with-lora-hf-c0d503b0'}]}, [e2e-llm-inference-service] 'status': None}, model_name='facebook/opt-125m') [e2e-llm-inference-service] timeout_seconds = 900 [e2e-llm-inference-service] extra_headers = {'X-Gateway-Model-Name': 'publishers/kserve-ci-e2e-test/models/facebook/opt-125m'} [e2e-llm-inference-service] [e2e-llm-inference-service] @log_execution [e2e-llm-inference-service] def wait_for_model_response( [e2e-llm-inference-service] kserve_client: KServeClient, [e2e-llm-inference-service] test_case: TestCase, # noqa: F811 [e2e-llm-inference-service] timeout_seconds: int = 900, [e2e-llm-inference-service] extra_headers: Optional[Dict[str, str]] = None, [e2e-llm-inference-service] ) -> str: [e2e-llm-inference-service] def get_successful_response(): [e2e-llm-inference-service] try: [e2e-llm-inference-service] if test_case.url_getter: [e2e-llm-inference-service] service_url = test_case.url_getter(kserve_client, test_case.llm_service) [e2e-llm-inference-service] else: [e2e-llm-inference-service] service_url = get_llm_service_url(kserve_client, test_case.llm_service) [e2e-llm-inference-service] except Exception as e: [e2e-llm-inference-service] raise AssertionError(f"❌ Failed to get service URL: {e}") from e [e2e-llm-inference-service] [e2e-llm-inference-service] model_url = service_url + test_case.endpoint [e2e-llm-inference-service] [e2e-llm-inference-service] headers = {"Content-Type": "application/json"} [e2e-llm-inference-service] if extra_headers: [e2e-llm-inference-service] headers.update(extra_headers) [e2e-llm-inference-service] [e2e-llm-inference-service] if test_case.payload_formatter is not None: [e2e-llm-inference-service] test_payload = test_case.payload_formatter(test_case) [e2e-llm-inference-service] elif test_case.prompt is not None: [e2e-llm-inference-service] test_payload = { [e2e-llm-inference-service] "model": test_case.model_name [e2e-llm-inference-service] if not extra_headers or MODEL_ROUTING_HEADER not in extra_headers [e2e-llm-inference-service] else extra_headers[MODEL_ROUTING_HEADER], [e2e-llm-inference-service] "prompt": test_case.prompt, [e2e-llm-inference-service] "max_tokens": test_case.max_tokens, [e2e-llm-inference-service] } [e2e-llm-inference-service] else: [e2e-llm-inference-service] test_payload = None [e2e-llm-inference-service] [e2e-llm-inference-service] logger.info(f"Calling LLM service at {model_url} with payload {test_payload}") [e2e-llm-inference-service] try: [e2e-llm-inference-service] if test_payload is not None: [e2e-llm-inference-service] response = post_with_retry( [e2e-llm-inference-service] model_url, [e2e-llm-inference-service] headers=headers, [e2e-llm-inference-service] json_data=test_payload, [e2e-llm-inference-service] timeout=test_case.response_timeout, [e2e-llm-inference-service] ) [e2e-llm-inference-service] else: [e2e-llm-inference-service] response = get_with_retry( [e2e-llm-inference-service] model_url, [e2e-llm-inference-service] headers=headers, [e2e-llm-inference-service] timeout=test_case.response_timeout, [e2e-llm-inference-service] ) [e2e-llm-inference-service] except Exception as e: [e2e-llm-inference-service] logger.error(f"❌ Failed to call model: {e}") [e2e-llm-inference-service] raise AssertionError(f"❌ Failed to call model: {e}") from e [e2e-llm-inference-service] [e2e-llm-inference-service] logger.info(f"Model response is {response.status_code}: {response.text[:500]}") [e2e-llm-inference-service] [e2e-llm-inference-service] if 200 <= response.status_code < 300: [e2e-llm-inference-service] return response [e2e-llm-inference-service] raise AssertionError( [e2e-llm-inference-service] f"Service returned {response.status_code}: {response.text}" [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] > response = wait_for(get_successful_response, timeout=timeout_seconds, interval=5.0) [e2e-llm-inference-service] [e2e-llm-inference-service] llmisvc/test_llm_inference_service.py:1119: [e2e-llm-inference-service] _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ [e2e-llm-inference-service] [e2e-llm-inference-service] assertion_fn = .get_successful_response at 0x7f1d17b737e0> [e2e-llm-inference-service] timeout = 900, interval = 5.0 [e2e-llm-inference-service] [e2e-llm-inference-service] def wait_for( [e2e-llm-inference-service] assertion_fn: Callable[[], Any], timeout: float = 5.0, interval: float = 0.1 [e2e-llm-inference-service] ) -> Any: [e2e-llm-inference-service] """Wait for the assertion to succeed within timeout.""" [e2e-llm-inference-service] deadline = time.time() + timeout [e2e-llm-inference-service] last_msg = None [e2e-llm-inference-service] while True: [e2e-llm-inference-service] try: [e2e-llm-inference-service] > return assertion_fn() [e2e-llm-inference-service] [e2e-llm-inference-service] llmisvc/test_llm_inference_service.py:1215: [e2e-llm-inference-service] _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ [e2e-llm-inference-service] [e2e-llm-inference-service] def get_successful_response(): [e2e-llm-inference-service] try: [e2e-llm-inference-service] if test_case.url_getter: [e2e-llm-inference-service] service_url = test_case.url_getter(kserve_client, test_case.llm_service) [e2e-llm-inference-service] else: [e2e-llm-inference-service] service_url = get_llm_service_url(kserve_client, test_case.llm_service) [e2e-llm-inference-service] except Exception as e: [e2e-llm-inference-service] raise AssertionError(f"❌ Failed to get service URL: {e}") from e [e2e-llm-inference-service] [e2e-llm-inference-service] model_url = service_url + test_case.endpoint [e2e-llm-inference-service] [e2e-llm-inference-service] headers = {"Content-Type": "application/json"} [e2e-llm-inference-service] if extra_headers: [e2e-llm-inference-service] headers.update(extra_headers) [e2e-llm-inference-service] [e2e-llm-inference-service] if test_case.payload_formatter is not None: [e2e-llm-inference-service] test_payload = test_case.payload_formatter(test_case) [e2e-llm-inference-service] elif test_case.prompt is not None: [e2e-llm-inference-service] test_payload = { [e2e-llm-inference-service] "model": test_case.model_name [e2e-llm-inference-service] if not extra_headers or MODEL_ROUTING_HEADER not in extra_headers [e2e-llm-inference-service] else extra_headers[MODEL_ROUTING_HEADER], [e2e-llm-inference-service] "prompt": test_case.prompt, [e2e-llm-inference-service] "max_tokens": test_case.max_tokens, [e2e-llm-inference-service] } [e2e-llm-inference-service] else: [e2e-llm-inference-service] test_payload = None [e2e-llm-inference-service] [e2e-llm-inference-service] logger.info(f"Calling LLM service at {model_url} with payload {test_payload}") [e2e-llm-inference-service] try: [e2e-llm-inference-service] if test_payload is not None: [e2e-llm-inference-service] response = post_with_retry( [e2e-llm-inference-service] model_url, [e2e-llm-inference-service] headers=headers, [e2e-llm-inference-service] json_data=test_payload, [e2e-llm-inference-service] timeout=test_case.response_timeout, [e2e-llm-inference-service] ) [e2e-llm-inference-service] else: [e2e-llm-inference-service] response = get_with_retry( [e2e-llm-inference-service] model_url, [e2e-llm-inference-service] headers=headers, [e2e-llm-inference-service] timeout=test_case.response_timeout, [e2e-llm-inference-service] ) [e2e-llm-inference-service] except Exception as e: [e2e-llm-inference-service] logger.error(f"❌ Failed to call model: {e}") [e2e-llm-inference-service] > raise AssertionError(f"❌ Failed to call model: {e}") from e [e2e-llm-inference-service] E AssertionError: ❌ Failed to call model: HTTPConnectionPool(host='ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com', port=80): Max retries exceeded with url: /v1/models (Caused by ReadTimeoutError("HTTPConnectionPool(host='ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com', port=80): Read timed out. (read timeout=60)")) [e2e-llm-inference-service] [e2e-llm-inference-service] llmisvc/test_llm_inference_service.py:1109: AssertionError [e2e-llm-inference-service] ------------------------------ Captured log setup ------------------------------ [e2e-llm-inference-service] INFO e2e.llmisvc.logging:fixtures.py:1637 No HTTP proxy configured for k8s client [e2e-llm-inference-service] INFO e2e.llmisvc.logging:fixtures.py:1586 Checking LLMInferenceServiceConfig router-managed-llmisvc-model-fb-66e80b02 in namespace kserve-ci-e2e-test [e2e-llm-inference-service] INFO e2e.llmisvc.logging:fixtures.py:1612 Resource not found, creating LLMInferenceServiceConfig router-managed-llmisvc-model-fb-66e80b02 [e2e-llm-inference-service] INFO e2e.llmisvc.logging:fixtures.py:1622 ✓ Successfully created LLMInferenceServiceConfig router-managed-llmisvc-model-fb-66e80b02 [e2e-llm-inference-service] INFO e2e.llmisvc.logging:fixtures.py:1586 Checking LLMInferenceServiceConfig workload-single-cpu-llmisvc-mod-1ae2b31a in namespace kserve-ci-e2e-test [e2e-llm-inference-service] INFO e2e.llmisvc.logging:fixtures.py:1612 Resource not found, creating LLMInferenceServiceConfig workload-single-cpu-llmisvc-mod-1ae2b31a [e2e-llm-inference-service] INFO e2e.llmisvc.logging:fixtures.py:1622 ✓ Successfully created LLMInferenceServiceConfig workload-single-cpu-llmisvc-mod-1ae2b31a [e2e-llm-inference-service] INFO e2e.llmisvc.logging:fixtures.py:1586 Checking LLMInferenceServiceConfig model-fb-opt-125m-with-lora-hf-c0d503b0 in namespace kserve-ci-e2e-test [e2e-llm-inference-service] INFO e2e.llmisvc.logging:fixtures.py:1612 Resource not found, creating LLMInferenceServiceConfig model-fb-opt-125m-with-lora-hf-c0d503b0 [e2e-llm-inference-service] INFO e2e.llmisvc.logging:fixtures.py:1622 ✓ Successfully created LLMInferenceServiceConfig model-fb-opt-125m-with-lora-hf-c0d503b0 [e2e-llm-inference-service] ------------------------------ Captured log call ------------------------------- [e2e-llm-inference-service] INFO e2e.llmisvc.logging:logging.py:34 [test_llm_inference_service] [2026-07-07T14:45:53.392221] start - args=(), kwargs={'test_case': TestCase(base_refs=['router-managed', 'workload-single-cpu', 'model-fb-opt-125m-with-lora-hf'], prompt=None, service_name='llmisvc-model-fb-opt-125m-with-ba4d693a', endpoint='/v1/models', max_tokens=20, payload_formatter=None, response_assertion=.response_assertion at 0x7f1d1813eb60>, wait_timeout=900, response_timeout=60, extra_headers={'X-Gateway-Model-Name': 'publishers/kserve-ci-e2e-test/models/facebook/opt-125m'}, url_getter=, expected_gateway=None, before_test=[], after_test=[], peers=[], llm_service={'api_version': 'serving.kserve.io/v1alpha1', [e2e-llm-inference-service] 'kind': 'LLMInferenceService', [e2e-llm-inference-service] 'metadata': {'annotations': None, [e2e-llm-inference-service] 'creation_timestamp': None, [e2e-llm-inference-service] 'deletion_grace_period_seconds': None, [e2e-llm-inference-service] 'deletion_timestamp': None, [e2e-llm-inference-service] 'finalizers': None, [e2e-llm-inference-service] 'generate_name': None, [e2e-llm-inference-service] 'generation': None, [e2e-llm-inference-service] 'labels': None, [e2e-llm-inference-service] 'managed_fields': None, [e2e-llm-inference-service] 'name': 'llmisvc-model-fb-opt-125m-with-ba4d693a', [e2e-llm-inference-service] 'namespace': 'kserve-ci-e2e-test', [e2e-llm-inference-service] 'owner_references': None, [e2e-llm-inference-service] 'resource_version': None, [e2e-llm-inference-service] 'self_link': None, [e2e-llm-inference-service] 'uid': None}, [e2e-llm-inference-service] 'spec': {'baseRefs': [{'name': 'router-managed-llmisvc-model-fb-66e80b02'}, [e2e-llm-inference-service] {'name': 'workload-single-cpu-llmisvc-mod-1ae2b31a'}, [e2e-llm-inference-service] {'name': 'model-fb-opt-125m-with-lora-hf-c0d503b0'}]}, [e2e-llm-inference-service] 'status': None}, model_name='facebook/opt-125m')} [e2e-llm-inference-service] INFO e2e.llmisvc.logging:fixtures.py:1637 No HTTP proxy configured for k8s client [e2e-llm-inference-service] INFO e2e.llmisvc.logging:logging.py:34 [create_llmisvc] [2026-07-07T14:45:53.406866] start - args=(, {'api_version': 'serving.kserve.io/v1alpha1', [e2e-llm-inference-service] 'kind': 'LLMInferenceService', [e2e-llm-inference-service] 'metadata': {'annotations': {'security.opendatahub.io/enable-auth': 'false'}, [e2e-llm-inference-service] 'creation_timestamp': None, [e2e-llm-inference-service] 'deletion_grace_period_seconds': None, [e2e-llm-inference-service] 'deletion_timestamp': None, [e2e-llm-inference-service] 'finalizers': None, [e2e-llm-inference-service] 'generate_name': None, [e2e-llm-inference-service] 'generation': None, [e2e-llm-inference-service] 'labels': None, [e2e-llm-inference-service] 'managed_fields': None, [e2e-llm-inference-service] 'name': 'llmisvc-model-fb-opt-125m-with-ba4d693a', [e2e-llm-inference-service] 'namespace': 'kserve-ci-e2e-test', [e2e-llm-inference-service] 'owner_references': None, [e2e-llm-inference-service] 'resource_version': None, [e2e-llm-inference-service] 'self_link': None, [e2e-llm-inference-service] 'uid': None}, [e2e-llm-inference-service] 'spec': {'baseRefs': [{'name': 'router-managed-llmisvc-model-fb-66e80b02'}, [e2e-llm-inference-service] {'name': 'workload-single-cpu-llmisvc-mod-1ae2b31a'}, [e2e-llm-inference-service] {'name': 'model-fb-opt-125m-with-lora-hf-c0d503b0'}]}, [e2e-llm-inference-service] 'status': None}), kwargs={} [e2e-llm-inference-service] INFO e2e.llmisvc.logging:logging.py:43 [create_llmisvc] [2026-07-07T14:45:53.518722] end - ✅ in 0.112s [e2e-llm-inference-service] INFO e2e.llmisvc.logging:logging.py:34 [wait_for_llm_isvc_ready] [2026-07-07T14:45:53.518890] start - args=(, {'api_version': 'serving.kserve.io/v1alpha1', [e2e-llm-inference-service] 'kind': 'LLMInferenceService', [e2e-llm-inference-service] 'metadata': {'annotations': {'security.opendatahub.io/enable-auth': 'false'}, [e2e-llm-inference-service] 'creation_timestamp': None, [e2e-llm-inference-service] 'deletion_grace_period_seconds': None, [e2e-llm-inference-service] 'deletion_timestamp': None, [e2e-llm-inference-service] 'finalizers': None, [e2e-llm-inference-service] 'generate_name': None, [e2e-llm-inference-service] 'generation': None, [e2e-llm-inference-service] 'labels': None, [e2e-llm-inference-service] 'managed_fields': None, [e2e-llm-inference-service] 'name': 'llmisvc-model-fb-opt-125m-with-ba4d693a', [e2e-llm-inference-service] 'namespace': 'kserve-ci-e2e-test', [e2e-llm-inference-service] 'owner_references': None, [e2e-llm-inference-service] 'resource_version': None, [e2e-llm-inference-service] 'self_link': None, [e2e-llm-inference-service] 'uid': None}, [e2e-llm-inference-service] 'spec': {'baseRefs': [{'name': 'router-managed-llmisvc-model-fb-66e80b02'}, [e2e-llm-inference-service] {'name': 'workload-single-cpu-llmisvc-mod-1ae2b31a'}, [e2e-llm-inference-service] {'name': 'model-fb-opt-125m-with-lora-hf-c0d503b0'}]}, [e2e-llm-inference-service] 'status': None}, 900), kwargs={} [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1222 Waiting: No conditions found in status [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1222 Waiting: Missing true conditions: {'RouterReady', 'Ready', 'WorkloadsReady'}, expected {'RouterReady', 'Ready', 'WorkloadsReady'}, got [{'lastTransitionTime': '2026-07-07T14:45:59Z', 'message': 'The following HTTPRoutes are not ready: [kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-with-ba4d693a-kserve-route: "False" (reason "InvalidKind", message "referencing unsupported backendRef: group \\"inference.networking.x-k8s.io\\" kind \\"InferencePool\\"")]', 'reason': 'HTTPRoutesNotReady', 'severity': 'Info', 'status': 'False', 'type': 'HTTPRoutesReady'}, {'lastTransitionTime': '2026-07-07T14:45:59Z', 'message': 'Inference Pool kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-with-ba4d693a-inference-pool exists but no Gateway controller has accepted it yet', 'reason': 'WaitingForGateway', 'severity': 'Info', 'status': 'False', 'type': 'InferencePoolReady'}, {'lastTransitionTime': '2026-07-07T14:45:59Z', 'reason': 'Progressing', 'severity': 'Info', 'status': 'False', 'type': 'MainWorkloadReady'}, {'lastTransitionTime': '2026-07-07T14:45:59Z', 'severity': 'Info', 'status': 'True', 'type': 'PresetsCombined'}, {'lastTransitionTime': '2026-07-07T14:45:59Z', 'message': 'The following HTTPRoutes are not ready: [kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-with-ba4d693a-kserve-route: "False" (reason "InvalidKind", message "referencing unsupported backendRef: group \\"inference.networking.x-k8s.io\\" kind \\"InferencePool\\"")]', 'reason': 'HTTPRoutesNotReady', 'status': 'False', 'type': 'Ready'}, {'lastTransitionTime': '2026-07-07T14:45:59Z', 'message': 'The following HTTPRoutes are not ready: [kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-with-ba4d693a-kserve-route: "False" (reason "InvalidKind", message "referencing unsupported backendRef: group \\"inference.networking.x-k8s.io\\" kind \\"InferencePool\\"")]', 'reason': 'HTTPRoutesNotReady', 'status': 'False', 'type': 'RouterReady'}, {'lastTransitionTime': '2026-07-07T14:45:59Z', 'message': 'Deployment rollout in progress', 'reason': 'Progressing', 'severity': 'Info', 'status': 'False', 'type': 'SchedulerWorkloadReady'}, {'lastTransitionTime': '2026-07-07T14:45:59Z', 'reason': 'Progressing', 'status': 'False', 'type': 'WorkloadsReady'}] [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1222 Waiting: Missing true conditions: {'RouterReady', 'Ready', 'WorkloadsReady'}, expected {'RouterReady', 'Ready', 'WorkloadsReady'}, got [{'lastTransitionTime': '2026-07-07T14:46:18Z', 'severity': 'Info', 'status': 'True', 'type': 'HTTPRoutesReady'}, {'lastTransitionTime': '2026-07-07T14:46:18Z', 'severity': 'Info', 'status': 'True', 'type': 'InferencePoolReady'}, {'lastTransitionTime': '2026-07-07T14:46:18Z', 'message': 'Deployment does not have minimum availability.', 'reason': 'MinimumReplicasUnavailable', 'severity': 'Info', 'status': 'False', 'type': 'MainWorkloadReady'}, {'lastTransitionTime': '2026-07-07T14:45:59Z', 'severity': 'Info', 'status': 'True', 'type': 'PresetsCombined'}, {'lastTransitionTime': '2026-07-07T14:46:18Z', 'message': 'Deployment does not have minimum availability.', 'reason': 'MinimumReplicasUnavailable', 'status': 'False', 'type': 'Ready'}, {'lastTransitionTime': '2026-07-07T14:46:18Z', 'message': 'Deployment does not have minimum availability.', 'reason': 'MinimumReplicasUnavailable', 'status': 'False', 'type': 'RouterReady'}, {'lastTransitionTime': '2026-07-07T14:46:18Z', 'message': 'Deployment does not have minimum availability.', 'reason': 'MinimumReplicasUnavailable', 'severity': 'Info', 'status': 'False', 'type': 'SchedulerWorkloadReady'}, {'lastTransitionTime': '2026-07-07T14:46:18Z', 'message': 'Deployment does not have minimum availability.', 'reason': 'MinimumReplicasUnavailable', 'status': 'False', 'type': 'WorkloadsReady'}] [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1222 Waiting: Missing true conditions: {'Ready', 'WorkloadsReady'}, expected {'RouterReady', 'Ready', 'WorkloadsReady'}, got [{'lastTransitionTime': '2026-07-07T14:46:18Z', 'severity': 'Info', 'status': 'True', 'type': 'HTTPRoutesReady'}, {'lastTransitionTime': '2026-07-07T14:46:18Z', 'severity': 'Info', 'status': 'True', 'type': 'InferencePoolReady'}, {'lastTransitionTime': '2026-07-07T14:46:18Z', 'message': 'Deployment does not have minimum availability.', 'reason': 'MinimumReplicasUnavailable', 'severity': 'Info', 'status': 'False', 'type': 'MainWorkloadReady'}, {'lastTransitionTime': '2026-07-07T14:45:59Z', 'severity': 'Info', 'status': 'True', 'type': 'PresetsCombined'}, {'lastTransitionTime': '2026-07-07T14:46:18Z', 'message': 'Deployment does not have minimum availability.', 'reason': 'MinimumReplicasUnavailable', 'status': 'False', 'type': 'Ready'}, {'lastTransitionTime': '2026-07-07T14:46:30Z', 'status': 'True', 'type': 'RouterReady'}, {'lastTransitionTime': '2026-07-07T14:46:30Z', 'severity': 'Info', 'status': 'True', 'type': 'SchedulerWorkloadReady'}, {'lastTransitionTime': '2026-07-07T14:46:18Z', 'message': 'Deployment does not have minimum availability.', 'reason': 'MinimumReplicasUnavailable', 'status': 'False', 'type': 'WorkloadsReady'}] [e2e-llm-inference-service] INFO e2e.llmisvc.logging:logging.py:43 [wait_for_llm_isvc_ready] [2026-07-07T14:47:17.842270] end - ✅ in 84.323s [e2e-llm-inference-service] INFO e2e.llmisvc.logging:logging.py:34 [wait_for_model_response] [2026-07-07T14:47:17.842425] start - args=(, TestCase(base_refs=['router-managed', 'workload-single-cpu', 'model-fb-opt-125m-with-lora-hf'], prompt=None, service_name='llmisvc-model-fb-opt-125m-with-ba4d693a', endpoint='/v1/models', max_tokens=20, payload_formatter=None, response_assertion=.response_assertion at 0x7f1d1813eb60>, wait_timeout=900, response_timeout=60, extra_headers={'X-Gateway-Model-Name': 'publishers/kserve-ci-e2e-test/models/facebook/opt-125m'}, url_getter=, expected_gateway=None, before_test=[], after_test=[], peers=[], llm_service={'api_version': 'serving.kserve.io/v1alpha1', [e2e-llm-inference-service] 'kind': 'LLMInferenceService', [e2e-llm-inference-service] 'metadata': {'annotations': {'security.opendatahub.io/enable-auth': 'false'}, [e2e-llm-inference-service] 'creation_timestamp': None, [e2e-llm-inference-service] 'deletion_grace_period_seconds': None, [e2e-llm-inference-service] 'deletion_timestamp': None, [e2e-llm-inference-service] 'finalizers': None, [e2e-llm-inference-service] 'generate_name': None, [e2e-llm-inference-service] 'generation': None, [e2e-llm-inference-service] 'labels': None, [e2e-llm-inference-service] 'managed_fields': None, [e2e-llm-inference-service] 'name': 'llmisvc-model-fb-opt-125m-with-ba4d693a', [e2e-llm-inference-service] 'namespace': 'kserve-ci-e2e-test', [e2e-llm-inference-service] 'owner_references': None, [e2e-llm-inference-service] 'resource_version': None, [e2e-llm-inference-service] 'self_link': None, [e2e-llm-inference-service] 'uid': None}, [e2e-llm-inference-service] 'spec': {'baseRefs': [{'name': 'router-managed-llmisvc-model-fb-66e80b02'}, [e2e-llm-inference-service] {'name': 'workload-single-cpu-llmisvc-mod-1ae2b31a'}, [e2e-llm-inference-service] {'name': 'model-fb-opt-125m-with-lora-hf-c0d503b0'}]}, [e2e-llm-inference-service] 'status': None}, model_name='facebook/opt-125m'), 900), kwargs={'extra_headers': {'X-Gateway-Model-Name': 'publishers/kserve-ci-e2e-test/models/facebook/opt-125m'}} [e2e-llm-inference-service] INFO e2e.llmisvc.logging:logging.py:34 [get_model_routing_url] [2026-07-07T14:47:17.842722] start - args=(, {'api_version': 'serving.kserve.io/v1alpha1', [e2e-llm-inference-service] 'kind': 'LLMInferenceService', [e2e-llm-inference-service] 'metadata': {'annotations': {'security.opendatahub.io/enable-auth': 'false'}, [e2e-llm-inference-service] 'creation_timestamp': None, [e2e-llm-inference-service] 'deletion_grace_period_seconds': None, [e2e-llm-inference-service] 'deletion_timestamp': None, [e2e-llm-inference-service] 'finalizers': None, [e2e-llm-inference-service] 'generate_name': None, [e2e-llm-inference-service] 'generation': None, [e2e-llm-inference-service] 'labels': None, [e2e-llm-inference-service] 'managed_fields': None, [e2e-llm-inference-service] 'name': 'llmisvc-model-fb-opt-125m-with-ba4d693a', [e2e-llm-inference-service] 'namespace': 'kserve-ci-e2e-test', [e2e-llm-inference-service] 'owner_references': None, [e2e-llm-inference-service] 'resource_version': None, [e2e-llm-inference-service] 'self_link': None, [e2e-llm-inference-service] 'uid': None}, [e2e-llm-inference-service] 'spec': {'baseRefs': [{'name': 'router-managed-llmisvc-model-fb-66e80b02'}, [e2e-llm-inference-service] {'name': 'workload-single-cpu-llmisvc-mod-1ae2b31a'}, [e2e-llm-inference-service] {'name': 'model-fb-opt-125m-with-lora-hf-c0d503b0'}]}, [e2e-llm-inference-service] 'status': None}), kwargs={} [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:181 Found model-routing URL for llmisvc-model-fb-opt-125m-with-ba4d693a: http://ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com/ (name='gateway-external-model-routing', path='/') [e2e-llm-inference-service] INFO e2e.llmisvc.logging:logging.py:43 [get_model_routing_url] [2026-07-07T14:47:17.850472] end - ✅ in 0.008s [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1092 Calling LLM service at http://ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com/v1/models with payload None [e2e-llm-inference-service] WARNING urllib3.connectionpool:connectionpool.py:868 Retrying (Retry(total=7, connect=None, read=None, redirect=None, status=None)) after connection broken by 'ReadTimeoutError("HTTPConnectionPool(host='ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com', port=80): Read timed out. (read timeout=60)")': /v1/models [e2e-llm-inference-service] WARNING urllib3.connectionpool:connectionpool.py:868 Retrying (Retry(total=6, connect=None, read=None, redirect=None, status=None)) after connection broken by 'RemoteDisconnected('Remote end closed connection without response')': /v1/models [e2e-llm-inference-service] WARNING urllib3.connectionpool:connectionpool.py:868 Retrying (Retry(total=5, connect=None, read=None, redirect=None, status=None)) after connection broken by 'ReadTimeoutError("HTTPConnectionPool(host='ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com', port=80): Read timed out. (read timeout=60)")': /v1/models [e2e-llm-inference-service] WARNING urllib3.connectionpool:connectionpool.py:868 Retrying (Retry(total=4, connect=None, read=None, redirect=None, status=None)) after connection broken by 'ReadTimeoutError("HTTPConnectionPool(host='ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com', port=80): Read timed out. (read timeout=60)")': /v1/models [e2e-llm-inference-service] WARNING urllib3.connectionpool:connectionpool.py:868 Retrying (Retry(total=3, connect=None, read=None, redirect=None, status=None)) after connection broken by 'ReadTimeoutError("HTTPConnectionPool(host='ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com', port=80): Read timed out. (read timeout=60)")': /v1/models [e2e-llm-inference-service] WARNING urllib3.connectionpool:connectionpool.py:868 Retrying (Retry(total=2, connect=None, read=None, redirect=None, status=None)) after connection broken by 'ReadTimeoutError("HTTPConnectionPool(host='ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com', port=80): Read timed out. (read timeout=60)")': /v1/models [e2e-llm-inference-service] WARNING urllib3.connectionpool:connectionpool.py:868 Retrying (Retry(total=1, connect=None, read=None, redirect=None, status=None)) after connection broken by 'ReadTimeoutError("HTTPConnectionPool(host='ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com', port=80): Read timed out. (read timeout=60)")': /v1/models [e2e-llm-inference-service] WARNING urllib3.connectionpool:connectionpool.py:868 Retrying (Retry(total=0, connect=None, read=None, redirect=None, status=None)) after connection broken by 'ReadTimeoutError("HTTPConnectionPool(host='ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com', port=80): Read timed out. (read timeout=60)")': /v1/models [e2e-llm-inference-service] ERROR e2e.llmisvc.logging:test_llm_inference_service.py:1108 ❌ Failed to call model: HTTPConnectionPool(host='ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com', port=80): Max retries exceeded with url: /v1/models (Caused by ReadTimeoutError("HTTPConnectionPool(host='ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com', port=80): Read timed out. (read timeout=60)")) [e2e-llm-inference-service] ERROR e2e.llmisvc.logging:test_llm_inference_service.py:1219 Timed out waiting: ❌ Failed to call model: HTTPConnectionPool(host='ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com', port=80): Max retries exceeded with url: /v1/models (Caused by ReadTimeoutError("HTTPConnectionPool(host='ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com', port=80): Read timed out. (read timeout=60)")) [e2e-llm-inference-service] ERROR e2e.llmisvc.logging:logging.py:48 [wait_for_model_response] [2026-07-07T15:02:22.470662] end - ❌ 904.628s: ❌ Failed to call model: HTTPConnectionPool(host='ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com', port=80): Max retries exceeded with url: /v1/models (Caused by ReadTimeoutError("HTTPConnectionPool(host='ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com', port=80): Read timed out. (read timeout=60)")) [e2e-llm-inference-service] ERROR e2e.llmisvc.logging:test_llm_inference_service.py:831 [router-managed-workload-single-cpu-model-fb-opt-125m-with-lora-hf] ❌ ERROR: Failed to call llm inference service llmisvc-model-fb-opt-125m-with-ba4d693a: ❌ Failed to call model: HTTPConnectionPool(host='ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com', port=80): Max retries exceeded with url: /v1/models (Caused by ReadTimeoutError("HTTPConnectionPool(host='ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com', port=80): Read timed out. (read timeout=60)")) [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1240 🔍 # Diagnostics for 'llmisvc-model-fb-opt-125m-with-ba4d693a' in 'kserve-ci-e2e-test' [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1241 --- [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1242 # LLMInferenceService llmisvc-model-fb-opt-125m-with-ba4d693a [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1245 apiVersion: serving.kserve.io/v1alpha1 [e2e-llm-inference-service] kind: LLMInferenceService [e2e-llm-inference-service] metadata: [e2e-llm-inference-service] annotations: [e2e-llm-inference-service] security.opendatahub.io/enable-auth: 'false' [e2e-llm-inference-service] creationTimestamp: '2026-07-07T14:45:53Z' [e2e-llm-inference-service] finalizers: [e2e-llm-inference-service] - serving.kserve.io/llmisvc-finalizer [e2e-llm-inference-service] generation: 1 [e2e-llm-inference-service] managedFields: [e2e-llm-inference-service] - apiVersion: serving.kserve.io/v1alpha1 [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:metadata: [e2e-llm-inference-service] f:annotations: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:security.opendatahub.io/enable-auth: {} [e2e-llm-inference-service] f:spec: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:baseRefs: {} [e2e-llm-inference-service] manager: OpenAPI-Generator [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] time: '2026-07-07T14:45:53Z' [e2e-llm-inference-service] - apiVersion: serving.kserve.io/v1alpha2 [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:metadata: [e2e-llm-inference-service] f:finalizers: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] v:"serving.kserve.io/llmisvc-finalizer": {} [e2e-llm-inference-service] manager: manager [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] time: '2026-07-07T14:45:53Z' [e2e-llm-inference-service] - apiVersion: serving.kserve.io/v1alpha2 [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:status: [e2e-llm-inference-service] f:addresses: {} [e2e-llm-inference-service] f:annotations: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:serving.kserve.io/config-llm-decode-template: {} [e2e-llm-inference-service] f:serving.kserve.io/config-llm-decode-worker-data-parallel: {} [e2e-llm-inference-service] f:serving.kserve.io/config-llm-prefill-template: {} [e2e-llm-inference-service] f:serving.kserve.io/config-llm-prefill-worker-data-parallel: {} [e2e-llm-inference-service] f:serving.kserve.io/config-llm-router-route: {} [e2e-llm-inference-service] f:serving.kserve.io/config-llm-scheduler: {} [e2e-llm-inference-service] f:serving.kserve.io/config-llm-template: {} [e2e-llm-inference-service] f:serving.kserve.io/config-llm-tracing: {} [e2e-llm-inference-service] f:serving.kserve.io/config-llm-worker-data-parallel: {} [e2e-llm-inference-service] f:appliedConfigs: {} [e2e-llm-inference-service] f:conditions: {} [e2e-llm-inference-service] f:observedGeneration: {} [e2e-llm-inference-service] f:router: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:gateways: {} [e2e-llm-inference-service] f:scheduler: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:inferencePool: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:group: {} [e2e-llm-inference-service] f:kind: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:service: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:group: {} [e2e-llm-inference-service] f:kind: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:url: {} [e2e-llm-inference-service] f:workloads: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:primary: {} [e2e-llm-inference-service] f:scheduler: {} [e2e-llm-inference-service] f:service: {} [e2e-llm-inference-service] manager: manager [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] subresource: status [e2e-llm-inference-service] time: '2026-07-07T14:47:17Z' [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-ba4d693a [e2e-llm-inference-service] namespace: kserve-ci-e2e-test [e2e-llm-inference-service] resourceVersion: '47375' [e2e-llm-inference-service] uid: 997d3cef-fabf-4b25-ad95-5ebe7a367fc5 [e2e-llm-inference-service] spec: [e2e-llm-inference-service] baseRefs: [e2e-llm-inference-service] - name: router-managed-llmisvc-model-fb-66e80b02 [e2e-llm-inference-service] - name: workload-single-cpu-llmisvc-mod-1ae2b31a [e2e-llm-inference-service] - name: model-fb-opt-125m-with-lora-hf-c0d503b0 [e2e-llm-inference-service] model: [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-ba4d693a [e2e-llm-inference-service] uri: '' [e2e-llm-inference-service] status: [e2e-llm-inference-service] addresses: [e2e-llm-inference-service] - name: gateway-external-model-routing [e2e-llm-inference-service] url: http://ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com/ [e2e-llm-inference-service] - name: gateway-external [e2e-llm-inference-service] url: http://ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com/kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-with-ba4d693a [e2e-llm-inference-service] - name: gateway-internal-model-routing [e2e-llm-inference-service] url: http://openshift-ai-inference-openshift-default.openshift-ingress.svc.cluster.local/ [e2e-llm-inference-service] - name: gateway-internal [e2e-llm-inference-service] url: http://openshift-ai-inference-openshift-default.openshift-ingress.svc.cluster.local/kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-with-ba4d693a [e2e-llm-inference-service] annotations: [e2e-llm-inference-service] serving.kserve.io/config-llm-decode-template: kserve-config-llm-decode-template [e2e-llm-inference-service] serving.kserve.io/config-llm-decode-worker-data-parallel: kserve-config-llm-decode-worker-data-parallel [e2e-llm-inference-service] serving.kserve.io/config-llm-prefill-template: kserve-config-llm-prefill-template [e2e-llm-inference-service] serving.kserve.io/config-llm-prefill-worker-data-parallel: kserve-config-llm-prefill-worker-data-parallel [e2e-llm-inference-service] serving.kserve.io/config-llm-router-route: kserve-config-llm-router-route [e2e-llm-inference-service] serving.kserve.io/config-llm-scheduler: kserve-config-llm-scheduler [e2e-llm-inference-service] serving.kserve.io/config-llm-template: kserve-config-llm-template [e2e-llm-inference-service] serving.kserve.io/config-llm-tracing: kserve-config-llm-tracing [e2e-llm-inference-service] serving.kserve.io/config-llm-worker-data-parallel: kserve-config-llm-worker-data-parallel [e2e-llm-inference-service] conditions: [e2e-llm-inference-service] - lastTransitionTime: '2026-07-07T14:46:18Z' [e2e-llm-inference-service] severity: Info [e2e-llm-inference-service] status: 'True' [e2e-llm-inference-service] type: HTTPRoutesReady [e2e-llm-inference-service] - lastTransitionTime: '2026-07-07T14:46:18Z' [e2e-llm-inference-service] severity: Info [e2e-llm-inference-service] status: 'True' [e2e-llm-inference-service] type: InferencePoolReady [e2e-llm-inference-service] - lastTransitionTime: '2026-07-07T14:47:17Z' [e2e-llm-inference-service] severity: Info [e2e-llm-inference-service] status: 'True' [e2e-llm-inference-service] type: MainWorkloadReady [e2e-llm-inference-service] - lastTransitionTime: '2026-07-07T14:45:59Z' [e2e-llm-inference-service] severity: Info [e2e-llm-inference-service] status: 'True' [e2e-llm-inference-service] type: PresetsCombined [e2e-llm-inference-service] - lastTransitionTime: '2026-07-07T14:47:17Z' [e2e-llm-inference-service] status: 'True' [e2e-llm-inference-service] type: Ready [e2e-llm-inference-service] - lastTransitionTime: '2026-07-07T14:46:30Z' [e2e-llm-inference-service] status: 'True' [e2e-llm-inference-service] type: RouterReady [e2e-llm-inference-service] - lastTransitionTime: '2026-07-07T14:46:30Z' [e2e-llm-inference-service] severity: Info [e2e-llm-inference-service] status: 'True' [e2e-llm-inference-service] type: SchedulerWorkloadReady [e2e-llm-inference-service] - lastTransitionTime: '2026-07-07T14:47:17Z' [e2e-llm-inference-service] status: 'True' [e2e-llm-inference-service] type: WorkloadsReady [e2e-llm-inference-service] observedGeneration: 1 [e2e-llm-inference-service] url: http://ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com/kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-with-ba4d693a [e2e-llm-inference-service] [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:44 TIME NAMESPACE SOURCE TYPE REASON MESSAGE [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:45 -------------------------------------------------------------------------------------------------- [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 None kserve-ci-e2e-test Normal Scheduled Successfully assigned kserve-ci-e2e-test/auth-disabled-test-kserve-658fdbccfc-7k7fm to ip-10-0-139-47.ec2.internal [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:28:27 kserve-ci-e2e-test multus Normal AddedInterface Add eth0 [10.134.0.29/23] from ovn-kubernetes [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:28:28 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Pulled Container image "quay.io/opendatahub/kserve-storage-initializer@sha256:13ad765c0f323200e1db7b45c99661287e65c5158d6b54b16139829636d20be4" already present on machine [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:28:28 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Created Created container: storage-initializer [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:28:28 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Started Started container storage-initializer [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:28:31 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Pulled Container image "public.ecr.aws/q9t5s3a7/vllm-cpu-release-repo:v0.19.0" already present on machine [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:28:32 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Created Created container: main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:28:32 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Started Started container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:30:17 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Warning Unhealthy Startup probe failed: Get "https://10.134.0.29:8000/health": dial tcp 10.134.0.29:8000: connect: connection refused [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:30:32 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Killing Stopping container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:30:47 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Warning Unhealthy Readiness probe failed: Get "https://10.134.0.29:8000/health": dial tcp 10.134.0.29:8000: connect: connection refused [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:28:27 kserve-ci-e2e-test replicaset-controller Normal SuccessfulCreate Created pod: auth-disabled-test-kserve-658fdbccfc-7k7fm [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 None kserve-ci-e2e-test Normal Scheduled Successfully assigned kserve-ci-e2e-test/auth-disabled-test-kserve-router-scheduler-58c676b848-mfsnv to ip-10-0-133-25.ec2.internal [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:28:28 kserve-ci-e2e-test multus Normal AddedInterface Add eth0 [10.132.0.51/23] from ovn-kubernetes [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:28:28 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Pulled Container image "quay.io/opendatahub/kserve-storage-initializer@sha256:13ad765c0f323200e1db7b45c99661287e65c5158d6b54b16139829636d20be4" already present on machine [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:28:28 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Created Created container: storage-initializer [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:28:28 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Started Started container storage-initializer [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:28:29 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Pulled Container image "ghcr.io/llm-d/llm-d-router-endpoint-picker:v0.9.0-rc.2" already present on machine [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:28:29 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Created Created container: main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:28:29 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Started Started container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:30:32 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Killing Stopping container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:30:32 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Warning FailedPreStopHook PreStopHook failed [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:28:28 kserve-ci-e2e-test replicaset-controller Normal SuccessfulCreate Created pod: auth-disabled-test-kserve-router-scheduler-58c676b848-mfsnv [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:28:27 kserve-ci-e2e-test deployment-controller Normal ScalingReplicaSet Scaled up replica set auth-disabled-test-kserve-router-scheduler-58c676b848 from 0 to 1 [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:28:27 kserve-ci-e2e-test deployment-controller Normal ScalingReplicaSet Scaled up replica set auth-disabled-test-kserve-658fdbccfc from 0 to 1 [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:28:27 kserve-ci-e2e-test OpenDataHubModelController Warning ReconcileError Failed to reconcile LLMInferenceService: 1 error occurred: * failed to get HTTPRoute for AuthPolicy auth-disabled-test-kserve-route-authn: failed to get HTTPRoute kserve-ci-e2e-test/auth-disabled-test-kserve-route: HTTPRoute.gateway.networking.k8s.io "auth-disabled-test-kserve-route" not found [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:28:27 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Secret kserve-ci-e2e-test/auth-disabled-test-kserve-self-signed-certs [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:28:27 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Deployment kserve-ci-e2e-test/auth-disabled-test-kserve [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:28:27 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Service kserve-ci-e2e-test/auth-disabled-test-kserve-workload-svc [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:28:27 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.ServiceAccount kserve-ci-e2e-test/auth-disabled-test-epp-sa [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:28:27 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.ClusterRoleBinding /kserve-ci-e2e-test-auth-disabled-test-epp-auth-rb [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:28:27 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Role kserve-ci-e2e-test/auth-disabled-test-epp-role [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:28:27 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.RoleBinding kserve-ci-e2e-test/auth-disabled-test-epp-rb [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:28:27 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Deployment kserve-ci-e2e-test/auth-disabled-test-kserve-router-scheduler [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:28:27 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Service kserve-ci-e2e-test/auth-disabled-test-epp-service [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:28:38 kserve-ci-e2e-test LLMInferenceServiceController Normal Created (combined from similar events): Created v1.DestinationRule kserve-ci-e2e-test/auth-disabled-test-kserve-shadow-svc [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:28:38 kserve-ci-e2e-test LLMInferenceServiceController Normal Updated Updated v1.Secret kserve-ci-e2e-test/auth-disabled-test-kserve-self-signed-certs [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:28:38 kserve-ci-e2e-test LLMInferenceServiceController Normal Updated Updated v1.HTTPRoute kserve-ci-e2e-test/auth-disabled-test-kserve-route [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:30:28 kserve-ci-e2e-test LLMInferenceServiceController Normal LLMInferenceServiceReady LLMInferenceService [auth-disabled-test] is Ready [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:30:32 kserve-ci-e2e-test LLMInferenceServiceController Normal Deleted Deleted v1.ClusterRoleBinding /kserve-ci-e2e-test-auth-disabled-test-epp-auth-rb [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 None kserve-ci-e2e-test Normal Scheduled Successfully assigned kserve-ci-e2e-test/auth-enabled-test-kserve-87bd45bf5-zxvsp to ip-10-0-139-47.ec2.internal [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:33 kserve-ci-e2e-test multus Normal AddedInterface Add eth0 [10.134.0.24/23] from ovn-kubernetes [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:33 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Pulled Container image "quay.io/opendatahub/kserve-storage-initializer@sha256:13ad765c0f323200e1db7b45c99661287e65c5158d6b54b16139829636d20be4" already present on machine [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:33 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Created Created container: storage-initializer [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:33 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Started Started container storage-initializer [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:39 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Pulling Pulling image "public.ecr.aws/q9t5s3a7/vllm-cpu-release-repo:v0.19.0" [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:23:08 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Pulled Successfully pulled image "public.ecr.aws/q9t5s3a7/vllm-cpu-release-repo:v0.19.0" in 28.615s (28.615s including waiting). Image size: 3531177328 bytes. [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:23:08 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Created Created container: main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:23:08 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Started Started container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:24:53 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Warning Unhealthy Startup probe failed: Get "https://10.134.0.24:8000/health": dial tcp 10.134.0.24:8000: connect: connection refused [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:10 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Killing Stopping container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:25 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Warning Unhealthy Readiness probe failed: Get "https://10.134.0.24:8000/health": dial tcp 10.134.0.24:8000: connect: connection refused [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:32 kserve-ci-e2e-test replicaset-controller Normal SuccessfulCreate Created pod: auth-enabled-test-kserve-87bd45bf5-zxvsp [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 None kserve-ci-e2e-test Normal Scheduled Successfully assigned kserve-ci-e2e-test/auth-enabled-test-kserve-router-scheduler-6bbc585b87-nf5lr to ip-10-0-133-25.ec2.internal [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:33 kserve-ci-e2e-test multus Normal AddedInterface Add eth0 [10.132.0.47/23] from ovn-kubernetes [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:33 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Pulled Container image "quay.io/opendatahub/kserve-storage-initializer@sha256:13ad765c0f323200e1db7b45c99661287e65c5158d6b54b16139829636d20be4" already present on machine [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:33 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Created Created container: storage-initializer [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:33 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Started Started container storage-initializer [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:34 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Pulling Pulling image "ghcr.io/llm-d/llm-d-router-endpoint-picker:v0.9.0-rc.2" [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:35 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Pulled Successfully pulled image "ghcr.io/llm-d/llm-d-router-endpoint-picker:v0.9.0-rc.2" in 1.131s (1.131s including waiting). Image size: 75073927 bytes. [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:35 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Created Created container: main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:35 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Started Started container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:10 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Killing Stopping container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:10 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Warning FailedPreStopHook PreStopHook failed [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:33 kserve-ci-e2e-test replicaset-controller Normal SuccessfulCreate Created pod: auth-enabled-test-kserve-router-scheduler-6bbc585b87-nf5lr [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:33 kserve-ci-e2e-test deployment-controller Normal ScalingReplicaSet Scaled up replica set auth-enabled-test-kserve-router-scheduler-6bbc585b87 from 0 to 1 [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:32 kserve-ci-e2e-test deployment-controller Normal ScalingReplicaSet Scaled up replica set auth-enabled-test-kserve-87bd45bf5 from 0 to 1 [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:32 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Secret kserve-ci-e2e-test/auth-enabled-test-kserve-self-signed-certs [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:32 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Deployment kserve-ci-e2e-test/auth-enabled-test-kserve [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:32 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Service kserve-ci-e2e-test/auth-enabled-test-kserve-workload-svc [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:32 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.ServiceAccount kserve-ci-e2e-test/auth-enabled-test-epp-sa [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:33 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.ClusterRoleBinding /kserve-ci-e2e-test-auth-enabled-test-epp-auth-rb [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:33 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Role kserve-ci-e2e-test/auth-enabled-test-epp-role [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:33 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.RoleBinding kserve-ci-e2e-test/auth-enabled-test-epp-rb [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:33 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Deployment kserve-ci-e2e-test/auth-enabled-test-kserve-router-scheduler [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:33 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Service kserve-ci-e2e-test/auth-enabled-test-epp-service [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:43 kserve-ci-e2e-test LLMInferenceServiceController Normal Created (combined from similar events): Created v1.DestinationRule kserve-ci-e2e-test/auth-enabled-test-kserve-shadow-svc [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:42 kserve-ci-e2e-test LLMInferenceServiceController Normal Updated Updated v1.Secret kserve-ci-e2e-test/auth-enabled-test-kserve-self-signed-certs [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:43 kserve-ci-e2e-test LLMInferenceServiceController Normal Updated Updated v1.HTTPRoute kserve-ci-e2e-test/auth-enabled-test-kserve-route [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:05 kserve-ci-e2e-test LLMInferenceServiceController Normal LLMInferenceServiceReady LLMInferenceService [auth-enabled-test] is Ready [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:10 kserve-ci-e2e-test LLMInferenceServiceController Normal Deleted Deleted v1.ClusterRoleBinding /kserve-ci-e2e-test-auth-enabled-test-epp-auth-rb [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 None kserve-ci-e2e-test Normal Scheduled Successfully assigned kserve-ci-e2e-test/auth-invalid-token-test-kserve-6467b8fbf9-jhlvl to ip-10-0-139-47.ec2.internal [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:44 kserve-ci-e2e-test multus Normal AddedInterface Add eth0 [10.134.0.27/23] from ovn-kubernetes [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:44 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Pulled Container image "quay.io/opendatahub/kserve-storage-initializer@sha256:13ad765c0f323200e1db7b45c99661287e65c5158d6b54b16139829636d20be4" already present on machine [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:44 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Created Created container: storage-initializer [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:44 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Started Started container storage-initializer [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:49 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Pulled Container image "public.ecr.aws/q9t5s3a7/vllm-cpu-release-repo:v0.19.0" already present on machine [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:49 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Created Created container: main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:49 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Started Started container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:34 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Warning Unhealthy Startup probe failed: Get "https://10.134.0.27:8000/health": dial tcp 10.134.0.27:8000: connect: connection refused [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:46 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Killing Stopping container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:28:03 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Warning Unhealthy Readiness probe failed: Get "https://10.134.0.27:8000/health": context deadline exceeded [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:43 kserve-ci-e2e-test replicaset-controller Normal SuccessfulCreate Created pod: auth-invalid-token-test-kserve-6467b8fbf9-jhlvl [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:43 kserve-ci-e2e-test replicaset-controller Normal SuccessfulCreate Created pod: auth-invalid-token-test-kserve-router-scheduler-7c68b57cd5h6v2p [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 None kserve-ci-e2e-test Normal Scheduled Successfully assigned kserve-ci-e2e-test/auth-invalid-token-test-kserve-router-scheduler-7c68b57cd5h6v2p to ip-10-0-133-25.ec2.internal [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:44 kserve-ci-e2e-test multus Normal AddedInterface Add eth0 [10.132.0.49/23] from ovn-kubernetes [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:44 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Pulled Container image "quay.io/opendatahub/kserve-storage-initializer@sha256:13ad765c0f323200e1db7b45c99661287e65c5158d6b54b16139829636d20be4" already present on machine [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:44 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Created Created container: storage-initializer [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:44 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Started Started container storage-initializer [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:45 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Pulled Container image "ghcr.io/llm-d/llm-d-router-endpoint-picker:v0.9.0-rc.2" already present on machine [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:45 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Created Created container: main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:45 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Started Started container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:46 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Killing Stopping container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:46 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Warning FailedPreStopHook PreStopHook failed [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:43 kserve-ci-e2e-test deployment-controller Normal ScalingReplicaSet Scaled up replica set auth-invalid-token-test-kserve-router-scheduler-7c68b57cd5 from 0 to 1 [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:43 kserve-ci-e2e-test deployment-controller Normal ScalingReplicaSet Scaled up replica set auth-invalid-token-test-kserve-6467b8fbf9 from 0 to 1 [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:43 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Secret kserve-ci-e2e-test/auth-invalid-token-test-kserve-self-signed-certs [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:43 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Deployment kserve-ci-e2e-test/auth-invalid-token-test-kserve [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:43 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Service kserve-ci-e2e-test/auth-invalid-token-test-kserve-workload-svc [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:43 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.ServiceAccount kserve-ci-e2e-test/auth-invalid-token-test-epp-sa [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:43 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.ClusterRoleBinding /kserve-ci-e2e-test-auth-invalid-token-test-epp-auth-rb [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:43 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Role kserve-ci-e2e-test/auth-invalid-token-test-epp-role [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:43 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.RoleBinding kserve-ci-e2e-test/auth-invalid-token-test-epp-rb [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:43 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Deployment kserve-ci-e2e-test/auth-invalid-token-test-kserve-router-scheduler [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:43 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Service kserve-ci-e2e-test/auth-invalid-token-test-epp-service [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:26:10 kserve-ci-e2e-test LLMInferenceServiceController Normal Created (combined from similar events): Created v1.DestinationRule kserve-ci-e2e-test/auth-invalid-token-test-kserve-shadow-svc [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:26:09 kserve-ci-e2e-test LLMInferenceServiceController Normal Updated Updated v1.Secret kserve-ci-e2e-test/auth-invalid-token-test-kserve-self-signed-certs [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:26:09 kserve-ci-e2e-test LLMInferenceServiceController Normal Updated Updated v1.HTTPRoute kserve-ci-e2e-test/auth-invalid-token-test-kserve-route [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:45 kserve-ci-e2e-test LLMInferenceServiceController Normal LLMInferenceServiceReady LLMInferenceService [auth-invalid-token-test] is Ready [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:46 kserve-ci-e2e-test LLMInferenceServiceController Normal Deleted Deleted v1.ClusterRoleBinding /kserve-ci-e2e-test-auth-invalid-token-test-epp-auth-rb [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 None kserve-ci-e2e-test Normal Scheduled Successfully assigned kserve-ci-e2e-test/custom-route-timeout-test-kserve-7f56b8c469-hkpbl to ip-10-0-139-47.ec2.internal [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:48:57 kserve-ci-e2e-test multus Normal AddedInterface Add eth0 [10.134.0.33/23] from ovn-kubernetes [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:48:57 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Pulled Container image "quay.io/opendatahub/kserve-storage-initializer@sha256:13ad765c0f323200e1db7b45c99661287e65c5158d6b54b16139829636d20be4" already present on machine [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:48:57 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Created Created container: storage-initializer [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:48:57 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Started Started container storage-initializer [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:49:02 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Pulled Container image "public.ecr.aws/q9t5s3a7/vllm-cpu-release-repo:v0.19.0" already present on machine [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:49:02 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Created Created container: main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:49:02 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Started Started container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:50:47 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Warning Unhealthy Startup probe failed: Get "https://10.134.0.33:8000/health": dial tcp 10.134.0.33:8000: connect: connection refused [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:48:57 kserve-ci-e2e-test replicaset-controller Normal SuccessfulCreate Created pod: custom-route-timeout-test-kserve-7f56b8c469-hkpbl [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:48:57 kserve-ci-e2e-test replicaset-controller Normal SuccessfulCreate Created pod: custom-route-timeout-test-kserve-router-scheduler-6578bddfnbxnb [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 None kserve-ci-e2e-test Normal Scheduled Successfully assigned kserve-ci-e2e-test/custom-route-timeout-test-kserve-router-scheduler-6578bddfnbxnb to ip-10-0-133-25.ec2.internal [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:48:58 kserve-ci-e2e-test multus Normal AddedInterface Add eth0 [10.132.0.55/23] from ovn-kubernetes [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:48:58 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Pulled Container image "quay.io/opendatahub/kserve-storage-initializer@sha256:13ad765c0f323200e1db7b45c99661287e65c5158d6b54b16139829636d20be4" already present on machine [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:48:58 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Created Created container: storage-initializer [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:48:58 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Started Started container storage-initializer [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:48:59 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Pulled Container image "ghcr.io/llm-d/llm-d-router-endpoint-picker:v0.9.0-rc.2" already present on machine [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:48:59 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Created Created container: main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:48:59 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Started Started container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:48:57 kserve-ci-e2e-test deployment-controller Normal ScalingReplicaSet Scaled up replica set custom-route-timeout-test-kserve-router-scheduler-6578bddf5b from 0 to 1 [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:48:57 kserve-ci-e2e-test deployment-controller Normal ScalingReplicaSet Scaled up replica set custom-route-timeout-test-kserve-7f56b8c469 from 0 to 1 [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:48:53 kserve-ci-e2e-test OpenDataHubModelController Warning ReconcileError Failed to reconcile LLMInferenceService: 1 error occurred: * failed to get HTTPRoute for AuthPolicy custom-route-timeout-test-kserve-route-authn: failed to get HTTPRoute kserve-ci-e2e-test/custom-route-timeout-test-kserve-route: HTTPRoute.gateway.networking.k8s.io "custom-route-timeout-test-kserve-route" not found [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:48:57 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Secret kserve-ci-e2e-test/custom-route-timeout-test-kserve-self-signed-certs [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:48:57 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Deployment kserve-ci-e2e-test/custom-route-timeout-test-kserve [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:48:57 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Service kserve-ci-e2e-test/custom-route-timeout-test-kserve-workload-svc [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:48:57 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.ServiceAccount kserve-ci-e2e-test/custom-route-timeout-test-epp-sa [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:48:57 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.ClusterRoleBinding /kserve-ci-e2e-test-custom-route-timeout-test-epp-auth-rb [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:48:57 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Role kserve-ci-e2e-test/custom-route-timeout-test-epp-role [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:48:57 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.RoleBinding kserve-ci-e2e-test/custom-route-timeout-test-epp-rb [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:48:57 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Deployment kserve-ci-e2e-test/custom-route-timeout-test-kserve-router-scheduler [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:48:57 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Service kserve-ci-e2e-test/custom-route-timeout-test-epp-service [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:49:12 kserve-ci-e2e-test LLMInferenceServiceController Normal Created (combined from similar events): Created v1.DestinationRule kserve-ci-e2e-test/custom-route-timeout-test-kserve-shadow-svc [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:49:11 kserve-ci-e2e-test LLMInferenceServiceController Normal Updated Updated v1.Secret kserve-ci-e2e-test/custom-route-timeout-test-kserve-self-signed-certs [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:49:11 kserve-ci-e2e-test LLMInferenceServiceController Normal Updated Updated v1.HTTPRoute kserve-ci-e2e-test/custom-route-timeout-test-kserve-route [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:50:58 kserve-ci-e2e-test LLMInferenceServiceController Normal LLMInferenceServiceReady LLMInferenceService [custom-route-timeout-test] is Ready [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:45 kserve-ci-e2e-test replicaset-controller Normal SuccessfulCreate Created pod: gw-section-name-router-with-gat-2f0a622e-kserve-7c9c8cffc558295 [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 None kserve-ci-e2e-test Normal Scheduled Successfully assigned kserve-ci-e2e-test/gw-section-name-router-with-gat-2f0a622e-kserve-7c9c8cffc558295 to ip-10-0-139-47.ec2.internal [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:46 kserve-ci-e2e-test multus Normal AddedInterface Add eth0 [10.134.0.21/23] from ovn-kubernetes [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:46 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Pulling Pulling image "ghcr.io/llm-d/llm-d-inference-sim:v0.8.2" [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:47 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Pulled Successfully pulled image "ghcr.io/llm-d/llm-d-inference-sim:v0.8.2" in 1.195s (1.195s including waiting). Image size: 98346788 bytes. [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:47 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Created Created container: main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:47 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Started Started container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:58 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Killing Stopping container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:45 kserve-ci-e2e-test deployment-controller Normal ScalingReplicaSet Scaled up replica set gw-section-name-router-with-gat-2f0a622e-kserve-7c9c8cffc5 from 0 to 1 [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:45 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Secret kserve-ci-e2e-test/gw-sec0c69dceeb48768325d1a53a749e65786-kserve-self-signed-certs [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:45 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Deployment kserve-ci-e2e-test/gw-section-name-router-with-gat-2f0a622e-kserve [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:45 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Service kserve-ci-e2e-test/gw-section-name-router-with-gat-2f0a622e-kserve-workload-svc [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:45 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.HTTPRoute kserve-ci-e2e-test/gw-section-name-router-with-gat-2f0a622e-kserve-route [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:46 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.DestinationRule kserve-ci-e2e-test/gw-section-name-router-with-gat-2f0a622e-kserve-workload-svc [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:10 kserve-ci-e2e-test replicaset-controller Normal SuccessfulCreate Created pod: gw-section-name-router-with-gat-f1d92d0f-kserve-7bc8dd6c5bcr4rf [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 None kserve-ci-e2e-test Normal Scheduled Successfully assigned kserve-ci-e2e-test/gw-section-name-router-with-gat-f1d92d0f-kserve-7bc8dd6c5bcr4rf to ip-10-0-133-25.ec2.internal [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:10 kserve-ci-e2e-test multus Normal AddedInterface Add eth0 [10.132.0.46/23] from ovn-kubernetes [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:10 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Pulled Container image "ghcr.io/llm-d/llm-d-inference-sim:v0.8.2" already present on machine [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:10 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Created Created container: main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:10 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Started Started container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:27 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Killing Stopping container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:10 kserve-ci-e2e-test deployment-controller Normal ScalingReplicaSet Scaled up replica set gw-section-name-router-with-gat-f1d92d0f-kserve-7bc8dd6c5b from 0 to 1 [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:10 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Secret kserve-ci-e2e-test/gw-sec2774c263d49959f50d9eebc552e13bf9-kserve-self-signed-certs [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:10 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Deployment kserve-ci-e2e-test/gw-section-name-router-with-gat-f1d92d0f-kserve [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:10 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Service kserve-ci-e2e-test/gw-section-name-router-with-gat-f1d92d0f-kserve-workload-svc [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:10 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.HTTPRoute kserve-ci-e2e-test/gw-section-name-router-with-gat-f1d92d0f-kserve-route [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:10 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.DestinationRule kserve-ci-e2e-test/gw-section-name-router-with-gat-f1d92d0f-kserve-workload-svc [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:31:40 kserve-ci-e2e-test replicaset-controller Normal SuccessfulCreate Created pod: llmisvc-model-fb-opt-125m-route-87882a8e-kserve-5bd7864c6ctnql9 [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 None kserve-ci-e2e-test Normal Scheduled Successfully assigned kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-route-87882a8e-kserve-5bd7864c6ctnql9 to ip-10-0-139-47.ec2.internal [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:31:40 kserve-ci-e2e-test multus Normal AddedInterface Add eth0 [10.134.0.31/23] from ovn-kubernetes [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:31:40 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Pulled Container image "quay.io/opendatahub/kserve-storage-initializer@sha256:13ad765c0f323200e1db7b45c99661287e65c5158d6b54b16139829636d20be4" already present on machine [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:31:40 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Created Created container: storage-initializer [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:31:40 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Started Started container storage-initializer [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:31:45 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Pulled Container image "public.ecr.aws/q9t5s3a7/vllm-cpu-release-repo:v0.19.0" already present on machine [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:31:45 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Created Created container: main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:31:45 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Started Started container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:33:30 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Warning Unhealthy Startup probe failed: Get "https://10.134.0.31:8000/health": dial tcp 10.134.0.31:8000: connect: connection refused [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:31:40 kserve-ci-e2e-test deployment-controller Normal ScalingReplicaSet Scaled up replica set llmisvc-model-fb-opt-125m-route-87882a8e-kserve-5bd7864c6c from 0 to 1 [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:31:38 kserve-ci-e2e-test OpenDataHubModelController Warning ReconcileError Failed to reconcile LLMInferenceService: 1 error occurred: * failed to get HTTPRoute for AuthPolicy llmisvc-model-fb-opt-125m-route-87882a8e-kserve-route-authn: failed to get HTTPRoute kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-route-87882a8e-kserve-route: HTTPRoute.gateway.networking.k8s.io "llmisvc-model-fb-opt-125m-route-87882a8e-kserve-route" not found [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:31:40 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Secret kserve-ci-e2e-test/llmisva690bbc929faec8bc98c767f16c003c1-kserve-self-signed-certs [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:31:40 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Deployment kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-route-87882a8e-kserve [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:31:40 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Service kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-route-87882a8e-kserve-workload-svc [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:31:40 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.ServiceAccount kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-route-87882a8e-epp-sa [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:31:40 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.ClusterRoleBinding /kserve-ci-e2e-test21fe6730fe484f3a92b1a16afe1bac8f-llmisvc-mode [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:31:40 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Role kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-route-87882a8e-epp-role [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:31:40 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.RoleBinding kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-route-87882a8e-epp-rb [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:31:40 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Deployment kserve-ci-e2e-test/llmisvca690bbc929faec8bc98c767f16c003c1-kserve-router-scheduler [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:31:40 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Service kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-route-87882a8e-epp-service [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:31:47 kserve-ci-e2e-test LLMInferenceServiceController Normal Created (combined from similar events): Created v1.DestinationRule kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-route-87882a8e-kserve-shadow-svc [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:31:45 kserve-ci-e2e-test LLMInferenceServiceController Normal Updated Updated v1.Secret kserve-ci-e2e-test/llmisva690bbc929faec8bc98c767f16c003c1-kserve-self-signed-certs [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:31:46 kserve-ci-e2e-test LLMInferenceServiceController Normal Updated Updated v1.HTTPRoute kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-route-87882a8e-kserve-route [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:33:40 kserve-ci-e2e-test LLMInferenceServiceController Normal LLMInferenceServiceReady LLMInferenceService [llmisvc-model-fb-opt-125m-route-87882a8e] is Ready [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:31:02 kserve-ci-e2e-test replicaset-controller Normal SuccessfulCreate Created pod: llmisvc-model-fb-opt-125m-route-e95b1dc1-kserve-66874c76d46pqbc [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 None kserve-ci-e2e-test Normal Scheduled Successfully assigned kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-route-e95b1dc1-kserve-66874c76d46pqbc to ip-10-0-135-220.ec2.internal [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:31:02 kserve-ci-e2e-test multus Normal AddedInterface Add eth0 [10.133.0.37/23] from ovn-kubernetes [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:31:02 kserve-ci-e2e-test kubelet/ip-10-0-135-220.ec2.internal Normal Pulled Container image "ghcr.io/llm-d/llm-d-inference-sim:v0.8.2" already present on machine [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:31:02 kserve-ci-e2e-test kubelet/ip-10-0-135-220.ec2.internal Normal Created Created container: main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:31:02 kserve-ci-e2e-test kubelet/ip-10-0-135-220.ec2.internal Normal Started Started container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:31:15 kserve-ci-e2e-test kubelet/ip-10-0-135-220.ec2.internal Normal Killing Stopping container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:31:02 kserve-ci-e2e-test deployment-controller Normal ScalingReplicaSet Scaled up replica set llmisvc-model-fb-opt-125m-route-e95b1dc1-kserve-66874c76d4 from 0 to 1 [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:30:58 kserve-ci-e2e-test OpenDataHubModelController Warning ReconcileError Failed to reconcile LLMInferenceService: 1 error occurred: * failed to get HTTPRoute for AuthPolicy llmisvc-model-fb-opt-125m-route-e95b1dc1-kserve-route-authn: failed to get HTTPRoute kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-route-e95b1dc1-kserve-route: HTTPRoute.gateway.networking.k8s.io "llmisvc-model-fb-opt-125m-route-e95b1dc1-kserve-route" not found [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:31:02 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Secret kserve-ci-e2e-test/llmisv122f03714c5bdf915a2917fdf1262b98-kserve-self-signed-certs [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:31:02 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Deployment kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-route-e95b1dc1-kserve [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:31:02 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Service kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-route-e95b1dc1-kserve-workload-svc [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:31:02 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.HTTPRoute kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-route-e95b1dc1-kserve-route [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:31:02 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.DestinationRule kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-route-e95b1dc1-kserve-workload-svc [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:31:14 kserve-ci-e2e-test LLMInferenceServiceController Normal Updated Updated v1.Secret kserve-ci-e2e-test/llmisv122f03714c5bdf915a2917fdf1262b98-kserve-self-signed-certs [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:31:14 kserve-ci-e2e-test LLMInferenceServiceController Normal LLMInferenceServiceReady LLMInferenceService [llmisvc-model-fb-opt-125m-route-e95b1dc1] is Ready [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 None kserve-ci-e2e-test Normal Scheduled Successfully assigned kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-with-7ca60146-kserve-7cfdf49d74-pd9px to ip-10-0-139-47.ec2.internal [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:29:25 kserve-ci-e2e-test multus Normal AddedInterface Add eth0 [10.134.0.30/23] from ovn-kubernetes [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:29:25 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Pulled Container image "quay.io/opendatahub/kserve-storage-initializer@sha256:13ad765c0f323200e1db7b45c99661287e65c5158d6b54b16139829636d20be4" already present on machine [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:29:25 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Created Created container: storage-initializer [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:29:25 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Started Started container storage-initializer [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:29:30 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Pulled Container image "public.ecr.aws/q9t5s3a7/vllm-cpu-release-repo:v0.19.0" already present on machine [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:29:30 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Created Created container: main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:29:30 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Started Started container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:30:34 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Warning Unhealthy Startup probe failed: Get "https://10.134.0.30:8000/health": dial tcp 10.134.0.30:8000: connect: connection refused [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:29:24 kserve-ci-e2e-test replicaset-controller Normal SuccessfulCreate Created pod: llmisvc-model-fb-opt-125m-with-7ca60146-kserve-7cfdf49d74-pd9px [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:29:24 kserve-ci-e2e-test replicaset-controller Normal SuccessfulCreate Created pod: llmisvc-model-fb-opt-125m-with-7ca60146-kserve-router-schenp5mk [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:29:24 kserve-ci-e2e-test deployment-controller Normal ScalingReplicaSet Scaled up replica set llmisvc-model-fb-opt-125m-with-7ca60146-kserve-router-scheduler-778547fbc from 0 to 1 [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 None kserve-ci-e2e-test Normal Scheduled Successfully assigned kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-with-7ca60146-kserve-router-schenp5mk to ip-10-0-133-25.ec2.internal [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:29:25 kserve-ci-e2e-test multus Normal AddedInterface Add eth0 [10.132.0.52/23] from ovn-kubernetes [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:29:25 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Pulled Container image "quay.io/opendatahub/kserve-storage-initializer@sha256:13ad765c0f323200e1db7b45c99661287e65c5158d6b54b16139829636d20be4" already present on machine [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:29:25 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Created Created container: storage-initializer [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:29:25 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Started Started container storage-initializer [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:29:26 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Pulled Container image "ghcr.io/llm-d/llm-d-router-endpoint-picker:v0.9.0-rc.2" already present on machine [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:29:27 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Created Created container: main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:29:27 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Started Started container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:29:24 kserve-ci-e2e-test deployment-controller Normal ScalingReplicaSet Scaled up replica set llmisvc-model-fb-opt-125m-with-7ca60146-kserve-7cfdf49d74 from 0 to 1 [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:29:23 kserve-ci-e2e-test OpenDataHubModelController Warning ReconcileError Failed to reconcile LLMInferenceService: 1 error occurred: * failed to get HTTPRoute for AuthPolicy llmisvc-model-fb-opt-125m-with-7ca60146-kserve-route-authn: failed to get HTTPRoute kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-with-7ca60146-kserve-route: HTTPRoute.gateway.networking.k8s.io "llmisvc-model-fb-opt-125m-with-7ca60146-kserve-route" not found [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:29:24 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Secret kserve-ci-e2e-test/llmisv3e414c2ba058a022dfd694dbcbac5b51-kserve-self-signed-certs [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:29:24 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Deployment kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-with-7ca60146-kserve [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:29:24 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Service kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-with-7ca60146-kserve-workload-svc [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:29:24 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.ServiceAccount kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-with-7ca60146-epp-sa [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:29:24 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.ClusterRoleBinding /kserve-ci-e2e-test5216bfd716f919dc046bc693ceb22e41-llmisvc-mode [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:29:24 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Role kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-with-7ca60146-epp-role [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:29:24 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.RoleBinding kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-with-7ca60146-epp-rb [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:29:24 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Deployment kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-with-7ca60146-kserve-router-scheduler [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:29:24 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Service kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-with-7ca60146-epp-service [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:29:35 kserve-ci-e2e-test LLMInferenceServiceController Normal Created (combined from similar events): Created v1.DestinationRule kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-with-7ca60146-kserve-shadow-svc [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:29:33 kserve-ci-e2e-test LLMInferenceServiceController Normal Updated Updated v1.Secret kserve-ci-e2e-test/llmisv3e414c2ba058a022dfd694dbcbac5b51-kserve-self-signed-certs [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:29:34 kserve-ci-e2e-test LLMInferenceServiceController Normal Updated Updated v1.HTTPRoute kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-with-7ca60146-kserve-route [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:30:45 kserve-ci-e2e-test LLMInferenceServiceController Normal LLMInferenceServiceReady LLMInferenceService [llmisvc-model-fb-opt-125m-with-7ca60146] is Ready [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 None kserve-ci-e2e-test Normal Scheduled Successfully assigned kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-with-ba4d693a-kserve-5b998d5789-t6wf4 to ip-10-0-139-47.ec2.internal [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:45:56 kserve-ci-e2e-test multus Normal AddedInterface Add eth0 [10.134.0.32/23] from ovn-kubernetes [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:45:56 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Pulled Container image "quay.io/opendatahub/kserve-storage-initializer@sha256:13ad765c0f323200e1db7b45c99661287e65c5158d6b54b16139829636d20be4" already present on machine [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:45:56 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Created Created container: storage-initializer [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:45:56 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Started Started container storage-initializer [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:46:01 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Pulled Container image "public.ecr.aws/q9t5s3a7/vllm-cpu-release-repo:v0.19.0" already present on machine [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:46:01 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Created Created container: main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:46:01 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Started Started container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:47:06 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Warning Unhealthy Startup probe failed: Get "https://10.134.0.32:8000/health": dial tcp 10.134.0.32:8000: connect: connection refused [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:45:56 kserve-ci-e2e-test replicaset-controller Normal SuccessfulCreate Created pod: llmisvc-model-fb-opt-125m-with-ba4d693a-kserve-5b998d5789-t6wf4 [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 None kserve-ci-e2e-test Normal Scheduled Successfully assigned kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-with-ba4d693a-kserve-router-sche4ghr9 to ip-10-0-133-25.ec2.internal [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:45:56 kserve-ci-e2e-test multus Normal AddedInterface Add eth0 [10.132.0.54/23] from ovn-kubernetes [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:45:56 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Pulled Container image "quay.io/opendatahub/kserve-storage-initializer@sha256:13ad765c0f323200e1db7b45c99661287e65c5158d6b54b16139829636d20be4" already present on machine [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:45:57 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Created Created container: storage-initializer [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:45:57 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Started Started container storage-initializer [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:45:57 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Pulled Container image "ghcr.io/llm-d/llm-d-router-endpoint-picker:v0.9.0-rc.2" already present on machine [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:45:57 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Created Created container: main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:45:57 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Started Started container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:45:56 kserve-ci-e2e-test replicaset-controller Normal SuccessfulCreate Created pod: llmisvc-model-fb-opt-125m-with-ba4d693a-kserve-router-sche4ghr9 [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:45:56 kserve-ci-e2e-test deployment-controller Normal ScalingReplicaSet Scaled up replica set llmisvc-model-fb-opt-125m-with-ba4d693a-kserve-router-scheduler-79f5455bb from 0 to 1 [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:45:56 kserve-ci-e2e-test deployment-controller Normal ScalingReplicaSet Scaled up replica set llmisvc-model-fb-opt-125m-with-ba4d693a-kserve-5b998d5789 from 0 to 1 [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:45:56 kserve-ci-e2e-test OpenDataHubModelController Warning ReconcileError Failed to reconcile LLMInferenceService: 1 error occurred: * failed to get HTTPRoute for AuthPolicy llmisvc-model-fb-opt-125m-with-ba4d693a-kserve-route-authn: failed to get HTTPRoute kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-with-ba4d693a-kserve-route: HTTPRoute.gateway.networking.k8s.io "llmisvc-model-fb-opt-125m-with-ba4d693a-kserve-route" not found [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:45:56 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Secret kserve-ci-e2e-test/llmisv77ff2528d3e9b4972cd9335229fce9f0-kserve-self-signed-certs [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:45:56 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Deployment kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-with-ba4d693a-kserve [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:45:56 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Service kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-with-ba4d693a-kserve-workload-svc [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:45:56 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.ServiceAccount kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-with-ba4d693a-epp-sa [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:45:56 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.ClusterRoleBinding /kserve-ci-e2e-test05addb65ba05195619f26ef266e8fc04-llmisvc-mode [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:45:56 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Role kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-with-ba4d693a-epp-role [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:45:56 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.RoleBinding kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-with-ba4d693a-epp-rb [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:45:56 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Deployment kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-with-ba4d693a-kserve-router-scheduler [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:45:56 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Service kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-with-ba4d693a-epp-service [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:46:17 kserve-ci-e2e-test LLMInferenceServiceController Normal Created (combined from similar events): Created v1.DestinationRule kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-with-ba4d693a-kserve-shadow-svc [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:46:16 kserve-ci-e2e-test LLMInferenceServiceController Normal Updated Updated v1.Secret kserve-ci-e2e-test/llmisv77ff2528d3e9b4972cd9335229fce9f0-kserve-self-signed-certs [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:46:16 kserve-ci-e2e-test LLMInferenceServiceController Normal Updated Updated v1.HTTPRoute kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-with-ba4d693a-kserve-route [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:47:17 kserve-ci-e2e-test LLMInferenceServiceController Normal LLMInferenceServiceReady LLMInferenceService [llmisvc-model-fb-opt-125m-with-ba4d693a] is Ready [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:54 kserve-ci-e2e-test replicaset-controller Normal SuccessfulCreate Created pod: llmisvc-model-qwen2-5-0-5b-rout-4f8c0978-kserve-5c54ddb98flf892 [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 None kserve-ci-e2e-test Normal Scheduled Successfully assigned kserve-ci-e2e-test/llmisvc-model-qwen2-5-0-5b-rout-4f8c0978-kserve-5c54ddb98flf892 to ip-10-0-135-220.ec2.internal [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:54 kserve-ci-e2e-test multus Normal AddedInterface Add eth0 [10.133.0.36/23] from ovn-kubernetes [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:54 kserve-ci-e2e-test kubelet/ip-10-0-135-220.ec2.internal Normal Pulled Container image "ghcr.io/llm-d/llm-d-inference-sim:v0.8.2" already present on machine [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:54 kserve-ci-e2e-test kubelet/ip-10-0-135-220.ec2.internal Normal Created Created container: main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:54 kserve-ci-e2e-test kubelet/ip-10-0-135-220.ec2.internal Normal Started Started container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:28:40 kserve-ci-e2e-test kubelet/ip-10-0-135-220.ec2.internal Normal Killing Stopping container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:28:56 kserve-ci-e2e-test kubelet/ip-10-0-135-220.ec2.internal Warning Unhealthy Readiness probe failed: Get "https://10.133.0.36:8000/health": context deadline exceeded [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:54 kserve-ci-e2e-test deployment-controller Normal ScalingReplicaSet Scaled up replica set llmisvc-model-qwen2-5-0-5b-rout-4f8c0978-kserve-5c54ddb98f from 0 to 1 [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:53 kserve-ci-e2e-test OpenDataHubModelController Warning ReconcileError Failed to reconcile LLMInferenceService: 1 error occurred: * failed to get HTTPRoute for AuthPolicy llmisvc-model-qwen2-5-0-5b-rout-4f8c0978-kserve-route-authn: failed to get HTTPRoute kserve-ci-e2e-test/llmisvc-model-qwen2-5-0-5b-rout-4f8c0978-kserve-route: HTTPRoute.gateway.networking.k8s.io "llmisvc-model-qwen2-5-0-5b-rout-4f8c0978-kserve-route" not found [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:54 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Secret kserve-ci-e2e-test/llmisv08544b88a8d963ffd553cc1f3ed82d16-kserve-self-signed-certs [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:54 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Deployment kserve-ci-e2e-test/llmisvc-model-qwen2-5-0-5b-rout-4f8c0978-kserve [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:54 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Service kserve-ci-e2e-test/llmisvc-model-qwen2-5-0-5b-rout-4f8c0978-kserve-workload-svc [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:54 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.ServiceAccount kserve-ci-e2e-test/llmisvc-model-qwen2-5-0-5b-rout-4f8c0978-epp-sa [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:54 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.ClusterRoleBinding /kserve-ci-e2e-test76d7f532acb694e4a7bcef75d32cd8a1-llmisvc-mode [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:54 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Role kserve-ci-e2e-test/llmisvc-model-qwen2-5-0-5b-rout-4f8c0978-epp-role [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:54 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.RoleBinding kserve-ci-e2e-test/llmisvc-model-qwen2-5-0-5b-rout-4f8c0978-epp-rb [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:54 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Deployment kserve-ci-e2e-test/llmisvc08544b88a8d963ffd553cc1f3ed82d16-kserve-router-scheduler [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:54 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Service kserve-ci-e2e-test/llmisvc-model-qwen2-5-0-5b-rout-4f8c0978-epp-service [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:28:06 kserve-ci-e2e-test LLMInferenceServiceController Normal Created (combined from similar events): Created v1.DestinationRule kserve-ci-e2e-test/llmisvc-model-qwen2-5-0-5b-rout-4f8c0978-kserve-shadow-svc [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:28:05 kserve-ci-e2e-test LLMInferenceServiceController Normal Updated Updated v1.Secret kserve-ci-e2e-test/llmisv08544b88a8d963ffd553cc1f3ed82d16-kserve-self-signed-certs [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:28:06 kserve-ci-e2e-test LLMInferenceServiceController Normal Updated Updated v1.HTTPRoute kserve-ci-e2e-test/llmisvc-model-qwen2-5-0-5b-rout-4f8c0978-kserve-route [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:28:31 kserve-ci-e2e-test LLMInferenceServiceController Normal LLMInferenceServiceReady LLMInferenceService [llmisvc-model-qwen2-5-0-5b-rout-4f8c0978] is Ready [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:28:40 kserve-ci-e2e-test LLMInferenceServiceController Normal Deleted Deleted v1.ClusterRoleBinding /kserve-ci-e2e-test76d7f532acb694e4a7bcef75d32cd8a1-llmisvc-mode [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:23 kserve-ci-e2e-test replicaset-controller Normal SuccessfulCreate Created pod: llmisvc-model-qwen2-5-0-5b-rout-a50492e9-kserve-5d8ffd58ddqnd8m [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 None kserve-ci-e2e-test Normal Scheduled Successfully assigned kserve-ci-e2e-test/llmisvc-model-qwen2-5-0-5b-rout-a50492e9-kserve-5d8ffd58ddqnd8m to ip-10-0-135-220.ec2.internal [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:23 kserve-ci-e2e-test multus Normal AddedInterface Add eth0 [10.133.0.34/23] from ovn-kubernetes [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:23 kserve-ci-e2e-test kubelet/ip-10-0-135-220.ec2.internal Normal Pulled Container image "ghcr.io/llm-d/llm-d-inference-sim:v0.8.2" already present on machine [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:23 kserve-ci-e2e-test kubelet/ip-10-0-135-220.ec2.internal Normal Created Created container: main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:23 kserve-ci-e2e-test kubelet/ip-10-0-135-220.ec2.internal Normal Started Started container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:26:13 kserve-ci-e2e-test kubelet/ip-10-0-135-220.ec2.internal Normal Killing Stopping container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:23 kserve-ci-e2e-test deployment-controller Normal ScalingReplicaSet Scaled up replica set llmisvc-model-qwen2-5-0-5b-rout-a50492e9-kserve-5d8ffd58dd from 0 to 1 [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:21 kserve-ci-e2e-test OpenDataHubModelController Warning ReconcileError Failed to reconcile LLMInferenceService: 1 error occurred: * failed to get HTTPRoute for AuthPolicy llmisvc-model-qwen2-5-0-5b-rout-a50492e9-kserve-route-authn: failed to get HTTPRoute kserve-ci-e2e-test/llmisvc-model-qwen2-5-0-5b-rout-a50492e9-kserve-route: HTTPRoute.gateway.networking.k8s.io "llmisvc-model-qwen2-5-0-5b-rout-a50492e9-kserve-route" not found [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:23 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Secret kserve-ci-e2e-test/llmisvb11a9c9832b99b016bc8f8e0ea095712-kserve-self-signed-certs [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:23 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Deployment kserve-ci-e2e-test/llmisvc-model-qwen2-5-0-5b-rout-a50492e9-kserve [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:23 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Service kserve-ci-e2e-test/llmisvc-model-qwen2-5-0-5b-rout-a50492e9-kserve-workload-svc [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:23 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.ServiceAccount kserve-ci-e2e-test/llmisvc-model-qwen2-5-0-5b-rout-a50492e9-epp-sa [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:23 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.ClusterRoleBinding /kserve-ci-e2e-testb7025ca4d8a6f8f5b2fd08b5581d2678-llmisvc-mode [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:23 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Role kserve-ci-e2e-test/llmisvc-model-qwen2-5-0-5b-rout-a50492e9-epp-role [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:23 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.RoleBinding kserve-ci-e2e-test/llmisvc-model-qwen2-5-0-5b-rout-a50492e9-epp-rb [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:23 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Deployment kserve-ci-e2e-test/llmisvcb11a9c9832b99b016bc8f8e0ea095712-kserve-router-scheduler [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:23 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Service kserve-ci-e2e-test/llmisvc-model-qwen2-5-0-5b-rout-a50492e9-epp-service [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:34 kserve-ci-e2e-test LLMInferenceServiceController Normal Created (combined from similar events): Created v1.DestinationRule kserve-ci-e2e-test/llmisvc-model-qwen2-5-0-5b-rout-a50492e9-kserve-shadow-svc [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:33 kserve-ci-e2e-test LLMInferenceServiceController Normal Updated Updated v1.Secret kserve-ci-e2e-test/llmisvb11a9c9832b99b016bc8f8e0ea095712-kserve-self-signed-certs [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:33 kserve-ci-e2e-test LLMInferenceServiceController Normal Updated Updated v1.HTTPRoute kserve-ci-e2e-test/llmisvc-model-qwen2-5-0-5b-rout-a50492e9-kserve-route [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:26:12 kserve-ci-e2e-test LLMInferenceServiceController Normal LLMInferenceServiceReady LLMInferenceService [llmisvc-model-qwen2-5-0-5b-rout-a50492e9] is Ready [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:26:13 kserve-ci-e2e-test LLMInferenceServiceController Normal Deleted Deleted v1.ClusterRoleBinding /kserve-ci-e2e-testb7025ca4d8a6f8f5b2fd08b5581d2678-llmisvc-mode [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 None kserve-ci-e2e-test Normal Scheduled Successfully assigned kserve-ci-e2e-test/llmisvc-router-managed-test-llm-4b931143-kserve-bd545d56-jcs4n to ip-10-0-135-220.ec2.internal [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:18 kserve-ci-e2e-test multus Normal AddedInterface Add eth0 [10.133.0.35/23] from ovn-kubernetes [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:18 kserve-ci-e2e-test kubelet/ip-10-0-135-220.ec2.internal Normal Pulled Container image "ghcr.io/llm-d/llm-d-inference-sim:v0.8.2" already present on machine [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:18 kserve-ci-e2e-test kubelet/ip-10-0-135-220.ec2.internal Normal Created Created container: main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:18 kserve-ci-e2e-test kubelet/ip-10-0-135-220.ec2.internal Normal Started Started container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:28:58 kserve-ci-e2e-test kubelet/ip-10-0-135-220.ec2.internal Normal Killing Stopping container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:17 kserve-ci-e2e-test replicaset-controller Normal SuccessfulCreate Created pod: llmisvc-router-managed-test-llm-4b931143-kserve-bd545d56-jcs4n [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:17 kserve-ci-e2e-test deployment-controller Normal ScalingReplicaSet Scaled up replica set llmisvc-router-managed-test-llm-4b931143-kserve-bd545d56 from 0 to 1 [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:14 kserve-ci-e2e-test OpenDataHubModelController Warning ReconcileError Failed to reconcile LLMInferenceService: 1 error occurred: * failed to get HTTPRoute for AuthPolicy llmisvc-router-managed-test-llm-4b931143-kserve-route-authn: failed to get HTTPRoute kserve-ci-e2e-test/llmisvc-router-managed-test-llm-4b931143-kserve-route: HTTPRoute.gateway.networking.k8s.io "llmisvc-router-managed-test-llm-4b931143-kserve-route" not found [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:17 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Secret kserve-ci-e2e-test/llmisvca2d2d7d499abb359505529ebe02c136-kserve-self-signed-certs [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:17 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Deployment kserve-ci-e2e-test/llmisvc-router-managed-test-llm-4b931143-kserve [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:17 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Service kserve-ci-e2e-test/llmisvc-router-managed-test-llm-4b931143-kserve-workload-svc [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:17 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.ServiceAccount kserve-ci-e2e-test/llmisvc-router-managed-test-llm-4b931143-epp-sa [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:17 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.ClusterRoleBinding /kserve-ci-e2e-test8ac8e3d2264ccb939eb021b0b835847c-llmisvc-rout [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:17 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Role kserve-ci-e2e-test/llmisvc-router-managed-test-llm-4b931143-epp-role [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:17 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.RoleBinding kserve-ci-e2e-test/llmisvc-router-managed-test-llm-4b931143-epp-rb [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:17 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Deployment kserve-ci-e2e-test/llmisvcca2d2d7d499abb359505529ebe02c136-kserve-router-scheduler [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:17 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Service kserve-ci-e2e-test/llmisvc-router-managed-test-llm-4b931143-epp-service [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:43 kserve-ci-e2e-test LLMInferenceServiceController Normal Created (combined from similar events): Created v1.DestinationRule kserve-ci-e2e-test/llmisvc-router-managed-test-llm-4b931143-kserve-shadow-svc [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:42 kserve-ci-e2e-test LLMInferenceServiceController Normal Updated Updated v1.Secret kserve-ci-e2e-test/llmisvca2d2d7d499abb359505529ebe02c136-kserve-self-signed-certs [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:43 kserve-ci-e2e-test LLMInferenceServiceController Normal Updated Updated v1.HTTPRoute kserve-ci-e2e-test/llmisvc-router-managed-test-llm-4b931143-kserve-route [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:50 kserve-ci-e2e-test LLMInferenceServiceController Normal LLMInferenceServiceReady LLMInferenceService [llmisvc-router-managed-test-llm-4b931143] is Ready [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:28:58 kserve-ci-e2e-test LLMInferenceServiceController Normal Deleted Deleted v1.ClusterRoleBinding /kserve-ci-e2e-test8ac8e3d2264ccb939eb021b0b835847c-llmisvc-rout [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:24:42 kserve-ci-e2e-test replicaset-controller Normal SuccessfulCreate Created pod: llmisvc-router-managed-test-llm-5b1e8f15-kserve-64df7bddb4xtf2m [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 None kserve-ci-e2e-test Normal Scheduled Successfully assigned kserve-ci-e2e-test/llmisvc-router-managed-test-llm-5b1e8f15-kserve-64df7bddb4xtf2m to ip-10-0-135-220.ec2.internal [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:24:43 kserve-ci-e2e-test multus Normal AddedInterface Add eth0 [10.133.0.33/23] from ovn-kubernetes [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:24:43 kserve-ci-e2e-test kubelet/ip-10-0-135-220.ec2.internal Normal Pulled Container image "ghcr.io/llm-d/llm-d-inference-sim:v0.8.2" already present on machine [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:24:43 kserve-ci-e2e-test kubelet/ip-10-0-135-220.ec2.internal Normal Created Created container: main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:24:43 kserve-ci-e2e-test kubelet/ip-10-0-135-220.ec2.internal Normal Started Started container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:26:33 kserve-ci-e2e-test kubelet/ip-10-0-135-220.ec2.internal Normal Killing Stopping container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:24:42 kserve-ci-e2e-test deployment-controller Normal ScalingReplicaSet Scaled up replica set llmisvc-router-managed-test-llm-5b1e8f15-kserve-64df7bddb4 from 0 to 1 [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:24:40 kserve-ci-e2e-test OpenDataHubModelController Warning ReconcileError Failed to reconcile LLMInferenceService: 1 error occurred: * failed to get HTTPRoute for AuthPolicy llmisvc-router-managed-test-llm-5b1e8f15-kserve-route-authn: failed to get HTTPRoute kserve-ci-e2e-test/llmisvc-router-managed-test-llm-5b1e8f15-kserve-route: HTTPRoute.gateway.networking.k8s.io "llmisvc-router-managed-test-llm-5b1e8f15-kserve-route" not found [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:24:42 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Secret kserve-ci-e2e-test/llmisve55ae740357a3a31a27cdb8b66ffe20f-kserve-self-signed-certs [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:24:42 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Deployment kserve-ci-e2e-test/llmisvc-router-managed-test-llm-5b1e8f15-kserve [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:24:42 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Service kserve-ci-e2e-test/llmisvc-router-managed-test-llm-5b1e8f15-kserve-workload-svc [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:24:42 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.ServiceAccount kserve-ci-e2e-test/llmisvc-router-managed-test-llm-5b1e8f15-epp-sa [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:24:42 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.ClusterRoleBinding /kserve-ci-e2e-test7f54e84970003a6e7372bdbcb574f7ed-llmisvc-rout [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:24:42 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Role kserve-ci-e2e-test/llmisvc-router-managed-test-llm-5b1e8f15-epp-role [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:24:43 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.RoleBinding kserve-ci-e2e-test/llmisvc-router-managed-test-llm-5b1e8f15-epp-rb [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:24:43 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Deployment kserve-ci-e2e-test/llmisvce55ae740357a3a31a27cdb8b66ffe20f-kserve-router-scheduler [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:24:43 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Service kserve-ci-e2e-test/llmisvc-router-managed-test-llm-5b1e8f15-epp-service [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:03 kserve-ci-e2e-test LLMInferenceServiceController Normal Created (combined from similar events): Created v1.DestinationRule kserve-ci-e2e-test/llmisvc-router-managed-test-llm-5b1e8f15-kserve-shadow-svc [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:03 kserve-ci-e2e-test LLMInferenceServiceController Normal Updated Updated v1.Secret kserve-ci-e2e-test/llmisve55ae740357a3a31a27cdb8b66ffe20f-kserve-self-signed-certs [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:03 kserve-ci-e2e-test LLMInferenceServiceController Normal Updated Updated v1.HTTPRoute kserve-ci-e2e-test/llmisvc-router-managed-test-llm-5b1e8f15-kserve-route [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:15 kserve-ci-e2e-test LLMInferenceServiceController Normal LLMInferenceServiceReady LLMInferenceService [llmisvc-router-managed-test-llm-5b1e8f15] is Ready [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:26:33 kserve-ci-e2e-test LLMInferenceServiceController Normal Deleted Deleted v1.ClusterRoleBinding /kserve-ci-e2e-test7f54e84970003a6e7372bdbcb574f7ed-llmisvc-rout [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:23:41 kserve-ci-e2e-test replicaset-controller Normal SuccessfulCreate Created pod: llmisvc-router-managed-test-llm-e45d1f79-kserve-7fdbbd4879ffz96 [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 None kserve-ci-e2e-test Normal Scheduled Successfully assigned kserve-ci-e2e-test/llmisvc-router-managed-test-llm-e45d1f79-kserve-7fdbbd4879ffz96 to ip-10-0-135-220.ec2.internal [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:23:41 kserve-ci-e2e-test multus Normal AddedInterface Add eth0 [10.133.0.32/23] from ovn-kubernetes [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:23:41 kserve-ci-e2e-test kubelet/ip-10-0-135-220.ec2.internal Normal Pulling Pulling image "ghcr.io/llm-d/llm-d-inference-sim:v0.8.2" [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:23:43 kserve-ci-e2e-test kubelet/ip-10-0-135-220.ec2.internal Normal Pulled Successfully pulled image "ghcr.io/llm-d/llm-d-inference-sim:v0.8.2" in 1.407s (1.407s including waiting). Image size: 98346788 bytes. [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:23:43 kserve-ci-e2e-test kubelet/ip-10-0-135-220.ec2.internal Normal Created Created container: main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:23:43 kserve-ci-e2e-test kubelet/ip-10-0-135-220.ec2.internal Normal Started Started container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:24:15 kserve-ci-e2e-test kubelet/ip-10-0-135-220.ec2.internal Normal Killing Stopping container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:23:41 kserve-ci-e2e-test deployment-controller Normal ScalingReplicaSet Scaled up replica set llmisvc-router-managed-test-llm-e45d1f79-kserve-7fdbbd4879 from 0 to 1 [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:23:36 kserve-ci-e2e-test OpenDataHubModelController Warning ReconcileError Failed to reconcile LLMInferenceService: 1 error occurred: * failed to get HTTPRoute for AuthPolicy llmisvc-router-managed-test-llm-e45d1f79-kserve-route-authn: failed to get HTTPRoute kserve-ci-e2e-test/llmisvc-router-managed-test-llm-e45d1f79-kserve-route: HTTPRoute.gateway.networking.k8s.io "llmisvc-router-managed-test-llm-e45d1f79-kserve-route" not found [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:23:41 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Secret kserve-ci-e2e-test/llmisv5c7e67b6c51568d1d6d13829a9337f2a-kserve-self-signed-certs [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:23:41 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Deployment kserve-ci-e2e-test/llmisvc-router-managed-test-llm-e45d1f79-kserve [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:23:41 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Service kserve-ci-e2e-test/llmisvc-router-managed-test-llm-e45d1f79-kserve-workload-svc [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:23:41 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.ServiceAccount kserve-ci-e2e-test/llmisvc-router-managed-test-llm-e45d1f79-epp-sa [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:23:41 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.ClusterRoleBinding /kserve-ci-e2e-testef4d2875be14b30dc1561ed84d0d4bde-llmisvc-rout [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:23:41 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Role kserve-ci-e2e-test/llmisvc-router-managed-test-llm-e45d1f79-epp-role [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:23:41 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.RoleBinding kserve-ci-e2e-test/llmisvc-router-managed-test-llm-e45d1f79-epp-rb [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:23:41 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Deployment kserve-ci-e2e-test/llmisvc5c7e67b6c51568d1d6d13829a9337f2a-kserve-router-scheduler [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:23:41 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Service kserve-ci-e2e-test/llmisvc-router-managed-test-llm-e45d1f79-epp-service [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:23:53 kserve-ci-e2e-test LLMInferenceServiceController Normal Created (combined from similar events): Created v1.DestinationRule kserve-ci-e2e-test/llmisvc-router-managed-test-llm-e45d1f79-kserve-shadow-svc [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:23:52 kserve-ci-e2e-test LLMInferenceServiceController Normal Updated Updated v1.Secret kserve-ci-e2e-test/llmisv5c7e67b6c51568d1d6d13829a9337f2a-kserve-self-signed-certs [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:23:52 kserve-ci-e2e-test LLMInferenceServiceController Normal Updated Updated v1.HTTPRoute kserve-ci-e2e-test/llmisvc-router-managed-test-llm-e45d1f79-kserve-route [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:24:14 kserve-ci-e2e-test LLMInferenceServiceController Normal LLMInferenceServiceReady LLMInferenceService [llmisvc-router-managed-test-llm-e45d1f79] is Ready [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:24:15 kserve-ci-e2e-test LLMInferenceServiceController Normal Deleted Deleted v1.ClusterRoleBinding /kserve-ci-e2e-testef4d2875be14b30dc1561ed84d0d4bde-llmisvc-rout [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:54 kserve-ci-e2e-test replicaset-controller Normal SuccessfulCreate Created pod: llmisvc08544b88a8d963ffd553cc1f3ed82d16-kserve-router-schefvtgh [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:54 kserve-ci-e2e-test deployment-controller Normal ScalingReplicaSet Scaled up replica set llmisvc08544b88a8d963ffd553cc1f3ed82d16-kserve-router-scheduler-6fcb489785 from 0 to 1 [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 None kserve-ci-e2e-test Normal Scheduled Successfully assigned kserve-ci-e2e-test/llmisvc08544b88a8d963ffd553cc1f3ed82d16-kserve-router-schefvtgh to ip-10-0-133-25.ec2.internal [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:55 kserve-ci-e2e-test multus Normal AddedInterface Add eth0 [10.132.0.50/23] from ovn-kubernetes [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:55 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Pulled Container image "ghcr.io/llm-d/llm-d-router-endpoint-picker:v0.9.0-rc.2" already present on machine [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:55 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Created Created container: main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:55 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Started Started container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:28:40 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Killing Stopping container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:28:40 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Warning FailedPreStopHook PreStopHook failed [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:23:41 kserve-ci-e2e-test replicaset-controller Normal SuccessfulCreate Created pod: llmisvc5c7e67b6c51568d1d6d13829a9337f2a-kserve-router-schexv898 [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:23:41 kserve-ci-e2e-test deployment-controller Normal ScalingReplicaSet Scaled up replica set llmisvc5c7e67b6c51568d1d6d13829a9337f2a-kserve-router-scheduler-5dd88bfbb7 from 0 to 1 [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 None kserve-ci-e2e-test Normal Scheduled Successfully assigned kserve-ci-e2e-test/llmisvc5c7e67b6c51568d1d6d13829a9337f2a-kserve-router-schexv898 to ip-10-0-139-47.ec2.internal [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:23:42 kserve-ci-e2e-test multus Normal AddedInterface Add eth0 [10.134.0.25/23] from ovn-kubernetes [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:23:42 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Pulled Container image "ghcr.io/llm-d/llm-d-router-endpoint-picker:v0.9.0-rc.2" already present on machine [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:23:42 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Created Created container: main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:23:42 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Started Started container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:24:15 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Killing Stopping container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:24:15 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Warning FailedPreStopHook PreStopHook failed [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:31:40 kserve-ci-e2e-test replicaset-controller Normal SuccessfulCreate Created pod: llmisvca690bbc929faec8bc98c767f16c003c1-kserve-router-scheq44kv [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:31:40 kserve-ci-e2e-test deployment-controller Normal ScalingReplicaSet Scaled up replica set llmisvca690bbc929faec8bc98c767f16c003c1-kserve-router-scheduler-5d6998448f from 0 to 1 [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 None kserve-ci-e2e-test Normal Scheduled Successfully assigned kserve-ci-e2e-test/llmisvca690bbc929faec8bc98c767f16c003c1-kserve-router-scheq44kv to ip-10-0-133-25.ec2.internal [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:31:40 kserve-ci-e2e-test multus Normal AddedInterface Add eth0 [10.132.0.53/23] from ovn-kubernetes [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:31:41 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Pulled Container image "quay.io/opendatahub/kserve-storage-initializer@sha256:13ad765c0f323200e1db7b45c99661287e65c5158d6b54b16139829636d20be4" already present on machine [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:31:41 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Created Created container: storage-initializer [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:31:41 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Started Started container storage-initializer [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:31:42 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Pulled Container image "ghcr.io/llm-d/llm-d-router-endpoint-picker:v0.9.0-rc.2" already present on machine [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:31:42 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Created Created container: main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:31:42 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Started Started container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:23 kserve-ci-e2e-test replicaset-controller Normal SuccessfulCreate Created pod: llmisvcb11a9c9832b99b016bc8f8e0ea095712-kserve-router-scheknzg7 [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:23 kserve-ci-e2e-test deployment-controller Normal ScalingReplicaSet Scaled up replica set llmisvcb11a9c9832b99b016bc8f8e0ea095712-kserve-router-scheduler-6d86bd4d9d from 0 to 1 [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 None kserve-ci-e2e-test Normal Scheduled Successfully assigned kserve-ci-e2e-test/llmisvcb11a9c9832b99b016bc8f8e0ea095712-kserve-router-scheknzg7 to ip-10-0-133-25.ec2.internal [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:24 kserve-ci-e2e-test multus Normal AddedInterface Add eth0 [10.132.0.48/23] from ovn-kubernetes [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:24 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Pulled Container image "ghcr.io/llm-d/llm-d-router-endpoint-picker:v0.9.0-rc.2" already present on machine [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:24 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Created Created container: main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:24 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Started Started container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:26:13 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Killing Stopping container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:26:13 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Warning FailedPreStopHook PreStopHook failed [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 None kserve-ci-e2e-test Normal Scheduled Successfully assigned kserve-ci-e2e-test/llmisvcca2d2d7d499abb359505529ebe02c136-kserve-router-scheb2glb to ip-10-0-139-47.ec2.internal [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:18 kserve-ci-e2e-test multus Normal AddedInterface Add eth0 [10.134.0.28/23] from ovn-kubernetes [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:18 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Pulled Container image "ghcr.io/llm-d/llm-d-router-endpoint-picker:v0.9.0-rc.2" already present on machine [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:18 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Created Created container: main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:18 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Started Started container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:28:58 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Killing Stopping container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:28:58 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Warning FailedPreStopHook PreStopHook failed [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:17 kserve-ci-e2e-test replicaset-controller Normal SuccessfulCreate Created pod: llmisvcca2d2d7d499abb359505529ebe02c136-kserve-router-scheb2glb [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:17 kserve-ci-e2e-test deployment-controller Normal ScalingReplicaSet Scaled up replica set llmisvcca2d2d7d499abb359505529ebe02c136-kserve-router-scheduler-68cc9685d6 from 0 to 1 [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:24:43 kserve-ci-e2e-test replicaset-controller Normal SuccessfulCreate Created pod: llmisvce55ae740357a3a31a27cdb8b66ffe20f-kserve-router-schex4lqs [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:24:43 kserve-ci-e2e-test deployment-controller Normal ScalingReplicaSet Scaled up replica set llmisvce55ae740357a3a31a27cdb8b66ffe20f-kserve-router-scheduler-749449dbc8 from 0 to 1 [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 None kserve-ci-e2e-test Normal Scheduled Successfully assigned kserve-ci-e2e-test/llmisvce55ae740357a3a31a27cdb8b66ffe20f-kserve-router-schex4lqs to ip-10-0-139-47.ec2.internal [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:24:43 kserve-ci-e2e-test multus Normal AddedInterface Add eth0 [10.134.0.26/23] from ovn-kubernetes [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:24:43 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Pulled Container image "ghcr.io/llm-d/llm-d-router-endpoint-picker:v0.9.0-rc.2" already present on machine [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:24:43 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Created Created container: main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:24:43 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Started Started container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:26:33 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Killing Stopping container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:26:33 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Warning FailedPreStopHook PreStopHook failed [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 None kserve-ci-e2e-test Normal Scheduled Successfully assigned kserve-ci-e2e-test/precise-prefix-cache-test-kserve-5755bc98d9-f88w6 to ip-10-0-139-47.ec2.internal [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:55 kserve-ci-e2e-test multus Normal AddedInterface Add eth0 [10.134.0.22/23] from ovn-kubernetes [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:55 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Pulled Container image "quay.io/opendatahub/kserve-storage-initializer@sha256:13ad765c0f323200e1db7b45c99661287e65c5158d6b54b16139829636d20be4" already present on machine [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:55 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Created Created container: storage-initializer [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:55 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Started Started container storage-initializer [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:00 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Pulled Container image "ghcr.io/llm-d/llm-d-inference-sim:v0.8.2" already present on machine [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:01 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Created Created container: main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:01 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Started Started container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:23:10 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Killing Stopping container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 None kserve-ci-e2e-test Normal Scheduled Successfully assigned kserve-ci-e2e-test/precise-prefix-cache-test-kserve-5755bc98d9-hdpjr to ip-10-0-133-25.ec2.internal [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:55 kserve-ci-e2e-test multus Normal AddedInterface Add eth0 [10.132.0.45/23] from ovn-kubernetes [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:55 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Pulling Pulling image "quay.io/opendatahub/kserve-storage-initializer@sha256:13ad765c0f323200e1db7b45c99661287e65c5158d6b54b16139829636d20be4" [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:00 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Pulled Successfully pulled image "quay.io/opendatahub/kserve-storage-initializer@sha256:13ad765c0f323200e1db7b45c99661287e65c5158d6b54b16139829636d20be4" in 5.546s (5.546s including waiting). Image size: 301879255 bytes. [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:00 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Created Created container: storage-initializer [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:00 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Started Started container storage-initializer [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:07 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Pulling Pulling image "ghcr.io/llm-d/llm-d-inference-sim:v0.8.2" [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:08 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Pulled Successfully pulled image "ghcr.io/llm-d/llm-d-inference-sim:v0.8.2" in 1.434s (1.434s including waiting). Image size: 98346788 bytes. [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:09 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Created Created container: main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:09 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Started Started container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:23:10 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Killing Stopping container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:54 kserve-ci-e2e-test replicaset-controller Normal SuccessfulCreate Created pod: precise-prefix-cache-test-kserve-5755bc98d9-f88w6 [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:54 kserve-ci-e2e-test replicaset-controller Normal SuccessfulCreate Created pod: precise-prefix-cache-test-kserve-5755bc98d9-hdpjr [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:55 kserve-ci-e2e-test replicaset-controller Normal SuccessfulCreate Created pod: precise-prefix-cache-test-kserve-router-scheduler-647bfdf7jdncq [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 None kserve-ci-e2e-test Normal Scheduled Successfully assigned kserve-ci-e2e-test/precise-prefix-cache-test-kserve-router-scheduler-647bfdf7jdncq to ip-10-0-139-47.ec2.internal [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:55 kserve-ci-e2e-test multus Normal AddedInterface Add eth0 [10.134.0.23/23] from ovn-kubernetes [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:55 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Pulled Container image "quay.io/opendatahub/kserve-storage-initializer@sha256:13ad765c0f323200e1db7b45c99661287e65c5158d6b54b16139829636d20be4" already present on machine [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:55 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Created Created container: storage-initializer [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:55 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Started Started container storage-initializer [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:57 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Pulling Pulling image "ghcr.io/llm-d/llm-d-router-endpoint-picker:v0.9.0-rc.2" [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:58 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Pulled Successfully pulled image "ghcr.io/llm-d/llm-d-router-endpoint-picker:v0.9.0-rc.2" in 1.604s (1.604s including waiting). Image size: 75073927 bytes. [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:30 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Created Created container: main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:30 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Started Started container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:58 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Pulling Pulling image "ghcr.io/llm-d/llm-d-uds-tokenizer:vllm-v0.19.1" [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:29 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Pulled Successfully pulled image "ghcr.io/llm-d/llm-d-uds-tokenizer:vllm-v0.19.1" in 30.722s (30.722s including waiting). Image size: 2989890188 bytes. [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:29 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Created Created container: tokenizer [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:29 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Started Started container tokenizer [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:30 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Pulled Container image "ghcr.io/llm-d/llm-d-router-endpoint-picker:v0.9.0-rc.2" already present on machine [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:35 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Warning Unhealthy Startup probe failed: Get "http://10.134.0.23:8082/healthz": dial tcp 10.134.0.23:8082: connect: connection refused [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:46 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Warning Unhealthy Liveness probe failed: timeout: failed to connect service "10.134.0.23:9003" within 1s: context deadline exceeded [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:23:10 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Killing Stopping container tokenizer [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:23:10 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Killing Stopping container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:23:10 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Warning FailedPreStopHook PreStopHook failed [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:55 kserve-ci-e2e-test deployment-controller Normal ScalingReplicaSet Scaled up replica set precise-prefix-cache-test-kserve-router-scheduler-647bfdf79c from 0 to 1 [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:54 kserve-ci-e2e-test deployment-controller Normal ScalingReplicaSet Scaled up replica set precise-prefix-cache-test-kserve-5755bc98d9 from 0 to 2 [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:47 kserve-ci-e2e-test OpenDataHubModelController Warning ReconcileError Failed to reconcile LLMInferenceService: 1 error occurred: * failed to get HTTPRoute for AuthPolicy precise-prefix-cache-test-kserve-route-authn: failed to get HTTPRoute kserve-ci-e2e-test/precise-prefix-cache-test-kserve-route: HTTPRoute.gateway.networking.k8s.io "precise-prefix-cache-test-kserve-route" not found [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:54 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Secret kserve-ci-e2e-test/precise-prefix-cache-test-kserve-self-signed-certs [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:54 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Deployment kserve-ci-e2e-test/precise-prefix-cache-test-kserve [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:54 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Service kserve-ci-e2e-test/precise-prefix-cache-test-kserve-workload-svc [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:54 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.ServiceAccount kserve-ci-e2e-test/precise-prefix-cache-test-epp-sa [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:54 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.ClusterRoleBinding /kserve-ci-e2e-test-precise-prefix-cache-test-epp-auth-rb [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:54 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Role kserve-ci-e2e-test/precise-prefix-cache-test-epp-role [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:55 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.RoleBinding kserve-ci-e2e-test/precise-prefix-cache-test-epp-rb [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:55 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Deployment kserve-ci-e2e-test/precise-prefix-cache-test-kserve-router-scheduler [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:55 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Service kserve-ci-e2e-test/precise-prefix-cache-test-epp-service [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:26 kserve-ci-e2e-test LLMInferenceServiceController Normal Created (combined from similar events): Created v1.DestinationRule kserve-ci-e2e-test/precise-prefix-cache-test-kserve-shadow-svc [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:25 kserve-ci-e2e-test LLMInferenceServiceController Normal Updated Updated v1.Secret kserve-ci-e2e-test/precise-prefix-cache-test-kserve-self-signed-certs [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:26 kserve-ci-e2e-test LLMInferenceServiceController Normal Updated Updated v1.HTTPRoute kserve-ci-e2e-test/precise-prefix-cache-test-kserve-route [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:23:09 kserve-ci-e2e-test LLMInferenceServiceController Normal LLMInferenceServiceReady LLMInferenceService [precise-prefix-cache-test] is Ready [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:23:10 kserve-ci-e2e-test LLMInferenceServiceController Normal Deleted Deleted v1.ClusterRoleBinding /kserve-ci-e2e-test-precise-prefix-cache-test-epp-auth-rb [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 None kserve-ci-e2e-test Normal Scheduled Successfully assigned kserve-ci-e2e-test/router-gateway-1-openshift-default-799f46c59b-vt5f9 to ip-10-0-133-25.ec2.internal [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:36 kserve-ci-e2e-test multus Normal AddedInterface Add eth0 [10.132.0.44/23] from ovn-kubernetes [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:56 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Pulled Container image "registry.redhat.io/openshift-service-mesh/istio-proxyv2-rhel9@sha256:7d15cebf9b62f3f235c0eab5158ac8ff2fda86a1d193490dc94c301402c99da8" already present on machine [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:56 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Created Created container: istio-proxy [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:56 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Started Started container istio-proxy [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:54 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Warning Unhealthy Readiness probe failed: Get "http://10.132.0.44:15021/healthz/ready": context deadline exceeded (Client.Timeout exceeded while awaiting headers) [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:02 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Warning Unhealthy Startup probe failed: Get "http://10.132.0.44:15021/healthz/ready": dial tcp 10.132.0.44:15021: connect: connection refused [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:35 kserve-ci-e2e-test replicaset-controller Normal SuccessfulCreate Created pod: router-gateway-1-openshift-default-799f46c59b-vt5f9 [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:35 kserve-ci-e2e-test deployment-controller Normal ScalingReplicaSet Scaled up replica set router-gateway-1-openshift-default-799f46c59b from 0 to 1 [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:35 kserve-ci-e2e-test service-controller Normal EnsuringLoadBalancer Ensuring load balancer [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:40 kserve-ci-e2e-test service-controller Normal EnsuredLoadBalancer Ensured load balancer [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:30:52 kserve-ci-e2e-test gateway_labeler_controller Normal AddedLabel Added label istio.io/rev=openshift-gateway to gateway router-gateway-1 [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:148 ### Pod llmisvc-model-fb-opt-125m-with-ba4d693a-kserve-5b998d5789-t6wf4 (phase=Running) [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:188 #### init-container 'storage-initializer' (restarts=0) [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:201 # -- logs (current) -- [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:202 2026-07-07 14:45:56.986 1 storage.initializer INFO [initializer-entrypoint:():17] Initializing, args: (src_uri, dest_path): [('hf://facebook/opt-125m', '/mnt/models'), ('hf://edbeeching/opt-125m-lora', '/mnt/lora/lora-adapter-1')] [e2e-llm-inference-service] 2026-07-07 14:45:56.987 1 storage.initializer INFO [kserve_storage.py:download():166] Copying contents of hf://facebook/opt-125m to local [e2e-llm-inference-service] 2026-07-07 14:46:00.491 1 storage.initializer INFO [kserve_storage.py:download():234] Successfully copied hf://facebook/opt-125m to /mnt/models [e2e-llm-inference-service] 2026-07-07 14:46:00.491 1 storage.initializer INFO [kserve_storage.py:download():235] Model downloaded in 3.504402551000112 seconds. [e2e-llm-inference-service] 2026-07-07 14:46:00.491 1 storage.initializer INFO [kserve_storage.py:download():166] Copying contents of hf://edbeeching/opt-125m-lora to local [e2e-llm-inference-service] 2026-07-07 14:46:00.850 1 storage.initializer INFO [kserve_storage.py:download():234] Successfully copied hf://edbeeching/opt-125m-lora to /mnt/lora/lora-adapter-1 [e2e-llm-inference-service] 2026-07-07 14:46:00.850 1 storage.initializer INFO [kserve_storage.py:download():235] Model downloaded in 0.35927644799994596 seconds. [e2e-llm-inference-service] [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:188 #### container 'main' (restarts=0) [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:201 # -- logs (current) -- [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:202 (APIServer pid=1) DEBUG 07-07 14:59:36 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:59:36 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:59:37 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:59:38 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:59:39 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:59:40 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:59:41 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:59:42 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:59:43 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:59:44 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:59:45 [v1/metrics/loggers.py:259] Engine 000: Avg prompt throughput: 0.0 tokens/s, Avg generation throughput: 0.0 tokens/s, Running: 0 reqs, Waiting: 0 reqs, GPU KV cache usage: 0.0%, Prefix cache hit rate: 0.0% [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:59:45 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:59:46 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:59:46 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:59:47 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:59:48 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:59:49 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:59:50 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:59:51 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:59:52 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:59:53 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:59:54 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:59:55 [v1/metrics/loggers.py:259] Engine 000: Avg prompt throughput: 0.0 tokens/s, Avg generation throughput: 0.0 tokens/s, Running: 0 reqs, Waiting: 0 reqs, GPU KV cache usage: 0.0%, Prefix cache hit rate: 0.0% [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:59:55 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:59:56 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:59:56 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:59:57 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:59:58 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 14:59:59 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:00:00 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:00:01 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:00:02 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:00:03 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:00:04 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:00:05 [v1/metrics/loggers.py:259] Engine 000: Avg prompt throughput: 0.0 tokens/s, Avg generation throughput: 0.0 tokens/s, Running: 0 reqs, Waiting: 0 reqs, GPU KV cache usage: 0.0%, Prefix cache hit rate: 0.0% [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:00:05 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:00:06 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:00:06 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:00:07 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:00:08 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:00:09 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:00:10 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:00:11 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:00:12 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:00:13 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:00:14 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:00:15 [v1/metrics/loggers.py:259] Engine 000: Avg prompt throughput: 0.0 tokens/s, Avg generation throughput: 0.0 tokens/s, Running: 0 reqs, Waiting: 0 reqs, GPU KV cache usage: 0.0%, Prefix cache hit rate: 0.0% [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:00:15 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:00:16 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:00:16 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:00:17 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:00:18 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:00:19 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:00:20 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:00:21 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:00:22 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:00:23 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:00:24 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:00:25 [v1/metrics/loggers.py:259] Engine 000: Avg prompt throughput: 0.0 tokens/s, Avg generation throughput: 0.0 tokens/s, Running: 0 reqs, Waiting: 0 reqs, GPU KV cache usage: 0.0%, Prefix cache hit rate: 0.0% [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:00:25 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:00:26 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:00:26 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:00:27 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:00:28 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:00:29 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:00:30 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:00:31 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:00:32 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:00:33 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:00:34 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:00:35 [v1/metrics/loggers.py:259] Engine 000: Avg prompt throughput: 0.0 tokens/s, Avg generation throughput: 0.0 tokens/s, Running: 0 reqs, Waiting: 0 reqs, GPU KV cache usage: 0.0%, Prefix cache hit rate: 0.0% [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:00:35 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:00:36 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:00:36 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:00:37 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:00:38 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:00:39 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:00:40 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:00:41 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:00:42 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:00:43 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:00:44 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:00:45 [v1/metrics/loggers.py:259] Engine 000: Avg prompt throughput: 0.0 tokens/s, Avg generation throughput: 0.0 tokens/s, Running: 0 reqs, Waiting: 0 reqs, GPU KV cache usage: 0.0%, Prefix cache hit rate: 0.0% [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:00:45 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:00:46 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:00:46 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:00:47 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:00:48 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:00:49 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:00:50 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:00:51 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:00:52 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:00:53 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:00:54 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:00:55 [v1/metrics/loggers.py:259] Engine 000: Avg prompt throughput: 0.0 tokens/s, Avg generation throughput: 0.0 tokens/s, Running: 0 reqs, Waiting: 0 reqs, GPU KV cache usage: 0.0%, Prefix cache hit rate: 0.0% [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:00:55 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:00:56 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:00:56 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:00:57 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:00:58 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:00:59 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:01:00 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:01:01 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:01:02 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:01:03 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:01:04 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:01:05 [v1/metrics/loggers.py:259] Engine 000: Avg prompt throughput: 0.0 tokens/s, Avg generation throughput: 0.0 tokens/s, Running: 0 reqs, Waiting: 0 reqs, GPU KV cache usage: 0.0%, Prefix cache hit rate: 0.0% [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:01:05 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:01:06 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:01:06 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:01:07 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:01:08 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:01:09 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:01:10 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:01:11 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:01:12 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:01:13 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:01:14 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:01:15 [v1/metrics/loggers.py:259] Engine 000: Avg prompt throughput: 0.0 tokens/s, Avg generation throughput: 0.0 tokens/s, Running: 0 reqs, Waiting: 0 reqs, GPU KV cache usage: 0.0%, Prefix cache hit rate: 0.0% [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:01:15 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:01:16 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:01:16 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:01:17 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:01:18 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:01:19 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:01:20 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:01:21 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:01:22 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:01:23 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:01:24 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:01:25 [v1/metrics/loggers.py:259] Engine 000: Avg prompt throughput: 0.0 tokens/s, Avg generation throughput: 0.0 tokens/s, Running: 0 reqs, Waiting: 0 reqs, GPU KV cache usage: 0.0%, Prefix cache hit rate: 0.0% [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:01:25 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:01:26 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:01:26 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:01:27 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:01:28 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:01:29 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:01:30 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:01:31 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:01:32 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:01:33 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:01:34 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:01:35 [v1/metrics/loggers.py:259] Engine 000: Avg prompt throughput: 0.0 tokens/s, Avg generation throughput: 0.0 tokens/s, Running: 0 reqs, Waiting: 0 reqs, GPU KV cache usage: 0.0%, Prefix cache hit rate: 0.0% [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:01:35 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:01:36 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:01:36 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:01:37 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:01:38 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:01:39 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:01:40 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:01:41 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:01:42 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:01:43 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:01:44 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:01:45 [v1/metrics/loggers.py:259] Engine 000: Avg prompt throughput: 0.0 tokens/s, Avg generation throughput: 0.0 tokens/s, Running: 0 reqs, Waiting: 0 reqs, GPU KV cache usage: 0.0%, Prefix cache hit rate: 0.0% [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:01:45 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:01:46 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:01:46 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:01:47 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:01:48 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:01:49 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:01:50 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:01:51 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:01:52 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:01:53 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:01:54 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:01:55 [v1/metrics/loggers.py:259] Engine 000: Avg prompt throughput: 0.0 tokens/s, Avg generation throughput: 0.0 tokens/s, Running: 0 reqs, Waiting: 0 reqs, GPU KV cache usage: 0.0%, Prefix cache hit rate: 0.0% [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:01:55 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:01:56 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:01:56 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:01:57 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:01:58 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:01:59 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:02:00 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:02:01 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:02:02 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:02:03 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:02:04 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:02:05 [v1/metrics/loggers.py:259] Engine 000: Avg prompt throughput: 0.0 tokens/s, Avg generation throughput: 0.0 tokens/s, Running: 0 reqs, Waiting: 0 reqs, GPU KV cache usage: 0.0%, Prefix cache hit rate: 0.0% [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:02:05 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:02:06 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:02:06 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:02:07 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:02:08 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:02:09 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:02:10 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:02:11 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:02:12 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:02:13 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:02:14 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:02:15 [v1/metrics/loggers.py:259] Engine 000: Avg prompt throughput: 0.0 tokens/s, Avg generation throughput: 0.0 tokens/s, Running: 0 reqs, Waiting: 0 reqs, GPU KV cache usage: 0.0%, Prefix cache hit rate: 0.0% [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:02:15 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:02:16 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:02:16 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:02:17 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:02:18 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:02:19 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:02:20 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:02:21 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:02:22 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:148 ### Pod llmisvc-model-fb-opt-125m-with-ba4d693a-kserve-router-sche4ghr9 (phase=Running) [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:188 #### init-container 'storage-initializer' (restarts=0) [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:201 # -- logs (current) -- [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:202 2026-07-07 14:45:57.257 1 storage.initializer INFO [initializer-entrypoint:():17] Initializing, args: (src_uri, dest_path): [('hf://facebook/opt-125m', '/mnt/models')] [e2e-llm-inference-service] 2026-07-07 14:45:57.257 1 storage.initializer INFO [kserve_storage.py:download():166] Copying contents of hf://facebook/opt-125m to local [e2e-llm-inference-service] 2026-07-07 14:45:57.257 1 storage.initializer INFO [kserve_storage.py:download():169] Allow patterns: ['tokenizer.json', 'tokenizer_config.json', 'special_tokens_map.json', 'vocab.json', 'merges.txt', 'config.json', 'generation_config.json'] [e2e-llm-inference-service] Could not set the permissions on the file '/mnt/models/.cache/huggingface/download/8_PA_wEVGiVa2goH2H4KQOQpvVY=.b3fb716a3024261980becb2382e31a3780985130.incomplete'. Error: [Errno 13] Permission denied: '/mnt/tmp_8974577d-59f3-425b-8729-dcdfcece1472'. [e2e-llm-inference-service] Continuing without setting permissions. [e2e-llm-inference-service] Could not set the permissions on the file '/mnt/models/.cache/huggingface/download/3EVKVggOldJcKSsGjSdoUCN1AyQ=.cf739e3ba86db7791ebab2828cc34b8a5acd3a86.incomplete'. Error: [Errno 13] Permission denied: '/mnt/tmp_71fe50b8-8e09-43d5-a119-98b92cde23a7'. [e2e-llm-inference-service] Continuing without setting permissions. [e2e-llm-inference-service] Could not set the permissions on the file '/mnt/models/.cache/huggingface/download/PtHk0z_I45atnj23IIRhTExwT3w=.226b0752cac7789c48f0cb3ec53eda48b7be36cc.incomplete'. Error: [Errno 13] Permission denied: '/mnt/tmp_4903002a-449a-4cbd-9ff4-9361e5c6578c'. [e2e-llm-inference-service] Continuing without setting permissions. [e2e-llm-inference-service] Could not set the permissions on the file '/mnt/models/.cache/huggingface/download/ahkChHUJFxEmOdq5GDFEmerRzCY=.5dfa36546b8eddce0e04df3133c30df43fcc3828.incomplete'. Error: [Errno 13] Permission denied: '/mnt/tmp_d063cd02-94e9-4362-a074-5091f0bdfef5'. [e2e-llm-inference-service] Continuing without setting permissions. [e2e-llm-inference-service] Could not set the permissions on the file '/mnt/models/.cache/huggingface/download/vzaExXFZNBay89bvlQv-ZcI6BTg=.27c24ca9d908d0b678b20c698aeb9e950c44d865.incomplete'. Error: [Errno 13] Permission denied: '/mnt/tmp_2bbc1d00-73cb-4500-95a3-7a9f6793f4ea'. [e2e-llm-inference-service] Continuing without setting permissions. [e2e-llm-inference-service] Could not set the permissions on the file '/mnt/models/.cache/huggingface/download/j3m-Hy6QvBddw8RXA1uSWl1AJ0c=.0a39732b2d8be8e493cab3da68b68cc3e28221de.incomplete'. Error: [Errno 13] Permission denied: '/mnt/tmp_876c00f3-3996-4af3-995a-107e3d1f50cd'. [e2e-llm-inference-service] Continuing without setting permissions. [e2e-llm-inference-service] 2026-07-07 14:45:57.689 1 storage.initializer INFO [kserve_storage.py:download():234] Successfully copied hf://facebook/opt-125m to /mnt/models [e2e-llm-inference-service] 2026-07-07 14:45:57.690 1 storage.initializer INFO [kserve_storage.py:download():235] Model downloaded in 0.43240728599994327 seconds. [e2e-llm-inference-service] [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:188 #### container 'main' (restarts=0) [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:201 # -- logs (current) -- [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:202 {"level":"info","ts":1783435557.8697639,"logger":"setup","caller":"runner/runner.go:196","msg":"GIE build","commit-sha":"181aa8358916e19b8844ccc752b2d6153d4b2ad6","build-ref":"v0.9.0-rc.2"} [e2e-llm-inference-service] Flag --model-server-metrics-scheme has been deprecated, This flag is deprecated. Configure via EndpointPickerConfig data layer plugin parameters instead. [e2e-llm-inference-service] {"level":"info","ts":1783435557.8699422,"logger":"setup","caller":"runner/runner.go:217","msg":"Flags processed","flags":{"cert-path":"/var/run/kserve/tls","config-file":"","config-text":"apiVersion: inference.networking.x-k8s.io/v1alpha1\nkind: EndpointPickerConfig\nplugins:\n- type: single-profile-handler\n- type: queue-scorer\n- type: prefix-cache-scorer\n- type: max-score-picker\n- type: lora-affinity-scorer\nschedulingProfiles:\n- name: default\n plugins:\n - pluginRef: lora-affinity-scorer\n weight: 4\n - pluginRef: queue-scorer\n weight: 2\n - pluginRef: prefix-cache-scorer\n weight: 3\n - pluginRef: max-score-picker\n","disable-endpoint-subset-filter":false,"enable-cert-reload":true,"enable-grpc-stream-metrics":false,"enable-pprof":true,"endpoint-selector":"","endpoint-target-ports":{},"grpc-health-port":9003,"grpc-max-recv-msg-size":"","grpc-max-send-msg-size":"","grpc-port":9002,"ha-enable-leader-election":false,"health-checking":false,"metrics-endpoint-auth":true,"metrics-port":9090,"metrics-staleness-threshold":2000000000,"model-server-metrics-https-insecure-skip-verify":true,"model-server-metrics-path":"/metrics","model-server-metrics-port":0,"model-server-metrics-scheme":"https","pool-group":"inference.networking.k8s.io","pool-name":"llmisvc-model-fb-opt-125m-with-ba4d693a-inference-pool","pool-namespace":"kserve-ci-e2e-test","refresh-metrics-interval":50000000,"refresh-prometheus-metrics-interval":5000000000,"secure-serving":true,"tracing":true,"v":2,"zap-devel":{},"zap-encoder":{},"zap-log-level":{},"zap-stacktrace-level":{},"zap-time-encoding":{}}} [e2e-llm-inference-service] {"level":"info","ts":1783435557.8700616,"logger":"setup.trace","caller":"tracing/telemetry.go:123","msg":"init OTel trace exporter","type":"console"} [e2e-llm-inference-service] {"level":"info","ts":1783435557.8704958,"caller":"loader/configloader.go:89","msg":"DEPRECATION: apiVersion inference.networking.x-k8s.io/v1alpha1/EndpointPickerConfig is deprecated","replacement":"llm-d.ai/v1alpha1/EndpointPickerConfig"} [e2e-llm-inference-service] {"level":"info","ts":1783435557.8705323,"caller":"loader/configloader.go:121","msg":"Loaded raw configuration","config":"{Plugins: [{Type: single-profile-handler} {Type: queue-scorer} {Type: prefix-cache-scorer} {Type: max-score-picker} {Type: lora-affinity-scorer}], SchedulingProfiles: [{Name: default, Plugins: [{PluginRef: lora-affinity-scorer, Weight: 4.00} {PluginRef: queue-scorer, Weight: 2.00} {PluginRef: prefix-cache-scorer, Weight: 3.00} {PluginRef: max-score-picker}]}]}"} [e2e-llm-inference-service] {"level":"info","ts":1783435557.8705447,"logger":"setup","caller":"runner/runner.go:622","msg":"Data layer: ENABLED"} [e2e-llm-inference-service] {"level":"info","ts":1783435557.870867,"logger":"setup","caller":"runner/runner.go:281","msg":"Raw config after phase one","config":{"apiVersion":"inference.networking.x-k8s.io/v1alpha1","dataLayer":null,"kind":"EndpointPickerConfig","plugins":[{"name":"single-profile-handler","parameters":null,"type":"single-profile-handler"},{"name":"queue-scorer","parameters":null,"type":"queue-scorer"},{"name":"prefix-cache-scorer","parameters":null,"type":"prefix-cache-scorer"},{"name":"max-score-picker","parameters":null,"type":"max-score-picker"},{"name":"lora-affinity-scorer","parameters":null,"type":"lora-affinity-scorer"}],"schedulingProfiles":[{"name":"default","plugins":[{"pluginRef":"lora-affinity-scorer","weight":4},{"pluginRef":"queue-scorer","weight":2},{"pluginRef":"prefix-cache-scorer","weight":3},{"pluginRef":"max-score-picker","weight":null}]}]}} [e2e-llm-inference-service] {"level":"info","ts":1783435557.8918867,"logger":"utilization-detector/utilization-detector","caller":"utilization/detector.go:83","msg":"Creating new UtilizationDetector","queueDepthThreshold":5,"kvCacheUtilThreshold":0.8,"metricsStalenessThreshold":"200ms","headroom":0} [e2e-llm-inference-service] {"level":"info","ts":1783435557.8919723,"caller":"metrics/factories.go:230","msg":"Registered engine mapping","engine":"vllm","mapping":"Mapping{all specs enabled}"} [e2e-llm-inference-service] {"level":"info","ts":1783435557.8920102,"caller":"metrics/factories.go:230","msg":"Registered engine mapping","engine":"sglang","mapping":"Mapping{disabled: [lora]}"} [e2e-llm-inference-service] {"level":"info","ts":1783435557.8920474,"caller":"metrics/factories.go:230","msg":"Registered engine mapping","engine":"trtllm-serve","mapping":"Mapping{disabled: [lora, cacheInfo]}"} [e2e-llm-inference-service] {"level":"info","ts":1783435557.8921108,"caller":"metrics/factories.go:230","msg":"Registered engine mapping","engine":"triton-tensorrt-llm","mapping":"Mapping{disabled: [lora, cacheInfo]}"} [e2e-llm-inference-service] {"level":"info","ts":1783435557.8921313,"caller":"metrics/factories.go:230","msg":"Registered engine mapping","engine":"triton","mapping":"Mapping{disabled: [kv, lora, cacheInfo]}"} [e2e-llm-inference-service] {"level":"info","ts":1783435557.8921897,"caller":"loader/configloader.go:154","msg":"Instantiated all plugins and applied system defaults. Effective raw configuration","config":"{Plugins: [{Name: single-profile-handler, Type: single-profile-handler} {Name: queue-scorer, Type: queue-scorer} {Name: prefix-cache-scorer, Type: prefix-cache-scorer} {Name: max-score-picker, Type: max-score-picker} {Name: lora-affinity-scorer, Type: lora-affinity-scorer} {Name: fcfs-ordering-policy, Type: fcfs-ordering-policy} {Name: global-strict-fairness-policy, Type: global-strict-fairness-policy} {Name: static-usage-limit-policy, Type: static-usage-limit-policy} {Name: openai-parser, Type: openai-parser} {Name: anthropic-parser, Type: anthropic-parser} {Name: vllmhttp-parser, Type: vllmhttp-parser} {Name: utilization-detector, Type: utilization-detector} {Name: metrics-data-source, Type: metrics-data-source} {Name: core-metrics-extractor, Type: core-metrics-extractor}], SchedulingProfiles: [{Name: default, Plugins: [{PluginRef: lora-affinity-scorer, Weight: 4.00} {PluginRef: queue-scorer, Weight: 2.00} {PluginRef: prefix-cache-scorer, Weight: 3.00} {PluginRef: max-score-picker}]}], DataLayer: {Sources: [{PluginRef: metrics-data-source, Extractors: [{PluginRef: core-metrics-extractor}]}], Discovery: }, FlowControl: {MaxBytes: unlimited, MaxRequests: unlimited, SaturationDetector: {PluginRef: utilization-detector}}, RequestHandler: {Parsers: [{PluginRef: openai-parser}, {PluginRef: anthropic-parser}, {PluginRef: vllmhttp-parser}]}}"} [e2e-llm-inference-service] {"level":"info","ts":1783435557.8922462,"caller":"approximateprefix/plugin.go:88","msg":"Prefix DataProducer initialized","config":{"autoTune":true,"blockSizeTokens":16,"blockSize":0,"maxPrefixBlocksToMatch":2048,"maxPrefixTokensToMatch":131072,"lruCapacityPerServer":31250}} [e2e-llm-inference-service] {"level":"info","ts":1783435557.8923209,"caller":"approximateprefix/plugin.go:111","msg":"WARNING: configured blockSizeTokens is below the recommended minimum, overriding it.","blockSizeTokens":16,"minimum":64,"issue":"https://github.com/llm-d/llm-d-router/issues/1158"} [e2e-llm-inference-service] {"level":"info","ts":1783435557.8923454,"caller":"datalayer/data_graph.go:116","msg":"auto-created default producer","producer":"approx-prefix-cache-producer/approx-prefix-cache-producer","dataKey":"PrefixCacheMatchInfoDataKey/approx-prefix-cache-producer","consumer":"prefix-cache-scorer"} [e2e-llm-inference-service] {"level":"info","ts":1783435557.8923733,"caller":"datalayer/data_graph.go:116","msg":"auto-created default producer","producer":"token-producer/token-producer","dataKey":"TokenizedPrompt/token-producer","consumer":"approx-prefix-cache-producer"} [e2e-llm-inference-service] {"level":"info","ts":1783435557.8924885,"caller":"runner/runner.go:685","msg":"loaded configuration from file/text successfully"} [e2e-llm-inference-service] {"level":"info","ts":1783435557.8924994,"logger":"setup","caller":"runner/runner.go:308","msg":"EPP config after phase two","config":"{SchedulerConfig:{ProfileHandler: single-profile-handler/single-profile-handler, Profiles: map[default:{Filters: [], Scorers: [lora-affinity-scorer/lora-affinity-scorer: 4.000000, queue-scorer/queue-scorer: 2.000000, prefix-cache-scorer/prefix-cache-scorer: 3.000000], Picker: max-score-picker/max-score-picker}]} SaturationDetector:0xc000325d00 DataConfig:{Sources:[{Plugin:0xc000534cf0 Extractors:[0xc00081e140]}]} FlowControlConfig: ParserRegistry:0xc00081e600}"} [e2e-llm-inference-service] {"level":"info","ts":1783435557.9121385,"logger":"setup","caller":"runner/runner.go:352","msg":"Setting pprof handlers"} [e2e-llm-inference-service] {"level":"info","ts":1783435557.9121761,"caller":"manager/internal.go:201","msg":"Registering metrics http server extra handler","path":"/debug/pprof/cmdline"} [e2e-llm-inference-service] {"level":"info","ts":1783435557.9121902,"caller":"manager/internal.go:201","msg":"Registering metrics http server extra handler","path":"/debug/pprof/profile"} [e2e-llm-inference-service] {"level":"info","ts":1783435557.9121957,"caller":"manager/internal.go:201","msg":"Registering metrics http server extra handler","path":"/debug/pprof/trace"} [e2e-llm-inference-service] {"level":"info","ts":1783435557.9122,"caller":"manager/internal.go:201","msg":"Registering metrics http server extra handler","path":"/debug/pprof/heap"} [e2e-llm-inference-service] {"level":"info","ts":1783435557.9122045,"caller":"manager/internal.go:201","msg":"Registering metrics http server extra handler","path":"/debug/pprof/goroutine"} [e2e-llm-inference-service] {"level":"info","ts":1783435557.912209,"caller":"manager/internal.go:201","msg":"Registering metrics http server extra handler","path":"/debug/pprof/threadcreate"} [e2e-llm-inference-service] {"level":"info","ts":1783435557.9122136,"caller":"manager/internal.go:201","msg":"Registering metrics http server extra handler","path":"/debug/pprof/block"} [e2e-llm-inference-service] {"level":"info","ts":1783435557.912218,"caller":"manager/internal.go:201","msg":"Registering metrics http server extra handler","path":"/debug/pprof/"} [e2e-llm-inference-service] {"level":"info","ts":1783435557.9122233,"caller":"manager/internal.go:201","msg":"Registering metrics http server extra handler","path":"/debug/pprof/symbol"} [e2e-llm-inference-service] {"level":"info","ts":1783435557.9122276,"caller":"manager/internal.go:201","msg":"Registering metrics http server extra handler","path":"/debug/pprof/allocs"} [e2e-llm-inference-service] {"level":"info","ts":1783435557.9122322,"caller":"manager/internal.go:201","msg":"Registering metrics http server extra handler","path":"/debug/pprof/mutex"} [e2e-llm-inference-service] {"level":"info","ts":1783435557.9122388,"caller":"manager/internal.go:201","msg":"Registering metrics http server extra handler","path":"/debug/plugins/state"} [e2e-llm-inference-service] {"level":"info","ts":1783435557.9122436,"logger":"setup","caller":"runner/runner.go:373","msg":"parsed config","scheduler-config":"{ProfileHandler: single-profile-handler/single-profile-handler, Profiles: map[default:{Filters: [], Scorers: [lora-affinity-scorer/lora-affinity-scorer: 4.000000, queue-scorer/queue-scorer: 2.000000, prefix-cache-scorer/prefix-cache-scorer: 3.000000], Picker: max-score-picker/max-score-picker}]}"} [e2e-llm-inference-service] {"level":"info","ts":1783435557.9122725,"logger":"setup","caller":"datalayer/runtime.go:99","msg":"Configuring datalayer runtime","numSources":1} [e2e-llm-inference-service] {"level":"info","ts":1783435557.9122808,"logger":"setup","caller":"datalayer/runtime.go:118","msg":"Processing source","source":"metrics-data-source","numExtractors":1} [e2e-llm-inference-service] {"level":"info","ts":1783435557.9122953,"logger":"setup","caller":"datalayer/runtime.go:147","msg":"Source configured","source":"metrics-data-source","extractors":["core-metrics-extractor/core-metrics-extractor"]} [e2e-llm-inference-service] {"level":"info","ts":1783435557.9123068,"logger":"setup","caller":"datalayer/runtime.go:206","msg":"Datalayer runtime configured","pollers":1,"notifiers":0,"endpointSources":0} [e2e-llm-inference-service] {"level":"info","ts":1783435557.9123156,"logger":"setup","caller":"runner/runner.go:833","msg":"Experimental Flow Control layer is disabled, using legacy admission control"} [e2e-llm-inference-service] {"level":"info","ts":1783435557.9124043,"logger":"setup","caller":"runner/runner.go:721","msg":"ExtProc server runner added to manager."} [e2e-llm-inference-service] {"level":"info","ts":1783435557.9124181,"logger":"setup","caller":"runner/runner.go:260","msg":"Controller manager starting"} [e2e-llm-inference-service] {"level":"info","ts":1783435557.9124434,"logger":"controller-runtime.metrics","caller":"server/server.go:208","msg":"Starting metrics server"} [e2e-llm-inference-service] {"level":"info","ts":1783435557.913001,"logger":"controller-runtime.metrics","caller":"server/server.go:247","msg":"Serving metrics server","bindAddress":":9090","secure":false} [e2e-llm-inference-service] {"level":"info","ts":1783435557.9130688,"caller":"runnable/grpc.go:35","msg":"gRPC server starting","name":"health"} [e2e-llm-inference-service] {"level":"info","ts":1783435557.9131496,"caller":"runnable/grpc.go:43","msg":"gRPC server listening","name":"health","port":9003} [e2e-llm-inference-service] {"level":"info","ts":1783435557.9137008,"caller":"controller/controller.go:370","msg":"Starting EventSource","controller":"inferencemodelrewrite","controllerGroup":"inference.networking.x-k8s.io","controllerKind":"InferenceModelRewrite","source":"kind source: *v1alpha2.InferenceModelRewrite"} [e2e-llm-inference-service] {"level":"info","ts":1783435557.9137294,"caller":"controller/controller.go:370","msg":"Starting EventSource","controller":"pod","controllerGroup":"","controllerKind":"Pod","source":"kind source: *v1.Pod"} [e2e-llm-inference-service] {"level":"info","ts":1783435557.9137223,"caller":"controller/controller.go:370","msg":"Starting EventSource","controller":"inferencepool","controllerGroup":"inference.networking.k8s.io","controllerKind":"InferencePool","source":"kind source: *v1.InferencePool"} [e2e-llm-inference-service] {"level":"info","ts":1783435557.9139583,"caller":"controller/controller.go:370","msg":"Starting EventSource","controller":"inferenceobjective","controllerGroup":"inference.networking.x-k8s.io","controllerKind":"InferenceObjective","source":"kind source: *v1alpha2.InferenceObjective"} [e2e-llm-inference-service] {"level":"info","ts":1783435557.9145608,"caller":"runnable/grpc.go:35","msg":"gRPC server starting","name":"ext-proc"} [e2e-llm-inference-service] {"level":"info","ts":1783435557.9146981,"caller":"runnable/grpc.go:43","msg":"gRPC server listening","name":"ext-proc","port":9002} [e2e-llm-inference-service] {"level":"info","ts":1783435557.9194705,"logger":"controller-runtime.cache","caller":"cache/reflector.go:446","msg":"Caches populated","type":"*v1alpha2.InferenceObjective","reflector":"pkg/mod/k8s.io/client-go@v0.35.6/tools/cache/reflector.go:289"} [e2e-llm-inference-service] {"level":"info","ts":1783435557.9194987,"logger":"controller-runtime.cache","caller":"cache/reflector.go:446","msg":"Caches populated","type":"*v1alpha2.InferenceModelRewrite","reflector":"pkg/mod/k8s.io/client-go@v0.35.6/tools/cache/reflector.go:289"} [e2e-llm-inference-service] {"level":"info","ts":1783435557.9197295,"logger":"controller-runtime.cache","caller":"cache/reflector.go:446","msg":"Caches populated","type":"*v1.InferencePool","reflector":"pkg/mod/k8s.io/client-go@v0.35.6/tools/cache/reflector.go:289"} [e2e-llm-inference-service] {"level":"info","ts":1783435557.921627,"logger":"controller-runtime.cache","caller":"cache/reflector.go:446","msg":"Caches populated","type":"*v1.Pod","reflector":"pkg/mod/k8s.io/client-go@v0.35.6/tools/cache/reflector.go:289"} [e2e-llm-inference-service] {"level":"info","ts":1783435558.0157726,"caller":"controller/controller.go:303","msg":"Starting Controller","controller":"inferencemodelrewrite","controllerGroup":"inference.networking.x-k8s.io","controllerKind":"InferenceModelRewrite"} [e2e-llm-inference-service] {"level":"info","ts":1783435558.0158186,"caller":"controller/controller.go:306","msg":"Starting workers","controller":"inferencemodelrewrite","controllerGroup":"inference.networking.x-k8s.io","controllerKind":"InferenceModelRewrite","worker count":1} [e2e-llm-inference-service] {"level":"info","ts":1783435558.015817,"caller":"controller/controller.go:303","msg":"Starting Controller","controller":"inferenceobjective","controllerGroup":"inference.networking.x-k8s.io","controllerKind":"InferenceObjective"} [e2e-llm-inference-service] {"level":"info","ts":1783435558.0158398,"caller":"controller/controller.go:306","msg":"Starting workers","controller":"inferenceobjective","controllerGroup":"inference.networking.x-k8s.io","controllerKind":"InferenceObjective","worker count":1} [e2e-llm-inference-service] {"level":"info","ts":1783435558.0158548,"caller":"controller/controller.go:303","msg":"Starting Controller","controller":"inferencepool","controllerGroup":"inference.networking.k8s.io","controllerKind":"InferencePool"} [e2e-llm-inference-service] {"level":"info","ts":1783435558.0158808,"caller":"controller/controller.go:306","msg":"Starting workers","controller":"inferencepool","controllerGroup":"inference.networking.k8s.io","controllerKind":"InferencePool","worker count":1} [e2e-llm-inference-service] {"level":"info","ts":1783435558.0160408,"caller":"controller/inferencepool_reconciler.go:46","msg":"Reconciling InferencePool","controller":"inferencepool","controllerGroup":"inference.networking.k8s.io","controllerKind":"InferencePool","InferencePool":{"name":"llmisvc-model-fb-opt-125m-with-ba4d693a-inference-pool","namespace":"kserve-ci-e2e-test"},"namespace":"kserve-ci-e2e-test","name":"llmisvc-model-fb-opt-125m-with-ba4d693a-inference-pool","reconcileID":"1bdb4988-cfda-4724-ae12-10c204d776a3"} [e2e-llm-inference-service] {"level":"info","ts":1783435558.1148396,"caller":"controller/controller.go:303","msg":"Starting Controller","controller":"pod","controllerGroup":"","controllerKind":"Pod"} [e2e-llm-inference-service] {"level":"info","ts":1783435558.1148746,"caller":"controller/controller.go:306","msg":"Starting workers","controller":"pod","controllerGroup":"","controllerKind":"Pod","worker count":1} [e2e-llm-inference-service] {"level":"info","ts":1783435576.977578,"caller":"controller/inferencepool_reconciler.go:46","msg":"Reconciling InferencePool","controller":"inferencepool","controllerGroup":"inference.networking.k8s.io","controllerKind":"InferencePool","InferencePool":{"name":"llmisvc-model-fb-opt-125m-with-ba4d693a-inference-pool","namespace":"kserve-ci-e2e-test"},"namespace":"kserve-ci-e2e-test","name":"llmisvc-model-fb-opt-125m-with-ba4d693a-inference-pool","reconcileID":"ab23a81b-3405-433b-a31c-c6de6fd228b2"} [e2e-llm-inference-service] {"level":"info","ts":1783435636.5951805,"caller":"controller/pod_reconciler.go:99","msg":"Pod already exists","controller":"pod","controllerGroup":"","controllerKind":"Pod","Pod":{"name":"llmisvc-model-fb-opt-125m-with-ba4d693a-kserve-5b998d5789-t6wf4","namespace":"kserve-ci-e2e-test"},"namespace":"kserve-ci-e2e-test","name":"llmisvc-model-fb-opt-125m-with-ba4d693a-kserve-5b998d5789-t6wf4","reconcileID":"c03da43b-7840-4124-9286-cee7e1378723"} [e2e-llm-inference-service] [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1254 --- [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1255 metadata: [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-ba4d693a-epp-service [e2e-llm-inference-service] namespace: kserve-ci-e2e-test [e2e-llm-inference-service] uid: 5cb29ef2-232f-4489-bcec-df2984218fec [e2e-llm-inference-service] resourceVersion: '46619' [e2e-llm-inference-service] creationTimestamp: '2026-07-07T14:45:56Z' [e2e-llm-inference-service] labels: [e2e-llm-inference-service] app.kubernetes.io/component: llminferenceservice-router-scheduler [e2e-llm-inference-service] app.kubernetes.io/name: llmisvc-model-fb-opt-125m-with-ba4d693a [e2e-llm-inference-service] app.kubernetes.io/part-of: llminferenceservice [e2e-llm-inference-service] endpoints.kubernetes.io/managed-by: endpoint-controller [e2e-llm-inference-service] annotations: [e2e-llm-inference-service] endpoints.kubernetes.io/last-change-trigger-time: '2026-07-07T14:46:29Z' [e2e-llm-inference-service] managedFields: [e2e-llm-inference-service] - manager: kube-controller-manager [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] apiVersion: v1 [e2e-llm-inference-service] time: '2026-07-07T14:46:29Z' [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:metadata: [e2e-llm-inference-service] f:annotations: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:endpoints.kubernetes.io/last-change-trigger-time: {} [e2e-llm-inference-service] f:labels: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:app.kubernetes.io/component: {} [e2e-llm-inference-service] f:app.kubernetes.io/name: {} [e2e-llm-inference-service] f:app.kubernetes.io/part-of: {} [e2e-llm-inference-service] f:endpoints.kubernetes.io/managed-by: {} [e2e-llm-inference-service] f:subsets: {} [e2e-llm-inference-service] subsets: [e2e-llm-inference-service] - addresses: [e2e-llm-inference-service] - ip: 10.132.0.54 [e2e-llm-inference-service] nodeName: ip-10-0-133-25.ec2.internal [e2e-llm-inference-service] targetRef: [e2e-llm-inference-service] kind: Pod [e2e-llm-inference-service] namespace: kserve-ci-e2e-test [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-ba4d693a-kserve-router-sche4ghr9 [e2e-llm-inference-service] uid: 66635d02-cb85-4e6d-8342-c235b66670ec [e2e-llm-inference-service] ports: [e2e-llm-inference-service] - name: grpc-health [e2e-llm-inference-service] port: 9003 [e2e-llm-inference-service] protocol: TCP [e2e-llm-inference-service] - name: metrics [e2e-llm-inference-service] port: 9090 [e2e-llm-inference-service] protocol: TCP [e2e-llm-inference-service] - name: zmq [e2e-llm-inference-service] port: 5557 [e2e-llm-inference-service] protocol: TCP [e2e-llm-inference-service] - name: grpc [e2e-llm-inference-service] port: 9002 [e2e-llm-inference-service] protocol: TCP [e2e-llm-inference-service] apiVersion: v1 [e2e-llm-inference-service] kind: Endpoints [e2e-llm-inference-service] [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1254 --- [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1255 metadata: [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-ba4d693a-kserve-workload-svc [e2e-llm-inference-service] namespace: kserve-ci-e2e-test [e2e-llm-inference-service] uid: 8fe2da55-aba1-42a3-8acd-b8be2a34d5ae [e2e-llm-inference-service] resourceVersion: '47361' [e2e-llm-inference-service] creationTimestamp: '2026-07-07T14:45:56Z' [e2e-llm-inference-service] labels: [e2e-llm-inference-service] app.kubernetes.io/component: llminferenceservice-workload [e2e-llm-inference-service] app.kubernetes.io/name: llmisvc-model-fb-opt-125m-with-ba4d693a [e2e-llm-inference-service] app.kubernetes.io/part-of: llminferenceservice [e2e-llm-inference-service] endpoints.kubernetes.io/managed-by: endpoint-controller [e2e-llm-inference-service] annotations: [e2e-llm-inference-service] endpoints.kubernetes.io/last-change-trigger-time: '2026-07-07T14:47:16Z' [e2e-llm-inference-service] managedFields: [e2e-llm-inference-service] - manager: kube-controller-manager [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] apiVersion: v1 [e2e-llm-inference-service] time: '2026-07-07T14:47:16Z' [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:metadata: [e2e-llm-inference-service] f:annotations: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:endpoints.kubernetes.io/last-change-trigger-time: {} [e2e-llm-inference-service] f:labels: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:app.kubernetes.io/component: {} [e2e-llm-inference-service] f:app.kubernetes.io/name: {} [e2e-llm-inference-service] f:app.kubernetes.io/part-of: {} [e2e-llm-inference-service] f:endpoints.kubernetes.io/managed-by: {} [e2e-llm-inference-service] f:subsets: {} [e2e-llm-inference-service] subsets: [e2e-llm-inference-service] - addresses: [e2e-llm-inference-service] - ip: 10.134.0.32 [e2e-llm-inference-service] nodeName: ip-10-0-139-47.ec2.internal [e2e-llm-inference-service] targetRef: [e2e-llm-inference-service] kind: Pod [e2e-llm-inference-service] namespace: kserve-ci-e2e-test [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-ba4d693a-kserve-5b998d5789-t6wf4 [e2e-llm-inference-service] uid: 68c5609b-442c-4a9c-837a-67c032fcf236 [e2e-llm-inference-service] ports: [e2e-llm-inference-service] - name: https [e2e-llm-inference-service] port: 8000 [e2e-llm-inference-service] protocol: TCP [e2e-llm-inference-service] appProtocol: https [e2e-llm-inference-service] apiVersion: v1 [e2e-llm-inference-service] kind: Endpoints [e2e-llm-inference-service] [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1254 --- [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1255 metadata: [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-ba4d693a-kserve-5b998d5789-t6wf4 [e2e-llm-inference-service] generateName: llmisvc-model-fb-opt-125m-with-ba4d693a-kserve-5b998d5789- [e2e-llm-inference-service] namespace: kserve-ci-e2e-test [e2e-llm-inference-service] uid: 68c5609b-442c-4a9c-837a-67c032fcf236 [e2e-llm-inference-service] resourceVersion: '47360' [e2e-llm-inference-service] generation: 1 [e2e-llm-inference-service] creationTimestamp: '2026-07-07T14:45:56Z' [e2e-llm-inference-service] labels: [e2e-llm-inference-service] app.kubernetes.io/component: llminferenceservice-workload [e2e-llm-inference-service] app.kubernetes.io/name: llmisvc-model-fb-opt-125m-with-ba4d693a [e2e-llm-inference-service] app.kubernetes.io/part-of: llminferenceservice [e2e-llm-inference-service] kserve.io/component: workload [e2e-llm-inference-service] llm-d.ai/role: both [e2e-llm-inference-service] pod-template-hash: 5b998d5789 [e2e-llm-inference-service] annotations: [e2e-llm-inference-service] k8s.ovn.org/pod-networks: '{"default":{"ip_addresses":["10.134.0.32/23"],"mac_address":"0a:58:0a:86:00:20","gateway_ips":["10.134.0.1"],"routes":[{"dest":"10.132.0.0/14","nextHop":"10.134.0.1"},{"dest":"172.31.0.0/16","nextHop":"10.134.0.1"},{"dest":"169.254.0.5/32","nextHop":"10.134.0.1"},{"dest":"100.64.0.0/16","nextHop":"10.134.0.1"}],"ip_address":"10.134.0.32/23","gateway_ip":"10.134.0.1","role":"primary"}}' [e2e-llm-inference-service] k8s.v1.cni.cncf.io/network-status: "[{\n \"name\": \"ovn-kubernetes\",\n \ [e2e-llm-inference-service] \ \"interface\": \"eth0\",\n \"ips\": [\n \"10.134.0.32\"\n ],\n\ [e2e-llm-inference-service] \ \"mac\": \"0a:58:0a:86:00:20\",\n \"default\": true,\n \"dns\": {}\n\ [e2e-llm-inference-service] }]" [e2e-llm-inference-service] openshift.io/scc: restricted-v2 [e2e-llm-inference-service] seccomp.security.alpha.kubernetes.io/pod: runtime/default [e2e-llm-inference-service] security.openshift.io/validated-scc-subject-type: user [e2e-llm-inference-service] ownerReferences: [e2e-llm-inference-service] - apiVersion: apps/v1 [e2e-llm-inference-service] kind: ReplicaSet [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-ba4d693a-kserve-5b998d5789 [e2e-llm-inference-service] uid: e69bc84c-145f-4492-a785-714fe7558007 [e2e-llm-inference-service] controller: true [e2e-llm-inference-service] blockOwnerDeletion: true [e2e-llm-inference-service] managedFields: [e2e-llm-inference-service] - manager: ip-10-0-139-47 [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] apiVersion: v1 [e2e-llm-inference-service] time: '2026-07-07T14:45:56Z' [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:metadata: [e2e-llm-inference-service] f:annotations: [e2e-llm-inference-service] f:k8s.ovn.org/pod-networks: {} [e2e-llm-inference-service] subresource: status [e2e-llm-inference-service] - manager: kube-controller-manager [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] apiVersion: v1 [e2e-llm-inference-service] time: '2026-07-07T14:45:56Z' [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:metadata: [e2e-llm-inference-service] f:generateName: {} [e2e-llm-inference-service] f:labels: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:app.kubernetes.io/component: {} [e2e-llm-inference-service] f:app.kubernetes.io/name: {} [e2e-llm-inference-service] f:app.kubernetes.io/part-of: {} [e2e-llm-inference-service] f:kserve.io/component: {} [e2e-llm-inference-service] f:llm-d.ai/role: {} [e2e-llm-inference-service] f:pod-template-hash: {} [e2e-llm-inference-service] f:ownerReferences: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"uid":"e69bc84c-145f-4492-a785-714fe7558007"}: {} [e2e-llm-inference-service] f:spec: [e2e-llm-inference-service] f:containers: [e2e-llm-inference-service] k:{"name":"main"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:args: {} [e2e-llm-inference-service] f:command: {} [e2e-llm-inference-service] f:env: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"name":"HF_HUB_CACHE"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"HOME"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"TORCHINDUCTOR_CACHE_DIR"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"USER"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"VLLM_CPU_KVCACHE_SPACE"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"VLLM_ENABLE_V1_MULTIPROCESSING"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"VLLM_LOGGING_LEVEL"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] f:image: {} [e2e-llm-inference-service] f:imagePullPolicy: {} [e2e-llm-inference-service] f:lifecycle: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:preStop: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:exec: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:command: {} [e2e-llm-inference-service] f:livenessProbe: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:failureThreshold: {} [e2e-llm-inference-service] f:httpGet: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:path: {} [e2e-llm-inference-service] f:port: {} [e2e-llm-inference-service] f:scheme: {} [e2e-llm-inference-service] f:periodSeconds: {} [e2e-llm-inference-service] f:successThreshold: {} [e2e-llm-inference-service] f:timeoutSeconds: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:ports: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"containerPort":8000,"protocol":"TCP"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:containerPort: {} [e2e-llm-inference-service] f:protocol: {} [e2e-llm-inference-service] f:readinessProbe: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:failureThreshold: {} [e2e-llm-inference-service] f:httpGet: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:path: {} [e2e-llm-inference-service] f:port: {} [e2e-llm-inference-service] f:scheme: {} [e2e-llm-inference-service] f:periodSeconds: {} [e2e-llm-inference-service] f:successThreshold: {} [e2e-llm-inference-service] f:timeoutSeconds: {} [e2e-llm-inference-service] f:resources: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:limits: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:cpu: {} [e2e-llm-inference-service] f:memory: {} [e2e-llm-inference-service] f:requests: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:cpu: {} [e2e-llm-inference-service] f:memory: {} [e2e-llm-inference-service] f:securityContext: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:allowPrivilegeEscalation: {} [e2e-llm-inference-service] f:capabilities: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:drop: {} [e2e-llm-inference-service] f:readOnlyRootFilesystem: {} [e2e-llm-inference-service] f:runAsNonRoot: {} [e2e-llm-inference-service] f:seccompProfile: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:type: {} [e2e-llm-inference-service] f:startupProbe: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:failureThreshold: {} [e2e-llm-inference-service] f:httpGet: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:path: {} [e2e-llm-inference-service] f:port: {} [e2e-llm-inference-service] f:scheme: {} [e2e-llm-inference-service] f:periodSeconds: {} [e2e-llm-inference-service] f:successThreshold: {} [e2e-llm-inference-service] f:timeoutSeconds: {} [e2e-llm-inference-service] f:terminationMessagePath: {} [e2e-llm-inference-service] f:terminationMessagePolicy: {} [e2e-llm-inference-service] f:volumeMounts: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"mountPath":"/dev/shm"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:mountPath: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] k:{"mountPath":"/home"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:mountPath: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] k:{"mountPath":"/mnt"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:mountPath: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:readOnly: {} [e2e-llm-inference-service] k:{"mountPath":"/models"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:mountPath: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] k:{"mountPath":"/tmp"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:mountPath: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] k:{"mountPath":"/var/run/kserve/tls"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:mountPath: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:readOnly: {} [e2e-llm-inference-service] f:dnsPolicy: {} [e2e-llm-inference-service] f:enableServiceLinks: {} [e2e-llm-inference-service] f:initContainers: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"name":"storage-initializer"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:args: {} [e2e-llm-inference-service] f:env: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"name":"AWS_ACCESS_KEY_ID"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:valueFrom: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:secretKeyRef: {} [e2e-llm-inference-service] k:{"name":"AWS_CA_BUNDLE"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"AWS_CA_BUNDLE_CONFIGMAP"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"AWS_ENDPOINT_URL"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"AWS_SECRET_ACCESS_KEY"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:valueFrom: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:secretKeyRef: {} [e2e-llm-inference-service] k:{"name":"CA_BUNDLE_CONFIGMAP_NAME"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"CA_BUNDLE_VOLUME_MOUNT_POINT"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"HF_HUB_ENABLE_HF_TRANSFER"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"HF_XET_HIGH_PERFORMANCE"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"HF_XET_NUM_CONCURRENT_RANGE_GETS"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"S3_ENDPOINT"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"S3_USE_HTTPS"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"S3_VERIFY_SSL"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] f:image: {} [e2e-llm-inference-service] f:imagePullPolicy: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:resources: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:limits: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:cpu: {} [e2e-llm-inference-service] f:memory: {} [e2e-llm-inference-service] f:requests: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:cpu: {} [e2e-llm-inference-service] f:memory: {} [e2e-llm-inference-service] f:terminationMessagePath: {} [e2e-llm-inference-service] f:terminationMessagePolicy: {} [e2e-llm-inference-service] f:volumeMounts: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"mountPath":"/etc/ssl/custom-certs"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:mountPath: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:readOnly: {} [e2e-llm-inference-service] k:{"mountPath":"/mnt"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:mountPath: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:restartPolicy: {} [e2e-llm-inference-service] f:schedulerName: {} [e2e-llm-inference-service] f:securityContext: {} [e2e-llm-inference-service] f:terminationGracePeriodSeconds: {} [e2e-llm-inference-service] f:volumes: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"name":"cabundle-cert"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:configMap: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:defaultMode: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] k:{"name":"dshm"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:emptyDir: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:medium: {} [e2e-llm-inference-service] f:sizeLimit: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] k:{"name":"home"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:emptyDir: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] k:{"name":"kserve-provision-location"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:emptyDir: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] k:{"name":"model-cache"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:emptyDir: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] k:{"name":"tls-certs"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:secret: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:defaultMode: {} [e2e-llm-inference-service] f:secretName: {} [e2e-llm-inference-service] k:{"name":"tmp-dir"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:emptyDir: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] - manager: multus-daemon [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] apiVersion: v1 [e2e-llm-inference-service] time: '2026-07-07T14:45:56Z' [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:metadata: [e2e-llm-inference-service] f:annotations: [e2e-llm-inference-service] f:k8s.v1.cni.cncf.io/network-status: {} [e2e-llm-inference-service] subresource: status [e2e-llm-inference-service] - manager: kubelet [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] apiVersion: v1 [e2e-llm-inference-service] time: '2026-07-07T14:47:16Z' [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:status: [e2e-llm-inference-service] f:conditions: [e2e-llm-inference-service] k:{"type":"ContainersReady"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:lastProbeTime: {} [e2e-llm-inference-service] f:lastTransitionTime: {} [e2e-llm-inference-service] f:observedGeneration: {} [e2e-llm-inference-service] f:status: {} [e2e-llm-inference-service] f:type: {} [e2e-llm-inference-service] k:{"type":"Initialized"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:lastProbeTime: {} [e2e-llm-inference-service] f:lastTransitionTime: {} [e2e-llm-inference-service] f:observedGeneration: {} [e2e-llm-inference-service] f:status: {} [e2e-llm-inference-service] f:type: {} [e2e-llm-inference-service] k:{"type":"PodReadyToStartContainers"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:lastProbeTime: {} [e2e-llm-inference-service] f:lastTransitionTime: {} [e2e-llm-inference-service] f:observedGeneration: {} [e2e-llm-inference-service] f:status: {} [e2e-llm-inference-service] f:type: {} [e2e-llm-inference-service] k:{"type":"PodScheduled"}: [e2e-llm-inference-service] f:observedGeneration: {} [e2e-llm-inference-service] k:{"type":"Ready"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:lastProbeTime: {} [e2e-llm-inference-service] f:lastTransitionTime: {} [e2e-llm-inference-service] f:observedGeneration: {} [e2e-llm-inference-service] f:status: {} [e2e-llm-inference-service] f:type: {} [e2e-llm-inference-service] f:containerStatuses: {} [e2e-llm-inference-service] f:hostIP: {} [e2e-llm-inference-service] f:hostIPs: {} [e2e-llm-inference-service] f:initContainerStatuses: {} [e2e-llm-inference-service] f:observedGeneration: {} [e2e-llm-inference-service] f:phase: {} [e2e-llm-inference-service] f:podIP: {} [e2e-llm-inference-service] f:podIPs: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"ip":"10.134.0.32"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:ip: {} [e2e-llm-inference-service] f:startTime: {} [e2e-llm-inference-service] subresource: status [e2e-llm-inference-service] spec: [e2e-llm-inference-service] volumes: [e2e-llm-inference-service] - name: home [e2e-llm-inference-service] emptyDir: {} [e2e-llm-inference-service] - name: dshm [e2e-llm-inference-service] emptyDir: [e2e-llm-inference-service] medium: Memory [e2e-llm-inference-service] sizeLimit: 1Gi [e2e-llm-inference-service] - name: model-cache [e2e-llm-inference-service] emptyDir: {} [e2e-llm-inference-service] - name: tmp-dir [e2e-llm-inference-service] emptyDir: {} [e2e-llm-inference-service] - name: tls-certs [e2e-llm-inference-service] secret: [e2e-llm-inference-service] secretName: llmisv77ff2528d3e9b4972cd9335229fce9f0-kserve-self-signed-certs [e2e-llm-inference-service] defaultMode: 420 [e2e-llm-inference-service] - name: kserve-provision-location [e2e-llm-inference-service] emptyDir: {} [e2e-llm-inference-service] - name: cabundle-cert [e2e-llm-inference-service] configMap: [e2e-llm-inference-service] name: odh-kserve-custom-ca-bundle [e2e-llm-inference-service] defaultMode: 420 [e2e-llm-inference-service] - name: kube-api-access-tqj64 [e2e-llm-inference-service] projected: [e2e-llm-inference-service] sources: [e2e-llm-inference-service] - serviceAccountToken: [e2e-llm-inference-service] expirationSeconds: 3607 [e2e-llm-inference-service] path: token [e2e-llm-inference-service] - configMap: [e2e-llm-inference-service] name: kube-root-ca.crt [e2e-llm-inference-service] items: [e2e-llm-inference-service] - key: ca.crt [e2e-llm-inference-service] path: ca.crt [e2e-llm-inference-service] - downwardAPI: [e2e-llm-inference-service] items: [e2e-llm-inference-service] - path: namespace [e2e-llm-inference-service] fieldRef: [e2e-llm-inference-service] apiVersion: v1 [e2e-llm-inference-service] fieldPath: metadata.namespace [e2e-llm-inference-service] - configMap: [e2e-llm-inference-service] name: openshift-service-ca.crt [e2e-llm-inference-service] items: [e2e-llm-inference-service] - key: service-ca.crt [e2e-llm-inference-service] path: service-ca.crt [e2e-llm-inference-service] defaultMode: 420 [e2e-llm-inference-service] initContainers: [e2e-llm-inference-service] - name: storage-initializer [e2e-llm-inference-service] image: quay.io/opendatahub/kserve-storage-initializer@sha256:13ad765c0f323200e1db7b45c99661287e65c5158d6b54b16139829636d20be4 [e2e-llm-inference-service] args: [e2e-llm-inference-service] - hf://facebook/opt-125m [e2e-llm-inference-service] - /mnt/models [e2e-llm-inference-service] - hf://edbeeching/opt-125m-lora [e2e-llm-inference-service] - /mnt/lora/lora-adapter-1 [e2e-llm-inference-service] env: [e2e-llm-inference-service] - name: AWS_ACCESS_KEY_ID [e2e-llm-inference-service] valueFrom: [e2e-llm-inference-service] secretKeyRef: [e2e-llm-inference-service] name: seaweedfs-s3-creds [e2e-llm-inference-service] key: AWS_ACCESS_KEY_ID [e2e-llm-inference-service] - name: AWS_SECRET_ACCESS_KEY [e2e-llm-inference-service] valueFrom: [e2e-llm-inference-service] secretKeyRef: [e2e-llm-inference-service] name: seaweedfs-s3-creds [e2e-llm-inference-service] key: AWS_SECRET_ACCESS_KEY [e2e-llm-inference-service] - name: S3_USE_HTTPS [e2e-llm-inference-service] value: '0' [e2e-llm-inference-service] - name: S3_ENDPOINT [e2e-llm-inference-service] value: s3-service.kserve:8333 [e2e-llm-inference-service] - name: AWS_ENDPOINT_URL [e2e-llm-inference-service] value: http://s3-service.kserve:8333 [e2e-llm-inference-service] - name: S3_VERIFY_SSL [e2e-llm-inference-service] value: '0' [e2e-llm-inference-service] - name: AWS_CA_BUNDLE [e2e-llm-inference-service] value: /etc/ssl/custom-certs/cabundle.crt [e2e-llm-inference-service] - name: AWS_CA_BUNDLE_CONFIGMAP [e2e-llm-inference-service] value: odh-kserve-custom-ca-bundle [e2e-llm-inference-service] - name: HF_HUB_ENABLE_HF_TRANSFER [e2e-llm-inference-service] value: '1' [e2e-llm-inference-service] - name: HF_XET_HIGH_PERFORMANCE [e2e-llm-inference-service] value: '1' [e2e-llm-inference-service] - name: HF_XET_NUM_CONCURRENT_RANGE_GETS [e2e-llm-inference-service] value: '8' [e2e-llm-inference-service] - name: CA_BUNDLE_CONFIGMAP_NAME [e2e-llm-inference-service] value: odh-kserve-custom-ca-bundle [e2e-llm-inference-service] - name: CA_BUNDLE_VOLUME_MOUNT_POINT [e2e-llm-inference-service] value: /etc/ssl/custom-certs [e2e-llm-inference-service] resources: [e2e-llm-inference-service] limits: [e2e-llm-inference-service] cpu: '1' [e2e-llm-inference-service] memory: 24Gi [e2e-llm-inference-service] requests: [e2e-llm-inference-service] cpu: 100m [e2e-llm-inference-service] memory: 100Mi [e2e-llm-inference-service] volumeMounts: [e2e-llm-inference-service] - name: kserve-provision-location [e2e-llm-inference-service] mountPath: /mnt [e2e-llm-inference-service] - name: cabundle-cert [e2e-llm-inference-service] readOnly: true [e2e-llm-inference-service] mountPath: /etc/ssl/custom-certs [e2e-llm-inference-service] - name: kube-api-access-tqj64 [e2e-llm-inference-service] readOnly: true [e2e-llm-inference-service] mountPath: /var/run/secrets/kubernetes.io/serviceaccount [e2e-llm-inference-service] terminationMessagePath: /dev/termination-log [e2e-llm-inference-service] terminationMessagePolicy: FallbackToLogsOnError [e2e-llm-inference-service] imagePullPolicy: IfNotPresent [e2e-llm-inference-service] securityContext: [e2e-llm-inference-service] capabilities: [e2e-llm-inference-service] drop: [e2e-llm-inference-service] - ALL [e2e-llm-inference-service] runAsUser: 1000690000 [e2e-llm-inference-service] runAsNonRoot: true [e2e-llm-inference-service] allowPrivilegeEscalation: false [e2e-llm-inference-service] containers: [e2e-llm-inference-service] - name: main [e2e-llm-inference-service] image: public.ecr.aws/q9t5s3a7/vllm-cpu-release-repo:v0.19.0 [e2e-llm-inference-service] command: [e2e-llm-inference-service] - /bin/bash [e2e-llm-inference-service] - -c [e2e-llm-inference-service] - "if [ -f /etc/profile.d/ibm-aiu-setup.sh ]; then\n source /etc/profile.d/ibm-aiu-setup.sh\n\ [e2e-llm-inference-service] fi\n\nif [ \"$KSERVE_INFER_ROCE\" = \"true\" ]; then\n echo \"Trying to infer\ [e2e-llm-inference-service] \ RoCE configs ... \"\n grep -H . /sys/class/infiniband/*/ports/*/gids/* 2>/dev/null\n\ [e2e-llm-inference-service] \ grep -H . /sys/class/infiniband/*/ports/*/gid_attrs/types/* 2>/dev/null\n\ [e2e-llm-inference-service] \n cat /proc/driver/nvidia/params\n\n KSERVE_INFER_IB_GID_INDEX_GREP=${KSERVE_INFER_IB_GID_INDEX_GREP:-\"\ [e2e-llm-inference-service] RoCE v2\"}\n\n echo \"[Infer RoCE] Discovering active HCAs ...\"\n active_hcas=()\n\ [e2e-llm-inference-service] \ # Loop through all mlx5 devices found in sysfs\n for hca_dir in /sys/class/infiniband/mlx5_*;\ [e2e-llm-inference-service] \ do\n # Ensure it's a directory before proceeding\n if [ -d \"$hca_dir\"\ [e2e-llm-inference-service] \ ]; then\n hca_name=$(basename \"$hca_dir\")\n port_state_file=\"\ [e2e-llm-inference-service] $hca_dir/ports/1/state\" # Assume port 1\n type_file=\"$hca_dir/ports/1/gid_attrs/types/*\"\ [e2e-llm-inference-service] \n\n echo \"[Infer RoCE] Check if the port state file ${port_state_file}\ [e2e-llm-inference-service] \ exists and contains 'ACTIVE'\"\n if [ -f \"$port_state_file\" ] &&\ [e2e-llm-inference-service] \ grep -q \"ACTIVE\" \"$port_state_file\" && grep -q \"${KSERVE_INFER_IB_GID_INDEX_GREP}\"\ [e2e-llm-inference-service] \ ${type_file} 2>/dev/null; then\n echo \"[Infer RoCE] Found active\ [e2e-llm-inference-service] \ HCA: $hca_name\"\n active_hcas+=(\"$hca_name\")\n else\n\ [e2e-llm-inference-service] \ echo \"[Infer RoCE] Skipping inactive or down HCA: $hca_name\"\ [e2e-llm-inference-service] \n fi\n fi\n done\n\n # Check if we found any active HCAs\n\ [e2e-llm-inference-service] \ if [ ${#active_hcas[@]} -gt 0 ]; then\n # Join the array elements with\ [e2e-llm-inference-service] \ a comma\n hca_port_pairs=()\n for hca in \"${active_hcas[@]}\";\ [e2e-llm-inference-service] \ do\n hca_port_pairs+=(\"${hca}:1\")\n done\n\n active_hca_list=$(IFS=,;\ [e2e-llm-inference-service] \ echo \"${active_hcas[*]}\")\n hca_port_pairs_list=$(IFS=,; echo \"${hca_port_pairs[*]}\"\ [e2e-llm-inference-service] )\n echo \"[Infer RoCE] Setting active HCAs: ${active_hca_list}\"\n \ [e2e-llm-inference-service] \ export NCCL_IB_HCA=${NCCL_IB_HCA:-${active_hca_list}}\n export NVSHMEM_HCA_LIST=${NVSHMEM_HCA_LIST:-${hca_port_pairs_list}}\n\ [e2e-llm-inference-service] \ export UCX_NET_DEVICES=${UCX_NET_DEVICES:-${hca_port_pairs_list}}\n\n\ [e2e-llm-inference-service] \ echo \"[Infer RoCE] NCCL_IB_HCA=${NCCL_IB_HCA}\"\n echo \"[Infer\ [e2e-llm-inference-service] \ RoCE] NVSHMEM_HCA_LIST=${NVSHMEM_HCA_LIST}\"\n echo \"[Infer RoCE] UCX_NET_DEVICES=${UCX_NET_DEVICES}\"\ [e2e-llm-inference-service] \n else\n echo \"[Infer RoCE] WARNING: No active RoCE HCAs found. NCCL_IB_HCA\ [e2e-llm-inference-service] \ will not be set.\"\n fi\n\n if [ ${#active_hcas[@]} -gt 0 ]; then\n \ [e2e-llm-inference-service] \ echo \"[Infer RoCE] Finding GID_INDEX for each active HCA (SR-IOV compatible)...\"\ [e2e-llm-inference-service] \n\n # For SR-IOV environments, find the most common IPv4 RoCE v2 GID index\ [e2e-llm-inference-service] \ across all HCAs\n declare -A gid_index_count\n declare -A hca_gid_index\n\ [e2e-llm-inference-service] \n for hca_name in \"${active_hcas[@]}\"; do\n echo \"[Infer RoCE]\ [e2e-llm-inference-service] \ Processing HCA: ${hca_name}\"\n\n # Find all RoCE v2 IPv4 GIDs for\ [e2e-llm-inference-service] \ this HCA and count by index\n for tpath in /sys/class/infiniband/${hca_name}/ports/1/gid_attrs/types/*;\ [e2e-llm-inference-service] \ do\n if grep -q \"${KSERVE_INFER_IB_GID_INDEX_GREP}\" \"$tpath\"\ [e2e-llm-inference-service] \ 2>/dev/null; then\n idx=$(basename \"$tpath\")\n \ [e2e-llm-inference-service] \ gid_file=\"/sys/class/infiniband/${hca_name}/ports/1/gids/${idx}\"\ [e2e-llm-inference-service] \n # Check for IPv4 GID (contains ffff:)\n \ [e2e-llm-inference-service] \ if [ -f \"$gid_file\" ] && grep -q \"ffff:\" \"$gid_file\"; then\n \ [e2e-llm-inference-service] \ gid_value=$(cat \"$gid_file\" 2>/dev/null || echo \"\")\n \ [e2e-llm-inference-service] \ echo \"[Infer RoCE] Found IPv4 RoCE v2 GID for ${hca_name}:\ [e2e-llm-inference-service] \ index=${idx}, gid=${gid_value}\"\n hca_gid_index[\"${hca_name}\"\ [e2e-llm-inference-service] ]=\"${idx}\"\n gid_index_count[\"${idx}\"]=$((${gid_index_count[\"\ [e2e-llm-inference-service] ${idx}\"]} + 1))\n break # Use first found IPv4 GID per\ [e2e-llm-inference-service] \ HCA\n fi\n fi\n done\n done\n\n\ [e2e-llm-inference-service] \ # Find the most common GID index (most likely to be consistent across\ [e2e-llm-inference-service] \ nodes)\n best_gid_index=\"\"\n max_count=0\n for idx in \"\ [e2e-llm-inference-service] ${!gid_index_count[@]}\"; do\n count=${gid_index_count[\"${idx}\"]}\n\ [e2e-llm-inference-service] \ echo \"[Infer RoCE] GID_INDEX ${idx} found on ${count} HCAs\"\n \ [e2e-llm-inference-service] \ if [ $count -gt $max_count ]; then\n max_count=$count\n\ [e2e-llm-inference-service] \ best_gid_index=\"$idx\"\n fi\n done\n\n #\ [e2e-llm-inference-service] \ Use deterministic fallback if tied - prefer index 3 (SR-IOV standard)\n \ [e2e-llm-inference-service] \ if [ ${#gid_index_count[@]} -gt 1 ]; then\n echo \"[Infer RoCE]\ [e2e-llm-inference-service] \ Multiple GID indices found, selecting most common: ${best_gid_index}\"\n \ [e2e-llm-inference-service] \ # If there's a tie, prefer index 3 as it's most common in SR-IOV setups\n\ [e2e-llm-inference-service] \ if [ -n \"${gid_index_count['3']}\" ] && [ \"${gid_index_count['3']}\"\ [e2e-llm-inference-service] \ -eq \"$max_count\" ]; then\n best_gid_index=\"3\"\n \ [e2e-llm-inference-service] \ echo \"[Infer RoCE] Using deterministic fallback: GID_INDEX=3 (SR-IOV\ [e2e-llm-inference-service] \ standard)\"\n fi\n fi\n\n # Check if GID_INDEX is already\ [e2e-llm-inference-service] \ set via environment variables\n if [ -n \"${NCCL_IB_GID_INDEX}\" ]; then\n\ [e2e-llm-inference-service] \ echo \"[Infer RoCE] Using pre-configured NCCL_IB_GID_INDEX=${NCCL_IB_GID_INDEX}\ [e2e-llm-inference-service] \ from environment\"\n export NVSHMEM_IB_GID_INDEX=${NVSHMEM_IB_GID_INDEX:-$NCCL_IB_GID_INDEX}\n\ [e2e-llm-inference-service] \ export UCX_IB_GID_INDEX=${UCX_IB_GID_INDEX:-$NCCL_IB_GID_INDEX}\n\ [e2e-llm-inference-service] \ echo \"[Infer RoCE] Using pre-configured GID_INDEX=${NCCL_IB_GID_INDEX}\ [e2e-llm-inference-service] \ for NCCL, NVSHMEM, and UCX\"\n elif [ -n \"$best_gid_index\" ]; then\n\ [e2e-llm-inference-service] \ echo \"[Infer RoCE] Selected GID_INDEX: ${best_gid_index} (found\ [e2e-llm-inference-service] \ on ${max_count} HCAs)\"\n\n export NCCL_IB_GID_INDEX=${NCCL_IB_GID_INDEX:-$best_gid_index}\n\ [e2e-llm-inference-service] \ export NVSHMEM_IB_GID_INDEX=${NVSHMEM_IB_GID_INDEX:-$best_gid_index}\n\ [e2e-llm-inference-service] \ export UCX_IB_GID_INDEX=${UCX_IB_GID_INDEX:-$best_gid_index}\n\n\ [e2e-llm-inference-service] \ echo \"[Infer RoCE] Exported GID_INDEX=${best_gid_index} for NCCL,\ [e2e-llm-inference-service] \ NVSHMEM, and UCX\"\n else\n echo \"[Infer RoCE] ERROR: No valid\ [e2e-llm-inference-service] \ IPv4 ${KSERVE_INFER_IB_GID_INDEX_GREP} GID_INDEX found on any HCA.\"\n \ [e2e-llm-inference-service] \ fi\n else\n echo \"[Infer RoCE] No active HCAs found, skipping GID_INDEX\ [e2e-llm-inference-service] \ inference.\"\n fi\nfi\n\n# --disable-access-log-for-endpoints landed in vLLM\ [e2e-llm-inference-service] \ 0.16.0 (vllm-project/vllm#30011).\n# Older versions still need the blanket\ [e2e-llm-inference-service] \ --disable-uvicorn-access-log.\nACCESS_LOG_ARGS=\"--disable-uvicorn-access-log\"\ [e2e-llm-inference-service] \nVLLM_VERSION=$(vllm --version 2>/dev/null | tail -1 | awk '{print $NF}')\n\ [e2e-llm-inference-service] echo \"[access-log-detect] vllm version='${VLLM_VERSION}'\"\nif [[ \"$VLLM_VERSION\"\ [e2e-llm-inference-service] \ =~ ^[0-9]+\\.[0-9]+ ]] && [ \"$(printf '%s\\n%s\\n' \"0.16.0\" \"${VLLM_VERSION}\"\ [e2e-llm-inference-service] \ | sort -V | head -1)\" = \"0.16.0\" ]; then\n ACCESS_LOG_ARGS=\"--disable-access-log-for-endpoints\ [e2e-llm-inference-service] \ /health,/metrics,/ping\"\nfi\necho \"[access-log-detect] selected ACCESS_LOG_ARGS='${ACCESS_LOG_ARGS}'\"\ [e2e-llm-inference-service] \n\n# --shutdown-timeout landed in vLLM 0.18.0 (vllm-project/vllm#36666).\n\ [e2e-llm-inference-service] SHUTDOWN_TIMEOUT_ARGS=\"\"\nif [[ \"$VLLM_VERSION\" =~ ^[0-9]+\\.[0-9]+ ]] &&\ [e2e-llm-inference-service] \ [ \"$(printf '%s\\n%s\\n' \"0.18.0\" \"${VLLM_VERSION}\" | sort -V | head\ [e2e-llm-inference-service] \ -1)\" = \"0.18.0\" ]; then\n SHUTDOWN_TIMEOUT_ARGS=\"--shutdown-timeout 40\"\ [e2e-llm-inference-service] \nfi\n\neval \"exec vllm serve /mnt/models \\\n --served-model-name \"facebook/opt-125m\"\ [e2e-llm-inference-service] \ \"publishers/kserve-ci-e2e-test/models/facebook/opt-125m\" \\\n --port 8000\ [e2e-llm-inference-service] \ \\\n ${ACCESS_LOG_ARGS} \\\n ${SHUTDOWN_TIMEOUT_ARGS} \\\n --enable-ssl-refresh\ [e2e-llm-inference-service] \ \\\n --ssl-certfile /var/run/kserve/tls/tls.crt \\\n --ssl-keyfile /var/run/kserve/tls/tls.key\ [e2e-llm-inference-service] \ \\\n ${VLLM_ADDITIONAL_ARGS} \\\n $@\"" [e2e-llm-inference-service] - -- [e2e-llm-inference-service] args: [e2e-llm-inference-service] - --enable-lora [e2e-llm-inference-service] - --lora-modules [e2e-llm-inference-service] - '''{"name":"lora-adapter-1","path":"/mnt/lora/lora-adapter-1"}''' [e2e-llm-inference-service] - '''{"name":"publishers/kserve-ci-e2e-test/models/lora-adapter-1","path":"/mnt/lora/lora-adapter-1"}''' [e2e-llm-inference-service] ports: [e2e-llm-inference-service] - containerPort: 8000 [e2e-llm-inference-service] protocol: TCP [e2e-llm-inference-service] env: [e2e-llm-inference-service] - name: HOME [e2e-llm-inference-service] value: /home [e2e-llm-inference-service] - name: VLLM_LOGGING_LEVEL [e2e-llm-inference-service] value: DEBUG [e2e-llm-inference-service] - name: VLLM_CPU_KVCACHE_SPACE [e2e-llm-inference-service] value: '1' [e2e-llm-inference-service] - name: VLLM_ENABLE_V1_MULTIPROCESSING [e2e-llm-inference-service] value: '0' [e2e-llm-inference-service] - name: USER [e2e-llm-inference-service] value: nonroot [e2e-llm-inference-service] - name: TORCHINDUCTOR_CACHE_DIR [e2e-llm-inference-service] value: /tmp/torchinductor-cache [e2e-llm-inference-service] - name: HF_HUB_CACHE [e2e-llm-inference-service] value: /models [e2e-llm-inference-service] resources: [e2e-llm-inference-service] limits: [e2e-llm-inference-service] cpu: '2' [e2e-llm-inference-service] memory: 7Gi [e2e-llm-inference-service] requests: [e2e-llm-inference-service] cpu: 200m [e2e-llm-inference-service] memory: 2Gi [e2e-llm-inference-service] volumeMounts: [e2e-llm-inference-service] - name: home [e2e-llm-inference-service] mountPath: /home [e2e-llm-inference-service] - name: tmp-dir [e2e-llm-inference-service] mountPath: /tmp [e2e-llm-inference-service] - name: dshm [e2e-llm-inference-service] mountPath: /dev/shm [e2e-llm-inference-service] - name: model-cache [e2e-llm-inference-service] mountPath: /models [e2e-llm-inference-service] - name: tls-certs [e2e-llm-inference-service] readOnly: true [e2e-llm-inference-service] mountPath: /var/run/kserve/tls [e2e-llm-inference-service] - name: kserve-provision-location [e2e-llm-inference-service] readOnly: true [e2e-llm-inference-service] mountPath: /mnt [e2e-llm-inference-service] - name: kube-api-access-tqj64 [e2e-llm-inference-service] readOnly: true [e2e-llm-inference-service] mountPath: /var/run/secrets/kubernetes.io/serviceaccount [e2e-llm-inference-service] livenessProbe: [e2e-llm-inference-service] httpGet: [e2e-llm-inference-service] path: /health [e2e-llm-inference-service] port: 8000 [e2e-llm-inference-service] scheme: HTTPS [e2e-llm-inference-service] timeoutSeconds: 1 [e2e-llm-inference-service] periodSeconds: 10 [e2e-llm-inference-service] successThreshold: 1 [e2e-llm-inference-service] failureThreshold: 10 [e2e-llm-inference-service] readinessProbe: [e2e-llm-inference-service] httpGet: [e2e-llm-inference-service] path: /health [e2e-llm-inference-service] port: 8000 [e2e-llm-inference-service] scheme: HTTPS [e2e-llm-inference-service] timeoutSeconds: 1 [e2e-llm-inference-service] periodSeconds: 1 [e2e-llm-inference-service] successThreshold: 1 [e2e-llm-inference-service] failureThreshold: 2 [e2e-llm-inference-service] startupProbe: [e2e-llm-inference-service] httpGet: [e2e-llm-inference-service] path: /health [e2e-llm-inference-service] port: 8000 [e2e-llm-inference-service] scheme: HTTPS [e2e-llm-inference-service] timeoutSeconds: 1 [e2e-llm-inference-service] periodSeconds: 10 [e2e-llm-inference-service] successThreshold: 1 [e2e-llm-inference-service] failureThreshold: 60 [e2e-llm-inference-service] lifecycle: [e2e-llm-inference-service] preStop: [e2e-llm-inference-service] exec: [e2e-llm-inference-service] command: [e2e-llm-inference-service] - /bin/sleep [e2e-llm-inference-service] - '15' [e2e-llm-inference-service] terminationMessagePath: /dev/termination-log [e2e-llm-inference-service] terminationMessagePolicy: FallbackToLogsOnError [e2e-llm-inference-service] imagePullPolicy: IfNotPresent [e2e-llm-inference-service] securityContext: [e2e-llm-inference-service] capabilities: [e2e-llm-inference-service] drop: [e2e-llm-inference-service] - ALL [e2e-llm-inference-service] runAsUser: 1000690000 [e2e-llm-inference-service] runAsNonRoot: true [e2e-llm-inference-service] readOnlyRootFilesystem: true [e2e-llm-inference-service] allowPrivilegeEscalation: false [e2e-llm-inference-service] seccompProfile: [e2e-llm-inference-service] type: RuntimeDefault [e2e-llm-inference-service] restartPolicy: Always [e2e-llm-inference-service] terminationGracePeriodSeconds: 60 [e2e-llm-inference-service] dnsPolicy: ClusterFirst [e2e-llm-inference-service] serviceAccountName: default [e2e-llm-inference-service] serviceAccount: default [e2e-llm-inference-service] nodeName: ip-10-0-139-47.ec2.internal [e2e-llm-inference-service] securityContext: [e2e-llm-inference-service] seLinuxOptions: [e2e-llm-inference-service] level: s0:c26,c20 [e2e-llm-inference-service] fsGroup: 1000690000 [e2e-llm-inference-service] seccompProfile: [e2e-llm-inference-service] type: RuntimeDefault [e2e-llm-inference-service] imagePullSecrets: [e2e-llm-inference-service] - name: default-dockercfg-tbbgr [e2e-llm-inference-service] schedulerName: default-scheduler [e2e-llm-inference-service] tolerations: [e2e-llm-inference-service] - key: node.kubernetes.io/not-ready [e2e-llm-inference-service] operator: Exists [e2e-llm-inference-service] effect: NoExecute [e2e-llm-inference-service] tolerationSeconds: 300 [e2e-llm-inference-service] - key: node.kubernetes.io/unreachable [e2e-llm-inference-service] operator: Exists [e2e-llm-inference-service] effect: NoExecute [e2e-llm-inference-service] tolerationSeconds: 300 [e2e-llm-inference-service] - key: node.kubernetes.io/memory-pressure [e2e-llm-inference-service] operator: Exists [e2e-llm-inference-service] effect: NoSchedule [e2e-llm-inference-service] priority: 0 [e2e-llm-inference-service] enableServiceLinks: true [e2e-llm-inference-service] preemptionPolicy: PreemptLowerPriority [e2e-llm-inference-service] status: [e2e-llm-inference-service] observedGeneration: 1 [e2e-llm-inference-service] phase: Running [e2e-llm-inference-service] conditions: [e2e-llm-inference-service] - type: PodReadyToStartContainers [e2e-llm-inference-service] observedGeneration: 1 [e2e-llm-inference-service] status: 'True' [e2e-llm-inference-service] lastProbeTime: null [e2e-llm-inference-service] lastTransitionTime: '2026-07-07T14:45:57Z' [e2e-llm-inference-service] - type: Initialized [e2e-llm-inference-service] observedGeneration: 1 [e2e-llm-inference-service] status: 'True' [e2e-llm-inference-service] lastProbeTime: null [e2e-llm-inference-service] lastTransitionTime: '2026-07-07T14:46:01Z' [e2e-llm-inference-service] - type: Ready [e2e-llm-inference-service] observedGeneration: 1 [e2e-llm-inference-service] status: 'True' [e2e-llm-inference-service] lastProbeTime: null [e2e-llm-inference-service] lastTransitionTime: '2026-07-07T14:47:16Z' [e2e-llm-inference-service] - type: ContainersReady [e2e-llm-inference-service] observedGeneration: 1 [e2e-llm-inference-service] status: 'True' [e2e-llm-inference-service] lastProbeTime: null [e2e-llm-inference-service] lastTransitionTime: '2026-07-07T14:47:16Z' [e2e-llm-inference-service] - type: PodScheduled [e2e-llm-inference-service] observedGeneration: 1 [e2e-llm-inference-service] status: 'True' [e2e-llm-inference-service] lastProbeTime: null [e2e-llm-inference-service] lastTransitionTime: '2026-07-07T14:45:56Z' [e2e-llm-inference-service] hostIP: 10.0.139.47 [e2e-llm-inference-service] hostIPs: [e2e-llm-inference-service] - ip: 10.0.139.47 [e2e-llm-inference-service] podIP: 10.134.0.32 [e2e-llm-inference-service] podIPs: [e2e-llm-inference-service] - ip: 10.134.0.32 [e2e-llm-inference-service] startTime: '2026-07-07T14:45:56Z' [e2e-llm-inference-service] initContainerStatuses: [e2e-llm-inference-service] - name: storage-initializer [e2e-llm-inference-service] state: [e2e-llm-inference-service] terminated: [e2e-llm-inference-service] exitCode: 0 [e2e-llm-inference-service] reason: Completed [e2e-llm-inference-service] startedAt: '2026-07-07T14:45:56Z' [e2e-llm-inference-service] finishedAt: '2026-07-07T14:46:00Z' [e2e-llm-inference-service] containerID: cri-o://0b6f16cd6d7b965405ab882e89176ad464b1d169ee2d10c3d7780e09b144d0f0 [e2e-llm-inference-service] lastState: {} [e2e-llm-inference-service] ready: true [e2e-llm-inference-service] restartCount: 0 [e2e-llm-inference-service] image: quay.io/opendatahub/kserve-storage-initializer@sha256:13ad765c0f323200e1db7b45c99661287e65c5158d6b54b16139829636d20be4 [e2e-llm-inference-service] imageID: quay.io/opendatahub/kserve-storage-initializer@sha256:13ad765c0f323200e1db7b45c99661287e65c5158d6b54b16139829636d20be4 [e2e-llm-inference-service] containerID: cri-o://0b6f16cd6d7b965405ab882e89176ad464b1d169ee2d10c3d7780e09b144d0f0 [e2e-llm-inference-service] started: false [e2e-llm-inference-service] allocatedResources: [e2e-llm-inference-service] cpu: 100m [e2e-llm-inference-service] memory: 100Mi [e2e-llm-inference-service] resources: [e2e-llm-inference-service] limits: [e2e-llm-inference-service] cpu: '1' [e2e-llm-inference-service] memory: 24Gi [e2e-llm-inference-service] requests: [e2e-llm-inference-service] cpu: 100m [e2e-llm-inference-service] memory: 100Mi [e2e-llm-inference-service] volumeMounts: [e2e-llm-inference-service] - name: kserve-provision-location [e2e-llm-inference-service] mountPath: /mnt [e2e-llm-inference-service] - name: cabundle-cert [e2e-llm-inference-service] mountPath: /etc/ssl/custom-certs [e2e-llm-inference-service] readOnly: true [e2e-llm-inference-service] recursiveReadOnly: Disabled [e2e-llm-inference-service] - name: kube-api-access-tqj64 [e2e-llm-inference-service] mountPath: /var/run/secrets/kubernetes.io/serviceaccount [e2e-llm-inference-service] readOnly: true [e2e-llm-inference-service] recursiveReadOnly: Disabled [e2e-llm-inference-service] user: [e2e-llm-inference-service] linux: [e2e-llm-inference-service] uid: 1000690000 [e2e-llm-inference-service] gid: 0 [e2e-llm-inference-service] supplementalGroups: [e2e-llm-inference-service] - 0 [e2e-llm-inference-service] - 1000690000 [e2e-llm-inference-service] containerStatuses: [e2e-llm-inference-service] - name: main [e2e-llm-inference-service] state: [e2e-llm-inference-service] running: [e2e-llm-inference-service] startedAt: '2026-07-07T14:46:01Z' [e2e-llm-inference-service] lastState: {} [e2e-llm-inference-service] ready: true [e2e-llm-inference-service] restartCount: 0 [e2e-llm-inference-service] image: public.ecr.aws/q9t5s3a7/vllm-cpu-release-repo:v0.19.0 [e2e-llm-inference-service] imageID: public.ecr.aws/q9t5s3a7/vllm-cpu-release-repo@sha256:afb39fca138b51d019d986229d546531b45a2a3deb73bcf59bd42406e13fbba0 [e2e-llm-inference-service] containerID: cri-o://53a794764aed4ea86da3675a236ae5c75a5ea9ee34ea2c2087cb2d019cbfddca [e2e-llm-inference-service] started: true [e2e-llm-inference-service] allocatedResources: [e2e-llm-inference-service] cpu: 200m [e2e-llm-inference-service] memory: 2Gi [e2e-llm-inference-service] resources: [e2e-llm-inference-service] limits: [e2e-llm-inference-service] cpu: '2' [e2e-llm-inference-service] memory: 7Gi [e2e-llm-inference-service] requests: [e2e-llm-inference-service] cpu: 200m [e2e-llm-inference-service] memory: 2Gi [e2e-llm-inference-service] volumeMounts: [e2e-llm-inference-service] - name: home [e2e-llm-inference-service] mountPath: /home [e2e-llm-inference-service] - name: tmp-dir [e2e-llm-inference-service] mountPath: /tmp [e2e-llm-inference-service] - name: dshm [e2e-llm-inference-service] mountPath: /dev/shm [e2e-llm-inference-service] - name: model-cache [e2e-llm-inference-service] mountPath: /models [e2e-llm-inference-service] - name: tls-certs [e2e-llm-inference-service] mountPath: /var/run/kserve/tls [e2e-llm-inference-service] readOnly: true [e2e-llm-inference-service] recursiveReadOnly: Disabled [e2e-llm-inference-service] - name: kserve-provision-location [e2e-llm-inference-service] mountPath: /mnt [e2e-llm-inference-service] readOnly: true [e2e-llm-inference-service] recursiveReadOnly: Disabled [e2e-llm-inference-service] - name: kube-api-access-tqj64 [e2e-llm-inference-service] mountPath: /var/run/secrets/kubernetes.io/serviceaccount [e2e-llm-inference-service] readOnly: true [e2e-llm-inference-service] recursiveReadOnly: Disabled [e2e-llm-inference-service] user: [e2e-llm-inference-service] linux: [e2e-llm-inference-service] uid: 1000690000 [e2e-llm-inference-service] gid: 0 [e2e-llm-inference-service] supplementalGroups: [e2e-llm-inference-service] - 0 [e2e-llm-inference-service] - 1000690000 [e2e-llm-inference-service] qosClass: Burstable [e2e-llm-inference-service] apiVersion: v1 [e2e-llm-inference-service] kind: Pod [e2e-llm-inference-service] [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1254 --- [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1255 metadata: [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-ba4d693a-kserve-router-sche4ghr9 [e2e-llm-inference-service] generateName: llmisvc-model-fb-opt-125m-with-ba4d693a-kserve-router-scheduler-79f5455bb- [e2e-llm-inference-service] namespace: kserve-ci-e2e-test [e2e-llm-inference-service] uid: 66635d02-cb85-4e6d-8342-c235b66670ec [e2e-llm-inference-service] resourceVersion: '46618' [e2e-llm-inference-service] generation: 1 [e2e-llm-inference-service] creationTimestamp: '2026-07-07T14:45:56Z' [e2e-llm-inference-service] labels: [e2e-llm-inference-service] app.kubernetes.io/component: llminferenceservice-router-scheduler [e2e-llm-inference-service] app.kubernetes.io/name: llmisvc-model-fb-opt-125m-with-ba4d693a [e2e-llm-inference-service] app.kubernetes.io/part-of: llminferenceservice [e2e-llm-inference-service] pod-template-hash: 79f5455bb [e2e-llm-inference-service] annotations: [e2e-llm-inference-service] app.kubernetes.io/version: 0.9.0 [e2e-llm-inference-service] certificates.kserve.io/expiration-v2: 'true' [e2e-llm-inference-service] k8s.ovn.org/pod-networks: '{"default":{"ip_addresses":["10.132.0.54/23"],"mac_address":"0a:58:0a:84:00:36","gateway_ips":["10.132.0.1"],"routes":[{"dest":"10.132.0.0/14","nextHop":"10.132.0.1"},{"dest":"172.31.0.0/16","nextHop":"10.132.0.1"},{"dest":"169.254.0.5/32","nextHop":"10.132.0.1"},{"dest":"100.64.0.0/16","nextHop":"10.132.0.1"}],"ip_address":"10.132.0.54/23","gateway_ip":"10.132.0.1","role":"primary"}}' [e2e-llm-inference-service] k8s.v1.cni.cncf.io/network-status: "[{\n \"name\": \"ovn-kubernetes\",\n \ [e2e-llm-inference-service] \ \"interface\": \"eth0\",\n \"ips\": [\n \"10.132.0.54\"\n ],\n\ [e2e-llm-inference-service] \ \"mac\": \"0a:58:0a:84:00:36\",\n \"default\": true,\n \"dns\": {}\n\ [e2e-llm-inference-service] }]" [e2e-llm-inference-service] openshift.io/scc: restricted-v2 [e2e-llm-inference-service] seccomp.security.alpha.kubernetes.io/pod: runtime/default [e2e-llm-inference-service] security.openshift.io/validated-scc-subject-type: user [e2e-llm-inference-service] ownerReferences: [e2e-llm-inference-service] - apiVersion: apps/v1 [e2e-llm-inference-service] kind: ReplicaSet [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-ba4d693a-kserve-router-scheduler-79f5455bb [e2e-llm-inference-service] uid: 05afd940-da33-41e2-a13a-10e5c0538851 [e2e-llm-inference-service] controller: true [e2e-llm-inference-service] blockOwnerDeletion: true [e2e-llm-inference-service] managedFields: [e2e-llm-inference-service] - manager: ip-10-0-133-25 [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] apiVersion: v1 [e2e-llm-inference-service] time: '2026-07-07T14:45:56Z' [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:metadata: [e2e-llm-inference-service] f:annotations: [e2e-llm-inference-service] f:k8s.ovn.org/pod-networks: {} [e2e-llm-inference-service] subresource: status [e2e-llm-inference-service] - manager: kube-controller-manager [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] apiVersion: v1 [e2e-llm-inference-service] time: '2026-07-07T14:45:56Z' [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:metadata: [e2e-llm-inference-service] f:annotations: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:app.kubernetes.io/version: {} [e2e-llm-inference-service] f:certificates.kserve.io/expiration-v2: {} [e2e-llm-inference-service] f:generateName: {} [e2e-llm-inference-service] f:labels: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:app.kubernetes.io/component: {} [e2e-llm-inference-service] f:app.kubernetes.io/name: {} [e2e-llm-inference-service] f:app.kubernetes.io/part-of: {} [e2e-llm-inference-service] f:pod-template-hash: {} [e2e-llm-inference-service] f:ownerReferences: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"uid":"05afd940-da33-41e2-a13a-10e5c0538851"}: {} [e2e-llm-inference-service] f:spec: [e2e-llm-inference-service] f:containers: [e2e-llm-inference-service] k:{"name":"main"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:args: {} [e2e-llm-inference-service] f:command: {} [e2e-llm-inference-service] f:env: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"name":"SSL_CERT_DIR"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] f:image: {} [e2e-llm-inference-service] f:imagePullPolicy: {} [e2e-llm-inference-service] f:lifecycle: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:preStop: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:exec: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:command: {} [e2e-llm-inference-service] f:livenessProbe: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:failureThreshold: {} [e2e-llm-inference-service] f:grpc: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:port: {} [e2e-llm-inference-service] f:service: {} [e2e-llm-inference-service] f:initialDelaySeconds: {} [e2e-llm-inference-service] f:periodSeconds: {} [e2e-llm-inference-service] f:successThreshold: {} [e2e-llm-inference-service] f:timeoutSeconds: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:ports: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"containerPort":5557,"protocol":"TCP"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:containerPort: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:protocol: {} [e2e-llm-inference-service] k:{"containerPort":9002,"protocol":"TCP"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:containerPort: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:protocol: {} [e2e-llm-inference-service] k:{"containerPort":9003,"protocol":"TCP"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:containerPort: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:protocol: {} [e2e-llm-inference-service] k:{"containerPort":9090,"protocol":"TCP"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:containerPort: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:protocol: {} [e2e-llm-inference-service] f:readinessProbe: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:failureThreshold: {} [e2e-llm-inference-service] f:grpc: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:port: {} [e2e-llm-inference-service] f:service: {} [e2e-llm-inference-service] f:initialDelaySeconds: {} [e2e-llm-inference-service] f:periodSeconds: {} [e2e-llm-inference-service] f:successThreshold: {} [e2e-llm-inference-service] f:timeoutSeconds: {} [e2e-llm-inference-service] f:resources: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:requests: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:cpu: {} [e2e-llm-inference-service] f:memory: {} [e2e-llm-inference-service] f:securityContext: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:allowPrivilegeEscalation: {} [e2e-llm-inference-service] f:capabilities: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:drop: {} [e2e-llm-inference-service] f:readOnlyRootFilesystem: {} [e2e-llm-inference-service] f:runAsNonRoot: {} [e2e-llm-inference-service] f:seccompProfile: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:type: {} [e2e-llm-inference-service] f:terminationMessagePath: {} [e2e-llm-inference-service] f:terminationMessagePolicy: {} [e2e-llm-inference-service] f:volumeMounts: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"mountPath":"/tmp/tokenizer"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:mountPath: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] k:{"mountPath":"/var/run/kserve/tls"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:mountPath: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:readOnly: {} [e2e-llm-inference-service] f:dnsPolicy: {} [e2e-llm-inference-service] f:enableServiceLinks: {} [e2e-llm-inference-service] f:initContainers: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"name":"storage-initializer"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:args: {} [e2e-llm-inference-service] f:env: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"name":"AWS_ACCESS_KEY_ID"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:valueFrom: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:secretKeyRef: {} [e2e-llm-inference-service] k:{"name":"AWS_CA_BUNDLE"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"AWS_CA_BUNDLE_CONFIGMAP"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"AWS_ENDPOINT_URL"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"AWS_SECRET_ACCESS_KEY"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:valueFrom: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:secretKeyRef: {} [e2e-llm-inference-service] k:{"name":"HF_HUB_ENABLE_HF_TRANSFER"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"HF_XET_HIGH_PERFORMANCE"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"HF_XET_NUM_CONCURRENT_RANGE_GETS"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"S3_ENDPOINT"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"S3_USE_HTTPS"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"S3_VERIFY_SSL"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"STORAGE_ALLOW_PATTERNS"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] f:image: {} [e2e-llm-inference-service] f:imagePullPolicy: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:resources: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:limits: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:cpu: {} [e2e-llm-inference-service] f:memory: {} [e2e-llm-inference-service] f:requests: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:cpu: {} [e2e-llm-inference-service] f:memory: {} [e2e-llm-inference-service] f:terminationMessagePath: {} [e2e-llm-inference-service] f:terminationMessagePolicy: {} [e2e-llm-inference-service] f:volumeMounts: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"mountPath":"/mnt/models"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:mountPath: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:restartPolicy: {} [e2e-llm-inference-service] f:schedulerName: {} [e2e-llm-inference-service] f:securityContext: {} [e2e-llm-inference-service] f:serviceAccount: {} [e2e-llm-inference-service] f:serviceAccountName: {} [e2e-llm-inference-service] f:terminationGracePeriodSeconds: {} [e2e-llm-inference-service] f:volumes: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"name":"kserve-provision-location"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:emptyDir: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] k:{"name":"tls-certs"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:secret: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:defaultMode: {} [e2e-llm-inference-service] f:secretName: {} [e2e-llm-inference-service] k:{"name":"tokenizer-cache"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:emptyDir: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] k:{"name":"tokenizer-tmp"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:emptyDir: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] k:{"name":"tokenizer-uds"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:emptyDir: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] - manager: multus-daemon [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] apiVersion: v1 [e2e-llm-inference-service] time: '2026-07-07T14:45:56Z' [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:metadata: [e2e-llm-inference-service] f:annotations: [e2e-llm-inference-service] f:k8s.v1.cni.cncf.io/network-status: {} [e2e-llm-inference-service] subresource: status [e2e-llm-inference-service] - manager: kubelet [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] apiVersion: v1 [e2e-llm-inference-service] time: '2026-07-07T14:46:29Z' [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:status: [e2e-llm-inference-service] f:conditions: [e2e-llm-inference-service] k:{"type":"ContainersReady"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:lastProbeTime: {} [e2e-llm-inference-service] f:lastTransitionTime: {} [e2e-llm-inference-service] f:observedGeneration: {} [e2e-llm-inference-service] f:status: {} [e2e-llm-inference-service] f:type: {} [e2e-llm-inference-service] k:{"type":"Initialized"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:lastProbeTime: {} [e2e-llm-inference-service] f:lastTransitionTime: {} [e2e-llm-inference-service] f:observedGeneration: {} [e2e-llm-inference-service] f:status: {} [e2e-llm-inference-service] f:type: {} [e2e-llm-inference-service] k:{"type":"PodReadyToStartContainers"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:lastProbeTime: {} [e2e-llm-inference-service] f:lastTransitionTime: {} [e2e-llm-inference-service] f:observedGeneration: {} [e2e-llm-inference-service] f:status: {} [e2e-llm-inference-service] f:type: {} [e2e-llm-inference-service] k:{"type":"PodScheduled"}: [e2e-llm-inference-service] f:observedGeneration: {} [e2e-llm-inference-service] k:{"type":"Ready"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:lastProbeTime: {} [e2e-llm-inference-service] f:lastTransitionTime: {} [e2e-llm-inference-service] f:observedGeneration: {} [e2e-llm-inference-service] f:status: {} [e2e-llm-inference-service] f:type: {} [e2e-llm-inference-service] f:containerStatuses: {} [e2e-llm-inference-service] f:hostIP: {} [e2e-llm-inference-service] f:hostIPs: {} [e2e-llm-inference-service] f:initContainerStatuses: {} [e2e-llm-inference-service] f:observedGeneration: {} [e2e-llm-inference-service] f:phase: {} [e2e-llm-inference-service] f:podIP: {} [e2e-llm-inference-service] f:podIPs: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"ip":"10.132.0.54"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:ip: {} [e2e-llm-inference-service] f:startTime: {} [e2e-llm-inference-service] subresource: status [e2e-llm-inference-service] spec: [e2e-llm-inference-service] volumes: [e2e-llm-inference-service] - name: tls-certs [e2e-llm-inference-service] secret: [e2e-llm-inference-service] secretName: llmisv77ff2528d3e9b4972cd9335229fce9f0-kserve-self-signed-certs [e2e-llm-inference-service] defaultMode: 420 [e2e-llm-inference-service] - name: tokenizer-uds [e2e-llm-inference-service] emptyDir: {} [e2e-llm-inference-service] - name: tokenizer-tmp [e2e-llm-inference-service] emptyDir: {} [e2e-llm-inference-service] - name: tokenizer-cache [e2e-llm-inference-service] emptyDir: {} [e2e-llm-inference-service] - name: kserve-provision-location [e2e-llm-inference-service] emptyDir: {} [e2e-llm-inference-service] - name: kube-api-access-247lk [e2e-llm-inference-service] projected: [e2e-llm-inference-service] sources: [e2e-llm-inference-service] - serviceAccountToken: [e2e-llm-inference-service] expirationSeconds: 3607 [e2e-llm-inference-service] path: token [e2e-llm-inference-service] - configMap: [e2e-llm-inference-service] name: kube-root-ca.crt [e2e-llm-inference-service] items: [e2e-llm-inference-service] - key: ca.crt [e2e-llm-inference-service] path: ca.crt [e2e-llm-inference-service] - downwardAPI: [e2e-llm-inference-service] items: [e2e-llm-inference-service] - path: namespace [e2e-llm-inference-service] fieldRef: [e2e-llm-inference-service] apiVersion: v1 [e2e-llm-inference-service] fieldPath: metadata.namespace [e2e-llm-inference-service] - configMap: [e2e-llm-inference-service] name: openshift-service-ca.crt [e2e-llm-inference-service] items: [e2e-llm-inference-service] - key: service-ca.crt [e2e-llm-inference-service] path: service-ca.crt [e2e-llm-inference-service] defaultMode: 420 [e2e-llm-inference-service] initContainers: [e2e-llm-inference-service] - name: storage-initializer [e2e-llm-inference-service] image: quay.io/opendatahub/kserve-storage-initializer@sha256:13ad765c0f323200e1db7b45c99661287e65c5158d6b54b16139829636d20be4 [e2e-llm-inference-service] args: [e2e-llm-inference-service] - hf://facebook/opt-125m [e2e-llm-inference-service] - /mnt/models [e2e-llm-inference-service] env: [e2e-llm-inference-service] - name: AWS_ACCESS_KEY_ID [e2e-llm-inference-service] valueFrom: [e2e-llm-inference-service] secretKeyRef: [e2e-llm-inference-service] name: seaweedfs-s3-creds [e2e-llm-inference-service] key: AWS_ACCESS_KEY_ID [e2e-llm-inference-service] - name: AWS_SECRET_ACCESS_KEY [e2e-llm-inference-service] valueFrom: [e2e-llm-inference-service] secretKeyRef: [e2e-llm-inference-service] name: seaweedfs-s3-creds [e2e-llm-inference-service] key: AWS_SECRET_ACCESS_KEY [e2e-llm-inference-service] - name: S3_USE_HTTPS [e2e-llm-inference-service] value: '0' [e2e-llm-inference-service] - name: S3_ENDPOINT [e2e-llm-inference-service] value: s3-service.kserve:8333 [e2e-llm-inference-service] - name: AWS_ENDPOINT_URL [e2e-llm-inference-service] value: http://s3-service.kserve:8333 [e2e-llm-inference-service] - name: S3_VERIFY_SSL [e2e-llm-inference-service] value: '0' [e2e-llm-inference-service] - name: AWS_CA_BUNDLE [e2e-llm-inference-service] value: /etc/ssl/custom-certs/cabundle.crt [e2e-llm-inference-service] - name: AWS_CA_BUNDLE_CONFIGMAP [e2e-llm-inference-service] value: odh-kserve-custom-ca-bundle [e2e-llm-inference-service] - name: HF_HUB_ENABLE_HF_TRANSFER [e2e-llm-inference-service] value: '1' [e2e-llm-inference-service] - name: HF_XET_HIGH_PERFORMANCE [e2e-llm-inference-service] value: '1' [e2e-llm-inference-service] - name: HF_XET_NUM_CONCURRENT_RANGE_GETS [e2e-llm-inference-service] value: '8' [e2e-llm-inference-service] - name: STORAGE_ALLOW_PATTERNS [e2e-llm-inference-service] value: '["tokenizer.json", "tokenizer_config.json", "special_tokens_map.json", [e2e-llm-inference-service] "vocab.json", "merges.txt", "config.json", "generation_config.json"]' [e2e-llm-inference-service] resources: [e2e-llm-inference-service] limits: [e2e-llm-inference-service] cpu: '1' [e2e-llm-inference-service] memory: 24Gi [e2e-llm-inference-service] requests: [e2e-llm-inference-service] cpu: 100m [e2e-llm-inference-service] memory: 100Mi [e2e-llm-inference-service] volumeMounts: [e2e-llm-inference-service] - name: kserve-provision-location [e2e-llm-inference-service] mountPath: /mnt/models [e2e-llm-inference-service] - name: kube-api-access-247lk [e2e-llm-inference-service] readOnly: true [e2e-llm-inference-service] mountPath: /var/run/secrets/kubernetes.io/serviceaccount [e2e-llm-inference-service] terminationMessagePath: /dev/termination-log [e2e-llm-inference-service] terminationMessagePolicy: FallbackToLogsOnError [e2e-llm-inference-service] imagePullPolicy: IfNotPresent [e2e-llm-inference-service] securityContext: [e2e-llm-inference-service] capabilities: [e2e-llm-inference-service] drop: [e2e-llm-inference-service] - ALL [e2e-llm-inference-service] runAsUser: 1000690000 [e2e-llm-inference-service] runAsNonRoot: true [e2e-llm-inference-service] allowPrivilegeEscalation: false [e2e-llm-inference-service] containers: [e2e-llm-inference-service] - name: main [e2e-llm-inference-service] image: ghcr.io/llm-d/llm-d-router-endpoint-picker:v0.9.0-rc.2 [e2e-llm-inference-service] command: [e2e-llm-inference-service] - /app/epp [e2e-llm-inference-service] - --pool-name [e2e-llm-inference-service] - llmisvc-model-fb-opt-125m-with-ba4d693a-inference-pool [e2e-llm-inference-service] - --pool-namespace [e2e-llm-inference-service] - kserve-ci-e2e-test [e2e-llm-inference-service] - --zap-encoder [e2e-llm-inference-service] - json [e2e-llm-inference-service] - --grpc-port [e2e-llm-inference-service] - '9002' [e2e-llm-inference-service] - --grpc-health-port [e2e-llm-inference-service] - '9003' [e2e-llm-inference-service] - --enable-cert-reload=true [e2e-llm-inference-service] - --secure-serving=true [e2e-llm-inference-service] - --model-server-metrics-scheme=https [e2e-llm-inference-service] - --cert-path=/var/run/kserve/tls [e2e-llm-inference-service] args: [e2e-llm-inference-service] - --config-text [e2e-llm-inference-service] - "apiVersion: inference.networking.x-k8s.io/v1alpha1\nkind: EndpointPickerConfig\n\ [e2e-llm-inference-service] plugins:\n- type: single-profile-handler\n- type: queue-scorer\n- type: prefix-cache-scorer\n\ [e2e-llm-inference-service] - type: max-score-picker\n- type: lora-affinity-scorer\nschedulingProfiles:\n\ [e2e-llm-inference-service] - name: default\n plugins:\n - pluginRef: lora-affinity-scorer\n weight:\ [e2e-llm-inference-service] \ 4\n - pluginRef: queue-scorer\n weight: 2\n - pluginRef: prefix-cache-scorer\n\ [e2e-llm-inference-service] \ weight: 3\n - pluginRef: max-score-picker\n" [e2e-llm-inference-service] ports: [e2e-llm-inference-service] - name: grpc [e2e-llm-inference-service] containerPort: 9002 [e2e-llm-inference-service] protocol: TCP [e2e-llm-inference-service] - name: grpc-health [e2e-llm-inference-service] containerPort: 9003 [e2e-llm-inference-service] protocol: TCP [e2e-llm-inference-service] - name: metrics [e2e-llm-inference-service] containerPort: 9090 [e2e-llm-inference-service] protocol: TCP [e2e-llm-inference-service] - name: zmq [e2e-llm-inference-service] containerPort: 5557 [e2e-llm-inference-service] protocol: TCP [e2e-llm-inference-service] env: [e2e-llm-inference-service] - name: SSL_CERT_DIR [e2e-llm-inference-service] value: /var/run/kserve/tls:/var/run/secrets/kubernetes.io/serviceaccount:/etc/pki/tls/certs [e2e-llm-inference-service] resources: [e2e-llm-inference-service] requests: [e2e-llm-inference-service] cpu: 256m [e2e-llm-inference-service] memory: 500Mi [e2e-llm-inference-service] volumeMounts: [e2e-llm-inference-service] - name: tls-certs [e2e-llm-inference-service] readOnly: true [e2e-llm-inference-service] mountPath: /var/run/kserve/tls [e2e-llm-inference-service] - name: tokenizer-uds [e2e-llm-inference-service] mountPath: /tmp/tokenizer [e2e-llm-inference-service] - name: kube-api-access-247lk [e2e-llm-inference-service] readOnly: true [e2e-llm-inference-service] mountPath: /var/run/secrets/kubernetes.io/serviceaccount [e2e-llm-inference-service] livenessProbe: [e2e-llm-inference-service] grpc: [e2e-llm-inference-service] port: 9003 [e2e-llm-inference-service] service: liveness [e2e-llm-inference-service] initialDelaySeconds: 5 [e2e-llm-inference-service] timeoutSeconds: 1 [e2e-llm-inference-service] periodSeconds: 10 [e2e-llm-inference-service] successThreshold: 1 [e2e-llm-inference-service] failureThreshold: 3 [e2e-llm-inference-service] readinessProbe: [e2e-llm-inference-service] grpc: [e2e-llm-inference-service] port: 9003 [e2e-llm-inference-service] service: readiness [e2e-llm-inference-service] initialDelaySeconds: 30 [e2e-llm-inference-service] timeoutSeconds: 1 [e2e-llm-inference-service] periodSeconds: 10 [e2e-llm-inference-service] successThreshold: 1 [e2e-llm-inference-service] failureThreshold: 3 [e2e-llm-inference-service] lifecycle: [e2e-llm-inference-service] preStop: [e2e-llm-inference-service] exec: [e2e-llm-inference-service] command: [e2e-llm-inference-service] - /bin/sleep [e2e-llm-inference-service] - '15' [e2e-llm-inference-service] terminationMessagePath: /dev/termination-log [e2e-llm-inference-service] terminationMessagePolicy: FallbackToLogsOnError [e2e-llm-inference-service] imagePullPolicy: IfNotPresent [e2e-llm-inference-service] securityContext: [e2e-llm-inference-service] capabilities: [e2e-llm-inference-service] drop: [e2e-llm-inference-service] - ALL [e2e-llm-inference-service] runAsUser: 1000690000 [e2e-llm-inference-service] runAsNonRoot: true [e2e-llm-inference-service] readOnlyRootFilesystem: true [e2e-llm-inference-service] allowPrivilegeEscalation: false [e2e-llm-inference-service] seccompProfile: [e2e-llm-inference-service] type: RuntimeDefault [e2e-llm-inference-service] restartPolicy: Always [e2e-llm-inference-service] terminationGracePeriodSeconds: 60 [e2e-llm-inference-service] dnsPolicy: ClusterFirst [e2e-llm-inference-service] serviceAccountName: llmisvc-model-fb-opt-125m-with-ba4d693a-epp-sa [e2e-llm-inference-service] serviceAccount: llmisvc-model-fb-opt-125m-with-ba4d693a-epp-sa [e2e-llm-inference-service] nodeName: ip-10-0-133-25.ec2.internal [e2e-llm-inference-service] securityContext: [e2e-llm-inference-service] seLinuxOptions: [e2e-llm-inference-service] level: s0:c26,c20 [e2e-llm-inference-service] fsGroup: 1000690000 [e2e-llm-inference-service] seccompProfile: [e2e-llm-inference-service] type: RuntimeDefault [e2e-llm-inference-service] imagePullSecrets: [e2e-llm-inference-service] - name: default-dockercfg-tbbgr [e2e-llm-inference-service] - name: llmisvc-model-fb-opt-125m-with-ba4d693a-epp-sa-dockercfg-fdstx [e2e-llm-inference-service] schedulerName: default-scheduler [e2e-llm-inference-service] tolerations: [e2e-llm-inference-service] - key: node.kubernetes.io/not-ready [e2e-llm-inference-service] operator: Exists [e2e-llm-inference-service] effect: NoExecute [e2e-llm-inference-service] tolerationSeconds: 300 [e2e-llm-inference-service] - key: node.kubernetes.io/unreachable [e2e-llm-inference-service] operator: Exists [e2e-llm-inference-service] effect: NoExecute [e2e-llm-inference-service] tolerationSeconds: 300 [e2e-llm-inference-service] - key: node.kubernetes.io/memory-pressure [e2e-llm-inference-service] operator: Exists [e2e-llm-inference-service] effect: NoSchedule [e2e-llm-inference-service] priority: 0 [e2e-llm-inference-service] enableServiceLinks: true [e2e-llm-inference-service] preemptionPolicy: PreemptLowerPriority [e2e-llm-inference-service] status: [e2e-llm-inference-service] observedGeneration: 1 [e2e-llm-inference-service] phase: Running [e2e-llm-inference-service] conditions: [e2e-llm-inference-service] - type: PodReadyToStartContainers [e2e-llm-inference-service] observedGeneration: 1 [e2e-llm-inference-service] status: 'True' [e2e-llm-inference-service] lastProbeTime: null [e2e-llm-inference-service] lastTransitionTime: '2026-07-07T14:45:57Z' [e2e-llm-inference-service] - type: Initialized [e2e-llm-inference-service] observedGeneration: 1 [e2e-llm-inference-service] status: 'True' [e2e-llm-inference-service] lastProbeTime: null [e2e-llm-inference-service] lastTransitionTime: '2026-07-07T14:45:57Z' [e2e-llm-inference-service] - type: Ready [e2e-llm-inference-service] observedGeneration: 1 [e2e-llm-inference-service] status: 'True' [e2e-llm-inference-service] lastProbeTime: null [e2e-llm-inference-service] lastTransitionTime: '2026-07-07T14:46:29Z' [e2e-llm-inference-service] - type: ContainersReady [e2e-llm-inference-service] observedGeneration: 1 [e2e-llm-inference-service] status: 'True' [e2e-llm-inference-service] lastProbeTime: null [e2e-llm-inference-service] lastTransitionTime: '2026-07-07T14:46:29Z' [e2e-llm-inference-service] - type: PodScheduled [e2e-llm-inference-service] observedGeneration: 1 [e2e-llm-inference-service] status: 'True' [e2e-llm-inference-service] lastProbeTime: null [e2e-llm-inference-service] lastTransitionTime: '2026-07-07T14:45:56Z' [e2e-llm-inference-service] hostIP: 10.0.133.25 [e2e-llm-inference-service] hostIPs: [e2e-llm-inference-service] - ip: 10.0.133.25 [e2e-llm-inference-service] podIP: 10.132.0.54 [e2e-llm-inference-service] podIPs: [e2e-llm-inference-service] - ip: 10.132.0.54 [e2e-llm-inference-service] startTime: '2026-07-07T14:45:56Z' [e2e-llm-inference-service] initContainerStatuses: [e2e-llm-inference-service] - name: storage-initializer [e2e-llm-inference-service] state: [e2e-llm-inference-service] terminated: [e2e-llm-inference-service] exitCode: 0 [e2e-llm-inference-service] reason: Completed [e2e-llm-inference-service] startedAt: '2026-07-07T14:45:57Z' [e2e-llm-inference-service] finishedAt: '2026-07-07T14:45:57Z' [e2e-llm-inference-service] containerID: cri-o://2215da3a43bfbbdc09dd2fbdabda7f0e6b83329fab9ff6135893ff86eb1dd393 [e2e-llm-inference-service] lastState: {} [e2e-llm-inference-service] ready: true [e2e-llm-inference-service] restartCount: 0 [e2e-llm-inference-service] image: quay.io/opendatahub/kserve-storage-initializer@sha256:13ad765c0f323200e1db7b45c99661287e65c5158d6b54b16139829636d20be4 [e2e-llm-inference-service] imageID: quay.io/opendatahub/kserve-storage-initializer@sha256:13ad765c0f323200e1db7b45c99661287e65c5158d6b54b16139829636d20be4 [e2e-llm-inference-service] containerID: cri-o://2215da3a43bfbbdc09dd2fbdabda7f0e6b83329fab9ff6135893ff86eb1dd393 [e2e-llm-inference-service] started: false [e2e-llm-inference-service] allocatedResources: [e2e-llm-inference-service] cpu: 100m [e2e-llm-inference-service] memory: 100Mi [e2e-llm-inference-service] resources: [e2e-llm-inference-service] limits: [e2e-llm-inference-service] cpu: '1' [e2e-llm-inference-service] memory: 24Gi [e2e-llm-inference-service] requests: [e2e-llm-inference-service] cpu: 100m [e2e-llm-inference-service] memory: 100Mi [e2e-llm-inference-service] volumeMounts: [e2e-llm-inference-service] - name: kserve-provision-location [e2e-llm-inference-service] mountPath: /mnt/models [e2e-llm-inference-service] - name: kube-api-access-247lk [e2e-llm-inference-service] mountPath: /var/run/secrets/kubernetes.io/serviceaccount [e2e-llm-inference-service] readOnly: true [e2e-llm-inference-service] recursiveReadOnly: Disabled [e2e-llm-inference-service] user: [e2e-llm-inference-service] linux: [e2e-llm-inference-service] uid: 1000690000 [e2e-llm-inference-service] gid: 0 [e2e-llm-inference-service] supplementalGroups: [e2e-llm-inference-service] - 0 [e2e-llm-inference-service] - 1000690000 [e2e-llm-inference-service] containerStatuses: [e2e-llm-inference-service] - name: main [e2e-llm-inference-service] state: [e2e-llm-inference-service] running: [e2e-llm-inference-service] startedAt: '2026-07-07T14:45:57Z' [e2e-llm-inference-service] lastState: {} [e2e-llm-inference-service] ready: true [e2e-llm-inference-service] restartCount: 0 [e2e-llm-inference-service] image: ghcr.io/llm-d/llm-d-router-endpoint-picker:v0.9.0-rc.2 [e2e-llm-inference-service] imageID: ghcr.io/llm-d/llm-d-router-endpoint-picker@sha256:06b6c75d77afd0e07053402752a9736c2dfbc12a306d0d37d963aac4c1d4e6a6 [e2e-llm-inference-service] containerID: cri-o://f4702b226fb44b6a901bd2f22bfb55a846e22d839de1160737a56e7863f1e77b [e2e-llm-inference-service] started: true [e2e-llm-inference-service] allocatedResources: [e2e-llm-inference-service] cpu: 256m [e2e-llm-inference-service] memory: 500Mi [e2e-llm-inference-service] resources: [e2e-llm-inference-service] requests: [e2e-llm-inference-service] cpu: 256m [e2e-llm-inference-service] memory: 500Mi [e2e-llm-inference-service] volumeMounts: [e2e-llm-inference-service] - name: tls-certs [e2e-llm-inference-service] mountPath: /var/run/kserve/tls [e2e-llm-inference-service] readOnly: true [e2e-llm-inference-service] recursiveReadOnly: Disabled [e2e-llm-inference-service] - name: tokenizer-uds [e2e-llm-inference-service] mountPath: /tmp/tokenizer [e2e-llm-inference-service] - name: kube-api-access-247lk [e2e-llm-inference-service] mountPath: /var/run/secrets/kubernetes.io/serviceaccount [e2e-llm-inference-service] readOnly: true [e2e-llm-inference-service] recursiveReadOnly: Disabled [e2e-llm-inference-service] user: [e2e-llm-inference-service] linux: [e2e-llm-inference-service] uid: 1000690000 [e2e-llm-inference-service] gid: 0 [e2e-llm-inference-service] supplementalGroups: [e2e-llm-inference-service] - 0 [e2e-llm-inference-service] - 1000690000 [e2e-llm-inference-service] qosClass: Burstable [e2e-llm-inference-service] apiVersion: v1 [e2e-llm-inference-service] kind: Pod [e2e-llm-inference-service] [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1254 --- [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1255 metadata: [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-ba4d693a-epp-sa [e2e-llm-inference-service] namespace: kserve-ci-e2e-test [e2e-llm-inference-service] uid: 8b31a56e-2b64-440f-aff7-0b28597eb860 [e2e-llm-inference-service] resourceVersion: '45955' [e2e-llm-inference-service] creationTimestamp: '2026-07-07T14:45:56Z' [e2e-llm-inference-service] labels: [e2e-llm-inference-service] app.kubernetes.io/component: llminferenceservice-router-scheduler [e2e-llm-inference-service] app.kubernetes.io/name: llmisvc-model-fb-opt-125m-with-ba4d693a [e2e-llm-inference-service] app.kubernetes.io/part-of: llminferenceservice [e2e-llm-inference-service] annotations: [e2e-llm-inference-service] openshift.io/internal-registry-pull-secret-ref: llmisvc-model-fb-opt-125m-with-ba4d693a-epp-sa-dockercfg-fdstx [e2e-llm-inference-service] ownerReferences: [e2e-llm-inference-service] - apiVersion: serving.kserve.io/v1alpha2 [e2e-llm-inference-service] kind: LLMInferenceService [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-ba4d693a [e2e-llm-inference-service] uid: 997d3cef-fabf-4b25-ad95-5ebe7a367fc5 [e2e-llm-inference-service] controller: true [e2e-llm-inference-service] blockOwnerDeletion: true [e2e-llm-inference-service] managedFields: [e2e-llm-inference-service] - manager: openshift.io/image-registry-pull-secrets_service-account-controller [e2e-llm-inference-service] operation: Apply [e2e-llm-inference-service] apiVersion: v1 [e2e-llm-inference-service] time: '2026-07-07T14:45:56Z' [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:imagePullSecrets: {} [e2e-llm-inference-service] f:metadata: [e2e-llm-inference-service] f:annotations: [e2e-llm-inference-service] f:openshift.io/internal-registry-pull-secret-ref: {} [e2e-llm-inference-service] f:secrets: [e2e-llm-inference-service] k:{"name":"llmisvc-model-fb-opt-125m-with-ba4d693a-epp-sa-dockercfg-fdstx"}: {} [e2e-llm-inference-service] - manager: manager [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] apiVersion: v1 [e2e-llm-inference-service] time: '2026-07-07T14:45:56Z' [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:metadata: [e2e-llm-inference-service] f:labels: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:app.kubernetes.io/component: {} [e2e-llm-inference-service] f:app.kubernetes.io/name: {} [e2e-llm-inference-service] f:app.kubernetes.io/part-of: {} [e2e-llm-inference-service] f:ownerReferences: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"uid":"997d3cef-fabf-4b25-ad95-5ebe7a367fc5"}: {} [e2e-llm-inference-service] f:secrets: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"name":"default-dockercfg-tbbgr"}: {} [e2e-llm-inference-service] k:{"name":"seaweedfs-s3-creds"}: {} [e2e-llm-inference-service] secrets: [e2e-llm-inference-service] - name: default-dockercfg-tbbgr [e2e-llm-inference-service] - name: seaweedfs-s3-creds [e2e-llm-inference-service] - name: llmisvc-model-fb-opt-125m-with-ba4d693a-epp-sa-dockercfg-fdstx [e2e-llm-inference-service] imagePullSecrets: [e2e-llm-inference-service] - name: default-dockercfg-tbbgr [e2e-llm-inference-service] - name: llmisvc-model-fb-opt-125m-with-ba4d693a-epp-sa-dockercfg-fdstx [e2e-llm-inference-service] apiVersion: v1 [e2e-llm-inference-service] kind: ServiceAccount [e2e-llm-inference-service] [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1254 --- [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1255 metadata: [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-ba4d693a-epp-service [e2e-llm-inference-service] namespace: kserve-ci-e2e-test [e2e-llm-inference-service] uid: 27ed2141-ccc0-4ffe-aeed-63126edda97e [e2e-llm-inference-service] resourceVersion: '45978' [e2e-llm-inference-service] creationTimestamp: '2026-07-07T14:45:56Z' [e2e-llm-inference-service] labels: [e2e-llm-inference-service] app.kubernetes.io/component: llminferenceservice-router-scheduler [e2e-llm-inference-service] app.kubernetes.io/name: llmisvc-model-fb-opt-125m-with-ba4d693a [e2e-llm-inference-service] app.kubernetes.io/part-of: llminferenceservice [e2e-llm-inference-service] ownerReferences: [e2e-llm-inference-service] - apiVersion: serving.kserve.io/v1alpha2 [e2e-llm-inference-service] kind: LLMInferenceService [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-ba4d693a [e2e-llm-inference-service] uid: 997d3cef-fabf-4b25-ad95-5ebe7a367fc5 [e2e-llm-inference-service] controller: true [e2e-llm-inference-service] blockOwnerDeletion: true [e2e-llm-inference-service] managedFields: [e2e-llm-inference-service] - manager: manager [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] apiVersion: v1 [e2e-llm-inference-service] time: '2026-07-07T14:45:56Z' [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:metadata: [e2e-llm-inference-service] f:labels: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:app.kubernetes.io/component: {} [e2e-llm-inference-service] f:app.kubernetes.io/name: {} [e2e-llm-inference-service] f:app.kubernetes.io/part-of: {} [e2e-llm-inference-service] f:ownerReferences: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"uid":"997d3cef-fabf-4b25-ad95-5ebe7a367fc5"}: {} [e2e-llm-inference-service] f:spec: [e2e-llm-inference-service] f:internalTrafficPolicy: {} [e2e-llm-inference-service] f:ports: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"port":5557,"protocol":"TCP"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:port: {} [e2e-llm-inference-service] f:protocol: {} [e2e-llm-inference-service] f:targetPort: {} [e2e-llm-inference-service] k:{"port":9002,"protocol":"TCP"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:port: {} [e2e-llm-inference-service] f:protocol: {} [e2e-llm-inference-service] f:targetPort: {} [e2e-llm-inference-service] k:{"port":9003,"protocol":"TCP"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:port: {} [e2e-llm-inference-service] f:protocol: {} [e2e-llm-inference-service] f:targetPort: {} [e2e-llm-inference-service] k:{"port":9090,"protocol":"TCP"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:port: {} [e2e-llm-inference-service] f:protocol: {} [e2e-llm-inference-service] f:targetPort: {} [e2e-llm-inference-service] f:selector: {} [e2e-llm-inference-service] f:sessionAffinity: {} [e2e-llm-inference-service] f:type: {} [e2e-llm-inference-service] spec: [e2e-llm-inference-service] ports: [e2e-llm-inference-service] - name: grpc [e2e-llm-inference-service] protocol: TCP [e2e-llm-inference-service] port: 9002 [e2e-llm-inference-service] targetPort: grpc [e2e-llm-inference-service] - name: grpc-health [e2e-llm-inference-service] protocol: TCP [e2e-llm-inference-service] port: 9003 [e2e-llm-inference-service] targetPort: grpc-health [e2e-llm-inference-service] - name: metrics [e2e-llm-inference-service] protocol: TCP [e2e-llm-inference-service] port: 9090 [e2e-llm-inference-service] targetPort: metrics [e2e-llm-inference-service] - name: zmq [e2e-llm-inference-service] protocol: TCP [e2e-llm-inference-service] port: 5557 [e2e-llm-inference-service] targetPort: zmq [e2e-llm-inference-service] selector: [e2e-llm-inference-service] app.kubernetes.io/component: llminferenceservice-router-scheduler [e2e-llm-inference-service] app.kubernetes.io/name: llmisvc-model-fb-opt-125m-with-ba4d693a [e2e-llm-inference-service] app.kubernetes.io/part-of: llminferenceservice [e2e-llm-inference-service] clusterIP: 172.31.119.80 [e2e-llm-inference-service] clusterIPs: [e2e-llm-inference-service] - 172.31.119.80 [e2e-llm-inference-service] type: ClusterIP [e2e-llm-inference-service] sessionAffinity: None [e2e-llm-inference-service] ipFamilies: [e2e-llm-inference-service] - IPv4 [e2e-llm-inference-service] ipFamilyPolicy: SingleStack [e2e-llm-inference-service] internalTrafficPolicy: Cluster [e2e-llm-inference-service] status: [e2e-llm-inference-service] loadBalancer: {} [e2e-llm-inference-service] apiVersion: v1 [e2e-llm-inference-service] kind: Service [e2e-llm-inference-service] [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1254 --- [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1255 metadata: [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-ba4d693a-kserve-workload-svc [e2e-llm-inference-service] namespace: kserve-ci-e2e-test [e2e-llm-inference-service] uid: 687542da-fbcf-42b8-ae26-8d19e584f912 [e2e-llm-inference-service] resourceVersion: '45940' [e2e-llm-inference-service] creationTimestamp: '2026-07-07T14:45:56Z' [e2e-llm-inference-service] labels: [e2e-llm-inference-service] app.kubernetes.io/component: llminferenceservice-workload [e2e-llm-inference-service] app.kubernetes.io/name: llmisvc-model-fb-opt-125m-with-ba4d693a [e2e-llm-inference-service] app.kubernetes.io/part-of: llminferenceservice [e2e-llm-inference-service] ownerReferences: [e2e-llm-inference-service] - apiVersion: serving.kserve.io/v1alpha2 [e2e-llm-inference-service] kind: LLMInferenceService [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-ba4d693a [e2e-llm-inference-service] uid: 997d3cef-fabf-4b25-ad95-5ebe7a367fc5 [e2e-llm-inference-service] controller: true [e2e-llm-inference-service] blockOwnerDeletion: true [e2e-llm-inference-service] managedFields: [e2e-llm-inference-service] - manager: manager [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] apiVersion: v1 [e2e-llm-inference-service] time: '2026-07-07T14:45:56Z' [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:metadata: [e2e-llm-inference-service] f:labels: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:app.kubernetes.io/component: {} [e2e-llm-inference-service] f:app.kubernetes.io/name: {} [e2e-llm-inference-service] f:app.kubernetes.io/part-of: {} [e2e-llm-inference-service] f:ownerReferences: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"uid":"997d3cef-fabf-4b25-ad95-5ebe7a367fc5"}: {} [e2e-llm-inference-service] f:spec: [e2e-llm-inference-service] f:internalTrafficPolicy: {} [e2e-llm-inference-service] f:ports: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"port":8000,"protocol":"TCP"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:appProtocol: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:port: {} [e2e-llm-inference-service] f:protocol: {} [e2e-llm-inference-service] f:targetPort: {} [e2e-llm-inference-service] f:selector: {} [e2e-llm-inference-service] f:sessionAffinity: {} [e2e-llm-inference-service] f:type: {} [e2e-llm-inference-service] spec: [e2e-llm-inference-service] ports: [e2e-llm-inference-service] - name: https [e2e-llm-inference-service] protocol: TCP [e2e-llm-inference-service] appProtocol: https [e2e-llm-inference-service] port: 8000 [e2e-llm-inference-service] targetPort: 8000 [e2e-llm-inference-service] selector: [e2e-llm-inference-service] app.kubernetes.io/name: llmisvc-model-fb-opt-125m-with-ba4d693a [e2e-llm-inference-service] app.kubernetes.io/part-of: llminferenceservice [e2e-llm-inference-service] kserve.io/component: workload [e2e-llm-inference-service] clusterIP: 172.31.251.172 [e2e-llm-inference-service] clusterIPs: [e2e-llm-inference-service] - 172.31.251.172 [e2e-llm-inference-service] type: ClusterIP [e2e-llm-inference-service] sessionAffinity: None [e2e-llm-inference-service] ipFamilies: [e2e-llm-inference-service] - IPv4 [e2e-llm-inference-service] ipFamilyPolicy: SingleStack [e2e-llm-inference-service] internalTrafficPolicy: Cluster [e2e-llm-inference-service] status: [e2e-llm-inference-service] loadBalancer: {} [e2e-llm-inference-service] apiVersion: v1 [e2e-llm-inference-service] kind: Service [e2e-llm-inference-service] [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1254 --- [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1255 metadata: [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-ba4d693a-kserve [e2e-llm-inference-service] namespace: kserve-ci-e2e-test [e2e-llm-inference-service] uid: 834ec5b2-f5f3-4037-88ef-04cb98318a8c [e2e-llm-inference-service] resourceVersion: '47366' [e2e-llm-inference-service] generation: 1 [e2e-llm-inference-service] creationTimestamp: '2026-07-07T14:45:56Z' [e2e-llm-inference-service] labels: [e2e-llm-inference-service] app.kubernetes.io/component: llminferenceservice-workload [e2e-llm-inference-service] app.kubernetes.io/name: llmisvc-model-fb-opt-125m-with-ba4d693a [e2e-llm-inference-service] app.kubernetes.io/part-of: llminferenceservice [e2e-llm-inference-service] kserve.io/component: workload [e2e-llm-inference-service] llm-d.ai/role: both [e2e-llm-inference-service] annotations: [e2e-llm-inference-service] deployment.kubernetes.io/revision: '1' [e2e-llm-inference-service] ownerReferences: [e2e-llm-inference-service] - apiVersion: serving.kserve.io/v1alpha2 [e2e-llm-inference-service] kind: LLMInferenceService [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-ba4d693a [e2e-llm-inference-service] uid: 997d3cef-fabf-4b25-ad95-5ebe7a367fc5 [e2e-llm-inference-service] controller: true [e2e-llm-inference-service] blockOwnerDeletion: true [e2e-llm-inference-service] managedFields: [e2e-llm-inference-service] - manager: manager [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] apiVersion: apps/v1 [e2e-llm-inference-service] time: '2026-07-07T14:45:56Z' [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:metadata: [e2e-llm-inference-service] f:labels: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:app.kubernetes.io/component: {} [e2e-llm-inference-service] f:app.kubernetes.io/name: {} [e2e-llm-inference-service] f:app.kubernetes.io/part-of: {} [e2e-llm-inference-service] f:kserve.io/component: {} [e2e-llm-inference-service] f:llm-d.ai/role: {} [e2e-llm-inference-service] f:ownerReferences: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"uid":"997d3cef-fabf-4b25-ad95-5ebe7a367fc5"}: {} [e2e-llm-inference-service] f:spec: [e2e-llm-inference-service] f:progressDeadlineSeconds: {} [e2e-llm-inference-service] f:replicas: {} [e2e-llm-inference-service] f:revisionHistoryLimit: {} [e2e-llm-inference-service] f:selector: {} [e2e-llm-inference-service] f:strategy: [e2e-llm-inference-service] f:rollingUpdate: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:maxSurge: {} [e2e-llm-inference-service] f:maxUnavailable: {} [e2e-llm-inference-service] f:type: {} [e2e-llm-inference-service] f:template: [e2e-llm-inference-service] f:metadata: [e2e-llm-inference-service] f:labels: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:app.kubernetes.io/component: {} [e2e-llm-inference-service] f:app.kubernetes.io/name: {} [e2e-llm-inference-service] f:app.kubernetes.io/part-of: {} [e2e-llm-inference-service] f:kserve.io/component: {} [e2e-llm-inference-service] f:llm-d.ai/role: {} [e2e-llm-inference-service] f:spec: [e2e-llm-inference-service] f:containers: [e2e-llm-inference-service] k:{"name":"main"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:args: {} [e2e-llm-inference-service] f:command: {} [e2e-llm-inference-service] f:env: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"name":"HF_HUB_CACHE"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"HOME"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"TORCHINDUCTOR_CACHE_DIR"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"USER"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"VLLM_CPU_KVCACHE_SPACE"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"VLLM_ENABLE_V1_MULTIPROCESSING"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"VLLM_LOGGING_LEVEL"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] f:image: {} [e2e-llm-inference-service] f:imagePullPolicy: {} [e2e-llm-inference-service] f:lifecycle: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:preStop: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:exec: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:command: {} [e2e-llm-inference-service] f:livenessProbe: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:failureThreshold: {} [e2e-llm-inference-service] f:httpGet: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:path: {} [e2e-llm-inference-service] f:port: {} [e2e-llm-inference-service] f:scheme: {} [e2e-llm-inference-service] f:periodSeconds: {} [e2e-llm-inference-service] f:successThreshold: {} [e2e-llm-inference-service] f:timeoutSeconds: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:ports: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"containerPort":8000,"protocol":"TCP"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:containerPort: {} [e2e-llm-inference-service] f:protocol: {} [e2e-llm-inference-service] f:readinessProbe: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:failureThreshold: {} [e2e-llm-inference-service] f:httpGet: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:path: {} [e2e-llm-inference-service] f:port: {} [e2e-llm-inference-service] f:scheme: {} [e2e-llm-inference-service] f:periodSeconds: {} [e2e-llm-inference-service] f:successThreshold: {} [e2e-llm-inference-service] f:timeoutSeconds: {} [e2e-llm-inference-service] f:resources: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:limits: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:cpu: {} [e2e-llm-inference-service] f:memory: {} [e2e-llm-inference-service] f:requests: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:cpu: {} [e2e-llm-inference-service] f:memory: {} [e2e-llm-inference-service] f:securityContext: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:allowPrivilegeEscalation: {} [e2e-llm-inference-service] f:capabilities: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:drop: {} [e2e-llm-inference-service] f:readOnlyRootFilesystem: {} [e2e-llm-inference-service] f:runAsNonRoot: {} [e2e-llm-inference-service] f:seccompProfile: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:type: {} [e2e-llm-inference-service] f:startupProbe: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:failureThreshold: {} [e2e-llm-inference-service] f:httpGet: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:path: {} [e2e-llm-inference-service] f:port: {} [e2e-llm-inference-service] f:scheme: {} [e2e-llm-inference-service] f:periodSeconds: {} [e2e-llm-inference-service] f:successThreshold: {} [e2e-llm-inference-service] f:timeoutSeconds: {} [e2e-llm-inference-service] f:terminationMessagePath: {} [e2e-llm-inference-service] f:terminationMessagePolicy: {} [e2e-llm-inference-service] f:volumeMounts: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"mountPath":"/dev/shm"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:mountPath: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] k:{"mountPath":"/home"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:mountPath: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] k:{"mountPath":"/mnt"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:mountPath: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:readOnly: {} [e2e-llm-inference-service] k:{"mountPath":"/models"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:mountPath: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] k:{"mountPath":"/tmp"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:mountPath: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] k:{"mountPath":"/var/run/kserve/tls"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:mountPath: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:readOnly: {} [e2e-llm-inference-service] f:dnsPolicy: {} [e2e-llm-inference-service] f:initContainers: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"name":"storage-initializer"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:args: {} [e2e-llm-inference-service] f:env: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"name":"AWS_ACCESS_KEY_ID"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:valueFrom: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:secretKeyRef: {} [e2e-llm-inference-service] k:{"name":"AWS_CA_BUNDLE"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"AWS_CA_BUNDLE_CONFIGMAP"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"AWS_ENDPOINT_URL"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"AWS_SECRET_ACCESS_KEY"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:valueFrom: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:secretKeyRef: {} [e2e-llm-inference-service] k:{"name":"CA_BUNDLE_CONFIGMAP_NAME"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"CA_BUNDLE_VOLUME_MOUNT_POINT"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"HF_HUB_ENABLE_HF_TRANSFER"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"HF_XET_HIGH_PERFORMANCE"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"HF_XET_NUM_CONCURRENT_RANGE_GETS"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"S3_ENDPOINT"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"S3_USE_HTTPS"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"S3_VERIFY_SSL"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] f:image: {} [e2e-llm-inference-service] f:imagePullPolicy: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:resources: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:limits: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:cpu: {} [e2e-llm-inference-service] f:memory: {} [e2e-llm-inference-service] f:requests: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:cpu: {} [e2e-llm-inference-service] f:memory: {} [e2e-llm-inference-service] f:terminationMessagePath: {} [e2e-llm-inference-service] f:terminationMessagePolicy: {} [e2e-llm-inference-service] f:volumeMounts: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"mountPath":"/etc/ssl/custom-certs"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:mountPath: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:readOnly: {} [e2e-llm-inference-service] k:{"mountPath":"/mnt"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:mountPath: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:restartPolicy: {} [e2e-llm-inference-service] f:schedulerName: {} [e2e-llm-inference-service] f:securityContext: {} [e2e-llm-inference-service] f:terminationGracePeriodSeconds: {} [e2e-llm-inference-service] f:volumes: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"name":"cabundle-cert"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:configMap: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:defaultMode: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] k:{"name":"dshm"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:emptyDir: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:medium: {} [e2e-llm-inference-service] f:sizeLimit: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] k:{"name":"home"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:emptyDir: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] k:{"name":"kserve-provision-location"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:emptyDir: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] k:{"name":"model-cache"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:emptyDir: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] k:{"name":"tls-certs"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:secret: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:defaultMode: {} [e2e-llm-inference-service] f:secretName: {} [e2e-llm-inference-service] k:{"name":"tmp-dir"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:emptyDir: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] - manager: kube-controller-manager [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] apiVersion: apps/v1 [e2e-llm-inference-service] time: '2026-07-07T14:47:16Z' [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:metadata: [e2e-llm-inference-service] f:annotations: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:deployment.kubernetes.io/revision: {} [e2e-llm-inference-service] f:status: [e2e-llm-inference-service] f:availableReplicas: {} [e2e-llm-inference-service] f:conditions: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"type":"Available"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:lastTransitionTime: {} [e2e-llm-inference-service] f:lastUpdateTime: {} [e2e-llm-inference-service] f:message: {} [e2e-llm-inference-service] f:reason: {} [e2e-llm-inference-service] f:status: {} [e2e-llm-inference-service] f:type: {} [e2e-llm-inference-service] k:{"type":"Progressing"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:lastTransitionTime: {} [e2e-llm-inference-service] f:lastUpdateTime: {} [e2e-llm-inference-service] f:message: {} [e2e-llm-inference-service] f:reason: {} [e2e-llm-inference-service] f:status: {} [e2e-llm-inference-service] f:type: {} [e2e-llm-inference-service] f:observedGeneration: {} [e2e-llm-inference-service] f:readyReplicas: {} [e2e-llm-inference-service] f:replicas: {} [e2e-llm-inference-service] f:updatedReplicas: {} [e2e-llm-inference-service] subresource: status [e2e-llm-inference-service] spec: [e2e-llm-inference-service] replicas: 1 [e2e-llm-inference-service] selector: [e2e-llm-inference-service] matchLabels: [e2e-llm-inference-service] app.kubernetes.io/component: llminferenceservice-workload [e2e-llm-inference-service] app.kubernetes.io/name: llmisvc-model-fb-opt-125m-with-ba4d693a [e2e-llm-inference-service] app.kubernetes.io/part-of: llminferenceservice [e2e-llm-inference-service] kserve.io/component: workload [e2e-llm-inference-service] llm-d.ai/role: both [e2e-llm-inference-service] template: [e2e-llm-inference-service] metadata: [e2e-llm-inference-service] labels: [e2e-llm-inference-service] app.kubernetes.io/component: llminferenceservice-workload [e2e-llm-inference-service] app.kubernetes.io/name: llmisvc-model-fb-opt-125m-with-ba4d693a [e2e-llm-inference-service] app.kubernetes.io/part-of: llminferenceservice [e2e-llm-inference-service] kserve.io/component: workload [e2e-llm-inference-service] llm-d.ai/role: both [e2e-llm-inference-service] spec: [e2e-llm-inference-service] volumes: [e2e-llm-inference-service] - name: home [e2e-llm-inference-service] emptyDir: {} [e2e-llm-inference-service] - name: dshm [e2e-llm-inference-service] emptyDir: [e2e-llm-inference-service] medium: Memory [e2e-llm-inference-service] sizeLimit: 1Gi [e2e-llm-inference-service] - name: model-cache [e2e-llm-inference-service] emptyDir: {} [e2e-llm-inference-service] - name: tmp-dir [e2e-llm-inference-service] emptyDir: {} [e2e-llm-inference-service] - name: tls-certs [e2e-llm-inference-service] secret: [e2e-llm-inference-service] secretName: llmisv77ff2528d3e9b4972cd9335229fce9f0-kserve-self-signed-certs [e2e-llm-inference-service] defaultMode: 420 [e2e-llm-inference-service] - name: kserve-provision-location [e2e-llm-inference-service] emptyDir: {} [e2e-llm-inference-service] - name: cabundle-cert [e2e-llm-inference-service] configMap: [e2e-llm-inference-service] name: odh-kserve-custom-ca-bundle [e2e-llm-inference-service] defaultMode: 420 [e2e-llm-inference-service] initContainers: [e2e-llm-inference-service] - name: storage-initializer [e2e-llm-inference-service] image: quay.io/opendatahub/kserve-storage-initializer@sha256:13ad765c0f323200e1db7b45c99661287e65c5158d6b54b16139829636d20be4 [e2e-llm-inference-service] args: [e2e-llm-inference-service] - hf://facebook/opt-125m [e2e-llm-inference-service] - /mnt/models [e2e-llm-inference-service] - hf://edbeeching/opt-125m-lora [e2e-llm-inference-service] - /mnt/lora/lora-adapter-1 [e2e-llm-inference-service] env: [e2e-llm-inference-service] - name: AWS_ACCESS_KEY_ID [e2e-llm-inference-service] valueFrom: [e2e-llm-inference-service] secretKeyRef: [e2e-llm-inference-service] name: seaweedfs-s3-creds [e2e-llm-inference-service] key: AWS_ACCESS_KEY_ID [e2e-llm-inference-service] - name: AWS_SECRET_ACCESS_KEY [e2e-llm-inference-service] valueFrom: [e2e-llm-inference-service] secretKeyRef: [e2e-llm-inference-service] name: seaweedfs-s3-creds [e2e-llm-inference-service] key: AWS_SECRET_ACCESS_KEY [e2e-llm-inference-service] - name: S3_USE_HTTPS [e2e-llm-inference-service] value: '0' [e2e-llm-inference-service] - name: S3_ENDPOINT [e2e-llm-inference-service] value: s3-service.kserve:8333 [e2e-llm-inference-service] - name: AWS_ENDPOINT_URL [e2e-llm-inference-service] value: http://s3-service.kserve:8333 [e2e-llm-inference-service] - name: S3_VERIFY_SSL [e2e-llm-inference-service] value: '0' [e2e-llm-inference-service] - name: AWS_CA_BUNDLE [e2e-llm-inference-service] value: /etc/ssl/custom-certs/cabundle.crt [e2e-llm-inference-service] - name: AWS_CA_BUNDLE_CONFIGMAP [e2e-llm-inference-service] value: odh-kserve-custom-ca-bundle [e2e-llm-inference-service] - name: HF_HUB_ENABLE_HF_TRANSFER [e2e-llm-inference-service] value: '1' [e2e-llm-inference-service] - name: HF_XET_HIGH_PERFORMANCE [e2e-llm-inference-service] value: '1' [e2e-llm-inference-service] - name: HF_XET_NUM_CONCURRENT_RANGE_GETS [e2e-llm-inference-service] value: '8' [e2e-llm-inference-service] - name: CA_BUNDLE_CONFIGMAP_NAME [e2e-llm-inference-service] value: odh-kserve-custom-ca-bundle [e2e-llm-inference-service] - name: CA_BUNDLE_VOLUME_MOUNT_POINT [e2e-llm-inference-service] value: /etc/ssl/custom-certs [e2e-llm-inference-service] resources: [e2e-llm-inference-service] limits: [e2e-llm-inference-service] cpu: '1' [e2e-llm-inference-service] memory: 24Gi [e2e-llm-inference-service] requests: [e2e-llm-inference-service] cpu: 100m [e2e-llm-inference-service] memory: 100Mi [e2e-llm-inference-service] volumeMounts: [e2e-llm-inference-service] - name: kserve-provision-location [e2e-llm-inference-service] mountPath: /mnt [e2e-llm-inference-service] - name: cabundle-cert [e2e-llm-inference-service] readOnly: true [e2e-llm-inference-service] mountPath: /etc/ssl/custom-certs [e2e-llm-inference-service] terminationMessagePath: /dev/termination-log [e2e-llm-inference-service] terminationMessagePolicy: FallbackToLogsOnError [e2e-llm-inference-service] imagePullPolicy: IfNotPresent [e2e-llm-inference-service] containers: [e2e-llm-inference-service] - name: main [e2e-llm-inference-service] image: public.ecr.aws/q9t5s3a7/vllm-cpu-release-repo:v0.19.0 [e2e-llm-inference-service] command: [e2e-llm-inference-service] - /bin/bash [e2e-llm-inference-service] - -c [e2e-llm-inference-service] - "if [ -f /etc/profile.d/ibm-aiu-setup.sh ]; then\n source /etc/profile.d/ibm-aiu-setup.sh\n\ [e2e-llm-inference-service] fi\n\nif [ \"$KSERVE_INFER_ROCE\" = \"true\" ]; then\n echo \"Trying to\ [e2e-llm-inference-service] \ infer RoCE configs ... \"\n grep -H . /sys/class/infiniband/*/ports/*/gids/*\ [e2e-llm-inference-service] \ 2>/dev/null\n grep -H . /sys/class/infiniband/*/ports/*/gid_attrs/types/*\ [e2e-llm-inference-service] \ 2>/dev/null\n\n cat /proc/driver/nvidia/params\n\n KSERVE_INFER_IB_GID_INDEX_GREP=${KSERVE_INFER_IB_GID_INDEX_GREP:-\"\ [e2e-llm-inference-service] RoCE v2\"}\n\n echo \"[Infer RoCE] Discovering active HCAs ...\"\n active_hcas=()\n\ [e2e-llm-inference-service] \ # Loop through all mlx5 devices found in sysfs\n for hca_dir in /sys/class/infiniband/mlx5_*;\ [e2e-llm-inference-service] \ do\n # Ensure it's a directory before proceeding\n if [ -d \"\ [e2e-llm-inference-service] $hca_dir\" ]; then\n hca_name=$(basename \"$hca_dir\")\n \ [e2e-llm-inference-service] \ port_state_file=\"$hca_dir/ports/1/state\" # Assume port 1\n \ [e2e-llm-inference-service] \ type_file=\"$hca_dir/ports/1/gid_attrs/types/*\"\n\n echo\ [e2e-llm-inference-service] \ \"[Infer RoCE] Check if the port state file ${port_state_file} exists\ [e2e-llm-inference-service] \ and contains 'ACTIVE'\"\n if [ -f \"$port_state_file\" ] && grep\ [e2e-llm-inference-service] \ -q \"ACTIVE\" \"$port_state_file\" && grep -q \"${KSERVE_INFER_IB_GID_INDEX_GREP}\"\ [e2e-llm-inference-service] \ ${type_file} 2>/dev/null; then\n echo \"[Infer RoCE] Found\ [e2e-llm-inference-service] \ active HCA: $hca_name\"\n active_hcas+=(\"$hca_name\")\n\ [e2e-llm-inference-service] \ else\n echo \"[Infer RoCE] Skipping inactive or\ [e2e-llm-inference-service] \ down HCA: $hca_name\"\n fi\n fi\n done\n\n # Check if\ [e2e-llm-inference-service] \ we found any active HCAs\n if [ ${#active_hcas[@]} -gt 0 ]; then\n \ [e2e-llm-inference-service] \ # Join the array elements with a comma\n hca_port_pairs=()\n \ [e2e-llm-inference-service] \ for hca in \"${active_hcas[@]}\"; do\n hca_port_pairs+=(\"\ [e2e-llm-inference-service] ${hca}:1\")\n done\n\n active_hca_list=$(IFS=,; echo \"${active_hcas[*]}\"\ [e2e-llm-inference-service] )\n hca_port_pairs_list=$(IFS=,; echo \"${hca_port_pairs[*]}\")\n \ [e2e-llm-inference-service] \ echo \"[Infer RoCE] Setting active HCAs: ${active_hca_list}\"\n \ [e2e-llm-inference-service] \ export NCCL_IB_HCA=${NCCL_IB_HCA:-${active_hca_list}}\n export\ [e2e-llm-inference-service] \ NVSHMEM_HCA_LIST=${NVSHMEM_HCA_LIST:-${hca_port_pairs_list}}\n export\ [e2e-llm-inference-service] \ UCX_NET_DEVICES=${UCX_NET_DEVICES:-${hca_port_pairs_list}}\n\n echo\ [e2e-llm-inference-service] \ \"[Infer RoCE] NCCL_IB_HCA=${NCCL_IB_HCA}\"\n echo \"[Infer RoCE]\ [e2e-llm-inference-service] \ NVSHMEM_HCA_LIST=${NVSHMEM_HCA_LIST}\"\n echo \"[Infer RoCE] UCX_NET_DEVICES=${UCX_NET_DEVICES}\"\ [e2e-llm-inference-service] \n else\n echo \"[Infer RoCE] WARNING: No active RoCE HCAs found.\ [e2e-llm-inference-service] \ NCCL_IB_HCA will not be set.\"\n fi\n\n if [ ${#active_hcas[@]} -gt\ [e2e-llm-inference-service] \ 0 ]; then\n echo \"[Infer RoCE] Finding GID_INDEX for each active\ [e2e-llm-inference-service] \ HCA (SR-IOV compatible)...\"\n\n # For SR-IOV environments, find\ [e2e-llm-inference-service] \ the most common IPv4 RoCE v2 GID index across all HCAs\n declare\ [e2e-llm-inference-service] \ -A gid_index_count\n declare -A hca_gid_index\n\n for hca_name\ [e2e-llm-inference-service] \ in \"${active_hcas[@]}\"; do\n echo \"[Infer RoCE] Processing\ [e2e-llm-inference-service] \ HCA: ${hca_name}\"\n\n # Find all RoCE v2 IPv4 GIDs for this\ [e2e-llm-inference-service] \ HCA and count by index\n for tpath in /sys/class/infiniband/${hca_name}/ports/1/gid_attrs/types/*;\ [e2e-llm-inference-service] \ do\n if grep -q \"${KSERVE_INFER_IB_GID_INDEX_GREP}\" \"\ [e2e-llm-inference-service] $tpath\" 2>/dev/null; then\n idx=$(basename \"$tpath\"\ [e2e-llm-inference-service] )\n gid_file=\"/sys/class/infiniband/${hca_name}/ports/1/gids/${idx}\"\ [e2e-llm-inference-service] \n # Check for IPv4 GID (contains ffff:)\n \ [e2e-llm-inference-service] \ if [ -f \"$gid_file\" ] && grep -q \"ffff:\" \"$gid_file\"; then\n\ [e2e-llm-inference-service] \ gid_value=$(cat \"$gid_file\" 2>/dev/null || echo\ [e2e-llm-inference-service] \ \"\")\n echo \"[Infer RoCE] Found IPv4 RoCE v2 GID\ [e2e-llm-inference-service] \ for ${hca_name}: index=${idx}, gid=${gid_value}\"\n \ [e2e-llm-inference-service] \ hca_gid_index[\"${hca_name}\"]=\"${idx}\"\n gid_index_count[\"\ [e2e-llm-inference-service] ${idx}\"]=$((${gid_index_count[\"${idx}\"]} + 1))\n \ [e2e-llm-inference-service] \ break # Use first found IPv4 GID per HCA\n fi\n \ [e2e-llm-inference-service] \ fi\n done\n done\n\n # Find the most common\ [e2e-llm-inference-service] \ GID index (most likely to be consistent across nodes)\n best_gid_index=\"\ [e2e-llm-inference-service] \"\n max_count=0\n for idx in \"${!gid_index_count[@]}\"; do\n\ [e2e-llm-inference-service] \ count=${gid_index_count[\"${idx}\"]}\n echo \"[Infer\ [e2e-llm-inference-service] \ RoCE] GID_INDEX ${idx} found on ${count} HCAs\"\n if [ $count\ [e2e-llm-inference-service] \ -gt $max_count ]; then\n max_count=$count\n \ [e2e-llm-inference-service] \ best_gid_index=\"$idx\"\n fi\n done\n\n # Use deterministic\ [e2e-llm-inference-service] \ fallback if tied - prefer index 3 (SR-IOV standard)\n if [ ${#gid_index_count[@]}\ [e2e-llm-inference-service] \ -gt 1 ]; then\n echo \"[Infer RoCE] Multiple GID indices found,\ [e2e-llm-inference-service] \ selecting most common: ${best_gid_index}\"\n # If there's a tie,\ [e2e-llm-inference-service] \ prefer index 3 as it's most common in SR-IOV setups\n if [ -n\ [e2e-llm-inference-service] \ \"${gid_index_count['3']}\" ] && [ \"${gid_index_count['3']}\" -eq \"\ [e2e-llm-inference-service] $max_count\" ]; then\n best_gid_index=\"3\"\n \ [e2e-llm-inference-service] \ echo \"[Infer RoCE] Using deterministic fallback: GID_INDEX=3 (SR-IOV\ [e2e-llm-inference-service] \ standard)\"\n fi\n fi\n\n # Check if GID_INDEX is already\ [e2e-llm-inference-service] \ set via environment variables\n if [ -n \"${NCCL_IB_GID_INDEX}\"\ [e2e-llm-inference-service] \ ]; then\n echo \"[Infer RoCE] Using pre-configured NCCL_IB_GID_INDEX=${NCCL_IB_GID_INDEX}\ [e2e-llm-inference-service] \ from environment\"\n export NVSHMEM_IB_GID_INDEX=${NVSHMEM_IB_GID_INDEX:-$NCCL_IB_GID_INDEX}\n\ [e2e-llm-inference-service] \ export UCX_IB_GID_INDEX=${UCX_IB_GID_INDEX:-$NCCL_IB_GID_INDEX}\n\ [e2e-llm-inference-service] \ echo \"[Infer RoCE] Using pre-configured GID_INDEX=${NCCL_IB_GID_INDEX}\ [e2e-llm-inference-service] \ for NCCL, NVSHMEM, and UCX\"\n elif [ -n \"$best_gid_index\" ]; then\n\ [e2e-llm-inference-service] \ echo \"[Infer RoCE] Selected GID_INDEX: ${best_gid_index} (found\ [e2e-llm-inference-service] \ on ${max_count} HCAs)\"\n\n export NCCL_IB_GID_INDEX=${NCCL_IB_GID_INDEX:-$best_gid_index}\n\ [e2e-llm-inference-service] \ export NVSHMEM_IB_GID_INDEX=${NVSHMEM_IB_GID_INDEX:-$best_gid_index}\n\ [e2e-llm-inference-service] \ export UCX_IB_GID_INDEX=${UCX_IB_GID_INDEX:-$best_gid_index}\n\ [e2e-llm-inference-service] \n echo \"[Infer RoCE] Exported GID_INDEX=${best_gid_index} for\ [e2e-llm-inference-service] \ NCCL, NVSHMEM, and UCX\"\n else\n echo \"[Infer RoCE] ERROR:\ [e2e-llm-inference-service] \ No valid IPv4 ${KSERVE_INFER_IB_GID_INDEX_GREP} GID_INDEX found on any\ [e2e-llm-inference-service] \ HCA.\"\n fi\n else\n echo \"[Infer RoCE] No active HCAs found,\ [e2e-llm-inference-service] \ skipping GID_INDEX inference.\"\n fi\nfi\n\n# --disable-access-log-for-endpoints\ [e2e-llm-inference-service] \ landed in vLLM 0.16.0 (vllm-project/vllm#30011).\n# Older versions still\ [e2e-llm-inference-service] \ need the blanket --disable-uvicorn-access-log.\nACCESS_LOG_ARGS=\"--disable-uvicorn-access-log\"\ [e2e-llm-inference-service] \nVLLM_VERSION=$(vllm --version 2>/dev/null | tail -1 | awk '{print $NF}')\n\ [e2e-llm-inference-service] echo \"[access-log-detect] vllm version='${VLLM_VERSION}'\"\nif [[ \"$VLLM_VERSION\"\ [e2e-llm-inference-service] \ =~ ^[0-9]+\\.[0-9]+ ]] && [ \"$(printf '%s\\n%s\\n' \"0.16.0\" \"${VLLM_VERSION}\"\ [e2e-llm-inference-service] \ | sort -V | head -1)\" = \"0.16.0\" ]; then\n ACCESS_LOG_ARGS=\"--disable-access-log-for-endpoints\ [e2e-llm-inference-service] \ /health,/metrics,/ping\"\nfi\necho \"[access-log-detect] selected ACCESS_LOG_ARGS='${ACCESS_LOG_ARGS}'\"\ [e2e-llm-inference-service] \n\n# --shutdown-timeout landed in vLLM 0.18.0 (vllm-project/vllm#36666).\n\ [e2e-llm-inference-service] SHUTDOWN_TIMEOUT_ARGS=\"\"\nif [[ \"$VLLM_VERSION\" =~ ^[0-9]+\\.[0-9]+\ [e2e-llm-inference-service] \ ]] && [ \"$(printf '%s\\n%s\\n' \"0.18.0\" \"${VLLM_VERSION}\" | sort\ [e2e-llm-inference-service] \ -V | head -1)\" = \"0.18.0\" ]; then\n SHUTDOWN_TIMEOUT_ARGS=\"--shutdown-timeout\ [e2e-llm-inference-service] \ 40\"\nfi\n\neval \"exec vllm serve /mnt/models \\\n --served-model-name\ [e2e-llm-inference-service] \ \"facebook/opt-125m\" \"publishers/kserve-ci-e2e-test/models/facebook/opt-125m\"\ [e2e-llm-inference-service] \ \\\n --port 8000 \\\n ${ACCESS_LOG_ARGS} \\\n ${SHUTDOWN_TIMEOUT_ARGS}\ [e2e-llm-inference-service] \ \\\n --enable-ssl-refresh \\\n --ssl-certfile /var/run/kserve/tls/tls.crt\ [e2e-llm-inference-service] \ \\\n --ssl-keyfile /var/run/kserve/tls/tls.key \\\n ${VLLM_ADDITIONAL_ARGS}\ [e2e-llm-inference-service] \ \\\n $@\"" [e2e-llm-inference-service] - -- [e2e-llm-inference-service] args: [e2e-llm-inference-service] - --enable-lora [e2e-llm-inference-service] - --lora-modules [e2e-llm-inference-service] - '''{"name":"lora-adapter-1","path":"/mnt/lora/lora-adapter-1"}''' [e2e-llm-inference-service] - '''{"name":"publishers/kserve-ci-e2e-test/models/lora-adapter-1","path":"/mnt/lora/lora-adapter-1"}''' [e2e-llm-inference-service] ports: [e2e-llm-inference-service] - containerPort: 8000 [e2e-llm-inference-service] protocol: TCP [e2e-llm-inference-service] env: [e2e-llm-inference-service] - name: HOME [e2e-llm-inference-service] value: /home [e2e-llm-inference-service] - name: VLLM_LOGGING_LEVEL [e2e-llm-inference-service] value: DEBUG [e2e-llm-inference-service] - name: VLLM_CPU_KVCACHE_SPACE [e2e-llm-inference-service] value: '1' [e2e-llm-inference-service] - name: VLLM_ENABLE_V1_MULTIPROCESSING [e2e-llm-inference-service] value: '0' [e2e-llm-inference-service] - name: USER [e2e-llm-inference-service] value: nonroot [e2e-llm-inference-service] - name: TORCHINDUCTOR_CACHE_DIR [e2e-llm-inference-service] value: /tmp/torchinductor-cache [e2e-llm-inference-service] - name: HF_HUB_CACHE [e2e-llm-inference-service] value: /models [e2e-llm-inference-service] resources: [e2e-llm-inference-service] limits: [e2e-llm-inference-service] cpu: '2' [e2e-llm-inference-service] memory: 7Gi [e2e-llm-inference-service] requests: [e2e-llm-inference-service] cpu: 200m [e2e-llm-inference-service] memory: 2Gi [e2e-llm-inference-service] volumeMounts: [e2e-llm-inference-service] - name: home [e2e-llm-inference-service] mountPath: /home [e2e-llm-inference-service] - name: tmp-dir [e2e-llm-inference-service] mountPath: /tmp [e2e-llm-inference-service] - name: dshm [e2e-llm-inference-service] mountPath: /dev/shm [e2e-llm-inference-service] - name: model-cache [e2e-llm-inference-service] mountPath: /models [e2e-llm-inference-service] - name: tls-certs [e2e-llm-inference-service] readOnly: true [e2e-llm-inference-service] mountPath: /var/run/kserve/tls [e2e-llm-inference-service] - name: kserve-provision-location [e2e-llm-inference-service] readOnly: true [e2e-llm-inference-service] mountPath: /mnt [e2e-llm-inference-service] livenessProbe: [e2e-llm-inference-service] httpGet: [e2e-llm-inference-service] path: /health [e2e-llm-inference-service] port: 8000 [e2e-llm-inference-service] scheme: HTTPS [e2e-llm-inference-service] timeoutSeconds: 1 [e2e-llm-inference-service] periodSeconds: 10 [e2e-llm-inference-service] successThreshold: 1 [e2e-llm-inference-service] failureThreshold: 10 [e2e-llm-inference-service] readinessProbe: [e2e-llm-inference-service] httpGet: [e2e-llm-inference-service] path: /health [e2e-llm-inference-service] port: 8000 [e2e-llm-inference-service] scheme: HTTPS [e2e-llm-inference-service] timeoutSeconds: 1 [e2e-llm-inference-service] periodSeconds: 1 [e2e-llm-inference-service] successThreshold: 1 [e2e-llm-inference-service] failureThreshold: 2 [e2e-llm-inference-service] startupProbe: [e2e-llm-inference-service] httpGet: [e2e-llm-inference-service] path: /health [e2e-llm-inference-service] port: 8000 [e2e-llm-inference-service] scheme: HTTPS [e2e-llm-inference-service] timeoutSeconds: 1 [e2e-llm-inference-service] periodSeconds: 10 [e2e-llm-inference-service] successThreshold: 1 [e2e-llm-inference-service] failureThreshold: 60 [e2e-llm-inference-service] lifecycle: [e2e-llm-inference-service] preStop: [e2e-llm-inference-service] exec: [e2e-llm-inference-service] command: [e2e-llm-inference-service] - /bin/sleep [e2e-llm-inference-service] - '15' [e2e-llm-inference-service] terminationMessagePath: /dev/termination-log [e2e-llm-inference-service] terminationMessagePolicy: FallbackToLogsOnError [e2e-llm-inference-service] imagePullPolicy: IfNotPresent [e2e-llm-inference-service] securityContext: [e2e-llm-inference-service] capabilities: [e2e-llm-inference-service] drop: [e2e-llm-inference-service] - ALL [e2e-llm-inference-service] runAsNonRoot: true [e2e-llm-inference-service] readOnlyRootFilesystem: true [e2e-llm-inference-service] allowPrivilegeEscalation: false [e2e-llm-inference-service] seccompProfile: [e2e-llm-inference-service] type: RuntimeDefault [e2e-llm-inference-service] restartPolicy: Always [e2e-llm-inference-service] terminationGracePeriodSeconds: 60 [e2e-llm-inference-service] dnsPolicy: ClusterFirst [e2e-llm-inference-service] securityContext: {} [e2e-llm-inference-service] schedulerName: default-scheduler [e2e-llm-inference-service] strategy: [e2e-llm-inference-service] type: RollingUpdate [e2e-llm-inference-service] rollingUpdate: [e2e-llm-inference-service] maxUnavailable: 25% [e2e-llm-inference-service] maxSurge: 25% [e2e-llm-inference-service] revisionHistoryLimit: 10 [e2e-llm-inference-service] progressDeadlineSeconds: 600 [e2e-llm-inference-service] status: [e2e-llm-inference-service] observedGeneration: 1 [e2e-llm-inference-service] replicas: 1 [e2e-llm-inference-service] updatedReplicas: 1 [e2e-llm-inference-service] readyReplicas: 1 [e2e-llm-inference-service] availableReplicas: 1 [e2e-llm-inference-service] conditions: [e2e-llm-inference-service] - type: Available [e2e-llm-inference-service] status: 'True' [e2e-llm-inference-service] lastUpdateTime: '2026-07-07T14:47:16Z' [e2e-llm-inference-service] lastTransitionTime: '2026-07-07T14:47:16Z' [e2e-llm-inference-service] reason: MinimumReplicasAvailable [e2e-llm-inference-service] message: Deployment has minimum availability. [e2e-llm-inference-service] - type: Progressing [e2e-llm-inference-service] status: 'True' [e2e-llm-inference-service] lastUpdateTime: '2026-07-07T14:47:16Z' [e2e-llm-inference-service] lastTransitionTime: '2026-07-07T14:45:56Z' [e2e-llm-inference-service] reason: NewReplicaSetAvailable [e2e-llm-inference-service] message: ReplicaSet "llmisvc-model-fb-opt-125m-with-ba4d693a-kserve-5b998d5789" [e2e-llm-inference-service] has successfully progressed. [e2e-llm-inference-service] apiVersion: apps/v1 [e2e-llm-inference-service] kind: Deployment [e2e-llm-inference-service] [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1254 --- [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1255 metadata: [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-ba4d693a-kserve-router-scheduler [e2e-llm-inference-service] namespace: kserve-ci-e2e-test [e2e-llm-inference-service] uid: 15262fe4-b55c-4249-ae3a-b00445a310f9 [e2e-llm-inference-service] resourceVersion: '46622' [e2e-llm-inference-service] generation: 1 [e2e-llm-inference-service] creationTimestamp: '2026-07-07T14:45:56Z' [e2e-llm-inference-service] labels: [e2e-llm-inference-service] app.kubernetes.io/component: llminferenceservice-router-scheduler [e2e-llm-inference-service] app.kubernetes.io/name: llmisvc-model-fb-opt-125m-with-ba4d693a [e2e-llm-inference-service] app.kubernetes.io/part-of: llminferenceservice [e2e-llm-inference-service] annotations: [e2e-llm-inference-service] deployment.kubernetes.io/revision: '1' [e2e-llm-inference-service] ownerReferences: [e2e-llm-inference-service] - apiVersion: serving.kserve.io/v1alpha2 [e2e-llm-inference-service] kind: LLMInferenceService [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-ba4d693a [e2e-llm-inference-service] uid: 997d3cef-fabf-4b25-ad95-5ebe7a367fc5 [e2e-llm-inference-service] controller: true [e2e-llm-inference-service] blockOwnerDeletion: true [e2e-llm-inference-service] managedFields: [e2e-llm-inference-service] - manager: manager [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] apiVersion: apps/v1 [e2e-llm-inference-service] time: '2026-07-07T14:45:56Z' [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:metadata: [e2e-llm-inference-service] f:labels: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:app.kubernetes.io/component: {} [e2e-llm-inference-service] f:app.kubernetes.io/name: {} [e2e-llm-inference-service] f:app.kubernetes.io/part-of: {} [e2e-llm-inference-service] f:ownerReferences: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"uid":"997d3cef-fabf-4b25-ad95-5ebe7a367fc5"}: {} [e2e-llm-inference-service] f:spec: [e2e-llm-inference-service] f:progressDeadlineSeconds: {} [e2e-llm-inference-service] f:replicas: {} [e2e-llm-inference-service] f:revisionHistoryLimit: {} [e2e-llm-inference-service] f:selector: {} [e2e-llm-inference-service] f:strategy: [e2e-llm-inference-service] f:type: {} [e2e-llm-inference-service] f:template: [e2e-llm-inference-service] f:metadata: [e2e-llm-inference-service] f:annotations: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:app.kubernetes.io/version: {} [e2e-llm-inference-service] f:certificates.kserve.io/expiration-v2: {} [e2e-llm-inference-service] f:labels: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:app.kubernetes.io/component: {} [e2e-llm-inference-service] f:app.kubernetes.io/name: {} [e2e-llm-inference-service] f:app.kubernetes.io/part-of: {} [e2e-llm-inference-service] f:spec: [e2e-llm-inference-service] f:containers: [e2e-llm-inference-service] k:{"name":"main"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:args: {} [e2e-llm-inference-service] f:command: {} [e2e-llm-inference-service] f:env: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"name":"SSL_CERT_DIR"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] f:image: {} [e2e-llm-inference-service] f:imagePullPolicy: {} [e2e-llm-inference-service] f:lifecycle: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:preStop: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:exec: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:command: {} [e2e-llm-inference-service] f:livenessProbe: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:failureThreshold: {} [e2e-llm-inference-service] f:grpc: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:port: {} [e2e-llm-inference-service] f:service: {} [e2e-llm-inference-service] f:initialDelaySeconds: {} [e2e-llm-inference-service] f:periodSeconds: {} [e2e-llm-inference-service] f:successThreshold: {} [e2e-llm-inference-service] f:timeoutSeconds: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:ports: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"containerPort":5557,"protocol":"TCP"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:containerPort: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:protocol: {} [e2e-llm-inference-service] k:{"containerPort":9002,"protocol":"TCP"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:containerPort: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:protocol: {} [e2e-llm-inference-service] k:{"containerPort":9003,"protocol":"TCP"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:containerPort: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:protocol: {} [e2e-llm-inference-service] k:{"containerPort":9090,"protocol":"TCP"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:containerPort: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:protocol: {} [e2e-llm-inference-service] f:readinessProbe: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:failureThreshold: {} [e2e-llm-inference-service] f:grpc: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:port: {} [e2e-llm-inference-service] f:service: {} [e2e-llm-inference-service] f:initialDelaySeconds: {} [e2e-llm-inference-service] f:periodSeconds: {} [e2e-llm-inference-service] f:successThreshold: {} [e2e-llm-inference-service] f:timeoutSeconds: {} [e2e-llm-inference-service] f:resources: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:requests: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:cpu: {} [e2e-llm-inference-service] f:memory: {} [e2e-llm-inference-service] f:securityContext: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:allowPrivilegeEscalation: {} [e2e-llm-inference-service] f:capabilities: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:drop: {} [e2e-llm-inference-service] f:readOnlyRootFilesystem: {} [e2e-llm-inference-service] f:runAsNonRoot: {} [e2e-llm-inference-service] f:seccompProfile: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:type: {} [e2e-llm-inference-service] f:terminationMessagePath: {} [e2e-llm-inference-service] f:terminationMessagePolicy: {} [e2e-llm-inference-service] f:volumeMounts: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"mountPath":"/tmp/tokenizer"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:mountPath: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] k:{"mountPath":"/var/run/kserve/tls"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:mountPath: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:readOnly: {} [e2e-llm-inference-service] f:dnsPolicy: {} [e2e-llm-inference-service] f:initContainers: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"name":"storage-initializer"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:args: {} [e2e-llm-inference-service] f:env: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"name":"AWS_ACCESS_KEY_ID"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:valueFrom: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:secretKeyRef: {} [e2e-llm-inference-service] k:{"name":"AWS_CA_BUNDLE"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"AWS_CA_BUNDLE_CONFIGMAP"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"AWS_ENDPOINT_URL"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"AWS_SECRET_ACCESS_KEY"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:valueFrom: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:secretKeyRef: {} [e2e-llm-inference-service] k:{"name":"HF_HUB_ENABLE_HF_TRANSFER"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"HF_XET_HIGH_PERFORMANCE"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"HF_XET_NUM_CONCURRENT_RANGE_GETS"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"S3_ENDPOINT"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"S3_USE_HTTPS"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"S3_VERIFY_SSL"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"STORAGE_ALLOW_PATTERNS"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] f:image: {} [e2e-llm-inference-service] f:imagePullPolicy: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:resources: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:limits: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:cpu: {} [e2e-llm-inference-service] f:memory: {} [e2e-llm-inference-service] f:requests: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:cpu: {} [e2e-llm-inference-service] f:memory: {} [e2e-llm-inference-service] f:terminationMessagePath: {} [e2e-llm-inference-service] f:terminationMessagePolicy: {} [e2e-llm-inference-service] f:volumeMounts: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"mountPath":"/mnt/models"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:mountPath: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:restartPolicy: {} [e2e-llm-inference-service] f:schedulerName: {} [e2e-llm-inference-service] f:securityContext: {} [e2e-llm-inference-service] f:serviceAccount: {} [e2e-llm-inference-service] f:serviceAccountName: {} [e2e-llm-inference-service] f:terminationGracePeriodSeconds: {} [e2e-llm-inference-service] f:volumes: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"name":"kserve-provision-location"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:emptyDir: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] k:{"name":"tls-certs"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:secret: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:defaultMode: {} [e2e-llm-inference-service] f:secretName: {} [e2e-llm-inference-service] k:{"name":"tokenizer-cache"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:emptyDir: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] k:{"name":"tokenizer-tmp"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:emptyDir: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] k:{"name":"tokenizer-uds"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:emptyDir: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] - manager: kube-controller-manager [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] apiVersion: apps/v1 [e2e-llm-inference-service] time: '2026-07-07T14:46:29Z' [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:metadata: [e2e-llm-inference-service] f:annotations: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:deployment.kubernetes.io/revision: {} [e2e-llm-inference-service] f:status: [e2e-llm-inference-service] f:availableReplicas: {} [e2e-llm-inference-service] f:conditions: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"type":"Available"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:lastTransitionTime: {} [e2e-llm-inference-service] f:lastUpdateTime: {} [e2e-llm-inference-service] f:message: {} [e2e-llm-inference-service] f:reason: {} [e2e-llm-inference-service] f:status: {} [e2e-llm-inference-service] f:type: {} [e2e-llm-inference-service] k:{"type":"Progressing"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:lastTransitionTime: {} [e2e-llm-inference-service] f:lastUpdateTime: {} [e2e-llm-inference-service] f:message: {} [e2e-llm-inference-service] f:reason: {} [e2e-llm-inference-service] f:status: {} [e2e-llm-inference-service] f:type: {} [e2e-llm-inference-service] f:observedGeneration: {} [e2e-llm-inference-service] f:readyReplicas: {} [e2e-llm-inference-service] f:replicas: {} [e2e-llm-inference-service] f:updatedReplicas: {} [e2e-llm-inference-service] subresource: status [e2e-llm-inference-service] spec: [e2e-llm-inference-service] replicas: 1 [e2e-llm-inference-service] selector: [e2e-llm-inference-service] matchLabels: [e2e-llm-inference-service] app.kubernetes.io/component: llminferenceservice-router-scheduler [e2e-llm-inference-service] app.kubernetes.io/name: llmisvc-model-fb-opt-125m-with-ba4d693a [e2e-llm-inference-service] app.kubernetes.io/part-of: llminferenceservice [e2e-llm-inference-service] template: [e2e-llm-inference-service] metadata: [e2e-llm-inference-service] labels: [e2e-llm-inference-service] app.kubernetes.io/component: llminferenceservice-router-scheduler [e2e-llm-inference-service] app.kubernetes.io/name: llmisvc-model-fb-opt-125m-with-ba4d693a [e2e-llm-inference-service] app.kubernetes.io/part-of: llminferenceservice [e2e-llm-inference-service] annotations: [e2e-llm-inference-service] app.kubernetes.io/version: 0.9.0 [e2e-llm-inference-service] certificates.kserve.io/expiration-v2: 'true' [e2e-llm-inference-service] spec: [e2e-llm-inference-service] volumes: [e2e-llm-inference-service] - name: tls-certs [e2e-llm-inference-service] secret: [e2e-llm-inference-service] secretName: llmisv77ff2528d3e9b4972cd9335229fce9f0-kserve-self-signed-certs [e2e-llm-inference-service] defaultMode: 420 [e2e-llm-inference-service] - name: tokenizer-uds [e2e-llm-inference-service] emptyDir: {} [e2e-llm-inference-service] - name: tokenizer-tmp [e2e-llm-inference-service] emptyDir: {} [e2e-llm-inference-service] - name: tokenizer-cache [e2e-llm-inference-service] emptyDir: {} [e2e-llm-inference-service] - name: kserve-provision-location [e2e-llm-inference-service] emptyDir: {} [e2e-llm-inference-service] initContainers: [e2e-llm-inference-service] - name: storage-initializer [e2e-llm-inference-service] image: quay.io/opendatahub/kserve-storage-initializer@sha256:13ad765c0f323200e1db7b45c99661287e65c5158d6b54b16139829636d20be4 [e2e-llm-inference-service] args: [e2e-llm-inference-service] - hf://facebook/opt-125m [e2e-llm-inference-service] - /mnt/models [e2e-llm-inference-service] env: [e2e-llm-inference-service] - name: AWS_ACCESS_KEY_ID [e2e-llm-inference-service] valueFrom: [e2e-llm-inference-service] secretKeyRef: [e2e-llm-inference-service] name: seaweedfs-s3-creds [e2e-llm-inference-service] key: AWS_ACCESS_KEY_ID [e2e-llm-inference-service] - name: AWS_SECRET_ACCESS_KEY [e2e-llm-inference-service] valueFrom: [e2e-llm-inference-service] secretKeyRef: [e2e-llm-inference-service] name: seaweedfs-s3-creds [e2e-llm-inference-service] key: AWS_SECRET_ACCESS_KEY [e2e-llm-inference-service] - name: S3_USE_HTTPS [e2e-llm-inference-service] value: '0' [e2e-llm-inference-service] - name: S3_ENDPOINT [e2e-llm-inference-service] value: s3-service.kserve:8333 [e2e-llm-inference-service] - name: AWS_ENDPOINT_URL [e2e-llm-inference-service] value: http://s3-service.kserve:8333 [e2e-llm-inference-service] - name: S3_VERIFY_SSL [e2e-llm-inference-service] value: '0' [e2e-llm-inference-service] - name: AWS_CA_BUNDLE [e2e-llm-inference-service] value: /etc/ssl/custom-certs/cabundle.crt [e2e-llm-inference-service] - name: AWS_CA_BUNDLE_CONFIGMAP [e2e-llm-inference-service] value: odh-kserve-custom-ca-bundle [e2e-llm-inference-service] - name: HF_HUB_ENABLE_HF_TRANSFER [e2e-llm-inference-service] value: '1' [e2e-llm-inference-service] - name: HF_XET_HIGH_PERFORMANCE [e2e-llm-inference-service] value: '1' [e2e-llm-inference-service] - name: HF_XET_NUM_CONCURRENT_RANGE_GETS [e2e-llm-inference-service] value: '8' [e2e-llm-inference-service] - name: STORAGE_ALLOW_PATTERNS [e2e-llm-inference-service] value: '["tokenizer.json", "tokenizer_config.json", "special_tokens_map.json", [e2e-llm-inference-service] "vocab.json", "merges.txt", "config.json", "generation_config.json"]' [e2e-llm-inference-service] resources: [e2e-llm-inference-service] limits: [e2e-llm-inference-service] cpu: '1' [e2e-llm-inference-service] memory: 24Gi [e2e-llm-inference-service] requests: [e2e-llm-inference-service] cpu: 100m [e2e-llm-inference-service] memory: 100Mi [e2e-llm-inference-service] volumeMounts: [e2e-llm-inference-service] - name: kserve-provision-location [e2e-llm-inference-service] mountPath: /mnt/models [e2e-llm-inference-service] terminationMessagePath: /dev/termination-log [e2e-llm-inference-service] terminationMessagePolicy: FallbackToLogsOnError [e2e-llm-inference-service] imagePullPolicy: IfNotPresent [e2e-llm-inference-service] containers: [e2e-llm-inference-service] - name: main [e2e-llm-inference-service] image: ghcr.io/llm-d/llm-d-router-endpoint-picker:v0.9.0-rc.2 [e2e-llm-inference-service] command: [e2e-llm-inference-service] - /app/epp [e2e-llm-inference-service] - --pool-name [e2e-llm-inference-service] - llmisvc-model-fb-opt-125m-with-ba4d693a-inference-pool [e2e-llm-inference-service] - --pool-namespace [e2e-llm-inference-service] - kserve-ci-e2e-test [e2e-llm-inference-service] - --zap-encoder [e2e-llm-inference-service] - json [e2e-llm-inference-service] - --grpc-port [e2e-llm-inference-service] - '9002' [e2e-llm-inference-service] - --grpc-health-port [e2e-llm-inference-service] - '9003' [e2e-llm-inference-service] - --enable-cert-reload=true [e2e-llm-inference-service] - --secure-serving=true [e2e-llm-inference-service] - --model-server-metrics-scheme=https [e2e-llm-inference-service] - --cert-path=/var/run/kserve/tls [e2e-llm-inference-service] args: [e2e-llm-inference-service] - --config-text [e2e-llm-inference-service] - "apiVersion: inference.networking.x-k8s.io/v1alpha1\nkind: EndpointPickerConfig\n\ [e2e-llm-inference-service] plugins:\n- type: single-profile-handler\n- type: queue-scorer\n- type:\ [e2e-llm-inference-service] \ prefix-cache-scorer\n- type: max-score-picker\n- type: lora-affinity-scorer\n\ [e2e-llm-inference-service] schedulingProfiles:\n- name: default\n plugins:\n - pluginRef: lora-affinity-scorer\n\ [e2e-llm-inference-service] \ weight: 4\n - pluginRef: queue-scorer\n weight: 2\n - pluginRef:\ [e2e-llm-inference-service] \ prefix-cache-scorer\n weight: 3\n - pluginRef: max-score-picker\n" [e2e-llm-inference-service] ports: [e2e-llm-inference-service] - name: grpc [e2e-llm-inference-service] containerPort: 9002 [e2e-llm-inference-service] protocol: TCP [e2e-llm-inference-service] - name: grpc-health [e2e-llm-inference-service] containerPort: 9003 [e2e-llm-inference-service] protocol: TCP [e2e-llm-inference-service] - name: metrics [e2e-llm-inference-service] containerPort: 9090 [e2e-llm-inference-service] protocol: TCP [e2e-llm-inference-service] - name: zmq [e2e-llm-inference-service] containerPort: 5557 [e2e-llm-inference-service] protocol: TCP [e2e-llm-inference-service] env: [e2e-llm-inference-service] - name: SSL_CERT_DIR [e2e-llm-inference-service] value: /var/run/kserve/tls:/var/run/secrets/kubernetes.io/serviceaccount:/etc/pki/tls/certs [e2e-llm-inference-service] resources: [e2e-llm-inference-service] requests: [e2e-llm-inference-service] cpu: 256m [e2e-llm-inference-service] memory: 500Mi [e2e-llm-inference-service] volumeMounts: [e2e-llm-inference-service] - name: tls-certs [e2e-llm-inference-service] readOnly: true [e2e-llm-inference-service] mountPath: /var/run/kserve/tls [e2e-llm-inference-service] - name: tokenizer-uds [e2e-llm-inference-service] mountPath: /tmp/tokenizer [e2e-llm-inference-service] livenessProbe: [e2e-llm-inference-service] grpc: [e2e-llm-inference-service] port: 9003 [e2e-llm-inference-service] service: liveness [e2e-llm-inference-service] initialDelaySeconds: 5 [e2e-llm-inference-service] timeoutSeconds: 1 [e2e-llm-inference-service] periodSeconds: 10 [e2e-llm-inference-service] successThreshold: 1 [e2e-llm-inference-service] failureThreshold: 3 [e2e-llm-inference-service] readinessProbe: [e2e-llm-inference-service] grpc: [e2e-llm-inference-service] port: 9003 [e2e-llm-inference-service] service: readiness [e2e-llm-inference-service] initialDelaySeconds: 30 [e2e-llm-inference-service] timeoutSeconds: 1 [e2e-llm-inference-service] periodSeconds: 10 [e2e-llm-inference-service] successThreshold: 1 [e2e-llm-inference-service] failureThreshold: 3 [e2e-llm-inference-service] lifecycle: [e2e-llm-inference-service] preStop: [e2e-llm-inference-service] exec: [e2e-llm-inference-service] command: [e2e-llm-inference-service] - /bin/sleep [e2e-llm-inference-service] - '15' [e2e-llm-inference-service] terminationMessagePath: /dev/termination-log [e2e-llm-inference-service] terminationMessagePolicy: FallbackToLogsOnError [e2e-llm-inference-service] imagePullPolicy: IfNotPresent [e2e-llm-inference-service] securityContext: [e2e-llm-inference-service] capabilities: [e2e-llm-inference-service] drop: [e2e-llm-inference-service] - ALL [e2e-llm-inference-service] runAsNonRoot: true [e2e-llm-inference-service] readOnlyRootFilesystem: true [e2e-llm-inference-service] allowPrivilegeEscalation: false [e2e-llm-inference-service] seccompProfile: [e2e-llm-inference-service] type: RuntimeDefault [e2e-llm-inference-service] restartPolicy: Always [e2e-llm-inference-service] terminationGracePeriodSeconds: 60 [e2e-llm-inference-service] dnsPolicy: ClusterFirst [e2e-llm-inference-service] serviceAccountName: llmisvc-model-fb-opt-125m-with-ba4d693a-epp-sa [e2e-llm-inference-service] serviceAccount: llmisvc-model-fb-opt-125m-with-ba4d693a-epp-sa [e2e-llm-inference-service] securityContext: {} [e2e-llm-inference-service] schedulerName: default-scheduler [e2e-llm-inference-service] strategy: [e2e-llm-inference-service] type: Recreate [e2e-llm-inference-service] revisionHistoryLimit: 10 [e2e-llm-inference-service] progressDeadlineSeconds: 600 [e2e-llm-inference-service] status: [e2e-llm-inference-service] observedGeneration: 1 [e2e-llm-inference-service] replicas: 1 [e2e-llm-inference-service] updatedReplicas: 1 [e2e-llm-inference-service] readyReplicas: 1 [e2e-llm-inference-service] availableReplicas: 1 [e2e-llm-inference-service] conditions: [e2e-llm-inference-service] - type: Available [e2e-llm-inference-service] status: 'True' [e2e-llm-inference-service] lastUpdateTime: '2026-07-07T14:46:29Z' [e2e-llm-inference-service] lastTransitionTime: '2026-07-07T14:46:29Z' [e2e-llm-inference-service] reason: MinimumReplicasAvailable [e2e-llm-inference-service] message: Deployment has minimum availability. [e2e-llm-inference-service] - type: Progressing [e2e-llm-inference-service] status: 'True' [e2e-llm-inference-service] lastUpdateTime: '2026-07-07T14:46:29Z' [e2e-llm-inference-service] lastTransitionTime: '2026-07-07T14:45:56Z' [e2e-llm-inference-service] reason: NewReplicaSetAvailable [e2e-llm-inference-service] message: ReplicaSet "llmisvc-model-fb-opt-125m-with-ba4d693a-kserve-router-scheduler-79f5455bb" [e2e-llm-inference-service] has successfully progressed. [e2e-llm-inference-service] apiVersion: apps/v1 [e2e-llm-inference-service] kind: Deployment [e2e-llm-inference-service] [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1254 --- [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1255 metadata: [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-ba4d693a-kserve-5b998d5789 [e2e-llm-inference-service] namespace: kserve-ci-e2e-test [e2e-llm-inference-service] uid: e69bc84c-145f-4492-a785-714fe7558007 [e2e-llm-inference-service] resourceVersion: '47365' [e2e-llm-inference-service] generation: 1 [e2e-llm-inference-service] creationTimestamp: '2026-07-07T14:45:56Z' [e2e-llm-inference-service] labels: [e2e-llm-inference-service] app.kubernetes.io/component: llminferenceservice-workload [e2e-llm-inference-service] app.kubernetes.io/name: llmisvc-model-fb-opt-125m-with-ba4d693a [e2e-llm-inference-service] app.kubernetes.io/part-of: llminferenceservice [e2e-llm-inference-service] kserve.io/component: workload [e2e-llm-inference-service] llm-d.ai/role: both [e2e-llm-inference-service] pod-template-hash: 5b998d5789 [e2e-llm-inference-service] annotations: [e2e-llm-inference-service] deployment.kubernetes.io/desired-replicas: '1' [e2e-llm-inference-service] deployment.kubernetes.io/max-replicas: '2' [e2e-llm-inference-service] deployment.kubernetes.io/revision: '1' [e2e-llm-inference-service] ownerReferences: [e2e-llm-inference-service] - apiVersion: apps/v1 [e2e-llm-inference-service] kind: Deployment [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-ba4d693a-kserve [e2e-llm-inference-service] uid: 834ec5b2-f5f3-4037-88ef-04cb98318a8c [e2e-llm-inference-service] controller: true [e2e-llm-inference-service] blockOwnerDeletion: true [e2e-llm-inference-service] managedFields: [e2e-llm-inference-service] - manager: kube-controller-manager [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] apiVersion: apps/v1 [e2e-llm-inference-service] time: '2026-07-07T14:45:56Z' [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:metadata: [e2e-llm-inference-service] f:annotations: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:deployment.kubernetes.io/desired-replicas: {} [e2e-llm-inference-service] f:deployment.kubernetes.io/max-replicas: {} [e2e-llm-inference-service] f:deployment.kubernetes.io/revision: {} [e2e-llm-inference-service] f:labels: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:app.kubernetes.io/component: {} [e2e-llm-inference-service] f:app.kubernetes.io/name: {} [e2e-llm-inference-service] f:app.kubernetes.io/part-of: {} [e2e-llm-inference-service] f:kserve.io/component: {} [e2e-llm-inference-service] f:llm-d.ai/role: {} [e2e-llm-inference-service] f:pod-template-hash: {} [e2e-llm-inference-service] f:ownerReferences: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"uid":"834ec5b2-f5f3-4037-88ef-04cb98318a8c"}: {} [e2e-llm-inference-service] f:spec: [e2e-llm-inference-service] f:replicas: {} [e2e-llm-inference-service] f:selector: {} [e2e-llm-inference-service] f:template: [e2e-llm-inference-service] f:metadata: [e2e-llm-inference-service] f:labels: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:app.kubernetes.io/component: {} [e2e-llm-inference-service] f:app.kubernetes.io/name: {} [e2e-llm-inference-service] f:app.kubernetes.io/part-of: {} [e2e-llm-inference-service] f:kserve.io/component: {} [e2e-llm-inference-service] f:llm-d.ai/role: {} [e2e-llm-inference-service] f:pod-template-hash: {} [e2e-llm-inference-service] f:spec: [e2e-llm-inference-service] f:containers: [e2e-llm-inference-service] k:{"name":"main"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:args: {} [e2e-llm-inference-service] f:command: {} [e2e-llm-inference-service] f:env: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"name":"HF_HUB_CACHE"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"HOME"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"TORCHINDUCTOR_CACHE_DIR"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"USER"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"VLLM_CPU_KVCACHE_SPACE"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"VLLM_ENABLE_V1_MULTIPROCESSING"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"VLLM_LOGGING_LEVEL"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] f:image: {} [e2e-llm-inference-service] f:imagePullPolicy: {} [e2e-llm-inference-service] f:lifecycle: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:preStop: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:exec: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:command: {} [e2e-llm-inference-service] f:livenessProbe: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:failureThreshold: {} [e2e-llm-inference-service] f:httpGet: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:path: {} [e2e-llm-inference-service] f:port: {} [e2e-llm-inference-service] f:scheme: {} [e2e-llm-inference-service] f:periodSeconds: {} [e2e-llm-inference-service] f:successThreshold: {} [e2e-llm-inference-service] f:timeoutSeconds: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:ports: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"containerPort":8000,"protocol":"TCP"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:containerPort: {} [e2e-llm-inference-service] f:protocol: {} [e2e-llm-inference-service] f:readinessProbe: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:failureThreshold: {} [e2e-llm-inference-service] f:httpGet: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:path: {} [e2e-llm-inference-service] f:port: {} [e2e-llm-inference-service] f:scheme: {} [e2e-llm-inference-service] f:periodSeconds: {} [e2e-llm-inference-service] f:successThreshold: {} [e2e-llm-inference-service] f:timeoutSeconds: {} [e2e-llm-inference-service] f:resources: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:limits: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:cpu: {} [e2e-llm-inference-service] f:memory: {} [e2e-llm-inference-service] f:requests: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:cpu: {} [e2e-llm-inference-service] f:memory: {} [e2e-llm-inference-service] f:securityContext: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:allowPrivilegeEscalation: {} [e2e-llm-inference-service] f:capabilities: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:drop: {} [e2e-llm-inference-service] f:readOnlyRootFilesystem: {} [e2e-llm-inference-service] f:runAsNonRoot: {} [e2e-llm-inference-service] f:seccompProfile: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:type: {} [e2e-llm-inference-service] f:startupProbe: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:failureThreshold: {} [e2e-llm-inference-service] f:httpGet: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:path: {} [e2e-llm-inference-service] f:port: {} [e2e-llm-inference-service] f:scheme: {} [e2e-llm-inference-service] f:periodSeconds: {} [e2e-llm-inference-service] f:successThreshold: {} [e2e-llm-inference-service] f:timeoutSeconds: {} [e2e-llm-inference-service] f:terminationMessagePath: {} [e2e-llm-inference-service] f:terminationMessagePolicy: {} [e2e-llm-inference-service] f:volumeMounts: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"mountPath":"/dev/shm"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:mountPath: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] k:{"mountPath":"/home"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:mountPath: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] k:{"mountPath":"/mnt"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:mountPath: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:readOnly: {} [e2e-llm-inference-service] k:{"mountPath":"/models"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:mountPath: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] k:{"mountPath":"/tmp"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:mountPath: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] k:{"mountPath":"/var/run/kserve/tls"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:mountPath: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:readOnly: {} [e2e-llm-inference-service] f:dnsPolicy: {} [e2e-llm-inference-service] f:initContainers: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"name":"storage-initializer"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:args: {} [e2e-llm-inference-service] f:env: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"name":"AWS_ACCESS_KEY_ID"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:valueFrom: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:secretKeyRef: {} [e2e-llm-inference-service] k:{"name":"AWS_CA_BUNDLE"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"AWS_CA_BUNDLE_CONFIGMAP"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"AWS_ENDPOINT_URL"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"AWS_SECRET_ACCESS_KEY"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:valueFrom: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:secretKeyRef: {} [e2e-llm-inference-service] k:{"name":"CA_BUNDLE_CONFIGMAP_NAME"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"CA_BUNDLE_VOLUME_MOUNT_POINT"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"HF_HUB_ENABLE_HF_TRANSFER"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"HF_XET_HIGH_PERFORMANCE"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"HF_XET_NUM_CONCURRENT_RANGE_GETS"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"S3_ENDPOINT"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"S3_USE_HTTPS"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"S3_VERIFY_SSL"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] f:image: {} [e2e-llm-inference-service] f:imagePullPolicy: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:resources: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:limits: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:cpu: {} [e2e-llm-inference-service] f:memory: {} [e2e-llm-inference-service] f:requests: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:cpu: {} [e2e-llm-inference-service] f:memory: {} [e2e-llm-inference-service] f:terminationMessagePath: {} [e2e-llm-inference-service] f:terminationMessagePolicy: {} [e2e-llm-inference-service] f:volumeMounts: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"mountPath":"/etc/ssl/custom-certs"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:mountPath: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:readOnly: {} [e2e-llm-inference-service] k:{"mountPath":"/mnt"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:mountPath: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:restartPolicy: {} [e2e-llm-inference-service] f:schedulerName: {} [e2e-llm-inference-service] f:securityContext: {} [e2e-llm-inference-service] f:terminationGracePeriodSeconds: {} [e2e-llm-inference-service] f:volumes: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"name":"cabundle-cert"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:configMap: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:defaultMode: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] k:{"name":"dshm"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:emptyDir: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:medium: {} [e2e-llm-inference-service] f:sizeLimit: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] k:{"name":"home"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:emptyDir: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] k:{"name":"kserve-provision-location"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:emptyDir: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] k:{"name":"model-cache"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:emptyDir: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] k:{"name":"tls-certs"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:secret: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:defaultMode: {} [e2e-llm-inference-service] f:secretName: {} [e2e-llm-inference-service] k:{"name":"tmp-dir"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:emptyDir: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] - manager: kube-controller-manager [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] apiVersion: apps/v1 [e2e-llm-inference-service] time: '2026-07-07T14:47:16Z' [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:status: [e2e-llm-inference-service] f:availableReplicas: {} [e2e-llm-inference-service] f:fullyLabeledReplicas: {} [e2e-llm-inference-service] f:observedGeneration: {} [e2e-llm-inference-service] f:readyReplicas: {} [e2e-llm-inference-service] f:replicas: {} [e2e-llm-inference-service] subresource: status [e2e-llm-inference-service] spec: [e2e-llm-inference-service] replicas: 1 [e2e-llm-inference-service] selector: [e2e-llm-inference-service] matchLabels: [e2e-llm-inference-service] app.kubernetes.io/component: llminferenceservice-workload [e2e-llm-inference-service] app.kubernetes.io/name: llmisvc-model-fb-opt-125m-with-ba4d693a [e2e-llm-inference-service] app.kubernetes.io/part-of: llminferenceservice [e2e-llm-inference-service] kserve.io/component: workload [e2e-llm-inference-service] llm-d.ai/role: both [e2e-llm-inference-service] pod-template-hash: 5b998d5789 [e2e-llm-inference-service] template: [e2e-llm-inference-service] metadata: [e2e-llm-inference-service] labels: [e2e-llm-inference-service] app.kubernetes.io/component: llminferenceservice-workload [e2e-llm-inference-service] app.kubernetes.io/name: llmisvc-model-fb-opt-125m-with-ba4d693a [e2e-llm-inference-service] app.kubernetes.io/part-of: llminferenceservice [e2e-llm-inference-service] kserve.io/component: workload [e2e-llm-inference-service] llm-d.ai/role: both [e2e-llm-inference-service] pod-template-hash: 5b998d5789 [e2e-llm-inference-service] spec: [e2e-llm-inference-service] volumes: [e2e-llm-inference-service] - name: home [e2e-llm-inference-service] emptyDir: {} [e2e-llm-inference-service] - name: dshm [e2e-llm-inference-service] emptyDir: [e2e-llm-inference-service] medium: Memory [e2e-llm-inference-service] sizeLimit: 1Gi [e2e-llm-inference-service] - name: model-cache [e2e-llm-inference-service] emptyDir: {} [e2e-llm-inference-service] - name: tmp-dir [e2e-llm-inference-service] emptyDir: {} [e2e-llm-inference-service] - name: tls-certs [e2e-llm-inference-service] secret: [e2e-llm-inference-service] secretName: llmisv77ff2528d3e9b4972cd9335229fce9f0-kserve-self-signed-certs [e2e-llm-inference-service] defaultMode: 420 [e2e-llm-inference-service] - name: kserve-provision-location [e2e-llm-inference-service] emptyDir: {} [e2e-llm-inference-service] - name: cabundle-cert [e2e-llm-inference-service] configMap: [e2e-llm-inference-service] name: odh-kserve-custom-ca-bundle [e2e-llm-inference-service] defaultMode: 420 [e2e-llm-inference-service] initContainers: [e2e-llm-inference-service] - name: storage-initializer [e2e-llm-inference-service] image: quay.io/opendatahub/kserve-storage-initializer@sha256:13ad765c0f323200e1db7b45c99661287e65c5158d6b54b16139829636d20be4 [e2e-llm-inference-service] args: [e2e-llm-inference-service] - hf://facebook/opt-125m [e2e-llm-inference-service] - /mnt/models [e2e-llm-inference-service] - hf://edbeeching/opt-125m-lora [e2e-llm-inference-service] - /mnt/lora/lora-adapter-1 [e2e-llm-inference-service] env: [e2e-llm-inference-service] - name: AWS_ACCESS_KEY_ID [e2e-llm-inference-service] valueFrom: [e2e-llm-inference-service] secretKeyRef: [e2e-llm-inference-service] name: seaweedfs-s3-creds [e2e-llm-inference-service] key: AWS_ACCESS_KEY_ID [e2e-llm-inference-service] - name: AWS_SECRET_ACCESS_KEY [e2e-llm-inference-service] valueFrom: [e2e-llm-inference-service] secretKeyRef: [e2e-llm-inference-service] name: seaweedfs-s3-creds [e2e-llm-inference-service] key: AWS_SECRET_ACCESS_KEY [e2e-llm-inference-service] - name: S3_USE_HTTPS [e2e-llm-inference-service] value: '0' [e2e-llm-inference-service] - name: S3_ENDPOINT [e2e-llm-inference-service] value: s3-service.kserve:8333 [e2e-llm-inference-service] - name: AWS_ENDPOINT_URL [e2e-llm-inference-service] value: http://s3-service.kserve:8333 [e2e-llm-inference-service] - name: S3_VERIFY_SSL [e2e-llm-inference-service] value: '0' [e2e-llm-inference-service] - name: AWS_CA_BUNDLE [e2e-llm-inference-service] value: /etc/ssl/custom-certs/cabundle.crt [e2e-llm-inference-service] - name: AWS_CA_BUNDLE_CONFIGMAP [e2e-llm-inference-service] value: odh-kserve-custom-ca-bundle [e2e-llm-inference-service] - name: HF_HUB_ENABLE_HF_TRANSFER [e2e-llm-inference-service] value: '1' [e2e-llm-inference-service] - name: HF_XET_HIGH_PERFORMANCE [e2e-llm-inference-service] value: '1' [e2e-llm-inference-service] - name: HF_XET_NUM_CONCURRENT_RANGE_GETS [e2e-llm-inference-service] value: '8' [e2e-llm-inference-service] - name: CA_BUNDLE_CONFIGMAP_NAME [e2e-llm-inference-service] value: odh-kserve-custom-ca-bundle [e2e-llm-inference-service] - name: CA_BUNDLE_VOLUME_MOUNT_POINT [e2e-llm-inference-service] value: /etc/ssl/custom-certs [e2e-llm-inference-service] resources: [e2e-llm-inference-service] limits: [e2e-llm-inference-service] cpu: '1' [e2e-llm-inference-service] memory: 24Gi [e2e-llm-inference-service] requests: [e2e-llm-inference-service] cpu: 100m [e2e-llm-inference-service] memory: 100Mi [e2e-llm-inference-service] volumeMounts: [e2e-llm-inference-service] - name: kserve-provision-location [e2e-llm-inference-service] mountPath: /mnt [e2e-llm-inference-service] - name: cabundle-cert [e2e-llm-inference-service] readOnly: true [e2e-llm-inference-service] mountPath: /etc/ssl/custom-certs [e2e-llm-inference-service] terminationMessagePath: /dev/termination-log [e2e-llm-inference-service] terminationMessagePolicy: FallbackToLogsOnError [e2e-llm-inference-service] imagePullPolicy: IfNotPresent [e2e-llm-inference-service] containers: [e2e-llm-inference-service] - name: main [e2e-llm-inference-service] image: public.ecr.aws/q9t5s3a7/vllm-cpu-release-repo:v0.19.0 [e2e-llm-inference-service] command: [e2e-llm-inference-service] - /bin/bash [e2e-llm-inference-service] - -c [e2e-llm-inference-service] - "if [ -f /etc/profile.d/ibm-aiu-setup.sh ]; then\n source /etc/profile.d/ibm-aiu-setup.sh\n\ [e2e-llm-inference-service] fi\n\nif [ \"$KSERVE_INFER_ROCE\" = \"true\" ]; then\n echo \"Trying to\ [e2e-llm-inference-service] \ infer RoCE configs ... \"\n grep -H . /sys/class/infiniband/*/ports/*/gids/*\ [e2e-llm-inference-service] \ 2>/dev/null\n grep -H . /sys/class/infiniband/*/ports/*/gid_attrs/types/*\ [e2e-llm-inference-service] \ 2>/dev/null\n\n cat /proc/driver/nvidia/params\n\n KSERVE_INFER_IB_GID_INDEX_GREP=${KSERVE_INFER_IB_GID_INDEX_GREP:-\"\ [e2e-llm-inference-service] RoCE v2\"}\n\n echo \"[Infer RoCE] Discovering active HCAs ...\"\n active_hcas=()\n\ [e2e-llm-inference-service] \ # Loop through all mlx5 devices found in sysfs\n for hca_dir in /sys/class/infiniband/mlx5_*;\ [e2e-llm-inference-service] \ do\n # Ensure it's a directory before proceeding\n if [ -d \"\ [e2e-llm-inference-service] $hca_dir\" ]; then\n hca_name=$(basename \"$hca_dir\")\n \ [e2e-llm-inference-service] \ port_state_file=\"$hca_dir/ports/1/state\" # Assume port 1\n \ [e2e-llm-inference-service] \ type_file=\"$hca_dir/ports/1/gid_attrs/types/*\"\n\n echo\ [e2e-llm-inference-service] \ \"[Infer RoCE] Check if the port state file ${port_state_file} exists\ [e2e-llm-inference-service] \ and contains 'ACTIVE'\"\n if [ -f \"$port_state_file\" ] && grep\ [e2e-llm-inference-service] \ -q \"ACTIVE\" \"$port_state_file\" && grep -q \"${KSERVE_INFER_IB_GID_INDEX_GREP}\"\ [e2e-llm-inference-service] \ ${type_file} 2>/dev/null; then\n echo \"[Infer RoCE] Found\ [e2e-llm-inference-service] \ active HCA: $hca_name\"\n active_hcas+=(\"$hca_name\")\n\ [e2e-llm-inference-service] \ else\n echo \"[Infer RoCE] Skipping inactive or\ [e2e-llm-inference-service] \ down HCA: $hca_name\"\n fi\n fi\n done\n\n # Check if\ [e2e-llm-inference-service] \ we found any active HCAs\n if [ ${#active_hcas[@]} -gt 0 ]; then\n \ [e2e-llm-inference-service] \ # Join the array elements with a comma\n hca_port_pairs=()\n \ [e2e-llm-inference-service] \ for hca in \"${active_hcas[@]}\"; do\n hca_port_pairs+=(\"\ [e2e-llm-inference-service] ${hca}:1\")\n done\n\n active_hca_list=$(IFS=,; echo \"${active_hcas[*]}\"\ [e2e-llm-inference-service] )\n hca_port_pairs_list=$(IFS=,; echo \"${hca_port_pairs[*]}\")\n \ [e2e-llm-inference-service] \ echo \"[Infer RoCE] Setting active HCAs: ${active_hca_list}\"\n \ [e2e-llm-inference-service] \ export NCCL_IB_HCA=${NCCL_IB_HCA:-${active_hca_list}}\n export\ [e2e-llm-inference-service] \ NVSHMEM_HCA_LIST=${NVSHMEM_HCA_LIST:-${hca_port_pairs_list}}\n export\ [e2e-llm-inference-service] \ UCX_NET_DEVICES=${UCX_NET_DEVICES:-${hca_port_pairs_list}}\n\n echo\ [e2e-llm-inference-service] \ \"[Infer RoCE] NCCL_IB_HCA=${NCCL_IB_HCA}\"\n echo \"[Infer RoCE]\ [e2e-llm-inference-service] \ NVSHMEM_HCA_LIST=${NVSHMEM_HCA_LIST}\"\n echo \"[Infer RoCE] UCX_NET_DEVICES=${UCX_NET_DEVICES}\"\ [e2e-llm-inference-service] \n else\n echo \"[Infer RoCE] WARNING: No active RoCE HCAs found.\ [e2e-llm-inference-service] \ NCCL_IB_HCA will not be set.\"\n fi\n\n if [ ${#active_hcas[@]} -gt\ [e2e-llm-inference-service] \ 0 ]; then\n echo \"[Infer RoCE] Finding GID_INDEX for each active\ [e2e-llm-inference-service] \ HCA (SR-IOV compatible)...\"\n\n # For SR-IOV environments, find\ [e2e-llm-inference-service] \ the most common IPv4 RoCE v2 GID index across all HCAs\n declare\ [e2e-llm-inference-service] \ -A gid_index_count\n declare -A hca_gid_index\n\n for hca_name\ [e2e-llm-inference-service] \ in \"${active_hcas[@]}\"; do\n echo \"[Infer RoCE] Processing\ [e2e-llm-inference-service] \ HCA: ${hca_name}\"\n\n # Find all RoCE v2 IPv4 GIDs for this\ [e2e-llm-inference-service] \ HCA and count by index\n for tpath in /sys/class/infiniband/${hca_name}/ports/1/gid_attrs/types/*;\ [e2e-llm-inference-service] \ do\n if grep -q \"${KSERVE_INFER_IB_GID_INDEX_GREP}\" \"\ [e2e-llm-inference-service] $tpath\" 2>/dev/null; then\n idx=$(basename \"$tpath\"\ [e2e-llm-inference-service] )\n gid_file=\"/sys/class/infiniband/${hca_name}/ports/1/gids/${idx}\"\ [e2e-llm-inference-service] \n # Check for IPv4 GID (contains ffff:)\n \ [e2e-llm-inference-service] \ if [ -f \"$gid_file\" ] && grep -q \"ffff:\" \"$gid_file\"; then\n\ [e2e-llm-inference-service] \ gid_value=$(cat \"$gid_file\" 2>/dev/null || echo\ [e2e-llm-inference-service] \ \"\")\n echo \"[Infer RoCE] Found IPv4 RoCE v2 GID\ [e2e-llm-inference-service] \ for ${hca_name}: index=${idx}, gid=${gid_value}\"\n \ [e2e-llm-inference-service] \ hca_gid_index[\"${hca_name}\"]=\"${idx}\"\n gid_index_count[\"\ [e2e-llm-inference-service] ${idx}\"]=$((${gid_index_count[\"${idx}\"]} + 1))\n \ [e2e-llm-inference-service] \ break # Use first found IPv4 GID per HCA\n fi\n \ [e2e-llm-inference-service] \ fi\n done\n done\n\n # Find the most common\ [e2e-llm-inference-service] \ GID index (most likely to be consistent across nodes)\n best_gid_index=\"\ [e2e-llm-inference-service] \"\n max_count=0\n for idx in \"${!gid_index_count[@]}\"; do\n\ [e2e-llm-inference-service] \ count=${gid_index_count[\"${idx}\"]}\n echo \"[Infer\ [e2e-llm-inference-service] \ RoCE] GID_INDEX ${idx} found on ${count} HCAs\"\n if [ $count\ [e2e-llm-inference-service] \ -gt $max_count ]; then\n max_count=$count\n \ [e2e-llm-inference-service] \ best_gid_index=\"$idx\"\n fi\n done\n\n # Use deterministic\ [e2e-llm-inference-service] \ fallback if tied - prefer index 3 (SR-IOV standard)\n if [ ${#gid_index_count[@]}\ [e2e-llm-inference-service] \ -gt 1 ]; then\n echo \"[Infer RoCE] Multiple GID indices found,\ [e2e-llm-inference-service] \ selecting most common: ${best_gid_index}\"\n # If there's a tie,\ [e2e-llm-inference-service] \ prefer index 3 as it's most common in SR-IOV setups\n if [ -n\ [e2e-llm-inference-service] \ \"${gid_index_count['3']}\" ] && [ \"${gid_index_count['3']}\" -eq \"\ [e2e-llm-inference-service] $max_count\" ]; then\n best_gid_index=\"3\"\n \ [e2e-llm-inference-service] \ echo \"[Infer RoCE] Using deterministic fallback: GID_INDEX=3 (SR-IOV\ [e2e-llm-inference-service] \ standard)\"\n fi\n fi\n\n # Check if GID_INDEX is already\ [e2e-llm-inference-service] \ set via environment variables\n if [ -n \"${NCCL_IB_GID_INDEX}\"\ [e2e-llm-inference-service] \ ]; then\n echo \"[Infer RoCE] Using pre-configured NCCL_IB_GID_INDEX=${NCCL_IB_GID_INDEX}\ [e2e-llm-inference-service] \ from environment\"\n export NVSHMEM_IB_GID_INDEX=${NVSHMEM_IB_GID_INDEX:-$NCCL_IB_GID_INDEX}\n\ [e2e-llm-inference-service] \ export UCX_IB_GID_INDEX=${UCX_IB_GID_INDEX:-$NCCL_IB_GID_INDEX}\n\ [e2e-llm-inference-service] \ echo \"[Infer RoCE] Using pre-configured GID_INDEX=${NCCL_IB_GID_INDEX}\ [e2e-llm-inference-service] \ for NCCL, NVSHMEM, and UCX\"\n elif [ -n \"$best_gid_index\" ]; then\n\ [e2e-llm-inference-service] \ echo \"[Infer RoCE] Selected GID_INDEX: ${best_gid_index} (found\ [e2e-llm-inference-service] \ on ${max_count} HCAs)\"\n\n export NCCL_IB_GID_INDEX=${NCCL_IB_GID_INDEX:-$best_gid_index}\n\ [e2e-llm-inference-service] \ export NVSHMEM_IB_GID_INDEX=${NVSHMEM_IB_GID_INDEX:-$best_gid_index}\n\ [e2e-llm-inference-service] \ export UCX_IB_GID_INDEX=${UCX_IB_GID_INDEX:-$best_gid_index}\n\ [e2e-llm-inference-service] \n echo \"[Infer RoCE] Exported GID_INDEX=${best_gid_index} for\ [e2e-llm-inference-service] \ NCCL, NVSHMEM, and UCX\"\n else\n echo \"[Infer RoCE] ERROR:\ [e2e-llm-inference-service] \ No valid IPv4 ${KSERVE_INFER_IB_GID_INDEX_GREP} GID_INDEX found on any\ [e2e-llm-inference-service] \ HCA.\"\n fi\n else\n echo \"[Infer RoCE] No active HCAs found,\ [e2e-llm-inference-service] \ skipping GID_INDEX inference.\"\n fi\nfi\n\n# --disable-access-log-for-endpoints\ [e2e-llm-inference-service] \ landed in vLLM 0.16.0 (vllm-project/vllm#30011).\n# Older versions still\ [e2e-llm-inference-service] \ need the blanket --disable-uvicorn-access-log.\nACCESS_LOG_ARGS=\"--disable-uvicorn-access-log\"\ [e2e-llm-inference-service] \nVLLM_VERSION=$(vllm --version 2>/dev/null | tail -1 | awk '{print $NF}')\n\ [e2e-llm-inference-service] echo \"[access-log-detect] vllm version='${VLLM_VERSION}'\"\nif [[ \"$VLLM_VERSION\"\ [e2e-llm-inference-service] \ =~ ^[0-9]+\\.[0-9]+ ]] && [ \"$(printf '%s\\n%s\\n' \"0.16.0\" \"${VLLM_VERSION}\"\ [e2e-llm-inference-service] \ | sort -V | head -1)\" = \"0.16.0\" ]; then\n ACCESS_LOG_ARGS=\"--disable-access-log-for-endpoints\ [e2e-llm-inference-service] \ /health,/metrics,/ping\"\nfi\necho \"[access-log-detect] selected ACCESS_LOG_ARGS='${ACCESS_LOG_ARGS}'\"\ [e2e-llm-inference-service] \n\n# --shutdown-timeout landed in vLLM 0.18.0 (vllm-project/vllm#36666).\n\ [e2e-llm-inference-service] SHUTDOWN_TIMEOUT_ARGS=\"\"\nif [[ \"$VLLM_VERSION\" =~ ^[0-9]+\\.[0-9]+\ [e2e-llm-inference-service] \ ]] && [ \"$(printf '%s\\n%s\\n' \"0.18.0\" \"${VLLM_VERSION}\" | sort\ [e2e-llm-inference-service] \ -V | head -1)\" = \"0.18.0\" ]; then\n SHUTDOWN_TIMEOUT_ARGS=\"--shutdown-timeout\ [e2e-llm-inference-service] \ 40\"\nfi\n\neval \"exec vllm serve /mnt/models \\\n --served-model-name\ [e2e-llm-inference-service] \ \"facebook/opt-125m\" \"publishers/kserve-ci-e2e-test/models/facebook/opt-125m\"\ [e2e-llm-inference-service] \ \\\n --port 8000 \\\n ${ACCESS_LOG_ARGS} \\\n ${SHUTDOWN_TIMEOUT_ARGS}\ [e2e-llm-inference-service] \ \\\n --enable-ssl-refresh \\\n --ssl-certfile /var/run/kserve/tls/tls.crt\ [e2e-llm-inference-service] \ \\\n --ssl-keyfile /var/run/kserve/tls/tls.key \\\n ${VLLM_ADDITIONAL_ARGS}\ [e2e-llm-inference-service] \ \\\n $@\"" [e2e-llm-inference-service] - -- [e2e-llm-inference-service] args: [e2e-llm-inference-service] - --enable-lora [e2e-llm-inference-service] - --lora-modules [e2e-llm-inference-service] - '''{"name":"lora-adapter-1","path":"/mnt/lora/lora-adapter-1"}''' [e2e-llm-inference-service] - '''{"name":"publishers/kserve-ci-e2e-test/models/lora-adapter-1","path":"/mnt/lora/lora-adapter-1"}''' [e2e-llm-inference-service] ports: [e2e-llm-inference-service] - containerPort: 8000 [e2e-llm-inference-service] protocol: TCP [e2e-llm-inference-service] env: [e2e-llm-inference-service] - name: HOME [e2e-llm-inference-service] value: /home [e2e-llm-inference-service] - name: VLLM_LOGGING_LEVEL [e2e-llm-inference-service] value: DEBUG [e2e-llm-inference-service] - name: VLLM_CPU_KVCACHE_SPACE [e2e-llm-inference-service] value: '1' [e2e-llm-inference-service] - name: VLLM_ENABLE_V1_MULTIPROCESSING [e2e-llm-inference-service] value: '0' [e2e-llm-inference-service] - name: USER [e2e-llm-inference-service] value: nonroot [e2e-llm-inference-service] - name: TORCHINDUCTOR_CACHE_DIR [e2e-llm-inference-service] value: /tmp/torchinductor-cache [e2e-llm-inference-service] - name: HF_HUB_CACHE [e2e-llm-inference-service] value: /models [e2e-llm-inference-service] resources: [e2e-llm-inference-service] limits: [e2e-llm-inference-service] cpu: '2' [e2e-llm-inference-service] memory: 7Gi [e2e-llm-inference-service] requests: [e2e-llm-inference-service] cpu: 200m [e2e-llm-inference-service] memory: 2Gi [e2e-llm-inference-service] volumeMounts: [e2e-llm-inference-service] - name: home [e2e-llm-inference-service] mountPath: /home [e2e-llm-inference-service] - name: tmp-dir [e2e-llm-inference-service] mountPath: /tmp [e2e-llm-inference-service] - name: dshm [e2e-llm-inference-service] mountPath: /dev/shm [e2e-llm-inference-service] - name: model-cache [e2e-llm-inference-service] mountPath: /models [e2e-llm-inference-service] - name: tls-certs [e2e-llm-inference-service] readOnly: true [e2e-llm-inference-service] mountPath: /var/run/kserve/tls [e2e-llm-inference-service] - name: kserve-provision-location [e2e-llm-inference-service] readOnly: true [e2e-llm-inference-service] mountPath: /mnt [e2e-llm-inference-service] livenessProbe: [e2e-llm-inference-service] httpGet: [e2e-llm-inference-service] path: /health [e2e-llm-inference-service] port: 8000 [e2e-llm-inference-service] scheme: HTTPS [e2e-llm-inference-service] timeoutSeconds: 1 [e2e-llm-inference-service] periodSeconds: 10 [e2e-llm-inference-service] successThreshold: 1 [e2e-llm-inference-service] failureThreshold: 10 [e2e-llm-inference-service] readinessProbe: [e2e-llm-inference-service] httpGet: [e2e-llm-inference-service] path: /health [e2e-llm-inference-service] port: 8000 [e2e-llm-inference-service] scheme: HTTPS [e2e-llm-inference-service] timeoutSeconds: 1 [e2e-llm-inference-service] periodSeconds: 1 [e2e-llm-inference-service] successThreshold: 1 [e2e-llm-inference-service] failureThreshold: 2 [e2e-llm-inference-service] startupProbe: [e2e-llm-inference-service] httpGet: [e2e-llm-inference-service] path: /health [e2e-llm-inference-service] port: 8000 [e2e-llm-inference-service] scheme: HTTPS [e2e-llm-inference-service] timeoutSeconds: 1 [e2e-llm-inference-service] periodSeconds: 10 [e2e-llm-inference-service] successThreshold: 1 [e2e-llm-inference-service] failureThreshold: 60 [e2e-llm-inference-service] lifecycle: [e2e-llm-inference-service] preStop: [e2e-llm-inference-service] exec: [e2e-llm-inference-service] command: [e2e-llm-inference-service] - /bin/sleep [e2e-llm-inference-service] - '15' [e2e-llm-inference-service] terminationMessagePath: /dev/termination-log [e2e-llm-inference-service] terminationMessagePolicy: FallbackToLogsOnError [e2e-llm-inference-service] imagePullPolicy: IfNotPresent [e2e-llm-inference-service] securityContext: [e2e-llm-inference-service] capabilities: [e2e-llm-inference-service] drop: [e2e-llm-inference-service] - ALL [e2e-llm-inference-service] runAsNonRoot: true [e2e-llm-inference-service] readOnlyRootFilesystem: true [e2e-llm-inference-service] allowPrivilegeEscalation: false [e2e-llm-inference-service] seccompProfile: [e2e-llm-inference-service] type: RuntimeDefault [e2e-llm-inference-service] restartPolicy: Always [e2e-llm-inference-service] terminationGracePeriodSeconds: 60 [e2e-llm-inference-service] dnsPolicy: ClusterFirst [e2e-llm-inference-service] securityContext: {} [e2e-llm-inference-service] schedulerName: default-scheduler [e2e-llm-inference-service] status: [e2e-llm-inference-service] replicas: 1 [e2e-llm-inference-service] fullyLabeledReplicas: 1 [e2e-llm-inference-service] readyReplicas: 1 [e2e-llm-inference-service] availableReplicas: 1 [e2e-llm-inference-service] observedGeneration: 1 [e2e-llm-inference-service] apiVersion: apps/v1 [e2e-llm-inference-service] kind: ReplicaSet [e2e-llm-inference-service] [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1254 --- [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1255 metadata: [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-ba4d693a-kserve-router-scheduler-79f5455bb [e2e-llm-inference-service] namespace: kserve-ci-e2e-test [e2e-llm-inference-service] uid: 05afd940-da33-41e2-a13a-10e5c0538851 [e2e-llm-inference-service] resourceVersion: '46621' [e2e-llm-inference-service] generation: 1 [e2e-llm-inference-service] creationTimestamp: '2026-07-07T14:45:56Z' [e2e-llm-inference-service] labels: [e2e-llm-inference-service] app.kubernetes.io/component: llminferenceservice-router-scheduler [e2e-llm-inference-service] app.kubernetes.io/name: llmisvc-model-fb-opt-125m-with-ba4d693a [e2e-llm-inference-service] app.kubernetes.io/part-of: llminferenceservice [e2e-llm-inference-service] pod-template-hash: 79f5455bb [e2e-llm-inference-service] annotations: [e2e-llm-inference-service] deployment.kubernetes.io/desired-replicas: '1' [e2e-llm-inference-service] deployment.kubernetes.io/max-replicas: '1' [e2e-llm-inference-service] deployment.kubernetes.io/revision: '1' [e2e-llm-inference-service] ownerReferences: [e2e-llm-inference-service] - apiVersion: apps/v1 [e2e-llm-inference-service] kind: Deployment [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-ba4d693a-kserve-router-scheduler [e2e-llm-inference-service] uid: 15262fe4-b55c-4249-ae3a-b00445a310f9 [e2e-llm-inference-service] controller: true [e2e-llm-inference-service] blockOwnerDeletion: true [e2e-llm-inference-service] managedFields: [e2e-llm-inference-service] - manager: kube-controller-manager [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] apiVersion: apps/v1 [e2e-llm-inference-service] time: '2026-07-07T14:45:56Z' [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:metadata: [e2e-llm-inference-service] f:annotations: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:deployment.kubernetes.io/desired-replicas: {} [e2e-llm-inference-service] f:deployment.kubernetes.io/max-replicas: {} [e2e-llm-inference-service] f:deployment.kubernetes.io/revision: {} [e2e-llm-inference-service] f:labels: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:app.kubernetes.io/component: {} [e2e-llm-inference-service] f:app.kubernetes.io/name: {} [e2e-llm-inference-service] f:app.kubernetes.io/part-of: {} [e2e-llm-inference-service] f:pod-template-hash: {} [e2e-llm-inference-service] f:ownerReferences: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"uid":"15262fe4-b55c-4249-ae3a-b00445a310f9"}: {} [e2e-llm-inference-service] f:spec: [e2e-llm-inference-service] f:replicas: {} [e2e-llm-inference-service] f:selector: {} [e2e-llm-inference-service] f:template: [e2e-llm-inference-service] f:metadata: [e2e-llm-inference-service] f:annotations: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:app.kubernetes.io/version: {} [e2e-llm-inference-service] f:certificates.kserve.io/expiration-v2: {} [e2e-llm-inference-service] f:labels: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:app.kubernetes.io/component: {} [e2e-llm-inference-service] f:app.kubernetes.io/name: {} [e2e-llm-inference-service] f:app.kubernetes.io/part-of: {} [e2e-llm-inference-service] f:pod-template-hash: {} [e2e-llm-inference-service] f:spec: [e2e-llm-inference-service] f:containers: [e2e-llm-inference-service] k:{"name":"main"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:args: {} [e2e-llm-inference-service] f:command: {} [e2e-llm-inference-service] f:env: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"name":"SSL_CERT_DIR"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] f:image: {} [e2e-llm-inference-service] f:imagePullPolicy: {} [e2e-llm-inference-service] f:lifecycle: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:preStop: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:exec: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:command: {} [e2e-llm-inference-service] f:livenessProbe: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:failureThreshold: {} [e2e-llm-inference-service] f:grpc: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:port: {} [e2e-llm-inference-service] f:service: {} [e2e-llm-inference-service] f:initialDelaySeconds: {} [e2e-llm-inference-service] f:periodSeconds: {} [e2e-llm-inference-service] f:successThreshold: {} [e2e-llm-inference-service] f:timeoutSeconds: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:ports: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"containerPort":5557,"protocol":"TCP"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:containerPort: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:protocol: {} [e2e-llm-inference-service] k:{"containerPort":9002,"protocol":"TCP"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:containerPort: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:protocol: {} [e2e-llm-inference-service] k:{"containerPort":9003,"protocol":"TCP"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:containerPort: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:protocol: {} [e2e-llm-inference-service] k:{"containerPort":9090,"protocol":"TCP"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:containerPort: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:protocol: {} [e2e-llm-inference-service] f:readinessProbe: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:failureThreshold: {} [e2e-llm-inference-service] f:grpc: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:port: {} [e2e-llm-inference-service] f:service: {} [e2e-llm-inference-service] f:initialDelaySeconds: {} [e2e-llm-inference-service] f:periodSeconds: {} [e2e-llm-inference-service] f:successThreshold: {} [e2e-llm-inference-service] f:timeoutSeconds: {} [e2e-llm-inference-service] f:resources: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:requests: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:cpu: {} [e2e-llm-inference-service] f:memory: {} [e2e-llm-inference-service] f:securityContext: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:allowPrivilegeEscalation: {} [e2e-llm-inference-service] f:capabilities: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:drop: {} [e2e-llm-inference-service] f:readOnlyRootFilesystem: {} [e2e-llm-inference-service] f:runAsNonRoot: {} [e2e-llm-inference-service] f:seccompProfile: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:type: {} [e2e-llm-inference-service] f:terminationMessagePath: {} [e2e-llm-inference-service] f:terminationMessagePolicy: {} [e2e-llm-inference-service] f:volumeMounts: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"mountPath":"/tmp/tokenizer"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:mountPath: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] k:{"mountPath":"/var/run/kserve/tls"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:mountPath: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:readOnly: {} [e2e-llm-inference-service] f:dnsPolicy: {} [e2e-llm-inference-service] f:initContainers: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"name":"storage-initializer"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:args: {} [e2e-llm-inference-service] f:env: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"name":"AWS_ACCESS_KEY_ID"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:valueFrom: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:secretKeyRef: {} [e2e-llm-inference-service] k:{"name":"AWS_CA_BUNDLE"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"AWS_CA_BUNDLE_CONFIGMAP"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"AWS_ENDPOINT_URL"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"AWS_SECRET_ACCESS_KEY"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:valueFrom: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:secretKeyRef: {} [e2e-llm-inference-service] k:{"name":"HF_HUB_ENABLE_HF_TRANSFER"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"HF_XET_HIGH_PERFORMANCE"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"HF_XET_NUM_CONCURRENT_RANGE_GETS"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"S3_ENDPOINT"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"S3_USE_HTTPS"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"S3_VERIFY_SSL"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"STORAGE_ALLOW_PATTERNS"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] f:image: {} [e2e-llm-inference-service] f:imagePullPolicy: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:resources: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:limits: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:cpu: {} [e2e-llm-inference-service] f:memory: {} [e2e-llm-inference-service] f:requests: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:cpu: {} [e2e-llm-inference-service] f:memory: {} [e2e-llm-inference-service] f:terminationMessagePath: {} [e2e-llm-inference-service] f:terminationMessagePolicy: {} [e2e-llm-inference-service] f:volumeMounts: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"mountPath":"/mnt/models"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:mountPath: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:restartPolicy: {} [e2e-llm-inference-service] f:schedulerName: {} [e2e-llm-inference-service] f:securityContext: {} [e2e-llm-inference-service] f:serviceAccount: {} [e2e-llm-inference-service] f:serviceAccountName: {} [e2e-llm-inference-service] f:terminationGracePeriodSeconds: {} [e2e-llm-inference-service] f:volumes: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"name":"kserve-provision-location"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:emptyDir: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] k:{"name":"tls-certs"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:secret: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:defaultMode: {} [e2e-llm-inference-service] f:secretName: {} [e2e-llm-inference-service] k:{"name":"tokenizer-cache"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:emptyDir: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] k:{"name":"tokenizer-tmp"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:emptyDir: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] k:{"name":"tokenizer-uds"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:emptyDir: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] - manager: kube-controller-manager [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] apiVersion: apps/v1 [e2e-llm-inference-service] time: '2026-07-07T14:46:29Z' [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:status: [e2e-llm-inference-service] f:availableReplicas: {} [e2e-llm-inference-service] f:fullyLabeledReplicas: {} [e2e-llm-inference-service] f:observedGeneration: {} [e2e-llm-inference-service] f:readyReplicas: {} [e2e-llm-inference-service] f:replicas: {} [e2e-llm-inference-service] subresource: status [e2e-llm-inference-service] spec: [e2e-llm-inference-service] replicas: 1 [e2e-llm-inference-service] selector: [e2e-llm-inference-service] matchLabels: [e2e-llm-inference-service] app.kubernetes.io/component: llminferenceservice-router-scheduler [e2e-llm-inference-service] app.kubernetes.io/name: llmisvc-model-fb-opt-125m-with-ba4d693a [e2e-llm-inference-service] app.kubernetes.io/part-of: llminferenceservice [e2e-llm-inference-service] pod-template-hash: 79f5455bb [e2e-llm-inference-service] template: [e2e-llm-inference-service] metadata: [e2e-llm-inference-service] labels: [e2e-llm-inference-service] app.kubernetes.io/component: llminferenceservice-router-scheduler [e2e-llm-inference-service] app.kubernetes.io/name: llmisvc-model-fb-opt-125m-with-ba4d693a [e2e-llm-inference-service] app.kubernetes.io/part-of: llminferenceservice [e2e-llm-inference-service] pod-template-hash: 79f5455bb [e2e-llm-inference-service] annotations: [e2e-llm-inference-service] app.kubernetes.io/version: 0.9.0 [e2e-llm-inference-service] certificates.kserve.io/expiration-v2: 'true' [e2e-llm-inference-service] spec: [e2e-llm-inference-service] volumes: [e2e-llm-inference-service] - name: tls-certs [e2e-llm-inference-service] secret: [e2e-llm-inference-service] secretName: llmisv77ff2528d3e9b4972cd9335229fce9f0-kserve-self-signed-certs [e2e-llm-inference-service] defaultMode: 420 [e2e-llm-inference-service] - name: tokenizer-uds [e2e-llm-inference-service] emptyDir: {} [e2e-llm-inference-service] - name: tokenizer-tmp [e2e-llm-inference-service] emptyDir: {} [e2e-llm-inference-service] - name: tokenizer-cache [e2e-llm-inference-service] emptyDir: {} [e2e-llm-inference-service] - name: kserve-provision-location [e2e-llm-inference-service] emptyDir: {} [e2e-llm-inference-service] initContainers: [e2e-llm-inference-service] - name: storage-initializer [e2e-llm-inference-service] image: quay.io/opendatahub/kserve-storage-initializer@sha256:13ad765c0f323200e1db7b45c99661287e65c5158d6b54b16139829636d20be4 [e2e-llm-inference-service] args: [e2e-llm-inference-service] - hf://facebook/opt-125m [e2e-llm-inference-service] - /mnt/models [e2e-llm-inference-service] env: [e2e-llm-inference-service] - name: AWS_ACCESS_KEY_ID [e2e-llm-inference-service] valueFrom: [e2e-llm-inference-service] secretKeyRef: [e2e-llm-inference-service] name: seaweedfs-s3-creds [e2e-llm-inference-service] key: AWS_ACCESS_KEY_ID [e2e-llm-inference-service] - name: AWS_SECRET_ACCESS_KEY [e2e-llm-inference-service] valueFrom: [e2e-llm-inference-service] secretKeyRef: [e2e-llm-inference-service] name: seaweedfs-s3-creds [e2e-llm-inference-service] key: AWS_SECRET_ACCESS_KEY [e2e-llm-inference-service] - name: S3_USE_HTTPS [e2e-llm-inference-service] value: '0' [e2e-llm-inference-service] - name: S3_ENDPOINT [e2e-llm-inference-service] value: s3-service.kserve:8333 [e2e-llm-inference-service] - name: AWS_ENDPOINT_URL [e2e-llm-inference-service] value: http://s3-service.kserve:8333 [e2e-llm-inference-service] - name: S3_VERIFY_SSL [e2e-llm-inference-service] value: '0' [e2e-llm-inference-service] - name: AWS_CA_BUNDLE [e2e-llm-inference-service] value: /etc/ssl/custom-certs/cabundle.crt [e2e-llm-inference-service] - name: AWS_CA_BUNDLE_CONFIGMAP [e2e-llm-inference-service] value: odh-kserve-custom-ca-bundle [e2e-llm-inference-service] - name: HF_HUB_ENABLE_HF_TRANSFER [e2e-llm-inference-service] value: '1' [e2e-llm-inference-service] - name: HF_XET_HIGH_PERFORMANCE [e2e-llm-inference-service] value: '1' [e2e-llm-inference-service] - name: HF_XET_NUM_CONCURRENT_RANGE_GETS [e2e-llm-inference-service] value: '8' [e2e-llm-inference-service] - name: STORAGE_ALLOW_PATTERNS [e2e-llm-inference-service] value: '["tokenizer.json", "tokenizer_config.json", "special_tokens_map.json", [e2e-llm-inference-service] "vocab.json", "merges.txt", "config.json", "generation_config.json"]' [e2e-llm-inference-service] resources: [e2e-llm-inference-service] limits: [e2e-llm-inference-service] cpu: '1' [e2e-llm-inference-service] memory: 24Gi [e2e-llm-inference-service] requests: [e2e-llm-inference-service] cpu: 100m [e2e-llm-inference-service] memory: 100Mi [e2e-llm-inference-service] volumeMounts: [e2e-llm-inference-service] - name: kserve-provision-location [e2e-llm-inference-service] mountPath: /mnt/models [e2e-llm-inference-service] terminationMessagePath: /dev/termination-log [e2e-llm-inference-service] terminationMessagePolicy: FallbackToLogsOnError [e2e-llm-inference-service] imagePullPolicy: IfNotPresent [e2e-llm-inference-service] containers: [e2e-llm-inference-service] - name: main [e2e-llm-inference-service] image: ghcr.io/llm-d/llm-d-router-endpoint-picker:v0.9.0-rc.2 [e2e-llm-inference-service] command: [e2e-llm-inference-service] - /app/epp [e2e-llm-inference-service] - --pool-name [e2e-llm-inference-service] - llmisvc-model-fb-opt-125m-with-ba4d693a-inference-pool [e2e-llm-inference-service] - --pool-namespace [e2e-llm-inference-service] - kserve-ci-e2e-test [e2e-llm-inference-service] - --zap-encoder [e2e-llm-inference-service] - json [e2e-llm-inference-service] - --grpc-port [e2e-llm-inference-service] - '9002' [e2e-llm-inference-service] - --grpc-health-port [e2e-llm-inference-service] - '9003' [e2e-llm-inference-service] - --enable-cert-reload=true [e2e-llm-inference-service] - --secure-serving=true [e2e-llm-inference-service] - --model-server-metrics-scheme=https [e2e-llm-inference-service] - --cert-path=/var/run/kserve/tls [e2e-llm-inference-service] args: [e2e-llm-inference-service] - --config-text [e2e-llm-inference-service] - "apiVersion: inference.networking.x-k8s.io/v1alpha1\nkind: EndpointPickerConfig\n\ [e2e-llm-inference-service] plugins:\n- type: single-profile-handler\n- type: queue-scorer\n- type:\ [e2e-llm-inference-service] \ prefix-cache-scorer\n- type: max-score-picker\n- type: lora-affinity-scorer\n\ [e2e-llm-inference-service] schedulingProfiles:\n- name: default\n plugins:\n - pluginRef: lora-affinity-scorer\n\ [e2e-llm-inference-service] \ weight: 4\n - pluginRef: queue-scorer\n weight: 2\n - pluginRef:\ [e2e-llm-inference-service] \ prefix-cache-scorer\n weight: 3\n - pluginRef: max-score-picker\n" [e2e-llm-inference-service] ports: [e2e-llm-inference-service] - name: grpc [e2e-llm-inference-service] containerPort: 9002 [e2e-llm-inference-service] protocol: TCP [e2e-llm-inference-service] - name: grpc-health [e2e-llm-inference-service] containerPort: 9003 [e2e-llm-inference-service] protocol: TCP [e2e-llm-inference-service] - name: metrics [e2e-llm-inference-service] containerPort: 9090 [e2e-llm-inference-service] protocol: TCP [e2e-llm-inference-service] - name: zmq [e2e-llm-inference-service] containerPort: 5557 [e2e-llm-inference-service] protocol: TCP [e2e-llm-inference-service] env: [e2e-llm-inference-service] - name: SSL_CERT_DIR [e2e-llm-inference-service] value: /var/run/kserve/tls:/var/run/secrets/kubernetes.io/serviceaccount:/etc/pki/tls/certs [e2e-llm-inference-service] resources: [e2e-llm-inference-service] requests: [e2e-llm-inference-service] cpu: 256m [e2e-llm-inference-service] memory: 500Mi [e2e-llm-inference-service] volumeMounts: [e2e-llm-inference-service] - name: tls-certs [e2e-llm-inference-service] readOnly: true [e2e-llm-inference-service] mountPath: /var/run/kserve/tls [e2e-llm-inference-service] - name: tokenizer-uds [e2e-llm-inference-service] mountPath: /tmp/tokenizer [e2e-llm-inference-service] livenessProbe: [e2e-llm-inference-service] grpc: [e2e-llm-inference-service] port: 9003 [e2e-llm-inference-service] service: liveness [e2e-llm-inference-service] initialDelaySeconds: 5 [e2e-llm-inference-service] timeoutSeconds: 1 [e2e-llm-inference-service] periodSeconds: 10 [e2e-llm-inference-service] successThreshold: 1 [e2e-llm-inference-service] failureThreshold: 3 [e2e-llm-inference-service] readinessProbe: [e2e-llm-inference-service] grpc: [e2e-llm-inference-service] port: 9003 [e2e-llm-inference-service] service: readiness [e2e-llm-inference-service] initialDelaySeconds: 30 [e2e-llm-inference-service] timeoutSeconds: 1 [e2e-llm-inference-service] periodSeconds: 10 [e2e-llm-inference-service] successThreshold: 1 [e2e-llm-inference-service] failureThreshold: 3 [e2e-llm-inference-service] lifecycle: [e2e-llm-inference-service] preStop: [e2e-llm-inference-service] exec: [e2e-llm-inference-service] command: [e2e-llm-inference-service] - /bin/sleep [e2e-llm-inference-service] - '15' [e2e-llm-inference-service] terminationMessagePath: /dev/termination-log [e2e-llm-inference-service] terminationMessagePolicy: FallbackToLogsOnError [e2e-llm-inference-service] imagePullPolicy: IfNotPresent [e2e-llm-inference-service] securityContext: [e2e-llm-inference-service] capabilities: [e2e-llm-inference-service] drop: [e2e-llm-inference-service] - ALL [e2e-llm-inference-service] runAsNonRoot: true [e2e-llm-inference-service] readOnlyRootFilesystem: true [e2e-llm-inference-service] allowPrivilegeEscalation: false [e2e-llm-inference-service] seccompProfile: [e2e-llm-inference-service] type: RuntimeDefault [e2e-llm-inference-service] restartPolicy: Always [e2e-llm-inference-service] terminationGracePeriodSeconds: 60 [e2e-llm-inference-service] dnsPolicy: ClusterFirst [e2e-llm-inference-service] serviceAccountName: llmisvc-model-fb-opt-125m-with-ba4d693a-epp-sa [e2e-llm-inference-service] serviceAccount: llmisvc-model-fb-opt-125m-with-ba4d693a-epp-sa [e2e-llm-inference-service] securityContext: {} [e2e-llm-inference-service] schedulerName: default-scheduler [e2e-llm-inference-service] status: [e2e-llm-inference-service] replicas: 1 [e2e-llm-inference-service] fullyLabeledReplicas: 1 [e2e-llm-inference-service] readyReplicas: 1 [e2e-llm-inference-service] availableReplicas: 1 [e2e-llm-inference-service] observedGeneration: 1 [e2e-llm-inference-service] apiVersion: apps/v1 [e2e-llm-inference-service] kind: ReplicaSet [e2e-llm-inference-service] [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1254 --- [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1255 metadata: [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-ba4d693a-epp-rb [e2e-llm-inference-service] namespace: kserve-ci-e2e-test [e2e-llm-inference-service] uid: 4c790913-83d0-4c04-9203-88f36e531295 [e2e-llm-inference-service] resourceVersion: '45967' [e2e-llm-inference-service] creationTimestamp: '2026-07-07T14:45:56Z' [e2e-llm-inference-service] labels: [e2e-llm-inference-service] app.kubernetes.io/component: llminferenceservice-router-scheduler [e2e-llm-inference-service] app.kubernetes.io/name: llmisvc-model-fb-opt-125m-with-ba4d693a [e2e-llm-inference-service] app.kubernetes.io/part-of: llminferenceservice [e2e-llm-inference-service] ownerReferences: [e2e-llm-inference-service] - apiVersion: serving.kserve.io/v1alpha2 [e2e-llm-inference-service] kind: LLMInferenceService [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-ba4d693a [e2e-llm-inference-service] uid: 997d3cef-fabf-4b25-ad95-5ebe7a367fc5 [e2e-llm-inference-service] controller: true [e2e-llm-inference-service] blockOwnerDeletion: true [e2e-llm-inference-service] managedFields: [e2e-llm-inference-service] - manager: manager [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] apiVersion: rbac.authorization.k8s.io/v1 [e2e-llm-inference-service] time: '2026-07-07T14:45:56Z' [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:metadata: [e2e-llm-inference-service] f:labels: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:app.kubernetes.io/component: {} [e2e-llm-inference-service] f:app.kubernetes.io/name: {} [e2e-llm-inference-service] f:app.kubernetes.io/part-of: {} [e2e-llm-inference-service] f:ownerReferences: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"uid":"997d3cef-fabf-4b25-ad95-5ebe7a367fc5"}: {} [e2e-llm-inference-service] f:roleRef: {} [e2e-llm-inference-service] f:subjects: {} [e2e-llm-inference-service] subjects: [e2e-llm-inference-service] - kind: ServiceAccount [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-ba4d693a-epp-sa [e2e-llm-inference-service] namespace: kserve-ci-e2e-test [e2e-llm-inference-service] roleRef: [e2e-llm-inference-service] apiGroup: rbac.authorization.k8s.io [e2e-llm-inference-service] kind: Role [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-ba4d693a-epp-role [e2e-llm-inference-service] apiVersion: rbac.authorization.k8s.io/v1 [e2e-llm-inference-service] kind: RoleBinding [e2e-llm-inference-service] [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1254 --- [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1255 metadata: [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-ba4d693a-epp-role [e2e-llm-inference-service] namespace: kserve-ci-e2e-test [e2e-llm-inference-service] uid: a3676052-cd32-4566-87ab-4320a6cc62e1 [e2e-llm-inference-service] resourceVersion: '45961' [e2e-llm-inference-service] creationTimestamp: '2026-07-07T14:45:56Z' [e2e-llm-inference-service] labels: [e2e-llm-inference-service] app.kubernetes.io/component: llminferenceservice-router-scheduler [e2e-llm-inference-service] app.kubernetes.io/name: llmisvc-model-fb-opt-125m-with-ba4d693a [e2e-llm-inference-service] app.kubernetes.io/part-of: llminferenceservice [e2e-llm-inference-service] ownerReferences: [e2e-llm-inference-service] - apiVersion: serving.kserve.io/v1alpha2 [e2e-llm-inference-service] kind: LLMInferenceService [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-ba4d693a [e2e-llm-inference-service] uid: 997d3cef-fabf-4b25-ad95-5ebe7a367fc5 [e2e-llm-inference-service] controller: true [e2e-llm-inference-service] blockOwnerDeletion: true [e2e-llm-inference-service] managedFields: [e2e-llm-inference-service] - manager: manager [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] apiVersion: rbac.authorization.k8s.io/v1 [e2e-llm-inference-service] time: '2026-07-07T14:45:56Z' [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:metadata: [e2e-llm-inference-service] f:labels: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:app.kubernetes.io/component: {} [e2e-llm-inference-service] f:app.kubernetes.io/name: {} [e2e-llm-inference-service] f:app.kubernetes.io/part-of: {} [e2e-llm-inference-service] f:ownerReferences: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"uid":"997d3cef-fabf-4b25-ad95-5ebe7a367fc5"}: {} [e2e-llm-inference-service] f:rules: {} [e2e-llm-inference-service] rules: [e2e-llm-inference-service] - verbs: [e2e-llm-inference-service] - get [e2e-llm-inference-service] - list [e2e-llm-inference-service] - watch [e2e-llm-inference-service] apiGroups: [e2e-llm-inference-service] - '' [e2e-llm-inference-service] resources: [e2e-llm-inference-service] - pods [e2e-llm-inference-service] - verbs: [e2e-llm-inference-service] - get [e2e-llm-inference-service] - list [e2e-llm-inference-service] - watch [e2e-llm-inference-service] apiGroups: [e2e-llm-inference-service] - inference.networking.k8s.io [e2e-llm-inference-service] - inference.networking.x-k8s.io [e2e-llm-inference-service] resources: [e2e-llm-inference-service] - inferencepools [e2e-llm-inference-service] - inferenceobjectives [e2e-llm-inference-service] - inferencemodels [e2e-llm-inference-service] - verbs: [e2e-llm-inference-service] - get [e2e-llm-inference-service] - list [e2e-llm-inference-service] - watch [e2e-llm-inference-service] apiGroups: [e2e-llm-inference-service] - inference.networking.x-k8s.io [e2e-llm-inference-service] resources: [e2e-llm-inference-service] - inferencemodelrewrites [e2e-llm-inference-service] - inferencepoolimports [e2e-llm-inference-service] - verbs: [e2e-llm-inference-service] - get [e2e-llm-inference-service] - list [e2e-llm-inference-service] - watch [e2e-llm-inference-service] apiGroups: [e2e-llm-inference-service] - discovery.k8s.io [e2e-llm-inference-service] resources: [e2e-llm-inference-service] - endpointslices [e2e-llm-inference-service] - verbs: [e2e-llm-inference-service] - get [e2e-llm-inference-service] - list [e2e-llm-inference-service] - watch [e2e-llm-inference-service] - create [e2e-llm-inference-service] - update [e2e-llm-inference-service] - patch [e2e-llm-inference-service] - delete [e2e-llm-inference-service] apiGroups: [e2e-llm-inference-service] - coordination.k8s.io [e2e-llm-inference-service] resources: [e2e-llm-inference-service] - leases [e2e-llm-inference-service] apiVersion: rbac.authorization.k8s.io/v1 [e2e-llm-inference-service] kind: Role [e2e-llm-inference-service] [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1254 --- [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1255 metadata: [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-ba4d693a-epp-service-grc5v [e2e-llm-inference-service] generateName: llmisvc-model-fb-opt-125m-with-ba4d693a-epp-service- [e2e-llm-inference-service] namespace: kserve-ci-e2e-test [e2e-llm-inference-service] uid: c997270b-fd83-4d44-9602-c88aa276d578 [e2e-llm-inference-service] resourceVersion: '46620' [e2e-llm-inference-service] generation: 3 [e2e-llm-inference-service] creationTimestamp: '2026-07-07T14:45:56Z' [e2e-llm-inference-service] labels: [e2e-llm-inference-service] app.kubernetes.io/component: llminferenceservice-router-scheduler [e2e-llm-inference-service] app.kubernetes.io/name: llmisvc-model-fb-opt-125m-with-ba4d693a [e2e-llm-inference-service] app.kubernetes.io/part-of: llminferenceservice [e2e-llm-inference-service] endpointslice.kubernetes.io/managed-by: endpointslice-controller.k8s.io [e2e-llm-inference-service] kubernetes.io/service-name: llmisvc-model-fb-opt-125m-with-ba4d693a-epp-service [e2e-llm-inference-service] annotations: [e2e-llm-inference-service] endpoints.kubernetes.io/last-change-trigger-time: '2026-07-07T14:46:29Z' [e2e-llm-inference-service] ownerReferences: [e2e-llm-inference-service] - apiVersion: v1 [e2e-llm-inference-service] kind: Service [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-ba4d693a-epp-service [e2e-llm-inference-service] uid: 27ed2141-ccc0-4ffe-aeed-63126edda97e [e2e-llm-inference-service] controller: true [e2e-llm-inference-service] blockOwnerDeletion: true [e2e-llm-inference-service] managedFields: [e2e-llm-inference-service] - manager: kube-controller-manager [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] apiVersion: discovery.k8s.io/v1 [e2e-llm-inference-service] time: '2026-07-07T14:46:29Z' [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:addressType: {} [e2e-llm-inference-service] f:endpoints: {} [e2e-llm-inference-service] f:metadata: [e2e-llm-inference-service] f:annotations: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:endpoints.kubernetes.io/last-change-trigger-time: {} [e2e-llm-inference-service] f:generateName: {} [e2e-llm-inference-service] f:labels: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:app.kubernetes.io/component: {} [e2e-llm-inference-service] f:app.kubernetes.io/name: {} [e2e-llm-inference-service] f:app.kubernetes.io/part-of: {} [e2e-llm-inference-service] f:endpointslice.kubernetes.io/managed-by: {} [e2e-llm-inference-service] f:kubernetes.io/service-name: {} [e2e-llm-inference-service] f:ownerReferences: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"uid":"27ed2141-ccc0-4ffe-aeed-63126edda97e"}: {} [e2e-llm-inference-service] f:ports: {} [e2e-llm-inference-service] addressType: IPv4 [e2e-llm-inference-service] endpoints: [e2e-llm-inference-service] - addresses: [e2e-llm-inference-service] - 10.132.0.54 [e2e-llm-inference-service] conditions: [e2e-llm-inference-service] ready: true [e2e-llm-inference-service] serving: true [e2e-llm-inference-service] terminating: false [e2e-llm-inference-service] targetRef: [e2e-llm-inference-service] kind: Pod [e2e-llm-inference-service] namespace: kserve-ci-e2e-test [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-ba4d693a-kserve-router-sche4ghr9 [e2e-llm-inference-service] uid: 66635d02-cb85-4e6d-8342-c235b66670ec [e2e-llm-inference-service] nodeName: ip-10-0-133-25.ec2.internal [e2e-llm-inference-service] zone: us-east-1a [e2e-llm-inference-service] ports: [e2e-llm-inference-service] - name: grpc [e2e-llm-inference-service] protocol: TCP [e2e-llm-inference-service] port: 9002 [e2e-llm-inference-service] - name: grpc-health [e2e-llm-inference-service] protocol: TCP [e2e-llm-inference-service] port: 9003 [e2e-llm-inference-service] - name: metrics [e2e-llm-inference-service] protocol: TCP [e2e-llm-inference-service] port: 9090 [e2e-llm-inference-service] - name: zmq [e2e-llm-inference-service] protocol: TCP [e2e-llm-inference-service] port: 5557 [e2e-llm-inference-service] apiVersion: discovery.k8s.io/v1 [e2e-llm-inference-service] kind: EndpointSlice [e2e-llm-inference-service] [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1254 --- [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1255 metadata: [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-ba4d693a-kserve-workload-svtcpc4 [e2e-llm-inference-service] generateName: llmisvc-model-fb-opt-125m-with-ba4d693a-kserve-workload-svc- [e2e-llm-inference-service] namespace: kserve-ci-e2e-test [e2e-llm-inference-service] uid: 4fbbb52f-d405-4de4-8b88-ed40d7cb7c21 [e2e-llm-inference-service] resourceVersion: '47363' [e2e-llm-inference-service] generation: 3 [e2e-llm-inference-service] creationTimestamp: '2026-07-07T14:45:56Z' [e2e-llm-inference-service] labels: [e2e-llm-inference-service] app.kubernetes.io/component: llminferenceservice-workload [e2e-llm-inference-service] app.kubernetes.io/name: llmisvc-model-fb-opt-125m-with-ba4d693a [e2e-llm-inference-service] app.kubernetes.io/part-of: llminferenceservice [e2e-llm-inference-service] endpointslice.kubernetes.io/managed-by: endpointslice-controller.k8s.io [e2e-llm-inference-service] kubernetes.io/service-name: llmisvc-model-fb-opt-125m-with-ba4d693a-kserve-workload-svc [e2e-llm-inference-service] annotations: [e2e-llm-inference-service] endpoints.kubernetes.io/last-change-trigger-time: '2026-07-07T14:47:16Z' [e2e-llm-inference-service] ownerReferences: [e2e-llm-inference-service] - apiVersion: v1 [e2e-llm-inference-service] kind: Service [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-ba4d693a-kserve-workload-svc [e2e-llm-inference-service] uid: 687542da-fbcf-42b8-ae26-8d19e584f912 [e2e-llm-inference-service] controller: true [e2e-llm-inference-service] blockOwnerDeletion: true [e2e-llm-inference-service] managedFields: [e2e-llm-inference-service] - manager: kube-controller-manager [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] apiVersion: discovery.k8s.io/v1 [e2e-llm-inference-service] time: '2026-07-07T14:47:16Z' [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:addressType: {} [e2e-llm-inference-service] f:endpoints: {} [e2e-llm-inference-service] f:metadata: [e2e-llm-inference-service] f:annotations: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:endpoints.kubernetes.io/last-change-trigger-time: {} [e2e-llm-inference-service] f:generateName: {} [e2e-llm-inference-service] f:labels: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:app.kubernetes.io/component: {} [e2e-llm-inference-service] f:app.kubernetes.io/name: {} [e2e-llm-inference-service] f:app.kubernetes.io/part-of: {} [e2e-llm-inference-service] f:endpointslice.kubernetes.io/managed-by: {} [e2e-llm-inference-service] f:kubernetes.io/service-name: {} [e2e-llm-inference-service] f:ownerReferences: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"uid":"687542da-fbcf-42b8-ae26-8d19e584f912"}: {} [e2e-llm-inference-service] f:ports: {} [e2e-llm-inference-service] addressType: IPv4 [e2e-llm-inference-service] endpoints: [e2e-llm-inference-service] - addresses: [e2e-llm-inference-service] - 10.134.0.32 [e2e-llm-inference-service] conditions: [e2e-llm-inference-service] ready: true [e2e-llm-inference-service] serving: true [e2e-llm-inference-service] terminating: false [e2e-llm-inference-service] targetRef: [e2e-llm-inference-service] kind: Pod [e2e-llm-inference-service] namespace: kserve-ci-e2e-test [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-ba4d693a-kserve-5b998d5789-t6wf4 [e2e-llm-inference-service] uid: 68c5609b-442c-4a9c-837a-67c032fcf236 [e2e-llm-inference-service] nodeName: ip-10-0-139-47.ec2.internal [e2e-llm-inference-service] zone: us-east-1a [e2e-llm-inference-service] ports: [e2e-llm-inference-service] - name: https [e2e-llm-inference-service] protocol: TCP [e2e-llm-inference-service] port: 8000 [e2e-llm-inference-service] appProtocol: https [e2e-llm-inference-service] apiVersion: discovery.k8s.io/v1 [e2e-llm-inference-service] kind: EndpointSlice [e2e-llm-inference-service] [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1254 --- [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1255 metadata: [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-ba4d693a-epp-rb [e2e-llm-inference-service] namespace: kserve-ci-e2e-test [e2e-llm-inference-service] uid: 4c790913-83d0-4c04-9203-88f36e531295 [e2e-llm-inference-service] resourceVersion: '45967' [e2e-llm-inference-service] creationTimestamp: '2026-07-07T14:45:56Z' [e2e-llm-inference-service] labels: [e2e-llm-inference-service] app.kubernetes.io/component: llminferenceservice-router-scheduler [e2e-llm-inference-service] app.kubernetes.io/name: llmisvc-model-fb-opt-125m-with-ba4d693a [e2e-llm-inference-service] app.kubernetes.io/part-of: llminferenceservice [e2e-llm-inference-service] ownerReferences: [e2e-llm-inference-service] - apiVersion: serving.kserve.io/v1alpha2 [e2e-llm-inference-service] kind: LLMInferenceService [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-ba4d693a [e2e-llm-inference-service] uid: 997d3cef-fabf-4b25-ad95-5ebe7a367fc5 [e2e-llm-inference-service] controller: true [e2e-llm-inference-service] blockOwnerDeletion: true [e2e-llm-inference-service] managedFields: [e2e-llm-inference-service] - manager: manager [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] apiVersion: rbac.authorization.k8s.io/v1 [e2e-llm-inference-service] time: '2026-07-07T14:45:56Z' [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:metadata: [e2e-llm-inference-service] f:labels: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:app.kubernetes.io/component: {} [e2e-llm-inference-service] f:app.kubernetes.io/name: {} [e2e-llm-inference-service] f:app.kubernetes.io/part-of: {} [e2e-llm-inference-service] f:ownerReferences: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"uid":"997d3cef-fabf-4b25-ad95-5ebe7a367fc5"}: {} [e2e-llm-inference-service] f:roleRef: {} [e2e-llm-inference-service] f:subjects: {} [e2e-llm-inference-service] userNames: [e2e-llm-inference-service] - system:serviceaccount:kserve-ci-e2e-test:llmisvc-model-fb-opt-125m-with-ba4d693a-epp-sa [e2e-llm-inference-service] groupNames: null [e2e-llm-inference-service] subjects: [e2e-llm-inference-service] - kind: ServiceAccount [e2e-llm-inference-service] namespace: kserve-ci-e2e-test [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-ba4d693a-epp-sa [e2e-llm-inference-service] roleRef: [e2e-llm-inference-service] namespace: kserve-ci-e2e-test [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-ba4d693a-epp-role [e2e-llm-inference-service] apiVersion: authorization.openshift.io/v1 [e2e-llm-inference-service] kind: RoleBinding [e2e-llm-inference-service] [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1254 --- [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1255 metadata: [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-ba4d693a-epp-role [e2e-llm-inference-service] namespace: kserve-ci-e2e-test [e2e-llm-inference-service] uid: a3676052-cd32-4566-87ab-4320a6cc62e1 [e2e-llm-inference-service] resourceVersion: '45961' [e2e-llm-inference-service] creationTimestamp: '2026-07-07T14:45:56Z' [e2e-llm-inference-service] labels: [e2e-llm-inference-service] app.kubernetes.io/component: llminferenceservice-router-scheduler [e2e-llm-inference-service] app.kubernetes.io/name: llmisvc-model-fb-opt-125m-with-ba4d693a [e2e-llm-inference-service] app.kubernetes.io/part-of: llminferenceservice [e2e-llm-inference-service] ownerReferences: [e2e-llm-inference-service] - apiVersion: serving.kserve.io/v1alpha2 [e2e-llm-inference-service] kind: LLMInferenceService [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-ba4d693a [e2e-llm-inference-service] uid: 997d3cef-fabf-4b25-ad95-5ebe7a367fc5 [e2e-llm-inference-service] controller: true [e2e-llm-inference-service] blockOwnerDeletion: true [e2e-llm-inference-service] managedFields: [e2e-llm-inference-service] - manager: manager [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] apiVersion: rbac.authorization.k8s.io/v1 [e2e-llm-inference-service] time: '2026-07-07T14:45:56Z' [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:metadata: [e2e-llm-inference-service] f:labels: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:app.kubernetes.io/component: {} [e2e-llm-inference-service] f:app.kubernetes.io/name: {} [e2e-llm-inference-service] f:app.kubernetes.io/part-of: {} [e2e-llm-inference-service] f:ownerReferences: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"uid":"997d3cef-fabf-4b25-ad95-5ebe7a367fc5"}: {} [e2e-llm-inference-service] f:rules: {} [e2e-llm-inference-service] rules: [e2e-llm-inference-service] - verbs: [e2e-llm-inference-service] - get [e2e-llm-inference-service] - list [e2e-llm-inference-service] - watch [e2e-llm-inference-service] attributeRestrictions: null [e2e-llm-inference-service] apiGroups: [e2e-llm-inference-service] - '' [e2e-llm-inference-service] resources: [e2e-llm-inference-service] - pods [e2e-llm-inference-service] - verbs: [e2e-llm-inference-service] - get [e2e-llm-inference-service] - list [e2e-llm-inference-service] - watch [e2e-llm-inference-service] attributeRestrictions: null [e2e-llm-inference-service] apiGroups: [e2e-llm-inference-service] - inference.networking.k8s.io [e2e-llm-inference-service] - inference.networking.x-k8s.io [e2e-llm-inference-service] resources: [e2e-llm-inference-service] - inferencemodels [e2e-llm-inference-service] - inferenceobjectives [e2e-llm-inference-service] - inferencepools [e2e-llm-inference-service] - verbs: [e2e-llm-inference-service] - get [e2e-llm-inference-service] - list [e2e-llm-inference-service] - watch [e2e-llm-inference-service] attributeRestrictions: null [e2e-llm-inference-service] apiGroups: [e2e-llm-inference-service] - inference.networking.x-k8s.io [e2e-llm-inference-service] resources: [e2e-llm-inference-service] - inferencemodelrewrites [e2e-llm-inference-service] - inferencepoolimports [e2e-llm-inference-service] - verbs: [e2e-llm-inference-service] - get [e2e-llm-inference-service] - list [e2e-llm-inference-service] - watch [e2e-llm-inference-service] attributeRestrictions: null [e2e-llm-inference-service] apiGroups: [e2e-llm-inference-service] - discovery.k8s.io [e2e-llm-inference-service] resources: [e2e-llm-inference-service] - endpointslices [e2e-llm-inference-service] - verbs: [e2e-llm-inference-service] - create [e2e-llm-inference-service] - delete [e2e-llm-inference-service] - get [e2e-llm-inference-service] - list [e2e-llm-inference-service] - patch [e2e-llm-inference-service] - update [e2e-llm-inference-service] - watch [e2e-llm-inference-service] attributeRestrictions: null [e2e-llm-inference-service] apiGroups: [e2e-llm-inference-service] - coordination.k8s.io [e2e-llm-inference-service] resources: [e2e-llm-inference-service] - leases [e2e-llm-inference-service] apiVersion: authorization.openshift.io/v1 [e2e-llm-inference-service] kind: Role [e2e-llm-inference-service] [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1254 --- [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1255 apiVersion: gateway.networking.k8s.io/v1 [e2e-llm-inference-service] kind: HTTPRoute [e2e-llm-inference-service] metadata: [e2e-llm-inference-service] annotations: [e2e-llm-inference-service] serving.kserve.io/inference-pool-migrated: v1 [e2e-llm-inference-service] creationTimestamp: '2026-07-07T14:45:56Z' [e2e-llm-inference-service] generation: 2 [e2e-llm-inference-service] labels: [e2e-llm-inference-service] app.kubernetes.io/component: llminferenceservice-router [e2e-llm-inference-service] app.kubernetes.io/name: llmisvc-model-fb-opt-125m-with-ba4d693a [e2e-llm-inference-service] app.kubernetes.io/part-of: llminferenceservice [e2e-llm-inference-service] managedFields: [e2e-llm-inference-service] - apiVersion: gateway.networking.k8s.io/v1 [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:metadata: [e2e-llm-inference-service] f:annotations: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:serving.kserve.io/inference-pool-migrated: {} [e2e-llm-inference-service] f:labels: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:app.kubernetes.io/component: {} [e2e-llm-inference-service] f:app.kubernetes.io/name: {} [e2e-llm-inference-service] f:app.kubernetes.io/part-of: {} [e2e-llm-inference-service] f:ownerReferences: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"uid":"997d3cef-fabf-4b25-ad95-5ebe7a367fc5"}: {} [e2e-llm-inference-service] f:spec: {} [e2e-llm-inference-service] manager: manager [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] time: '2026-07-07T14:46:16Z' [e2e-llm-inference-service] - apiVersion: gateway.networking.k8s.io/v1 [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:status: [e2e-llm-inference-service] f:parents: {} [e2e-llm-inference-service] manager: manager [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] subresource: status [e2e-llm-inference-service] time: '2026-07-07T14:46:17Z' [e2e-llm-inference-service] - apiVersion: gateway.networking.k8s.io/v1beta1 [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:status: {} [e2e-llm-inference-service] manager: pilot-discovery [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] subresource: status [e2e-llm-inference-service] time: '2026-07-07T14:46:17Z' [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-ba4d693a-kserve-route [e2e-llm-inference-service] namespace: kserve-ci-e2e-test [e2e-llm-inference-service] ownerReferences: [e2e-llm-inference-service] - apiVersion: serving.kserve.io/v1alpha2 [e2e-llm-inference-service] blockOwnerDeletion: true [e2e-llm-inference-service] controller: true [e2e-llm-inference-service] kind: LLMInferenceService [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-ba4d693a [e2e-llm-inference-service] uid: 997d3cef-fabf-4b25-ad95-5ebe7a367fc5 [e2e-llm-inference-service] resourceVersion: '46438' [e2e-llm-inference-service] uid: f21bbfa3-4d7b-4115-8b05-b9d581d11772 [e2e-llm-inference-service] spec: [e2e-llm-inference-service] parentRefs: [e2e-llm-inference-service] - group: gateway.networking.k8s.io [e2e-llm-inference-service] kind: Gateway [e2e-llm-inference-service] name: openshift-ai-inference [e2e-llm-inference-service] namespace: openshift-ingress [e2e-llm-inference-service] rules: [e2e-llm-inference-service] - backendRefs: [e2e-llm-inference-service] - group: inference.networking.k8s.io [e2e-llm-inference-service] kind: InferencePool [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-ba4d693a-inference-pool [e2e-llm-inference-service] port: 8000 [e2e-llm-inference-service] weight: 1 [e2e-llm-inference-service] filters: [e2e-llm-inference-service] - type: URLRewrite [e2e-llm-inference-service] urlRewrite: [e2e-llm-inference-service] path: [e2e-llm-inference-service] replacePrefixMatch: /v1/completions [e2e-llm-inference-service] type: ReplacePrefixMatch [e2e-llm-inference-service] matches: [e2e-llm-inference-service] - path: [e2e-llm-inference-service] type: PathPrefix [e2e-llm-inference-service] value: /kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-with-ba4d693a/v1/completions [e2e-llm-inference-service] timeouts: [e2e-llm-inference-service] backendRequest: 0s [e2e-llm-inference-service] request: 0s [e2e-llm-inference-service] - backendRefs: [e2e-llm-inference-service] - group: inference.networking.k8s.io [e2e-llm-inference-service] kind: InferencePool [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-ba4d693a-inference-pool [e2e-llm-inference-service] port: 8000 [e2e-llm-inference-service] weight: 1 [e2e-llm-inference-service] matches: [e2e-llm-inference-service] - headers: [e2e-llm-inference-service] - name: X-Gateway-Model-Name [e2e-llm-inference-service] type: Exact [e2e-llm-inference-service] value: publishers/kserve-ci-e2e-test/models/facebook/opt-125m [e2e-llm-inference-service] path: [e2e-llm-inference-service] type: Exact [e2e-llm-inference-service] value: /v1/completions [e2e-llm-inference-service] - headers: [e2e-llm-inference-service] - name: X-Gateway-Model-Name [e2e-llm-inference-service] type: Exact [e2e-llm-inference-service] value: publishers/kserve-ci-e2e-test/models/facebook/opt-125m [e2e-llm-inference-service] path: [e2e-llm-inference-service] type: Exact [e2e-llm-inference-service] value: /v1/completions/ [e2e-llm-inference-service] - headers: [e2e-llm-inference-service] - name: X-Gateway-Model-Name [e2e-llm-inference-service] type: Exact [e2e-llm-inference-service] value: publishers/kserve-ci-e2e-test/models/lora-adapter-1 [e2e-llm-inference-service] path: [e2e-llm-inference-service] type: Exact [e2e-llm-inference-service] value: /v1/completions [e2e-llm-inference-service] - headers: [e2e-llm-inference-service] - name: X-Gateway-Model-Name [e2e-llm-inference-service] type: Exact [e2e-llm-inference-service] value: publishers/kserve-ci-e2e-test/models/lora-adapter-1 [e2e-llm-inference-service] path: [e2e-llm-inference-service] type: Exact [e2e-llm-inference-service] value: /v1/completions/ [e2e-llm-inference-service] timeouts: [e2e-llm-inference-service] backendRequest: 0s [e2e-llm-inference-service] request: 0s [e2e-llm-inference-service] - backendRefs: [e2e-llm-inference-service] - group: inference.networking.k8s.io [e2e-llm-inference-service] kind: InferencePool [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-ba4d693a-inference-pool [e2e-llm-inference-service] port: 8000 [e2e-llm-inference-service] weight: 1 [e2e-llm-inference-service] filters: [e2e-llm-inference-service] - type: URLRewrite [e2e-llm-inference-service] urlRewrite: [e2e-llm-inference-service] path: [e2e-llm-inference-service] replacePrefixMatch: /v1/chat/completions [e2e-llm-inference-service] type: ReplacePrefixMatch [e2e-llm-inference-service] matches: [e2e-llm-inference-service] - path: [e2e-llm-inference-service] type: PathPrefix [e2e-llm-inference-service] value: /kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-with-ba4d693a/v1/chat/completions [e2e-llm-inference-service] timeouts: [e2e-llm-inference-service] backendRequest: 0s [e2e-llm-inference-service] request: 0s [e2e-llm-inference-service] - backendRefs: [e2e-llm-inference-service] - group: inference.networking.k8s.io [e2e-llm-inference-service] kind: InferencePool [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-ba4d693a-inference-pool [e2e-llm-inference-service] port: 8000 [e2e-llm-inference-service] weight: 1 [e2e-llm-inference-service] matches: [e2e-llm-inference-service] - headers: [e2e-llm-inference-service] - name: X-Gateway-Model-Name [e2e-llm-inference-service] type: Exact [e2e-llm-inference-service] value: publishers/kserve-ci-e2e-test/models/facebook/opt-125m [e2e-llm-inference-service] path: [e2e-llm-inference-service] type: Exact [e2e-llm-inference-service] value: /v1/chat/completions [e2e-llm-inference-service] - headers: [e2e-llm-inference-service] - name: X-Gateway-Model-Name [e2e-llm-inference-service] type: Exact [e2e-llm-inference-service] value: publishers/kserve-ci-e2e-test/models/facebook/opt-125m [e2e-llm-inference-service] path: [e2e-llm-inference-service] type: Exact [e2e-llm-inference-service] value: /v1/chat/completions/ [e2e-llm-inference-service] - headers: [e2e-llm-inference-service] - name: X-Gateway-Model-Name [e2e-llm-inference-service] type: Exact [e2e-llm-inference-service] value: publishers/kserve-ci-e2e-test/models/lora-adapter-1 [e2e-llm-inference-service] path: [e2e-llm-inference-service] type: Exact [e2e-llm-inference-service] value: /v1/chat/completions [e2e-llm-inference-service] - headers: [e2e-llm-inference-service] - name: X-Gateway-Model-Name [e2e-llm-inference-service] type: Exact [e2e-llm-inference-service] value: publishers/kserve-ci-e2e-test/models/lora-adapter-1 [e2e-llm-inference-service] path: [e2e-llm-inference-service] type: Exact [e2e-llm-inference-service] value: /v1/chat/completions/ [e2e-llm-inference-service] timeouts: [e2e-llm-inference-service] backendRequest: 0s [e2e-llm-inference-service] request: 0s [e2e-llm-inference-service] - backendRefs: [e2e-llm-inference-service] - group: inference.networking.k8s.io [e2e-llm-inference-service] kind: InferencePool [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-ba4d693a-inference-pool [e2e-llm-inference-service] port: 8000 [e2e-llm-inference-service] weight: 1 [e2e-llm-inference-service] filters: [e2e-llm-inference-service] - type: URLRewrite [e2e-llm-inference-service] urlRewrite: [e2e-llm-inference-service] path: [e2e-llm-inference-service] replacePrefixMatch: /v1/responses [e2e-llm-inference-service] type: ReplacePrefixMatch [e2e-llm-inference-service] matches: [e2e-llm-inference-service] - path: [e2e-llm-inference-service] type: PathPrefix [e2e-llm-inference-service] value: /kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-with-ba4d693a/v1/responses [e2e-llm-inference-service] timeouts: [e2e-llm-inference-service] backendRequest: 0s [e2e-llm-inference-service] request: 0s [e2e-llm-inference-service] - backendRefs: [e2e-llm-inference-service] - group: inference.networking.k8s.io [e2e-llm-inference-service] kind: InferencePool [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-ba4d693a-inference-pool [e2e-llm-inference-service] port: 8000 [e2e-llm-inference-service] weight: 1 [e2e-llm-inference-service] matches: [e2e-llm-inference-service] - headers: [e2e-llm-inference-service] - name: X-Gateway-Model-Name [e2e-llm-inference-service] type: Exact [e2e-llm-inference-service] value: publishers/kserve-ci-e2e-test/models/facebook/opt-125m [e2e-llm-inference-service] path: [e2e-llm-inference-service] type: Exact [e2e-llm-inference-service] value: /v1/responses [e2e-llm-inference-service] - headers: [e2e-llm-inference-service] - name: X-Gateway-Model-Name [e2e-llm-inference-service] type: Exact [e2e-llm-inference-service] value: publishers/kserve-ci-e2e-test/models/facebook/opt-125m [e2e-llm-inference-service] path: [e2e-llm-inference-service] type: Exact [e2e-llm-inference-service] value: /v1/responses/ [e2e-llm-inference-service] - headers: [e2e-llm-inference-service] - name: X-Gateway-Model-Name [e2e-llm-inference-service] type: Exact [e2e-llm-inference-service] value: publishers/kserve-ci-e2e-test/models/lora-adapter-1 [e2e-llm-inference-service] path: [e2e-llm-inference-service] type: Exact [e2e-llm-inference-service] value: /v1/responses [e2e-llm-inference-service] - headers: [e2e-llm-inference-service] - name: X-Gateway-Model-Name [e2e-llm-inference-service] type: Exact [e2e-llm-inference-service] value: publishers/kserve-ci-e2e-test/models/lora-adapter-1 [e2e-llm-inference-service] path: [e2e-llm-inference-service] type: Exact [e2e-llm-inference-service] value: /v1/responses/ [e2e-llm-inference-service] timeouts: [e2e-llm-inference-service] backendRequest: 0s [e2e-llm-inference-service] request: 0s [e2e-llm-inference-service] - backendRefs: [e2e-llm-inference-service] - group: inference.networking.k8s.io [e2e-llm-inference-service] kind: InferencePool [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-ba4d693a-inference-pool [e2e-llm-inference-service] port: 8000 [e2e-llm-inference-service] weight: 1 [e2e-llm-inference-service] filters: [e2e-llm-inference-service] - type: URLRewrite [e2e-llm-inference-service] urlRewrite: [e2e-llm-inference-service] path: [e2e-llm-inference-service] replacePrefixMatch: /v1/messages [e2e-llm-inference-service] type: ReplacePrefixMatch [e2e-llm-inference-service] matches: [e2e-llm-inference-service] - path: [e2e-llm-inference-service] type: PathPrefix [e2e-llm-inference-service] value: /kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-with-ba4d693a/v1/messages [e2e-llm-inference-service] timeouts: [e2e-llm-inference-service] backendRequest: 0s [e2e-llm-inference-service] request: 0s [e2e-llm-inference-service] - backendRefs: [e2e-llm-inference-service] - group: inference.networking.k8s.io [e2e-llm-inference-service] kind: InferencePool [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-ba4d693a-inference-pool [e2e-llm-inference-service] port: 8000 [e2e-llm-inference-service] weight: 1 [e2e-llm-inference-service] matches: [e2e-llm-inference-service] - headers: [e2e-llm-inference-service] - name: X-Gateway-Model-Name [e2e-llm-inference-service] type: Exact [e2e-llm-inference-service] value: publishers/kserve-ci-e2e-test/models/facebook/opt-125m [e2e-llm-inference-service] path: [e2e-llm-inference-service] type: Exact [e2e-llm-inference-service] value: /v1/messages [e2e-llm-inference-service] - headers: [e2e-llm-inference-service] - name: X-Gateway-Model-Name [e2e-llm-inference-service] type: Exact [e2e-llm-inference-service] value: publishers/kserve-ci-e2e-test/models/facebook/opt-125m [e2e-llm-inference-service] path: [e2e-llm-inference-service] type: Exact [e2e-llm-inference-service] value: /v1/messages/ [e2e-llm-inference-service] - headers: [e2e-llm-inference-service] - name: X-Gateway-Model-Name [e2e-llm-inference-service] type: Exact [e2e-llm-inference-service] value: publishers/kserve-ci-e2e-test/models/lora-adapter-1 [e2e-llm-inference-service] path: [e2e-llm-inference-service] type: Exact [e2e-llm-inference-service] value: /v1/messages [e2e-llm-inference-service] - headers: [e2e-llm-inference-service] - name: X-Gateway-Model-Name [e2e-llm-inference-service] type: Exact [e2e-llm-inference-service] value: publishers/kserve-ci-e2e-test/models/lora-adapter-1 [e2e-llm-inference-service] path: [e2e-llm-inference-service] type: Exact [e2e-llm-inference-service] value: /v1/messages/ [e2e-llm-inference-service] timeouts: [e2e-llm-inference-service] backendRequest: 0s [e2e-llm-inference-service] request: 0s [e2e-llm-inference-service] - backendRefs: [e2e-llm-inference-service] - group: '' [e2e-llm-inference-service] kind: Service [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-ba4d693a-kserve-workload-svc [e2e-llm-inference-service] port: 8000 [e2e-llm-inference-service] weight: 1 [e2e-llm-inference-service] filters: [e2e-llm-inference-service] - type: URLRewrite [e2e-llm-inference-service] urlRewrite: [e2e-llm-inference-service] path: [e2e-llm-inference-service] replacePrefixMatch: / [e2e-llm-inference-service] type: ReplacePrefixMatch [e2e-llm-inference-service] matches: [e2e-llm-inference-service] - path: [e2e-llm-inference-service] type: PathPrefix [e2e-llm-inference-service] value: /kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-with-ba4d693a [e2e-llm-inference-service] timeouts: [e2e-llm-inference-service] backendRequest: 0s [e2e-llm-inference-service] request: 0s [e2e-llm-inference-service] - backendRefs: [e2e-llm-inference-service] - group: '' [e2e-llm-inference-service] kind: Service [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-ba4d693a-kserve-workload-svc [e2e-llm-inference-service] port: 8000 [e2e-llm-inference-service] weight: 1 [e2e-llm-inference-service] matches: [e2e-llm-inference-service] - headers: [e2e-llm-inference-service] - name: X-Gateway-Model-Name [e2e-llm-inference-service] type: Exact [e2e-llm-inference-service] value: publishers/kserve-ci-e2e-test/models/facebook/opt-125m [e2e-llm-inference-service] path: [e2e-llm-inference-service] type: PathPrefix [e2e-llm-inference-service] value: / [e2e-llm-inference-service] - headers: [e2e-llm-inference-service] - name: X-Gateway-Model-Name [e2e-llm-inference-service] type: Exact [e2e-llm-inference-service] value: publishers/kserve-ci-e2e-test/models/lora-adapter-1 [e2e-llm-inference-service] path: [e2e-llm-inference-service] type: PathPrefix [e2e-llm-inference-service] value: / [e2e-llm-inference-service] timeouts: [e2e-llm-inference-service] backendRequest: 0s [e2e-llm-inference-service] request: 0s [e2e-llm-inference-service] status: [e2e-llm-inference-service] parents: [e2e-llm-inference-service] - conditions: [e2e-llm-inference-service] - lastTransitionTime: '2026-07-07T14:45:56Z' [e2e-llm-inference-service] message: Route was valid [e2e-llm-inference-service] observedGeneration: 2 [e2e-llm-inference-service] reason: Accepted [e2e-llm-inference-service] status: 'True' [e2e-llm-inference-service] type: Accepted [e2e-llm-inference-service] - lastTransitionTime: '2026-07-07T14:46:17Z' [e2e-llm-inference-service] message: All references resolved [e2e-llm-inference-service] observedGeneration: 2 [e2e-llm-inference-service] reason: ResolvedRefs [e2e-llm-inference-service] status: 'True' [e2e-llm-inference-service] type: ResolvedRefs [e2e-llm-inference-service] controllerName: openshift.io/gateway-controller/v1 [e2e-llm-inference-service] parentRef: [e2e-llm-inference-service] group: gateway.networking.k8s.io [e2e-llm-inference-service] kind: Gateway [e2e-llm-inference-service] name: openshift-ai-inference [e2e-llm-inference-service] namespace: openshift-ingress [e2e-llm-inference-service] - conditions: [e2e-llm-inference-service] - lastTransitionTime: '2026-07-07T14:45:57Z' [e2e-llm-inference-service] message: Object affected by AuthPolicy [kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-with-ba4d693a-kserve-route-authn [e2e-llm-inference-service] openshift-ingress/openshift-ai-inference-authn] [e2e-llm-inference-service] observedGeneration: 2 [e2e-llm-inference-service] reason: Accepted [e2e-llm-inference-service] status: 'True' [e2e-llm-inference-service] type: kuadrant.io/AuthPolicyAffected [e2e-llm-inference-service] controllerName: kuadrant.io/policy-controller [e2e-llm-inference-service] parentRef: [e2e-llm-inference-service] group: gateway.networking.k8s.io [e2e-llm-inference-service] kind: Gateway [e2e-llm-inference-service] name: openshift-ai-inference [e2e-llm-inference-service] namespace: openshift-ingress [e2e-llm-inference-service] [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1254 --- [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1255 apiVersion: gateway.networking.k8s.io/v1beta1 [e2e-llm-inference-service] kind: HTTPRoute [e2e-llm-inference-service] metadata: [e2e-llm-inference-service] annotations: [e2e-llm-inference-service] serving.kserve.io/inference-pool-migrated: v1 [e2e-llm-inference-service] creationTimestamp: '2026-07-07T14:45:56Z' [e2e-llm-inference-service] generation: 2 [e2e-llm-inference-service] labels: [e2e-llm-inference-service] app.kubernetes.io/component: llminferenceservice-router [e2e-llm-inference-service] app.kubernetes.io/name: llmisvc-model-fb-opt-125m-with-ba4d693a [e2e-llm-inference-service] app.kubernetes.io/part-of: llminferenceservice [e2e-llm-inference-service] managedFields: [e2e-llm-inference-service] - apiVersion: gateway.networking.k8s.io/v1 [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:metadata: [e2e-llm-inference-service] f:annotations: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:serving.kserve.io/inference-pool-migrated: {} [e2e-llm-inference-service] f:labels: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:app.kubernetes.io/component: {} [e2e-llm-inference-service] f:app.kubernetes.io/name: {} [e2e-llm-inference-service] f:app.kubernetes.io/part-of: {} [e2e-llm-inference-service] f:ownerReferences: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"uid":"997d3cef-fabf-4b25-ad95-5ebe7a367fc5"}: {} [e2e-llm-inference-service] f:spec: {} [e2e-llm-inference-service] manager: manager [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] time: '2026-07-07T14:46:16Z' [e2e-llm-inference-service] - apiVersion: gateway.networking.k8s.io/v1 [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:status: [e2e-llm-inference-service] f:parents: {} [e2e-llm-inference-service] manager: manager [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] subresource: status [e2e-llm-inference-service] time: '2026-07-07T14:46:17Z' [e2e-llm-inference-service] - apiVersion: gateway.networking.k8s.io/v1beta1 [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:status: {} [e2e-llm-inference-service] manager: pilot-discovery [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] subresource: status [e2e-llm-inference-service] time: '2026-07-07T14:46:17Z' [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-ba4d693a-kserve-route [e2e-llm-inference-service] namespace: kserve-ci-e2e-test [e2e-llm-inference-service] ownerReferences: [e2e-llm-inference-service] - apiVersion: serving.kserve.io/v1alpha2 [e2e-llm-inference-service] blockOwnerDeletion: true [e2e-llm-inference-service] controller: true [e2e-llm-inference-service] kind: LLMInferenceService [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-ba4d693a [e2e-llm-inference-service] uid: 997d3cef-fabf-4b25-ad95-5ebe7a367fc5 [e2e-llm-inference-service] resourceVersion: '46438' [e2e-llm-inference-service] uid: f21bbfa3-4d7b-4115-8b05-b9d581d11772 [e2e-llm-inference-service] spec: [e2e-llm-inference-service] parentRefs: [e2e-llm-inference-service] - group: gateway.networking.k8s.io [e2e-llm-inference-service] kind: Gateway [e2e-llm-inference-service] name: openshift-ai-inference [e2e-llm-inference-service] namespace: openshift-ingress [e2e-llm-inference-service] rules: [e2e-llm-inference-service] - backendRefs: [e2e-llm-inference-service] - group: inference.networking.k8s.io [e2e-llm-inference-service] kind: InferencePool [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-ba4d693a-inference-pool [e2e-llm-inference-service] port: 8000 [e2e-llm-inference-service] weight: 1 [e2e-llm-inference-service] filters: [e2e-llm-inference-service] - type: URLRewrite [e2e-llm-inference-service] urlRewrite: [e2e-llm-inference-service] path: [e2e-llm-inference-service] replacePrefixMatch: /v1/completions [e2e-llm-inference-service] type: ReplacePrefixMatch [e2e-llm-inference-service] matches: [e2e-llm-inference-service] - path: [e2e-llm-inference-service] type: PathPrefix [e2e-llm-inference-service] value: /kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-with-ba4d693a/v1/completions [e2e-llm-inference-service] timeouts: [e2e-llm-inference-service] backendRequest: 0s [e2e-llm-inference-service] request: 0s [e2e-llm-inference-service] - backendRefs: [e2e-llm-inference-service] - group: inference.networking.k8s.io [e2e-llm-inference-service] kind: InferencePool [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-ba4d693a-inference-pool [e2e-llm-inference-service] port: 8000 [e2e-llm-inference-service] weight: 1 [e2e-llm-inference-service] matches: [e2e-llm-inference-service] - headers: [e2e-llm-inference-service] - name: X-Gateway-Model-Name [e2e-llm-inference-service] type: Exact [e2e-llm-inference-service] value: publishers/kserve-ci-e2e-test/models/facebook/opt-125m [e2e-llm-inference-service] path: [e2e-llm-inference-service] type: Exact [e2e-llm-inference-service] value: /v1/completions [e2e-llm-inference-service] - headers: [e2e-llm-inference-service] - name: X-Gateway-Model-Name [e2e-llm-inference-service] type: Exact [e2e-llm-inference-service] value: publishers/kserve-ci-e2e-test/models/facebook/opt-125m [e2e-llm-inference-service] path: [e2e-llm-inference-service] type: Exact [e2e-llm-inference-service] value: /v1/completions/ [e2e-llm-inference-service] - headers: [e2e-llm-inference-service] - name: X-Gateway-Model-Name [e2e-llm-inference-service] type: Exact [e2e-llm-inference-service] value: publishers/kserve-ci-e2e-test/models/lora-adapter-1 [e2e-llm-inference-service] path: [e2e-llm-inference-service] type: Exact [e2e-llm-inference-service] value: /v1/completions [e2e-llm-inference-service] - headers: [e2e-llm-inference-service] - name: X-Gateway-Model-Name [e2e-llm-inference-service] type: Exact [e2e-llm-inference-service] value: publishers/kserve-ci-e2e-test/models/lora-adapter-1 [e2e-llm-inference-service] path: [e2e-llm-inference-service] type: Exact [e2e-llm-inference-service] value: /v1/completions/ [e2e-llm-inference-service] timeouts: [e2e-llm-inference-service] backendRequest: 0s [e2e-llm-inference-service] request: 0s [e2e-llm-inference-service] - backendRefs: [e2e-llm-inference-service] - group: inference.networking.k8s.io [e2e-llm-inference-service] kind: InferencePool [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-ba4d693a-inference-pool [e2e-llm-inference-service] port: 8000 [e2e-llm-inference-service] weight: 1 [e2e-llm-inference-service] filters: [e2e-llm-inference-service] - type: URLRewrite [e2e-llm-inference-service] urlRewrite: [e2e-llm-inference-service] path: [e2e-llm-inference-service] replacePrefixMatch: /v1/chat/completions [e2e-llm-inference-service] type: ReplacePrefixMatch [e2e-llm-inference-service] matches: [e2e-llm-inference-service] - path: [e2e-llm-inference-service] type: PathPrefix [e2e-llm-inference-service] value: /kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-with-ba4d693a/v1/chat/completions [e2e-llm-inference-service] timeouts: [e2e-llm-inference-service] backendRequest: 0s [e2e-llm-inference-service] request: 0s [e2e-llm-inference-service] - backendRefs: [e2e-llm-inference-service] - group: inference.networking.k8s.io [e2e-llm-inference-service] kind: InferencePool [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-ba4d693a-inference-pool [e2e-llm-inference-service] port: 8000 [e2e-llm-inference-service] weight: 1 [e2e-llm-inference-service] matches: [e2e-llm-inference-service] - headers: [e2e-llm-inference-service] - name: X-Gateway-Model-Name [e2e-llm-inference-service] type: Exact [e2e-llm-inference-service] value: publishers/kserve-ci-e2e-test/models/facebook/opt-125m [e2e-llm-inference-service] path: [e2e-llm-inference-service] type: Exact [e2e-llm-inference-service] value: /v1/chat/completions [e2e-llm-inference-service] - headers: [e2e-llm-inference-service] - name: X-Gateway-Model-Name [e2e-llm-inference-service] type: Exact [e2e-llm-inference-service] value: publishers/kserve-ci-e2e-test/models/facebook/opt-125m [e2e-llm-inference-service] path: [e2e-llm-inference-service] type: Exact [e2e-llm-inference-service] value: /v1/chat/completions/ [e2e-llm-inference-service] - headers: [e2e-llm-inference-service] - name: X-Gateway-Model-Name [e2e-llm-inference-service] type: Exact [e2e-llm-inference-service] value: publishers/kserve-ci-e2e-test/models/lora-adapter-1 [e2e-llm-inference-service] path: [e2e-llm-inference-service] type: Exact [e2e-llm-inference-service] value: /v1/chat/completions [e2e-llm-inference-service] - headers: [e2e-llm-inference-service] - name: X-Gateway-Model-Name [e2e-llm-inference-service] type: Exact [e2e-llm-inference-service] value: publishers/kserve-ci-e2e-test/models/lora-adapter-1 [e2e-llm-inference-service] path: [e2e-llm-inference-service] type: Exact [e2e-llm-inference-service] value: /v1/chat/completions/ [e2e-llm-inference-service] timeouts: [e2e-llm-inference-service] backendRequest: 0s [e2e-llm-inference-service] request: 0s [e2e-llm-inference-service] - backendRefs: [e2e-llm-inference-service] - group: inference.networking.k8s.io [e2e-llm-inference-service] kind: InferencePool [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-ba4d693a-inference-pool [e2e-llm-inference-service] port: 8000 [e2e-llm-inference-service] weight: 1 [e2e-llm-inference-service] filters: [e2e-llm-inference-service] - type: URLRewrite [e2e-llm-inference-service] urlRewrite: [e2e-llm-inference-service] path: [e2e-llm-inference-service] replacePrefixMatch: /v1/responses [e2e-llm-inference-service] type: ReplacePrefixMatch [e2e-llm-inference-service] matches: [e2e-llm-inference-service] - path: [e2e-llm-inference-service] type: PathPrefix [e2e-llm-inference-service] value: /kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-with-ba4d693a/v1/responses [e2e-llm-inference-service] timeouts: [e2e-llm-inference-service] backendRequest: 0s [e2e-llm-inference-service] request: 0s [e2e-llm-inference-service] - backendRefs: [e2e-llm-inference-service] - group: inference.networking.k8s.io [e2e-llm-inference-service] kind: InferencePool [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-ba4d693a-inference-pool [e2e-llm-inference-service] port: 8000 [e2e-llm-inference-service] weight: 1 [e2e-llm-inference-service] matches: [e2e-llm-inference-service] - headers: [e2e-llm-inference-service] - name: X-Gateway-Model-Name [e2e-llm-inference-service] type: Exact [e2e-llm-inference-service] value: publishers/kserve-ci-e2e-test/models/facebook/opt-125m [e2e-llm-inference-service] path: [e2e-llm-inference-service] type: Exact [e2e-llm-inference-service] value: /v1/responses [e2e-llm-inference-service] - headers: [e2e-llm-inference-service] - name: X-Gateway-Model-Name [e2e-llm-inference-service] type: Exact [e2e-llm-inference-service] value: publishers/kserve-ci-e2e-test/models/facebook/opt-125m [e2e-llm-inference-service] path: [e2e-llm-inference-service] type: Exact [e2e-llm-inference-service] value: /v1/responses/ [e2e-llm-inference-service] - headers: [e2e-llm-inference-service] - name: X-Gateway-Model-Name [e2e-llm-inference-service] type: Exact [e2e-llm-inference-service] value: publishers/kserve-ci-e2e-test/models/lora-adapter-1 [e2e-llm-inference-service] path: [e2e-llm-inference-service] type: Exact [e2e-llm-inference-service] value: /v1/responses [e2e-llm-inference-service] - headers: [e2e-llm-inference-service] - name: X-Gateway-Model-Name [e2e-llm-inference-service] type: Exact [e2e-llm-inference-service] value: publishers/kserve-ci-e2e-test/models/lora-adapter-1 [e2e-llm-inference-service] path: [e2e-llm-inference-service] type: Exact [e2e-llm-inference-service] value: /v1/responses/ [e2e-llm-inference-service] timeouts: [e2e-llm-inference-service] backendRequest: 0s [e2e-llm-inference-service] request: 0s [e2e-llm-inference-service] - backendRefs: [e2e-llm-inference-service] - group: inference.networking.k8s.io [e2e-llm-inference-service] kind: InferencePool [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-ba4d693a-inference-pool [e2e-llm-inference-service] port: 8000 [e2e-llm-inference-service] weight: 1 [e2e-llm-inference-service] filters: [e2e-llm-inference-service] - type: URLRewrite [e2e-llm-inference-service] urlRewrite: [e2e-llm-inference-service] path: [e2e-llm-inference-service] replacePrefixMatch: /v1/messages [e2e-llm-inference-service] type: ReplacePrefixMatch [e2e-llm-inference-service] matches: [e2e-llm-inference-service] - path: [e2e-llm-inference-service] type: PathPrefix [e2e-llm-inference-service] value: /kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-with-ba4d693a/v1/messages [e2e-llm-inference-service] timeouts: [e2e-llm-inference-service] backendRequest: 0s [e2e-llm-inference-service] request: 0s [e2e-llm-inference-service] - backendRefs: [e2e-llm-inference-service] - group: inference.networking.k8s.io [e2e-llm-inference-service] kind: InferencePool [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-ba4d693a-inference-pool [e2e-llm-inference-service] port: 8000 [e2e-llm-inference-service] weight: 1 [e2e-llm-inference-service] matches: [e2e-llm-inference-service] - headers: [e2e-llm-inference-service] - name: X-Gateway-Model-Name [e2e-llm-inference-service] type: Exact [e2e-llm-inference-service] value: publishers/kserve-ci-e2e-test/models/facebook/opt-125m [e2e-llm-inference-service] path: [e2e-llm-inference-service] type: Exact [e2e-llm-inference-service] value: /v1/messages [e2e-llm-inference-service] - headers: [e2e-llm-inference-service] - name: X-Gateway-Model-Name [e2e-llm-inference-service] type: Exact [e2e-llm-inference-service] value: publishers/kserve-ci-e2e-test/models/facebook/opt-125m [e2e-llm-inference-service] path: [e2e-llm-inference-service] type: Exact [e2e-llm-inference-service] value: /v1/messages/ [e2e-llm-inference-service] - headers: [e2e-llm-inference-service] - name: X-Gateway-Model-Name [e2e-llm-inference-service] type: Exact [e2e-llm-inference-service] value: publishers/kserve-ci-e2e-test/models/lora-adapter-1 [e2e-llm-inference-service] path: [e2e-llm-inference-service] type: Exact [e2e-llm-inference-service] value: /v1/messages [e2e-llm-inference-service] - headers: [e2e-llm-inference-service] - name: X-Gateway-Model-Name [e2e-llm-inference-service] type: Exact [e2e-llm-inference-service] value: publishers/kserve-ci-e2e-test/models/lora-adapter-1 [e2e-llm-inference-service] path: [e2e-llm-inference-service] type: Exact [e2e-llm-inference-service] value: /v1/messages/ [e2e-llm-inference-service] timeouts: [e2e-llm-inference-service] backendRequest: 0s [e2e-llm-inference-service] request: 0s [e2e-llm-inference-service] - backendRefs: [e2e-llm-inference-service] - group: '' [e2e-llm-inference-service] kind: Service [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-ba4d693a-kserve-workload-svc [e2e-llm-inference-service] port: 8000 [e2e-llm-inference-service] weight: 1 [e2e-llm-inference-service] filters: [e2e-llm-inference-service] - type: URLRewrite [e2e-llm-inference-service] urlRewrite: [e2e-llm-inference-service] path: [e2e-llm-inference-service] replacePrefixMatch: / [e2e-llm-inference-service] type: ReplacePrefixMatch [e2e-llm-inference-service] matches: [e2e-llm-inference-service] - path: [e2e-llm-inference-service] type: PathPrefix [e2e-llm-inference-service] value: /kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-with-ba4d693a [e2e-llm-inference-service] timeouts: [e2e-llm-inference-service] backendRequest: 0s [e2e-llm-inference-service] request: 0s [e2e-llm-inference-service] - backendRefs: [e2e-llm-inference-service] - group: '' [e2e-llm-inference-service] kind: Service [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-ba4d693a-kserve-workload-svc [e2e-llm-inference-service] port: 8000 [e2e-llm-inference-service] weight: 1 [e2e-llm-inference-service] matches: [e2e-llm-inference-service] - headers: [e2e-llm-inference-service] - name: X-Gateway-Model-Name [e2e-llm-inference-service] type: Exact [e2e-llm-inference-service] value: publishers/kserve-ci-e2e-test/models/facebook/opt-125m [e2e-llm-inference-service] path: [e2e-llm-inference-service] type: PathPrefix [e2e-llm-inference-service] value: / [e2e-llm-inference-service] - headers: [e2e-llm-inference-service] - name: X-Gateway-Model-Name [e2e-llm-inference-service] type: Exact [e2e-llm-inference-service] value: publishers/kserve-ci-e2e-test/models/lora-adapter-1 [e2e-llm-inference-service] path: [e2e-llm-inference-service] type: PathPrefix [e2e-llm-inference-service] value: / [e2e-llm-inference-service] timeouts: [e2e-llm-inference-service] backendRequest: 0s [e2e-llm-inference-service] request: 0s [e2e-llm-inference-service] status: [e2e-llm-inference-service] parents: [e2e-llm-inference-service] - conditions: [e2e-llm-inference-service] - lastTransitionTime: '2026-07-07T14:45:56Z' [e2e-llm-inference-service] message: Route was valid [e2e-llm-inference-service] observedGeneration: 2 [e2e-llm-inference-service] reason: Accepted [e2e-llm-inference-service] status: 'True' [e2e-llm-inference-service] type: Accepted [e2e-llm-inference-service] - lastTransitionTime: '2026-07-07T14:46:17Z' [e2e-llm-inference-service] message: All references resolved [e2e-llm-inference-service] observedGeneration: 2 [e2e-llm-inference-service] reason: ResolvedRefs [e2e-llm-inference-service] status: 'True' [e2e-llm-inference-service] type: ResolvedRefs [e2e-llm-inference-service] controllerName: openshift.io/gateway-controller/v1 [e2e-llm-inference-service] parentRef: [e2e-llm-inference-service] group: gateway.networking.k8s.io [e2e-llm-inference-service] kind: Gateway [e2e-llm-inference-service] name: openshift-ai-inference [e2e-llm-inference-service] namespace: openshift-ingress [e2e-llm-inference-service] - conditions: [e2e-llm-inference-service] - lastTransitionTime: '2026-07-07T14:45:57Z' [e2e-llm-inference-service] message: Object affected by AuthPolicy [kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-with-ba4d693a-kserve-route-authn [e2e-llm-inference-service] openshift-ingress/openshift-ai-inference-authn] [e2e-llm-inference-service] observedGeneration: 2 [e2e-llm-inference-service] reason: Accepted [e2e-llm-inference-service] status: 'True' [e2e-llm-inference-service] type: kuadrant.io/AuthPolicyAffected [e2e-llm-inference-service] controllerName: kuadrant.io/policy-controller [e2e-llm-inference-service] parentRef: [e2e-llm-inference-service] group: gateway.networking.k8s.io [e2e-llm-inference-service] kind: Gateway [e2e-llm-inference-service] name: openshift-ai-inference [e2e-llm-inference-service] namespace: openshift-ingress [e2e-llm-inference-service] [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1254 --- [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1255 apiVersion: inference.networking.k8s.io/v1 [e2e-llm-inference-service] kind: InferencePool [e2e-llm-inference-service] metadata: [e2e-llm-inference-service] creationTimestamp: '2026-07-07T14:45:56Z' [e2e-llm-inference-service] generation: 1 [e2e-llm-inference-service] labels: [e2e-llm-inference-service] app.kubernetes.io/component: llminferenceservice-router-scheduler [e2e-llm-inference-service] app.kubernetes.io/name: llmisvc-model-fb-opt-125m-with-ba4d693a [e2e-llm-inference-service] app.kubernetes.io/part-of: llminferenceservice [e2e-llm-inference-service] managedFields: [e2e-llm-inference-service] - apiVersion: inference.networking.k8s.io/v1 [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:metadata: [e2e-llm-inference-service] f:labels: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:app.kubernetes.io/component: {} [e2e-llm-inference-service] f:app.kubernetes.io/name: {} [e2e-llm-inference-service] f:app.kubernetes.io/part-of: {} [e2e-llm-inference-service] f:ownerReferences: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"uid":"997d3cef-fabf-4b25-ad95-5ebe7a367fc5"}: {} [e2e-llm-inference-service] f:spec: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:appProtocol: {} [e2e-llm-inference-service] f:endpointPickerRef: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:failureMode: {} [e2e-llm-inference-service] f:group: {} [e2e-llm-inference-service] f:kind: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:port: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:number: {} [e2e-llm-inference-service] f:selector: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:matchLabels: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:app.kubernetes.io/name: {} [e2e-llm-inference-service] f:app.kubernetes.io/part-of: {} [e2e-llm-inference-service] f:kserve.io/component: {} [e2e-llm-inference-service] f:targetPorts: {} [e2e-llm-inference-service] manager: manager [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] time: '2026-07-07T14:45:56Z' [e2e-llm-inference-service] - apiVersion: inference.networking.k8s.io/v1 [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:status: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:parents: {} [e2e-llm-inference-service] manager: pilot-discovery [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] subresource: status [e2e-llm-inference-service] time: '2026-07-07T14:46:16Z' [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-ba4d693a-inference-pool [e2e-llm-inference-service] namespace: kserve-ci-e2e-test [e2e-llm-inference-service] ownerReferences: [e2e-llm-inference-service] - apiVersion: serving.kserve.io/v1alpha2 [e2e-llm-inference-service] blockOwnerDeletion: true [e2e-llm-inference-service] controller: true [e2e-llm-inference-service] kind: LLMInferenceService [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-ba4d693a [e2e-llm-inference-service] uid: 997d3cef-fabf-4b25-ad95-5ebe7a367fc5 [e2e-llm-inference-service] resourceVersion: '46422' [e2e-llm-inference-service] uid: 180f1d9a-1547-4eaf-9bd7-d4e85723761b [e2e-llm-inference-service] spec: [e2e-llm-inference-service] appProtocol: http [e2e-llm-inference-service] endpointPickerRef: [e2e-llm-inference-service] failureMode: FailOpen [e2e-llm-inference-service] group: '' [e2e-llm-inference-service] kind: Service [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-ba4d693a-epp-service [e2e-llm-inference-service] port: [e2e-llm-inference-service] number: 9002 [e2e-llm-inference-service] selector: [e2e-llm-inference-service] matchLabels: [e2e-llm-inference-service] app.kubernetes.io/name: llmisvc-model-fb-opt-125m-with-ba4d693a [e2e-llm-inference-service] app.kubernetes.io/part-of: llminferenceservice [e2e-llm-inference-service] kserve.io/component: workload [e2e-llm-inference-service] targetPorts: [e2e-llm-inference-service] - number: 8000 [e2e-llm-inference-service] status: [e2e-llm-inference-service] parents: [e2e-llm-inference-service] - conditions: [e2e-llm-inference-service] - lastTransitionTime: '2026-07-07T14:46:16Z' [e2e-llm-inference-service] message: Referenced by an HTTPRoute accepted by the parentRef Gateway [e2e-llm-inference-service] observedGeneration: 1 [e2e-llm-inference-service] reason: Accepted [e2e-llm-inference-service] status: 'True' [e2e-llm-inference-service] type: Accepted [e2e-llm-inference-service] - lastTransitionTime: '2026-07-07T14:46:16Z' [e2e-llm-inference-service] message: Referenced ExtensionRef resolved successfully [e2e-llm-inference-service] observedGeneration: 1 [e2e-llm-inference-service] reason: ResolvedRefs [e2e-llm-inference-service] status: 'True' [e2e-llm-inference-service] type: ResolvedRefs [e2e-llm-inference-service] parentRef: [e2e-llm-inference-service] group: networking.istio.io [e2e-llm-inference-service] kind: Gateway [e2e-llm-inference-service] name: openshift-ai-inference [e2e-llm-inference-service] namespace: openshift-ingress [e2e-llm-inference-service] [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1254 --- [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1255 apiVersion: kuadrant.io/v1 [e2e-llm-inference-service] kind: AuthPolicy [e2e-llm-inference-service] metadata: [e2e-llm-inference-service] creationTimestamp: '2026-07-07T14:45:58Z' [e2e-llm-inference-service] generation: 1 [e2e-llm-inference-service] labels: [e2e-llm-inference-service] app.kubernetes.io/component: llminferenceservice-policies [e2e-llm-inference-service] app.kubernetes.io/managed-by: odh-model-controller [e2e-llm-inference-service] app.kubernetes.io/name: llmisvc-model-fb-opt-125m-with-ba4d693a [e2e-llm-inference-service] app.kubernetes.io/part-of: llminferenceservice [e2e-llm-inference-service] managedFields: [e2e-llm-inference-service] - apiVersion: kuadrant.io/v1 [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:metadata: [e2e-llm-inference-service] f:labels: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:app.kubernetes.io/component: {} [e2e-llm-inference-service] f:app.kubernetes.io/managed-by: {} [e2e-llm-inference-service] f:app.kubernetes.io/name: {} [e2e-llm-inference-service] f:app.kubernetes.io/part-of: {} [e2e-llm-inference-service] f:ownerReferences: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"uid":"997d3cef-fabf-4b25-ad95-5ebe7a367fc5"}: {} [e2e-llm-inference-service] f:spec: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:rules: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:authentication: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:public: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:anonymous: {} [e2e-llm-inference-service] f:credentials: {} [e2e-llm-inference-service] f:metrics: {} [e2e-llm-inference-service] f:overrides: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:fairness: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] f:objective: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] f:priority: {} [e2e-llm-inference-service] f:response: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:success: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:headers: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:x-gateway-inference-fairness-id: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:metrics: {} [e2e-llm-inference-service] f:plain: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:expression: {} [e2e-llm-inference-service] f:priority: {} [e2e-llm-inference-service] f:x-gateway-inference-objective: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:metrics: {} [e2e-llm-inference-service] f:plain: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:expression: {} [e2e-llm-inference-service] f:priority: {} [e2e-llm-inference-service] f:targetRef: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:group: {} [e2e-llm-inference-service] f:kind: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] manager: manager [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] time: '2026-07-07T14:45:58Z' [e2e-llm-inference-service] - apiVersion: kuadrant.io/v1 [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:status: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:conditions: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"type":"Accepted"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:lastTransitionTime: {} [e2e-llm-inference-service] f:message: {} [e2e-llm-inference-service] f:reason: {} [e2e-llm-inference-service] f:status: {} [e2e-llm-inference-service] f:type: {} [e2e-llm-inference-service] k:{"type":"Enforced"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:lastTransitionTime: {} [e2e-llm-inference-service] f:message: {} [e2e-llm-inference-service] f:reason: {} [e2e-llm-inference-service] f:status: {} [e2e-llm-inference-service] f:type: {} [e2e-llm-inference-service] f:observedGeneration: {} [e2e-llm-inference-service] manager: manager [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] subresource: status [e2e-llm-inference-service] time: '2026-07-07T14:46:00Z' [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-ba4d693a-kserve-route-authn [e2e-llm-inference-service] namespace: kserve-ci-e2e-test [e2e-llm-inference-service] ownerReferences: [e2e-llm-inference-service] - apiVersion: serving.kserve.io/v1alpha2 [e2e-llm-inference-service] blockOwnerDeletion: true [e2e-llm-inference-service] controller: true [e2e-llm-inference-service] kind: LLMInferenceService [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-ba4d693a [e2e-llm-inference-service] uid: 997d3cef-fabf-4b25-ad95-5ebe7a367fc5 [e2e-llm-inference-service] resourceVersion: '46150' [e2e-llm-inference-service] uid: 26d4b3f4-dddd-4c8d-9389-5ded634e963a [e2e-llm-inference-service] spec: [e2e-llm-inference-service] rules: [e2e-llm-inference-service] authentication: [e2e-llm-inference-service] public: [e2e-llm-inference-service] anonymous: {} [e2e-llm-inference-service] credentials: {} [e2e-llm-inference-service] metrics: false [e2e-llm-inference-service] overrides: [e2e-llm-inference-service] fairness: [e2e-llm-inference-service] value: unauthenticated [e2e-llm-inference-service] objective: [e2e-llm-inference-service] value: unauthenticated [e2e-llm-inference-service] priority: 0 [e2e-llm-inference-service] response: [e2e-llm-inference-service] success: [e2e-llm-inference-service] headers: [e2e-llm-inference-service] x-gateway-inference-fairness-id: [e2e-llm-inference-service] metrics: false [e2e-llm-inference-service] plain: [e2e-llm-inference-service] expression: auth.identity.fairness [e2e-llm-inference-service] priority: 0 [e2e-llm-inference-service] x-gateway-inference-objective: [e2e-llm-inference-service] metrics: false [e2e-llm-inference-service] plain: [e2e-llm-inference-service] expression: auth.identity.objective [e2e-llm-inference-service] priority: 0 [e2e-llm-inference-service] targetRef: [e2e-llm-inference-service] group: gateway.networking.k8s.io [e2e-llm-inference-service] kind: HTTPRoute [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-ba4d693a-kserve-route [e2e-llm-inference-service] status: [e2e-llm-inference-service] conditions: [e2e-llm-inference-service] - lastTransitionTime: '2026-07-07T14:45:59Z' [e2e-llm-inference-service] message: AuthPolicy has been accepted [e2e-llm-inference-service] reason: Accepted [e2e-llm-inference-service] status: 'True' [e2e-llm-inference-service] type: Accepted [e2e-llm-inference-service] - lastTransitionTime: '2026-07-07T14:46:00Z' [e2e-llm-inference-service] message: AuthPolicy has been successfully enforced [e2e-llm-inference-service] reason: Enforced [e2e-llm-inference-service] status: 'True' [e2e-llm-inference-service] type: Enforced [e2e-llm-inference-service] observedGeneration: 1 [e2e-llm-inference-service] [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1254 --- [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1255 apiVersion: networking.istio.io/v1 [e2e-llm-inference-service] kind: DestinationRule [e2e-llm-inference-service] metadata: [e2e-llm-inference-service] creationTimestamp: '2026-07-07T14:45:57Z' [e2e-llm-inference-service] generation: 1 [e2e-llm-inference-service] labels: [e2e-llm-inference-service] app.kubernetes.io/component: llminferenceservice-router-scheduler [e2e-llm-inference-service] app.kubernetes.io/name: llmisvc-model-fb-opt-125m-with-ba4d693a [e2e-llm-inference-service] app.kubernetes.io/part-of: llminferenceservice [e2e-llm-inference-service] llm-d.ai/managed: 'true' [e2e-llm-inference-service] managedFields: [e2e-llm-inference-service] - apiVersion: networking.istio.io/v1 [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:metadata: [e2e-llm-inference-service] f:labels: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:app.kubernetes.io/component: {} [e2e-llm-inference-service] f:app.kubernetes.io/name: {} [e2e-llm-inference-service] f:app.kubernetes.io/part-of: {} [e2e-llm-inference-service] f:llm-d.ai/managed: {} [e2e-llm-inference-service] f:ownerReferences: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"uid":"997d3cef-fabf-4b25-ad95-5ebe7a367fc5"}: {} [e2e-llm-inference-service] f:spec: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:exportTo: {} [e2e-llm-inference-service] f:host: {} [e2e-llm-inference-service] f:trafficPolicy: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:tls: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:insecureSkipVerify: {} [e2e-llm-inference-service] f:mode: {} [e2e-llm-inference-service] f:sni: {} [e2e-llm-inference-service] manager: manager [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] time: '2026-07-07T14:45:57Z' [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-ba4d693a-kserve-scheduler [e2e-llm-inference-service] namespace: kserve-ci-e2e-test [e2e-llm-inference-service] ownerReferences: [e2e-llm-inference-service] - apiVersion: serving.kserve.io/v1alpha2 [e2e-llm-inference-service] blockOwnerDeletion: true [e2e-llm-inference-service] controller: true [e2e-llm-inference-service] kind: LLMInferenceService [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-ba4d693a [e2e-llm-inference-service] uid: 997d3cef-fabf-4b25-ad95-5ebe7a367fc5 [e2e-llm-inference-service] resourceVersion: '46016' [e2e-llm-inference-service] uid: 50d64387-a88e-4d07-8099-27ab66edd189 [e2e-llm-inference-service] spec: [e2e-llm-inference-service] exportTo: [e2e-llm-inference-service] - '*' [e2e-llm-inference-service] host: llmisvc-model-fb-opt-125m-with-ba4d693a-epp-service.kserve-ci-e2e-test.svc.cluster.local [e2e-llm-inference-service] trafficPolicy: [e2e-llm-inference-service] tls: [e2e-llm-inference-service] insecureSkipVerify: true [e2e-llm-inference-service] mode: SIMPLE [e2e-llm-inference-service] sni: llmisvc-model-fb-opt-125m-with-ba4d693a-epp-service.kserve-ci-e2e-test.svc.cluster.local [e2e-llm-inference-service] [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1254 --- [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1255 apiVersion: networking.istio.io/v1 [e2e-llm-inference-service] kind: DestinationRule [e2e-llm-inference-service] metadata: [e2e-llm-inference-service] creationTimestamp: '2026-07-07T14:46:17Z' [e2e-llm-inference-service] generation: 1 [e2e-llm-inference-service] labels: [e2e-llm-inference-service] app.kubernetes.io/component: llminferenceservice-shadow-service [e2e-llm-inference-service] app.kubernetes.io/name: llmisvc-model-fb-opt-125m-with-ba4d693a [e2e-llm-inference-service] app.kubernetes.io/part-of: llminferenceservice [e2e-llm-inference-service] llm-d.ai/managed: 'true' [e2e-llm-inference-service] managedFields: [e2e-llm-inference-service] - apiVersion: networking.istio.io/v1 [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:metadata: [e2e-llm-inference-service] f:labels: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:app.kubernetes.io/component: {} [e2e-llm-inference-service] f:app.kubernetes.io/name: {} [e2e-llm-inference-service] f:app.kubernetes.io/part-of: {} [e2e-llm-inference-service] f:llm-d.ai/managed: {} [e2e-llm-inference-service] f:ownerReferences: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"uid":"997d3cef-fabf-4b25-ad95-5ebe7a367fc5"}: {} [e2e-llm-inference-service] f:spec: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:exportTo: {} [e2e-llm-inference-service] f:host: {} [e2e-llm-inference-service] f:trafficPolicy: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:tls: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:insecureSkipVerify: {} [e2e-llm-inference-service] f:mode: {} [e2e-llm-inference-service] f:sni: {} [e2e-llm-inference-service] manager: manager [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] time: '2026-07-07T14:46:17Z' [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-ba4d693a-kserve-shadow-svc [e2e-llm-inference-service] namespace: kserve-ci-e2e-test [e2e-llm-inference-service] ownerReferences: [e2e-llm-inference-service] - apiVersion: serving.kserve.io/v1alpha2 [e2e-llm-inference-service] blockOwnerDeletion: true [e2e-llm-inference-service] controller: true [e2e-llm-inference-service] kind: LLMInferenceService [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-ba4d693a [e2e-llm-inference-service] uid: 997d3cef-fabf-4b25-ad95-5ebe7a367fc5 [e2e-llm-inference-service] resourceVersion: '46432' [e2e-llm-inference-service] uid: 19f7f636-ffdd-4340-9a70-055e419f6bf5 [e2e-llm-inference-service] spec: [e2e-llm-inference-service] exportTo: [e2e-llm-inference-service] - '*' [e2e-llm-inference-service] host: llmisvc-model-fb-opt-125m-with-ba4d693a-inference-p-ip-f5162b44.kserve-ci-e2e-test.svc.cluster.local [e2e-llm-inference-service] trafficPolicy: [e2e-llm-inference-service] tls: [e2e-llm-inference-service] insecureSkipVerify: true [e2e-llm-inference-service] mode: SIMPLE [e2e-llm-inference-service] sni: llmisvc-model-fb-opt-125m-with-ba4d693a-kserve-workload-svc.kserve-ci-e2e-test.svc.cluster.local [e2e-llm-inference-service] [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1254 --- [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1255 apiVersion: networking.istio.io/v1 [e2e-llm-inference-service] kind: DestinationRule [e2e-llm-inference-service] metadata: [e2e-llm-inference-service] creationTimestamp: '2026-07-07T14:45:57Z' [e2e-llm-inference-service] generation: 1 [e2e-llm-inference-service] labels: [e2e-llm-inference-service] app.kubernetes.io/component: llminferenceservice-workload [e2e-llm-inference-service] app.kubernetes.io/name: llmisvc-model-fb-opt-125m-with-ba4d693a [e2e-llm-inference-service] app.kubernetes.io/part-of: llminferenceservice [e2e-llm-inference-service] llm-d.ai/managed: 'true' [e2e-llm-inference-service] managedFields: [e2e-llm-inference-service] - apiVersion: networking.istio.io/v1 [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:metadata: [e2e-llm-inference-service] f:labels: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:app.kubernetes.io/component: {} [e2e-llm-inference-service] f:app.kubernetes.io/name: {} [e2e-llm-inference-service] f:app.kubernetes.io/part-of: {} [e2e-llm-inference-service] f:llm-d.ai/managed: {} [e2e-llm-inference-service] f:ownerReferences: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"uid":"997d3cef-fabf-4b25-ad95-5ebe7a367fc5"}: {} [e2e-llm-inference-service] f:spec: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:exportTo: {} [e2e-llm-inference-service] f:host: {} [e2e-llm-inference-service] f:trafficPolicy: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:tls: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:caCertificates: {} [e2e-llm-inference-service] f:insecureSkipVerify: {} [e2e-llm-inference-service] f:mode: {} [e2e-llm-inference-service] f:sni: {} [e2e-llm-inference-service] manager: manager [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] time: '2026-07-07T14:45:57Z' [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-ba4d693a-kserve-workload-svc [e2e-llm-inference-service] namespace: kserve-ci-e2e-test [e2e-llm-inference-service] ownerReferences: [e2e-llm-inference-service] - apiVersion: serving.kserve.io/v1alpha2 [e2e-llm-inference-service] blockOwnerDeletion: true [e2e-llm-inference-service] controller: true [e2e-llm-inference-service] kind: LLMInferenceService [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-ba4d693a [e2e-llm-inference-service] uid: 997d3cef-fabf-4b25-ad95-5ebe7a367fc5 [e2e-llm-inference-service] resourceVersion: '46026' [e2e-llm-inference-service] uid: a335ca74-2dd7-444b-81eb-c4ca2f17d5ca [e2e-llm-inference-service] spec: [e2e-llm-inference-service] exportTo: [e2e-llm-inference-service] - '*' [e2e-llm-inference-service] host: llmisvc-model-fb-opt-125m-with-ba4d693a-kserve-workload-svc.kserve-ci-e2e-test.svc.cluster.local [e2e-llm-inference-service] trafficPolicy: [e2e-llm-inference-service] tls: [e2e-llm-inference-service] caCertificates: /var/run/secrets/kubernetes.io/serviceaccount/service-ca.crt [e2e-llm-inference-service] insecureSkipVerify: false [e2e-llm-inference-service] mode: SIMPLE [e2e-llm-inference-service] sni: llmisvc-model-fb-opt-125m-with-ba4d693a-kserve-workload-svc.kserve-ci-e2e-test.svc.cluster.local [e2e-llm-inference-service] [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1254 --- [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1255 apiVersion: networking.istio.io/v1beta1 [e2e-llm-inference-service] kind: DestinationRule [e2e-llm-inference-service] metadata: [e2e-llm-inference-service] creationTimestamp: '2026-07-07T14:45:57Z' [e2e-llm-inference-service] generation: 1 [e2e-llm-inference-service] labels: [e2e-llm-inference-service] app.kubernetes.io/component: llminferenceservice-router-scheduler [e2e-llm-inference-service] app.kubernetes.io/name: llmisvc-model-fb-opt-125m-with-ba4d693a [e2e-llm-inference-service] app.kubernetes.io/part-of: llminferenceservice [e2e-llm-inference-service] llm-d.ai/managed: 'true' [e2e-llm-inference-service] managedFields: [e2e-llm-inference-service] - apiVersion: networking.istio.io/v1 [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:metadata: [e2e-llm-inference-service] f:labels: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:app.kubernetes.io/component: {} [e2e-llm-inference-service] f:app.kubernetes.io/name: {} [e2e-llm-inference-service] f:app.kubernetes.io/part-of: {} [e2e-llm-inference-service] f:llm-d.ai/managed: {} [e2e-llm-inference-service] f:ownerReferences: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"uid":"997d3cef-fabf-4b25-ad95-5ebe7a367fc5"}: {} [e2e-llm-inference-service] f:spec: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:exportTo: {} [e2e-llm-inference-service] f:host: {} [e2e-llm-inference-service] f:trafficPolicy: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:tls: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:insecureSkipVerify: {} [e2e-llm-inference-service] f:mode: {} [e2e-llm-inference-service] f:sni: {} [e2e-llm-inference-service] manager: manager [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] time: '2026-07-07T14:45:57Z' [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-ba4d693a-kserve-scheduler [e2e-llm-inference-service] namespace: kserve-ci-e2e-test [e2e-llm-inference-service] ownerReferences: [e2e-llm-inference-service] - apiVersion: serving.kserve.io/v1alpha2 [e2e-llm-inference-service] blockOwnerDeletion: true [e2e-llm-inference-service] controller: true [e2e-llm-inference-service] kind: LLMInferenceService [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-ba4d693a [e2e-llm-inference-service] uid: 997d3cef-fabf-4b25-ad95-5ebe7a367fc5 [e2e-llm-inference-service] resourceVersion: '46016' [e2e-llm-inference-service] uid: 50d64387-a88e-4d07-8099-27ab66edd189 [e2e-llm-inference-service] spec: [e2e-llm-inference-service] exportTo: [e2e-llm-inference-service] - '*' [e2e-llm-inference-service] host: llmisvc-model-fb-opt-125m-with-ba4d693a-epp-service.kserve-ci-e2e-test.svc.cluster.local [e2e-llm-inference-service] trafficPolicy: [e2e-llm-inference-service] tls: [e2e-llm-inference-service] insecureSkipVerify: true [e2e-llm-inference-service] mode: SIMPLE [e2e-llm-inference-service] sni: llmisvc-model-fb-opt-125m-with-ba4d693a-epp-service.kserve-ci-e2e-test.svc.cluster.local [e2e-llm-inference-service] [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1254 --- [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1255 apiVersion: networking.istio.io/v1beta1 [e2e-llm-inference-service] kind: DestinationRule [e2e-llm-inference-service] metadata: [e2e-llm-inference-service] creationTimestamp: '2026-07-07T14:46:17Z' [e2e-llm-inference-service] generation: 1 [e2e-llm-inference-service] labels: [e2e-llm-inference-service] app.kubernetes.io/component: llminferenceservice-shadow-service [e2e-llm-inference-service] app.kubernetes.io/name: llmisvc-model-fb-opt-125m-with-ba4d693a [e2e-llm-inference-service] app.kubernetes.io/part-of: llminferenceservice [e2e-llm-inference-service] llm-d.ai/managed: 'true' [e2e-llm-inference-service] managedFields: [e2e-llm-inference-service] - apiVersion: networking.istio.io/v1 [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:metadata: [e2e-llm-inference-service] f:labels: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:app.kubernetes.io/component: {} [e2e-llm-inference-service] f:app.kubernetes.io/name: {} [e2e-llm-inference-service] f:app.kubernetes.io/part-of: {} [e2e-llm-inference-service] f:llm-d.ai/managed: {} [e2e-llm-inference-service] f:ownerReferences: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"uid":"997d3cef-fabf-4b25-ad95-5ebe7a367fc5"}: {} [e2e-llm-inference-service] f:spec: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:exportTo: {} [e2e-llm-inference-service] f:host: {} [e2e-llm-inference-service] f:trafficPolicy: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:tls: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:insecureSkipVerify: {} [e2e-llm-inference-service] f:mode: {} [e2e-llm-inference-service] f:sni: {} [e2e-llm-inference-service] manager: manager [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] time: '2026-07-07T14:46:17Z' [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-ba4d693a-kserve-shadow-svc [e2e-llm-inference-service] namespace: kserve-ci-e2e-test [e2e-llm-inference-service] ownerReferences: [e2e-llm-inference-service] - apiVersion: serving.kserve.io/v1alpha2 [e2e-llm-inference-service] blockOwnerDeletion: true [e2e-llm-inference-service] controller: true [e2e-llm-inference-service] kind: LLMInferenceService [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-ba4d693a [e2e-llm-inference-service] uid: 997d3cef-fabf-4b25-ad95-5ebe7a367fc5 [e2e-llm-inference-service] resourceVersion: '46432' [e2e-llm-inference-service] uid: 19f7f636-ffdd-4340-9a70-055e419f6bf5 [e2e-llm-inference-service] spec: [e2e-llm-inference-service] exportTo: [e2e-llm-inference-service] - '*' [e2e-llm-inference-service] host: llmisvc-model-fb-opt-125m-with-ba4d693a-inference-p-ip-f5162b44.kserve-ci-e2e-test.svc.cluster.local [e2e-llm-inference-service] trafficPolicy: [e2e-llm-inference-service] tls: [e2e-llm-inference-service] insecureSkipVerify: true [e2e-llm-inference-service] mode: SIMPLE [e2e-llm-inference-service] sni: llmisvc-model-fb-opt-125m-with-ba4d693a-kserve-workload-svc.kserve-ci-e2e-test.svc.cluster.local [e2e-llm-inference-service] [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1254 --- [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1255 apiVersion: networking.istio.io/v1beta1 [e2e-llm-inference-service] kind: DestinationRule [e2e-llm-inference-service] metadata: [e2e-llm-inference-service] creationTimestamp: '2026-07-07T14:45:57Z' [e2e-llm-inference-service] generation: 1 [e2e-llm-inference-service] labels: [e2e-llm-inference-service] app.kubernetes.io/component: llminferenceservice-workload [e2e-llm-inference-service] app.kubernetes.io/name: llmisvc-model-fb-opt-125m-with-ba4d693a [e2e-llm-inference-service] app.kubernetes.io/part-of: llminferenceservice [e2e-llm-inference-service] llm-d.ai/managed: 'true' [e2e-llm-inference-service] managedFields: [e2e-llm-inference-service] - apiVersion: networking.istio.io/v1 [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:metadata: [e2e-llm-inference-service] f:labels: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:app.kubernetes.io/component: {} [e2e-llm-inference-service] f:app.kubernetes.io/name: {} [e2e-llm-inference-service] f:app.kubernetes.io/part-of: {} [e2e-llm-inference-service] f:llm-d.ai/managed: {} [e2e-llm-inference-service] f:ownerReferences: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"uid":"997d3cef-fabf-4b25-ad95-5ebe7a367fc5"}: {} [e2e-llm-inference-service] f:spec: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:exportTo: {} [e2e-llm-inference-service] f:host: {} [e2e-llm-inference-service] f:trafficPolicy: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:tls: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:caCertificates: {} [e2e-llm-inference-service] f:insecureSkipVerify: {} [e2e-llm-inference-service] f:mode: {} [e2e-llm-inference-service] f:sni: {} [e2e-llm-inference-service] manager: manager [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] time: '2026-07-07T14:45:57Z' [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-ba4d693a-kserve-workload-svc [e2e-llm-inference-service] namespace: kserve-ci-e2e-test [e2e-llm-inference-service] ownerReferences: [e2e-llm-inference-service] - apiVersion: serving.kserve.io/v1alpha2 [e2e-llm-inference-service] blockOwnerDeletion: true [e2e-llm-inference-service] controller: true [e2e-llm-inference-service] kind: LLMInferenceService [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-ba4d693a [e2e-llm-inference-service] uid: 997d3cef-fabf-4b25-ad95-5ebe7a367fc5 [e2e-llm-inference-service] resourceVersion: '46026' [e2e-llm-inference-service] uid: a335ca74-2dd7-444b-81eb-c4ca2f17d5ca [e2e-llm-inference-service] spec: [e2e-llm-inference-service] exportTo: [e2e-llm-inference-service] - '*' [e2e-llm-inference-service] host: llmisvc-model-fb-opt-125m-with-ba4d693a-kserve-workload-svc.kserve-ci-e2e-test.svc.cluster.local [e2e-llm-inference-service] trafficPolicy: [e2e-llm-inference-service] tls: [e2e-llm-inference-service] caCertificates: /var/run/secrets/kubernetes.io/serviceaccount/service-ca.crt [e2e-llm-inference-service] insecureSkipVerify: false [e2e-llm-inference-service] mode: SIMPLE [e2e-llm-inference-service] sni: llmisvc-model-fb-opt-125m-with-ba4d693a-kserve-workload-svc.kserve-ci-e2e-test.svc.cluster.local [e2e-llm-inference-service] [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1254 --- [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1255 apiVersion: networking.istio.io/v1alpha3 [e2e-llm-inference-service] kind: DestinationRule [e2e-llm-inference-service] metadata: [e2e-llm-inference-service] creationTimestamp: '2026-07-07T14:45:57Z' [e2e-llm-inference-service] generation: 1 [e2e-llm-inference-service] labels: [e2e-llm-inference-service] app.kubernetes.io/component: llminferenceservice-router-scheduler [e2e-llm-inference-service] app.kubernetes.io/name: llmisvc-model-fb-opt-125m-with-ba4d693a [e2e-llm-inference-service] app.kubernetes.io/part-of: llminferenceservice [e2e-llm-inference-service] llm-d.ai/managed: 'true' [e2e-llm-inference-service] managedFields: [e2e-llm-inference-service] - apiVersion: networking.istio.io/v1 [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:metadata: [e2e-llm-inference-service] f:labels: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:app.kubernetes.io/component: {} [e2e-llm-inference-service] f:app.kubernetes.io/name: {} [e2e-llm-inference-service] f:app.kubernetes.io/part-of: {} [e2e-llm-inference-service] f:llm-d.ai/managed: {} [e2e-llm-inference-service] f:ownerReferences: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"uid":"997d3cef-fabf-4b25-ad95-5ebe7a367fc5"}: {} [e2e-llm-inference-service] f:spec: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:exportTo: {} [e2e-llm-inference-service] f:host: {} [e2e-llm-inference-service] f:trafficPolicy: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:tls: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:insecureSkipVerify: {} [e2e-llm-inference-service] f:mode: {} [e2e-llm-inference-service] f:sni: {} [e2e-llm-inference-service] manager: manager [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] time: '2026-07-07T14:45:57Z' [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-ba4d693a-kserve-scheduler [e2e-llm-inference-service] namespace: kserve-ci-e2e-test [e2e-llm-inference-service] ownerReferences: [e2e-llm-inference-service] - apiVersion: serving.kserve.io/v1alpha2 [e2e-llm-inference-service] blockOwnerDeletion: true [e2e-llm-inference-service] controller: true [e2e-llm-inference-service] kind: LLMInferenceService [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-ba4d693a [e2e-llm-inference-service] uid: 997d3cef-fabf-4b25-ad95-5ebe7a367fc5 [e2e-llm-inference-service] resourceVersion: '46016' [e2e-llm-inference-service] uid: 50d64387-a88e-4d07-8099-27ab66edd189 [e2e-llm-inference-service] spec: [e2e-llm-inference-service] exportTo: [e2e-llm-inference-service] - '*' [e2e-llm-inference-service] host: llmisvc-model-fb-opt-125m-with-ba4d693a-epp-service.kserve-ci-e2e-test.svc.cluster.local [e2e-llm-inference-service] trafficPolicy: [e2e-llm-inference-service] tls: [e2e-llm-inference-service] insecureSkipVerify: true [e2e-llm-inference-service] mode: SIMPLE [e2e-llm-inference-service] sni: llmisvc-model-fb-opt-125m-with-ba4d693a-epp-service.kserve-ci-e2e-test.svc.cluster.local [e2e-llm-inference-service] [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1254 --- [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1255 apiVersion: networking.istio.io/v1alpha3 [e2e-llm-inference-service] kind: DestinationRule [e2e-llm-inference-service] metadata: [e2e-llm-inference-service] creationTimestamp: '2026-07-07T14:46:17Z' [e2e-llm-inference-service] generation: 1 [e2e-llm-inference-service] labels: [e2e-llm-inference-service] app.kubernetes.io/component: llminferenceservice-shadow-service [e2e-llm-inference-service] app.kubernetes.io/name: llmisvc-model-fb-opt-125m-with-ba4d693a [e2e-llm-inference-service] app.kubernetes.io/part-of: llminferenceservice [e2e-llm-inference-service] llm-d.ai/managed: 'true' [e2e-llm-inference-service] managedFields: [e2e-llm-inference-service] - apiVersion: networking.istio.io/v1 [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:metadata: [e2e-llm-inference-service] f:labels: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:app.kubernetes.io/component: {} [e2e-llm-inference-service] f:app.kubernetes.io/name: {} [e2e-llm-inference-service] f:app.kubernetes.io/part-of: {} [e2e-llm-inference-service] f:llm-d.ai/managed: {} [e2e-llm-inference-service] f:ownerReferences: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"uid":"997d3cef-fabf-4b25-ad95-5ebe7a367fc5"}: {} [e2e-llm-inference-service] f:spec: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:exportTo: {} [e2e-llm-inference-service] f:host: {} [e2e-llm-inference-service] f:trafficPolicy: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:tls: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:insecureSkipVerify: {} [e2e-llm-inference-service] f:mode: {} [e2e-llm-inference-service] f:sni: {} [e2e-llm-inference-service] manager: manager [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] time: '2026-07-07T14:46:17Z' [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-ba4d693a-kserve-shadow-svc [e2e-llm-inference-service] namespace: kserve-ci-e2e-test [e2e-llm-inference-service] ownerReferences: [e2e-llm-inference-service] - apiVersion: serving.kserve.io/v1alpha2 [e2e-llm-inference-service] blockOwnerDeletion: true [e2e-llm-inference-service] controller: true [e2e-llm-inference-service] kind: LLMInferenceService [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-ba4d693a [e2e-llm-inference-service] uid: 997d3cef-fabf-4b25-ad95-5ebe7a367fc5 [e2e-llm-inference-service] resourceVersion: '46432' [e2e-llm-inference-service] uid: 19f7f636-ffdd-4340-9a70-055e419f6bf5 [e2e-llm-inference-service] spec: [e2e-llm-inference-service] exportTo: [e2e-llm-inference-service] - '*' [e2e-llm-inference-service] host: llmisvc-model-fb-opt-125m-with-ba4d693a-inference-p-ip-f5162b44.kserve-ci-e2e-test.svc.cluster.local [e2e-llm-inference-service] trafficPolicy: [e2e-llm-inference-service] tls: [e2e-llm-inference-service] insecureSkipVerify: true [e2e-llm-inference-service] mode: SIMPLE [e2e-llm-inference-service] sni: llmisvc-model-fb-opt-125m-with-ba4d693a-kserve-workload-svc.kserve-ci-e2e-test.svc.cluster.local [e2e-llm-inference-service] [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1254 --- [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1255 apiVersion: networking.istio.io/v1alpha3 [e2e-llm-inference-service] kind: DestinationRule [e2e-llm-inference-service] metadata: [e2e-llm-inference-service] creationTimestamp: '2026-07-07T14:45:57Z' [e2e-llm-inference-service] generation: 1 [e2e-llm-inference-service] labels: [e2e-llm-inference-service] app.kubernetes.io/component: llminferenceservice-workload [e2e-llm-inference-service] app.kubernetes.io/name: llmisvc-model-fb-opt-125m-with-ba4d693a [e2e-llm-inference-service] app.kubernetes.io/part-of: llminferenceservice [e2e-llm-inference-service] llm-d.ai/managed: 'true' [e2e-llm-inference-service] managedFields: [e2e-llm-inference-service] - apiVersion: networking.istio.io/v1 [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:metadata: [e2e-llm-inference-service] f:labels: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:app.kubernetes.io/component: {} [e2e-llm-inference-service] f:app.kubernetes.io/name: {} [e2e-llm-inference-service] f:app.kubernetes.io/part-of: {} [e2e-llm-inference-service] f:llm-d.ai/managed: {} [e2e-llm-inference-service] f:ownerReferences: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"uid":"997d3cef-fabf-4b25-ad95-5ebe7a367fc5"}: {} [e2e-llm-inference-service] f:spec: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:exportTo: {} [e2e-llm-inference-service] f:host: {} [e2e-llm-inference-service] f:trafficPolicy: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:tls: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:caCertificates: {} [e2e-llm-inference-service] f:insecureSkipVerify: {} [e2e-llm-inference-service] f:mode: {} [e2e-llm-inference-service] f:sni: {} [e2e-llm-inference-service] manager: manager [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] time: '2026-07-07T14:45:57Z' [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-ba4d693a-kserve-workload-svc [e2e-llm-inference-service] namespace: kserve-ci-e2e-test [e2e-llm-inference-service] ownerReferences: [e2e-llm-inference-service] - apiVersion: serving.kserve.io/v1alpha2 [e2e-llm-inference-service] blockOwnerDeletion: true [e2e-llm-inference-service] controller: true [e2e-llm-inference-service] kind: LLMInferenceService [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-ba4d693a [e2e-llm-inference-service] uid: 997d3cef-fabf-4b25-ad95-5ebe7a367fc5 [e2e-llm-inference-service] resourceVersion: '46026' [e2e-llm-inference-service] uid: a335ca74-2dd7-444b-81eb-c4ca2f17d5ca [e2e-llm-inference-service] spec: [e2e-llm-inference-service] exportTo: [e2e-llm-inference-service] - '*' [e2e-llm-inference-service] host: llmisvc-model-fb-opt-125m-with-ba4d693a-kserve-workload-svc.kserve-ci-e2e-test.svc.cluster.local [e2e-llm-inference-service] trafficPolicy: [e2e-llm-inference-service] tls: [e2e-llm-inference-service] caCertificates: /var/run/secrets/kubernetes.io/serviceaccount/service-ca.crt [e2e-llm-inference-service] insecureSkipVerify: false [e2e-llm-inference-service] mode: SIMPLE [e2e-llm-inference-service] sni: llmisvc-model-fb-opt-125m-with-ba4d693a-kserve-workload-svc.kserve-ci-e2e-test.svc.cluster.local [e2e-llm-inference-service] [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1254 --- [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1255 apiVersion: inference.networking.x-k8s.io/v1alpha2 [e2e-llm-inference-service] kind: InferencePool [e2e-llm-inference-service] metadata: [e2e-llm-inference-service] creationTimestamp: '2026-07-07T14:45:56Z' [e2e-llm-inference-service] generation: 1 [e2e-llm-inference-service] labels: [e2e-llm-inference-service] app.kubernetes.io/component: llminferenceservice-router-scheduler [e2e-llm-inference-service] app.kubernetes.io/name: llmisvc-model-fb-opt-125m-with-ba4d693a [e2e-llm-inference-service] app.kubernetes.io/part-of: llminferenceservice [e2e-llm-inference-service] managedFields: [e2e-llm-inference-service] - apiVersion: inference.networking.x-k8s.io/v1alpha2 [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:metadata: [e2e-llm-inference-service] f:labels: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:app.kubernetes.io/component: {} [e2e-llm-inference-service] f:app.kubernetes.io/name: {} [e2e-llm-inference-service] f:app.kubernetes.io/part-of: {} [e2e-llm-inference-service] f:ownerReferences: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"uid":"997d3cef-fabf-4b25-ad95-5ebe7a367fc5"}: {} [e2e-llm-inference-service] f:spec: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:extensionRef: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:failureMode: {} [e2e-llm-inference-service] f:group: {} [e2e-llm-inference-service] f:kind: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:portNumber: {} [e2e-llm-inference-service] f:selector: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:app.kubernetes.io/name: {} [e2e-llm-inference-service] f:app.kubernetes.io/part-of: {} [e2e-llm-inference-service] f:kserve.io/component: {} [e2e-llm-inference-service] f:targetPortNumber: {} [e2e-llm-inference-service] manager: manager [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] time: '2026-07-07T14:45:56Z' [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-ba4d693a-inference-pool [e2e-llm-inference-service] namespace: kserve-ci-e2e-test [e2e-llm-inference-service] ownerReferences: [e2e-llm-inference-service] - apiVersion: serving.kserve.io/v1alpha2 [e2e-llm-inference-service] blockOwnerDeletion: true [e2e-llm-inference-service] controller: true [e2e-llm-inference-service] kind: LLMInferenceService [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-ba4d693a [e2e-llm-inference-service] uid: 997d3cef-fabf-4b25-ad95-5ebe7a367fc5 [e2e-llm-inference-service] resourceVersion: '45990' [e2e-llm-inference-service] uid: 500886f1-ce07-4cf4-87f1-5f390bb11538 [e2e-llm-inference-service] spec: [e2e-llm-inference-service] extensionRef: [e2e-llm-inference-service] failureMode: FailOpen [e2e-llm-inference-service] group: '' [e2e-llm-inference-service] kind: Service [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-ba4d693a-epp-service [e2e-llm-inference-service] portNumber: 9002 [e2e-llm-inference-service] selector: [e2e-llm-inference-service] app.kubernetes.io/name: llmisvc-model-fb-opt-125m-with-ba4d693a [e2e-llm-inference-service] app.kubernetes.io/part-of: llminferenceservice [e2e-llm-inference-service] kserve.io/component: workload [e2e-llm-inference-service] targetPortNumber: 8000 [e2e-llm-inference-service] status: [e2e-llm-inference-service] parent: [e2e-llm-inference-service] - conditions: [e2e-llm-inference-service] - lastTransitionTime: '1970-01-01T00:00:00Z' [e2e-llm-inference-service] message: Waiting for controller [e2e-llm-inference-service] reason: Pending [e2e-llm-inference-service] status: Unknown [e2e-llm-inference-service] type: Accepted [e2e-llm-inference-service] parentRef: [e2e-llm-inference-service] group: gateway.networking.k8s.io [e2e-llm-inference-service] kind: Status [e2e-llm-inference-service] name: default [e2e-llm-inference-service] [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1254 --- [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1255 metadata: [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-ba4d693a-kserve-5b998d5789-t6wf4 [e2e-llm-inference-service] namespace: kserve-ci-e2e-test [e2e-llm-inference-service] creationTimestamp: '2026-07-07T15:02:24Z' [e2e-llm-inference-service] labels: [e2e-llm-inference-service] app.kubernetes.io/component: llminferenceservice-workload [e2e-llm-inference-service] app.kubernetes.io/name: llmisvc-model-fb-opt-125m-with-ba4d693a [e2e-llm-inference-service] app.kubernetes.io/part-of: llminferenceservice [e2e-llm-inference-service] kserve.io/component: workload [e2e-llm-inference-service] llm-d.ai/role: both [e2e-llm-inference-service] pod-template-hash: 5b998d5789 [e2e-llm-inference-service] timestamp: '2026-07-07T15:02:09Z' [e2e-llm-inference-service] window: 16.889s [e2e-llm-inference-service] containers: [e2e-llm-inference-service] - name: main [e2e-llm-inference-service] usage: [e2e-llm-inference-service] cpu: 100082479n [e2e-llm-inference-service] memory: 2320864Ki [e2e-llm-inference-service] apiVersion: metrics.k8s.io/v1beta1 [e2e-llm-inference-service] kind: PodMetrics [e2e-llm-inference-service] [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1254 --- [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1255 metadata: [e2e-llm-inference-service] name: llmisvc-model-fb-opt-125m-with-ba4d693a-kserve-router-sche4ghr9 [e2e-llm-inference-service] namespace: kserve-ci-e2e-test [e2e-llm-inference-service] creationTimestamp: '2026-07-07T15:02:24Z' [e2e-llm-inference-service] labels: [e2e-llm-inference-service] app.kubernetes.io/component: llminferenceservice-router-scheduler [e2e-llm-inference-service] app.kubernetes.io/name: llmisvc-model-fb-opt-125m-with-ba4d693a [e2e-llm-inference-service] app.kubernetes.io/part-of: llminferenceservice [e2e-llm-inference-service] pod-template-hash: 79f5455bb [e2e-llm-inference-service] timestamp: '2026-07-07T15:02:04Z' [e2e-llm-inference-service] window: 16.354s [e2e-llm-inference-service] containers: [e2e-llm-inference-service] - name: main [e2e-llm-inference-service] usage: [e2e-llm-inference-service] cpu: 56666931n [e2e-llm-inference-service] memory: 30400Ki [e2e-llm-inference-service] apiVersion: metrics.k8s.io/v1beta1 [e2e-llm-inference-service] kind: PodMetrics [e2e-llm-inference-service] [e2e-llm-inference-service] ERROR e2e.llmisvc.logging:logging.py:48 [test_llm_inference_service] [2026-07-07T15:02:24.222610] end - ❌ 990.830s: ❌ Failed to call model: HTTPConnectionPool(host='ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com', port=80): Max retries exceeded with url: /v1/models (Caused by ReadTimeoutError("HTTPConnectionPool(host='ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com', port=80): Read timed out. (read timeout=60)")) [e2e-llm-inference-service] _ test_llm_inference_service[router-custom-route-timeout-scheduler-managed-workload-single-cpu-model-fb-opt-125m] _ [e2e-llm-inference-service] [gw0] linux -- Python 3.11.13 /workspace/source/python/kserve/.venv/bin/python [e2e-llm-inference-service] [e2e-llm-inference-service] self = [e2e-llm-inference-service] conn = [e2e-llm-inference-service] method = 'POST' [e2e-llm-inference-service] url = '/kserve-ci-e2e-test/custom-route-timeout-test/v1/completions' [e2e-llm-inference-service] body = b'{"model": "facebook/opt-125m", "prompt": "KServe is a", "max_tokens": 20}' [e2e-llm-inference-service] headers = {'User-Agent': 'python-requests/2.32.3', 'Accept-Encoding': 'gzip, deflate', 'Accept': '*/*', 'Connection': 'keep-alive', 'Content-Type': 'application/json', 'Content-Length': '73'} [e2e-llm-inference-service] retries = Retry(total=0, connect=None, read=None, redirect=None, status=None) [e2e-llm-inference-service] timeout = Timeout(connect=60, read=60, total=None), chunked = False [e2e-llm-inference-service] response_conn = [e2e-llm-inference-service] preload_content = False, decode_content = False, enforce_content_length = True [e2e-llm-inference-service] [e2e-llm-inference-service] def _make_request( [e2e-llm-inference-service] self, [e2e-llm-inference-service] conn: BaseHTTPConnection, [e2e-llm-inference-service] method: str, [e2e-llm-inference-service] url: str, [e2e-llm-inference-service] body: _TYPE_BODY | None = None, [e2e-llm-inference-service] headers: typing.Mapping[str, str] | None = None, [e2e-llm-inference-service] retries: Retry | None = None, [e2e-llm-inference-service] timeout: _TYPE_TIMEOUT = _DEFAULT_TIMEOUT, [e2e-llm-inference-service] chunked: bool = False, [e2e-llm-inference-service] response_conn: BaseHTTPConnection | None = None, [e2e-llm-inference-service] preload_content: bool = True, [e2e-llm-inference-service] decode_content: bool = True, [e2e-llm-inference-service] enforce_content_length: bool = True, [e2e-llm-inference-service] ) -> BaseHTTPResponse: [e2e-llm-inference-service] """ [e2e-llm-inference-service] Perform a request on a given urllib connection object taken from our [e2e-llm-inference-service] pool. [e2e-llm-inference-service] [e2e-llm-inference-service] :param conn: [e2e-llm-inference-service] a connection from one of our connection pools [e2e-llm-inference-service] [e2e-llm-inference-service] :param method: [e2e-llm-inference-service] HTTP request method (such as GET, POST, PUT, etc.) [e2e-llm-inference-service] [e2e-llm-inference-service] :param url: [e2e-llm-inference-service] The URL to perform the request on. [e2e-llm-inference-service] [e2e-llm-inference-service] :param body: [e2e-llm-inference-service] Data to send in the request body, either :class:`str`, :class:`bytes`, [e2e-llm-inference-service] an iterable of :class:`str`/:class:`bytes`, or a file-like object. [e2e-llm-inference-service] [e2e-llm-inference-service] :param headers: [e2e-llm-inference-service] Dictionary of custom headers to send, such as User-Agent, [e2e-llm-inference-service] If-None-Match, etc. If None, pool headers are used. If provided, [e2e-llm-inference-service] these headers completely replace any pool-specific headers. [e2e-llm-inference-service] [e2e-llm-inference-service] :param retries: [e2e-llm-inference-service] Configure the number of retries to allow before raising a [e2e-llm-inference-service] :class:`~urllib3.exceptions.MaxRetryError` exception. [e2e-llm-inference-service] [e2e-llm-inference-service] Pass ``None`` to retry until you receive a response. Pass a [e2e-llm-inference-service] :class:`~urllib3.util.retry.Retry` object for fine-grained control [e2e-llm-inference-service] over different types of retries. [e2e-llm-inference-service] Pass an integer number to retry connection errors that many times, [e2e-llm-inference-service] but no other types of errors. Pass zero to never retry. [e2e-llm-inference-service] [e2e-llm-inference-service] If ``False``, then retries are disabled and any exception is raised [e2e-llm-inference-service] immediately. Also, instead of raising a MaxRetryError on redirects, [e2e-llm-inference-service] the redirect response will be returned. [e2e-llm-inference-service] [e2e-llm-inference-service] :type retries: :class:`~urllib3.util.retry.Retry`, False, or an int. [e2e-llm-inference-service] [e2e-llm-inference-service] :param timeout: [e2e-llm-inference-service] If specified, overrides the default timeout for this one [e2e-llm-inference-service] request. It may be a float (in seconds) or an instance of [e2e-llm-inference-service] :class:`urllib3.util.Timeout`. [e2e-llm-inference-service] [e2e-llm-inference-service] :param chunked: [e2e-llm-inference-service] If True, urllib3 will send the body using chunked transfer [e2e-llm-inference-service] encoding. Otherwise, urllib3 will send the body using the standard [e2e-llm-inference-service] content-length form. Defaults to False. [e2e-llm-inference-service] [e2e-llm-inference-service] :param response_conn: [e2e-llm-inference-service] Set this to ``None`` if you will handle releasing the connection or [e2e-llm-inference-service] set the connection to have the response release it. [e2e-llm-inference-service] [e2e-llm-inference-service] :param preload_content: [e2e-llm-inference-service] If True, the response's body will be preloaded during construction. [e2e-llm-inference-service] [e2e-llm-inference-service] :param decode_content: [e2e-llm-inference-service] If True, will attempt to decode the body based on the [e2e-llm-inference-service] 'content-encoding' header. [e2e-llm-inference-service] [e2e-llm-inference-service] :param enforce_content_length: [e2e-llm-inference-service] Enforce content length checking. Body returned by server must match [e2e-llm-inference-service] value of Content-Length header, if present. Otherwise, raise error. [e2e-llm-inference-service] """ [e2e-llm-inference-service] self.num_requests += 1 [e2e-llm-inference-service] [e2e-llm-inference-service] timeout_obj = self._get_timeout(timeout) [e2e-llm-inference-service] timeout_obj.start_connect() [e2e-llm-inference-service] conn.timeout = Timeout.resolve_default_timeout(timeout_obj.connect_timeout) [e2e-llm-inference-service] [e2e-llm-inference-service] try: [e2e-llm-inference-service] # Trigger any extra validation we need to do. [e2e-llm-inference-service] try: [e2e-llm-inference-service] self._validate_conn(conn) [e2e-llm-inference-service] except (SocketTimeout, BaseSSLError) as e: [e2e-llm-inference-service] self._raise_timeout(err=e, url=url, timeout_value=conn.timeout) [e2e-llm-inference-service] raise [e2e-llm-inference-service] [e2e-llm-inference-service] # _validate_conn() starts the connection to an HTTPS proxy [e2e-llm-inference-service] # so we need to wrap errors with 'ProxyError' here too. [e2e-llm-inference-service] except ( [e2e-llm-inference-service] OSError, [e2e-llm-inference-service] NewConnectionError, [e2e-llm-inference-service] TimeoutError, [e2e-llm-inference-service] BaseSSLError, [e2e-llm-inference-service] CertificateError, [e2e-llm-inference-service] SSLError, [e2e-llm-inference-service] ) as e: [e2e-llm-inference-service] new_e: Exception = e [e2e-llm-inference-service] if isinstance(e, (BaseSSLError, CertificateError)): [e2e-llm-inference-service] new_e = SSLError(e) [e2e-llm-inference-service] # If the connection didn't successfully connect to it's proxy [e2e-llm-inference-service] # then there [e2e-llm-inference-service] if isinstance( [e2e-llm-inference-service] new_e, (OSError, NewConnectionError, TimeoutError, SSLError) [e2e-llm-inference-service] ) and (conn and conn.proxy and not conn.has_connected_to_proxy): [e2e-llm-inference-service] new_e = _wrap_proxy_error(new_e, conn.proxy.scheme) [e2e-llm-inference-service] raise new_e [e2e-llm-inference-service] [e2e-llm-inference-service] # conn.request() calls http.client.*.request, not the method in [e2e-llm-inference-service] # urllib3.request. It also calls makefile (recv) on the socket. [e2e-llm-inference-service] try: [e2e-llm-inference-service] conn.request( [e2e-llm-inference-service] method, [e2e-llm-inference-service] url, [e2e-llm-inference-service] body=body, [e2e-llm-inference-service] headers=headers, [e2e-llm-inference-service] chunked=chunked, [e2e-llm-inference-service] preload_content=preload_content, [e2e-llm-inference-service] decode_content=decode_content, [e2e-llm-inference-service] enforce_content_length=enforce_content_length, [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] # We are swallowing BrokenPipeError (errno.EPIPE) since the server is [e2e-llm-inference-service] # legitimately able to close the connection after sending a valid response. [e2e-llm-inference-service] # With this behaviour, the received response is still readable. [e2e-llm-inference-service] except BrokenPipeError: [e2e-llm-inference-service] pass [e2e-llm-inference-service] except OSError as e: [e2e-llm-inference-service] # MacOS/Linux [e2e-llm-inference-service] # EPROTOTYPE and ECONNRESET are needed on macOS [e2e-llm-inference-service] # https://erickt.github.io/blog/2014/11/19/adventures-in-debugging-a-potential-osx-kernel-bug/ [e2e-llm-inference-service] # Condition changed later to emit ECONNRESET instead of only EPROTOTYPE. [e2e-llm-inference-service] if e.errno != errno.EPROTOTYPE and e.errno != errno.ECONNRESET: [e2e-llm-inference-service] raise [e2e-llm-inference-service] [e2e-llm-inference-service] # Reset the timeout for the recv() on the socket [e2e-llm-inference-service] read_timeout = timeout_obj.read_timeout [e2e-llm-inference-service] [e2e-llm-inference-service] if not conn.is_closed: [e2e-llm-inference-service] # In Python 3 socket.py will catch EAGAIN and return None when you [e2e-llm-inference-service] # try and read into the file pointer created by http.client, which [e2e-llm-inference-service] # instead raises a BadStatusLine exception. Instead of catching [e2e-llm-inference-service] # the exception and assuming all BadStatusLine exceptions are read [e2e-llm-inference-service] # timeouts, check for a zero timeout before making the request. [e2e-llm-inference-service] if read_timeout == 0: [e2e-llm-inference-service] raise ReadTimeoutError( [e2e-llm-inference-service] self, url, f"Read timed out. (read timeout={read_timeout})" [e2e-llm-inference-service] ) [e2e-llm-inference-service] conn.timeout = read_timeout [e2e-llm-inference-service] [e2e-llm-inference-service] # Receive the response from the server [e2e-llm-inference-service] try: [e2e-llm-inference-service] > response = conn.getresponse() [e2e-llm-inference-service] [e2e-llm-inference-service] ../../python/kserve/.venv/lib64/python3.11/site-packages/urllib3/connectionpool.py:534: [e2e-llm-inference-service] _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ [e2e-llm-inference-service] [e2e-llm-inference-service] self = [e2e-llm-inference-service] [e2e-llm-inference-service] def getresponse( # type: ignore[override] [e2e-llm-inference-service] self, [e2e-llm-inference-service] ) -> HTTPResponse: [e2e-llm-inference-service] """ [e2e-llm-inference-service] Get the response from the server. [e2e-llm-inference-service] [e2e-llm-inference-service] If the HTTPConnection is in the correct state, returns an instance of HTTPResponse or of whatever object is returned by the response_class variable. [e2e-llm-inference-service] [e2e-llm-inference-service] If a request has not been sent or if a previous response has not be handled, ResponseNotReady is raised. If the HTTP response indicates that the connection should be closed, then it will be closed before the response is returned. When the connection is closed, the underlying socket is closed. [e2e-llm-inference-service] """ [e2e-llm-inference-service] # Raise the same error as http.client.HTTPConnection [e2e-llm-inference-service] if self._response_options is None: [e2e-llm-inference-service] raise ResponseNotReady() [e2e-llm-inference-service] [e2e-llm-inference-service] # Reset this attribute for being used again. [e2e-llm-inference-service] resp_options = self._response_options [e2e-llm-inference-service] self._response_options = None [e2e-llm-inference-service] [e2e-llm-inference-service] # Since the connection's timeout value may have been updated [e2e-llm-inference-service] # we need to set the timeout on the socket. [e2e-llm-inference-service] self.sock.settimeout(self.timeout) [e2e-llm-inference-service] [e2e-llm-inference-service] # This is needed here to avoid circular import errors [e2e-llm-inference-service] from .response import HTTPResponse [e2e-llm-inference-service] [e2e-llm-inference-service] # Save a reference to the shutdown function before ownership is passed [e2e-llm-inference-service] # to httplib_response [e2e-llm-inference-service] # TODO should we implement it everywhere? [e2e-llm-inference-service] _shutdown = getattr(self.sock, "shutdown", None) [e2e-llm-inference-service] [e2e-llm-inference-service] # Get the response from http.client.HTTPConnection [e2e-llm-inference-service] > httplib_response = super().getresponse() [e2e-llm-inference-service] [e2e-llm-inference-service] ../../python/kserve/.venv/lib64/python3.11/site-packages/urllib3/connection.py:571: [e2e-llm-inference-service] _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ [e2e-llm-inference-service] [e2e-llm-inference-service] self = [e2e-llm-inference-service] [e2e-llm-inference-service] def getresponse(self): [e2e-llm-inference-service] """Get the response from the server. [e2e-llm-inference-service] [e2e-llm-inference-service] If the HTTPConnection is in the correct state, returns an [e2e-llm-inference-service] instance of HTTPResponse or of whatever object is returned by [e2e-llm-inference-service] the response_class variable. [e2e-llm-inference-service] [e2e-llm-inference-service] If a request has not been sent or if a previous response has [e2e-llm-inference-service] not be handled, ResponseNotReady is raised. If the HTTP [e2e-llm-inference-service] response indicates that the connection should be closed, then [e2e-llm-inference-service] it will be closed before the response is returned. When the [e2e-llm-inference-service] connection is closed, the underlying socket is closed. [e2e-llm-inference-service] """ [e2e-llm-inference-service] [e2e-llm-inference-service] # if a prior response has been completed, then forget about it. [e2e-llm-inference-service] if self.__response and self.__response.isclosed(): [e2e-llm-inference-service] self.__response = None [e2e-llm-inference-service] [e2e-llm-inference-service] # if a prior response exists, then it must be completed (otherwise, we [e2e-llm-inference-service] # cannot read this response's header to determine the connection-close [e2e-llm-inference-service] # behavior) [e2e-llm-inference-service] # [e2e-llm-inference-service] # note: if a prior response existed, but was connection-close, then the [e2e-llm-inference-service] # socket and response were made independent of this HTTPConnection [e2e-llm-inference-service] # object since a new request requires that we open a whole new [e2e-llm-inference-service] # connection [e2e-llm-inference-service] # [e2e-llm-inference-service] # this means the prior response had one of two states: [e2e-llm-inference-service] # 1) will_close: this connection was reset and the prior socket and [e2e-llm-inference-service] # response operate independently [e2e-llm-inference-service] # 2) persistent: the response was retained and we await its [e2e-llm-inference-service] # isclosed() status to become true. [e2e-llm-inference-service] # [e2e-llm-inference-service] if self.__state != _CS_REQ_SENT or self.__response: [e2e-llm-inference-service] raise ResponseNotReady(self.__state) [e2e-llm-inference-service] [e2e-llm-inference-service] if self.debuglevel > 0: [e2e-llm-inference-service] response = self.response_class(self.sock, self.debuglevel, [e2e-llm-inference-service] method=self._method) [e2e-llm-inference-service] else: [e2e-llm-inference-service] response = self.response_class(self.sock, method=self._method) [e2e-llm-inference-service] [e2e-llm-inference-service] try: [e2e-llm-inference-service] try: [e2e-llm-inference-service] > response.begin() [e2e-llm-inference-service] [e2e-llm-inference-service] /usr/lib64/python3.11/http/client.py:1395: [e2e-llm-inference-service] _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ [e2e-llm-inference-service] [e2e-llm-inference-service] self = [e2e-llm-inference-service] [e2e-llm-inference-service] def begin(self): [e2e-llm-inference-service] if self.headers is not None: [e2e-llm-inference-service] # we've already started reading the response [e2e-llm-inference-service] return [e2e-llm-inference-service] [e2e-llm-inference-service] # read until we get a non-100 response [e2e-llm-inference-service] while True: [e2e-llm-inference-service] > version, status, reason = self._read_status() [e2e-llm-inference-service] [e2e-llm-inference-service] /usr/lib64/python3.11/http/client.py:325: [e2e-llm-inference-service] _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ [e2e-llm-inference-service] [e2e-llm-inference-service] self = [e2e-llm-inference-service] [e2e-llm-inference-service] def _read_status(self): [e2e-llm-inference-service] > line = str(self.fp.readline(_MAXLINE + 1), "iso-8859-1") [e2e-llm-inference-service] [e2e-llm-inference-service] /usr/lib64/python3.11/http/client.py:286: [e2e-llm-inference-service] _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ [e2e-llm-inference-service] [e2e-llm-inference-service] self = [e2e-llm-inference-service] b = [e2e-llm-inference-service] [e2e-llm-inference-service] def readinto(self, b): [e2e-llm-inference-service] """Read up to len(b) bytes into the writable buffer *b* and return [e2e-llm-inference-service] the number of bytes read. If the socket is non-blocking and no bytes [e2e-llm-inference-service] are available, None is returned. [e2e-llm-inference-service] [e2e-llm-inference-service] If *b* is non-empty, a 0 return value indicates that the connection [e2e-llm-inference-service] was shutdown at the other end. [e2e-llm-inference-service] """ [e2e-llm-inference-service] self._checkClosed() [e2e-llm-inference-service] self._checkReadable() [e2e-llm-inference-service] if self._timeout_occurred: [e2e-llm-inference-service] raise OSError("cannot read from timed out object") [e2e-llm-inference-service] while True: [e2e-llm-inference-service] try: [e2e-llm-inference-service] > return self._sock.recv_into(b) [e2e-llm-inference-service] E TimeoutError: timed out [e2e-llm-inference-service] [e2e-llm-inference-service] /usr/lib64/python3.11/socket.py:718: TimeoutError [e2e-llm-inference-service] [e2e-llm-inference-service] The above exception was the direct cause of the following exception: [e2e-llm-inference-service] [e2e-llm-inference-service] self = [e2e-llm-inference-service] method = 'POST' [e2e-llm-inference-service] url = '/kserve-ci-e2e-test/custom-route-timeout-test/v1/completions' [e2e-llm-inference-service] body = b'{"model": "facebook/opt-125m", "prompt": "KServe is a", "max_tokens": 20}' [e2e-llm-inference-service] headers = {'User-Agent': 'python-requests/2.32.3', 'Accept-Encoding': 'gzip, deflate', 'Accept': '*/*', 'Connection': 'keep-alive', 'Content-Type': 'application/json', 'Content-Length': '73'} [e2e-llm-inference-service] retries = Retry(total=0, connect=None, read=None, redirect=None, status=None) [e2e-llm-inference-service] redirect = False, assert_same_host = False [e2e-llm-inference-service] timeout = Timeout(connect=60, read=60, total=None), pool_timeout = None [e2e-llm-inference-service] release_conn = False, chunked = False, body_pos = None, preload_content = False [e2e-llm-inference-service] decode_content = False, response_kw = {} [e2e-llm-inference-service] parsed_url = Url(scheme=None, auth=None, host=None, port=None, path='/kserve-ci-e2e-test/custom-route-timeout-test/v1/completions', query=None, fragment=None) [e2e-llm-inference-service] destination_scheme = None, conn = None, release_this_conn = True [e2e-llm-inference-service] http_tunnel_required = False, err = None, clean_exit = False [e2e-llm-inference-service] [e2e-llm-inference-service] def urlopen( # type: ignore[override] [e2e-llm-inference-service] self, [e2e-llm-inference-service] method: str, [e2e-llm-inference-service] url: str, [e2e-llm-inference-service] body: _TYPE_BODY | None = None, [e2e-llm-inference-service] headers: typing.Mapping[str, str] | None = None, [e2e-llm-inference-service] retries: Retry | bool | int | None = None, [e2e-llm-inference-service] redirect: bool = True, [e2e-llm-inference-service] assert_same_host: bool = True, [e2e-llm-inference-service] timeout: _TYPE_TIMEOUT = _DEFAULT_TIMEOUT, [e2e-llm-inference-service] pool_timeout: int | None = None, [e2e-llm-inference-service] release_conn: bool | None = None, [e2e-llm-inference-service] chunked: bool = False, [e2e-llm-inference-service] body_pos: _TYPE_BODY_POSITION | None = None, [e2e-llm-inference-service] preload_content: bool = True, [e2e-llm-inference-service] decode_content: bool = True, [e2e-llm-inference-service] **response_kw: typing.Any, [e2e-llm-inference-service] ) -> BaseHTTPResponse: [e2e-llm-inference-service] """ [e2e-llm-inference-service] Get a connection from the pool and perform an HTTP request. This is the [e2e-llm-inference-service] lowest level call for making a request, so you'll need to specify all [e2e-llm-inference-service] the raw details. [e2e-llm-inference-service] [e2e-llm-inference-service] .. note:: [e2e-llm-inference-service] [e2e-llm-inference-service] More commonly, it's appropriate to use a convenience method [e2e-llm-inference-service] such as :meth:`request`. [e2e-llm-inference-service] [e2e-llm-inference-service] .. note:: [e2e-llm-inference-service] [e2e-llm-inference-service] `release_conn` will only behave as expected if [e2e-llm-inference-service] `preload_content=False` because we want to make [e2e-llm-inference-service] `preload_content=False` the default behaviour someday soon without [e2e-llm-inference-service] breaking backwards compatibility. [e2e-llm-inference-service] [e2e-llm-inference-service] :param method: [e2e-llm-inference-service] HTTP request method (such as GET, POST, PUT, etc.) [e2e-llm-inference-service] [e2e-llm-inference-service] :param url: [e2e-llm-inference-service] The URL to perform the request on. [e2e-llm-inference-service] [e2e-llm-inference-service] :param body: [e2e-llm-inference-service] Data to send in the request body, either :class:`str`, :class:`bytes`, [e2e-llm-inference-service] an iterable of :class:`str`/:class:`bytes`, or a file-like object. [e2e-llm-inference-service] [e2e-llm-inference-service] :param headers: [e2e-llm-inference-service] Dictionary of custom headers to send, such as User-Agent, [e2e-llm-inference-service] If-None-Match, etc. If None, pool headers are used. If provided, [e2e-llm-inference-service] these headers completely replace any pool-specific headers. [e2e-llm-inference-service] [e2e-llm-inference-service] :param retries: [e2e-llm-inference-service] Configure the number of retries to allow before raising a [e2e-llm-inference-service] :class:`~urllib3.exceptions.MaxRetryError` exception. [e2e-llm-inference-service] [e2e-llm-inference-service] If ``None`` (default) will retry 3 times, see ``Retry.DEFAULT``. Pass a [e2e-llm-inference-service] :class:`~urllib3.util.retry.Retry` object for fine-grained control [e2e-llm-inference-service] over different types of retries. [e2e-llm-inference-service] Pass an integer number to retry connection errors that many times, [e2e-llm-inference-service] but no other types of errors. Pass zero to never retry. [e2e-llm-inference-service] [e2e-llm-inference-service] If ``False``, then retries are disabled and any exception is raised [e2e-llm-inference-service] immediately. Also, instead of raising a MaxRetryError on redirects, [e2e-llm-inference-service] the redirect response will be returned. [e2e-llm-inference-service] [e2e-llm-inference-service] :type retries: :class:`~urllib3.util.retry.Retry`, False, or an int. [e2e-llm-inference-service] [e2e-llm-inference-service] :param redirect: [e2e-llm-inference-service] If True, automatically handle redirects (status codes 301, 302, [e2e-llm-inference-service] 303, 307, 308). Each redirect counts as a retry. Disabling retries [e2e-llm-inference-service] will disable redirect, too. [e2e-llm-inference-service] [e2e-llm-inference-service] :param assert_same_host: [e2e-llm-inference-service] If ``True``, will make sure that the host of the pool requests is [e2e-llm-inference-service] consistent else will raise HostChangedError. When ``False``, you can [e2e-llm-inference-service] use the pool on an HTTP proxy and request foreign hosts. [e2e-llm-inference-service] [e2e-llm-inference-service] :param timeout: [e2e-llm-inference-service] If specified, overrides the default timeout for this one [e2e-llm-inference-service] request. It may be a float (in seconds) or an instance of [e2e-llm-inference-service] :class:`urllib3.util.Timeout`. [e2e-llm-inference-service] [e2e-llm-inference-service] :param pool_timeout: [e2e-llm-inference-service] If set and the pool is set to block=True, then this method will [e2e-llm-inference-service] block for ``pool_timeout`` seconds and raise EmptyPoolError if no [e2e-llm-inference-service] connection is available within the time period. [e2e-llm-inference-service] [e2e-llm-inference-service] :param bool preload_content: [e2e-llm-inference-service] If True, the response's body will be preloaded into memory. [e2e-llm-inference-service] [e2e-llm-inference-service] :param bool decode_content: [e2e-llm-inference-service] If True, will attempt to decode the body based on the [e2e-llm-inference-service] 'content-encoding' header. [e2e-llm-inference-service] [e2e-llm-inference-service] :param release_conn: [e2e-llm-inference-service] If False, then the urlopen call will not release the connection [e2e-llm-inference-service] back into the pool once a response is received (but will release if [e2e-llm-inference-service] you read the entire contents of the response such as when [e2e-llm-inference-service] `preload_content=True`). This is useful if you're not preloading [e2e-llm-inference-service] the response's content immediately. You will need to call [e2e-llm-inference-service] ``r.release_conn()`` on the response ``r`` to return the connection [e2e-llm-inference-service] back into the pool. If None, it takes the value of ``preload_content`` [e2e-llm-inference-service] which defaults to ``True``. [e2e-llm-inference-service] [e2e-llm-inference-service] :param bool chunked: [e2e-llm-inference-service] If True, urllib3 will send the body using chunked transfer [e2e-llm-inference-service] encoding. Otherwise, urllib3 will send the body using the standard [e2e-llm-inference-service] content-length form. Defaults to False. [e2e-llm-inference-service] [e2e-llm-inference-service] :param int body_pos: [e2e-llm-inference-service] Position to seek to in file-like body in the event of a retry or [e2e-llm-inference-service] redirect. Typically this won't need to be set because urllib3 will [e2e-llm-inference-service] auto-populate the value when needed. [e2e-llm-inference-service] """ [e2e-llm-inference-service] parsed_url = parse_url(url) [e2e-llm-inference-service] destination_scheme = parsed_url.scheme [e2e-llm-inference-service] [e2e-llm-inference-service] if headers is None: [e2e-llm-inference-service] headers = self.headers [e2e-llm-inference-service] [e2e-llm-inference-service] if not isinstance(retries, Retry): [e2e-llm-inference-service] retries = Retry.from_int(retries, redirect=redirect, default=self.retries) [e2e-llm-inference-service] [e2e-llm-inference-service] if release_conn is None: [e2e-llm-inference-service] release_conn = preload_content [e2e-llm-inference-service] [e2e-llm-inference-service] # Check host [e2e-llm-inference-service] if assert_same_host and not self.is_same_host(url): [e2e-llm-inference-service] raise HostChangedError(self, url, retries) [e2e-llm-inference-service] [e2e-llm-inference-service] # Ensure that the URL we're connecting to is properly encoded [e2e-llm-inference-service] if url.startswith("/"): [e2e-llm-inference-service] url = to_str(_encode_target(url)) [e2e-llm-inference-service] else: [e2e-llm-inference-service] url = to_str(parsed_url.url) [e2e-llm-inference-service] [e2e-llm-inference-service] conn = None [e2e-llm-inference-service] [e2e-llm-inference-service] # Track whether `conn` needs to be released before [e2e-llm-inference-service] # returning/raising/recursing. Update this variable if necessary, and [e2e-llm-inference-service] # leave `release_conn` constant throughout the function. That way, if [e2e-llm-inference-service] # the function recurses, the original value of `release_conn` will be [e2e-llm-inference-service] # passed down into the recursive call, and its value will be respected. [e2e-llm-inference-service] # [e2e-llm-inference-service] # See issue #651 [1] for details. [e2e-llm-inference-service] # [e2e-llm-inference-service] # [1] [e2e-llm-inference-service] release_this_conn = release_conn [e2e-llm-inference-service] [e2e-llm-inference-service] http_tunnel_required = connection_requires_http_tunnel( [e2e-llm-inference-service] self.proxy, self.proxy_config, destination_scheme [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] # Merge the proxy headers. Only done when not using HTTP CONNECT. We [e2e-llm-inference-service] # have to copy the headers dict so we can safely change it without those [e2e-llm-inference-service] # changes being reflected in anyone else's copy. [e2e-llm-inference-service] if not http_tunnel_required: [e2e-llm-inference-service] headers = headers.copy() # type: ignore[attr-defined] [e2e-llm-inference-service] headers.update(self.proxy_headers) # type: ignore[union-attr] [e2e-llm-inference-service] [e2e-llm-inference-service] # Must keep the exception bound to a separate variable or else Python 3 [e2e-llm-inference-service] # complains about UnboundLocalError. [e2e-llm-inference-service] err = None [e2e-llm-inference-service] [e2e-llm-inference-service] # Keep track of whether we cleanly exited the except block. This [e2e-llm-inference-service] # ensures we do proper cleanup in finally. [e2e-llm-inference-service] clean_exit = False [e2e-llm-inference-service] [e2e-llm-inference-service] # Rewind body position, if needed. Record current position [e2e-llm-inference-service] # for future rewinds in the event of a redirect/retry. [e2e-llm-inference-service] body_pos = set_file_position(body, body_pos) [e2e-llm-inference-service] [e2e-llm-inference-service] try: [e2e-llm-inference-service] # Request a connection from the queue. [e2e-llm-inference-service] timeout_obj = self._get_timeout(timeout) [e2e-llm-inference-service] conn = self._get_conn(timeout=pool_timeout) [e2e-llm-inference-service] [e2e-llm-inference-service] conn.timeout = timeout_obj.connect_timeout # type: ignore[assignment] [e2e-llm-inference-service] [e2e-llm-inference-service] # Is this a closed/new connection that requires CONNECT tunnelling? [e2e-llm-inference-service] if self.proxy is not None and http_tunnel_required and conn.is_closed: [e2e-llm-inference-service] try: [e2e-llm-inference-service] self._prepare_proxy(conn) [e2e-llm-inference-service] except (BaseSSLError, OSError, SocketTimeout) as e: [e2e-llm-inference-service] self._raise_timeout( [e2e-llm-inference-service] err=e, url=self.proxy.url, timeout_value=conn.timeout [e2e-llm-inference-service] ) [e2e-llm-inference-service] raise [e2e-llm-inference-service] [e2e-llm-inference-service] # If we're going to release the connection in ``finally:``, then [e2e-llm-inference-service] # the response doesn't need to know about the connection. Otherwise [e2e-llm-inference-service] # it will also try to release it and we'll have a double-release [e2e-llm-inference-service] # mess. [e2e-llm-inference-service] response_conn = conn if not release_conn else None [e2e-llm-inference-service] [e2e-llm-inference-service] # Make the request on the HTTPConnection object [e2e-llm-inference-service] > response = self._make_request( [e2e-llm-inference-service] conn, [e2e-llm-inference-service] method, [e2e-llm-inference-service] url, [e2e-llm-inference-service] timeout=timeout_obj, [e2e-llm-inference-service] body=body, [e2e-llm-inference-service] headers=headers, [e2e-llm-inference-service] chunked=chunked, [e2e-llm-inference-service] retries=retries, [e2e-llm-inference-service] response_conn=response_conn, [e2e-llm-inference-service] preload_content=preload_content, [e2e-llm-inference-service] decode_content=decode_content, [e2e-llm-inference-service] **response_kw, [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] ../../python/kserve/.venv/lib64/python3.11/site-packages/urllib3/connectionpool.py:787: [e2e-llm-inference-service] _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ [e2e-llm-inference-service] [e2e-llm-inference-service] self = [e2e-llm-inference-service] conn = [e2e-llm-inference-service] method = 'POST' [e2e-llm-inference-service] url = '/kserve-ci-e2e-test/custom-route-timeout-test/v1/completions' [e2e-llm-inference-service] body = b'{"model": "facebook/opt-125m", "prompt": "KServe is a", "max_tokens": 20}' [e2e-llm-inference-service] headers = {'User-Agent': 'python-requests/2.32.3', 'Accept-Encoding': 'gzip, deflate', 'Accept': '*/*', 'Connection': 'keep-alive', 'Content-Type': 'application/json', 'Content-Length': '73'} [e2e-llm-inference-service] retries = Retry(total=0, connect=None, read=None, redirect=None, status=None) [e2e-llm-inference-service] timeout = Timeout(connect=60, read=60, total=None), chunked = False [e2e-llm-inference-service] response_conn = [e2e-llm-inference-service] preload_content = False, decode_content = False, enforce_content_length = True [e2e-llm-inference-service] [e2e-llm-inference-service] def _make_request( [e2e-llm-inference-service] self, [e2e-llm-inference-service] conn: BaseHTTPConnection, [e2e-llm-inference-service] method: str, [e2e-llm-inference-service] url: str, [e2e-llm-inference-service] body: _TYPE_BODY | None = None, [e2e-llm-inference-service] headers: typing.Mapping[str, str] | None = None, [e2e-llm-inference-service] retries: Retry | None = None, [e2e-llm-inference-service] timeout: _TYPE_TIMEOUT = _DEFAULT_TIMEOUT, [e2e-llm-inference-service] chunked: bool = False, [e2e-llm-inference-service] response_conn: BaseHTTPConnection | None = None, [e2e-llm-inference-service] preload_content: bool = True, [e2e-llm-inference-service] decode_content: bool = True, [e2e-llm-inference-service] enforce_content_length: bool = True, [e2e-llm-inference-service] ) -> BaseHTTPResponse: [e2e-llm-inference-service] """ [e2e-llm-inference-service] Perform a request on a given urllib connection object taken from our [e2e-llm-inference-service] pool. [e2e-llm-inference-service] [e2e-llm-inference-service] :param conn: [e2e-llm-inference-service] a connection from one of our connection pools [e2e-llm-inference-service] [e2e-llm-inference-service] :param method: [e2e-llm-inference-service] HTTP request method (such as GET, POST, PUT, etc.) [e2e-llm-inference-service] [e2e-llm-inference-service] :param url: [e2e-llm-inference-service] The URL to perform the request on. [e2e-llm-inference-service] [e2e-llm-inference-service] :param body: [e2e-llm-inference-service] Data to send in the request body, either :class:`str`, :class:`bytes`, [e2e-llm-inference-service] an iterable of :class:`str`/:class:`bytes`, or a file-like object. [e2e-llm-inference-service] [e2e-llm-inference-service] :param headers: [e2e-llm-inference-service] Dictionary of custom headers to send, such as User-Agent, [e2e-llm-inference-service] If-None-Match, etc. If None, pool headers are used. If provided, [e2e-llm-inference-service] these headers completely replace any pool-specific headers. [e2e-llm-inference-service] [e2e-llm-inference-service] :param retries: [e2e-llm-inference-service] Configure the number of retries to allow before raising a [e2e-llm-inference-service] :class:`~urllib3.exceptions.MaxRetryError` exception. [e2e-llm-inference-service] [e2e-llm-inference-service] Pass ``None`` to retry until you receive a response. Pass a [e2e-llm-inference-service] :class:`~urllib3.util.retry.Retry` object for fine-grained control [e2e-llm-inference-service] over different types of retries. [e2e-llm-inference-service] Pass an integer number to retry connection errors that many times, [e2e-llm-inference-service] but no other types of errors. Pass zero to never retry. [e2e-llm-inference-service] [e2e-llm-inference-service] If ``False``, then retries are disabled and any exception is raised [e2e-llm-inference-service] immediately. Also, instead of raising a MaxRetryError on redirects, [e2e-llm-inference-service] the redirect response will be returned. [e2e-llm-inference-service] [e2e-llm-inference-service] :type retries: :class:`~urllib3.util.retry.Retry`, False, or an int. [e2e-llm-inference-service] [e2e-llm-inference-service] :param timeout: [e2e-llm-inference-service] If specified, overrides the default timeout for this one [e2e-llm-inference-service] request. It may be a float (in seconds) or an instance of [e2e-llm-inference-service] :class:`urllib3.util.Timeout`. [e2e-llm-inference-service] [e2e-llm-inference-service] :param chunked: [e2e-llm-inference-service] If True, urllib3 will send the body using chunked transfer [e2e-llm-inference-service] encoding. Otherwise, urllib3 will send the body using the standard [e2e-llm-inference-service] content-length form. Defaults to False. [e2e-llm-inference-service] [e2e-llm-inference-service] :param response_conn: [e2e-llm-inference-service] Set this to ``None`` if you will handle releasing the connection or [e2e-llm-inference-service] set the connection to have the response release it. [e2e-llm-inference-service] [e2e-llm-inference-service] :param preload_content: [e2e-llm-inference-service] If True, the response's body will be preloaded during construction. [e2e-llm-inference-service] [e2e-llm-inference-service] :param decode_content: [e2e-llm-inference-service] If True, will attempt to decode the body based on the [e2e-llm-inference-service] 'content-encoding' header. [e2e-llm-inference-service] [e2e-llm-inference-service] :param enforce_content_length: [e2e-llm-inference-service] Enforce content length checking. Body returned by server must match [e2e-llm-inference-service] value of Content-Length header, if present. Otherwise, raise error. [e2e-llm-inference-service] """ [e2e-llm-inference-service] self.num_requests += 1 [e2e-llm-inference-service] [e2e-llm-inference-service] timeout_obj = self._get_timeout(timeout) [e2e-llm-inference-service] timeout_obj.start_connect() [e2e-llm-inference-service] conn.timeout = Timeout.resolve_default_timeout(timeout_obj.connect_timeout) [e2e-llm-inference-service] [e2e-llm-inference-service] try: [e2e-llm-inference-service] # Trigger any extra validation we need to do. [e2e-llm-inference-service] try: [e2e-llm-inference-service] self._validate_conn(conn) [e2e-llm-inference-service] except (SocketTimeout, BaseSSLError) as e: [e2e-llm-inference-service] self._raise_timeout(err=e, url=url, timeout_value=conn.timeout) [e2e-llm-inference-service] raise [e2e-llm-inference-service] [e2e-llm-inference-service] # _validate_conn() starts the connection to an HTTPS proxy [e2e-llm-inference-service] # so we need to wrap errors with 'ProxyError' here too. [e2e-llm-inference-service] except ( [e2e-llm-inference-service] OSError, [e2e-llm-inference-service] NewConnectionError, [e2e-llm-inference-service] TimeoutError, [e2e-llm-inference-service] BaseSSLError, [e2e-llm-inference-service] CertificateError, [e2e-llm-inference-service] SSLError, [e2e-llm-inference-service] ) as e: [e2e-llm-inference-service] new_e: Exception = e [e2e-llm-inference-service] if isinstance(e, (BaseSSLError, CertificateError)): [e2e-llm-inference-service] new_e = SSLError(e) [e2e-llm-inference-service] # If the connection didn't successfully connect to it's proxy [e2e-llm-inference-service] # then there [e2e-llm-inference-service] if isinstance( [e2e-llm-inference-service] new_e, (OSError, NewConnectionError, TimeoutError, SSLError) [e2e-llm-inference-service] ) and (conn and conn.proxy and not conn.has_connected_to_proxy): [e2e-llm-inference-service] new_e = _wrap_proxy_error(new_e, conn.proxy.scheme) [e2e-llm-inference-service] raise new_e [e2e-llm-inference-service] [e2e-llm-inference-service] # conn.request() calls http.client.*.request, not the method in [e2e-llm-inference-service] # urllib3.request. It also calls makefile (recv) on the socket. [e2e-llm-inference-service] try: [e2e-llm-inference-service] conn.request( [e2e-llm-inference-service] method, [e2e-llm-inference-service] url, [e2e-llm-inference-service] body=body, [e2e-llm-inference-service] headers=headers, [e2e-llm-inference-service] chunked=chunked, [e2e-llm-inference-service] preload_content=preload_content, [e2e-llm-inference-service] decode_content=decode_content, [e2e-llm-inference-service] enforce_content_length=enforce_content_length, [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] # We are swallowing BrokenPipeError (errno.EPIPE) since the server is [e2e-llm-inference-service] # legitimately able to close the connection after sending a valid response. [e2e-llm-inference-service] # With this behaviour, the received response is still readable. [e2e-llm-inference-service] except BrokenPipeError: [e2e-llm-inference-service] pass [e2e-llm-inference-service] except OSError as e: [e2e-llm-inference-service] # MacOS/Linux [e2e-llm-inference-service] # EPROTOTYPE and ECONNRESET are needed on macOS [e2e-llm-inference-service] # https://erickt.github.io/blog/2014/11/19/adventures-in-debugging-a-potential-osx-kernel-bug/ [e2e-llm-inference-service] # Condition changed later to emit ECONNRESET instead of only EPROTOTYPE. [e2e-llm-inference-service] if e.errno != errno.EPROTOTYPE and e.errno != errno.ECONNRESET: [e2e-llm-inference-service] raise [e2e-llm-inference-service] [e2e-llm-inference-service] # Reset the timeout for the recv() on the socket [e2e-llm-inference-service] read_timeout = timeout_obj.read_timeout [e2e-llm-inference-service] [e2e-llm-inference-service] if not conn.is_closed: [e2e-llm-inference-service] # In Python 3 socket.py will catch EAGAIN and return None when you [e2e-llm-inference-service] # try and read into the file pointer created by http.client, which [e2e-llm-inference-service] # instead raises a BadStatusLine exception. Instead of catching [e2e-llm-inference-service] # the exception and assuming all BadStatusLine exceptions are read [e2e-llm-inference-service] # timeouts, check for a zero timeout before making the request. [e2e-llm-inference-service] if read_timeout == 0: [e2e-llm-inference-service] raise ReadTimeoutError( [e2e-llm-inference-service] self, url, f"Read timed out. (read timeout={read_timeout})" [e2e-llm-inference-service] ) [e2e-llm-inference-service] conn.timeout = read_timeout [e2e-llm-inference-service] [e2e-llm-inference-service] # Receive the response from the server [e2e-llm-inference-service] try: [e2e-llm-inference-service] response = conn.getresponse() [e2e-llm-inference-service] except (BaseSSLError, OSError) as e: [e2e-llm-inference-service] > self._raise_timeout(err=e, url=url, timeout_value=read_timeout) [e2e-llm-inference-service] [e2e-llm-inference-service] ../../python/kserve/.venv/lib64/python3.11/site-packages/urllib3/connectionpool.py:536: [e2e-llm-inference-service] _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ [e2e-llm-inference-service] [e2e-llm-inference-service] self = [e2e-llm-inference-service] err = TimeoutError('timed out') [e2e-llm-inference-service] url = '/kserve-ci-e2e-test/custom-route-timeout-test/v1/completions' [e2e-llm-inference-service] timeout_value = 60 [e2e-llm-inference-service] [e2e-llm-inference-service] def _raise_timeout( [e2e-llm-inference-service] self, [e2e-llm-inference-service] err: BaseSSLError | OSError | SocketTimeout, [e2e-llm-inference-service] url: str, [e2e-llm-inference-service] timeout_value: _TYPE_TIMEOUT | None, [e2e-llm-inference-service] ) -> None: [e2e-llm-inference-service] """Is the error actually a timeout? Will raise a ReadTimeout or pass""" [e2e-llm-inference-service] [e2e-llm-inference-service] if isinstance(err, SocketTimeout): [e2e-llm-inference-service] > raise ReadTimeoutError( [e2e-llm-inference-service] self, url, f"Read timed out. (read timeout={timeout_value})" [e2e-llm-inference-service] ) from err [e2e-llm-inference-service] E urllib3.exceptions.ReadTimeoutError: HTTPConnectionPool(host='ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com', port=80): Read timed out. (read timeout=60) [e2e-llm-inference-service] [e2e-llm-inference-service] ../../python/kserve/.venv/lib64/python3.11/site-packages/urllib3/connectionpool.py:367: ReadTimeoutError [e2e-llm-inference-service] [e2e-llm-inference-service] The above exception was the direct cause of the following exception: [e2e-llm-inference-service] [e2e-llm-inference-service] self = [e2e-llm-inference-service] request = , stream = False [e2e-llm-inference-service] timeout = Timeout(connect=60, read=60, total=None), verify = '/tmp/ca.crt' [e2e-llm-inference-service] cert = None, proxies = OrderedDict() [e2e-llm-inference-service] [e2e-llm-inference-service] def send( [e2e-llm-inference-service] self, request, stream=False, timeout=None, verify=True, cert=None, proxies=None [e2e-llm-inference-service] ): [e2e-llm-inference-service] """Sends PreparedRequest object. Returns Response object. [e2e-llm-inference-service] [e2e-llm-inference-service] :param request: The :class:`PreparedRequest ` being sent. [e2e-llm-inference-service] :param stream: (optional) Whether to stream the request content. [e2e-llm-inference-service] :param timeout: (optional) How long to wait for the server to send [e2e-llm-inference-service] data before giving up, as a float, or a :ref:`(connect timeout, [e2e-llm-inference-service] read timeout) ` tuple. [e2e-llm-inference-service] :type timeout: float or tuple or urllib3 Timeout object [e2e-llm-inference-service] :param verify: (optional) Either a boolean, in which case it controls whether [e2e-llm-inference-service] we verify the server's TLS certificate, or a string, in which case it [e2e-llm-inference-service] must be a path to a CA bundle to use [e2e-llm-inference-service] :param cert: (optional) Any user-provided SSL certificate to be trusted. [e2e-llm-inference-service] :param proxies: (optional) The proxies dictionary to apply to the request. [e2e-llm-inference-service] :rtype: requests.Response [e2e-llm-inference-service] """ [e2e-llm-inference-service] [e2e-llm-inference-service] try: [e2e-llm-inference-service] conn = self.get_connection_with_tls_context( [e2e-llm-inference-service] request, verify, proxies=proxies, cert=cert [e2e-llm-inference-service] ) [e2e-llm-inference-service] except LocationValueError as e: [e2e-llm-inference-service] raise InvalidURL(e, request=request) [e2e-llm-inference-service] [e2e-llm-inference-service] self.cert_verify(conn, request.url, verify, cert) [e2e-llm-inference-service] url = self.request_url(request, proxies) [e2e-llm-inference-service] self.add_headers( [e2e-llm-inference-service] request, [e2e-llm-inference-service] stream=stream, [e2e-llm-inference-service] timeout=timeout, [e2e-llm-inference-service] verify=verify, [e2e-llm-inference-service] cert=cert, [e2e-llm-inference-service] proxies=proxies, [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] chunked = not (request.body is None or "Content-Length" in request.headers) [e2e-llm-inference-service] [e2e-llm-inference-service] if isinstance(timeout, tuple): [e2e-llm-inference-service] try: [e2e-llm-inference-service] connect, read = timeout [e2e-llm-inference-service] timeout = TimeoutSauce(connect=connect, read=read) [e2e-llm-inference-service] except ValueError: [e2e-llm-inference-service] raise ValueError( [e2e-llm-inference-service] f"Invalid timeout {timeout}. Pass a (connect, read) timeout tuple, " [e2e-llm-inference-service] f"or a single float to set both timeouts to the same value." [e2e-llm-inference-service] ) [e2e-llm-inference-service] elif isinstance(timeout, TimeoutSauce): [e2e-llm-inference-service] pass [e2e-llm-inference-service] else: [e2e-llm-inference-service] timeout = TimeoutSauce(connect=timeout, read=timeout) [e2e-llm-inference-service] [e2e-llm-inference-service] try: [e2e-llm-inference-service] > resp = conn.urlopen( [e2e-llm-inference-service] method=request.method, [e2e-llm-inference-service] url=url, [e2e-llm-inference-service] body=request.body, [e2e-llm-inference-service] headers=request.headers, [e2e-llm-inference-service] redirect=False, [e2e-llm-inference-service] assert_same_host=False, [e2e-llm-inference-service] preload_content=False, [e2e-llm-inference-service] decode_content=False, [e2e-llm-inference-service] retries=self.max_retries, [e2e-llm-inference-service] timeout=timeout, [e2e-llm-inference-service] chunked=chunked, [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] ../../python/kserve/.venv/lib64/python3.11/site-packages/requests/adapters.py:667: [e2e-llm-inference-service] _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ [e2e-llm-inference-service] [e2e-llm-inference-service] self = [e2e-llm-inference-service] method = 'POST' [e2e-llm-inference-service] url = '/kserve-ci-e2e-test/custom-route-timeout-test/v1/completions' [e2e-llm-inference-service] body = b'{"model": "facebook/opt-125m", "prompt": "KServe is a", "max_tokens": 20}' [e2e-llm-inference-service] headers = {'User-Agent': 'python-requests/2.32.3', 'Accept-Encoding': 'gzip, deflate', 'Accept': '*/*', 'Connection': 'keep-alive', 'Content-Type': 'application/json', 'Content-Length': '73'} [e2e-llm-inference-service] retries = Retry(total=7, connect=None, read=None, redirect=None, status=None) [e2e-llm-inference-service] redirect = False, assert_same_host = False [e2e-llm-inference-service] timeout = Timeout(connect=60, read=60, total=None), pool_timeout = None [e2e-llm-inference-service] release_conn = False, chunked = False, body_pos = None, preload_content = False [e2e-llm-inference-service] decode_content = False, response_kw = {} [e2e-llm-inference-service] parsed_url = Url(scheme=None, auth=None, host=None, port=None, path='/kserve-ci-e2e-test/custom-route-timeout-test/v1/completions', query=None, fragment=None) [e2e-llm-inference-service] destination_scheme = None, conn = None, release_this_conn = True [e2e-llm-inference-service] http_tunnel_required = False [e2e-llm-inference-service] err = ReadTimeoutError("HTTPConnectionPool(host='ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com', port=80): Read timed out. (read timeout=60)") [e2e-llm-inference-service] clean_exit = False [e2e-llm-inference-service] [e2e-llm-inference-service] def urlopen( # type: ignore[override] [e2e-llm-inference-service] self, [e2e-llm-inference-service] method: str, [e2e-llm-inference-service] url: str, [e2e-llm-inference-service] body: _TYPE_BODY | None = None, [e2e-llm-inference-service] headers: typing.Mapping[str, str] | None = None, [e2e-llm-inference-service] retries: Retry | bool | int | None = None, [e2e-llm-inference-service] redirect: bool = True, [e2e-llm-inference-service] assert_same_host: bool = True, [e2e-llm-inference-service] timeout: _TYPE_TIMEOUT = _DEFAULT_TIMEOUT, [e2e-llm-inference-service] pool_timeout: int | None = None, [e2e-llm-inference-service] release_conn: bool | None = None, [e2e-llm-inference-service] chunked: bool = False, [e2e-llm-inference-service] body_pos: _TYPE_BODY_POSITION | None = None, [e2e-llm-inference-service] preload_content: bool = True, [e2e-llm-inference-service] decode_content: bool = True, [e2e-llm-inference-service] **response_kw: typing.Any, [e2e-llm-inference-service] ) -> BaseHTTPResponse: [e2e-llm-inference-service] """ [e2e-llm-inference-service] Get a connection from the pool and perform an HTTP request. This is the [e2e-llm-inference-service] lowest level call for making a request, so you'll need to specify all [e2e-llm-inference-service] the raw details. [e2e-llm-inference-service] [e2e-llm-inference-service] .. note:: [e2e-llm-inference-service] [e2e-llm-inference-service] More commonly, it's appropriate to use a convenience method [e2e-llm-inference-service] such as :meth:`request`. [e2e-llm-inference-service] [e2e-llm-inference-service] .. note:: [e2e-llm-inference-service] [e2e-llm-inference-service] `release_conn` will only behave as expected if [e2e-llm-inference-service] `preload_content=False` because we want to make [e2e-llm-inference-service] `preload_content=False` the default behaviour someday soon without [e2e-llm-inference-service] breaking backwards compatibility. [e2e-llm-inference-service] [e2e-llm-inference-service] :param method: [e2e-llm-inference-service] HTTP request method (such as GET, POST, PUT, etc.) [e2e-llm-inference-service] [e2e-llm-inference-service] :param url: [e2e-llm-inference-service] The URL to perform the request on. [e2e-llm-inference-service] [e2e-llm-inference-service] :param body: [e2e-llm-inference-service] Data to send in the request body, either :class:`str`, :class:`bytes`, [e2e-llm-inference-service] an iterable of :class:`str`/:class:`bytes`, or a file-like object. [e2e-llm-inference-service] [e2e-llm-inference-service] :param headers: [e2e-llm-inference-service] Dictionary of custom headers to send, such as User-Agent, [e2e-llm-inference-service] If-None-Match, etc. If None, pool headers are used. If provided, [e2e-llm-inference-service] these headers completely replace any pool-specific headers. [e2e-llm-inference-service] [e2e-llm-inference-service] :param retries: [e2e-llm-inference-service] Configure the number of retries to allow before raising a [e2e-llm-inference-service] :class:`~urllib3.exceptions.MaxRetryError` exception. [e2e-llm-inference-service] [e2e-llm-inference-service] If ``None`` (default) will retry 3 times, see ``Retry.DEFAULT``. Pass a [e2e-llm-inference-service] :class:`~urllib3.util.retry.Retry` object for fine-grained control [e2e-llm-inference-service] over different types of retries. [e2e-llm-inference-service] Pass an integer number to retry connection errors that many times, [e2e-llm-inference-service] but no other types of errors. Pass zero to never retry. [e2e-llm-inference-service] [e2e-llm-inference-service] If ``False``, then retries are disabled and any exception is raised [e2e-llm-inference-service] immediately. Also, instead of raising a MaxRetryError on redirects, [e2e-llm-inference-service] the redirect response will be returned. [e2e-llm-inference-service] [e2e-llm-inference-service] :type retries: :class:`~urllib3.util.retry.Retry`, False, or an int. [e2e-llm-inference-service] [e2e-llm-inference-service] :param redirect: [e2e-llm-inference-service] If True, automatically handle redirects (status codes 301, 302, [e2e-llm-inference-service] 303, 307, 308). Each redirect counts as a retry. Disabling retries [e2e-llm-inference-service] will disable redirect, too. [e2e-llm-inference-service] [e2e-llm-inference-service] :param assert_same_host: [e2e-llm-inference-service] If ``True``, will make sure that the host of the pool requests is [e2e-llm-inference-service] consistent else will raise HostChangedError. When ``False``, you can [e2e-llm-inference-service] use the pool on an HTTP proxy and request foreign hosts. [e2e-llm-inference-service] [e2e-llm-inference-service] :param timeout: [e2e-llm-inference-service] If specified, overrides the default timeout for this one [e2e-llm-inference-service] request. It may be a float (in seconds) or an instance of [e2e-llm-inference-service] :class:`urllib3.util.Timeout`. [e2e-llm-inference-service] [e2e-llm-inference-service] :param pool_timeout: [e2e-llm-inference-service] If set and the pool is set to block=True, then this method will [e2e-llm-inference-service] block for ``pool_timeout`` seconds and raise EmptyPoolError if no [e2e-llm-inference-service] connection is available within the time period. [e2e-llm-inference-service] [e2e-llm-inference-service] :param bool preload_content: [e2e-llm-inference-service] If True, the response's body will be preloaded into memory. [e2e-llm-inference-service] [e2e-llm-inference-service] :param bool decode_content: [e2e-llm-inference-service] If True, will attempt to decode the body based on the [e2e-llm-inference-service] 'content-encoding' header. [e2e-llm-inference-service] [e2e-llm-inference-service] :param release_conn: [e2e-llm-inference-service] If False, then the urlopen call will not release the connection [e2e-llm-inference-service] back into the pool once a response is received (but will release if [e2e-llm-inference-service] you read the entire contents of the response such as when [e2e-llm-inference-service] `preload_content=True`). This is useful if you're not preloading [e2e-llm-inference-service] the response's content immediately. You will need to call [e2e-llm-inference-service] ``r.release_conn()`` on the response ``r`` to return the connection [e2e-llm-inference-service] back into the pool. If None, it takes the value of ``preload_content`` [e2e-llm-inference-service] which defaults to ``True``. [e2e-llm-inference-service] [e2e-llm-inference-service] :param bool chunked: [e2e-llm-inference-service] If True, urllib3 will send the body using chunked transfer [e2e-llm-inference-service] encoding. Otherwise, urllib3 will send the body using the standard [e2e-llm-inference-service] content-length form. Defaults to False. [e2e-llm-inference-service] [e2e-llm-inference-service] :param int body_pos: [e2e-llm-inference-service] Position to seek to in file-like body in the event of a retry or [e2e-llm-inference-service] redirect. Typically this won't need to be set because urllib3 will [e2e-llm-inference-service] auto-populate the value when needed. [e2e-llm-inference-service] """ [e2e-llm-inference-service] parsed_url = parse_url(url) [e2e-llm-inference-service] destination_scheme = parsed_url.scheme [e2e-llm-inference-service] [e2e-llm-inference-service] if headers is None: [e2e-llm-inference-service] headers = self.headers [e2e-llm-inference-service] [e2e-llm-inference-service] if not isinstance(retries, Retry): [e2e-llm-inference-service] retries = Retry.from_int(retries, redirect=redirect, default=self.retries) [e2e-llm-inference-service] [e2e-llm-inference-service] if release_conn is None: [e2e-llm-inference-service] release_conn = preload_content [e2e-llm-inference-service] [e2e-llm-inference-service] # Check host [e2e-llm-inference-service] if assert_same_host and not self.is_same_host(url): [e2e-llm-inference-service] raise HostChangedError(self, url, retries) [e2e-llm-inference-service] [e2e-llm-inference-service] # Ensure that the URL we're connecting to is properly encoded [e2e-llm-inference-service] if url.startswith("/"): [e2e-llm-inference-service] url = to_str(_encode_target(url)) [e2e-llm-inference-service] else: [e2e-llm-inference-service] url = to_str(parsed_url.url) [e2e-llm-inference-service] [e2e-llm-inference-service] conn = None [e2e-llm-inference-service] [e2e-llm-inference-service] # Track whether `conn` needs to be released before [e2e-llm-inference-service] # returning/raising/recursing. Update this variable if necessary, and [e2e-llm-inference-service] # leave `release_conn` constant throughout the function. That way, if [e2e-llm-inference-service] # the function recurses, the original value of `release_conn` will be [e2e-llm-inference-service] # passed down into the recursive call, and its value will be respected. [e2e-llm-inference-service] # [e2e-llm-inference-service] # See issue #651 [1] for details. [e2e-llm-inference-service] # [e2e-llm-inference-service] # [1] [e2e-llm-inference-service] release_this_conn = release_conn [e2e-llm-inference-service] [e2e-llm-inference-service] http_tunnel_required = connection_requires_http_tunnel( [e2e-llm-inference-service] self.proxy, self.proxy_config, destination_scheme [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] # Merge the proxy headers. Only done when not using HTTP CONNECT. We [e2e-llm-inference-service] # have to copy the headers dict so we can safely change it without those [e2e-llm-inference-service] # changes being reflected in anyone else's copy. [e2e-llm-inference-service] if not http_tunnel_required: [e2e-llm-inference-service] headers = headers.copy() # type: ignore[attr-defined] [e2e-llm-inference-service] headers.update(self.proxy_headers) # type: ignore[union-attr] [e2e-llm-inference-service] [e2e-llm-inference-service] # Must keep the exception bound to a separate variable or else Python 3 [e2e-llm-inference-service] # complains about UnboundLocalError. [e2e-llm-inference-service] err = None [e2e-llm-inference-service] [e2e-llm-inference-service] # Keep track of whether we cleanly exited the except block. This [e2e-llm-inference-service] # ensures we do proper cleanup in finally. [e2e-llm-inference-service] clean_exit = False [e2e-llm-inference-service] [e2e-llm-inference-service] # Rewind body position, if needed. Record current position [e2e-llm-inference-service] # for future rewinds in the event of a redirect/retry. [e2e-llm-inference-service] body_pos = set_file_position(body, body_pos) [e2e-llm-inference-service] [e2e-llm-inference-service] try: [e2e-llm-inference-service] # Request a connection from the queue. [e2e-llm-inference-service] timeout_obj = self._get_timeout(timeout) [e2e-llm-inference-service] conn = self._get_conn(timeout=pool_timeout) [e2e-llm-inference-service] [e2e-llm-inference-service] conn.timeout = timeout_obj.connect_timeout # type: ignore[assignment] [e2e-llm-inference-service] [e2e-llm-inference-service] # Is this a closed/new connection that requires CONNECT tunnelling? [e2e-llm-inference-service] if self.proxy is not None and http_tunnel_required and conn.is_closed: [e2e-llm-inference-service] try: [e2e-llm-inference-service] self._prepare_proxy(conn) [e2e-llm-inference-service] except (BaseSSLError, OSError, SocketTimeout) as e: [e2e-llm-inference-service] self._raise_timeout( [e2e-llm-inference-service] err=e, url=self.proxy.url, timeout_value=conn.timeout [e2e-llm-inference-service] ) [e2e-llm-inference-service] raise [e2e-llm-inference-service] [e2e-llm-inference-service] # If we're going to release the connection in ``finally:``, then [e2e-llm-inference-service] # the response doesn't need to know about the connection. Otherwise [e2e-llm-inference-service] # it will also try to release it and we'll have a double-release [e2e-llm-inference-service] # mess. [e2e-llm-inference-service] response_conn = conn if not release_conn else None [e2e-llm-inference-service] [e2e-llm-inference-service] # Make the request on the HTTPConnection object [e2e-llm-inference-service] response = self._make_request( [e2e-llm-inference-service] conn, [e2e-llm-inference-service] method, [e2e-llm-inference-service] url, [e2e-llm-inference-service] timeout=timeout_obj, [e2e-llm-inference-service] body=body, [e2e-llm-inference-service] headers=headers, [e2e-llm-inference-service] chunked=chunked, [e2e-llm-inference-service] retries=retries, [e2e-llm-inference-service] response_conn=response_conn, [e2e-llm-inference-service] preload_content=preload_content, [e2e-llm-inference-service] decode_content=decode_content, [e2e-llm-inference-service] **response_kw, [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] # Everything went great! [e2e-llm-inference-service] clean_exit = True [e2e-llm-inference-service] [e2e-llm-inference-service] except EmptyPoolError: [e2e-llm-inference-service] # Didn't get a connection from the pool, no need to clean up [e2e-llm-inference-service] clean_exit = True [e2e-llm-inference-service] release_this_conn = False [e2e-llm-inference-service] raise [e2e-llm-inference-service] [e2e-llm-inference-service] except ( [e2e-llm-inference-service] TimeoutError, [e2e-llm-inference-service] HTTPException, [e2e-llm-inference-service] OSError, [e2e-llm-inference-service] ProtocolError, [e2e-llm-inference-service] BaseSSLError, [e2e-llm-inference-service] SSLError, [e2e-llm-inference-service] CertificateError, [e2e-llm-inference-service] ProxyError, [e2e-llm-inference-service] ) as e: [e2e-llm-inference-service] # Discard the connection for these exceptions. It will be [e2e-llm-inference-service] # replaced during the next _get_conn() call. [e2e-llm-inference-service] clean_exit = False [e2e-llm-inference-service] new_e: Exception = e [e2e-llm-inference-service] if isinstance(e, (BaseSSLError, CertificateError)): [e2e-llm-inference-service] new_e = SSLError(e) [e2e-llm-inference-service] if isinstance( [e2e-llm-inference-service] new_e, [e2e-llm-inference-service] ( [e2e-llm-inference-service] OSError, [e2e-llm-inference-service] NewConnectionError, [e2e-llm-inference-service] TimeoutError, [e2e-llm-inference-service] SSLError, [e2e-llm-inference-service] HTTPException, [e2e-llm-inference-service] ), [e2e-llm-inference-service] ) and (conn and conn.proxy and not conn.has_connected_to_proxy): [e2e-llm-inference-service] new_e = _wrap_proxy_error(new_e, conn.proxy.scheme) [e2e-llm-inference-service] elif isinstance(new_e, (OSError, HTTPException)): [e2e-llm-inference-service] new_e = ProtocolError("Connection aborted.", new_e) [e2e-llm-inference-service] [e2e-llm-inference-service] retries = retries.increment( [e2e-llm-inference-service] method, url, error=new_e, _pool=self, _stacktrace=sys.exc_info()[2] [e2e-llm-inference-service] ) [e2e-llm-inference-service] retries.sleep() [e2e-llm-inference-service] [e2e-llm-inference-service] # Keep track of the error for the retry warning. [e2e-llm-inference-service] err = e [e2e-llm-inference-service] [e2e-llm-inference-service] finally: [e2e-llm-inference-service] if not clean_exit: [e2e-llm-inference-service] # We hit some kind of exception, handled or otherwise. We need [e2e-llm-inference-service] # to throw the connection away unless explicitly told not to. [e2e-llm-inference-service] # Close the connection, set the variable to None, and make sure [e2e-llm-inference-service] # we put the None back in the pool to avoid leaking it. [e2e-llm-inference-service] if conn: [e2e-llm-inference-service] conn.close() [e2e-llm-inference-service] conn = None [e2e-llm-inference-service] release_this_conn = True [e2e-llm-inference-service] [e2e-llm-inference-service] if release_this_conn: [e2e-llm-inference-service] # Put the connection back to be reused. If the connection is [e2e-llm-inference-service] # expired then it will be None, which will get replaced with a [e2e-llm-inference-service] # fresh connection during _get_conn. [e2e-llm-inference-service] self._put_conn(conn) [e2e-llm-inference-service] [e2e-llm-inference-service] if not conn: [e2e-llm-inference-service] # Try again [e2e-llm-inference-service] log.warning( [e2e-llm-inference-service] "Retrying (%r) after connection broken by '%r': %s", retries, err, url [e2e-llm-inference-service] ) [e2e-llm-inference-service] > return self.urlopen( [e2e-llm-inference-service] method, [e2e-llm-inference-service] url, [e2e-llm-inference-service] body, [e2e-llm-inference-service] headers, [e2e-llm-inference-service] retries, [e2e-llm-inference-service] redirect, [e2e-llm-inference-service] assert_same_host, [e2e-llm-inference-service] timeout=timeout, [e2e-llm-inference-service] pool_timeout=pool_timeout, [e2e-llm-inference-service] release_conn=release_conn, [e2e-llm-inference-service] chunked=chunked, [e2e-llm-inference-service] body_pos=body_pos, [e2e-llm-inference-service] preload_content=preload_content, [e2e-llm-inference-service] decode_content=decode_content, [e2e-llm-inference-service] **response_kw, [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] ../../python/kserve/.venv/lib64/python3.11/site-packages/urllib3/connectionpool.py:871: [e2e-llm-inference-service] _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ [e2e-llm-inference-service] [e2e-llm-inference-service] self = [e2e-llm-inference-service] method = 'POST' [e2e-llm-inference-service] url = '/kserve-ci-e2e-test/custom-route-timeout-test/v1/completions' [e2e-llm-inference-service] body = b'{"model": "facebook/opt-125m", "prompt": "KServe is a", "max_tokens": 20}' [e2e-llm-inference-service] headers = {'User-Agent': 'python-requests/2.32.3', 'Accept-Encoding': 'gzip, deflate', 'Accept': '*/*', 'Connection': 'keep-alive', 'Content-Type': 'application/json', 'Content-Length': '73'} [e2e-llm-inference-service] retries = Retry(total=6, connect=None, read=None, redirect=None, status=None) [e2e-llm-inference-service] redirect = False, assert_same_host = False [e2e-llm-inference-service] timeout = Timeout(connect=60, read=60, total=None), pool_timeout = None [e2e-llm-inference-service] release_conn = False, chunked = False, body_pos = None, preload_content = False [e2e-llm-inference-service] decode_content = False, response_kw = {} [e2e-llm-inference-service] parsed_url = Url(scheme=None, auth=None, host=None, port=None, path='/kserve-ci-e2e-test/custom-route-timeout-test/v1/completions', query=None, fragment=None) [e2e-llm-inference-service] destination_scheme = None, conn = None, release_this_conn = True [e2e-llm-inference-service] http_tunnel_required = False [e2e-llm-inference-service] err = ReadTimeoutError("HTTPConnectionPool(host='ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com', port=80): Read timed out. (read timeout=60)") [e2e-llm-inference-service] clean_exit = False [e2e-llm-inference-service] [e2e-llm-inference-service] def urlopen( # type: ignore[override] [e2e-llm-inference-service] self, [e2e-llm-inference-service] method: str, [e2e-llm-inference-service] url: str, [e2e-llm-inference-service] body: _TYPE_BODY | None = None, [e2e-llm-inference-service] headers: typing.Mapping[str, str] | None = None, [e2e-llm-inference-service] retries: Retry | bool | int | None = None, [e2e-llm-inference-service] redirect: bool = True, [e2e-llm-inference-service] assert_same_host: bool = True, [e2e-llm-inference-service] timeout: _TYPE_TIMEOUT = _DEFAULT_TIMEOUT, [e2e-llm-inference-service] pool_timeout: int | None = None, [e2e-llm-inference-service] release_conn: bool | None = None, [e2e-llm-inference-service] chunked: bool = False, [e2e-llm-inference-service] body_pos: _TYPE_BODY_POSITION | None = None, [e2e-llm-inference-service] preload_content: bool = True, [e2e-llm-inference-service] decode_content: bool = True, [e2e-llm-inference-service] **response_kw: typing.Any, [e2e-llm-inference-service] ) -> BaseHTTPResponse: [e2e-llm-inference-service] """ [e2e-llm-inference-service] Get a connection from the pool and perform an HTTP request. This is the [e2e-llm-inference-service] lowest level call for making a request, so you'll need to specify all [e2e-llm-inference-service] the raw details. [e2e-llm-inference-service] [e2e-llm-inference-service] .. note:: [e2e-llm-inference-service] [e2e-llm-inference-service] More commonly, it's appropriate to use a convenience method [e2e-llm-inference-service] such as :meth:`request`. [e2e-llm-inference-service] [e2e-llm-inference-service] .. note:: [e2e-llm-inference-service] [e2e-llm-inference-service] `release_conn` will only behave as expected if [e2e-llm-inference-service] `preload_content=False` because we want to make [e2e-llm-inference-service] `preload_content=False` the default behaviour someday soon without [e2e-llm-inference-service] breaking backwards compatibility. [e2e-llm-inference-service] [e2e-llm-inference-service] :param method: [e2e-llm-inference-service] HTTP request method (such as GET, POST, PUT, etc.) [e2e-llm-inference-service] [e2e-llm-inference-service] :param url: [e2e-llm-inference-service] The URL to perform the request on. [e2e-llm-inference-service] [e2e-llm-inference-service] :param body: [e2e-llm-inference-service] Data to send in the request body, either :class:`str`, :class:`bytes`, [e2e-llm-inference-service] an iterable of :class:`str`/:class:`bytes`, or a file-like object. [e2e-llm-inference-service] [e2e-llm-inference-service] :param headers: [e2e-llm-inference-service] Dictionary of custom headers to send, such as User-Agent, [e2e-llm-inference-service] If-None-Match, etc. If None, pool headers are used. If provided, [e2e-llm-inference-service] these headers completely replace any pool-specific headers. [e2e-llm-inference-service] [e2e-llm-inference-service] :param retries: [e2e-llm-inference-service] Configure the number of retries to allow before raising a [e2e-llm-inference-service] :class:`~urllib3.exceptions.MaxRetryError` exception. [e2e-llm-inference-service] [e2e-llm-inference-service] If ``None`` (default) will retry 3 times, see ``Retry.DEFAULT``. Pass a [e2e-llm-inference-service] :class:`~urllib3.util.retry.Retry` object for fine-grained control [e2e-llm-inference-service] over different types of retries. [e2e-llm-inference-service] Pass an integer number to retry connection errors that many times, [e2e-llm-inference-service] but no other types of errors. Pass zero to never retry. [e2e-llm-inference-service] [e2e-llm-inference-service] If ``False``, then retries are disabled and any exception is raised [e2e-llm-inference-service] immediately. Also, instead of raising a MaxRetryError on redirects, [e2e-llm-inference-service] the redirect response will be returned. [e2e-llm-inference-service] [e2e-llm-inference-service] :type retries: :class:`~urllib3.util.retry.Retry`, False, or an int. [e2e-llm-inference-service] [e2e-llm-inference-service] :param redirect: [e2e-llm-inference-service] If True, automatically handle redirects (status codes 301, 302, [e2e-llm-inference-service] 303, 307, 308). Each redirect counts as a retry. Disabling retries [e2e-llm-inference-service] will disable redirect, too. [e2e-llm-inference-service] [e2e-llm-inference-service] :param assert_same_host: [e2e-llm-inference-service] If ``True``, will make sure that the host of the pool requests is [e2e-llm-inference-service] consistent else will raise HostChangedError. When ``False``, you can [e2e-llm-inference-service] use the pool on an HTTP proxy and request foreign hosts. [e2e-llm-inference-service] [e2e-llm-inference-service] :param timeout: [e2e-llm-inference-service] If specified, overrides the default timeout for this one [e2e-llm-inference-service] request. It may be a float (in seconds) or an instance of [e2e-llm-inference-service] :class:`urllib3.util.Timeout`. [e2e-llm-inference-service] [e2e-llm-inference-service] :param pool_timeout: [e2e-llm-inference-service] If set and the pool is set to block=True, then this method will [e2e-llm-inference-service] block for ``pool_timeout`` seconds and raise EmptyPoolError if no [e2e-llm-inference-service] connection is available within the time period. [e2e-llm-inference-service] [e2e-llm-inference-service] :param bool preload_content: [e2e-llm-inference-service] If True, the response's body will be preloaded into memory. [e2e-llm-inference-service] [e2e-llm-inference-service] :param bool decode_content: [e2e-llm-inference-service] If True, will attempt to decode the body based on the [e2e-llm-inference-service] 'content-encoding' header. [e2e-llm-inference-service] [e2e-llm-inference-service] :param release_conn: [e2e-llm-inference-service] If False, then the urlopen call will not release the connection [e2e-llm-inference-service] back into the pool once a response is received (but will release if [e2e-llm-inference-service] you read the entire contents of the response such as when [e2e-llm-inference-service] `preload_content=True`). This is useful if you're not preloading [e2e-llm-inference-service] the response's content immediately. You will need to call [e2e-llm-inference-service] ``r.release_conn()`` on the response ``r`` to return the connection [e2e-llm-inference-service] back into the pool. If None, it takes the value of ``preload_content`` [e2e-llm-inference-service] which defaults to ``True``. [e2e-llm-inference-service] [e2e-llm-inference-service] :param bool chunked: [e2e-llm-inference-service] If True, urllib3 will send the body using chunked transfer [e2e-llm-inference-service] encoding. Otherwise, urllib3 will send the body using the standard [e2e-llm-inference-service] content-length form. Defaults to False. [e2e-llm-inference-service] [e2e-llm-inference-service] :param int body_pos: [e2e-llm-inference-service] Position to seek to in file-like body in the event of a retry or [e2e-llm-inference-service] redirect. Typically this won't need to be set because urllib3 will [e2e-llm-inference-service] auto-populate the value when needed. [e2e-llm-inference-service] """ [e2e-llm-inference-service] parsed_url = parse_url(url) [e2e-llm-inference-service] destination_scheme = parsed_url.scheme [e2e-llm-inference-service] [e2e-llm-inference-service] if headers is None: [e2e-llm-inference-service] headers = self.headers [e2e-llm-inference-service] [e2e-llm-inference-service] if not isinstance(retries, Retry): [e2e-llm-inference-service] retries = Retry.from_int(retries, redirect=redirect, default=self.retries) [e2e-llm-inference-service] [e2e-llm-inference-service] if release_conn is None: [e2e-llm-inference-service] release_conn = preload_content [e2e-llm-inference-service] [e2e-llm-inference-service] # Check host [e2e-llm-inference-service] if assert_same_host and not self.is_same_host(url): [e2e-llm-inference-service] raise HostChangedError(self, url, retries) [e2e-llm-inference-service] [e2e-llm-inference-service] # Ensure that the URL we're connecting to is properly encoded [e2e-llm-inference-service] if url.startswith("/"): [e2e-llm-inference-service] url = to_str(_encode_target(url)) [e2e-llm-inference-service] else: [e2e-llm-inference-service] url = to_str(parsed_url.url) [e2e-llm-inference-service] [e2e-llm-inference-service] conn = None [e2e-llm-inference-service] [e2e-llm-inference-service] # Track whether `conn` needs to be released before [e2e-llm-inference-service] # returning/raising/recursing. Update this variable if necessary, and [e2e-llm-inference-service] # leave `release_conn` constant throughout the function. That way, if [e2e-llm-inference-service] # the function recurses, the original value of `release_conn` will be [e2e-llm-inference-service] # passed down into the recursive call, and its value will be respected. [e2e-llm-inference-service] # [e2e-llm-inference-service] # See issue #651 [1] for details. [e2e-llm-inference-service] # [e2e-llm-inference-service] # [1] [e2e-llm-inference-service] release_this_conn = release_conn [e2e-llm-inference-service] [e2e-llm-inference-service] http_tunnel_required = connection_requires_http_tunnel( [e2e-llm-inference-service] self.proxy, self.proxy_config, destination_scheme [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] # Merge the proxy headers. Only done when not using HTTP CONNECT. We [e2e-llm-inference-service] # have to copy the headers dict so we can safely change it without those [e2e-llm-inference-service] # changes being reflected in anyone else's copy. [e2e-llm-inference-service] if not http_tunnel_required: [e2e-llm-inference-service] headers = headers.copy() # type: ignore[attr-defined] [e2e-llm-inference-service] headers.update(self.proxy_headers) # type: ignore[union-attr] [e2e-llm-inference-service] [e2e-llm-inference-service] # Must keep the exception bound to a separate variable or else Python 3 [e2e-llm-inference-service] # complains about UnboundLocalError. [e2e-llm-inference-service] err = None [e2e-llm-inference-service] [e2e-llm-inference-service] # Keep track of whether we cleanly exited the except block. This [e2e-llm-inference-service] # ensures we do proper cleanup in finally. [e2e-llm-inference-service] clean_exit = False [e2e-llm-inference-service] [e2e-llm-inference-service] # Rewind body position, if needed. Record current position [e2e-llm-inference-service] # for future rewinds in the event of a redirect/retry. [e2e-llm-inference-service] body_pos = set_file_position(body, body_pos) [e2e-llm-inference-service] [e2e-llm-inference-service] try: [e2e-llm-inference-service] # Request a connection from the queue. [e2e-llm-inference-service] timeout_obj = self._get_timeout(timeout) [e2e-llm-inference-service] conn = self._get_conn(timeout=pool_timeout) [e2e-llm-inference-service] [e2e-llm-inference-service] conn.timeout = timeout_obj.connect_timeout # type: ignore[assignment] [e2e-llm-inference-service] [e2e-llm-inference-service] # Is this a closed/new connection that requires CONNECT tunnelling? [e2e-llm-inference-service] if self.proxy is not None and http_tunnel_required and conn.is_closed: [e2e-llm-inference-service] try: [e2e-llm-inference-service] self._prepare_proxy(conn) [e2e-llm-inference-service] except (BaseSSLError, OSError, SocketTimeout) as e: [e2e-llm-inference-service] self._raise_timeout( [e2e-llm-inference-service] err=e, url=self.proxy.url, timeout_value=conn.timeout [e2e-llm-inference-service] ) [e2e-llm-inference-service] raise [e2e-llm-inference-service] [e2e-llm-inference-service] # If we're going to release the connection in ``finally:``, then [e2e-llm-inference-service] # the response doesn't need to know about the connection. Otherwise [e2e-llm-inference-service] # it will also try to release it and we'll have a double-release [e2e-llm-inference-service] # mess. [e2e-llm-inference-service] response_conn = conn if not release_conn else None [e2e-llm-inference-service] [e2e-llm-inference-service] # Make the request on the HTTPConnection object [e2e-llm-inference-service] response = self._make_request( [e2e-llm-inference-service] conn, [e2e-llm-inference-service] method, [e2e-llm-inference-service] url, [e2e-llm-inference-service] timeout=timeout_obj, [e2e-llm-inference-service] body=body, [e2e-llm-inference-service] headers=headers, [e2e-llm-inference-service] chunked=chunked, [e2e-llm-inference-service] retries=retries, [e2e-llm-inference-service] response_conn=response_conn, [e2e-llm-inference-service] preload_content=preload_content, [e2e-llm-inference-service] decode_content=decode_content, [e2e-llm-inference-service] **response_kw, [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] # Everything went great! [e2e-llm-inference-service] clean_exit = True [e2e-llm-inference-service] [e2e-llm-inference-service] except EmptyPoolError: [e2e-llm-inference-service] # Didn't get a connection from the pool, no need to clean up [e2e-llm-inference-service] clean_exit = True [e2e-llm-inference-service] release_this_conn = False [e2e-llm-inference-service] raise [e2e-llm-inference-service] [e2e-llm-inference-service] except ( [e2e-llm-inference-service] TimeoutError, [e2e-llm-inference-service] HTTPException, [e2e-llm-inference-service] OSError, [e2e-llm-inference-service] ProtocolError, [e2e-llm-inference-service] BaseSSLError, [e2e-llm-inference-service] SSLError, [e2e-llm-inference-service] CertificateError, [e2e-llm-inference-service] ProxyError, [e2e-llm-inference-service] ) as e: [e2e-llm-inference-service] # Discard the connection for these exceptions. It will be [e2e-llm-inference-service] # replaced during the next _get_conn() call. [e2e-llm-inference-service] clean_exit = False [e2e-llm-inference-service] new_e: Exception = e [e2e-llm-inference-service] if isinstance(e, (BaseSSLError, CertificateError)): [e2e-llm-inference-service] new_e = SSLError(e) [e2e-llm-inference-service] if isinstance( [e2e-llm-inference-service] new_e, [e2e-llm-inference-service] ( [e2e-llm-inference-service] OSError, [e2e-llm-inference-service] NewConnectionError, [e2e-llm-inference-service] TimeoutError, [e2e-llm-inference-service] SSLError, [e2e-llm-inference-service] HTTPException, [e2e-llm-inference-service] ), [e2e-llm-inference-service] ) and (conn and conn.proxy and not conn.has_connected_to_proxy): [e2e-llm-inference-service] new_e = _wrap_proxy_error(new_e, conn.proxy.scheme) [e2e-llm-inference-service] elif isinstance(new_e, (OSError, HTTPException)): [e2e-llm-inference-service] new_e = ProtocolError("Connection aborted.", new_e) [e2e-llm-inference-service] [e2e-llm-inference-service] retries = retries.increment( [e2e-llm-inference-service] method, url, error=new_e, _pool=self, _stacktrace=sys.exc_info()[2] [e2e-llm-inference-service] ) [e2e-llm-inference-service] retries.sleep() [e2e-llm-inference-service] [e2e-llm-inference-service] # Keep track of the error for the retry warning. [e2e-llm-inference-service] err = e [e2e-llm-inference-service] [e2e-llm-inference-service] finally: [e2e-llm-inference-service] if not clean_exit: [e2e-llm-inference-service] # We hit some kind of exception, handled or otherwise. We need [e2e-llm-inference-service] # to throw the connection away unless explicitly told not to. [e2e-llm-inference-service] # Close the connection, set the variable to None, and make sure [e2e-llm-inference-service] # we put the None back in the pool to avoid leaking it. [e2e-llm-inference-service] if conn: [e2e-llm-inference-service] conn.close() [e2e-llm-inference-service] conn = None [e2e-llm-inference-service] release_this_conn = True [e2e-llm-inference-service] [e2e-llm-inference-service] if release_this_conn: [e2e-llm-inference-service] # Put the connection back to be reused. If the connection is [e2e-llm-inference-service] # expired then it will be None, which will get replaced with a [e2e-llm-inference-service] # fresh connection during _get_conn. [e2e-llm-inference-service] self._put_conn(conn) [e2e-llm-inference-service] [e2e-llm-inference-service] if not conn: [e2e-llm-inference-service] # Try again [e2e-llm-inference-service] log.warning( [e2e-llm-inference-service] "Retrying (%r) after connection broken by '%r': %s", retries, err, url [e2e-llm-inference-service] ) [e2e-llm-inference-service] > return self.urlopen( [e2e-llm-inference-service] method, [e2e-llm-inference-service] url, [e2e-llm-inference-service] body, [e2e-llm-inference-service] headers, [e2e-llm-inference-service] retries, [e2e-llm-inference-service] redirect, [e2e-llm-inference-service] assert_same_host, [e2e-llm-inference-service] timeout=timeout, [e2e-llm-inference-service] pool_timeout=pool_timeout, [e2e-llm-inference-service] release_conn=release_conn, [e2e-llm-inference-service] chunked=chunked, [e2e-llm-inference-service] body_pos=body_pos, [e2e-llm-inference-service] preload_content=preload_content, [e2e-llm-inference-service] decode_content=decode_content, [e2e-llm-inference-service] **response_kw, [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] ../../python/kserve/.venv/lib64/python3.11/site-packages/urllib3/connectionpool.py:871: [e2e-llm-inference-service] _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ [e2e-llm-inference-service] [e2e-llm-inference-service] self = [e2e-llm-inference-service] method = 'POST' [e2e-llm-inference-service] url = '/kserve-ci-e2e-test/custom-route-timeout-test/v1/completions' [e2e-llm-inference-service] body = b'{"model": "facebook/opt-125m", "prompt": "KServe is a", "max_tokens": 20}' [e2e-llm-inference-service] headers = {'User-Agent': 'python-requests/2.32.3', 'Accept-Encoding': 'gzip, deflate', 'Accept': '*/*', 'Connection': 'keep-alive', 'Content-Type': 'application/json', 'Content-Length': '73'} [e2e-llm-inference-service] retries = Retry(total=5, connect=None, read=None, redirect=None, status=None) [e2e-llm-inference-service] redirect = False, assert_same_host = False [e2e-llm-inference-service] timeout = Timeout(connect=60, read=60, total=None), pool_timeout = None [e2e-llm-inference-service] release_conn = False, chunked = False, body_pos = None, preload_content = False [e2e-llm-inference-service] decode_content = False, response_kw = {} [e2e-llm-inference-service] parsed_url = Url(scheme=None, auth=None, host=None, port=None, path='/kserve-ci-e2e-test/custom-route-timeout-test/v1/completions', query=None, fragment=None) [e2e-llm-inference-service] destination_scheme = None, conn = None, release_this_conn = True [e2e-llm-inference-service] http_tunnel_required = False [e2e-llm-inference-service] err = ReadTimeoutError("HTTPConnectionPool(host='ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com', port=80): Read timed out. (read timeout=60)") [e2e-llm-inference-service] clean_exit = False [e2e-llm-inference-service] [e2e-llm-inference-service] def urlopen( # type: ignore[override] [e2e-llm-inference-service] self, [e2e-llm-inference-service] method: str, [e2e-llm-inference-service] url: str, [e2e-llm-inference-service] body: _TYPE_BODY | None = None, [e2e-llm-inference-service] headers: typing.Mapping[str, str] | None = None, [e2e-llm-inference-service] retries: Retry | bool | int | None = None, [e2e-llm-inference-service] redirect: bool = True, [e2e-llm-inference-service] assert_same_host: bool = True, [e2e-llm-inference-service] timeout: _TYPE_TIMEOUT = _DEFAULT_TIMEOUT, [e2e-llm-inference-service] pool_timeout: int | None = None, [e2e-llm-inference-service] release_conn: bool | None = None, [e2e-llm-inference-service] chunked: bool = False, [e2e-llm-inference-service] body_pos: _TYPE_BODY_POSITION | None = None, [e2e-llm-inference-service] preload_content: bool = True, [e2e-llm-inference-service] decode_content: bool = True, [e2e-llm-inference-service] **response_kw: typing.Any, [e2e-llm-inference-service] ) -> BaseHTTPResponse: [e2e-llm-inference-service] """ [e2e-llm-inference-service] Get a connection from the pool and perform an HTTP request. This is the [e2e-llm-inference-service] lowest level call for making a request, so you'll need to specify all [e2e-llm-inference-service] the raw details. [e2e-llm-inference-service] [e2e-llm-inference-service] .. note:: [e2e-llm-inference-service] [e2e-llm-inference-service] More commonly, it's appropriate to use a convenience method [e2e-llm-inference-service] such as :meth:`request`. [e2e-llm-inference-service] [e2e-llm-inference-service] .. note:: [e2e-llm-inference-service] [e2e-llm-inference-service] `release_conn` will only behave as expected if [e2e-llm-inference-service] `preload_content=False` because we want to make [e2e-llm-inference-service] `preload_content=False` the default behaviour someday soon without [e2e-llm-inference-service] breaking backwards compatibility. [e2e-llm-inference-service] [e2e-llm-inference-service] :param method: [e2e-llm-inference-service] HTTP request method (such as GET, POST, PUT, etc.) [e2e-llm-inference-service] [e2e-llm-inference-service] :param url: [e2e-llm-inference-service] The URL to perform the request on. [e2e-llm-inference-service] [e2e-llm-inference-service] :param body: [e2e-llm-inference-service] Data to send in the request body, either :class:`str`, :class:`bytes`, [e2e-llm-inference-service] an iterable of :class:`str`/:class:`bytes`, or a file-like object. [e2e-llm-inference-service] [e2e-llm-inference-service] :param headers: [e2e-llm-inference-service] Dictionary of custom headers to send, such as User-Agent, [e2e-llm-inference-service] If-None-Match, etc. If None, pool headers are used. If provided, [e2e-llm-inference-service] these headers completely replace any pool-specific headers. [e2e-llm-inference-service] [e2e-llm-inference-service] :param retries: [e2e-llm-inference-service] Configure the number of retries to allow before raising a [e2e-llm-inference-service] :class:`~urllib3.exceptions.MaxRetryError` exception. [e2e-llm-inference-service] [e2e-llm-inference-service] If ``None`` (default) will retry 3 times, see ``Retry.DEFAULT``. Pass a [e2e-llm-inference-service] :class:`~urllib3.util.retry.Retry` object for fine-grained control [e2e-llm-inference-service] over different types of retries. [e2e-llm-inference-service] Pass an integer number to retry connection errors that many times, [e2e-llm-inference-service] but no other types of errors. Pass zero to never retry. [e2e-llm-inference-service] [e2e-llm-inference-service] If ``False``, then retries are disabled and any exception is raised [e2e-llm-inference-service] immediately. Also, instead of raising a MaxRetryError on redirects, [e2e-llm-inference-service] the redirect response will be returned. [e2e-llm-inference-service] [e2e-llm-inference-service] :type retries: :class:`~urllib3.util.retry.Retry`, False, or an int. [e2e-llm-inference-service] [e2e-llm-inference-service] :param redirect: [e2e-llm-inference-service] If True, automatically handle redirects (status codes 301, 302, [e2e-llm-inference-service] 303, 307, 308). Each redirect counts as a retry. Disabling retries [e2e-llm-inference-service] will disable redirect, too. [e2e-llm-inference-service] [e2e-llm-inference-service] :param assert_same_host: [e2e-llm-inference-service] If ``True``, will make sure that the host of the pool requests is [e2e-llm-inference-service] consistent else will raise HostChangedError. When ``False``, you can [e2e-llm-inference-service] use the pool on an HTTP proxy and request foreign hosts. [e2e-llm-inference-service] [e2e-llm-inference-service] :param timeout: [e2e-llm-inference-service] If specified, overrides the default timeout for this one [e2e-llm-inference-service] request. It may be a float (in seconds) or an instance of [e2e-llm-inference-service] :class:`urllib3.util.Timeout`. [e2e-llm-inference-service] [e2e-llm-inference-service] :param pool_timeout: [e2e-llm-inference-service] If set and the pool is set to block=True, then this method will [e2e-llm-inference-service] block for ``pool_timeout`` seconds and raise EmptyPoolError if no [e2e-llm-inference-service] connection is available within the time period. [e2e-llm-inference-service] [e2e-llm-inference-service] :param bool preload_content: [e2e-llm-inference-service] If True, the response's body will be preloaded into memory. [e2e-llm-inference-service] [e2e-llm-inference-service] :param bool decode_content: [e2e-llm-inference-service] If True, will attempt to decode the body based on the [e2e-llm-inference-service] 'content-encoding' header. [e2e-llm-inference-service] [e2e-llm-inference-service] :param release_conn: [e2e-llm-inference-service] If False, then the urlopen call will not release the connection [e2e-llm-inference-service] back into the pool once a response is received (but will release if [e2e-llm-inference-service] you read the entire contents of the response such as when [e2e-llm-inference-service] `preload_content=True`). This is useful if you're not preloading [e2e-llm-inference-service] the response's content immediately. You will need to call [e2e-llm-inference-service] ``r.release_conn()`` on the response ``r`` to return the connection [e2e-llm-inference-service] back into the pool. If None, it takes the value of ``preload_content`` [e2e-llm-inference-service] which defaults to ``True``. [e2e-llm-inference-service] [e2e-llm-inference-service] :param bool chunked: [e2e-llm-inference-service] If True, urllib3 will send the body using chunked transfer [e2e-llm-inference-service] encoding. Otherwise, urllib3 will send the body using the standard [e2e-llm-inference-service] content-length form. Defaults to False. [e2e-llm-inference-service] [e2e-llm-inference-service] :param int body_pos: [e2e-llm-inference-service] Position to seek to in file-like body in the event of a retry or [e2e-llm-inference-service] redirect. Typically this won't need to be set because urllib3 will [e2e-llm-inference-service] auto-populate the value when needed. [e2e-llm-inference-service] """ [e2e-llm-inference-service] parsed_url = parse_url(url) [e2e-llm-inference-service] destination_scheme = parsed_url.scheme [e2e-llm-inference-service] [e2e-llm-inference-service] if headers is None: [e2e-llm-inference-service] headers = self.headers [e2e-llm-inference-service] [e2e-llm-inference-service] if not isinstance(retries, Retry): [e2e-llm-inference-service] retries = Retry.from_int(retries, redirect=redirect, default=self.retries) [e2e-llm-inference-service] [e2e-llm-inference-service] if release_conn is None: [e2e-llm-inference-service] release_conn = preload_content [e2e-llm-inference-service] [e2e-llm-inference-service] # Check host [e2e-llm-inference-service] if assert_same_host and not self.is_same_host(url): [e2e-llm-inference-service] raise HostChangedError(self, url, retries) [e2e-llm-inference-service] [e2e-llm-inference-service] # Ensure that the URL we're connecting to is properly encoded [e2e-llm-inference-service] if url.startswith("/"): [e2e-llm-inference-service] url = to_str(_encode_target(url)) [e2e-llm-inference-service] else: [e2e-llm-inference-service] url = to_str(parsed_url.url) [e2e-llm-inference-service] [e2e-llm-inference-service] conn = None [e2e-llm-inference-service] [e2e-llm-inference-service] # Track whether `conn` needs to be released before [e2e-llm-inference-service] # returning/raising/recursing. Update this variable if necessary, and [e2e-llm-inference-service] # leave `release_conn` constant throughout the function. That way, if [e2e-llm-inference-service] # the function recurses, the original value of `release_conn` will be [e2e-llm-inference-service] # passed down into the recursive call, and its value will be respected. [e2e-llm-inference-service] # [e2e-llm-inference-service] # See issue #651 [1] for details. [e2e-llm-inference-service] # [e2e-llm-inference-service] # [1] [e2e-llm-inference-service] release_this_conn = release_conn [e2e-llm-inference-service] [e2e-llm-inference-service] http_tunnel_required = connection_requires_http_tunnel( [e2e-llm-inference-service] self.proxy, self.proxy_config, destination_scheme [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] # Merge the proxy headers. Only done when not using HTTP CONNECT. We [e2e-llm-inference-service] # have to copy the headers dict so we can safely change it without those [e2e-llm-inference-service] # changes being reflected in anyone else's copy. [e2e-llm-inference-service] if not http_tunnel_required: [e2e-llm-inference-service] headers = headers.copy() # type: ignore[attr-defined] [e2e-llm-inference-service] headers.update(self.proxy_headers) # type: ignore[union-attr] [e2e-llm-inference-service] [e2e-llm-inference-service] # Must keep the exception bound to a separate variable or else Python 3 [e2e-llm-inference-service] # complains about UnboundLocalError. [e2e-llm-inference-service] err = None [e2e-llm-inference-service] [e2e-llm-inference-service] # Keep track of whether we cleanly exited the except block. This [e2e-llm-inference-service] # ensures we do proper cleanup in finally. [e2e-llm-inference-service] clean_exit = False [e2e-llm-inference-service] [e2e-llm-inference-service] # Rewind body position, if needed. Record current position [e2e-llm-inference-service] # for future rewinds in the event of a redirect/retry. [e2e-llm-inference-service] body_pos = set_file_position(body, body_pos) [e2e-llm-inference-service] [e2e-llm-inference-service] try: [e2e-llm-inference-service] # Request a connection from the queue. [e2e-llm-inference-service] timeout_obj = self._get_timeout(timeout) [e2e-llm-inference-service] conn = self._get_conn(timeout=pool_timeout) [e2e-llm-inference-service] [e2e-llm-inference-service] conn.timeout = timeout_obj.connect_timeout # type: ignore[assignment] [e2e-llm-inference-service] [e2e-llm-inference-service] # Is this a closed/new connection that requires CONNECT tunnelling? [e2e-llm-inference-service] if self.proxy is not None and http_tunnel_required and conn.is_closed: [e2e-llm-inference-service] try: [e2e-llm-inference-service] self._prepare_proxy(conn) [e2e-llm-inference-service] except (BaseSSLError, OSError, SocketTimeout) as e: [e2e-llm-inference-service] self._raise_timeout( [e2e-llm-inference-service] err=e, url=self.proxy.url, timeout_value=conn.timeout [e2e-llm-inference-service] ) [e2e-llm-inference-service] raise [e2e-llm-inference-service] [e2e-llm-inference-service] # If we're going to release the connection in ``finally:``, then [e2e-llm-inference-service] # the response doesn't need to know about the connection. Otherwise [e2e-llm-inference-service] # it will also try to release it and we'll have a double-release [e2e-llm-inference-service] # mess. [e2e-llm-inference-service] response_conn = conn if not release_conn else None [e2e-llm-inference-service] [e2e-llm-inference-service] # Make the request on the HTTPConnection object [e2e-llm-inference-service] response = self._make_request( [e2e-llm-inference-service] conn, [e2e-llm-inference-service] method, [e2e-llm-inference-service] url, [e2e-llm-inference-service] timeout=timeout_obj, [e2e-llm-inference-service] body=body, [e2e-llm-inference-service] headers=headers, [e2e-llm-inference-service] chunked=chunked, [e2e-llm-inference-service] retries=retries, [e2e-llm-inference-service] response_conn=response_conn, [e2e-llm-inference-service] preload_content=preload_content, [e2e-llm-inference-service] decode_content=decode_content, [e2e-llm-inference-service] **response_kw, [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] # Everything went great! [e2e-llm-inference-service] clean_exit = True [e2e-llm-inference-service] [e2e-llm-inference-service] except EmptyPoolError: [e2e-llm-inference-service] # Didn't get a connection from the pool, no need to clean up [e2e-llm-inference-service] clean_exit = True [e2e-llm-inference-service] release_this_conn = False [e2e-llm-inference-service] raise [e2e-llm-inference-service] [e2e-llm-inference-service] except ( [e2e-llm-inference-service] TimeoutError, [e2e-llm-inference-service] HTTPException, [e2e-llm-inference-service] OSError, [e2e-llm-inference-service] ProtocolError, [e2e-llm-inference-service] BaseSSLError, [e2e-llm-inference-service] SSLError, [e2e-llm-inference-service] CertificateError, [e2e-llm-inference-service] ProxyError, [e2e-llm-inference-service] ) as e: [e2e-llm-inference-service] # Discard the connection for these exceptions. It will be [e2e-llm-inference-service] # replaced during the next _get_conn() call. [e2e-llm-inference-service] clean_exit = False [e2e-llm-inference-service] new_e: Exception = e [e2e-llm-inference-service] if isinstance(e, (BaseSSLError, CertificateError)): [e2e-llm-inference-service] new_e = SSLError(e) [e2e-llm-inference-service] if isinstance( [e2e-llm-inference-service] new_e, [e2e-llm-inference-service] ( [e2e-llm-inference-service] OSError, [e2e-llm-inference-service] NewConnectionError, [e2e-llm-inference-service] TimeoutError, [e2e-llm-inference-service] SSLError, [e2e-llm-inference-service] HTTPException, [e2e-llm-inference-service] ), [e2e-llm-inference-service] ) and (conn and conn.proxy and not conn.has_connected_to_proxy): [e2e-llm-inference-service] new_e = _wrap_proxy_error(new_e, conn.proxy.scheme) [e2e-llm-inference-service] elif isinstance(new_e, (OSError, HTTPException)): [e2e-llm-inference-service] new_e = ProtocolError("Connection aborted.", new_e) [e2e-llm-inference-service] [e2e-llm-inference-service] retries = retries.increment( [e2e-llm-inference-service] method, url, error=new_e, _pool=self, _stacktrace=sys.exc_info()[2] [e2e-llm-inference-service] ) [e2e-llm-inference-service] retries.sleep() [e2e-llm-inference-service] [e2e-llm-inference-service] # Keep track of the error for the retry warning. [e2e-llm-inference-service] err = e [e2e-llm-inference-service] [e2e-llm-inference-service] finally: [e2e-llm-inference-service] if not clean_exit: [e2e-llm-inference-service] # We hit some kind of exception, handled or otherwise. We need [e2e-llm-inference-service] # to throw the connection away unless explicitly told not to. [e2e-llm-inference-service] # Close the connection, set the variable to None, and make sure [e2e-llm-inference-service] # we put the None back in the pool to avoid leaking it. [e2e-llm-inference-service] if conn: [e2e-llm-inference-service] conn.close() [e2e-llm-inference-service] conn = None [e2e-llm-inference-service] release_this_conn = True [e2e-llm-inference-service] [e2e-llm-inference-service] if release_this_conn: [e2e-llm-inference-service] # Put the connection back to be reused. If the connection is [e2e-llm-inference-service] # expired then it will be None, which will get replaced with a [e2e-llm-inference-service] # fresh connection during _get_conn. [e2e-llm-inference-service] self._put_conn(conn) [e2e-llm-inference-service] [e2e-llm-inference-service] if not conn: [e2e-llm-inference-service] # Try again [e2e-llm-inference-service] log.warning( [e2e-llm-inference-service] "Retrying (%r) after connection broken by '%r': %s", retries, err, url [e2e-llm-inference-service] ) [e2e-llm-inference-service] > return self.urlopen( [e2e-llm-inference-service] method, [e2e-llm-inference-service] url, [e2e-llm-inference-service] body, [e2e-llm-inference-service] headers, [e2e-llm-inference-service] retries, [e2e-llm-inference-service] redirect, [e2e-llm-inference-service] assert_same_host, [e2e-llm-inference-service] timeout=timeout, [e2e-llm-inference-service] pool_timeout=pool_timeout, [e2e-llm-inference-service] release_conn=release_conn, [e2e-llm-inference-service] chunked=chunked, [e2e-llm-inference-service] body_pos=body_pos, [e2e-llm-inference-service] preload_content=preload_content, [e2e-llm-inference-service] decode_content=decode_content, [e2e-llm-inference-service] **response_kw, [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] ../../python/kserve/.venv/lib64/python3.11/site-packages/urllib3/connectionpool.py:871: [e2e-llm-inference-service] _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ [e2e-llm-inference-service] [e2e-llm-inference-service] self = [e2e-llm-inference-service] method = 'POST' [e2e-llm-inference-service] url = '/kserve-ci-e2e-test/custom-route-timeout-test/v1/completions' [e2e-llm-inference-service] body = b'{"model": "facebook/opt-125m", "prompt": "KServe is a", "max_tokens": 20}' [e2e-llm-inference-service] headers = {'User-Agent': 'python-requests/2.32.3', 'Accept-Encoding': 'gzip, deflate', 'Accept': '*/*', 'Connection': 'keep-alive', 'Content-Type': 'application/json', 'Content-Length': '73'} [e2e-llm-inference-service] retries = Retry(total=4, connect=None, read=None, redirect=None, status=None) [e2e-llm-inference-service] redirect = False, assert_same_host = False [e2e-llm-inference-service] timeout = Timeout(connect=60, read=60, total=None), pool_timeout = None [e2e-llm-inference-service] release_conn = False, chunked = False, body_pos = None, preload_content = False [e2e-llm-inference-service] decode_content = False, response_kw = {} [e2e-llm-inference-service] parsed_url = Url(scheme=None, auth=None, host=None, port=None, path='/kserve-ci-e2e-test/custom-route-timeout-test/v1/completions', query=None, fragment=None) [e2e-llm-inference-service] destination_scheme = None, conn = None, release_this_conn = True [e2e-llm-inference-service] http_tunnel_required = False [e2e-llm-inference-service] err = ReadTimeoutError("HTTPConnectionPool(host='ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com', port=80): Read timed out. (read timeout=60)") [e2e-llm-inference-service] clean_exit = False [e2e-llm-inference-service] [e2e-llm-inference-service] def urlopen( # type: ignore[override] [e2e-llm-inference-service] self, [e2e-llm-inference-service] method: str, [e2e-llm-inference-service] url: str, [e2e-llm-inference-service] body: _TYPE_BODY | None = None, [e2e-llm-inference-service] headers: typing.Mapping[str, str] | None = None, [e2e-llm-inference-service] retries: Retry | bool | int | None = None, [e2e-llm-inference-service] redirect: bool = True, [e2e-llm-inference-service] assert_same_host: bool = True, [e2e-llm-inference-service] timeout: _TYPE_TIMEOUT = _DEFAULT_TIMEOUT, [e2e-llm-inference-service] pool_timeout: int | None = None, [e2e-llm-inference-service] release_conn: bool | None = None, [e2e-llm-inference-service] chunked: bool = False, [e2e-llm-inference-service] body_pos: _TYPE_BODY_POSITION | None = None, [e2e-llm-inference-service] preload_content: bool = True, [e2e-llm-inference-service] decode_content: bool = True, [e2e-llm-inference-service] **response_kw: typing.Any, [e2e-llm-inference-service] ) -> BaseHTTPResponse: [e2e-llm-inference-service] """ [e2e-llm-inference-service] Get a connection from the pool and perform an HTTP request. This is the [e2e-llm-inference-service] lowest level call for making a request, so you'll need to specify all [e2e-llm-inference-service] the raw details. [e2e-llm-inference-service] [e2e-llm-inference-service] .. note:: [e2e-llm-inference-service] [e2e-llm-inference-service] More commonly, it's appropriate to use a convenience method [e2e-llm-inference-service] such as :meth:`request`. [e2e-llm-inference-service] [e2e-llm-inference-service] .. note:: [e2e-llm-inference-service] [e2e-llm-inference-service] `release_conn` will only behave as expected if [e2e-llm-inference-service] `preload_content=False` because we want to make [e2e-llm-inference-service] `preload_content=False` the default behaviour someday soon without [e2e-llm-inference-service] breaking backwards compatibility. [e2e-llm-inference-service] [e2e-llm-inference-service] :param method: [e2e-llm-inference-service] HTTP request method (such as GET, POST, PUT, etc.) [e2e-llm-inference-service] [e2e-llm-inference-service] :param url: [e2e-llm-inference-service] The URL to perform the request on. [e2e-llm-inference-service] [e2e-llm-inference-service] :param body: [e2e-llm-inference-service] Data to send in the request body, either :class:`str`, :class:`bytes`, [e2e-llm-inference-service] an iterable of :class:`str`/:class:`bytes`, or a file-like object. [e2e-llm-inference-service] [e2e-llm-inference-service] :param headers: [e2e-llm-inference-service] Dictionary of custom headers to send, such as User-Agent, [e2e-llm-inference-service] If-None-Match, etc. If None, pool headers are used. If provided, [e2e-llm-inference-service] these headers completely replace any pool-specific headers. [e2e-llm-inference-service] [e2e-llm-inference-service] :param retries: [e2e-llm-inference-service] Configure the number of retries to allow before raising a [e2e-llm-inference-service] :class:`~urllib3.exceptions.MaxRetryError` exception. [e2e-llm-inference-service] [e2e-llm-inference-service] If ``None`` (default) will retry 3 times, see ``Retry.DEFAULT``. Pass a [e2e-llm-inference-service] :class:`~urllib3.util.retry.Retry` object for fine-grained control [e2e-llm-inference-service] over different types of retries. [e2e-llm-inference-service] Pass an integer number to retry connection errors that many times, [e2e-llm-inference-service] but no other types of errors. Pass zero to never retry. [e2e-llm-inference-service] [e2e-llm-inference-service] If ``False``, then retries are disabled and any exception is raised [e2e-llm-inference-service] immediately. Also, instead of raising a MaxRetryError on redirects, [e2e-llm-inference-service] the redirect response will be returned. [e2e-llm-inference-service] [e2e-llm-inference-service] :type retries: :class:`~urllib3.util.retry.Retry`, False, or an int. [e2e-llm-inference-service] [e2e-llm-inference-service] :param redirect: [e2e-llm-inference-service] If True, automatically handle redirects (status codes 301, 302, [e2e-llm-inference-service] 303, 307, 308). Each redirect counts as a retry. Disabling retries [e2e-llm-inference-service] will disable redirect, too. [e2e-llm-inference-service] [e2e-llm-inference-service] :param assert_same_host: [e2e-llm-inference-service] If ``True``, will make sure that the host of the pool requests is [e2e-llm-inference-service] consistent else will raise HostChangedError. When ``False``, you can [e2e-llm-inference-service] use the pool on an HTTP proxy and request foreign hosts. [e2e-llm-inference-service] [e2e-llm-inference-service] :param timeout: [e2e-llm-inference-service] If specified, overrides the default timeout for this one [e2e-llm-inference-service] request. It may be a float (in seconds) or an instance of [e2e-llm-inference-service] :class:`urllib3.util.Timeout`. [e2e-llm-inference-service] [e2e-llm-inference-service] :param pool_timeout: [e2e-llm-inference-service] If set and the pool is set to block=True, then this method will [e2e-llm-inference-service] block for ``pool_timeout`` seconds and raise EmptyPoolError if no [e2e-llm-inference-service] connection is available within the time period. [e2e-llm-inference-service] [e2e-llm-inference-service] :param bool preload_content: [e2e-llm-inference-service] If True, the response's body will be preloaded into memory. [e2e-llm-inference-service] [e2e-llm-inference-service] :param bool decode_content: [e2e-llm-inference-service] If True, will attempt to decode the body based on the [e2e-llm-inference-service] 'content-encoding' header. [e2e-llm-inference-service] [e2e-llm-inference-service] :param release_conn: [e2e-llm-inference-service] If False, then the urlopen call will not release the connection [e2e-llm-inference-service] back into the pool once a response is received (but will release if [e2e-llm-inference-service] you read the entire contents of the response such as when [e2e-llm-inference-service] `preload_content=True`). This is useful if you're not preloading [e2e-llm-inference-service] the response's content immediately. You will need to call [e2e-llm-inference-service] ``r.release_conn()`` on the response ``r`` to return the connection [e2e-llm-inference-service] back into the pool. If None, it takes the value of ``preload_content`` [e2e-llm-inference-service] which defaults to ``True``. [e2e-llm-inference-service] [e2e-llm-inference-service] :param bool chunked: [e2e-llm-inference-service] If True, urllib3 will send the body using chunked transfer [e2e-llm-inference-service] encoding. Otherwise, urllib3 will send the body using the standard [e2e-llm-inference-service] content-length form. Defaults to False. [e2e-llm-inference-service] [e2e-llm-inference-service] :param int body_pos: [e2e-llm-inference-service] Position to seek to in file-like body in the event of a retry or [e2e-llm-inference-service] redirect. Typically this won't need to be set because urllib3 will [e2e-llm-inference-service] auto-populate the value when needed. [e2e-llm-inference-service] """ [e2e-llm-inference-service] parsed_url = parse_url(url) [e2e-llm-inference-service] destination_scheme = parsed_url.scheme [e2e-llm-inference-service] [e2e-llm-inference-service] if headers is None: [e2e-llm-inference-service] headers = self.headers [e2e-llm-inference-service] [e2e-llm-inference-service] if not isinstance(retries, Retry): [e2e-llm-inference-service] retries = Retry.from_int(retries, redirect=redirect, default=self.retries) [e2e-llm-inference-service] [e2e-llm-inference-service] if release_conn is None: [e2e-llm-inference-service] release_conn = preload_content [e2e-llm-inference-service] [e2e-llm-inference-service] # Check host [e2e-llm-inference-service] if assert_same_host and not self.is_same_host(url): [e2e-llm-inference-service] raise HostChangedError(self, url, retries) [e2e-llm-inference-service] [e2e-llm-inference-service] # Ensure that the URL we're connecting to is properly encoded [e2e-llm-inference-service] if url.startswith("/"): [e2e-llm-inference-service] url = to_str(_encode_target(url)) [e2e-llm-inference-service] else: [e2e-llm-inference-service] url = to_str(parsed_url.url) [e2e-llm-inference-service] [e2e-llm-inference-service] conn = None [e2e-llm-inference-service] [e2e-llm-inference-service] # Track whether `conn` needs to be released before [e2e-llm-inference-service] # returning/raising/recursing. Update this variable if necessary, and [e2e-llm-inference-service] # leave `release_conn` constant throughout the function. That way, if [e2e-llm-inference-service] # the function recurses, the original value of `release_conn` will be [e2e-llm-inference-service] # passed down into the recursive call, and its value will be respected. [e2e-llm-inference-service] # [e2e-llm-inference-service] # See issue #651 [1] for details. [e2e-llm-inference-service] # [e2e-llm-inference-service] # [1] [e2e-llm-inference-service] release_this_conn = release_conn [e2e-llm-inference-service] [e2e-llm-inference-service] http_tunnel_required = connection_requires_http_tunnel( [e2e-llm-inference-service] self.proxy, self.proxy_config, destination_scheme [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] # Merge the proxy headers. Only done when not using HTTP CONNECT. We [e2e-llm-inference-service] # have to copy the headers dict so we can safely change it without those [e2e-llm-inference-service] # changes being reflected in anyone else's copy. [e2e-llm-inference-service] if not http_tunnel_required: [e2e-llm-inference-service] headers = headers.copy() # type: ignore[attr-defined] [e2e-llm-inference-service] headers.update(self.proxy_headers) # type: ignore[union-attr] [e2e-llm-inference-service] [e2e-llm-inference-service] # Must keep the exception bound to a separate variable or else Python 3 [e2e-llm-inference-service] # complains about UnboundLocalError. [e2e-llm-inference-service] err = None [e2e-llm-inference-service] [e2e-llm-inference-service] # Keep track of whether we cleanly exited the except block. This [e2e-llm-inference-service] # ensures we do proper cleanup in finally. [e2e-llm-inference-service] clean_exit = False [e2e-llm-inference-service] [e2e-llm-inference-service] # Rewind body position, if needed. Record current position [e2e-llm-inference-service] # for future rewinds in the event of a redirect/retry. [e2e-llm-inference-service] body_pos = set_file_position(body, body_pos) [e2e-llm-inference-service] [e2e-llm-inference-service] try: [e2e-llm-inference-service] # Request a connection from the queue. [e2e-llm-inference-service] timeout_obj = self._get_timeout(timeout) [e2e-llm-inference-service] conn = self._get_conn(timeout=pool_timeout) [e2e-llm-inference-service] [e2e-llm-inference-service] conn.timeout = timeout_obj.connect_timeout # type: ignore[assignment] [e2e-llm-inference-service] [e2e-llm-inference-service] # Is this a closed/new connection that requires CONNECT tunnelling? [e2e-llm-inference-service] if self.proxy is not None and http_tunnel_required and conn.is_closed: [e2e-llm-inference-service] try: [e2e-llm-inference-service] self._prepare_proxy(conn) [e2e-llm-inference-service] except (BaseSSLError, OSError, SocketTimeout) as e: [e2e-llm-inference-service] self._raise_timeout( [e2e-llm-inference-service] err=e, url=self.proxy.url, timeout_value=conn.timeout [e2e-llm-inference-service] ) [e2e-llm-inference-service] raise [e2e-llm-inference-service] [e2e-llm-inference-service] # If we're going to release the connection in ``finally:``, then [e2e-llm-inference-service] # the response doesn't need to know about the connection. Otherwise [e2e-llm-inference-service] # it will also try to release it and we'll have a double-release [e2e-llm-inference-service] # mess. [e2e-llm-inference-service] response_conn = conn if not release_conn else None [e2e-llm-inference-service] [e2e-llm-inference-service] # Make the request on the HTTPConnection object [e2e-llm-inference-service] response = self._make_request( [e2e-llm-inference-service] conn, [e2e-llm-inference-service] method, [e2e-llm-inference-service] url, [e2e-llm-inference-service] timeout=timeout_obj, [e2e-llm-inference-service] body=body, [e2e-llm-inference-service] headers=headers, [e2e-llm-inference-service] chunked=chunked, [e2e-llm-inference-service] retries=retries, [e2e-llm-inference-service] response_conn=response_conn, [e2e-llm-inference-service] preload_content=preload_content, [e2e-llm-inference-service] decode_content=decode_content, [e2e-llm-inference-service] **response_kw, [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] # Everything went great! [e2e-llm-inference-service] clean_exit = True [e2e-llm-inference-service] [e2e-llm-inference-service] except EmptyPoolError: [e2e-llm-inference-service] # Didn't get a connection from the pool, no need to clean up [e2e-llm-inference-service] clean_exit = True [e2e-llm-inference-service] release_this_conn = False [e2e-llm-inference-service] raise [e2e-llm-inference-service] [e2e-llm-inference-service] except ( [e2e-llm-inference-service] TimeoutError, [e2e-llm-inference-service] HTTPException, [e2e-llm-inference-service] OSError, [e2e-llm-inference-service] ProtocolError, [e2e-llm-inference-service] BaseSSLError, [e2e-llm-inference-service] SSLError, [e2e-llm-inference-service] CertificateError, [e2e-llm-inference-service] ProxyError, [e2e-llm-inference-service] ) as e: [e2e-llm-inference-service] # Discard the connection for these exceptions. It will be [e2e-llm-inference-service] # replaced during the next _get_conn() call. [e2e-llm-inference-service] clean_exit = False [e2e-llm-inference-service] new_e: Exception = e [e2e-llm-inference-service] if isinstance(e, (BaseSSLError, CertificateError)): [e2e-llm-inference-service] new_e = SSLError(e) [e2e-llm-inference-service] if isinstance( [e2e-llm-inference-service] new_e, [e2e-llm-inference-service] ( [e2e-llm-inference-service] OSError, [e2e-llm-inference-service] NewConnectionError, [e2e-llm-inference-service] TimeoutError, [e2e-llm-inference-service] SSLError, [e2e-llm-inference-service] HTTPException, [e2e-llm-inference-service] ), [e2e-llm-inference-service] ) and (conn and conn.proxy and not conn.has_connected_to_proxy): [e2e-llm-inference-service] new_e = _wrap_proxy_error(new_e, conn.proxy.scheme) [e2e-llm-inference-service] elif isinstance(new_e, (OSError, HTTPException)): [e2e-llm-inference-service] new_e = ProtocolError("Connection aborted.", new_e) [e2e-llm-inference-service] [e2e-llm-inference-service] retries = retries.increment( [e2e-llm-inference-service] method, url, error=new_e, _pool=self, _stacktrace=sys.exc_info()[2] [e2e-llm-inference-service] ) [e2e-llm-inference-service] retries.sleep() [e2e-llm-inference-service] [e2e-llm-inference-service] # Keep track of the error for the retry warning. [e2e-llm-inference-service] err = e [e2e-llm-inference-service] [e2e-llm-inference-service] finally: [e2e-llm-inference-service] if not clean_exit: [e2e-llm-inference-service] # We hit some kind of exception, handled or otherwise. We need [e2e-llm-inference-service] # to throw the connection away unless explicitly told not to. [e2e-llm-inference-service] # Close the connection, set the variable to None, and make sure [e2e-llm-inference-service] # we put the None back in the pool to avoid leaking it. [e2e-llm-inference-service] if conn: [e2e-llm-inference-service] conn.close() [e2e-llm-inference-service] conn = None [e2e-llm-inference-service] release_this_conn = True [e2e-llm-inference-service] [e2e-llm-inference-service] if release_this_conn: [e2e-llm-inference-service] # Put the connection back to be reused. If the connection is [e2e-llm-inference-service] # expired then it will be None, which will get replaced with a [e2e-llm-inference-service] # fresh connection during _get_conn. [e2e-llm-inference-service] self._put_conn(conn) [e2e-llm-inference-service] [e2e-llm-inference-service] if not conn: [e2e-llm-inference-service] # Try again [e2e-llm-inference-service] log.warning( [e2e-llm-inference-service] "Retrying (%r) after connection broken by '%r': %s", retries, err, url [e2e-llm-inference-service] ) [e2e-llm-inference-service] > return self.urlopen( [e2e-llm-inference-service] method, [e2e-llm-inference-service] url, [e2e-llm-inference-service] body, [e2e-llm-inference-service] headers, [e2e-llm-inference-service] retries, [e2e-llm-inference-service] redirect, [e2e-llm-inference-service] assert_same_host, [e2e-llm-inference-service] timeout=timeout, [e2e-llm-inference-service] pool_timeout=pool_timeout, [e2e-llm-inference-service] release_conn=release_conn, [e2e-llm-inference-service] chunked=chunked, [e2e-llm-inference-service] body_pos=body_pos, [e2e-llm-inference-service] preload_content=preload_content, [e2e-llm-inference-service] decode_content=decode_content, [e2e-llm-inference-service] **response_kw, [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] ../../python/kserve/.venv/lib64/python3.11/site-packages/urllib3/connectionpool.py:871: [e2e-llm-inference-service] _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ [e2e-llm-inference-service] [e2e-llm-inference-service] self = [e2e-llm-inference-service] method = 'POST' [e2e-llm-inference-service] url = '/kserve-ci-e2e-test/custom-route-timeout-test/v1/completions' [e2e-llm-inference-service] body = b'{"model": "facebook/opt-125m", "prompt": "KServe is a", "max_tokens": 20}' [e2e-llm-inference-service] headers = {'User-Agent': 'python-requests/2.32.3', 'Accept-Encoding': 'gzip, deflate', 'Accept': '*/*', 'Connection': 'keep-alive', 'Content-Type': 'application/json', 'Content-Length': '73'} [e2e-llm-inference-service] retries = Retry(total=3, connect=None, read=None, redirect=None, status=None) [e2e-llm-inference-service] redirect = False, assert_same_host = False [e2e-llm-inference-service] timeout = Timeout(connect=60, read=60, total=None), pool_timeout = None [e2e-llm-inference-service] release_conn = False, chunked = False, body_pos = None, preload_content = False [e2e-llm-inference-service] decode_content = False, response_kw = {} [e2e-llm-inference-service] parsed_url = Url(scheme=None, auth=None, host=None, port=None, path='/kserve-ci-e2e-test/custom-route-timeout-test/v1/completions', query=None, fragment=None) [e2e-llm-inference-service] destination_scheme = None, conn = None, release_this_conn = True [e2e-llm-inference-service] http_tunnel_required = False [e2e-llm-inference-service] err = ReadTimeoutError("HTTPConnectionPool(host='ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com', port=80): Read timed out. (read timeout=60)") [e2e-llm-inference-service] clean_exit = False [e2e-llm-inference-service] [e2e-llm-inference-service] def urlopen( # type: ignore[override] [e2e-llm-inference-service] self, [e2e-llm-inference-service] method: str, [e2e-llm-inference-service] url: str, [e2e-llm-inference-service] body: _TYPE_BODY | None = None, [e2e-llm-inference-service] headers: typing.Mapping[str, str] | None = None, [e2e-llm-inference-service] retries: Retry | bool | int | None = None, [e2e-llm-inference-service] redirect: bool = True, [e2e-llm-inference-service] assert_same_host: bool = True, [e2e-llm-inference-service] timeout: _TYPE_TIMEOUT = _DEFAULT_TIMEOUT, [e2e-llm-inference-service] pool_timeout: int | None = None, [e2e-llm-inference-service] release_conn: bool | None = None, [e2e-llm-inference-service] chunked: bool = False, [e2e-llm-inference-service] body_pos: _TYPE_BODY_POSITION | None = None, [e2e-llm-inference-service] preload_content: bool = True, [e2e-llm-inference-service] decode_content: bool = True, [e2e-llm-inference-service] **response_kw: typing.Any, [e2e-llm-inference-service] ) -> BaseHTTPResponse: [e2e-llm-inference-service] """ [e2e-llm-inference-service] Get a connection from the pool and perform an HTTP request. This is the [e2e-llm-inference-service] lowest level call for making a request, so you'll need to specify all [e2e-llm-inference-service] the raw details. [e2e-llm-inference-service] [e2e-llm-inference-service] .. note:: [e2e-llm-inference-service] [e2e-llm-inference-service] More commonly, it's appropriate to use a convenience method [e2e-llm-inference-service] such as :meth:`request`. [e2e-llm-inference-service] [e2e-llm-inference-service] .. note:: [e2e-llm-inference-service] [e2e-llm-inference-service] `release_conn` will only behave as expected if [e2e-llm-inference-service] `preload_content=False` because we want to make [e2e-llm-inference-service] `preload_content=False` the default behaviour someday soon without [e2e-llm-inference-service] breaking backwards compatibility. [e2e-llm-inference-service] [e2e-llm-inference-service] :param method: [e2e-llm-inference-service] HTTP request method (such as GET, POST, PUT, etc.) [e2e-llm-inference-service] [e2e-llm-inference-service] :param url: [e2e-llm-inference-service] The URL to perform the request on. [e2e-llm-inference-service] [e2e-llm-inference-service] :param body: [e2e-llm-inference-service] Data to send in the request body, either :class:`str`, :class:`bytes`, [e2e-llm-inference-service] an iterable of :class:`str`/:class:`bytes`, or a file-like object. [e2e-llm-inference-service] [e2e-llm-inference-service] :param headers: [e2e-llm-inference-service] Dictionary of custom headers to send, such as User-Agent, [e2e-llm-inference-service] If-None-Match, etc. If None, pool headers are used. If provided, [e2e-llm-inference-service] these headers completely replace any pool-specific headers. [e2e-llm-inference-service] [e2e-llm-inference-service] :param retries: [e2e-llm-inference-service] Configure the number of retries to allow before raising a [e2e-llm-inference-service] :class:`~urllib3.exceptions.MaxRetryError` exception. [e2e-llm-inference-service] [e2e-llm-inference-service] If ``None`` (default) will retry 3 times, see ``Retry.DEFAULT``. Pass a [e2e-llm-inference-service] :class:`~urllib3.util.retry.Retry` object for fine-grained control [e2e-llm-inference-service] over different types of retries. [e2e-llm-inference-service] Pass an integer number to retry connection errors that many times, [e2e-llm-inference-service] but no other types of errors. Pass zero to never retry. [e2e-llm-inference-service] [e2e-llm-inference-service] If ``False``, then retries are disabled and any exception is raised [e2e-llm-inference-service] immediately. Also, instead of raising a MaxRetryError on redirects, [e2e-llm-inference-service] the redirect response will be returned. [e2e-llm-inference-service] [e2e-llm-inference-service] :type retries: :class:`~urllib3.util.retry.Retry`, False, or an int. [e2e-llm-inference-service] [e2e-llm-inference-service] :param redirect: [e2e-llm-inference-service] If True, automatically handle redirects (status codes 301, 302, [e2e-llm-inference-service] 303, 307, 308). Each redirect counts as a retry. Disabling retries [e2e-llm-inference-service] will disable redirect, too. [e2e-llm-inference-service] [e2e-llm-inference-service] :param assert_same_host: [e2e-llm-inference-service] If ``True``, will make sure that the host of the pool requests is [e2e-llm-inference-service] consistent else will raise HostChangedError. When ``False``, you can [e2e-llm-inference-service] use the pool on an HTTP proxy and request foreign hosts. [e2e-llm-inference-service] [e2e-llm-inference-service] :param timeout: [e2e-llm-inference-service] If specified, overrides the default timeout for this one [e2e-llm-inference-service] request. It may be a float (in seconds) or an instance of [e2e-llm-inference-service] :class:`urllib3.util.Timeout`. [e2e-llm-inference-service] [e2e-llm-inference-service] :param pool_timeout: [e2e-llm-inference-service] If set and the pool is set to block=True, then this method will [e2e-llm-inference-service] block for ``pool_timeout`` seconds and raise EmptyPoolError if no [e2e-llm-inference-service] connection is available within the time period. [e2e-llm-inference-service] [e2e-llm-inference-service] :param bool preload_content: [e2e-llm-inference-service] If True, the response's body will be preloaded into memory. [e2e-llm-inference-service] [e2e-llm-inference-service] :param bool decode_content: [e2e-llm-inference-service] If True, will attempt to decode the body based on the [e2e-llm-inference-service] 'content-encoding' header. [e2e-llm-inference-service] [e2e-llm-inference-service] :param release_conn: [e2e-llm-inference-service] If False, then the urlopen call will not release the connection [e2e-llm-inference-service] back into the pool once a response is received (but will release if [e2e-llm-inference-service] you read the entire contents of the response such as when [e2e-llm-inference-service] `preload_content=True`). This is useful if you're not preloading [e2e-llm-inference-service] the response's content immediately. You will need to call [e2e-llm-inference-service] ``r.release_conn()`` on the response ``r`` to return the connection [e2e-llm-inference-service] back into the pool. If None, it takes the value of ``preload_content`` [e2e-llm-inference-service] which defaults to ``True``. [e2e-llm-inference-service] [e2e-llm-inference-service] :param bool chunked: [e2e-llm-inference-service] If True, urllib3 will send the body using chunked transfer [e2e-llm-inference-service] encoding. Otherwise, urllib3 will send the body using the standard [e2e-llm-inference-service] content-length form. Defaults to False. [e2e-llm-inference-service] [e2e-llm-inference-service] :param int body_pos: [e2e-llm-inference-service] Position to seek to in file-like body in the event of a retry or [e2e-llm-inference-service] redirect. Typically this won't need to be set because urllib3 will [e2e-llm-inference-service] auto-populate the value when needed. [e2e-llm-inference-service] """ [e2e-llm-inference-service] parsed_url = parse_url(url) [e2e-llm-inference-service] destination_scheme = parsed_url.scheme [e2e-llm-inference-service] [e2e-llm-inference-service] if headers is None: [e2e-llm-inference-service] headers = self.headers [e2e-llm-inference-service] [e2e-llm-inference-service] if not isinstance(retries, Retry): [e2e-llm-inference-service] retries = Retry.from_int(retries, redirect=redirect, default=self.retries) [e2e-llm-inference-service] [e2e-llm-inference-service] if release_conn is None: [e2e-llm-inference-service] release_conn = preload_content [e2e-llm-inference-service] [e2e-llm-inference-service] # Check host [e2e-llm-inference-service] if assert_same_host and not self.is_same_host(url): [e2e-llm-inference-service] raise HostChangedError(self, url, retries) [e2e-llm-inference-service] [e2e-llm-inference-service] # Ensure that the URL we're connecting to is properly encoded [e2e-llm-inference-service] if url.startswith("/"): [e2e-llm-inference-service] url = to_str(_encode_target(url)) [e2e-llm-inference-service] else: [e2e-llm-inference-service] url = to_str(parsed_url.url) [e2e-llm-inference-service] [e2e-llm-inference-service] conn = None [e2e-llm-inference-service] [e2e-llm-inference-service] # Track whether `conn` needs to be released before [e2e-llm-inference-service] # returning/raising/recursing. Update this variable if necessary, and [e2e-llm-inference-service] # leave `release_conn` constant throughout the function. That way, if [e2e-llm-inference-service] # the function recurses, the original value of `release_conn` will be [e2e-llm-inference-service] # passed down into the recursive call, and its value will be respected. [e2e-llm-inference-service] # [e2e-llm-inference-service] # See issue #651 [1] for details. [e2e-llm-inference-service] # [e2e-llm-inference-service] # [1] [e2e-llm-inference-service] release_this_conn = release_conn [e2e-llm-inference-service] [e2e-llm-inference-service] http_tunnel_required = connection_requires_http_tunnel( [e2e-llm-inference-service] self.proxy, self.proxy_config, destination_scheme [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] # Merge the proxy headers. Only done when not using HTTP CONNECT. We [e2e-llm-inference-service] # have to copy the headers dict so we can safely change it without those [e2e-llm-inference-service] # changes being reflected in anyone else's copy. [e2e-llm-inference-service] if not http_tunnel_required: [e2e-llm-inference-service] headers = headers.copy() # type: ignore[attr-defined] [e2e-llm-inference-service] headers.update(self.proxy_headers) # type: ignore[union-attr] [e2e-llm-inference-service] [e2e-llm-inference-service] # Must keep the exception bound to a separate variable or else Python 3 [e2e-llm-inference-service] # complains about UnboundLocalError. [e2e-llm-inference-service] err = None [e2e-llm-inference-service] [e2e-llm-inference-service] # Keep track of whether we cleanly exited the except block. This [e2e-llm-inference-service] # ensures we do proper cleanup in finally. [e2e-llm-inference-service] clean_exit = False [e2e-llm-inference-service] [e2e-llm-inference-service] # Rewind body position, if needed. Record current position [e2e-llm-inference-service] # for future rewinds in the event of a redirect/retry. [e2e-llm-inference-service] body_pos = set_file_position(body, body_pos) [e2e-llm-inference-service] [e2e-llm-inference-service] try: [e2e-llm-inference-service] # Request a connection from the queue. [e2e-llm-inference-service] timeout_obj = self._get_timeout(timeout) [e2e-llm-inference-service] conn = self._get_conn(timeout=pool_timeout) [e2e-llm-inference-service] [e2e-llm-inference-service] conn.timeout = timeout_obj.connect_timeout # type: ignore[assignment] [e2e-llm-inference-service] [e2e-llm-inference-service] # Is this a closed/new connection that requires CONNECT tunnelling? [e2e-llm-inference-service] if self.proxy is not None and http_tunnel_required and conn.is_closed: [e2e-llm-inference-service] try: [e2e-llm-inference-service] self._prepare_proxy(conn) [e2e-llm-inference-service] except (BaseSSLError, OSError, SocketTimeout) as e: [e2e-llm-inference-service] self._raise_timeout( [e2e-llm-inference-service] err=e, url=self.proxy.url, timeout_value=conn.timeout [e2e-llm-inference-service] ) [e2e-llm-inference-service] raise [e2e-llm-inference-service] [e2e-llm-inference-service] # If we're going to release the connection in ``finally:``, then [e2e-llm-inference-service] # the response doesn't need to know about the connection. Otherwise [e2e-llm-inference-service] # it will also try to release it and we'll have a double-release [e2e-llm-inference-service] # mess. [e2e-llm-inference-service] response_conn = conn if not release_conn else None [e2e-llm-inference-service] [e2e-llm-inference-service] # Make the request on the HTTPConnection object [e2e-llm-inference-service] response = self._make_request( [e2e-llm-inference-service] conn, [e2e-llm-inference-service] method, [e2e-llm-inference-service] url, [e2e-llm-inference-service] timeout=timeout_obj, [e2e-llm-inference-service] body=body, [e2e-llm-inference-service] headers=headers, [e2e-llm-inference-service] chunked=chunked, [e2e-llm-inference-service] retries=retries, [e2e-llm-inference-service] response_conn=response_conn, [e2e-llm-inference-service] preload_content=preload_content, [e2e-llm-inference-service] decode_content=decode_content, [e2e-llm-inference-service] **response_kw, [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] # Everything went great! [e2e-llm-inference-service] clean_exit = True [e2e-llm-inference-service] [e2e-llm-inference-service] except EmptyPoolError: [e2e-llm-inference-service] # Didn't get a connection from the pool, no need to clean up [e2e-llm-inference-service] clean_exit = True [e2e-llm-inference-service] release_this_conn = False [e2e-llm-inference-service] raise [e2e-llm-inference-service] [e2e-llm-inference-service] except ( [e2e-llm-inference-service] TimeoutError, [e2e-llm-inference-service] HTTPException, [e2e-llm-inference-service] OSError, [e2e-llm-inference-service] ProtocolError, [e2e-llm-inference-service] BaseSSLError, [e2e-llm-inference-service] SSLError, [e2e-llm-inference-service] CertificateError, [e2e-llm-inference-service] ProxyError, [e2e-llm-inference-service] ) as e: [e2e-llm-inference-service] # Discard the connection for these exceptions. It will be [e2e-llm-inference-service] # replaced during the next _get_conn() call. [e2e-llm-inference-service] clean_exit = False [e2e-llm-inference-service] new_e: Exception = e [e2e-llm-inference-service] if isinstance(e, (BaseSSLError, CertificateError)): [e2e-llm-inference-service] new_e = SSLError(e) [e2e-llm-inference-service] if isinstance( [e2e-llm-inference-service] new_e, [e2e-llm-inference-service] ( [e2e-llm-inference-service] OSError, [e2e-llm-inference-service] NewConnectionError, [e2e-llm-inference-service] TimeoutError, [e2e-llm-inference-service] SSLError, [e2e-llm-inference-service] HTTPException, [e2e-llm-inference-service] ), [e2e-llm-inference-service] ) and (conn and conn.proxy and not conn.has_connected_to_proxy): [e2e-llm-inference-service] new_e = _wrap_proxy_error(new_e, conn.proxy.scheme) [e2e-llm-inference-service] elif isinstance(new_e, (OSError, HTTPException)): [e2e-llm-inference-service] new_e = ProtocolError("Connection aborted.", new_e) [e2e-llm-inference-service] [e2e-llm-inference-service] retries = retries.increment( [e2e-llm-inference-service] method, url, error=new_e, _pool=self, _stacktrace=sys.exc_info()[2] [e2e-llm-inference-service] ) [e2e-llm-inference-service] retries.sleep() [e2e-llm-inference-service] [e2e-llm-inference-service] # Keep track of the error for the retry warning. [e2e-llm-inference-service] err = e [e2e-llm-inference-service] [e2e-llm-inference-service] finally: [e2e-llm-inference-service] if not clean_exit: [e2e-llm-inference-service] # We hit some kind of exception, handled or otherwise. We need [e2e-llm-inference-service] # to throw the connection away unless explicitly told not to. [e2e-llm-inference-service] # Close the connection, set the variable to None, and make sure [e2e-llm-inference-service] # we put the None back in the pool to avoid leaking it. [e2e-llm-inference-service] if conn: [e2e-llm-inference-service] conn.close() [e2e-llm-inference-service] conn = None [e2e-llm-inference-service] release_this_conn = True [e2e-llm-inference-service] [e2e-llm-inference-service] if release_this_conn: [e2e-llm-inference-service] # Put the connection back to be reused. If the connection is [e2e-llm-inference-service] # expired then it will be None, which will get replaced with a [e2e-llm-inference-service] # fresh connection during _get_conn. [e2e-llm-inference-service] self._put_conn(conn) [e2e-llm-inference-service] [e2e-llm-inference-service] if not conn: [e2e-llm-inference-service] # Try again [e2e-llm-inference-service] log.warning( [e2e-llm-inference-service] "Retrying (%r) after connection broken by '%r': %s", retries, err, url [e2e-llm-inference-service] ) [e2e-llm-inference-service] > return self.urlopen( [e2e-llm-inference-service] method, [e2e-llm-inference-service] url, [e2e-llm-inference-service] body, [e2e-llm-inference-service] headers, [e2e-llm-inference-service] retries, [e2e-llm-inference-service] redirect, [e2e-llm-inference-service] assert_same_host, [e2e-llm-inference-service] timeout=timeout, [e2e-llm-inference-service] pool_timeout=pool_timeout, [e2e-llm-inference-service] release_conn=release_conn, [e2e-llm-inference-service] chunked=chunked, [e2e-llm-inference-service] body_pos=body_pos, [e2e-llm-inference-service] preload_content=preload_content, [e2e-llm-inference-service] decode_content=decode_content, [e2e-llm-inference-service] **response_kw, [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] ../../python/kserve/.venv/lib64/python3.11/site-packages/urllib3/connectionpool.py:871: [e2e-llm-inference-service] _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ [e2e-llm-inference-service] [e2e-llm-inference-service] self = [e2e-llm-inference-service] method = 'POST' [e2e-llm-inference-service] url = '/kserve-ci-e2e-test/custom-route-timeout-test/v1/completions' [e2e-llm-inference-service] body = b'{"model": "facebook/opt-125m", "prompt": "KServe is a", "max_tokens": 20}' [e2e-llm-inference-service] headers = {'User-Agent': 'python-requests/2.32.3', 'Accept-Encoding': 'gzip, deflate', 'Accept': '*/*', 'Connection': 'keep-alive', 'Content-Type': 'application/json', 'Content-Length': '73'} [e2e-llm-inference-service] retries = Retry(total=2, connect=None, read=None, redirect=None, status=None) [e2e-llm-inference-service] redirect = False, assert_same_host = False [e2e-llm-inference-service] timeout = Timeout(connect=60, read=60, total=None), pool_timeout = None [e2e-llm-inference-service] release_conn = False, chunked = False, body_pos = None, preload_content = False [e2e-llm-inference-service] decode_content = False, response_kw = {} [e2e-llm-inference-service] parsed_url = Url(scheme=None, auth=None, host=None, port=None, path='/kserve-ci-e2e-test/custom-route-timeout-test/v1/completions', query=None, fragment=None) [e2e-llm-inference-service] destination_scheme = None, conn = None, release_this_conn = True [e2e-llm-inference-service] http_tunnel_required = False [e2e-llm-inference-service] err = ReadTimeoutError("HTTPConnectionPool(host='ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com', port=80): Read timed out. (read timeout=60)") [e2e-llm-inference-service] clean_exit = False [e2e-llm-inference-service] [e2e-llm-inference-service] def urlopen( # type: ignore[override] [e2e-llm-inference-service] self, [e2e-llm-inference-service] method: str, [e2e-llm-inference-service] url: str, [e2e-llm-inference-service] body: _TYPE_BODY | None = None, [e2e-llm-inference-service] headers: typing.Mapping[str, str] | None = None, [e2e-llm-inference-service] retries: Retry | bool | int | None = None, [e2e-llm-inference-service] redirect: bool = True, [e2e-llm-inference-service] assert_same_host: bool = True, [e2e-llm-inference-service] timeout: _TYPE_TIMEOUT = _DEFAULT_TIMEOUT, [e2e-llm-inference-service] pool_timeout: int | None = None, [e2e-llm-inference-service] release_conn: bool | None = None, [e2e-llm-inference-service] chunked: bool = False, [e2e-llm-inference-service] body_pos: _TYPE_BODY_POSITION | None = None, [e2e-llm-inference-service] preload_content: bool = True, [e2e-llm-inference-service] decode_content: bool = True, [e2e-llm-inference-service] **response_kw: typing.Any, [e2e-llm-inference-service] ) -> BaseHTTPResponse: [e2e-llm-inference-service] """ [e2e-llm-inference-service] Get a connection from the pool and perform an HTTP request. This is the [e2e-llm-inference-service] lowest level call for making a request, so you'll need to specify all [e2e-llm-inference-service] the raw details. [e2e-llm-inference-service] [e2e-llm-inference-service] .. note:: [e2e-llm-inference-service] [e2e-llm-inference-service] More commonly, it's appropriate to use a convenience method [e2e-llm-inference-service] such as :meth:`request`. [e2e-llm-inference-service] [e2e-llm-inference-service] .. note:: [e2e-llm-inference-service] [e2e-llm-inference-service] `release_conn` will only behave as expected if [e2e-llm-inference-service] `preload_content=False` because we want to make [e2e-llm-inference-service] `preload_content=False` the default behaviour someday soon without [e2e-llm-inference-service] breaking backwards compatibility. [e2e-llm-inference-service] [e2e-llm-inference-service] :param method: [e2e-llm-inference-service] HTTP request method (such as GET, POST, PUT, etc.) [e2e-llm-inference-service] [e2e-llm-inference-service] :param url: [e2e-llm-inference-service] The URL to perform the request on. [e2e-llm-inference-service] [e2e-llm-inference-service] :param body: [e2e-llm-inference-service] Data to send in the request body, either :class:`str`, :class:`bytes`, [e2e-llm-inference-service] an iterable of :class:`str`/:class:`bytes`, or a file-like object. [e2e-llm-inference-service] [e2e-llm-inference-service] :param headers: [e2e-llm-inference-service] Dictionary of custom headers to send, such as User-Agent, [e2e-llm-inference-service] If-None-Match, etc. If None, pool headers are used. If provided, [e2e-llm-inference-service] these headers completely replace any pool-specific headers. [e2e-llm-inference-service] [e2e-llm-inference-service] :param retries: [e2e-llm-inference-service] Configure the number of retries to allow before raising a [e2e-llm-inference-service] :class:`~urllib3.exceptions.MaxRetryError` exception. [e2e-llm-inference-service] [e2e-llm-inference-service] If ``None`` (default) will retry 3 times, see ``Retry.DEFAULT``. Pass a [e2e-llm-inference-service] :class:`~urllib3.util.retry.Retry` object for fine-grained control [e2e-llm-inference-service] over different types of retries. [e2e-llm-inference-service] Pass an integer number to retry connection errors that many times, [e2e-llm-inference-service] but no other types of errors. Pass zero to never retry. [e2e-llm-inference-service] [e2e-llm-inference-service] If ``False``, then retries are disabled and any exception is raised [e2e-llm-inference-service] immediately. Also, instead of raising a MaxRetryError on redirects, [e2e-llm-inference-service] the redirect response will be returned. [e2e-llm-inference-service] [e2e-llm-inference-service] :type retries: :class:`~urllib3.util.retry.Retry`, False, or an int. [e2e-llm-inference-service] [e2e-llm-inference-service] :param redirect: [e2e-llm-inference-service] If True, automatically handle redirects (status codes 301, 302, [e2e-llm-inference-service] 303, 307, 308). Each redirect counts as a retry. Disabling retries [e2e-llm-inference-service] will disable redirect, too. [e2e-llm-inference-service] [e2e-llm-inference-service] :param assert_same_host: [e2e-llm-inference-service] If ``True``, will make sure that the host of the pool requests is [e2e-llm-inference-service] consistent else will raise HostChangedError. When ``False``, you can [e2e-llm-inference-service] use the pool on an HTTP proxy and request foreign hosts. [e2e-llm-inference-service] [e2e-llm-inference-service] :param timeout: [e2e-llm-inference-service] If specified, overrides the default timeout for this one [e2e-llm-inference-service] request. It may be a float (in seconds) or an instance of [e2e-llm-inference-service] :class:`urllib3.util.Timeout`. [e2e-llm-inference-service] [e2e-llm-inference-service] :param pool_timeout: [e2e-llm-inference-service] If set and the pool is set to block=True, then this method will [e2e-llm-inference-service] block for ``pool_timeout`` seconds and raise EmptyPoolError if no [e2e-llm-inference-service] connection is available within the time period. [e2e-llm-inference-service] [e2e-llm-inference-service] :param bool preload_content: [e2e-llm-inference-service] If True, the response's body will be preloaded into memory. [e2e-llm-inference-service] [e2e-llm-inference-service] :param bool decode_content: [e2e-llm-inference-service] If True, will attempt to decode the body based on the [e2e-llm-inference-service] 'content-encoding' header. [e2e-llm-inference-service] [e2e-llm-inference-service] :param release_conn: [e2e-llm-inference-service] If False, then the urlopen call will not release the connection [e2e-llm-inference-service] back into the pool once a response is received (but will release if [e2e-llm-inference-service] you read the entire contents of the response such as when [e2e-llm-inference-service] `preload_content=True`). This is useful if you're not preloading [e2e-llm-inference-service] the response's content immediately. You will need to call [e2e-llm-inference-service] ``r.release_conn()`` on the response ``r`` to return the connection [e2e-llm-inference-service] back into the pool. If None, it takes the value of ``preload_content`` [e2e-llm-inference-service] which defaults to ``True``. [e2e-llm-inference-service] [e2e-llm-inference-service] :param bool chunked: [e2e-llm-inference-service] If True, urllib3 will send the body using chunked transfer [e2e-llm-inference-service] encoding. Otherwise, urllib3 will send the body using the standard [e2e-llm-inference-service] content-length form. Defaults to False. [e2e-llm-inference-service] [e2e-llm-inference-service] :param int body_pos: [e2e-llm-inference-service] Position to seek to in file-like body in the event of a retry or [e2e-llm-inference-service] redirect. Typically this won't need to be set because urllib3 will [e2e-llm-inference-service] auto-populate the value when needed. [e2e-llm-inference-service] """ [e2e-llm-inference-service] parsed_url = parse_url(url) [e2e-llm-inference-service] destination_scheme = parsed_url.scheme [e2e-llm-inference-service] [e2e-llm-inference-service] if headers is None: [e2e-llm-inference-service] headers = self.headers [e2e-llm-inference-service] [e2e-llm-inference-service] if not isinstance(retries, Retry): [e2e-llm-inference-service] retries = Retry.from_int(retries, redirect=redirect, default=self.retries) [e2e-llm-inference-service] [e2e-llm-inference-service] if release_conn is None: [e2e-llm-inference-service] release_conn = preload_content [e2e-llm-inference-service] [e2e-llm-inference-service] # Check host [e2e-llm-inference-service] if assert_same_host and not self.is_same_host(url): [e2e-llm-inference-service] raise HostChangedError(self, url, retries) [e2e-llm-inference-service] [e2e-llm-inference-service] # Ensure that the URL we're connecting to is properly encoded [e2e-llm-inference-service] if url.startswith("/"): [e2e-llm-inference-service] url = to_str(_encode_target(url)) [e2e-llm-inference-service] else: [e2e-llm-inference-service] url = to_str(parsed_url.url) [e2e-llm-inference-service] [e2e-llm-inference-service] conn = None [e2e-llm-inference-service] [e2e-llm-inference-service] # Track whether `conn` needs to be released before [e2e-llm-inference-service] # returning/raising/recursing. Update this variable if necessary, and [e2e-llm-inference-service] # leave `release_conn` constant throughout the function. That way, if [e2e-llm-inference-service] # the function recurses, the original value of `release_conn` will be [e2e-llm-inference-service] # passed down into the recursive call, and its value will be respected. [e2e-llm-inference-service] # [e2e-llm-inference-service] # See issue #651 [1] for details. [e2e-llm-inference-service] # [e2e-llm-inference-service] # [1] [e2e-llm-inference-service] release_this_conn = release_conn [e2e-llm-inference-service] [e2e-llm-inference-service] http_tunnel_required = connection_requires_http_tunnel( [e2e-llm-inference-service] self.proxy, self.proxy_config, destination_scheme [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] # Merge the proxy headers. Only done when not using HTTP CONNECT. We [e2e-llm-inference-service] # have to copy the headers dict so we can safely change it without those [e2e-llm-inference-service] # changes being reflected in anyone else's copy. [e2e-llm-inference-service] if not http_tunnel_required: [e2e-llm-inference-service] headers = headers.copy() # type: ignore[attr-defined] [e2e-llm-inference-service] headers.update(self.proxy_headers) # type: ignore[union-attr] [e2e-llm-inference-service] [e2e-llm-inference-service] # Must keep the exception bound to a separate variable or else Python 3 [e2e-llm-inference-service] # complains about UnboundLocalError. [e2e-llm-inference-service] err = None [e2e-llm-inference-service] [e2e-llm-inference-service] # Keep track of whether we cleanly exited the except block. This [e2e-llm-inference-service] # ensures we do proper cleanup in finally. [e2e-llm-inference-service] clean_exit = False [e2e-llm-inference-service] [e2e-llm-inference-service] # Rewind body position, if needed. Record current position [e2e-llm-inference-service] # for future rewinds in the event of a redirect/retry. [e2e-llm-inference-service] body_pos = set_file_position(body, body_pos) [e2e-llm-inference-service] [e2e-llm-inference-service] try: [e2e-llm-inference-service] # Request a connection from the queue. [e2e-llm-inference-service] timeout_obj = self._get_timeout(timeout) [e2e-llm-inference-service] conn = self._get_conn(timeout=pool_timeout) [e2e-llm-inference-service] [e2e-llm-inference-service] conn.timeout = timeout_obj.connect_timeout # type: ignore[assignment] [e2e-llm-inference-service] [e2e-llm-inference-service] # Is this a closed/new connection that requires CONNECT tunnelling? [e2e-llm-inference-service] if self.proxy is not None and http_tunnel_required and conn.is_closed: [e2e-llm-inference-service] try: [e2e-llm-inference-service] self._prepare_proxy(conn) [e2e-llm-inference-service] except (BaseSSLError, OSError, SocketTimeout) as e: [e2e-llm-inference-service] self._raise_timeout( [e2e-llm-inference-service] err=e, url=self.proxy.url, timeout_value=conn.timeout [e2e-llm-inference-service] ) [e2e-llm-inference-service] raise [e2e-llm-inference-service] [e2e-llm-inference-service] # If we're going to release the connection in ``finally:``, then [e2e-llm-inference-service] # the response doesn't need to know about the connection. Otherwise [e2e-llm-inference-service] # it will also try to release it and we'll have a double-release [e2e-llm-inference-service] # mess. [e2e-llm-inference-service] response_conn = conn if not release_conn else None [e2e-llm-inference-service] [e2e-llm-inference-service] # Make the request on the HTTPConnection object [e2e-llm-inference-service] response = self._make_request( [e2e-llm-inference-service] conn, [e2e-llm-inference-service] method, [e2e-llm-inference-service] url, [e2e-llm-inference-service] timeout=timeout_obj, [e2e-llm-inference-service] body=body, [e2e-llm-inference-service] headers=headers, [e2e-llm-inference-service] chunked=chunked, [e2e-llm-inference-service] retries=retries, [e2e-llm-inference-service] response_conn=response_conn, [e2e-llm-inference-service] preload_content=preload_content, [e2e-llm-inference-service] decode_content=decode_content, [e2e-llm-inference-service] **response_kw, [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] # Everything went great! [e2e-llm-inference-service] clean_exit = True [e2e-llm-inference-service] [e2e-llm-inference-service] except EmptyPoolError: [e2e-llm-inference-service] # Didn't get a connection from the pool, no need to clean up [e2e-llm-inference-service] clean_exit = True [e2e-llm-inference-service] release_this_conn = False [e2e-llm-inference-service] raise [e2e-llm-inference-service] [e2e-llm-inference-service] except ( [e2e-llm-inference-service] TimeoutError, [e2e-llm-inference-service] HTTPException, [e2e-llm-inference-service] OSError, [e2e-llm-inference-service] ProtocolError, [e2e-llm-inference-service] BaseSSLError, [e2e-llm-inference-service] SSLError, [e2e-llm-inference-service] CertificateError, [e2e-llm-inference-service] ProxyError, [e2e-llm-inference-service] ) as e: [e2e-llm-inference-service] # Discard the connection for these exceptions. It will be [e2e-llm-inference-service] # replaced during the next _get_conn() call. [e2e-llm-inference-service] clean_exit = False [e2e-llm-inference-service] new_e: Exception = e [e2e-llm-inference-service] if isinstance(e, (BaseSSLError, CertificateError)): [e2e-llm-inference-service] new_e = SSLError(e) [e2e-llm-inference-service] if isinstance( [e2e-llm-inference-service] new_e, [e2e-llm-inference-service] ( [e2e-llm-inference-service] OSError, [e2e-llm-inference-service] NewConnectionError, [e2e-llm-inference-service] TimeoutError, [e2e-llm-inference-service] SSLError, [e2e-llm-inference-service] HTTPException, [e2e-llm-inference-service] ), [e2e-llm-inference-service] ) and (conn and conn.proxy and not conn.has_connected_to_proxy): [e2e-llm-inference-service] new_e = _wrap_proxy_error(new_e, conn.proxy.scheme) [e2e-llm-inference-service] elif isinstance(new_e, (OSError, HTTPException)): [e2e-llm-inference-service] new_e = ProtocolError("Connection aborted.", new_e) [e2e-llm-inference-service] [e2e-llm-inference-service] retries = retries.increment( [e2e-llm-inference-service] method, url, error=new_e, _pool=self, _stacktrace=sys.exc_info()[2] [e2e-llm-inference-service] ) [e2e-llm-inference-service] retries.sleep() [e2e-llm-inference-service] [e2e-llm-inference-service] # Keep track of the error for the retry warning. [e2e-llm-inference-service] err = e [e2e-llm-inference-service] [e2e-llm-inference-service] finally: [e2e-llm-inference-service] if not clean_exit: [e2e-llm-inference-service] # We hit some kind of exception, handled or otherwise. We need [e2e-llm-inference-service] # to throw the connection away unless explicitly told not to. [e2e-llm-inference-service] # Close the connection, set the variable to None, and make sure [e2e-llm-inference-service] # we put the None back in the pool to avoid leaking it. [e2e-llm-inference-service] if conn: [e2e-llm-inference-service] conn.close() [e2e-llm-inference-service] conn = None [e2e-llm-inference-service] release_this_conn = True [e2e-llm-inference-service] [e2e-llm-inference-service] if release_this_conn: [e2e-llm-inference-service] # Put the connection back to be reused. If the connection is [e2e-llm-inference-service] # expired then it will be None, which will get replaced with a [e2e-llm-inference-service] # fresh connection during _get_conn. [e2e-llm-inference-service] self._put_conn(conn) [e2e-llm-inference-service] [e2e-llm-inference-service] if not conn: [e2e-llm-inference-service] # Try again [e2e-llm-inference-service] log.warning( [e2e-llm-inference-service] "Retrying (%r) after connection broken by '%r': %s", retries, err, url [e2e-llm-inference-service] ) [e2e-llm-inference-service] > return self.urlopen( [e2e-llm-inference-service] method, [e2e-llm-inference-service] url, [e2e-llm-inference-service] body, [e2e-llm-inference-service] headers, [e2e-llm-inference-service] retries, [e2e-llm-inference-service] redirect, [e2e-llm-inference-service] assert_same_host, [e2e-llm-inference-service] timeout=timeout, [e2e-llm-inference-service] pool_timeout=pool_timeout, [e2e-llm-inference-service] release_conn=release_conn, [e2e-llm-inference-service] chunked=chunked, [e2e-llm-inference-service] body_pos=body_pos, [e2e-llm-inference-service] preload_content=preload_content, [e2e-llm-inference-service] decode_content=decode_content, [e2e-llm-inference-service] **response_kw, [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] ../../python/kserve/.venv/lib64/python3.11/site-packages/urllib3/connectionpool.py:871: [e2e-llm-inference-service] _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ [e2e-llm-inference-service] [e2e-llm-inference-service] self = [e2e-llm-inference-service] method = 'POST' [e2e-llm-inference-service] url = '/kserve-ci-e2e-test/custom-route-timeout-test/v1/completions' [e2e-llm-inference-service] body = b'{"model": "facebook/opt-125m", "prompt": "KServe is a", "max_tokens": 20}' [e2e-llm-inference-service] headers = {'User-Agent': 'python-requests/2.32.3', 'Accept-Encoding': 'gzip, deflate', 'Accept': '*/*', 'Connection': 'keep-alive', 'Content-Type': 'application/json', 'Content-Length': '73'} [e2e-llm-inference-service] retries = Retry(total=1, connect=None, read=None, redirect=None, status=None) [e2e-llm-inference-service] redirect = False, assert_same_host = False [e2e-llm-inference-service] timeout = Timeout(connect=60, read=60, total=None), pool_timeout = None [e2e-llm-inference-service] release_conn = False, chunked = False, body_pos = None, preload_content = False [e2e-llm-inference-service] decode_content = False, response_kw = {} [e2e-llm-inference-service] parsed_url = Url(scheme=None, auth=None, host=None, port=None, path='/kserve-ci-e2e-test/custom-route-timeout-test/v1/completions', query=None, fragment=None) [e2e-llm-inference-service] destination_scheme = None, conn = None, release_this_conn = True [e2e-llm-inference-service] http_tunnel_required = False [e2e-llm-inference-service] err = ReadTimeoutError("HTTPConnectionPool(host='ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com', port=80): Read timed out. (read timeout=60)") [e2e-llm-inference-service] clean_exit = False [e2e-llm-inference-service] [e2e-llm-inference-service] def urlopen( # type: ignore[override] [e2e-llm-inference-service] self, [e2e-llm-inference-service] method: str, [e2e-llm-inference-service] url: str, [e2e-llm-inference-service] body: _TYPE_BODY | None = None, [e2e-llm-inference-service] headers: typing.Mapping[str, str] | None = None, [e2e-llm-inference-service] retries: Retry | bool | int | None = None, [e2e-llm-inference-service] redirect: bool = True, [e2e-llm-inference-service] assert_same_host: bool = True, [e2e-llm-inference-service] timeout: _TYPE_TIMEOUT = _DEFAULT_TIMEOUT, [e2e-llm-inference-service] pool_timeout: int | None = None, [e2e-llm-inference-service] release_conn: bool | None = None, [e2e-llm-inference-service] chunked: bool = False, [e2e-llm-inference-service] body_pos: _TYPE_BODY_POSITION | None = None, [e2e-llm-inference-service] preload_content: bool = True, [e2e-llm-inference-service] decode_content: bool = True, [e2e-llm-inference-service] **response_kw: typing.Any, [e2e-llm-inference-service] ) -> BaseHTTPResponse: [e2e-llm-inference-service] """ [e2e-llm-inference-service] Get a connection from the pool and perform an HTTP request. This is the [e2e-llm-inference-service] lowest level call for making a request, so you'll need to specify all [e2e-llm-inference-service] the raw details. [e2e-llm-inference-service] [e2e-llm-inference-service] .. note:: [e2e-llm-inference-service] [e2e-llm-inference-service] More commonly, it's appropriate to use a convenience method [e2e-llm-inference-service] such as :meth:`request`. [e2e-llm-inference-service] [e2e-llm-inference-service] .. note:: [e2e-llm-inference-service] [e2e-llm-inference-service] `release_conn` will only behave as expected if [e2e-llm-inference-service] `preload_content=False` because we want to make [e2e-llm-inference-service] `preload_content=False` the default behaviour someday soon without [e2e-llm-inference-service] breaking backwards compatibility. [e2e-llm-inference-service] [e2e-llm-inference-service] :param method: [e2e-llm-inference-service] HTTP request method (such as GET, POST, PUT, etc.) [e2e-llm-inference-service] [e2e-llm-inference-service] :param url: [e2e-llm-inference-service] The URL to perform the request on. [e2e-llm-inference-service] [e2e-llm-inference-service] :param body: [e2e-llm-inference-service] Data to send in the request body, either :class:`str`, :class:`bytes`, [e2e-llm-inference-service] an iterable of :class:`str`/:class:`bytes`, or a file-like object. [e2e-llm-inference-service] [e2e-llm-inference-service] :param headers: [e2e-llm-inference-service] Dictionary of custom headers to send, such as User-Agent, [e2e-llm-inference-service] If-None-Match, etc. If None, pool headers are used. If provided, [e2e-llm-inference-service] these headers completely replace any pool-specific headers. [e2e-llm-inference-service] [e2e-llm-inference-service] :param retries: [e2e-llm-inference-service] Configure the number of retries to allow before raising a [e2e-llm-inference-service] :class:`~urllib3.exceptions.MaxRetryError` exception. [e2e-llm-inference-service] [e2e-llm-inference-service] If ``None`` (default) will retry 3 times, see ``Retry.DEFAULT``. Pass a [e2e-llm-inference-service] :class:`~urllib3.util.retry.Retry` object for fine-grained control [e2e-llm-inference-service] over different types of retries. [e2e-llm-inference-service] Pass an integer number to retry connection errors that many times, [e2e-llm-inference-service] but no other types of errors. Pass zero to never retry. [e2e-llm-inference-service] [e2e-llm-inference-service] If ``False``, then retries are disabled and any exception is raised [e2e-llm-inference-service] immediately. Also, instead of raising a MaxRetryError on redirects, [e2e-llm-inference-service] the redirect response will be returned. [e2e-llm-inference-service] [e2e-llm-inference-service] :type retries: :class:`~urllib3.util.retry.Retry`, False, or an int. [e2e-llm-inference-service] [e2e-llm-inference-service] :param redirect: [e2e-llm-inference-service] If True, automatically handle redirects (status codes 301, 302, [e2e-llm-inference-service] 303, 307, 308). Each redirect counts as a retry. Disabling retries [e2e-llm-inference-service] will disable redirect, too. [e2e-llm-inference-service] [e2e-llm-inference-service] :param assert_same_host: [e2e-llm-inference-service] If ``True``, will make sure that the host of the pool requests is [e2e-llm-inference-service] consistent else will raise HostChangedError. When ``False``, you can [e2e-llm-inference-service] use the pool on an HTTP proxy and request foreign hosts. [e2e-llm-inference-service] [e2e-llm-inference-service] :param timeout: [e2e-llm-inference-service] If specified, overrides the default timeout for this one [e2e-llm-inference-service] request. It may be a float (in seconds) or an instance of [e2e-llm-inference-service] :class:`urllib3.util.Timeout`. [e2e-llm-inference-service] [e2e-llm-inference-service] :param pool_timeout: [e2e-llm-inference-service] If set and the pool is set to block=True, then this method will [e2e-llm-inference-service] block for ``pool_timeout`` seconds and raise EmptyPoolError if no [e2e-llm-inference-service] connection is available within the time period. [e2e-llm-inference-service] [e2e-llm-inference-service] :param bool preload_content: [e2e-llm-inference-service] If True, the response's body will be preloaded into memory. [e2e-llm-inference-service] [e2e-llm-inference-service] :param bool decode_content: [e2e-llm-inference-service] If True, will attempt to decode the body based on the [e2e-llm-inference-service] 'content-encoding' header. [e2e-llm-inference-service] [e2e-llm-inference-service] :param release_conn: [e2e-llm-inference-service] If False, then the urlopen call will not release the connection [e2e-llm-inference-service] back into the pool once a response is received (but will release if [e2e-llm-inference-service] you read the entire contents of the response such as when [e2e-llm-inference-service] `preload_content=True`). This is useful if you're not preloading [e2e-llm-inference-service] the response's content immediately. You will need to call [e2e-llm-inference-service] ``r.release_conn()`` on the response ``r`` to return the connection [e2e-llm-inference-service] back into the pool. If None, it takes the value of ``preload_content`` [e2e-llm-inference-service] which defaults to ``True``. [e2e-llm-inference-service] [e2e-llm-inference-service] :param bool chunked: [e2e-llm-inference-service] If True, urllib3 will send the body using chunked transfer [e2e-llm-inference-service] encoding. Otherwise, urllib3 will send the body using the standard [e2e-llm-inference-service] content-length form. Defaults to False. [e2e-llm-inference-service] [e2e-llm-inference-service] :param int body_pos: [e2e-llm-inference-service] Position to seek to in file-like body in the event of a retry or [e2e-llm-inference-service] redirect. Typically this won't need to be set because urllib3 will [e2e-llm-inference-service] auto-populate the value when needed. [e2e-llm-inference-service] """ [e2e-llm-inference-service] parsed_url = parse_url(url) [e2e-llm-inference-service] destination_scheme = parsed_url.scheme [e2e-llm-inference-service] [e2e-llm-inference-service] if headers is None: [e2e-llm-inference-service] headers = self.headers [e2e-llm-inference-service] [e2e-llm-inference-service] if not isinstance(retries, Retry): [e2e-llm-inference-service] retries = Retry.from_int(retries, redirect=redirect, default=self.retries) [e2e-llm-inference-service] [e2e-llm-inference-service] if release_conn is None: [e2e-llm-inference-service] release_conn = preload_content [e2e-llm-inference-service] [e2e-llm-inference-service] # Check host [e2e-llm-inference-service] if assert_same_host and not self.is_same_host(url): [e2e-llm-inference-service] raise HostChangedError(self, url, retries) [e2e-llm-inference-service] [e2e-llm-inference-service] # Ensure that the URL we're connecting to is properly encoded [e2e-llm-inference-service] if url.startswith("/"): [e2e-llm-inference-service] url = to_str(_encode_target(url)) [e2e-llm-inference-service] else: [e2e-llm-inference-service] url = to_str(parsed_url.url) [e2e-llm-inference-service] [e2e-llm-inference-service] conn = None [e2e-llm-inference-service] [e2e-llm-inference-service] # Track whether `conn` needs to be released before [e2e-llm-inference-service] # returning/raising/recursing. Update this variable if necessary, and [e2e-llm-inference-service] # leave `release_conn` constant throughout the function. That way, if [e2e-llm-inference-service] # the function recurses, the original value of `release_conn` will be [e2e-llm-inference-service] # passed down into the recursive call, and its value will be respected. [e2e-llm-inference-service] # [e2e-llm-inference-service] # See issue #651 [1] for details. [e2e-llm-inference-service] # [e2e-llm-inference-service] # [1] [e2e-llm-inference-service] release_this_conn = release_conn [e2e-llm-inference-service] [e2e-llm-inference-service] http_tunnel_required = connection_requires_http_tunnel( [e2e-llm-inference-service] self.proxy, self.proxy_config, destination_scheme [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] # Merge the proxy headers. Only done when not using HTTP CONNECT. We [e2e-llm-inference-service] # have to copy the headers dict so we can safely change it without those [e2e-llm-inference-service] # changes being reflected in anyone else's copy. [e2e-llm-inference-service] if not http_tunnel_required: [e2e-llm-inference-service] headers = headers.copy() # type: ignore[attr-defined] [e2e-llm-inference-service] headers.update(self.proxy_headers) # type: ignore[union-attr] [e2e-llm-inference-service] [e2e-llm-inference-service] # Must keep the exception bound to a separate variable or else Python 3 [e2e-llm-inference-service] # complains about UnboundLocalError. [e2e-llm-inference-service] err = None [e2e-llm-inference-service] [e2e-llm-inference-service] # Keep track of whether we cleanly exited the except block. This [e2e-llm-inference-service] # ensures we do proper cleanup in finally. [e2e-llm-inference-service] clean_exit = False [e2e-llm-inference-service] [e2e-llm-inference-service] # Rewind body position, if needed. Record current position [e2e-llm-inference-service] # for future rewinds in the event of a redirect/retry. [e2e-llm-inference-service] body_pos = set_file_position(body, body_pos) [e2e-llm-inference-service] [e2e-llm-inference-service] try: [e2e-llm-inference-service] # Request a connection from the queue. [e2e-llm-inference-service] timeout_obj = self._get_timeout(timeout) [e2e-llm-inference-service] conn = self._get_conn(timeout=pool_timeout) [e2e-llm-inference-service] [e2e-llm-inference-service] conn.timeout = timeout_obj.connect_timeout # type: ignore[assignment] [e2e-llm-inference-service] [e2e-llm-inference-service] # Is this a closed/new connection that requires CONNECT tunnelling? [e2e-llm-inference-service] if self.proxy is not None and http_tunnel_required and conn.is_closed: [e2e-llm-inference-service] try: [e2e-llm-inference-service] self._prepare_proxy(conn) [e2e-llm-inference-service] except (BaseSSLError, OSError, SocketTimeout) as e: [e2e-llm-inference-service] self._raise_timeout( [e2e-llm-inference-service] err=e, url=self.proxy.url, timeout_value=conn.timeout [e2e-llm-inference-service] ) [e2e-llm-inference-service] raise [e2e-llm-inference-service] [e2e-llm-inference-service] # If we're going to release the connection in ``finally:``, then [e2e-llm-inference-service] # the response doesn't need to know about the connection. Otherwise [e2e-llm-inference-service] # it will also try to release it and we'll have a double-release [e2e-llm-inference-service] # mess. [e2e-llm-inference-service] response_conn = conn if not release_conn else None [e2e-llm-inference-service] [e2e-llm-inference-service] # Make the request on the HTTPConnection object [e2e-llm-inference-service] response = self._make_request( [e2e-llm-inference-service] conn, [e2e-llm-inference-service] method, [e2e-llm-inference-service] url, [e2e-llm-inference-service] timeout=timeout_obj, [e2e-llm-inference-service] body=body, [e2e-llm-inference-service] headers=headers, [e2e-llm-inference-service] chunked=chunked, [e2e-llm-inference-service] retries=retries, [e2e-llm-inference-service] response_conn=response_conn, [e2e-llm-inference-service] preload_content=preload_content, [e2e-llm-inference-service] decode_content=decode_content, [e2e-llm-inference-service] **response_kw, [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] # Everything went great! [e2e-llm-inference-service] clean_exit = True [e2e-llm-inference-service] [e2e-llm-inference-service] except EmptyPoolError: [e2e-llm-inference-service] # Didn't get a connection from the pool, no need to clean up [e2e-llm-inference-service] clean_exit = True [e2e-llm-inference-service] release_this_conn = False [e2e-llm-inference-service] raise [e2e-llm-inference-service] [e2e-llm-inference-service] except ( [e2e-llm-inference-service] TimeoutError, [e2e-llm-inference-service] HTTPException, [e2e-llm-inference-service] OSError, [e2e-llm-inference-service] ProtocolError, [e2e-llm-inference-service] BaseSSLError, [e2e-llm-inference-service] SSLError, [e2e-llm-inference-service] CertificateError, [e2e-llm-inference-service] ProxyError, [e2e-llm-inference-service] ) as e: [e2e-llm-inference-service] # Discard the connection for these exceptions. It will be [e2e-llm-inference-service] # replaced during the next _get_conn() call. [e2e-llm-inference-service] clean_exit = False [e2e-llm-inference-service] new_e: Exception = e [e2e-llm-inference-service] if isinstance(e, (BaseSSLError, CertificateError)): [e2e-llm-inference-service] new_e = SSLError(e) [e2e-llm-inference-service] if isinstance( [e2e-llm-inference-service] new_e, [e2e-llm-inference-service] ( [e2e-llm-inference-service] OSError, [e2e-llm-inference-service] NewConnectionError, [e2e-llm-inference-service] TimeoutError, [e2e-llm-inference-service] SSLError, [e2e-llm-inference-service] HTTPException, [e2e-llm-inference-service] ), [e2e-llm-inference-service] ) and (conn and conn.proxy and not conn.has_connected_to_proxy): [e2e-llm-inference-service] new_e = _wrap_proxy_error(new_e, conn.proxy.scheme) [e2e-llm-inference-service] elif isinstance(new_e, (OSError, HTTPException)): [e2e-llm-inference-service] new_e = ProtocolError("Connection aborted.", new_e) [e2e-llm-inference-service] [e2e-llm-inference-service] retries = retries.increment( [e2e-llm-inference-service] method, url, error=new_e, _pool=self, _stacktrace=sys.exc_info()[2] [e2e-llm-inference-service] ) [e2e-llm-inference-service] retries.sleep() [e2e-llm-inference-service] [e2e-llm-inference-service] # Keep track of the error for the retry warning. [e2e-llm-inference-service] err = e [e2e-llm-inference-service] [e2e-llm-inference-service] finally: [e2e-llm-inference-service] if not clean_exit: [e2e-llm-inference-service] # We hit some kind of exception, handled or otherwise. We need [e2e-llm-inference-service] # to throw the connection away unless explicitly told not to. [e2e-llm-inference-service] # Close the connection, set the variable to None, and make sure [e2e-llm-inference-service] # we put the None back in the pool to avoid leaking it. [e2e-llm-inference-service] if conn: [e2e-llm-inference-service] conn.close() [e2e-llm-inference-service] conn = None [e2e-llm-inference-service] release_this_conn = True [e2e-llm-inference-service] [e2e-llm-inference-service] if release_this_conn: [e2e-llm-inference-service] # Put the connection back to be reused. If the connection is [e2e-llm-inference-service] # expired then it will be None, which will get replaced with a [e2e-llm-inference-service] # fresh connection during _get_conn. [e2e-llm-inference-service] self._put_conn(conn) [e2e-llm-inference-service] [e2e-llm-inference-service] if not conn: [e2e-llm-inference-service] # Try again [e2e-llm-inference-service] log.warning( [e2e-llm-inference-service] "Retrying (%r) after connection broken by '%r': %s", retries, err, url [e2e-llm-inference-service] ) [e2e-llm-inference-service] > return self.urlopen( [e2e-llm-inference-service] method, [e2e-llm-inference-service] url, [e2e-llm-inference-service] body, [e2e-llm-inference-service] headers, [e2e-llm-inference-service] retries, [e2e-llm-inference-service] redirect, [e2e-llm-inference-service] assert_same_host, [e2e-llm-inference-service] timeout=timeout, [e2e-llm-inference-service] pool_timeout=pool_timeout, [e2e-llm-inference-service] release_conn=release_conn, [e2e-llm-inference-service] chunked=chunked, [e2e-llm-inference-service] body_pos=body_pos, [e2e-llm-inference-service] preload_content=preload_content, [e2e-llm-inference-service] decode_content=decode_content, [e2e-llm-inference-service] **response_kw, [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] ../../python/kserve/.venv/lib64/python3.11/site-packages/urllib3/connectionpool.py:871: [e2e-llm-inference-service] _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ [e2e-llm-inference-service] [e2e-llm-inference-service] self = [e2e-llm-inference-service] method = 'POST' [e2e-llm-inference-service] url = '/kserve-ci-e2e-test/custom-route-timeout-test/v1/completions' [e2e-llm-inference-service] body = b'{"model": "facebook/opt-125m", "prompt": "KServe is a", "max_tokens": 20}' [e2e-llm-inference-service] headers = {'User-Agent': 'python-requests/2.32.3', 'Accept-Encoding': 'gzip, deflate', 'Accept': '*/*', 'Connection': 'keep-alive', 'Content-Type': 'application/json', 'Content-Length': '73'} [e2e-llm-inference-service] retries = Retry(total=0, connect=None, read=None, redirect=None, status=None) [e2e-llm-inference-service] redirect = False, assert_same_host = False [e2e-llm-inference-service] timeout = Timeout(connect=60, read=60, total=None), pool_timeout = None [e2e-llm-inference-service] release_conn = False, chunked = False, body_pos = None, preload_content = False [e2e-llm-inference-service] decode_content = False, response_kw = {} [e2e-llm-inference-service] parsed_url = Url(scheme=None, auth=None, host=None, port=None, path='/kserve-ci-e2e-test/custom-route-timeout-test/v1/completions', query=None, fragment=None) [e2e-llm-inference-service] destination_scheme = None, conn = None, release_this_conn = True [e2e-llm-inference-service] http_tunnel_required = False [e2e-llm-inference-service] err = RemoteDisconnected('Remote end closed connection without response') [e2e-llm-inference-service] clean_exit = False [e2e-llm-inference-service] [e2e-llm-inference-service] def urlopen( # type: ignore[override] [e2e-llm-inference-service] self, [e2e-llm-inference-service] method: str, [e2e-llm-inference-service] url: str, [e2e-llm-inference-service] body: _TYPE_BODY | None = None, [e2e-llm-inference-service] headers: typing.Mapping[str, str] | None = None, [e2e-llm-inference-service] retries: Retry | bool | int | None = None, [e2e-llm-inference-service] redirect: bool = True, [e2e-llm-inference-service] assert_same_host: bool = True, [e2e-llm-inference-service] timeout: _TYPE_TIMEOUT = _DEFAULT_TIMEOUT, [e2e-llm-inference-service] pool_timeout: int | None = None, [e2e-llm-inference-service] release_conn: bool | None = None, [e2e-llm-inference-service] chunked: bool = False, [e2e-llm-inference-service] body_pos: _TYPE_BODY_POSITION | None = None, [e2e-llm-inference-service] preload_content: bool = True, [e2e-llm-inference-service] decode_content: bool = True, [e2e-llm-inference-service] **response_kw: typing.Any, [e2e-llm-inference-service] ) -> BaseHTTPResponse: [e2e-llm-inference-service] """ [e2e-llm-inference-service] Get a connection from the pool and perform an HTTP request. This is the [e2e-llm-inference-service] lowest level call for making a request, so you'll need to specify all [e2e-llm-inference-service] the raw details. [e2e-llm-inference-service] [e2e-llm-inference-service] .. note:: [e2e-llm-inference-service] [e2e-llm-inference-service] More commonly, it's appropriate to use a convenience method [e2e-llm-inference-service] such as :meth:`request`. [e2e-llm-inference-service] [e2e-llm-inference-service] .. note:: [e2e-llm-inference-service] [e2e-llm-inference-service] `release_conn` will only behave as expected if [e2e-llm-inference-service] `preload_content=False` because we want to make [e2e-llm-inference-service] `preload_content=False` the default behaviour someday soon without [e2e-llm-inference-service] breaking backwards compatibility. [e2e-llm-inference-service] [e2e-llm-inference-service] :param method: [e2e-llm-inference-service] HTTP request method (such as GET, POST, PUT, etc.) [e2e-llm-inference-service] [e2e-llm-inference-service] :param url: [e2e-llm-inference-service] The URL to perform the request on. [e2e-llm-inference-service] [e2e-llm-inference-service] :param body: [e2e-llm-inference-service] Data to send in the request body, either :class:`str`, :class:`bytes`, [e2e-llm-inference-service] an iterable of :class:`str`/:class:`bytes`, or a file-like object. [e2e-llm-inference-service] [e2e-llm-inference-service] :param headers: [e2e-llm-inference-service] Dictionary of custom headers to send, such as User-Agent, [e2e-llm-inference-service] If-None-Match, etc. If None, pool headers are used. If provided, [e2e-llm-inference-service] these headers completely replace any pool-specific headers. [e2e-llm-inference-service] [e2e-llm-inference-service] :param retries: [e2e-llm-inference-service] Configure the number of retries to allow before raising a [e2e-llm-inference-service] :class:`~urllib3.exceptions.MaxRetryError` exception. [e2e-llm-inference-service] [e2e-llm-inference-service] If ``None`` (default) will retry 3 times, see ``Retry.DEFAULT``. Pass a [e2e-llm-inference-service] :class:`~urllib3.util.retry.Retry` object for fine-grained control [e2e-llm-inference-service] over different types of retries. [e2e-llm-inference-service] Pass an integer number to retry connection errors that many times, [e2e-llm-inference-service] but no other types of errors. Pass zero to never retry. [e2e-llm-inference-service] [e2e-llm-inference-service] If ``False``, then retries are disabled and any exception is raised [e2e-llm-inference-service] immediately. Also, instead of raising a MaxRetryError on redirects, [e2e-llm-inference-service] the redirect response will be returned. [e2e-llm-inference-service] [e2e-llm-inference-service] :type retries: :class:`~urllib3.util.retry.Retry`, False, or an int. [e2e-llm-inference-service] [e2e-llm-inference-service] :param redirect: [e2e-llm-inference-service] If True, automatically handle redirects (status codes 301, 302, [e2e-llm-inference-service] 303, 307, 308). Each redirect counts as a retry. Disabling retries [e2e-llm-inference-service] will disable redirect, too. [e2e-llm-inference-service] [e2e-llm-inference-service] :param assert_same_host: [e2e-llm-inference-service] If ``True``, will make sure that the host of the pool requests is [e2e-llm-inference-service] consistent else will raise HostChangedError. When ``False``, you can [e2e-llm-inference-service] use the pool on an HTTP proxy and request foreign hosts. [e2e-llm-inference-service] [e2e-llm-inference-service] :param timeout: [e2e-llm-inference-service] If specified, overrides the default timeout for this one [e2e-llm-inference-service] request. It may be a float (in seconds) or an instance of [e2e-llm-inference-service] :class:`urllib3.util.Timeout`. [e2e-llm-inference-service] [e2e-llm-inference-service] :param pool_timeout: [e2e-llm-inference-service] If set and the pool is set to block=True, then this method will [e2e-llm-inference-service] block for ``pool_timeout`` seconds and raise EmptyPoolError if no [e2e-llm-inference-service] connection is available within the time period. [e2e-llm-inference-service] [e2e-llm-inference-service] :param bool preload_content: [e2e-llm-inference-service] If True, the response's body will be preloaded into memory. [e2e-llm-inference-service] [e2e-llm-inference-service] :param bool decode_content: [e2e-llm-inference-service] If True, will attempt to decode the body based on the [e2e-llm-inference-service] 'content-encoding' header. [e2e-llm-inference-service] [e2e-llm-inference-service] :param release_conn: [e2e-llm-inference-service] If False, then the urlopen call will not release the connection [e2e-llm-inference-service] back into the pool once a response is received (but will release if [e2e-llm-inference-service] you read the entire contents of the response such as when [e2e-llm-inference-service] `preload_content=True`). This is useful if you're not preloading [e2e-llm-inference-service] the response's content immediately. You will need to call [e2e-llm-inference-service] ``r.release_conn()`` on the response ``r`` to return the connection [e2e-llm-inference-service] back into the pool. If None, it takes the value of ``preload_content`` [e2e-llm-inference-service] which defaults to ``True``. [e2e-llm-inference-service] [e2e-llm-inference-service] :param bool chunked: [e2e-llm-inference-service] If True, urllib3 will send the body using chunked transfer [e2e-llm-inference-service] encoding. Otherwise, urllib3 will send the body using the standard [e2e-llm-inference-service] content-length form. Defaults to False. [e2e-llm-inference-service] [e2e-llm-inference-service] :param int body_pos: [e2e-llm-inference-service] Position to seek to in file-like body in the event of a retry or [e2e-llm-inference-service] redirect. Typically this won't need to be set because urllib3 will [e2e-llm-inference-service] auto-populate the value when needed. [e2e-llm-inference-service] """ [e2e-llm-inference-service] parsed_url = parse_url(url) [e2e-llm-inference-service] destination_scheme = parsed_url.scheme [e2e-llm-inference-service] [e2e-llm-inference-service] if headers is None: [e2e-llm-inference-service] headers = self.headers [e2e-llm-inference-service] [e2e-llm-inference-service] if not isinstance(retries, Retry): [e2e-llm-inference-service] retries = Retry.from_int(retries, redirect=redirect, default=self.retries) [e2e-llm-inference-service] [e2e-llm-inference-service] if release_conn is None: [e2e-llm-inference-service] release_conn = preload_content [e2e-llm-inference-service] [e2e-llm-inference-service] # Check host [e2e-llm-inference-service] if assert_same_host and not self.is_same_host(url): [e2e-llm-inference-service] raise HostChangedError(self, url, retries) [e2e-llm-inference-service] [e2e-llm-inference-service] # Ensure that the URL we're connecting to is properly encoded [e2e-llm-inference-service] if url.startswith("/"): [e2e-llm-inference-service] url = to_str(_encode_target(url)) [e2e-llm-inference-service] else: [e2e-llm-inference-service] url = to_str(parsed_url.url) [e2e-llm-inference-service] [e2e-llm-inference-service] conn = None [e2e-llm-inference-service] [e2e-llm-inference-service] # Track whether `conn` needs to be released before [e2e-llm-inference-service] # returning/raising/recursing. Update this variable if necessary, and [e2e-llm-inference-service] # leave `release_conn` constant throughout the function. That way, if [e2e-llm-inference-service] # the function recurses, the original value of `release_conn` will be [e2e-llm-inference-service] # passed down into the recursive call, and its value will be respected. [e2e-llm-inference-service] # [e2e-llm-inference-service] # See issue #651 [1] for details. [e2e-llm-inference-service] # [e2e-llm-inference-service] # [1] [e2e-llm-inference-service] release_this_conn = release_conn [e2e-llm-inference-service] [e2e-llm-inference-service] http_tunnel_required = connection_requires_http_tunnel( [e2e-llm-inference-service] self.proxy, self.proxy_config, destination_scheme [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] # Merge the proxy headers. Only done when not using HTTP CONNECT. We [e2e-llm-inference-service] # have to copy the headers dict so we can safely change it without those [e2e-llm-inference-service] # changes being reflected in anyone else's copy. [e2e-llm-inference-service] if not http_tunnel_required: [e2e-llm-inference-service] headers = headers.copy() # type: ignore[attr-defined] [e2e-llm-inference-service] headers.update(self.proxy_headers) # type: ignore[union-attr] [e2e-llm-inference-service] [e2e-llm-inference-service] # Must keep the exception bound to a separate variable or else Python 3 [e2e-llm-inference-service] # complains about UnboundLocalError. [e2e-llm-inference-service] err = None [e2e-llm-inference-service] [e2e-llm-inference-service] # Keep track of whether we cleanly exited the except block. This [e2e-llm-inference-service] # ensures we do proper cleanup in finally. [e2e-llm-inference-service] clean_exit = False [e2e-llm-inference-service] [e2e-llm-inference-service] # Rewind body position, if needed. Record current position [e2e-llm-inference-service] # for future rewinds in the event of a redirect/retry. [e2e-llm-inference-service] body_pos = set_file_position(body, body_pos) [e2e-llm-inference-service] [e2e-llm-inference-service] try: [e2e-llm-inference-service] # Request a connection from the queue. [e2e-llm-inference-service] timeout_obj = self._get_timeout(timeout) [e2e-llm-inference-service] conn = self._get_conn(timeout=pool_timeout) [e2e-llm-inference-service] [e2e-llm-inference-service] conn.timeout = timeout_obj.connect_timeout # type: ignore[assignment] [e2e-llm-inference-service] [e2e-llm-inference-service] # Is this a closed/new connection that requires CONNECT tunnelling? [e2e-llm-inference-service] if self.proxy is not None and http_tunnel_required and conn.is_closed: [e2e-llm-inference-service] try: [e2e-llm-inference-service] self._prepare_proxy(conn) [e2e-llm-inference-service] except (BaseSSLError, OSError, SocketTimeout) as e: [e2e-llm-inference-service] self._raise_timeout( [e2e-llm-inference-service] err=e, url=self.proxy.url, timeout_value=conn.timeout [e2e-llm-inference-service] ) [e2e-llm-inference-service] raise [e2e-llm-inference-service] [e2e-llm-inference-service] # If we're going to release the connection in ``finally:``, then [e2e-llm-inference-service] # the response doesn't need to know about the connection. Otherwise [e2e-llm-inference-service] # it will also try to release it and we'll have a double-release [e2e-llm-inference-service] # mess. [e2e-llm-inference-service] response_conn = conn if not release_conn else None [e2e-llm-inference-service] [e2e-llm-inference-service] # Make the request on the HTTPConnection object [e2e-llm-inference-service] response = self._make_request( [e2e-llm-inference-service] conn, [e2e-llm-inference-service] method, [e2e-llm-inference-service] url, [e2e-llm-inference-service] timeout=timeout_obj, [e2e-llm-inference-service] body=body, [e2e-llm-inference-service] headers=headers, [e2e-llm-inference-service] chunked=chunked, [e2e-llm-inference-service] retries=retries, [e2e-llm-inference-service] response_conn=response_conn, [e2e-llm-inference-service] preload_content=preload_content, [e2e-llm-inference-service] decode_content=decode_content, [e2e-llm-inference-service] **response_kw, [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] # Everything went great! [e2e-llm-inference-service] clean_exit = True [e2e-llm-inference-service] [e2e-llm-inference-service] except EmptyPoolError: [e2e-llm-inference-service] # Didn't get a connection from the pool, no need to clean up [e2e-llm-inference-service] clean_exit = True [e2e-llm-inference-service] release_this_conn = False [e2e-llm-inference-service] raise [e2e-llm-inference-service] [e2e-llm-inference-service] except ( [e2e-llm-inference-service] TimeoutError, [e2e-llm-inference-service] HTTPException, [e2e-llm-inference-service] OSError, [e2e-llm-inference-service] ProtocolError, [e2e-llm-inference-service] BaseSSLError, [e2e-llm-inference-service] SSLError, [e2e-llm-inference-service] CertificateError, [e2e-llm-inference-service] ProxyError, [e2e-llm-inference-service] ) as e: [e2e-llm-inference-service] # Discard the connection for these exceptions. It will be [e2e-llm-inference-service] # replaced during the next _get_conn() call. [e2e-llm-inference-service] clean_exit = False [e2e-llm-inference-service] new_e: Exception = e [e2e-llm-inference-service] if isinstance(e, (BaseSSLError, CertificateError)): [e2e-llm-inference-service] new_e = SSLError(e) [e2e-llm-inference-service] if isinstance( [e2e-llm-inference-service] new_e, [e2e-llm-inference-service] ( [e2e-llm-inference-service] OSError, [e2e-llm-inference-service] NewConnectionError, [e2e-llm-inference-service] TimeoutError, [e2e-llm-inference-service] SSLError, [e2e-llm-inference-service] HTTPException, [e2e-llm-inference-service] ), [e2e-llm-inference-service] ) and (conn and conn.proxy and not conn.has_connected_to_proxy): [e2e-llm-inference-service] new_e = _wrap_proxy_error(new_e, conn.proxy.scheme) [e2e-llm-inference-service] elif isinstance(new_e, (OSError, HTTPException)): [e2e-llm-inference-service] new_e = ProtocolError("Connection aborted.", new_e) [e2e-llm-inference-service] [e2e-llm-inference-service] retries = retries.increment( [e2e-llm-inference-service] method, url, error=new_e, _pool=self, _stacktrace=sys.exc_info()[2] [e2e-llm-inference-service] ) [e2e-llm-inference-service] retries.sleep() [e2e-llm-inference-service] [e2e-llm-inference-service] # Keep track of the error for the retry warning. [e2e-llm-inference-service] err = e [e2e-llm-inference-service] [e2e-llm-inference-service] finally: [e2e-llm-inference-service] if not clean_exit: [e2e-llm-inference-service] # We hit some kind of exception, handled or otherwise. We need [e2e-llm-inference-service] # to throw the connection away unless explicitly told not to. [e2e-llm-inference-service] # Close the connection, set the variable to None, and make sure [e2e-llm-inference-service] # we put the None back in the pool to avoid leaking it. [e2e-llm-inference-service] if conn: [e2e-llm-inference-service] conn.close() [e2e-llm-inference-service] conn = None [e2e-llm-inference-service] release_this_conn = True [e2e-llm-inference-service] [e2e-llm-inference-service] if release_this_conn: [e2e-llm-inference-service] # Put the connection back to be reused. If the connection is [e2e-llm-inference-service] # expired then it will be None, which will get replaced with a [e2e-llm-inference-service] # fresh connection during _get_conn. [e2e-llm-inference-service] self._put_conn(conn) [e2e-llm-inference-service] [e2e-llm-inference-service] if not conn: [e2e-llm-inference-service] # Try again [e2e-llm-inference-service] log.warning( [e2e-llm-inference-service] "Retrying (%r) after connection broken by '%r': %s", retries, err, url [e2e-llm-inference-service] ) [e2e-llm-inference-service] > return self.urlopen( [e2e-llm-inference-service] method, [e2e-llm-inference-service] url, [e2e-llm-inference-service] body, [e2e-llm-inference-service] headers, [e2e-llm-inference-service] retries, [e2e-llm-inference-service] redirect, [e2e-llm-inference-service] assert_same_host, [e2e-llm-inference-service] timeout=timeout, [e2e-llm-inference-service] pool_timeout=pool_timeout, [e2e-llm-inference-service] release_conn=release_conn, [e2e-llm-inference-service] chunked=chunked, [e2e-llm-inference-service] body_pos=body_pos, [e2e-llm-inference-service] preload_content=preload_content, [e2e-llm-inference-service] decode_content=decode_content, [e2e-llm-inference-service] **response_kw, [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] ../../python/kserve/.venv/lib64/python3.11/site-packages/urllib3/connectionpool.py:871: [e2e-llm-inference-service] _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ [e2e-llm-inference-service] [e2e-llm-inference-service] self = [e2e-llm-inference-service] method = 'POST' [e2e-llm-inference-service] url = '/kserve-ci-e2e-test/custom-route-timeout-test/v1/completions' [e2e-llm-inference-service] body = b'{"model": "facebook/opt-125m", "prompt": "KServe is a", "max_tokens": 20}' [e2e-llm-inference-service] headers = {'User-Agent': 'python-requests/2.32.3', 'Accept-Encoding': 'gzip, deflate', 'Accept': '*/*', 'Connection': 'keep-alive', 'Content-Type': 'application/json', 'Content-Length': '73'} [e2e-llm-inference-service] retries = Retry(total=0, connect=None, read=None, redirect=None, status=None) [e2e-llm-inference-service] redirect = False, assert_same_host = False [e2e-llm-inference-service] timeout = Timeout(connect=60, read=60, total=None), pool_timeout = None [e2e-llm-inference-service] release_conn = False, chunked = False, body_pos = None, preload_content = False [e2e-llm-inference-service] decode_content = False, response_kw = {} [e2e-llm-inference-service] parsed_url = Url(scheme=None, auth=None, host=None, port=None, path='/kserve-ci-e2e-test/custom-route-timeout-test/v1/completions', query=None, fragment=None) [e2e-llm-inference-service] destination_scheme = None, conn = None, release_this_conn = True [e2e-llm-inference-service] http_tunnel_required = False, err = None, clean_exit = False [e2e-llm-inference-service] [e2e-llm-inference-service] def urlopen( # type: ignore[override] [e2e-llm-inference-service] self, [e2e-llm-inference-service] method: str, [e2e-llm-inference-service] url: str, [e2e-llm-inference-service] body: _TYPE_BODY | None = None, [e2e-llm-inference-service] headers: typing.Mapping[str, str] | None = None, [e2e-llm-inference-service] retries: Retry | bool | int | None = None, [e2e-llm-inference-service] redirect: bool = True, [e2e-llm-inference-service] assert_same_host: bool = True, [e2e-llm-inference-service] timeout: _TYPE_TIMEOUT = _DEFAULT_TIMEOUT, [e2e-llm-inference-service] pool_timeout: int | None = None, [e2e-llm-inference-service] release_conn: bool | None = None, [e2e-llm-inference-service] chunked: bool = False, [e2e-llm-inference-service] body_pos: _TYPE_BODY_POSITION | None = None, [e2e-llm-inference-service] preload_content: bool = True, [e2e-llm-inference-service] decode_content: bool = True, [e2e-llm-inference-service] **response_kw: typing.Any, [e2e-llm-inference-service] ) -> BaseHTTPResponse: [e2e-llm-inference-service] """ [e2e-llm-inference-service] Get a connection from the pool and perform an HTTP request. This is the [e2e-llm-inference-service] lowest level call for making a request, so you'll need to specify all [e2e-llm-inference-service] the raw details. [e2e-llm-inference-service] [e2e-llm-inference-service] .. note:: [e2e-llm-inference-service] [e2e-llm-inference-service] More commonly, it's appropriate to use a convenience method [e2e-llm-inference-service] such as :meth:`request`. [e2e-llm-inference-service] [e2e-llm-inference-service] .. note:: [e2e-llm-inference-service] [e2e-llm-inference-service] `release_conn` will only behave as expected if [e2e-llm-inference-service] `preload_content=False` because we want to make [e2e-llm-inference-service] `preload_content=False` the default behaviour someday soon without [e2e-llm-inference-service] breaking backwards compatibility. [e2e-llm-inference-service] [e2e-llm-inference-service] :param method: [e2e-llm-inference-service] HTTP request method (such as GET, POST, PUT, etc.) [e2e-llm-inference-service] [e2e-llm-inference-service] :param url: [e2e-llm-inference-service] The URL to perform the request on. [e2e-llm-inference-service] [e2e-llm-inference-service] :param body: [e2e-llm-inference-service] Data to send in the request body, either :class:`str`, :class:`bytes`, [e2e-llm-inference-service] an iterable of :class:`str`/:class:`bytes`, or a file-like object. [e2e-llm-inference-service] [e2e-llm-inference-service] :param headers: [e2e-llm-inference-service] Dictionary of custom headers to send, such as User-Agent, [e2e-llm-inference-service] If-None-Match, etc. If None, pool headers are used. If provided, [e2e-llm-inference-service] these headers completely replace any pool-specific headers. [e2e-llm-inference-service] [e2e-llm-inference-service] :param retries: [e2e-llm-inference-service] Configure the number of retries to allow before raising a [e2e-llm-inference-service] :class:`~urllib3.exceptions.MaxRetryError` exception. [e2e-llm-inference-service] [e2e-llm-inference-service] If ``None`` (default) will retry 3 times, see ``Retry.DEFAULT``. Pass a [e2e-llm-inference-service] :class:`~urllib3.util.retry.Retry` object for fine-grained control [e2e-llm-inference-service] over different types of retries. [e2e-llm-inference-service] Pass an integer number to retry connection errors that many times, [e2e-llm-inference-service] but no other types of errors. Pass zero to never retry. [e2e-llm-inference-service] [e2e-llm-inference-service] If ``False``, then retries are disabled and any exception is raised [e2e-llm-inference-service] immediately. Also, instead of raising a MaxRetryError on redirects, [e2e-llm-inference-service] the redirect response will be returned. [e2e-llm-inference-service] [e2e-llm-inference-service] :type retries: :class:`~urllib3.util.retry.Retry`, False, or an int. [e2e-llm-inference-service] [e2e-llm-inference-service] :param redirect: [e2e-llm-inference-service] If True, automatically handle redirects (status codes 301, 302, [e2e-llm-inference-service] 303, 307, 308). Each redirect counts as a retry. Disabling retries [e2e-llm-inference-service] will disable redirect, too. [e2e-llm-inference-service] [e2e-llm-inference-service] :param assert_same_host: [e2e-llm-inference-service] If ``True``, will make sure that the host of the pool requests is [e2e-llm-inference-service] consistent else will raise HostChangedError. When ``False``, you can [e2e-llm-inference-service] use the pool on an HTTP proxy and request foreign hosts. [e2e-llm-inference-service] [e2e-llm-inference-service] :param timeout: [e2e-llm-inference-service] If specified, overrides the default timeout for this one [e2e-llm-inference-service] request. It may be a float (in seconds) or an instance of [e2e-llm-inference-service] :class:`urllib3.util.Timeout`. [e2e-llm-inference-service] [e2e-llm-inference-service] :param pool_timeout: [e2e-llm-inference-service] If set and the pool is set to block=True, then this method will [e2e-llm-inference-service] block for ``pool_timeout`` seconds and raise EmptyPoolError if no [e2e-llm-inference-service] connection is available within the time period. [e2e-llm-inference-service] [e2e-llm-inference-service] :param bool preload_content: [e2e-llm-inference-service] If True, the response's body will be preloaded into memory. [e2e-llm-inference-service] [e2e-llm-inference-service] :param bool decode_content: [e2e-llm-inference-service] If True, will attempt to decode the body based on the [e2e-llm-inference-service] 'content-encoding' header. [e2e-llm-inference-service] [e2e-llm-inference-service] :param release_conn: [e2e-llm-inference-service] If False, then the urlopen call will not release the connection [e2e-llm-inference-service] back into the pool once a response is received (but will release if [e2e-llm-inference-service] you read the entire contents of the response such as when [e2e-llm-inference-service] `preload_content=True`). This is useful if you're not preloading [e2e-llm-inference-service] the response's content immediately. You will need to call [e2e-llm-inference-service] ``r.release_conn()`` on the response ``r`` to return the connection [e2e-llm-inference-service] back into the pool. If None, it takes the value of ``preload_content`` [e2e-llm-inference-service] which defaults to ``True``. [e2e-llm-inference-service] [e2e-llm-inference-service] :param bool chunked: [e2e-llm-inference-service] If True, urllib3 will send the body using chunked transfer [e2e-llm-inference-service] encoding. Otherwise, urllib3 will send the body using the standard [e2e-llm-inference-service] content-length form. Defaults to False. [e2e-llm-inference-service] [e2e-llm-inference-service] :param int body_pos: [e2e-llm-inference-service] Position to seek to in file-like body in the event of a retry or [e2e-llm-inference-service] redirect. Typically this won't need to be set because urllib3 will [e2e-llm-inference-service] auto-populate the value when needed. [e2e-llm-inference-service] """ [e2e-llm-inference-service] parsed_url = parse_url(url) [e2e-llm-inference-service] destination_scheme = parsed_url.scheme [e2e-llm-inference-service] [e2e-llm-inference-service] if headers is None: [e2e-llm-inference-service] headers = self.headers [e2e-llm-inference-service] [e2e-llm-inference-service] if not isinstance(retries, Retry): [e2e-llm-inference-service] retries = Retry.from_int(retries, redirect=redirect, default=self.retries) [e2e-llm-inference-service] [e2e-llm-inference-service] if release_conn is None: [e2e-llm-inference-service] release_conn = preload_content [e2e-llm-inference-service] [e2e-llm-inference-service] # Check host [e2e-llm-inference-service] if assert_same_host and not self.is_same_host(url): [e2e-llm-inference-service] raise HostChangedError(self, url, retries) [e2e-llm-inference-service] [e2e-llm-inference-service] # Ensure that the URL we're connecting to is properly encoded [e2e-llm-inference-service] if url.startswith("/"): [e2e-llm-inference-service] url = to_str(_encode_target(url)) [e2e-llm-inference-service] else: [e2e-llm-inference-service] url = to_str(parsed_url.url) [e2e-llm-inference-service] [e2e-llm-inference-service] conn = None [e2e-llm-inference-service] [e2e-llm-inference-service] # Track whether `conn` needs to be released before [e2e-llm-inference-service] # returning/raising/recursing. Update this variable if necessary, and [e2e-llm-inference-service] # leave `release_conn` constant throughout the function. That way, if [e2e-llm-inference-service] # the function recurses, the original value of `release_conn` will be [e2e-llm-inference-service] # passed down into the recursive call, and its value will be respected. [e2e-llm-inference-service] # [e2e-llm-inference-service] # See issue #651 [1] for details. [e2e-llm-inference-service] # [e2e-llm-inference-service] # [1] [e2e-llm-inference-service] release_this_conn = release_conn [e2e-llm-inference-service] [e2e-llm-inference-service] http_tunnel_required = connection_requires_http_tunnel( [e2e-llm-inference-service] self.proxy, self.proxy_config, destination_scheme [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] # Merge the proxy headers. Only done when not using HTTP CONNECT. We [e2e-llm-inference-service] # have to copy the headers dict so we can safely change it without those [e2e-llm-inference-service] # changes being reflected in anyone else's copy. [e2e-llm-inference-service] if not http_tunnel_required: [e2e-llm-inference-service] headers = headers.copy() # type: ignore[attr-defined] [e2e-llm-inference-service] headers.update(self.proxy_headers) # type: ignore[union-attr] [e2e-llm-inference-service] [e2e-llm-inference-service] # Must keep the exception bound to a separate variable or else Python 3 [e2e-llm-inference-service] # complains about UnboundLocalError. [e2e-llm-inference-service] err = None [e2e-llm-inference-service] [e2e-llm-inference-service] # Keep track of whether we cleanly exited the except block. This [e2e-llm-inference-service] # ensures we do proper cleanup in finally. [e2e-llm-inference-service] clean_exit = False [e2e-llm-inference-service] [e2e-llm-inference-service] # Rewind body position, if needed. Record current position [e2e-llm-inference-service] # for future rewinds in the event of a redirect/retry. [e2e-llm-inference-service] body_pos = set_file_position(body, body_pos) [e2e-llm-inference-service] [e2e-llm-inference-service] try: [e2e-llm-inference-service] # Request a connection from the queue. [e2e-llm-inference-service] timeout_obj = self._get_timeout(timeout) [e2e-llm-inference-service] conn = self._get_conn(timeout=pool_timeout) [e2e-llm-inference-service] [e2e-llm-inference-service] conn.timeout = timeout_obj.connect_timeout # type: ignore[assignment] [e2e-llm-inference-service] [e2e-llm-inference-service] # Is this a closed/new connection that requires CONNECT tunnelling? [e2e-llm-inference-service] if self.proxy is not None and http_tunnel_required and conn.is_closed: [e2e-llm-inference-service] try: [e2e-llm-inference-service] self._prepare_proxy(conn) [e2e-llm-inference-service] except (BaseSSLError, OSError, SocketTimeout) as e: [e2e-llm-inference-service] self._raise_timeout( [e2e-llm-inference-service] err=e, url=self.proxy.url, timeout_value=conn.timeout [e2e-llm-inference-service] ) [e2e-llm-inference-service] raise [e2e-llm-inference-service] [e2e-llm-inference-service] # If we're going to release the connection in ``finally:``, then [e2e-llm-inference-service] # the response doesn't need to know about the connection. Otherwise [e2e-llm-inference-service] # it will also try to release it and we'll have a double-release [e2e-llm-inference-service] # mess. [e2e-llm-inference-service] response_conn = conn if not release_conn else None [e2e-llm-inference-service] [e2e-llm-inference-service] # Make the request on the HTTPConnection object [e2e-llm-inference-service] response = self._make_request( [e2e-llm-inference-service] conn, [e2e-llm-inference-service] method, [e2e-llm-inference-service] url, [e2e-llm-inference-service] timeout=timeout_obj, [e2e-llm-inference-service] body=body, [e2e-llm-inference-service] headers=headers, [e2e-llm-inference-service] chunked=chunked, [e2e-llm-inference-service] retries=retries, [e2e-llm-inference-service] response_conn=response_conn, [e2e-llm-inference-service] preload_content=preload_content, [e2e-llm-inference-service] decode_content=decode_content, [e2e-llm-inference-service] **response_kw, [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] # Everything went great! [e2e-llm-inference-service] clean_exit = True [e2e-llm-inference-service] [e2e-llm-inference-service] except EmptyPoolError: [e2e-llm-inference-service] # Didn't get a connection from the pool, no need to clean up [e2e-llm-inference-service] clean_exit = True [e2e-llm-inference-service] release_this_conn = False [e2e-llm-inference-service] raise [e2e-llm-inference-service] [e2e-llm-inference-service] except ( [e2e-llm-inference-service] TimeoutError, [e2e-llm-inference-service] HTTPException, [e2e-llm-inference-service] OSError, [e2e-llm-inference-service] ProtocolError, [e2e-llm-inference-service] BaseSSLError, [e2e-llm-inference-service] SSLError, [e2e-llm-inference-service] CertificateError, [e2e-llm-inference-service] ProxyError, [e2e-llm-inference-service] ) as e: [e2e-llm-inference-service] # Discard the connection for these exceptions. It will be [e2e-llm-inference-service] # replaced during the next _get_conn() call. [e2e-llm-inference-service] clean_exit = False [e2e-llm-inference-service] new_e: Exception = e [e2e-llm-inference-service] if isinstance(e, (BaseSSLError, CertificateError)): [e2e-llm-inference-service] new_e = SSLError(e) [e2e-llm-inference-service] if isinstance( [e2e-llm-inference-service] new_e, [e2e-llm-inference-service] ( [e2e-llm-inference-service] OSError, [e2e-llm-inference-service] NewConnectionError, [e2e-llm-inference-service] TimeoutError, [e2e-llm-inference-service] SSLError, [e2e-llm-inference-service] HTTPException, [e2e-llm-inference-service] ), [e2e-llm-inference-service] ) and (conn and conn.proxy and not conn.has_connected_to_proxy): [e2e-llm-inference-service] new_e = _wrap_proxy_error(new_e, conn.proxy.scheme) [e2e-llm-inference-service] elif isinstance(new_e, (OSError, HTTPException)): [e2e-llm-inference-service] new_e = ProtocolError("Connection aborted.", new_e) [e2e-llm-inference-service] [e2e-llm-inference-service] > retries = retries.increment( [e2e-llm-inference-service] method, url, error=new_e, _pool=self, _stacktrace=sys.exc_info()[2] [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] ../../python/kserve/.venv/lib64/python3.11/site-packages/urllib3/connectionpool.py:841: [e2e-llm-inference-service] _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ [e2e-llm-inference-service] [e2e-llm-inference-service] self = Retry(total=0, connect=None, read=None, redirect=None, status=None) [e2e-llm-inference-service] method = 'POST' [e2e-llm-inference-service] url = '/kserve-ci-e2e-test/custom-route-timeout-test/v1/completions' [e2e-llm-inference-service] response = None [e2e-llm-inference-service] error = ReadTimeoutError("HTTPConnectionPool(host='ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com', port=80): Read timed out. (read timeout=60)") [e2e-llm-inference-service] _pool = [e2e-llm-inference-service] _stacktrace = [e2e-llm-inference-service] [e2e-llm-inference-service] def increment( [e2e-llm-inference-service] self, [e2e-llm-inference-service] method: str | None = None, [e2e-llm-inference-service] url: str | None = None, [e2e-llm-inference-service] response: BaseHTTPResponse | None = None, [e2e-llm-inference-service] error: Exception | None = None, [e2e-llm-inference-service] _pool: ConnectionPool | None = None, [e2e-llm-inference-service] _stacktrace: TracebackType | None = None, [e2e-llm-inference-service] ) -> Self: [e2e-llm-inference-service] """Return a new Retry object with incremented retry counters. [e2e-llm-inference-service] [e2e-llm-inference-service] :param response: A response object, or None, if the server did not [e2e-llm-inference-service] return a response. [e2e-llm-inference-service] :type response: :class:`~urllib3.response.BaseHTTPResponse` [e2e-llm-inference-service] :param Exception error: An error encountered during the request, or [e2e-llm-inference-service] None if the response was received successfully. [e2e-llm-inference-service] [e2e-llm-inference-service] :return: A new ``Retry`` object. [e2e-llm-inference-service] """ [e2e-llm-inference-service] if self.total is False and error: [e2e-llm-inference-service] # Disabled, indicate to re-raise the error. [e2e-llm-inference-service] raise reraise(type(error), error, _stacktrace) [e2e-llm-inference-service] [e2e-llm-inference-service] total = self.total [e2e-llm-inference-service] if total is not None: [e2e-llm-inference-service] total -= 1 [e2e-llm-inference-service] [e2e-llm-inference-service] connect = self.connect [e2e-llm-inference-service] read = self.read [e2e-llm-inference-service] redirect = self.redirect [e2e-llm-inference-service] status_count = self.status [e2e-llm-inference-service] other = self.other [e2e-llm-inference-service] cause = "unknown" [e2e-llm-inference-service] status = None [e2e-llm-inference-service] redirect_location = None [e2e-llm-inference-service] [e2e-llm-inference-service] if error and self._is_connection_error(error): [e2e-llm-inference-service] # Connect retry? [e2e-llm-inference-service] if connect is False: [e2e-llm-inference-service] raise reraise(type(error), error, _stacktrace) [e2e-llm-inference-service] elif connect is not None: [e2e-llm-inference-service] connect -= 1 [e2e-llm-inference-service] [e2e-llm-inference-service] elif error and self._is_read_error(error): [e2e-llm-inference-service] # Read retry? [e2e-llm-inference-service] if read is False or method is None or not self._is_method_retryable(method): [e2e-llm-inference-service] raise reraise(type(error), error, _stacktrace) [e2e-llm-inference-service] elif read is not None: [e2e-llm-inference-service] read -= 1 [e2e-llm-inference-service] [e2e-llm-inference-service] elif error: [e2e-llm-inference-service] # Other retry? [e2e-llm-inference-service] if other is not None: [e2e-llm-inference-service] other -= 1 [e2e-llm-inference-service] [e2e-llm-inference-service] elif response and response.get_redirect_location(): [e2e-llm-inference-service] # Redirect retry? [e2e-llm-inference-service] if redirect is not None: [e2e-llm-inference-service] redirect -= 1 [e2e-llm-inference-service] cause = "too many redirects" [e2e-llm-inference-service] response_redirect_location = response.get_redirect_location() [e2e-llm-inference-service] if response_redirect_location: [e2e-llm-inference-service] redirect_location = response_redirect_location [e2e-llm-inference-service] status = response.status [e2e-llm-inference-service] [e2e-llm-inference-service] else: [e2e-llm-inference-service] # Incrementing because of a server error like a 500 in [e2e-llm-inference-service] # status_forcelist and the given method is in the allowed_methods [e2e-llm-inference-service] cause = ResponseError.GENERIC_ERROR [e2e-llm-inference-service] if response and response.status: [e2e-llm-inference-service] if status_count is not None: [e2e-llm-inference-service] status_count -= 1 [e2e-llm-inference-service] cause = ResponseError.SPECIFIC_ERROR.format(status_code=response.status) [e2e-llm-inference-service] status = response.status [e2e-llm-inference-service] [e2e-llm-inference-service] history = self.history + ( [e2e-llm-inference-service] RequestHistory(method, url, error, status, redirect_location), [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] new_retry = self.new( [e2e-llm-inference-service] total=total, [e2e-llm-inference-service] connect=connect, [e2e-llm-inference-service] read=read, [e2e-llm-inference-service] redirect=redirect, [e2e-llm-inference-service] status=status_count, [e2e-llm-inference-service] other=other, [e2e-llm-inference-service] history=history, [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] if new_retry.is_exhausted(): [e2e-llm-inference-service] reason = error or ResponseError(cause) [e2e-llm-inference-service] > raise MaxRetryError(_pool, url, reason) from reason # type: ignore[arg-type] [e2e-llm-inference-service] E urllib3.exceptions.MaxRetryError: HTTPConnectionPool(host='ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com', port=80): Max retries exceeded with url: /kserve-ci-e2e-test/custom-route-timeout-test/v1/completions (Caused by ReadTimeoutError("HTTPConnectionPool(host='ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com', port=80): Read timed out. (read timeout=60)")) [e2e-llm-inference-service] [e2e-llm-inference-service] ../../python/kserve/.venv/lib64/python3.11/site-packages/urllib3/util/retry.py:519: MaxRetryError [e2e-llm-inference-service] [e2e-llm-inference-service] During handling of the above exception, another exception occurred: [e2e-llm-inference-service] [e2e-llm-inference-service] def get_successful_response(): [e2e-llm-inference-service] try: [e2e-llm-inference-service] if test_case.url_getter: [e2e-llm-inference-service] service_url = test_case.url_getter(kserve_client, test_case.llm_service) [e2e-llm-inference-service] else: [e2e-llm-inference-service] service_url = get_llm_service_url(kserve_client, test_case.llm_service) [e2e-llm-inference-service] except Exception as e: [e2e-llm-inference-service] raise AssertionError(f"❌ Failed to get service URL: {e}") from e [e2e-llm-inference-service] [e2e-llm-inference-service] model_url = service_url + test_case.endpoint [e2e-llm-inference-service] [e2e-llm-inference-service] headers = {"Content-Type": "application/json"} [e2e-llm-inference-service] if extra_headers: [e2e-llm-inference-service] headers.update(extra_headers) [e2e-llm-inference-service] [e2e-llm-inference-service] if test_case.payload_formatter is not None: [e2e-llm-inference-service] test_payload = test_case.payload_formatter(test_case) [e2e-llm-inference-service] elif test_case.prompt is not None: [e2e-llm-inference-service] test_payload = { [e2e-llm-inference-service] "model": test_case.model_name [e2e-llm-inference-service] if not extra_headers or MODEL_ROUTING_HEADER not in extra_headers [e2e-llm-inference-service] else extra_headers[MODEL_ROUTING_HEADER], [e2e-llm-inference-service] "prompt": test_case.prompt, [e2e-llm-inference-service] "max_tokens": test_case.max_tokens, [e2e-llm-inference-service] } [e2e-llm-inference-service] else: [e2e-llm-inference-service] test_payload = None [e2e-llm-inference-service] [e2e-llm-inference-service] logger.info(f"Calling LLM service at {model_url} with payload {test_payload}") [e2e-llm-inference-service] try: [e2e-llm-inference-service] if test_payload is not None: [e2e-llm-inference-service] > response = post_with_retry( [e2e-llm-inference-service] model_url, [e2e-llm-inference-service] headers=headers, [e2e-llm-inference-service] json_data=test_payload, [e2e-llm-inference-service] timeout=test_case.response_timeout, [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] llmisvc/test_llm_inference_service.py:1095: [e2e-llm-inference-service] _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ [e2e-llm-inference-service] [e2e-llm-inference-service] url = 'http://ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com/kserve-ci-e2e-test/custom-route-timeout-test/v1/completions' [e2e-llm-inference-service] [e2e-llm-inference-service] def post_with_retry( [e2e-llm-inference-service] url: str, [e2e-llm-inference-service] *, [e2e-llm-inference-service] headers: Dict = None, [e2e-llm-inference-service] json_data: Union[Dict, List] = None, [e2e-llm-inference-service] data: Union[str, bytes] = None, [e2e-llm-inference-service] stream: bool = False, [e2e-llm-inference-service] timeout: float = None, [e2e-llm-inference-service] total_retries: int = DEFAULT_RETRY_TOTAL, [e2e-llm-inference-service] backoff_factor: float = DEFAULT_RETRY_BACKOFF_FACTOR, [e2e-llm-inference-service] retry_status_codes=DEFAULT_RETRY_STATUS_CODES, [e2e-llm-inference-service] ) -> requests.Response: [e2e-llm-inference-service] """ [e2e-llm-inference-service] Send POST request with retries for transient HTTP and network failures. [e2e-llm-inference-service] """ [e2e-llm-inference-service] if json_data is not None and data is not None: [e2e-llm-inference-service] raise ValueError("Only one of json_data or data can be provided.") [e2e-llm-inference-service] [e2e-llm-inference-service] with _retry_session( [e2e-llm-inference-service] ["POST"], total_retries, backoff_factor, retry_status_codes [e2e-llm-inference-service] ) as session: [e2e-llm-inference-service] > return session.post( [e2e-llm-inference-service] url, [e2e-llm-inference-service] json=json_data, [e2e-llm-inference-service] data=data, [e2e-llm-inference-service] headers=headers, [e2e-llm-inference-service] stream=stream, [e2e-llm-inference-service] timeout=timeout, [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] common/http_retry.py:70: [e2e-llm-inference-service] _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ [e2e-llm-inference-service] [e2e-llm-inference-service] self = [e2e-llm-inference-service] url = 'http://ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com/kserve-ci-e2e-test/custom-route-timeout-test/v1/completions' [e2e-llm-inference-service] data = None [e2e-llm-inference-service] json = {'max_tokens': 20, 'model': 'facebook/opt-125m', 'prompt': 'KServe is a'} [e2e-llm-inference-service] kwargs = {'headers': {'Content-Type': 'application/json'}, 'stream': False, 'timeout': 60} [e2e-llm-inference-service] [e2e-llm-inference-service] def post(self, url, data=None, json=None, **kwargs): [e2e-llm-inference-service] r"""Sends a POST request. Returns :class:`Response` object. [e2e-llm-inference-service] [e2e-llm-inference-service] :param url: URL for the new :class:`Request` object. [e2e-llm-inference-service] :param data: (optional) Dictionary, list of tuples, bytes, or file-like [e2e-llm-inference-service] object to send in the body of the :class:`Request`. [e2e-llm-inference-service] :param json: (optional) json to send in the body of the :class:`Request`. [e2e-llm-inference-service] :param \*\*kwargs: Optional arguments that ``request`` takes. [e2e-llm-inference-service] :rtype: requests.Response [e2e-llm-inference-service] """ [e2e-llm-inference-service] [e2e-llm-inference-service] > return self.request("POST", url, data=data, json=json, **kwargs) [e2e-llm-inference-service] [e2e-llm-inference-service] ../../python/kserve/.venv/lib64/python3.11/site-packages/requests/sessions.py:637: [e2e-llm-inference-service] _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ [e2e-llm-inference-service] [e2e-llm-inference-service] self = , method = 'POST' [e2e-llm-inference-service] url = 'http://ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com/kserve-ci-e2e-test/custom-route-timeout-test/v1/completions' [e2e-llm-inference-service] params = None, data = None, headers = {'Content-Type': 'application/json'} [e2e-llm-inference-service] cookies = None, files = None, auth = None, timeout = 60, allow_redirects = True [e2e-llm-inference-service] proxies = {}, hooks = None, stream = False, verify = None, cert = None [e2e-llm-inference-service] json = {'max_tokens': 20, 'model': 'facebook/opt-125m', 'prompt': 'KServe is a'} [e2e-llm-inference-service] [e2e-llm-inference-service] def request( [e2e-llm-inference-service] self, [e2e-llm-inference-service] method, [e2e-llm-inference-service] url, [e2e-llm-inference-service] params=None, [e2e-llm-inference-service] data=None, [e2e-llm-inference-service] headers=None, [e2e-llm-inference-service] cookies=None, [e2e-llm-inference-service] files=None, [e2e-llm-inference-service] auth=None, [e2e-llm-inference-service] timeout=None, [e2e-llm-inference-service] allow_redirects=True, [e2e-llm-inference-service] proxies=None, [e2e-llm-inference-service] hooks=None, [e2e-llm-inference-service] stream=None, [e2e-llm-inference-service] verify=None, [e2e-llm-inference-service] cert=None, [e2e-llm-inference-service] json=None, [e2e-llm-inference-service] ): [e2e-llm-inference-service] """Constructs a :class:`Request `, prepares it and sends it. [e2e-llm-inference-service] Returns :class:`Response ` object. [e2e-llm-inference-service] [e2e-llm-inference-service] :param method: method for the new :class:`Request` object. [e2e-llm-inference-service] :param url: URL for the new :class:`Request` object. [e2e-llm-inference-service] :param params: (optional) Dictionary or bytes to be sent in the query [e2e-llm-inference-service] string for the :class:`Request`. [e2e-llm-inference-service] :param data: (optional) Dictionary, list of tuples, bytes, or file-like [e2e-llm-inference-service] object to send in the body of the :class:`Request`. [e2e-llm-inference-service] :param json: (optional) json to send in the body of the [e2e-llm-inference-service] :class:`Request`. [e2e-llm-inference-service] :param headers: (optional) Dictionary of HTTP Headers to send with the [e2e-llm-inference-service] :class:`Request`. [e2e-llm-inference-service] :param cookies: (optional) Dict or CookieJar object to send with the [e2e-llm-inference-service] :class:`Request`. [e2e-llm-inference-service] :param files: (optional) Dictionary of ``'filename': file-like-objects`` [e2e-llm-inference-service] for multipart encoding upload. [e2e-llm-inference-service] :param auth: (optional) Auth tuple or callable to enable [e2e-llm-inference-service] Basic/Digest/Custom HTTP Auth. [e2e-llm-inference-service] :param timeout: (optional) How long to wait for the server to send [e2e-llm-inference-service] data before giving up, as a float, or a :ref:`(connect timeout, [e2e-llm-inference-service] read timeout) ` tuple. [e2e-llm-inference-service] :type timeout: float or tuple [e2e-llm-inference-service] :param allow_redirects: (optional) Set to True by default. [e2e-llm-inference-service] :type allow_redirects: bool [e2e-llm-inference-service] :param proxies: (optional) Dictionary mapping protocol or protocol and [e2e-llm-inference-service] hostname to the URL of the proxy. [e2e-llm-inference-service] :param hooks: (optional) Dictionary mapping hook name to one event or [e2e-llm-inference-service] list of events, event must be callable. [e2e-llm-inference-service] :param stream: (optional) whether to immediately download the response [e2e-llm-inference-service] content. Defaults to ``False``. [e2e-llm-inference-service] :param verify: (optional) Either a boolean, in which case it controls whether we verify [e2e-llm-inference-service] the server's TLS certificate, or a string, in which case it must be a path [e2e-llm-inference-service] to a CA bundle to use. Defaults to ``True``. When set to [e2e-llm-inference-service] ``False``, requests will accept any TLS certificate presented by [e2e-llm-inference-service] the server, and will ignore hostname mismatches and/or expired [e2e-llm-inference-service] certificates, which will make your application vulnerable to [e2e-llm-inference-service] man-in-the-middle (MitM) attacks. Setting verify to ``False`` [e2e-llm-inference-service] may be useful during local development or testing. [e2e-llm-inference-service] :param cert: (optional) if String, path to ssl client cert file (.pem). [e2e-llm-inference-service] If Tuple, ('cert', 'key') pair. [e2e-llm-inference-service] :rtype: requests.Response [e2e-llm-inference-service] """ [e2e-llm-inference-service] # Create the Request. [e2e-llm-inference-service] req = Request( [e2e-llm-inference-service] method=method.upper(), [e2e-llm-inference-service] url=url, [e2e-llm-inference-service] headers=headers, [e2e-llm-inference-service] files=files, [e2e-llm-inference-service] data=data or {}, [e2e-llm-inference-service] json=json, [e2e-llm-inference-service] params=params or {}, [e2e-llm-inference-service] auth=auth, [e2e-llm-inference-service] cookies=cookies, [e2e-llm-inference-service] hooks=hooks, [e2e-llm-inference-service] ) [e2e-llm-inference-service] prep = self.prepare_request(req) [e2e-llm-inference-service] [e2e-llm-inference-service] proxies = proxies or {} [e2e-llm-inference-service] [e2e-llm-inference-service] settings = self.merge_environment_settings( [e2e-llm-inference-service] prep.url, proxies, stream, verify, cert [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] # Send the request. [e2e-llm-inference-service] send_kwargs = { [e2e-llm-inference-service] "timeout": timeout, [e2e-llm-inference-service] "allow_redirects": allow_redirects, [e2e-llm-inference-service] } [e2e-llm-inference-service] send_kwargs.update(settings) [e2e-llm-inference-service] > resp = self.send(prep, **send_kwargs) [e2e-llm-inference-service] [e2e-llm-inference-service] ../../python/kserve/.venv/lib64/python3.11/site-packages/requests/sessions.py:589: [e2e-llm-inference-service] _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ [e2e-llm-inference-service] [e2e-llm-inference-service] self = [e2e-llm-inference-service] request = [e2e-llm-inference-service] kwargs = {'cert': None, 'proxies': OrderedDict(), 'stream': False, 'timeout': 60, ...} [e2e-llm-inference-service] allow_redirects = True, stream = False, hooks = {'response': []} [e2e-llm-inference-service] adapter = [e2e-llm-inference-service] start = 1783435858.5349567 [e2e-llm-inference-service] [e2e-llm-inference-service] def send(self, request, **kwargs): [e2e-llm-inference-service] """Send a given PreparedRequest. [e2e-llm-inference-service] [e2e-llm-inference-service] :rtype: requests.Response [e2e-llm-inference-service] """ [e2e-llm-inference-service] # Set defaults that the hooks can utilize to ensure they always have [e2e-llm-inference-service] # the correct parameters to reproduce the previous request. [e2e-llm-inference-service] kwargs.setdefault("stream", self.stream) [e2e-llm-inference-service] kwargs.setdefault("verify", self.verify) [e2e-llm-inference-service] kwargs.setdefault("cert", self.cert) [e2e-llm-inference-service] if "proxies" not in kwargs: [e2e-llm-inference-service] kwargs["proxies"] = resolve_proxies(request, self.proxies, self.trust_env) [e2e-llm-inference-service] [e2e-llm-inference-service] # It's possible that users might accidentally send a Request object. [e2e-llm-inference-service] # Guard against that specific failure case. [e2e-llm-inference-service] if isinstance(request, Request): [e2e-llm-inference-service] raise ValueError("You can only send PreparedRequests.") [e2e-llm-inference-service] [e2e-llm-inference-service] # Set up variables needed for resolve_redirects and dispatching of hooks [e2e-llm-inference-service] allow_redirects = kwargs.pop("allow_redirects", True) [e2e-llm-inference-service] stream = kwargs.get("stream") [e2e-llm-inference-service] hooks = request.hooks [e2e-llm-inference-service] [e2e-llm-inference-service] # Get the appropriate adapter to use [e2e-llm-inference-service] adapter = self.get_adapter(url=request.url) [e2e-llm-inference-service] [e2e-llm-inference-service] # Start time (approximately) of the request [e2e-llm-inference-service] start = preferred_clock() [e2e-llm-inference-service] [e2e-llm-inference-service] # Send the request [e2e-llm-inference-service] > r = adapter.send(request, **kwargs) [e2e-llm-inference-service] [e2e-llm-inference-service] ../../python/kserve/.venv/lib64/python3.11/site-packages/requests/sessions.py:703: [e2e-llm-inference-service] _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ [e2e-llm-inference-service] [e2e-llm-inference-service] self = [e2e-llm-inference-service] request = , stream = False [e2e-llm-inference-service] timeout = Timeout(connect=60, read=60, total=None), verify = '/tmp/ca.crt' [e2e-llm-inference-service] cert = None, proxies = OrderedDict() [e2e-llm-inference-service] [e2e-llm-inference-service] def send( [e2e-llm-inference-service] self, request, stream=False, timeout=None, verify=True, cert=None, proxies=None [e2e-llm-inference-service] ): [e2e-llm-inference-service] """Sends PreparedRequest object. Returns Response object. [e2e-llm-inference-service] [e2e-llm-inference-service] :param request: The :class:`PreparedRequest ` being sent. [e2e-llm-inference-service] :param stream: (optional) Whether to stream the request content. [e2e-llm-inference-service] :param timeout: (optional) How long to wait for the server to send [e2e-llm-inference-service] data before giving up, as a float, or a :ref:`(connect timeout, [e2e-llm-inference-service] read timeout) ` tuple. [e2e-llm-inference-service] :type timeout: float or tuple or urllib3 Timeout object [e2e-llm-inference-service] :param verify: (optional) Either a boolean, in which case it controls whether [e2e-llm-inference-service] we verify the server's TLS certificate, or a string, in which case it [e2e-llm-inference-service] must be a path to a CA bundle to use [e2e-llm-inference-service] :param cert: (optional) Any user-provided SSL certificate to be trusted. [e2e-llm-inference-service] :param proxies: (optional) The proxies dictionary to apply to the request. [e2e-llm-inference-service] :rtype: requests.Response [e2e-llm-inference-service] """ [e2e-llm-inference-service] [e2e-llm-inference-service] try: [e2e-llm-inference-service] conn = self.get_connection_with_tls_context( [e2e-llm-inference-service] request, verify, proxies=proxies, cert=cert [e2e-llm-inference-service] ) [e2e-llm-inference-service] except LocationValueError as e: [e2e-llm-inference-service] raise InvalidURL(e, request=request) [e2e-llm-inference-service] [e2e-llm-inference-service] self.cert_verify(conn, request.url, verify, cert) [e2e-llm-inference-service] url = self.request_url(request, proxies) [e2e-llm-inference-service] self.add_headers( [e2e-llm-inference-service] request, [e2e-llm-inference-service] stream=stream, [e2e-llm-inference-service] timeout=timeout, [e2e-llm-inference-service] verify=verify, [e2e-llm-inference-service] cert=cert, [e2e-llm-inference-service] proxies=proxies, [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] chunked = not (request.body is None or "Content-Length" in request.headers) [e2e-llm-inference-service] [e2e-llm-inference-service] if isinstance(timeout, tuple): [e2e-llm-inference-service] try: [e2e-llm-inference-service] connect, read = timeout [e2e-llm-inference-service] timeout = TimeoutSauce(connect=connect, read=read) [e2e-llm-inference-service] except ValueError: [e2e-llm-inference-service] raise ValueError( [e2e-llm-inference-service] f"Invalid timeout {timeout}. Pass a (connect, read) timeout tuple, " [e2e-llm-inference-service] f"or a single float to set both timeouts to the same value." [e2e-llm-inference-service] ) [e2e-llm-inference-service] elif isinstance(timeout, TimeoutSauce): [e2e-llm-inference-service] pass [e2e-llm-inference-service] else: [e2e-llm-inference-service] timeout = TimeoutSauce(connect=timeout, read=timeout) [e2e-llm-inference-service] [e2e-llm-inference-service] try: [e2e-llm-inference-service] resp = conn.urlopen( [e2e-llm-inference-service] method=request.method, [e2e-llm-inference-service] url=url, [e2e-llm-inference-service] body=request.body, [e2e-llm-inference-service] headers=request.headers, [e2e-llm-inference-service] redirect=False, [e2e-llm-inference-service] assert_same_host=False, [e2e-llm-inference-service] preload_content=False, [e2e-llm-inference-service] decode_content=False, [e2e-llm-inference-service] retries=self.max_retries, [e2e-llm-inference-service] timeout=timeout, [e2e-llm-inference-service] chunked=chunked, [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] except (ProtocolError, OSError) as err: [e2e-llm-inference-service] raise ConnectionError(err, request=request) [e2e-llm-inference-service] [e2e-llm-inference-service] except MaxRetryError as e: [e2e-llm-inference-service] if isinstance(e.reason, ConnectTimeoutError): [e2e-llm-inference-service] # TODO: Remove this in 3.0.0: see #2811 [e2e-llm-inference-service] if not isinstance(e.reason, NewConnectionError): [e2e-llm-inference-service] raise ConnectTimeout(e, request=request) [e2e-llm-inference-service] [e2e-llm-inference-service] if isinstance(e.reason, ResponseError): [e2e-llm-inference-service] raise RetryError(e, request=request) [e2e-llm-inference-service] [e2e-llm-inference-service] if isinstance(e.reason, _ProxyError): [e2e-llm-inference-service] raise ProxyError(e, request=request) [e2e-llm-inference-service] [e2e-llm-inference-service] if isinstance(e.reason, _SSLError): [e2e-llm-inference-service] # This branch is for urllib3 v1.22 and later. [e2e-llm-inference-service] raise SSLError(e, request=request) [e2e-llm-inference-service] [e2e-llm-inference-service] > raise ConnectionError(e, request=request) [e2e-llm-inference-service] E requests.exceptions.ConnectionError: HTTPConnectionPool(host='ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com', port=80): Max retries exceeded with url: /kserve-ci-e2e-test/custom-route-timeout-test/v1/completions (Caused by ReadTimeoutError("HTTPConnectionPool(host='ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com', port=80): Read timed out. (read timeout=60)")) [e2e-llm-inference-service] [e2e-llm-inference-service] ../../python/kserve/.venv/lib64/python3.11/site-packages/requests/adapters.py:700: ConnectionError [e2e-llm-inference-service] [e2e-llm-inference-service] The above exception was the direct cause of the following exception: [e2e-llm-inference-service] [e2e-llm-inference-service] test_case = TestCase(base_refs=['router-custom-route-timeout', 'scheduler-managed', 'workload-single-cpu', 'model-fb-opt-125m'], p... {'name': 'model-fb-opt-125m-custom-route-928a8601'}]}, [e2e-llm-inference-service] 'status': None}, model_name='facebook/opt-125m') [e2e-llm-inference-service] [e2e-llm-inference-service] @pytest.mark.llminferenceservice [e2e-llm-inference-service] @pytest.mark.asyncio(loop_scope="session") [e2e-llm-inference-service] @pytest.mark.parametrize( [e2e-llm-inference-service] "test_case", [e2e-llm-inference-service] [ [e2e-llm-inference-service] pytest.param( [e2e-llm-inference-service] TestCase( [e2e-llm-inference-service] base_refs=[ [e2e-llm-inference-service] "router-with-gateway-ref", [e2e-llm-inference-service] "router-with-managed-route", [e2e-llm-inference-service] "model-fb-opt-125m", [e2e-llm-inference-service] "workload-llmd-simulator", [e2e-llm-inference-service] ], [e2e-llm-inference-service] endpoint="/v1/completions", [e2e-llm-inference-service] prompt="KServe is a", [e2e-llm-inference-service] payload_formatter=completions_payload, [e2e-llm-inference-service] response_assertion=create_response_assertion(with_field="choices"), [e2e-llm-inference-service] expected_gateway=ROUTER_GATEWAYS[0], [e2e-llm-inference-service] before_test=[ [e2e-llm-inference-service] lambda: create_router_resources( [e2e-llm-inference-service] gateways=[ROUTER_GATEWAYS[0]], [e2e-llm-inference-service] ) [e2e-llm-inference-service] ], [e2e-llm-inference-service] ), [e2e-llm-inference-service] marks=[ [e2e-llm-inference-service] pytest.mark.cluster_cpu, [e2e-llm-inference-service] pytest.mark.cluster_single_node, [e2e-llm-inference-service] pytest.mark.llmd_simulator, [e2e-llm-inference-service] pytest.mark.custom_gateway, [e2e-llm-inference-service] ], [e2e-llm-inference-service] ), [e2e-llm-inference-service] pytest.param( [e2e-llm-inference-service] TestCase( [e2e-llm-inference-service] base_refs=[ [e2e-llm-inference-service] "router-managed", [e2e-llm-inference-service] "workload-single-cpu", [e2e-llm-inference-service] "model-fb-opt-125m", [e2e-llm-inference-service] ], [e2e-llm-inference-service] prompt="KServe is a", [e2e-llm-inference-service] payload_formatter=completions_payload, [e2e-llm-inference-service] response_assertion=assert_200_with_choices, [e2e-llm-inference-service] ), [e2e-llm-inference-service] marks=[pytest.mark.cluster_cpu, pytest.mark.cluster_single_node], [e2e-llm-inference-service] ), [e2e-llm-inference-service] pytest.param( [e2e-llm-inference-service] TestCase( [e2e-llm-inference-service] base_refs=[ [e2e-llm-inference-service] "router-custom-route-timeout", [e2e-llm-inference-service] "scheduler-managed", [e2e-llm-inference-service] "workload-single-cpu", [e2e-llm-inference-service] "model-fb-opt-125m", [e2e-llm-inference-service] ], [e2e-llm-inference-service] prompt="KServe is a", [e2e-llm-inference-service] service_name="custom-route-timeout-test", [e2e-llm-inference-service] ), [e2e-llm-inference-service] marks=[pytest.mark.cluster_cpu, pytest.mark.cluster_single_node], [e2e-llm-inference-service] ), [e2e-llm-inference-service] pytest.param( [e2e-llm-inference-service] TestCase( [e2e-llm-inference-service] base_refs=[ [e2e-llm-inference-service] "router-with-refs", [e2e-llm-inference-service] "scheduler-managed", [e2e-llm-inference-service] "workload-single-cpu", [e2e-llm-inference-service] "model-fb-opt-125m", [e2e-llm-inference-service] ], [e2e-llm-inference-service] prompt="KServe is a", [e2e-llm-inference-service] service_name="router-with-refs-test", [e2e-llm-inference-service] expected_gateway=ROUTER_GATEWAYS[0], [e2e-llm-inference-service] before_test=[ [e2e-llm-inference-service] lambda: create_router_resources( [e2e-llm-inference-service] gateways=[ROUTER_GATEWAYS[0]], [e2e-llm-inference-service] routes=[ROUTER_ROUTES[0], ROUTER_ROUTES[1]], [e2e-llm-inference-service] ) [e2e-llm-inference-service] ], [e2e-llm-inference-service] ), [e2e-llm-inference-service] marks=[ [e2e-llm-inference-service] pytest.mark.cluster_cpu, [e2e-llm-inference-service] pytest.mark.cluster_single_node, [e2e-llm-inference-service] pytest.mark.custom_gateway, [e2e-llm-inference-service] ], [e2e-llm-inference-service] ), [e2e-llm-inference-service] pytest.param( [e2e-llm-inference-service] TestCase( [e2e-llm-inference-service] base_refs=["router-managed", "workload-pd-cpu", "model-fb-opt-125m"], [e2e-llm-inference-service] prompt="You are an expert in Kubernetes-native machine learning serving platforms, with deep knowledge of the KServe project. " [e2e-llm-inference-service] "Explain the challenges of serving large-scale models, GPU scheduling, and how KServe integrates with capabilities like multi-model serving. " [e2e-llm-inference-service] "Provide a detailed comparison with open source alternatives, focusing on operational trade-offs.", [e2e-llm-inference-service] response_assertion=assert_200_with_choices, [e2e-llm-inference-service] ), [e2e-llm-inference-service] marks=[pytest.mark.cluster_cpu, pytest.mark.cluster_single_node], [e2e-llm-inference-service] ), [e2e-llm-inference-service] pytest.param( [e2e-llm-inference-service] TestCase( [e2e-llm-inference-service] base_refs=[ [e2e-llm-inference-service] "router-custom-route-timeout-pd", [e2e-llm-inference-service] "scheduler-managed", [e2e-llm-inference-service] "workload-pd-cpu", [e2e-llm-inference-service] "model-fb-opt-125m", [e2e-llm-inference-service] ], [e2e-llm-inference-service] prompt="You are an expert in Kubernetes-native machine learning serving platforms, with deep knowledge of the KServe project. " [e2e-llm-inference-service] "Explain the challenges of serving large-scale models, GPU scheduling, and how KServe integrates with capabilities like multi-model serving. " [e2e-llm-inference-service] "Provide a detailed comparison with open source alternatives, focusing on operational trade-offs.", [e2e-llm-inference-service] service_name="custom-route-timeout-pd-test", [e2e-llm-inference-service] response_assertion=assert_200_with_choices, [e2e-llm-inference-service] ), [e2e-llm-inference-service] marks=[pytest.mark.cluster_cpu, pytest.mark.cluster_single_node], [e2e-llm-inference-service] ), [e2e-llm-inference-service] pytest.param( [e2e-llm-inference-service] TestCase( [e2e-llm-inference-service] base_refs=[ [e2e-llm-inference-service] "router-with-refs-pd", [e2e-llm-inference-service] "scheduler-managed", [e2e-llm-inference-service] "workload-pd-cpu", [e2e-llm-inference-service] "model-fb-opt-125m", [e2e-llm-inference-service] ], [e2e-llm-inference-service] prompt="You are an expert in Kubernetes-native machine learning serving platforms, with deep knowledge of the KServe project. " [e2e-llm-inference-service] "Explain the challenges of serving large-scale models, GPU scheduling, and how KServe integrates with capabilities like multi-model serving. " [e2e-llm-inference-service] "Provide a detailed comparison with open source alternatives, focusing on operational trade-offs.", [e2e-llm-inference-service] service_name="router-with-refs-pd-test", [e2e-llm-inference-service] response_assertion=assert_200_with_choices, [e2e-llm-inference-service] expected_gateway=ROUTER_GATEWAYS[1], [e2e-llm-inference-service] before_test=[ [e2e-llm-inference-service] lambda: create_router_resources( [e2e-llm-inference-service] gateways=[ROUTER_GATEWAYS[1]], [e2e-llm-inference-service] routes=[ROUTER_ROUTES[2], ROUTER_ROUTES[3]], [e2e-llm-inference-service] ) [e2e-llm-inference-service] ], [e2e-llm-inference-service] ), [e2e-llm-inference-service] marks=[ [e2e-llm-inference-service] pytest.mark.cluster_cpu, [e2e-llm-inference-service] pytest.mark.cluster_single_node, [e2e-llm-inference-service] pytest.mark.custom_gateway, [e2e-llm-inference-service] ], [e2e-llm-inference-service] ), [e2e-llm-inference-service] pytest.param( [e2e-llm-inference-service] TestCase( [e2e-llm-inference-service] base_refs=[ [e2e-llm-inference-service] "router-managed", [e2e-llm-inference-service] "workload-dp-ep-gpu", [e2e-llm-inference-service] "workload-dp-ep-prefill-gpu", [e2e-llm-inference-service] "model-deepseek-v2-lite", [e2e-llm-inference-service] ], [e2e-llm-inference-service] prompt="Delve into the multifaceted implications of a fully disaggregated cloud architecture, specifically " [e2e-llm-inference-service] "where the compute plane (P) and the data plane (D) are independently deployed and managed for a " [e2e-llm-inference-service] "geographically distributed, high-throughput, low-latency microservices ecosystem. Beyond the " [e2e-llm-inference-service] "fundamental challenges of network latency and data consistency, elaborate on the advanced " [e2e-llm-inference-service] "considerations and trade-offs inherent in such a setup: 1. Network Architecture and Protocols: " [e2e-llm-inference-service] "How would the network fabric and underlying protocols (e.g., RDMA, custom transport layers) need to " [e2e-llm-inference-service] "evolve to support optimal performance and minimize inter-plane communication overhead, especially for " [e2e-llm-inference-service] "synchronous operations? Discuss the role of network programmability (e.g., SDN, P4) in dynamically " [e2e-llm-inference-service] "optimizing routing and traffic flow between P and D. 2. Advanced Data Consistency and Durability: " [e2e-llm-inference-service] "Explore sophisticated data consistency models (e.g., causal consistency, strong eventual consistency) " [e2e-llm-inference-service] "and their applicability in balancing performance and data integrity across a globally distributed data plane. " [e2e-llm-inference-service] "Detail strategies for ensuring data durability and fault tolerance, including multi-region replication, " [e2e-llm-inference-service] "intelligent partitioning, and recovery mechanisms in the event of partial or full plane failures. " [e2e-llm-inference-service] "3. Dynamic Resource Orchestration and Cost Optimization: Analyze how an orchestration layer would intelligently " [e2e-llm-inference-service] "manage the independent scaling of compute (P) and data (D) resources, considering fluctuating workloads, " [e2e-llm-inference-service] "cost efficiency, and performance targets (e.g., using predictive analytics for resource provisioning). " [e2e-llm-inference-service] "Discuss mechanisms for dynamically reallocating compute nodes to different data partitions based on " [e2e-llm-inference-service] "workload patterns and data locality, potentially involving live migration strategies. " [e2e-llm-inference-service] "4. Security and Compliance in a Distributed Landscape: Address the enhanced security perimeter " [e2e-llm-inference-service] "challenges, including securing communication channels between P and D (encryption in transit, mutual TLS), " [e2e-llm-inference-service] "fine-grained access control to data at rest and in motion, and identity management across disaggregated " [e2e-llm-inference-service] "components. Discuss how such an architecture impacts compliance with regulatory frameworks (e.g., GDPR, HIPAA) " [e2e-llm-inference-service] "concerning data sovereignty, privacy, and auditability. 5. Operational Complexity and Observability: " [e2e-llm-inference-service] "Examine the increased complexity in monitoring, logging, and tracing across highly decoupled compute and " [e2e-llm-inference-service] "data planes. What specialized tooling and practices (e.g., distributed tracing with OpenTelemetry, advanced AIOps) " [e2e-llm-inference-service] "would be essential? How would incident response and troubleshooting differ in this disaggregated environment " [e2e-llm-inference-service] "compared to traditional integrated systems? Consider the challenges of pinpointing root causes across " [e2e-llm-inference-service] "independent failures. 6. Real-world Applicability and Future Trends: Identify specific industries " [e2e-llm-inference-service] "or use cases (e.g., high-frequency trading, IoT edge processing, large language model inference) " [e2e-llm-inference-service] "where the benefits of P/D disaggregation would strongly outweigh its complexities. " [e2e-llm-inference-service] "Conclude by speculating on emerging technologies or paradigms (e.g., serverless compute functions " [e2e-llm-inference-service] "directly interacting with object storage, in-memory disaggregation) that could further drive or " [e2e-llm-inference-service] "transform P/D disaggregation in cloud computing.", [e2e-llm-inference-service] max_tokens=2000, [e2e-llm-inference-service] ), [e2e-llm-inference-service] marks=[ [e2e-llm-inference-service] pytest.mark.cluster_gpu, [e2e-llm-inference-service] pytest.mark.cluster_nvidia, [e2e-llm-inference-service] pytest.mark.cluster_nvidia_roce, [e2e-llm-inference-service] ], [e2e-llm-inference-service] ), [e2e-llm-inference-service] pytest.param( [e2e-llm-inference-service] TestCase( [e2e-llm-inference-service] base_refs=[ [e2e-llm-inference-service] "router-no-scheduler", [e2e-llm-inference-service] "workload-single-cpu", [e2e-llm-inference-service] "model-fb-opt-125m", [e2e-llm-inference-service] ], [e2e-llm-inference-service] prompt="What is KServe?", [e2e-llm-inference-service] ), [e2e-llm-inference-service] marks=[ [e2e-llm-inference-service] pytest.mark.cluster_cpu, [e2e-llm-inference-service] pytest.mark.cluster_single_node, [e2e-llm-inference-service] pytest.mark.no_scheduler, [e2e-llm-inference-service] ], [e2e-llm-inference-service] ), [e2e-llm-inference-service] pytest.param( [e2e-llm-inference-service] TestCase( [e2e-llm-inference-service] base_refs=[ [e2e-llm-inference-service] "router-managed", [e2e-llm-inference-service] "workload-simulated-dp-ep-cpu", [e2e-llm-inference-service] "model-fb-opt-125m", [e2e-llm-inference-service] ], [e2e-llm-inference-service] prompt="This test simulates DP+EP that can run on CPU, the idea is to test the LWS-based deployment, " [e2e-llm-inference-service] "but without the resources requirements for DP+EP (GPUs and ROCe/IB).", [e2e-llm-inference-service] ), [e2e-llm-inference-service] marks=[pytest.mark.cluster_cpu, pytest.mark.cluster_multi_node], [e2e-llm-inference-service] ), [e2e-llm-inference-service] # Scheduler config tests [e2e-llm-inference-service] pytest.param( [e2e-llm-inference-service] TestCase( [e2e-llm-inference-service] base_refs=[ [e2e-llm-inference-service] "router-managed", [e2e-llm-inference-service] "scheduler-with-inline-config", [e2e-llm-inference-service] "workload-llmd-simulator", [e2e-llm-inference-service] ], [e2e-llm-inference-service] prompt="KServe is a", [e2e-llm-inference-service] service_name="scheduler-inline-config-test", [e2e-llm-inference-service] ), [e2e-llm-inference-service] marks=[pytest.mark.cluster_cpu, pytest.mark.cluster_single_node], [e2e-llm-inference-service] ), [e2e-llm-inference-service] # Chat completions endpoint coverage [e2e-llm-inference-service] pytest.param( [e2e-llm-inference-service] TestCase( [e2e-llm-inference-service] base_refs=[ [e2e-llm-inference-service] "router-managed", [e2e-llm-inference-service] "workload-llmd-simulator", [e2e-llm-inference-service] "model-qwen2.5-0.5b", [e2e-llm-inference-service] ], [e2e-llm-inference-service] model_name="Qwen/Qwen2.5-0.5B-Instruct", [e2e-llm-inference-service] endpoint="/v1/chat/completions", [e2e-llm-inference-service] prompt="What is KServe?", [e2e-llm-inference-service] payload_formatter=chat_completions_payload, [e2e-llm-inference-service] response_assertion=create_response_assertion(with_field="choices"), [e2e-llm-inference-service] ), [e2e-llm-inference-service] marks=[ [e2e-llm-inference-service] pytest.mark.cluster_cpu, [e2e-llm-inference-service] pytest.mark.cluster_single_node, [e2e-llm-inference-service] pytest.mark.llmd_simulator, [e2e-llm-inference-service] ], [e2e-llm-inference-service] ), [e2e-llm-inference-service] pytest.param( [e2e-llm-inference-service] TestCase( [e2e-llm-inference-service] base_refs=[ [e2e-llm-inference-service] "router-managed", [e2e-llm-inference-service] "scheduler-with-configmap-ref", [e2e-llm-inference-service] "workload-llmd-simulator", [e2e-llm-inference-service] ], [e2e-llm-inference-service] prompt="KServe is a", [e2e-llm-inference-service] service_name="scheduler-configmap-ref-test", [e2e-llm-inference-service] before_test=[create_scheduler_configmap], [e2e-llm-inference-service] after_test=[delete_scheduler_configmap], [e2e-llm-inference-service] ), [e2e-llm-inference-service] marks=[pytest.mark.cluster_cpu, pytest.mark.cluster_single_node], [e2e-llm-inference-service] ), [e2e-llm-inference-service] pytest.param( [e2e-llm-inference-service] TestCase( [e2e-llm-inference-service] base_refs=[ [e2e-llm-inference-service] "router-managed", [e2e-llm-inference-service] "scheduler-with-replicas", [e2e-llm-inference-service] "workload-llmd-simulator", [e2e-llm-inference-service] ], [e2e-llm-inference-service] prompt="KServe is a", [e2e-llm-inference-service] service_name="scheduler-ha-replicas-test", [e2e-llm-inference-service] ), [e2e-llm-inference-service] marks=[pytest.mark.cluster_cpu, pytest.mark.cluster_single_node], [e2e-llm-inference-service] ), [e2e-llm-inference-service] pytest.param( [e2e-llm-inference-service] TestCase( [e2e-llm-inference-service] base_refs=[ [e2e-llm-inference-service] "router-managed", [e2e-llm-inference-service] "scheduler-with-custom-template", [e2e-llm-inference-service] "workload-llmd-simulator", [e2e-llm-inference-service] ], [e2e-llm-inference-service] prompt="KServe is a", [e2e-llm-inference-service] service_name="scheduler-custom-template-test", [e2e-llm-inference-service] ), [e2e-llm-inference-service] marks=[pytest.mark.cluster_cpu, pytest.mark.cluster_single_node], [e2e-llm-inference-service] ), [e2e-llm-inference-service] # Scheduler v0.6 → v0.7 migration tests. [e2e-llm-inference-service] # Deploy v0.6-style configs and verify the controller migrates them [e2e-llm-inference-service] # so the v0.7 scheduler boots successfully. [e2e-llm-inference-service] pytest.param( [e2e-llm-inference-service] TestCase( [e2e-llm-inference-service] base_refs=[ [e2e-llm-inference-service] "router-managed", [e2e-llm-inference-service] "scheduler-v06-pd-config-migration", [e2e-llm-inference-service] "workload-llmd-simulator-pd", [e2e-llm-inference-service] ], [e2e-llm-inference-service] prompt="KServe is a", [e2e-llm-inference-service] service_name="scheduler-v06-pd-migration-test", [e2e-llm-inference-service] response_assertion=assert_200_with_choices, [e2e-llm-inference-service] ), [e2e-llm-inference-service] marks=[ [e2e-llm-inference-service] pytest.mark.cluster_cpu, [e2e-llm-inference-service] pytest.mark.cluster_single_node, [e2e-llm-inference-service] pytest.mark.llmd_simulator, [e2e-llm-inference-service] ], [e2e-llm-inference-service] ), [e2e-llm-inference-service] pytest.param( [e2e-llm-inference-service] TestCase( [e2e-llm-inference-service] base_refs=[ [e2e-llm-inference-service] "router-managed", [e2e-llm-inference-service] "scheduler-v06-nonzero-threshold-migration", [e2e-llm-inference-service] "workload-llmd-simulator-pd", [e2e-llm-inference-service] ], [e2e-llm-inference-service] prompt="KServe is a", [e2e-llm-inference-service] service_name="scheduler-v06-threshold-migration-test", [e2e-llm-inference-service] response_assertion=assert_200_with_choices, [e2e-llm-inference-service] ), [e2e-llm-inference-service] marks=[ [e2e-llm-inference-service] pytest.mark.cluster_cpu, [e2e-llm-inference-service] pytest.mark.cluster_single_node, [e2e-llm-inference-service] pytest.mark.llmd_simulator, [e2e-llm-inference-service] ], [e2e-llm-inference-service] ), [e2e-llm-inference-service] # Precise prefix KV cache routing test [e2e-llm-inference-service] pytest.param( [e2e-llm-inference-service] TestCase( [e2e-llm-inference-service] base_refs=[ [e2e-llm-inference-service] "router-managed", [e2e-llm-inference-service] "scheduler-with-precise-prefix-cache-inline-config", [e2e-llm-inference-service] "workload-llmd-simulator-kvcache", [e2e-llm-inference-service] ], [e2e-llm-inference-service] prompt="KServe is a", [e2e-llm-inference-service] service_name="precise-prefix-cache-test", [e2e-llm-inference-service] ), [e2e-llm-inference-service] marks=[ [e2e-llm-inference-service] pytest.mark.cluster_cpu, [e2e-llm-inference-service] pytest.mark.cluster_single_node, [e2e-llm-inference-service] pytest.mark.llmd_simulator, [e2e-llm-inference-service] ], [e2e-llm-inference-service] ), [e2e-llm-inference-service] # Models endpoint coverage [e2e-llm-inference-service] pytest.param( [e2e-llm-inference-service] TestCase( [e2e-llm-inference-service] base_refs=[ [e2e-llm-inference-service] "router-managed", [e2e-llm-inference-service] "workload-llmd-simulator", [e2e-llm-inference-service] ], [e2e-llm-inference-service] endpoint="/v1/models", [e2e-llm-inference-service] response_assertion=create_response_assertion(with_field="data"), [e2e-llm-inference-service] ), [e2e-llm-inference-service] marks=[ [e2e-llm-inference-service] pytest.mark.cluster_cpu, [e2e-llm-inference-service] pytest.mark.cluster_single_node, [e2e-llm-inference-service] pytest.mark.llmd_simulator, [e2e-llm-inference-service] ], [e2e-llm-inference-service] ), [e2e-llm-inference-service] # Model-based routing via X-Gateway-Model-Name header — /v1/completions [e2e-llm-inference-service] pytest.param( [e2e-llm-inference-service] TestCase( [e2e-llm-inference-service] base_refs=[ [e2e-llm-inference-service] "router-managed", [e2e-llm-inference-service] "workload-llmd-simulator", [e2e-llm-inference-service] ], [e2e-llm-inference-service] endpoint="/v1/completions", [e2e-llm-inference-service] prompt="KServe is a", [e2e-llm-inference-service] payload_formatter=completions_payload, [e2e-llm-inference-service] response_assertion=assert_model_field_matches("facebook/opt-125m"), [e2e-llm-inference-service] url_getter=get_model_routing_url, [e2e-llm-inference-service] extra_headers={ [e2e-llm-inference-service] MODEL_ROUTING_HEADER: f"publishers/{KSERVE_TEST_NAMESPACE}/models/facebook/opt-125m", [e2e-llm-inference-service] }, [e2e-llm-inference-service] peers=[ [e2e-llm-inference-service] TestCase( [e2e-llm-inference-service] base_refs=[ [e2e-llm-inference-service] "router-managed", [e2e-llm-inference-service] "workload-llmd-simulator", [e2e-llm-inference-service] "model-qwen2.5-0.5b", [e2e-llm-inference-service] ], [e2e-llm-inference-service] endpoint="/v1/completions", [e2e-llm-inference-service] prompt="KServe is a", [e2e-llm-inference-service] payload_formatter=completions_payload, [e2e-llm-inference-service] response_assertion=assert_model_field_matches( [e2e-llm-inference-service] "Qwen/Qwen2.5-0.5B-Instruct" [e2e-llm-inference-service] ), [e2e-llm-inference-service] url_getter=get_model_routing_url, [e2e-llm-inference-service] extra_headers={ [e2e-llm-inference-service] MODEL_ROUTING_HEADER: f"publishers/{KSERVE_TEST_NAMESPACE}/models/Qwen/Qwen2.5-0.5B-Instruct", [e2e-llm-inference-service] }, [e2e-llm-inference-service] ), [e2e-llm-inference-service] ], [e2e-llm-inference-service] ), [e2e-llm-inference-service] marks=[ [e2e-llm-inference-service] pytest.mark.cluster_cpu, [e2e-llm-inference-service] pytest.mark.cluster_single_node, [e2e-llm-inference-service] pytest.mark.llmd_simulator, [e2e-llm-inference-service] pytest.mark.model_routing, [e2e-llm-inference-service] ], [e2e-llm-inference-service] ), [e2e-llm-inference-service] # Model-based routing via X-Gateway-Model-Name header — /v1/chat/completions [e2e-llm-inference-service] pytest.param( [e2e-llm-inference-service] TestCase( [e2e-llm-inference-service] base_refs=[ [e2e-llm-inference-service] "router-managed", [e2e-llm-inference-service] "workload-llmd-simulator", [e2e-llm-inference-service] ], [e2e-llm-inference-service] endpoint="/v1/chat/completions", [e2e-llm-inference-service] prompt="What is KServe?", [e2e-llm-inference-service] payload_formatter=chat_completions_payload, [e2e-llm-inference-service] response_assertion=assert_model_field_matches("facebook/opt-125m"), [e2e-llm-inference-service] url_getter=get_model_routing_url, [e2e-llm-inference-service] extra_headers={ [e2e-llm-inference-service] MODEL_ROUTING_HEADER: f"publishers/{KSERVE_TEST_NAMESPACE}/models/facebook/opt-125m", [e2e-llm-inference-service] }, [e2e-llm-inference-service] peers=[ [e2e-llm-inference-service] TestCase( [e2e-llm-inference-service] base_refs=[ [e2e-llm-inference-service] "router-managed", [e2e-llm-inference-service] "workload-llmd-simulator", [e2e-llm-inference-service] "model-qwen2.5-0.5b", [e2e-llm-inference-service] ], [e2e-llm-inference-service] endpoint="/v1/chat/completions", [e2e-llm-inference-service] prompt="What is KServe?", [e2e-llm-inference-service] payload_formatter=chat_completions_payload, [e2e-llm-inference-service] response_assertion=assert_model_field_matches( [e2e-llm-inference-service] "Qwen/Qwen2.5-0.5B-Instruct" [e2e-llm-inference-service] ), [e2e-llm-inference-service] url_getter=get_model_routing_url, [e2e-llm-inference-service] extra_headers={ [e2e-llm-inference-service] MODEL_ROUTING_HEADER: f"publishers/{KSERVE_TEST_NAMESPACE}/models/Qwen/Qwen2.5-0.5B-Instruct", [e2e-llm-inference-service] }, [e2e-llm-inference-service] ), [e2e-llm-inference-service] ], [e2e-llm-inference-service] ), [e2e-llm-inference-service] marks=[ [e2e-llm-inference-service] pytest.mark.cluster_cpu, [e2e-llm-inference-service] pytest.mark.cluster_single_node, [e2e-llm-inference-service] pytest.mark.llmd_simulator, [e2e-llm-inference-service] pytest.mark.model_routing, [e2e-llm-inference-service] ], [e2e-llm-inference-service] ), [e2e-llm-inference-service] # Model-based routing via X-Gateway-Model-Name header — LoRA adapter [e2e-llm-inference-service] pytest.param( [e2e-llm-inference-service] TestCase( [e2e-llm-inference-service] base_refs=[ [e2e-llm-inference-service] "router-managed", [e2e-llm-inference-service] "workload-single-cpu", [e2e-llm-inference-service] "model-fb-opt-125m-with-lora-hf", [e2e-llm-inference-service] ], [e2e-llm-inference-service] endpoint="/v1/completions", [e2e-llm-inference-service] prompt="KServe is a", [e2e-llm-inference-service] model_name=f"publishers/{KSERVE_TEST_NAMESPACE}/models/lora-adapter-1", [e2e-llm-inference-service] payload_formatter=completions_payload, [e2e-llm-inference-service] response_assertion=assert_model_field_matches( [e2e-llm-inference-service] f"publishers/{KSERVE_TEST_NAMESPACE}/models/lora-adapter-1" [e2e-llm-inference-service] ), [e2e-llm-inference-service] url_getter=get_model_routing_url, [e2e-llm-inference-service] extra_headers={ [e2e-llm-inference-service] MODEL_ROUTING_HEADER: f"publishers/{KSERVE_TEST_NAMESPACE}/models/lora-adapter-1", [e2e-llm-inference-service] }, [e2e-llm-inference-service] ), [e2e-llm-inference-service] marks=[ [e2e-llm-inference-service] pytest.mark.cluster_cpu, [e2e-llm-inference-service] pytest.mark.cluster_single_node, [e2e-llm-inference-service] pytest.mark.model_routing, [e2e-llm-inference-service] pytest.mark.lora, [e2e-llm-inference-service] ], [e2e-llm-inference-service] ), [e2e-llm-inference-service] # Model-based routing via X-Gateway-Model-Name header — /v1/models (base + LoRA) [e2e-llm-inference-service] pytest.param( [e2e-llm-inference-service] TestCase( [e2e-llm-inference-service] base_refs=[ [e2e-llm-inference-service] "router-managed", [e2e-llm-inference-service] "workload-single-cpu", [e2e-llm-inference-service] "model-fb-opt-125m-with-lora-hf", [e2e-llm-inference-service] ], [e2e-llm-inference-service] endpoint="/v1/models", [e2e-llm-inference-service] response_assertion=assert_models_contains( [e2e-llm-inference-service] "facebook/opt-125m", [e2e-llm-inference-service] f"publishers/{KSERVE_TEST_NAMESPACE}/models/facebook/opt-125m", [e2e-llm-inference-service] "lora-adapter-1", [e2e-llm-inference-service] f"publishers/{KSERVE_TEST_NAMESPACE}/models/lora-adapter-1", [e2e-llm-inference-service] ), [e2e-llm-inference-service] url_getter=get_model_routing_url, [e2e-llm-inference-service] extra_headers={ [e2e-llm-inference-service] MODEL_ROUTING_HEADER: f"publishers/{KSERVE_TEST_NAMESPACE}/models/facebook/opt-125m", [e2e-llm-inference-service] }, [e2e-llm-inference-service] ), [e2e-llm-inference-service] marks=[ [e2e-llm-inference-service] pytest.mark.cluster_cpu, [e2e-llm-inference-service] pytest.mark.cluster_single_node, [e2e-llm-inference-service] pytest.mark.model_routing, [e2e-llm-inference-service] pytest.mark.lora, [e2e-llm-inference-service] ], [e2e-llm-inference-service] ), [e2e-llm-inference-service] # PVC storage tests -- validate direct PVC volume mount with real vLLM serving [e2e-llm-inference-service] pytest.param( [e2e-llm-inference-service] TestCase( [e2e-llm-inference-service] base_refs=[ [e2e-llm-inference-service] "router-managed", [e2e-llm-inference-service] "workload-single-cpu", [e2e-llm-inference-service] "model-pvc", [e2e-llm-inference-service] ], [e2e-llm-inference-service] prompt="KServe is a", [e2e-llm-inference-service] response_assertion=assert_200_with_choices, [e2e-llm-inference-service] before_test=[ensure_pvc_with_model], [e2e-llm-inference-service] ), [e2e-llm-inference-service] marks=[ [e2e-llm-inference-service] pytest.mark.cluster_cpu, [e2e-llm-inference-service] pytest.mark.cluster_single_node, [e2e-llm-inference-service] pytest.mark.pvc_storage, [e2e-llm-inference-service] ], [e2e-llm-inference-service] ), [e2e-llm-inference-service] pytest.param( [e2e-llm-inference-service] TestCase( [e2e-llm-inference-service] base_refs=[ [e2e-llm-inference-service] "router-managed", [e2e-llm-inference-service] "workload-pd-cpu", [e2e-llm-inference-service] "model-pvc", [e2e-llm-inference-service] ], [e2e-llm-inference-service] prompt="KServe is a", [e2e-llm-inference-service] response_assertion=assert_200_with_choices, [e2e-llm-inference-service] before_test=[ensure_pvc_with_model], [e2e-llm-inference-service] ), [e2e-llm-inference-service] marks=[ [e2e-llm-inference-service] pytest.mark.cluster_cpu, [e2e-llm-inference-service] pytest.mark.cluster_single_node, [e2e-llm-inference-service] pytest.mark.pvc_storage, [e2e-llm-inference-service] ], [e2e-llm-inference-service] ), [e2e-llm-inference-service] pytest.param( [e2e-llm-inference-service] TestCase( [e2e-llm-inference-service] base_refs=[ [e2e-llm-inference-service] "router-managed", [e2e-llm-inference-service] "workload-simulated-dp-ep-cpu", [e2e-llm-inference-service] "model-pvc", [e2e-llm-inference-service] ], [e2e-llm-inference-service] prompt="KServe is a", [e2e-llm-inference-service] before_test=[ensure_pvc_with_model], [e2e-llm-inference-service] ), [e2e-llm-inference-service] marks=[ [e2e-llm-inference-service] pytest.mark.cluster_cpu, [e2e-llm-inference-service] pytest.mark.cluster_multi_node, [e2e-llm-inference-service] pytest.mark.pvc_storage, [e2e-llm-inference-service] ], [e2e-llm-inference-service] ), [e2e-llm-inference-service] ], [e2e-llm-inference-service] indirect=["test_case"], [e2e-llm-inference-service] ids=generate_test_id, [e2e-llm-inference-service] ) [e2e-llm-inference-service] @log_execution [e2e-llm-inference-service] def test_llm_inference_service(test_case: TestCase): # noqa: F811 [e2e-llm-inference-service] inject_k8s_proxy() [e2e-llm-inference-service] [e2e-llm-inference-service] kserve_client = KServeClient( [e2e-llm-inference-service] config_file=os.environ.get("KUBECONFIG", "~/.kube/config"), [e2e-llm-inference-service] client_configuration=client.Configuration(), [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] service_name = test_case.llm_service.metadata.name [e2e-llm-inference-service] if not test_case.llm_service.metadata.annotations: [e2e-llm-inference-service] test_case.llm_service.metadata.annotations = {} [e2e-llm-inference-service] [e2e-llm-inference-service] test_case.llm_service.metadata.annotations[ [e2e-llm-inference-service] "security.opendatahub.io/enable-auth" [e2e-llm-inference-service] ] = "false" [e2e-llm-inference-service] prefix = test_case.log_prefix [e2e-llm-inference-service] [e2e-llm-inference-service] test_failed = False [e2e-llm-inference-service] try: [e2e-llm-inference-service] print(f"{prefix} Creating LLMInferenceService {service_name}") [e2e-llm-inference-service] create_llmisvc(kserve_client, test_case.llm_service) [e2e-llm-inference-service] print(f"{prefix} Waiting for LLMInferenceService {service_name} to be ready") [e2e-llm-inference-service] wait_for_llm_isvc_ready( [e2e-llm-inference-service] kserve_client, test_case.llm_service, test_case.wait_timeout [e2e-llm-inference-service] ) [e2e-llm-inference-service] print(f"{prefix} Waiting for model response from {service_name}") [e2e-llm-inference-service] > wait_for_model_response( [e2e-llm-inference-service] kserve_client, [e2e-llm-inference-service] test_case, [e2e-llm-inference-service] test_case.wait_timeout, [e2e-llm-inference-service] extra_headers=test_case.extra_headers, [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] llmisvc/test_llm_inference_service.py:816: [e2e-llm-inference-service] _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ [e2e-llm-inference-service] [e2e-llm-inference-service] args = (, TestCase(base_refs=['router-custom-route-timeout', ... {'name': 'model-fb-opt-125m-custom-route-928a8601'}]}, [e2e-llm-inference-service] 'status': None}, model_name='facebook/opt-125m'), 900) [e2e-llm-inference-service] kwargs = {'extra_headers': None}, func_name = 'wait_for_model_response' [e2e-llm-inference-service] timestamp_start = '2026-07-07T14:50:58.433371', start_time = 1783435858.4336658 [e2e-llm-inference-service] duration = 904.6219441890717, timestamp_end = '2026-07-07T15:06:03.055613' [e2e-llm-inference-service] [e2e-llm-inference-service] @functools.wraps(func) [e2e-llm-inference-service] def wrapper(*args, **kwargs): [e2e-llm-inference-service] func_name = func.__name__ [e2e-llm-inference-service] [e2e-llm-inference-service] timestamp_start = datetime.now().isoformat() [e2e-llm-inference-service] logger.info( [e2e-llm-inference-service] f"[{func_name}] [{timestamp_start}] start - args={args}, kwargs={kwargs}" [e2e-llm-inference-service] ) [e2e-llm-inference-service] start_time = time.time() [e2e-llm-inference-service] [e2e-llm-inference-service] try: [e2e-llm-inference-service] > result = func(*args, **kwargs) [e2e-llm-inference-service] [e2e-llm-inference-service] llmisvc/logging.py:40: [e2e-llm-inference-service] _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ [e2e-llm-inference-service] [e2e-llm-inference-service] kserve_client = [e2e-llm-inference-service] test_case = TestCase(base_refs=['router-custom-route-timeout', 'scheduler-managed', 'workload-single-cpu', 'model-fb-opt-125m'], p... {'name': 'model-fb-opt-125m-custom-route-928a8601'}]}, [e2e-llm-inference-service] 'status': None}, model_name='facebook/opt-125m') [e2e-llm-inference-service] timeout_seconds = 900, extra_headers = None [e2e-llm-inference-service] [e2e-llm-inference-service] @log_execution [e2e-llm-inference-service] def wait_for_model_response( [e2e-llm-inference-service] kserve_client: KServeClient, [e2e-llm-inference-service] test_case: TestCase, # noqa: F811 [e2e-llm-inference-service] timeout_seconds: int = 900, [e2e-llm-inference-service] extra_headers: Optional[Dict[str, str]] = None, [e2e-llm-inference-service] ) -> str: [e2e-llm-inference-service] def get_successful_response(): [e2e-llm-inference-service] try: [e2e-llm-inference-service] if test_case.url_getter: [e2e-llm-inference-service] service_url = test_case.url_getter(kserve_client, test_case.llm_service) [e2e-llm-inference-service] else: [e2e-llm-inference-service] service_url = get_llm_service_url(kserve_client, test_case.llm_service) [e2e-llm-inference-service] except Exception as e: [e2e-llm-inference-service] raise AssertionError(f"❌ Failed to get service URL: {e}") from e [e2e-llm-inference-service] [e2e-llm-inference-service] model_url = service_url + test_case.endpoint [e2e-llm-inference-service] [e2e-llm-inference-service] headers = {"Content-Type": "application/json"} [e2e-llm-inference-service] if extra_headers: [e2e-llm-inference-service] headers.update(extra_headers) [e2e-llm-inference-service] [e2e-llm-inference-service] if test_case.payload_formatter is not None: [e2e-llm-inference-service] test_payload = test_case.payload_formatter(test_case) [e2e-llm-inference-service] elif test_case.prompt is not None: [e2e-llm-inference-service] test_payload = { [e2e-llm-inference-service] "model": test_case.model_name [e2e-llm-inference-service] if not extra_headers or MODEL_ROUTING_HEADER not in extra_headers [e2e-llm-inference-service] else extra_headers[MODEL_ROUTING_HEADER], [e2e-llm-inference-service] "prompt": test_case.prompt, [e2e-llm-inference-service] "max_tokens": test_case.max_tokens, [e2e-llm-inference-service] } [e2e-llm-inference-service] else: [e2e-llm-inference-service] test_payload = None [e2e-llm-inference-service] [e2e-llm-inference-service] logger.info(f"Calling LLM service at {model_url} with payload {test_payload}") [e2e-llm-inference-service] try: [e2e-llm-inference-service] if test_payload is not None: [e2e-llm-inference-service] response = post_with_retry( [e2e-llm-inference-service] model_url, [e2e-llm-inference-service] headers=headers, [e2e-llm-inference-service] json_data=test_payload, [e2e-llm-inference-service] timeout=test_case.response_timeout, [e2e-llm-inference-service] ) [e2e-llm-inference-service] else: [e2e-llm-inference-service] response = get_with_retry( [e2e-llm-inference-service] model_url, [e2e-llm-inference-service] headers=headers, [e2e-llm-inference-service] timeout=test_case.response_timeout, [e2e-llm-inference-service] ) [e2e-llm-inference-service] except Exception as e: [e2e-llm-inference-service] logger.error(f"❌ Failed to call model: {e}") [e2e-llm-inference-service] raise AssertionError(f"❌ Failed to call model: {e}") from e [e2e-llm-inference-service] [e2e-llm-inference-service] logger.info(f"Model response is {response.status_code}: {response.text[:500]}") [e2e-llm-inference-service] [e2e-llm-inference-service] if 200 <= response.status_code < 300: [e2e-llm-inference-service] return response [e2e-llm-inference-service] raise AssertionError( [e2e-llm-inference-service] f"Service returned {response.status_code}: {response.text}" [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] > response = wait_for(get_successful_response, timeout=timeout_seconds, interval=5.0) [e2e-llm-inference-service] [e2e-llm-inference-service] llmisvc/test_llm_inference_service.py:1119: [e2e-llm-inference-service] _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ [e2e-llm-inference-service] [e2e-llm-inference-service] assertion_fn = .get_successful_response at 0x7f4f0db94860> [e2e-llm-inference-service] timeout = 900, interval = 5.0 [e2e-llm-inference-service] [e2e-llm-inference-service] def wait_for( [e2e-llm-inference-service] assertion_fn: Callable[[], Any], timeout: float = 5.0, interval: float = 0.1 [e2e-llm-inference-service] ) -> Any: [e2e-llm-inference-service] """Wait for the assertion to succeed within timeout.""" [e2e-llm-inference-service] deadline = time.time() + timeout [e2e-llm-inference-service] last_msg = None [e2e-llm-inference-service] while True: [e2e-llm-inference-service] try: [e2e-llm-inference-service] > return assertion_fn() [e2e-llm-inference-service] [e2e-llm-inference-service] llmisvc/test_llm_inference_service.py:1215: [e2e-llm-inference-service] _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ [e2e-llm-inference-service] [e2e-llm-inference-service] def get_successful_response(): [e2e-llm-inference-service] try: [e2e-llm-inference-service] if test_case.url_getter: [e2e-llm-inference-service] service_url = test_case.url_getter(kserve_client, test_case.llm_service) [e2e-llm-inference-service] else: [e2e-llm-inference-service] service_url = get_llm_service_url(kserve_client, test_case.llm_service) [e2e-llm-inference-service] except Exception as e: [e2e-llm-inference-service] raise AssertionError(f"❌ Failed to get service URL: {e}") from e [e2e-llm-inference-service] [e2e-llm-inference-service] model_url = service_url + test_case.endpoint [e2e-llm-inference-service] [e2e-llm-inference-service] headers = {"Content-Type": "application/json"} [e2e-llm-inference-service] if extra_headers: [e2e-llm-inference-service] headers.update(extra_headers) [e2e-llm-inference-service] [e2e-llm-inference-service] if test_case.payload_formatter is not None: [e2e-llm-inference-service] test_payload = test_case.payload_formatter(test_case) [e2e-llm-inference-service] elif test_case.prompt is not None: [e2e-llm-inference-service] test_payload = { [e2e-llm-inference-service] "model": test_case.model_name [e2e-llm-inference-service] if not extra_headers or MODEL_ROUTING_HEADER not in extra_headers [e2e-llm-inference-service] else extra_headers[MODEL_ROUTING_HEADER], [e2e-llm-inference-service] "prompt": test_case.prompt, [e2e-llm-inference-service] "max_tokens": test_case.max_tokens, [e2e-llm-inference-service] } [e2e-llm-inference-service] else: [e2e-llm-inference-service] test_payload = None [e2e-llm-inference-service] [e2e-llm-inference-service] logger.info(f"Calling LLM service at {model_url} with payload {test_payload}") [e2e-llm-inference-service] try: [e2e-llm-inference-service] if test_payload is not None: [e2e-llm-inference-service] response = post_with_retry( [e2e-llm-inference-service] model_url, [e2e-llm-inference-service] headers=headers, [e2e-llm-inference-service] json_data=test_payload, [e2e-llm-inference-service] timeout=test_case.response_timeout, [e2e-llm-inference-service] ) [e2e-llm-inference-service] else: [e2e-llm-inference-service] response = get_with_retry( [e2e-llm-inference-service] model_url, [e2e-llm-inference-service] headers=headers, [e2e-llm-inference-service] timeout=test_case.response_timeout, [e2e-llm-inference-service] ) [e2e-llm-inference-service] except Exception as e: [e2e-llm-inference-service] logger.error(f"❌ Failed to call model: {e}") [e2e-llm-inference-service] > raise AssertionError(f"❌ Failed to call model: {e}") from e [e2e-llm-inference-service] E AssertionError: ❌ Failed to call model: HTTPConnectionPool(host='ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com', port=80): Max retries exceeded with url: /kserve-ci-e2e-test/custom-route-timeout-test/v1/completions (Caused by ReadTimeoutError("HTTPConnectionPool(host='ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com', port=80): Read timed out. (read timeout=60)")) [e2e-llm-inference-service] [e2e-llm-inference-service] llmisvc/test_llm_inference_service.py:1109: AssertionError [e2e-llm-inference-service] ------------------------------ Captured log setup ------------------------------ [e2e-llm-inference-service] INFO e2e.llmisvc.logging:fixtures.py:1637 No HTTP proxy configured for k8s client [e2e-llm-inference-service] INFO e2e.llmisvc.logging:fixtures.py:1586 Checking LLMInferenceServiceConfig router-custom-route-timeout-cus-3c5d4892 in namespace kserve-ci-e2e-test [e2e-llm-inference-service] INFO e2e.llmisvc.logging:fixtures.py:1612 Resource not found, creating LLMInferenceServiceConfig router-custom-route-timeout-cus-3c5d4892 [e2e-llm-inference-service] INFO e2e.llmisvc.logging:fixtures.py:1622 ✓ Successfully created LLMInferenceServiceConfig router-custom-route-timeout-cus-3c5d4892 [e2e-llm-inference-service] INFO e2e.llmisvc.logging:fixtures.py:1586 Checking LLMInferenceServiceConfig scheduler-managed-custom-route-c8a904b6 in namespace kserve-ci-e2e-test [e2e-llm-inference-service] INFO e2e.llmisvc.logging:fixtures.py:1612 Resource not found, creating LLMInferenceServiceConfig scheduler-managed-custom-route-c8a904b6 [e2e-llm-inference-service] INFO e2e.llmisvc.logging:fixtures.py:1622 ✓ Successfully created LLMInferenceServiceConfig scheduler-managed-custom-route-c8a904b6 [e2e-llm-inference-service] INFO e2e.llmisvc.logging:fixtures.py:1586 Checking LLMInferenceServiceConfig workload-single-cpu-custom-rout-4c4b9f6e in namespace kserve-ci-e2e-test [e2e-llm-inference-service] INFO e2e.llmisvc.logging:fixtures.py:1612 Resource not found, creating LLMInferenceServiceConfig workload-single-cpu-custom-rout-4c4b9f6e [e2e-llm-inference-service] INFO e2e.llmisvc.logging:fixtures.py:1622 ✓ Successfully created LLMInferenceServiceConfig workload-single-cpu-custom-rout-4c4b9f6e [e2e-llm-inference-service] INFO e2e.llmisvc.logging:fixtures.py:1586 Checking LLMInferenceServiceConfig model-fb-opt-125m-custom-route-928a8601 in namespace kserve-ci-e2e-test [e2e-llm-inference-service] INFO e2e.llmisvc.logging:fixtures.py:1612 Resource not found, creating LLMInferenceServiceConfig model-fb-opt-125m-custom-route-928a8601 [e2e-llm-inference-service] INFO e2e.llmisvc.logging:fixtures.py:1622 ✓ Successfully created LLMInferenceServiceConfig model-fb-opt-125m-custom-route-928a8601 [e2e-llm-inference-service] ------------------------------ Captured log call ------------------------------- [e2e-llm-inference-service] INFO e2e.llmisvc.logging:logging.py:34 [test_llm_inference_service] [2026-07-07T14:48:48.660682] start - args=(), kwargs={'test_case': TestCase(base_refs=['router-custom-route-timeout', 'scheduler-managed', 'workload-single-cpu', 'model-fb-opt-125m'], prompt='KServe is a', service_name='custom-route-timeout-test', endpoint='/v1/completions', max_tokens=20, payload_formatter=None, response_assertion=, wait_timeout=900, response_timeout=60, extra_headers=None, url_getter=None, expected_gateway=None, before_test=[], after_test=[], peers=[], llm_service={'api_version': 'serving.kserve.io/v1alpha1', [e2e-llm-inference-service] 'kind': 'LLMInferenceService', [e2e-llm-inference-service] 'metadata': {'annotations': None, [e2e-llm-inference-service] 'creation_timestamp': None, [e2e-llm-inference-service] 'deletion_grace_period_seconds': None, [e2e-llm-inference-service] 'deletion_timestamp': None, [e2e-llm-inference-service] 'finalizers': None, [e2e-llm-inference-service] 'generate_name': None, [e2e-llm-inference-service] 'generation': None, [e2e-llm-inference-service] 'labels': None, [e2e-llm-inference-service] 'managed_fields': None, [e2e-llm-inference-service] 'name': 'custom-route-timeout-test', [e2e-llm-inference-service] 'namespace': 'kserve-ci-e2e-test', [e2e-llm-inference-service] 'owner_references': None, [e2e-llm-inference-service] 'resource_version': None, [e2e-llm-inference-service] 'self_link': None, [e2e-llm-inference-service] 'uid': None}, [e2e-llm-inference-service] 'spec': {'baseRefs': [{'name': 'router-custom-route-timeout-cus-3c5d4892'}, [e2e-llm-inference-service] {'name': 'scheduler-managed-custom-route-c8a904b6'}, [e2e-llm-inference-service] {'name': 'workload-single-cpu-custom-rout-4c4b9f6e'}, [e2e-llm-inference-service] {'name': 'model-fb-opt-125m-custom-route-928a8601'}]}, [e2e-llm-inference-service] 'status': None}, model_name='facebook/opt-125m')} [e2e-llm-inference-service] INFO e2e.llmisvc.logging:fixtures.py:1637 No HTTP proxy configured for k8s client [e2e-llm-inference-service] INFO e2e.llmisvc.logging:logging.py:34 [create_llmisvc] [2026-07-07T14:48:48.674078] start - args=(, {'api_version': 'serving.kserve.io/v1alpha1', [e2e-llm-inference-service] 'kind': 'LLMInferenceService', [e2e-llm-inference-service] 'metadata': {'annotations': {'security.opendatahub.io/enable-auth': 'false'}, [e2e-llm-inference-service] 'creation_timestamp': None, [e2e-llm-inference-service] 'deletion_grace_period_seconds': None, [e2e-llm-inference-service] 'deletion_timestamp': None, [e2e-llm-inference-service] 'finalizers': None, [e2e-llm-inference-service] 'generate_name': None, [e2e-llm-inference-service] 'generation': None, [e2e-llm-inference-service] 'labels': None, [e2e-llm-inference-service] 'managed_fields': None, [e2e-llm-inference-service] 'name': 'custom-route-timeout-test', [e2e-llm-inference-service] 'namespace': 'kserve-ci-e2e-test', [e2e-llm-inference-service] 'owner_references': None, [e2e-llm-inference-service] 'resource_version': None, [e2e-llm-inference-service] 'self_link': None, [e2e-llm-inference-service] 'uid': None}, [e2e-llm-inference-service] 'spec': {'baseRefs': [{'name': 'router-custom-route-timeout-cus-3c5d4892'}, [e2e-llm-inference-service] {'name': 'scheduler-managed-custom-route-c8a904b6'}, [e2e-llm-inference-service] {'name': 'workload-single-cpu-custom-rout-4c4b9f6e'}, [e2e-llm-inference-service] {'name': 'model-fb-opt-125m-custom-route-928a8601'}]}, [e2e-llm-inference-service] 'status': None}), kwargs={} [e2e-llm-inference-service] INFO e2e.llmisvc.logging:logging.py:43 [create_llmisvc] [2026-07-07T14:48:48.822332] end - ✅ in 0.148s [e2e-llm-inference-service] INFO e2e.llmisvc.logging:logging.py:34 [wait_for_llm_isvc_ready] [2026-07-07T14:48:48.822486] start - args=(, {'api_version': 'serving.kserve.io/v1alpha1', [e2e-llm-inference-service] 'kind': 'LLMInferenceService', [e2e-llm-inference-service] 'metadata': {'annotations': {'security.opendatahub.io/enable-auth': 'false'}, [e2e-llm-inference-service] 'creation_timestamp': None, [e2e-llm-inference-service] 'deletion_grace_period_seconds': None, [e2e-llm-inference-service] 'deletion_timestamp': None, [e2e-llm-inference-service] 'finalizers': None, [e2e-llm-inference-service] 'generate_name': None, [e2e-llm-inference-service] 'generation': None, [e2e-llm-inference-service] 'labels': None, [e2e-llm-inference-service] 'managed_fields': None, [e2e-llm-inference-service] 'name': 'custom-route-timeout-test', [e2e-llm-inference-service] 'namespace': 'kserve-ci-e2e-test', [e2e-llm-inference-service] 'owner_references': None, [e2e-llm-inference-service] 'resource_version': None, [e2e-llm-inference-service] 'self_link': None, [e2e-llm-inference-service] 'uid': None}, [e2e-llm-inference-service] 'spec': {'baseRefs': [{'name': 'router-custom-route-timeout-cus-3c5d4892'}, [e2e-llm-inference-service] {'name': 'scheduler-managed-custom-route-c8a904b6'}, [e2e-llm-inference-service] {'name': 'workload-single-cpu-custom-rout-4c4b9f6e'}, [e2e-llm-inference-service] {'name': 'model-fb-opt-125m-custom-route-928a8601'}]}, [e2e-llm-inference-service] 'status': None}, 900), kwargs={} [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1222 Waiting: No conditions found in status [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1222 Waiting: Missing true conditions: {'Ready', 'RouterReady', 'WorkloadsReady'}, expected {'Ready', 'RouterReady', 'WorkloadsReady'}, got [{'lastTransitionTime': '2026-07-07T14:49:00Z', 'message': 'The following HTTPRoutes are not ready: [kserve-ci-e2e-test/custom-route-timeout-test-kserve-route: "False" (reason "InvalidKind", message "referencing unsupported backendRef: group \\"inference.networking.x-k8s.io\\" kind \\"InferencePool\\"")]', 'reason': 'HTTPRoutesNotReady', 'severity': 'Info', 'status': 'False', 'type': 'HTTPRoutesReady'}, {'lastTransitionTime': '2026-07-07T14:49:00Z', 'message': 'Inference Pool kserve-ci-e2e-test/custom-route-timeout-test-inference-pool exists but no Gateway controller has accepted it yet', 'reason': 'WaitingForGateway', 'severity': 'Info', 'status': 'False', 'type': 'InferencePoolReady'}, {'lastTransitionTime': '2026-07-07T14:49:00Z', 'reason': 'Progressing', 'severity': 'Info', 'status': 'False', 'type': 'MainWorkloadReady'}, {'lastTransitionTime': '2026-07-07T14:49:00Z', 'severity': 'Info', 'status': 'True', 'type': 'PresetsCombined'}, {'lastTransitionTime': '2026-07-07T14:49:00Z', 'message': 'The following HTTPRoutes are not ready: [kserve-ci-e2e-test/custom-route-timeout-test-kserve-route: "False" (reason "InvalidKind", message "referencing unsupported backendRef: group \\"inference.networking.x-k8s.io\\" kind \\"InferencePool\\"")]', 'reason': 'HTTPRoutesNotReady', 'status': 'False', 'type': 'Ready'}, {'lastTransitionTime': '2026-07-07T14:49:00Z', 'message': 'The following HTTPRoutes are not ready: [kserve-ci-e2e-test/custom-route-timeout-test-kserve-route: "False" (reason "InvalidKind", message "referencing unsupported backendRef: group \\"inference.networking.x-k8s.io\\" kind \\"InferencePool\\"")]', 'reason': 'HTTPRoutesNotReady', 'status': 'False', 'type': 'RouterReady'}, {'lastTransitionTime': '2026-07-07T14:49:00Z', 'message': 'Deployment rollout in progress', 'reason': 'Progressing', 'severity': 'Info', 'status': 'False', 'type': 'SchedulerWorkloadReady'}, {'lastTransitionTime': '2026-07-07T14:49:00Z', 'reason': 'Progressing', 'status': 'False', 'type': 'WorkloadsReady'}] [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1222 Waiting: Missing true conditions: {'Ready', 'RouterReady', 'WorkloadsReady'}, expected {'Ready', 'RouterReady', 'WorkloadsReady'}, got [{'lastTransitionTime': '2026-07-07T14:49:13Z', 'message': 'The following HTTPRoutes are not ready: [kserve-ci-e2e-test/custom-route-timeout-test-kserve-route: "False" (reason "BackendNotFound", message "backend(custom-route-timeout-test-inference-pool-ip-45342e47.kserve-ci-e2e-test.svc.cluster.local) not found")]', 'reason': 'HTTPRoutesNotReady', 'severity': 'Info', 'status': 'False', 'type': 'HTTPRoutesReady'}, {'lastTransitionTime': '2026-07-07T14:49:13Z', 'severity': 'Info', 'status': 'True', 'type': 'InferencePoolReady'}, {'lastTransitionTime': '2026-07-07T14:49:13Z', 'message': 'Deployment does not have minimum availability.', 'reason': 'MinimumReplicasUnavailable', 'severity': 'Info', 'status': 'False', 'type': 'MainWorkloadReady'}, {'lastTransitionTime': '2026-07-07T14:49:00Z', 'severity': 'Info', 'status': 'True', 'type': 'PresetsCombined'}, {'lastTransitionTime': '2026-07-07T14:49:13Z', 'message': 'The following HTTPRoutes are not ready: [kserve-ci-e2e-test/custom-route-timeout-test-kserve-route: "False" (reason "BackendNotFound", message "backend(custom-route-timeout-test-inference-pool-ip-45342e47.kserve-ci-e2e-test.svc.cluster.local) not found")]', 'reason': 'HTTPRoutesNotReady', 'status': 'False', 'type': 'Ready'}, {'lastTransitionTime': '2026-07-07T14:49:13Z', 'message': 'The following HTTPRoutes are not ready: [kserve-ci-e2e-test/custom-route-timeout-test-kserve-route: "False" (reason "BackendNotFound", message "backend(custom-route-timeout-test-inference-pool-ip-45342e47.kserve-ci-e2e-test.svc.cluster.local) not found")]', 'reason': 'HTTPRoutesNotReady', 'status': 'False', 'type': 'RouterReady'}, {'lastTransitionTime': '2026-07-07T14:49:13Z', 'message': 'Deployment does not have minimum availability.', 'reason': 'MinimumReplicasUnavailable', 'severity': 'Info', 'status': 'False', 'type': 'SchedulerWorkloadReady'}, {'lastTransitionTime': '2026-07-07T14:49:13Z', 'message': 'Deployment does not have minimum availability.', 'reason': 'MinimumReplicasUnavailable', 'status': 'False', 'type': 'WorkloadsReady'}] [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1222 Waiting: Missing true conditions: {'Ready', 'RouterReady', 'WorkloadsReady'}, expected {'Ready', 'RouterReady', 'WorkloadsReady'}, got [{'lastTransitionTime': '2026-07-07T14:49:15Z', 'severity': 'Info', 'status': 'True', 'type': 'HTTPRoutesReady'}, {'lastTransitionTime': '2026-07-07T14:49:13Z', 'severity': 'Info', 'status': 'True', 'type': 'InferencePoolReady'}, {'lastTransitionTime': '2026-07-07T14:49:13Z', 'message': 'Deployment does not have minimum availability.', 'reason': 'MinimumReplicasUnavailable', 'severity': 'Info', 'status': 'False', 'type': 'MainWorkloadReady'}, {'lastTransitionTime': '2026-07-07T14:49:00Z', 'severity': 'Info', 'status': 'True', 'type': 'PresetsCombined'}, {'lastTransitionTime': '2026-07-07T14:49:15Z', 'message': 'Deployment does not have minimum availability.', 'reason': 'MinimumReplicasUnavailable', 'status': 'False', 'type': 'Ready'}, {'lastTransitionTime': '2026-07-07T14:49:15Z', 'message': 'Deployment does not have minimum availability.', 'reason': 'MinimumReplicasUnavailable', 'status': 'False', 'type': 'RouterReady'}, {'lastTransitionTime': '2026-07-07T14:49:13Z', 'message': 'Deployment does not have minimum availability.', 'reason': 'MinimumReplicasUnavailable', 'severity': 'Info', 'status': 'False', 'type': 'SchedulerWorkloadReady'}, {'lastTransitionTime': '2026-07-07T14:49:13Z', 'message': 'Deployment does not have minimum availability.', 'reason': 'MinimumReplicasUnavailable', 'status': 'False', 'type': 'WorkloadsReady'}] [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1222 Waiting: Missing true conditions: {'Ready', 'WorkloadsReady'}, expected {'Ready', 'RouterReady', 'WorkloadsReady'}, got [{'lastTransitionTime': '2026-07-07T14:49:15Z', 'severity': 'Info', 'status': 'True', 'type': 'HTTPRoutesReady'}, {'lastTransitionTime': '2026-07-07T14:49:13Z', 'severity': 'Info', 'status': 'True', 'type': 'InferencePoolReady'}, {'lastTransitionTime': '2026-07-07T14:49:13Z', 'message': 'Deployment does not have minimum availability.', 'reason': 'MinimumReplicasUnavailable', 'severity': 'Info', 'status': 'False', 'type': 'MainWorkloadReady'}, {'lastTransitionTime': '2026-07-07T14:49:00Z', 'severity': 'Info', 'status': 'True', 'type': 'PresetsCombined'}, {'lastTransitionTime': '2026-07-07T14:49:15Z', 'message': 'Deployment does not have minimum availability.', 'reason': 'MinimumReplicasUnavailable', 'status': 'False', 'type': 'Ready'}, {'lastTransitionTime': '2026-07-07T14:49:32Z', 'status': 'True', 'type': 'RouterReady'}, {'lastTransitionTime': '2026-07-07T14:49:32Z', 'severity': 'Info', 'status': 'True', 'type': 'SchedulerWorkloadReady'}, {'lastTransitionTime': '2026-07-07T14:49:13Z', 'message': 'Deployment does not have minimum availability.', 'reason': 'MinimumReplicasUnavailable', 'status': 'False', 'type': 'WorkloadsReady'}] [e2e-llm-inference-service] INFO e2e.llmisvc.logging:logging.py:43 [wait_for_llm_isvc_ready] [2026-07-07T14:50:58.433209] end - ✅ in 129.610s [e2e-llm-inference-service] INFO e2e.llmisvc.logging:logging.py:34 [wait_for_model_response] [2026-07-07T14:50:58.433371] start - args=(, TestCase(base_refs=['router-custom-route-timeout', 'scheduler-managed', 'workload-single-cpu', 'model-fb-opt-125m'], prompt='KServe is a', service_name='custom-route-timeout-test', endpoint='/v1/completions', max_tokens=20, payload_formatter=None, response_assertion=, wait_timeout=900, response_timeout=60, extra_headers=None, url_getter=None, expected_gateway=None, before_test=[], after_test=[], peers=[], llm_service={'api_version': 'serving.kserve.io/v1alpha1', [e2e-llm-inference-service] 'kind': 'LLMInferenceService', [e2e-llm-inference-service] 'metadata': {'annotations': {'security.opendatahub.io/enable-auth': 'false'}, [e2e-llm-inference-service] 'creation_timestamp': None, [e2e-llm-inference-service] 'deletion_grace_period_seconds': None, [e2e-llm-inference-service] 'deletion_timestamp': None, [e2e-llm-inference-service] 'finalizers': None, [e2e-llm-inference-service] 'generate_name': None, [e2e-llm-inference-service] 'generation': None, [e2e-llm-inference-service] 'labels': None, [e2e-llm-inference-service] 'managed_fields': None, [e2e-llm-inference-service] 'name': 'custom-route-timeout-test', [e2e-llm-inference-service] 'namespace': 'kserve-ci-e2e-test', [e2e-llm-inference-service] 'owner_references': None, [e2e-llm-inference-service] 'resource_version': None, [e2e-llm-inference-service] 'self_link': None, [e2e-llm-inference-service] 'uid': None}, [e2e-llm-inference-service] 'spec': {'baseRefs': [{'name': 'router-custom-route-timeout-cus-3c5d4892'}, [e2e-llm-inference-service] {'name': 'scheduler-managed-custom-route-c8a904b6'}, [e2e-llm-inference-service] {'name': 'workload-single-cpu-custom-rout-4c4b9f6e'}, [e2e-llm-inference-service] {'name': 'model-fb-opt-125m-custom-route-928a8601'}]}, [e2e-llm-inference-service] 'status': None}, model_name='facebook/opt-125m'), 900), kwargs={'extra_headers': None} [e2e-llm-inference-service] INFO e2e.llmisvc.logging:logging.py:34 [get_llm_service_url] [2026-07-07T14:50:58.433675] start - args=(, {'api_version': 'serving.kserve.io/v1alpha1', [e2e-llm-inference-service] 'kind': 'LLMInferenceService', [e2e-llm-inference-service] 'metadata': {'annotations': {'security.opendatahub.io/enable-auth': 'false'}, [e2e-llm-inference-service] 'creation_timestamp': None, [e2e-llm-inference-service] 'deletion_grace_period_seconds': None, [e2e-llm-inference-service] 'deletion_timestamp': None, [e2e-llm-inference-service] 'finalizers': None, [e2e-llm-inference-service] 'generate_name': None, [e2e-llm-inference-service] 'generation': None, [e2e-llm-inference-service] 'labels': None, [e2e-llm-inference-service] 'managed_fields': None, [e2e-llm-inference-service] 'name': 'custom-route-timeout-test', [e2e-llm-inference-service] 'namespace': 'kserve-ci-e2e-test', [e2e-llm-inference-service] 'owner_references': None, [e2e-llm-inference-service] 'resource_version': None, [e2e-llm-inference-service] 'self_link': None, [e2e-llm-inference-service] 'uid': None}, [e2e-llm-inference-service] 'spec': {'baseRefs': [{'name': 'router-custom-route-timeout-cus-3c5d4892'}, [e2e-llm-inference-service] {'name': 'scheduler-managed-custom-route-c8a904b6'}, [e2e-llm-inference-service] {'name': 'workload-single-cpu-custom-rout-4c4b9f6e'}, [e2e-llm-inference-service] {'name': 'model-fb-opt-125m-custom-route-928a8601'}]}, [e2e-llm-inference-service] 'status': None}), kwargs={} [e2e-llm-inference-service] INFO e2e.llmisvc.logging:logging.py:43 [get_llm_service_url] [2026-07-07T14:50:58.533778] end - ✅ in 0.100s [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1092 Calling LLM service at http://ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com/kserve-ci-e2e-test/custom-route-timeout-test/v1/completions with payload {'model': 'facebook/opt-125m', 'prompt': 'KServe is a', 'max_tokens': 20} [e2e-llm-inference-service] WARNING urllib3.connectionpool:connectionpool.py:868 Retrying (Retry(total=7, connect=None, read=None, redirect=None, status=None)) after connection broken by 'ReadTimeoutError("HTTPConnectionPool(host='ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com', port=80): Read timed out. (read timeout=60)")': /kserve-ci-e2e-test/custom-route-timeout-test/v1/completions [e2e-llm-inference-service] WARNING urllib3.connectionpool:connectionpool.py:868 Retrying (Retry(total=6, connect=None, read=None, redirect=None, status=None)) after connection broken by 'ReadTimeoutError("HTTPConnectionPool(host='ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com', port=80): Read timed out. (read timeout=60)")': /kserve-ci-e2e-test/custom-route-timeout-test/v1/completions [e2e-llm-inference-service] WARNING urllib3.connectionpool:connectionpool.py:868 Retrying (Retry(total=5, connect=None, read=None, redirect=None, status=None)) after connection broken by 'ReadTimeoutError("HTTPConnectionPool(host='ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com', port=80): Read timed out. (read timeout=60)")': /kserve-ci-e2e-test/custom-route-timeout-test/v1/completions [e2e-llm-inference-service] WARNING urllib3.connectionpool:connectionpool.py:868 Retrying (Retry(total=4, connect=None, read=None, redirect=None, status=None)) after connection broken by 'ReadTimeoutError("HTTPConnectionPool(host='ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com', port=80): Read timed out. (read timeout=60)")': /kserve-ci-e2e-test/custom-route-timeout-test/v1/completions [e2e-llm-inference-service] WARNING urllib3.connectionpool:connectionpool.py:868 Retrying (Retry(total=3, connect=None, read=None, redirect=None, status=None)) after connection broken by 'ReadTimeoutError("HTTPConnectionPool(host='ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com', port=80): Read timed out. (read timeout=60)")': /kserve-ci-e2e-test/custom-route-timeout-test/v1/completions [e2e-llm-inference-service] WARNING urllib3.connectionpool:connectionpool.py:868 Retrying (Retry(total=2, connect=None, read=None, redirect=None, status=None)) after connection broken by 'ReadTimeoutError("HTTPConnectionPool(host='ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com', port=80): Read timed out. (read timeout=60)")': /kserve-ci-e2e-test/custom-route-timeout-test/v1/completions [e2e-llm-inference-service] WARNING urllib3.connectionpool:connectionpool.py:868 Retrying (Retry(total=1, connect=None, read=None, redirect=None, status=None)) after connection broken by 'ReadTimeoutError("HTTPConnectionPool(host='ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com', port=80): Read timed out. (read timeout=60)")': /kserve-ci-e2e-test/custom-route-timeout-test/v1/completions [e2e-llm-inference-service] WARNING urllib3.connectionpool:connectionpool.py:868 Retrying (Retry(total=0, connect=None, read=None, redirect=None, status=None)) after connection broken by 'RemoteDisconnected('Remote end closed connection without response')': /kserve-ci-e2e-test/custom-route-timeout-test/v1/completions [e2e-llm-inference-service] ERROR e2e.llmisvc.logging:test_llm_inference_service.py:1108 ❌ Failed to call model: HTTPConnectionPool(host='ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com', port=80): Max retries exceeded with url: /kserve-ci-e2e-test/custom-route-timeout-test/v1/completions (Caused by ReadTimeoutError("HTTPConnectionPool(host='ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com', port=80): Read timed out. (read timeout=60)")) [e2e-llm-inference-service] ERROR e2e.llmisvc.logging:test_llm_inference_service.py:1219 Timed out waiting: ❌ Failed to call model: HTTPConnectionPool(host='ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com', port=80): Max retries exceeded with url: /kserve-ci-e2e-test/custom-route-timeout-test/v1/completions (Caused by ReadTimeoutError("HTTPConnectionPool(host='ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com', port=80): Read timed out. (read timeout=60)")) [e2e-llm-inference-service] ERROR e2e.llmisvc.logging:logging.py:48 [wait_for_model_response] [2026-07-07T15:06:03.055613] end - ❌ 904.622s: ❌ Failed to call model: HTTPConnectionPool(host='ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com', port=80): Max retries exceeded with url: /kserve-ci-e2e-test/custom-route-timeout-test/v1/completions (Caused by ReadTimeoutError("HTTPConnectionPool(host='ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com', port=80): Read timed out. (read timeout=60)")) [e2e-llm-inference-service] ERROR e2e.llmisvc.logging:test_llm_inference_service.py:831 [router-custom-route-timeout-scheduler-managed-workload-single-cpu-model-fb-opt-125m] ❌ ERROR: Failed to call llm inference service custom-route-timeout-test: ❌ Failed to call model: HTTPConnectionPool(host='ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com', port=80): Max retries exceeded with url: /kserve-ci-e2e-test/custom-route-timeout-test/v1/completions (Caused by ReadTimeoutError("HTTPConnectionPool(host='ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com', port=80): Read timed out. (read timeout=60)")) [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1240 🔍 # Diagnostics for 'custom-route-timeout-test' in 'kserve-ci-e2e-test' [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1241 --- [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1242 # LLMInferenceService custom-route-timeout-test [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1245 apiVersion: serving.kserve.io/v1alpha1 [e2e-llm-inference-service] kind: LLMInferenceService [e2e-llm-inference-service] metadata: [e2e-llm-inference-service] annotations: [e2e-llm-inference-service] security.opendatahub.io/enable-auth: 'false' [e2e-llm-inference-service] creationTimestamp: '2026-07-07T14:48:48Z' [e2e-llm-inference-service] finalizers: [e2e-llm-inference-service] - serving.kserve.io/llmisvc-finalizer [e2e-llm-inference-service] generation: 1 [e2e-llm-inference-service] managedFields: [e2e-llm-inference-service] - apiVersion: serving.kserve.io/v1alpha1 [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:metadata: [e2e-llm-inference-service] f:annotations: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:security.opendatahub.io/enable-auth: {} [e2e-llm-inference-service] f:spec: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:baseRefs: {} [e2e-llm-inference-service] manager: OpenAPI-Generator [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] time: '2026-07-07T14:48:48Z' [e2e-llm-inference-service] - apiVersion: serving.kserve.io/v1alpha2 [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:metadata: [e2e-llm-inference-service] f:finalizers: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] v:"serving.kserve.io/llmisvc-finalizer": {} [e2e-llm-inference-service] manager: manager [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] time: '2026-07-07T14:48:48Z' [e2e-llm-inference-service] - apiVersion: serving.kserve.io/v1alpha2 [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:status: [e2e-llm-inference-service] f:addresses: {} [e2e-llm-inference-service] f:annotations: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:serving.kserve.io/config-llm-decode-template: {} [e2e-llm-inference-service] f:serving.kserve.io/config-llm-decode-worker-data-parallel: {} [e2e-llm-inference-service] f:serving.kserve.io/config-llm-prefill-template: {} [e2e-llm-inference-service] f:serving.kserve.io/config-llm-prefill-worker-data-parallel: {} [e2e-llm-inference-service] f:serving.kserve.io/config-llm-router-route: {} [e2e-llm-inference-service] f:serving.kserve.io/config-llm-scheduler: {} [e2e-llm-inference-service] f:serving.kserve.io/config-llm-template: {} [e2e-llm-inference-service] f:serving.kserve.io/config-llm-tracing: {} [e2e-llm-inference-service] f:serving.kserve.io/config-llm-worker-data-parallel: {} [e2e-llm-inference-service] f:appliedConfigs: {} [e2e-llm-inference-service] f:conditions: {} [e2e-llm-inference-service] f:observedGeneration: {} [e2e-llm-inference-service] f:router: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:gateways: {} [e2e-llm-inference-service] f:scheduler: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:inferencePool: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:group: {} [e2e-llm-inference-service] f:kind: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:service: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:group: {} [e2e-llm-inference-service] f:kind: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:url: {} [e2e-llm-inference-service] f:workloads: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:primary: {} [e2e-llm-inference-service] f:scheduler: {} [e2e-llm-inference-service] f:service: {} [e2e-llm-inference-service] manager: manager [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] subresource: status [e2e-llm-inference-service] time: '2026-07-07T14:50:58Z' [e2e-llm-inference-service] name: custom-route-timeout-test [e2e-llm-inference-service] namespace: kserve-ci-e2e-test [e2e-llm-inference-service] resourceVersion: '50907' [e2e-llm-inference-service] uid: 986f2d2c-a6d1-463d-9850-3aad6459eace [e2e-llm-inference-service] spec: [e2e-llm-inference-service] baseRefs: [e2e-llm-inference-service] - name: router-custom-route-timeout-cus-3c5d4892 [e2e-llm-inference-service] - name: scheduler-managed-custom-route-c8a904b6 [e2e-llm-inference-service] - name: workload-single-cpu-custom-rout-4c4b9f6e [e2e-llm-inference-service] - name: model-fb-opt-125m-custom-route-928a8601 [e2e-llm-inference-service] model: [e2e-llm-inference-service] name: custom-route-timeout-test [e2e-llm-inference-service] uri: '' [e2e-llm-inference-service] status: [e2e-llm-inference-service] addresses: [e2e-llm-inference-service] - name: gateway-external [e2e-llm-inference-service] url: http://ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com/kserve-ci-e2e-test/custom-route-timeout-test [e2e-llm-inference-service] - name: gateway-internal [e2e-llm-inference-service] url: http://openshift-ai-inference-openshift-default.openshift-ingress.svc.cluster.local/kserve-ci-e2e-test/custom-route-timeout-test [e2e-llm-inference-service] annotations: [e2e-llm-inference-service] serving.kserve.io/config-llm-decode-template: kserve-config-llm-decode-template [e2e-llm-inference-service] serving.kserve.io/config-llm-decode-worker-data-parallel: kserve-config-llm-decode-worker-data-parallel [e2e-llm-inference-service] serving.kserve.io/config-llm-prefill-template: kserve-config-llm-prefill-template [e2e-llm-inference-service] serving.kserve.io/config-llm-prefill-worker-data-parallel: kserve-config-llm-prefill-worker-data-parallel [e2e-llm-inference-service] serving.kserve.io/config-llm-router-route: kserve-config-llm-router-route [e2e-llm-inference-service] serving.kserve.io/config-llm-scheduler: kserve-config-llm-scheduler [e2e-llm-inference-service] serving.kserve.io/config-llm-template: kserve-config-llm-template [e2e-llm-inference-service] serving.kserve.io/config-llm-tracing: kserve-config-llm-tracing [e2e-llm-inference-service] serving.kserve.io/config-llm-worker-data-parallel: kserve-config-llm-worker-data-parallel [e2e-llm-inference-service] conditions: [e2e-llm-inference-service] - lastTransitionTime: '2026-07-07T14:49:15Z' [e2e-llm-inference-service] severity: Info [e2e-llm-inference-service] status: 'True' [e2e-llm-inference-service] type: HTTPRoutesReady [e2e-llm-inference-service] - lastTransitionTime: '2026-07-07T14:49:13Z' [e2e-llm-inference-service] severity: Info [e2e-llm-inference-service] status: 'True' [e2e-llm-inference-service] type: InferencePoolReady [e2e-llm-inference-service] - lastTransitionTime: '2026-07-07T14:50:58Z' [e2e-llm-inference-service] severity: Info [e2e-llm-inference-service] status: 'True' [e2e-llm-inference-service] type: MainWorkloadReady [e2e-llm-inference-service] - lastTransitionTime: '2026-07-07T14:49:00Z' [e2e-llm-inference-service] severity: Info [e2e-llm-inference-service] status: 'True' [e2e-llm-inference-service] type: PresetsCombined [e2e-llm-inference-service] - lastTransitionTime: '2026-07-07T14:50:58Z' [e2e-llm-inference-service] status: 'True' [e2e-llm-inference-service] type: Ready [e2e-llm-inference-service] - lastTransitionTime: '2026-07-07T14:49:32Z' [e2e-llm-inference-service] status: 'True' [e2e-llm-inference-service] type: RouterReady [e2e-llm-inference-service] - lastTransitionTime: '2026-07-07T14:49:32Z' [e2e-llm-inference-service] severity: Info [e2e-llm-inference-service] status: 'True' [e2e-llm-inference-service] type: SchedulerWorkloadReady [e2e-llm-inference-service] - lastTransitionTime: '2026-07-07T14:50:58Z' [e2e-llm-inference-service] status: 'True' [e2e-llm-inference-service] type: WorkloadsReady [e2e-llm-inference-service] observedGeneration: 1 [e2e-llm-inference-service] url: http://ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com/kserve-ci-e2e-test/custom-route-timeout-test [e2e-llm-inference-service] [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:44 TIME NAMESPACE SOURCE TYPE REASON MESSAGE [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:45 -------------------------------------------------------------------------------------------------- [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 None kserve-ci-e2e-test Normal Scheduled Successfully assigned kserve-ci-e2e-test/auth-disabled-test-kserve-658fdbccfc-7k7fm to ip-10-0-139-47.ec2.internal [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:28:27 kserve-ci-e2e-test multus Normal AddedInterface Add eth0 [10.134.0.29/23] from ovn-kubernetes [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:28:28 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Pulled Container image "quay.io/opendatahub/kserve-storage-initializer@sha256:13ad765c0f323200e1db7b45c99661287e65c5158d6b54b16139829636d20be4" already present on machine [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:28:28 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Created Created container: storage-initializer [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:28:28 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Started Started container storage-initializer [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:28:31 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Pulled Container image "public.ecr.aws/q9t5s3a7/vllm-cpu-release-repo:v0.19.0" already present on machine [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:28:32 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Created Created container: main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:28:32 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Started Started container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:30:17 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Warning Unhealthy Startup probe failed: Get "https://10.134.0.29:8000/health": dial tcp 10.134.0.29:8000: connect: connection refused [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:30:32 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Killing Stopping container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:30:47 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Warning Unhealthy Readiness probe failed: Get "https://10.134.0.29:8000/health": dial tcp 10.134.0.29:8000: connect: connection refused [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:28:27 kserve-ci-e2e-test replicaset-controller Normal SuccessfulCreate Created pod: auth-disabled-test-kserve-658fdbccfc-7k7fm [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 None kserve-ci-e2e-test Normal Scheduled Successfully assigned kserve-ci-e2e-test/auth-disabled-test-kserve-router-scheduler-58c676b848-mfsnv to ip-10-0-133-25.ec2.internal [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:28:28 kserve-ci-e2e-test multus Normal AddedInterface Add eth0 [10.132.0.51/23] from ovn-kubernetes [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:28:28 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Pulled Container image "quay.io/opendatahub/kserve-storage-initializer@sha256:13ad765c0f323200e1db7b45c99661287e65c5158d6b54b16139829636d20be4" already present on machine [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:28:28 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Created Created container: storage-initializer [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:28:28 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Started Started container storage-initializer [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:28:29 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Pulled Container image "ghcr.io/llm-d/llm-d-router-endpoint-picker:v0.9.0-rc.2" already present on machine [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:28:29 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Created Created container: main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:28:29 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Started Started container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:30:32 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Killing Stopping container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:30:32 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Warning FailedPreStopHook PreStopHook failed [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:28:28 kserve-ci-e2e-test replicaset-controller Normal SuccessfulCreate Created pod: auth-disabled-test-kserve-router-scheduler-58c676b848-mfsnv [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:28:27 kserve-ci-e2e-test deployment-controller Normal ScalingReplicaSet Scaled up replica set auth-disabled-test-kserve-router-scheduler-58c676b848 from 0 to 1 [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:28:27 kserve-ci-e2e-test deployment-controller Normal ScalingReplicaSet Scaled up replica set auth-disabled-test-kserve-658fdbccfc from 0 to 1 [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:28:27 kserve-ci-e2e-test OpenDataHubModelController Warning ReconcileError Failed to reconcile LLMInferenceService: 1 error occurred: * failed to get HTTPRoute for AuthPolicy auth-disabled-test-kserve-route-authn: failed to get HTTPRoute kserve-ci-e2e-test/auth-disabled-test-kserve-route: HTTPRoute.gateway.networking.k8s.io "auth-disabled-test-kserve-route" not found [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:28:27 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Secret kserve-ci-e2e-test/auth-disabled-test-kserve-self-signed-certs [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:28:27 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Deployment kserve-ci-e2e-test/auth-disabled-test-kserve [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:28:27 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Service kserve-ci-e2e-test/auth-disabled-test-kserve-workload-svc [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:28:27 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.ServiceAccount kserve-ci-e2e-test/auth-disabled-test-epp-sa [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:28:27 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.ClusterRoleBinding /kserve-ci-e2e-test-auth-disabled-test-epp-auth-rb [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:28:27 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Role kserve-ci-e2e-test/auth-disabled-test-epp-role [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:28:27 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.RoleBinding kserve-ci-e2e-test/auth-disabled-test-epp-rb [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:28:27 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Deployment kserve-ci-e2e-test/auth-disabled-test-kserve-router-scheduler [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:28:27 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Service kserve-ci-e2e-test/auth-disabled-test-epp-service [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:28:38 kserve-ci-e2e-test LLMInferenceServiceController Normal Created (combined from similar events): Created v1.DestinationRule kserve-ci-e2e-test/auth-disabled-test-kserve-shadow-svc [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:28:38 kserve-ci-e2e-test LLMInferenceServiceController Normal Updated Updated v1.Secret kserve-ci-e2e-test/auth-disabled-test-kserve-self-signed-certs [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:28:38 kserve-ci-e2e-test LLMInferenceServiceController Normal Updated Updated v1.HTTPRoute kserve-ci-e2e-test/auth-disabled-test-kserve-route [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:30:28 kserve-ci-e2e-test LLMInferenceServiceController Normal LLMInferenceServiceReady LLMInferenceService [auth-disabled-test] is Ready [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:30:32 kserve-ci-e2e-test LLMInferenceServiceController Normal Deleted Deleted v1.ClusterRoleBinding /kserve-ci-e2e-test-auth-disabled-test-epp-auth-rb [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 None kserve-ci-e2e-test Normal Scheduled Successfully assigned kserve-ci-e2e-test/auth-enabled-test-kserve-87bd45bf5-zxvsp to ip-10-0-139-47.ec2.internal [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:33 kserve-ci-e2e-test multus Normal AddedInterface Add eth0 [10.134.0.24/23] from ovn-kubernetes [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:33 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Pulled Container image "quay.io/opendatahub/kserve-storage-initializer@sha256:13ad765c0f323200e1db7b45c99661287e65c5158d6b54b16139829636d20be4" already present on machine [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:33 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Created Created container: storage-initializer [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:33 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Started Started container storage-initializer [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:39 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Pulling Pulling image "public.ecr.aws/q9t5s3a7/vllm-cpu-release-repo:v0.19.0" [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:23:08 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Pulled Successfully pulled image "public.ecr.aws/q9t5s3a7/vllm-cpu-release-repo:v0.19.0" in 28.615s (28.615s including waiting). Image size: 3531177328 bytes. [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:23:08 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Created Created container: main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:23:08 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Started Started container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:24:53 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Warning Unhealthy Startup probe failed: Get "https://10.134.0.24:8000/health": dial tcp 10.134.0.24:8000: connect: connection refused [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:10 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Killing Stopping container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:25 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Warning Unhealthy Readiness probe failed: Get "https://10.134.0.24:8000/health": dial tcp 10.134.0.24:8000: connect: connection refused [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:32 kserve-ci-e2e-test replicaset-controller Normal SuccessfulCreate Created pod: auth-enabled-test-kserve-87bd45bf5-zxvsp [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 None kserve-ci-e2e-test Normal Scheduled Successfully assigned kserve-ci-e2e-test/auth-enabled-test-kserve-router-scheduler-6bbc585b87-nf5lr to ip-10-0-133-25.ec2.internal [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:33 kserve-ci-e2e-test multus Normal AddedInterface Add eth0 [10.132.0.47/23] from ovn-kubernetes [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:33 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Pulled Container image "quay.io/opendatahub/kserve-storage-initializer@sha256:13ad765c0f323200e1db7b45c99661287e65c5158d6b54b16139829636d20be4" already present on machine [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:33 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Created Created container: storage-initializer [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:33 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Started Started container storage-initializer [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:34 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Pulling Pulling image "ghcr.io/llm-d/llm-d-router-endpoint-picker:v0.9.0-rc.2" [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:35 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Pulled Successfully pulled image "ghcr.io/llm-d/llm-d-router-endpoint-picker:v0.9.0-rc.2" in 1.131s (1.131s including waiting). Image size: 75073927 bytes. [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:35 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Created Created container: main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:35 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Started Started container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:10 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Killing Stopping container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:10 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Warning FailedPreStopHook PreStopHook failed [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:33 kserve-ci-e2e-test replicaset-controller Normal SuccessfulCreate Created pod: auth-enabled-test-kserve-router-scheduler-6bbc585b87-nf5lr [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:33 kserve-ci-e2e-test deployment-controller Normal ScalingReplicaSet Scaled up replica set auth-enabled-test-kserve-router-scheduler-6bbc585b87 from 0 to 1 [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:32 kserve-ci-e2e-test deployment-controller Normal ScalingReplicaSet Scaled up replica set auth-enabled-test-kserve-87bd45bf5 from 0 to 1 [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:32 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Secret kserve-ci-e2e-test/auth-enabled-test-kserve-self-signed-certs [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:32 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Deployment kserve-ci-e2e-test/auth-enabled-test-kserve [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:32 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Service kserve-ci-e2e-test/auth-enabled-test-kserve-workload-svc [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:32 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.ServiceAccount kserve-ci-e2e-test/auth-enabled-test-epp-sa [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:33 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.ClusterRoleBinding /kserve-ci-e2e-test-auth-enabled-test-epp-auth-rb [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:33 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Role kserve-ci-e2e-test/auth-enabled-test-epp-role [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:33 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.RoleBinding kserve-ci-e2e-test/auth-enabled-test-epp-rb [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:33 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Deployment kserve-ci-e2e-test/auth-enabled-test-kserve-router-scheduler [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:33 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Service kserve-ci-e2e-test/auth-enabled-test-epp-service [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:43 kserve-ci-e2e-test LLMInferenceServiceController Normal Created (combined from similar events): Created v1.DestinationRule kserve-ci-e2e-test/auth-enabled-test-kserve-shadow-svc [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:42 kserve-ci-e2e-test LLMInferenceServiceController Normal Updated Updated v1.Secret kserve-ci-e2e-test/auth-enabled-test-kserve-self-signed-certs [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:43 kserve-ci-e2e-test LLMInferenceServiceController Normal Updated Updated v1.HTTPRoute kserve-ci-e2e-test/auth-enabled-test-kserve-route [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:05 kserve-ci-e2e-test LLMInferenceServiceController Normal LLMInferenceServiceReady LLMInferenceService [auth-enabled-test] is Ready [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:10 kserve-ci-e2e-test LLMInferenceServiceController Normal Deleted Deleted v1.ClusterRoleBinding /kserve-ci-e2e-test-auth-enabled-test-epp-auth-rb [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 None kserve-ci-e2e-test Normal Scheduled Successfully assigned kserve-ci-e2e-test/auth-invalid-token-test-kserve-6467b8fbf9-jhlvl to ip-10-0-139-47.ec2.internal [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:44 kserve-ci-e2e-test multus Normal AddedInterface Add eth0 [10.134.0.27/23] from ovn-kubernetes [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:44 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Pulled Container image "quay.io/opendatahub/kserve-storage-initializer@sha256:13ad765c0f323200e1db7b45c99661287e65c5158d6b54b16139829636d20be4" already present on machine [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:44 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Created Created container: storage-initializer [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:44 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Started Started container storage-initializer [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:49 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Pulled Container image "public.ecr.aws/q9t5s3a7/vllm-cpu-release-repo:v0.19.0" already present on machine [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:49 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Created Created container: main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:49 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Started Started container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:34 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Warning Unhealthy Startup probe failed: Get "https://10.134.0.27:8000/health": dial tcp 10.134.0.27:8000: connect: connection refused [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:46 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Killing Stopping container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:28:03 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Warning Unhealthy Readiness probe failed: Get "https://10.134.0.27:8000/health": context deadline exceeded [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:43 kserve-ci-e2e-test replicaset-controller Normal SuccessfulCreate Created pod: auth-invalid-token-test-kserve-6467b8fbf9-jhlvl [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:43 kserve-ci-e2e-test replicaset-controller Normal SuccessfulCreate Created pod: auth-invalid-token-test-kserve-router-scheduler-7c68b57cd5h6v2p [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 None kserve-ci-e2e-test Normal Scheduled Successfully assigned kserve-ci-e2e-test/auth-invalid-token-test-kserve-router-scheduler-7c68b57cd5h6v2p to ip-10-0-133-25.ec2.internal [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:44 kserve-ci-e2e-test multus Normal AddedInterface Add eth0 [10.132.0.49/23] from ovn-kubernetes [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:44 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Pulled Container image "quay.io/opendatahub/kserve-storage-initializer@sha256:13ad765c0f323200e1db7b45c99661287e65c5158d6b54b16139829636d20be4" already present on machine [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:44 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Created Created container: storage-initializer [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:44 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Started Started container storage-initializer [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:45 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Pulled Container image "ghcr.io/llm-d/llm-d-router-endpoint-picker:v0.9.0-rc.2" already present on machine [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:45 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Created Created container: main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:45 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Started Started container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:46 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Killing Stopping container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:46 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Warning FailedPreStopHook PreStopHook failed [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:43 kserve-ci-e2e-test deployment-controller Normal ScalingReplicaSet Scaled up replica set auth-invalid-token-test-kserve-router-scheduler-7c68b57cd5 from 0 to 1 [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:43 kserve-ci-e2e-test deployment-controller Normal ScalingReplicaSet Scaled up replica set auth-invalid-token-test-kserve-6467b8fbf9 from 0 to 1 [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:43 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Secret kserve-ci-e2e-test/auth-invalid-token-test-kserve-self-signed-certs [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:43 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Deployment kserve-ci-e2e-test/auth-invalid-token-test-kserve [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:43 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Service kserve-ci-e2e-test/auth-invalid-token-test-kserve-workload-svc [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:43 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.ServiceAccount kserve-ci-e2e-test/auth-invalid-token-test-epp-sa [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:43 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.ClusterRoleBinding /kserve-ci-e2e-test-auth-invalid-token-test-epp-auth-rb [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:43 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Role kserve-ci-e2e-test/auth-invalid-token-test-epp-role [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:43 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.RoleBinding kserve-ci-e2e-test/auth-invalid-token-test-epp-rb [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:43 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Deployment kserve-ci-e2e-test/auth-invalid-token-test-kserve-router-scheduler [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:43 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Service kserve-ci-e2e-test/auth-invalid-token-test-epp-service [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:26:10 kserve-ci-e2e-test LLMInferenceServiceController Normal Created (combined from similar events): Created v1.DestinationRule kserve-ci-e2e-test/auth-invalid-token-test-kserve-shadow-svc [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:26:09 kserve-ci-e2e-test LLMInferenceServiceController Normal Updated Updated v1.Secret kserve-ci-e2e-test/auth-invalid-token-test-kserve-self-signed-certs [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:26:09 kserve-ci-e2e-test LLMInferenceServiceController Normal Updated Updated v1.HTTPRoute kserve-ci-e2e-test/auth-invalid-token-test-kserve-route [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:45 kserve-ci-e2e-test LLMInferenceServiceController Normal LLMInferenceServiceReady LLMInferenceService [auth-invalid-token-test] is Ready [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:46 kserve-ci-e2e-test LLMInferenceServiceController Normal Deleted Deleted v1.ClusterRoleBinding /kserve-ci-e2e-test-auth-invalid-token-test-epp-auth-rb [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 None kserve-ci-e2e-test Normal Scheduled Successfully assigned kserve-ci-e2e-test/custom-route-timeout-test-kserve-7f56b8c469-hkpbl to ip-10-0-139-47.ec2.internal [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:48:57 kserve-ci-e2e-test multus Normal AddedInterface Add eth0 [10.134.0.33/23] from ovn-kubernetes [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:48:57 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Pulled Container image "quay.io/opendatahub/kserve-storage-initializer@sha256:13ad765c0f323200e1db7b45c99661287e65c5158d6b54b16139829636d20be4" already present on machine [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:48:57 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Created Created container: storage-initializer [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:48:57 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Started Started container storage-initializer [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:49:02 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Pulled Container image "public.ecr.aws/q9t5s3a7/vllm-cpu-release-repo:v0.19.0" already present on machine [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:49:02 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Created Created container: main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:49:02 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Started Started container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:50:47 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Warning Unhealthy Startup probe failed: Get "https://10.134.0.33:8000/health": dial tcp 10.134.0.33:8000: connect: connection refused [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:48:57 kserve-ci-e2e-test replicaset-controller Normal SuccessfulCreate Created pod: custom-route-timeout-test-kserve-7f56b8c469-hkpbl [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:48:57 kserve-ci-e2e-test replicaset-controller Normal SuccessfulCreate Created pod: custom-route-timeout-test-kserve-router-scheduler-6578bddfnbxnb [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 None kserve-ci-e2e-test Normal Scheduled Successfully assigned kserve-ci-e2e-test/custom-route-timeout-test-kserve-router-scheduler-6578bddfnbxnb to ip-10-0-133-25.ec2.internal [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:48:58 kserve-ci-e2e-test multus Normal AddedInterface Add eth0 [10.132.0.55/23] from ovn-kubernetes [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:48:58 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Pulled Container image "quay.io/opendatahub/kserve-storage-initializer@sha256:13ad765c0f323200e1db7b45c99661287e65c5158d6b54b16139829636d20be4" already present on machine [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:48:58 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Created Created container: storage-initializer [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:48:58 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Started Started container storage-initializer [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:48:59 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Pulled Container image "ghcr.io/llm-d/llm-d-router-endpoint-picker:v0.9.0-rc.2" already present on machine [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:48:59 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Created Created container: main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:48:59 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Started Started container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:48:57 kserve-ci-e2e-test deployment-controller Normal ScalingReplicaSet Scaled up replica set custom-route-timeout-test-kserve-router-scheduler-6578bddf5b from 0 to 1 [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:48:57 kserve-ci-e2e-test deployment-controller Normal ScalingReplicaSet Scaled up replica set custom-route-timeout-test-kserve-7f56b8c469 from 0 to 1 [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:48:53 kserve-ci-e2e-test OpenDataHubModelController Warning ReconcileError Failed to reconcile LLMInferenceService: 1 error occurred: * failed to get HTTPRoute for AuthPolicy custom-route-timeout-test-kserve-route-authn: failed to get HTTPRoute kserve-ci-e2e-test/custom-route-timeout-test-kserve-route: HTTPRoute.gateway.networking.k8s.io "custom-route-timeout-test-kserve-route" not found [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:48:57 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Secret kserve-ci-e2e-test/custom-route-timeout-test-kserve-self-signed-certs [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:48:57 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Deployment kserve-ci-e2e-test/custom-route-timeout-test-kserve [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:48:57 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Service kserve-ci-e2e-test/custom-route-timeout-test-kserve-workload-svc [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:48:57 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.ServiceAccount kserve-ci-e2e-test/custom-route-timeout-test-epp-sa [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:48:57 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.ClusterRoleBinding /kserve-ci-e2e-test-custom-route-timeout-test-epp-auth-rb [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:48:57 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Role kserve-ci-e2e-test/custom-route-timeout-test-epp-role [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:48:57 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.RoleBinding kserve-ci-e2e-test/custom-route-timeout-test-epp-rb [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:48:57 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Deployment kserve-ci-e2e-test/custom-route-timeout-test-kserve-router-scheduler [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:48:57 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Service kserve-ci-e2e-test/custom-route-timeout-test-epp-service [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:49:12 kserve-ci-e2e-test LLMInferenceServiceController Normal Created (combined from similar events): Created v1.DestinationRule kserve-ci-e2e-test/custom-route-timeout-test-kserve-shadow-svc [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:49:11 kserve-ci-e2e-test LLMInferenceServiceController Normal Updated Updated v1.Secret kserve-ci-e2e-test/custom-route-timeout-test-kserve-self-signed-certs [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:49:11 kserve-ci-e2e-test LLMInferenceServiceController Normal Updated Updated v1.HTTPRoute kserve-ci-e2e-test/custom-route-timeout-test-kserve-route [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:50:58 kserve-ci-e2e-test LLMInferenceServiceController Normal LLMInferenceServiceReady LLMInferenceService [custom-route-timeout-test] is Ready [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 None kserve-ci-e2e-test Normal Scheduled Successfully assigned kserve-ci-e2e-test/e2e-pvc-model-download-6ls5n to ip-10-0-133-25.ec2.internal [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 15:02:31 kserve-ci-e2e-test attachdetach-controller Normal SuccessfulAttachVolume AttachVolume.Attach succeeded for volume "pvc-20b17380-3f46-45a4-a8e5-36f5cb935bca" [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 15:02:34 kserve-ci-e2e-test multus Normal AddedInterface Add eth0 [10.132.0.56/23] from ovn-kubernetes [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 15:02:34 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Pulled Container image "quay.io/opendatahub/kserve-storage-initializer@sha256:13ad765c0f323200e1db7b45c99661287e65c5158d6b54b16139829636d20be4" already present on machine [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 15:02:34 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Created Created container: storage-initializer [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 15:02:34 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Started Started container storage-initializer [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 15:02:24 kserve-ci-e2e-test job-controller Normal SuccessfulCreate Created pod: e2e-pvc-model-download-6ls5n [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 15:02:43 kserve-ci-e2e-test job-controller Normal Completed Job completed [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 15:02:24 kserve-ci-e2e-test persistentvolume-controller Normal WaitForFirstConsumer waiting for first consumer to be created before binding [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 15:02:24 kserve-ci-e2e-test persistentvolume-controller Normal ExternalProvisioning Waiting for a volume to be created either by the external provisioner 'ebs.csi.aws.com' or manually by the system administrator. If volume creation is delayed, please verify that the provisioner is running and correctly registered. [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 15:02:24 kserve-ci-e2e-test ebs.csi.aws.com_aws-ebs-csi-driver-controller-68ddddd7bb-bfzsm_6bcd944a-2011-4dc8-a6f1-9995389692e0 Normal Provisioning External provisioner is provisioning volume for claim "kserve-ci-e2e-test/e2e-pvc-model-storage" [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 15:02:27 kserve-ci-e2e-test ebs.csi.aws.com_aws-ebs-csi-driver-controller-68ddddd7bb-bfzsm_6bcd944a-2011-4dc8-a6f1-9995389692e0 Normal ProvisioningSucceeded Successfully provisioned volume pvc-20b17380-3f46-45a4-a8e5-36f5cb935bca [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:45 kserve-ci-e2e-test replicaset-controller Normal SuccessfulCreate Created pod: gw-section-name-router-with-gat-2f0a622e-kserve-7c9c8cffc558295 [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 None kserve-ci-e2e-test Normal Scheduled Successfully assigned kserve-ci-e2e-test/gw-section-name-router-with-gat-2f0a622e-kserve-7c9c8cffc558295 to ip-10-0-139-47.ec2.internal [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:46 kserve-ci-e2e-test multus Normal AddedInterface Add eth0 [10.134.0.21/23] from ovn-kubernetes [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:46 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Pulling Pulling image "ghcr.io/llm-d/llm-d-inference-sim:v0.8.2" [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:47 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Pulled Successfully pulled image "ghcr.io/llm-d/llm-d-inference-sim:v0.8.2" in 1.195s (1.195s including waiting). Image size: 98346788 bytes. [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:47 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Created Created container: main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:47 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Started Started container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:58 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Killing Stopping container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:45 kserve-ci-e2e-test deployment-controller Normal ScalingReplicaSet Scaled up replica set gw-section-name-router-with-gat-2f0a622e-kserve-7c9c8cffc5 from 0 to 1 [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:45 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Secret kserve-ci-e2e-test/gw-sec0c69dceeb48768325d1a53a749e65786-kserve-self-signed-certs [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:45 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Deployment kserve-ci-e2e-test/gw-section-name-router-with-gat-2f0a622e-kserve [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:45 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Service kserve-ci-e2e-test/gw-section-name-router-with-gat-2f0a622e-kserve-workload-svc [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:45 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.HTTPRoute kserve-ci-e2e-test/gw-section-name-router-with-gat-2f0a622e-kserve-route [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:46 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.DestinationRule kserve-ci-e2e-test/gw-section-name-router-with-gat-2f0a622e-kserve-workload-svc [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:10 kserve-ci-e2e-test replicaset-controller Normal SuccessfulCreate Created pod: gw-section-name-router-with-gat-f1d92d0f-kserve-7bc8dd6c5bcr4rf [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 None kserve-ci-e2e-test Normal Scheduled Successfully assigned kserve-ci-e2e-test/gw-section-name-router-with-gat-f1d92d0f-kserve-7bc8dd6c5bcr4rf to ip-10-0-133-25.ec2.internal [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:10 kserve-ci-e2e-test multus Normal AddedInterface Add eth0 [10.132.0.46/23] from ovn-kubernetes [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:10 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Pulled Container image "ghcr.io/llm-d/llm-d-inference-sim:v0.8.2" already present on machine [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:10 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Created Created container: main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:10 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Started Started container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:27 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Killing Stopping container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:10 kserve-ci-e2e-test deployment-controller Normal ScalingReplicaSet Scaled up replica set gw-section-name-router-with-gat-f1d92d0f-kserve-7bc8dd6c5b from 0 to 1 [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:10 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Secret kserve-ci-e2e-test/gw-sec2774c263d49959f50d9eebc552e13bf9-kserve-self-signed-certs [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:10 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Deployment kserve-ci-e2e-test/gw-section-name-router-with-gat-f1d92d0f-kserve [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:10 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Service kserve-ci-e2e-test/gw-section-name-router-with-gat-f1d92d0f-kserve-workload-svc [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:10 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.HTTPRoute kserve-ci-e2e-test/gw-section-name-router-with-gat-f1d92d0f-kserve-route [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:10 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.DestinationRule kserve-ci-e2e-test/gw-section-name-router-with-gat-f1d92d0f-kserve-workload-svc [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:31:40 kserve-ci-e2e-test replicaset-controller Normal SuccessfulCreate Created pod: llmisvc-model-fb-opt-125m-route-87882a8e-kserve-5bd7864c6ctnql9 [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 None kserve-ci-e2e-test Normal Scheduled Successfully assigned kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-route-87882a8e-kserve-5bd7864c6ctnql9 to ip-10-0-139-47.ec2.internal [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:31:40 kserve-ci-e2e-test multus Normal AddedInterface Add eth0 [10.134.0.31/23] from ovn-kubernetes [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:31:40 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Pulled Container image "quay.io/opendatahub/kserve-storage-initializer@sha256:13ad765c0f323200e1db7b45c99661287e65c5158d6b54b16139829636d20be4" already present on machine [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:31:40 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Created Created container: storage-initializer [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:31:40 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Started Started container storage-initializer [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:31:45 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Pulled Container image "public.ecr.aws/q9t5s3a7/vllm-cpu-release-repo:v0.19.0" already present on machine [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:31:45 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Created Created container: main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:31:45 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Started Started container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:33:30 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Warning Unhealthy Startup probe failed: Get "https://10.134.0.31:8000/health": dial tcp 10.134.0.31:8000: connect: connection refused [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:31:40 kserve-ci-e2e-test deployment-controller Normal ScalingReplicaSet Scaled up replica set llmisvc-model-fb-opt-125m-route-87882a8e-kserve-5bd7864c6c from 0 to 1 [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:31:38 kserve-ci-e2e-test OpenDataHubModelController Warning ReconcileError Failed to reconcile LLMInferenceService: 1 error occurred: * failed to get HTTPRoute for AuthPolicy llmisvc-model-fb-opt-125m-route-87882a8e-kserve-route-authn: failed to get HTTPRoute kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-route-87882a8e-kserve-route: HTTPRoute.gateway.networking.k8s.io "llmisvc-model-fb-opt-125m-route-87882a8e-kserve-route" not found [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:31:40 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Secret kserve-ci-e2e-test/llmisva690bbc929faec8bc98c767f16c003c1-kserve-self-signed-certs [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:31:40 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Deployment kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-route-87882a8e-kserve [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:31:40 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Service kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-route-87882a8e-kserve-workload-svc [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:31:40 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.ServiceAccount kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-route-87882a8e-epp-sa [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:31:40 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.ClusterRoleBinding /kserve-ci-e2e-test21fe6730fe484f3a92b1a16afe1bac8f-llmisvc-mode [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:31:40 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Role kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-route-87882a8e-epp-role [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:31:40 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.RoleBinding kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-route-87882a8e-epp-rb [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:31:40 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Deployment kserve-ci-e2e-test/llmisvca690bbc929faec8bc98c767f16c003c1-kserve-router-scheduler [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:31:40 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Service kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-route-87882a8e-epp-service [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:31:47 kserve-ci-e2e-test LLMInferenceServiceController Normal Created (combined from similar events): Created v1.DestinationRule kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-route-87882a8e-kserve-shadow-svc [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:31:45 kserve-ci-e2e-test LLMInferenceServiceController Normal Updated Updated v1.Secret kserve-ci-e2e-test/llmisva690bbc929faec8bc98c767f16c003c1-kserve-self-signed-certs [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:31:46 kserve-ci-e2e-test LLMInferenceServiceController Normal Updated Updated v1.HTTPRoute kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-route-87882a8e-kserve-route [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:33:40 kserve-ci-e2e-test LLMInferenceServiceController Normal LLMInferenceServiceReady LLMInferenceService [llmisvc-model-fb-opt-125m-route-87882a8e] is Ready [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:31:02 kserve-ci-e2e-test replicaset-controller Normal SuccessfulCreate Created pod: llmisvc-model-fb-opt-125m-route-e95b1dc1-kserve-66874c76d46pqbc [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 None kserve-ci-e2e-test Normal Scheduled Successfully assigned kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-route-e95b1dc1-kserve-66874c76d46pqbc to ip-10-0-135-220.ec2.internal [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:31:02 kserve-ci-e2e-test multus Normal AddedInterface Add eth0 [10.133.0.37/23] from ovn-kubernetes [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:31:02 kserve-ci-e2e-test kubelet/ip-10-0-135-220.ec2.internal Normal Pulled Container image "ghcr.io/llm-d/llm-d-inference-sim:v0.8.2" already present on machine [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:31:02 kserve-ci-e2e-test kubelet/ip-10-0-135-220.ec2.internal Normal Created Created container: main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:31:02 kserve-ci-e2e-test kubelet/ip-10-0-135-220.ec2.internal Normal Started Started container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:31:15 kserve-ci-e2e-test kubelet/ip-10-0-135-220.ec2.internal Normal Killing Stopping container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:31:02 kserve-ci-e2e-test deployment-controller Normal ScalingReplicaSet Scaled up replica set llmisvc-model-fb-opt-125m-route-e95b1dc1-kserve-66874c76d4 from 0 to 1 [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:30:58 kserve-ci-e2e-test OpenDataHubModelController Warning ReconcileError Failed to reconcile LLMInferenceService: 1 error occurred: * failed to get HTTPRoute for AuthPolicy llmisvc-model-fb-opt-125m-route-e95b1dc1-kserve-route-authn: failed to get HTTPRoute kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-route-e95b1dc1-kserve-route: HTTPRoute.gateway.networking.k8s.io "llmisvc-model-fb-opt-125m-route-e95b1dc1-kserve-route" not found [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:31:02 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Secret kserve-ci-e2e-test/llmisv122f03714c5bdf915a2917fdf1262b98-kserve-self-signed-certs [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:31:02 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Deployment kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-route-e95b1dc1-kserve [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:31:02 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Service kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-route-e95b1dc1-kserve-workload-svc [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:31:02 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.HTTPRoute kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-route-e95b1dc1-kserve-route [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:31:02 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.DestinationRule kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-route-e95b1dc1-kserve-workload-svc [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:31:14 kserve-ci-e2e-test LLMInferenceServiceController Normal Updated Updated v1.Secret kserve-ci-e2e-test/llmisv122f03714c5bdf915a2917fdf1262b98-kserve-self-signed-certs [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:31:14 kserve-ci-e2e-test LLMInferenceServiceController Normal LLMInferenceServiceReady LLMInferenceService [llmisvc-model-fb-opt-125m-route-e95b1dc1] is Ready [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 None kserve-ci-e2e-test Normal Scheduled Successfully assigned kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-with-7ca60146-kserve-7cfdf49d74-pd9px to ip-10-0-139-47.ec2.internal [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:29:25 kserve-ci-e2e-test multus Normal AddedInterface Add eth0 [10.134.0.30/23] from ovn-kubernetes [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:29:25 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Pulled Container image "quay.io/opendatahub/kserve-storage-initializer@sha256:13ad765c0f323200e1db7b45c99661287e65c5158d6b54b16139829636d20be4" already present on machine [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:29:25 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Created Created container: storage-initializer [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:29:25 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Started Started container storage-initializer [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:29:30 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Pulled Container image "public.ecr.aws/q9t5s3a7/vllm-cpu-release-repo:v0.19.0" already present on machine [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:29:30 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Created Created container: main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:29:30 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Started Started container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:30:34 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Warning Unhealthy Startup probe failed: Get "https://10.134.0.30:8000/health": dial tcp 10.134.0.30:8000: connect: connection refused [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:29:24 kserve-ci-e2e-test replicaset-controller Normal SuccessfulCreate Created pod: llmisvc-model-fb-opt-125m-with-7ca60146-kserve-7cfdf49d74-pd9px [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:29:24 kserve-ci-e2e-test replicaset-controller Normal SuccessfulCreate Created pod: llmisvc-model-fb-opt-125m-with-7ca60146-kserve-router-schenp5mk [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:29:24 kserve-ci-e2e-test deployment-controller Normal ScalingReplicaSet Scaled up replica set llmisvc-model-fb-opt-125m-with-7ca60146-kserve-router-scheduler-778547fbc from 0 to 1 [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 None kserve-ci-e2e-test Normal Scheduled Successfully assigned kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-with-7ca60146-kserve-router-schenp5mk to ip-10-0-133-25.ec2.internal [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:29:25 kserve-ci-e2e-test multus Normal AddedInterface Add eth0 [10.132.0.52/23] from ovn-kubernetes [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:29:25 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Pulled Container image "quay.io/opendatahub/kserve-storage-initializer@sha256:13ad765c0f323200e1db7b45c99661287e65c5158d6b54b16139829636d20be4" already present on machine [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:29:25 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Created Created container: storage-initializer [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:29:25 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Started Started container storage-initializer [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:29:26 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Pulled Container image "ghcr.io/llm-d/llm-d-router-endpoint-picker:v0.9.0-rc.2" already present on machine [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:29:27 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Created Created container: main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:29:27 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Started Started container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:29:24 kserve-ci-e2e-test deployment-controller Normal ScalingReplicaSet Scaled up replica set llmisvc-model-fb-opt-125m-with-7ca60146-kserve-7cfdf49d74 from 0 to 1 [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:29:23 kserve-ci-e2e-test OpenDataHubModelController Warning ReconcileError Failed to reconcile LLMInferenceService: 1 error occurred: * failed to get HTTPRoute for AuthPolicy llmisvc-model-fb-opt-125m-with-7ca60146-kserve-route-authn: failed to get HTTPRoute kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-with-7ca60146-kserve-route: HTTPRoute.gateway.networking.k8s.io "llmisvc-model-fb-opt-125m-with-7ca60146-kserve-route" not found [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:29:24 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Secret kserve-ci-e2e-test/llmisv3e414c2ba058a022dfd694dbcbac5b51-kserve-self-signed-certs [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:29:24 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Deployment kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-with-7ca60146-kserve [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:29:24 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Service kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-with-7ca60146-kserve-workload-svc [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:29:24 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.ServiceAccount kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-with-7ca60146-epp-sa [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:29:24 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.ClusterRoleBinding /kserve-ci-e2e-test5216bfd716f919dc046bc693ceb22e41-llmisvc-mode [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:29:24 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Role kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-with-7ca60146-epp-role [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:29:24 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.RoleBinding kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-with-7ca60146-epp-rb [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:29:24 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Deployment kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-with-7ca60146-kserve-router-scheduler [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:29:24 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Service kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-with-7ca60146-epp-service [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:29:35 kserve-ci-e2e-test LLMInferenceServiceController Normal Created (combined from similar events): Created v1.DestinationRule kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-with-7ca60146-kserve-shadow-svc [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:29:33 kserve-ci-e2e-test LLMInferenceServiceController Normal Updated Updated v1.Secret kserve-ci-e2e-test/llmisv3e414c2ba058a022dfd694dbcbac5b51-kserve-self-signed-certs [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:29:34 kserve-ci-e2e-test LLMInferenceServiceController Normal Updated Updated v1.HTTPRoute kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-with-7ca60146-kserve-route [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:30:45 kserve-ci-e2e-test LLMInferenceServiceController Normal LLMInferenceServiceReady LLMInferenceService [llmisvc-model-fb-opt-125m-with-7ca60146] is Ready [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 None kserve-ci-e2e-test Normal Scheduled Successfully assigned kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-with-ba4d693a-kserve-5b998d5789-t6wf4 to ip-10-0-139-47.ec2.internal [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:45:56 kserve-ci-e2e-test multus Normal AddedInterface Add eth0 [10.134.0.32/23] from ovn-kubernetes [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:45:56 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Pulled Container image "quay.io/opendatahub/kserve-storage-initializer@sha256:13ad765c0f323200e1db7b45c99661287e65c5158d6b54b16139829636d20be4" already present on machine [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:45:56 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Created Created container: storage-initializer [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:45:56 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Started Started container storage-initializer [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:46:01 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Pulled Container image "public.ecr.aws/q9t5s3a7/vllm-cpu-release-repo:v0.19.0" already present on machine [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:46:01 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Created Created container: main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:46:01 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Started Started container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:47:06 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Warning Unhealthy Startup probe failed: Get "https://10.134.0.32:8000/health": dial tcp 10.134.0.32:8000: connect: connection refused [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:45:56 kserve-ci-e2e-test replicaset-controller Normal SuccessfulCreate Created pod: llmisvc-model-fb-opt-125m-with-ba4d693a-kserve-5b998d5789-t6wf4 [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 None kserve-ci-e2e-test Normal Scheduled Successfully assigned kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-with-ba4d693a-kserve-router-sche4ghr9 to ip-10-0-133-25.ec2.internal [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:45:56 kserve-ci-e2e-test multus Normal AddedInterface Add eth0 [10.132.0.54/23] from ovn-kubernetes [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:45:56 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Pulled Container image "quay.io/opendatahub/kserve-storage-initializer@sha256:13ad765c0f323200e1db7b45c99661287e65c5158d6b54b16139829636d20be4" already present on machine [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:45:57 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Created Created container: storage-initializer [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:45:57 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Started Started container storage-initializer [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:45:57 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Pulled Container image "ghcr.io/llm-d/llm-d-router-endpoint-picker:v0.9.0-rc.2" already present on machine [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:45:57 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Created Created container: main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:45:57 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Started Started container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:45:56 kserve-ci-e2e-test replicaset-controller Normal SuccessfulCreate Created pod: llmisvc-model-fb-opt-125m-with-ba4d693a-kserve-router-sche4ghr9 [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:45:56 kserve-ci-e2e-test deployment-controller Normal ScalingReplicaSet Scaled up replica set llmisvc-model-fb-opt-125m-with-ba4d693a-kserve-router-scheduler-79f5455bb from 0 to 1 [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:45:56 kserve-ci-e2e-test deployment-controller Normal ScalingReplicaSet Scaled up replica set llmisvc-model-fb-opt-125m-with-ba4d693a-kserve-5b998d5789 from 0 to 1 [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:45:56 kserve-ci-e2e-test OpenDataHubModelController Warning ReconcileError Failed to reconcile LLMInferenceService: 1 error occurred: * failed to get HTTPRoute for AuthPolicy llmisvc-model-fb-opt-125m-with-ba4d693a-kserve-route-authn: failed to get HTTPRoute kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-with-ba4d693a-kserve-route: HTTPRoute.gateway.networking.k8s.io "llmisvc-model-fb-opt-125m-with-ba4d693a-kserve-route" not found [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:45:56 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Secret kserve-ci-e2e-test/llmisv77ff2528d3e9b4972cd9335229fce9f0-kserve-self-signed-certs [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:45:56 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Deployment kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-with-ba4d693a-kserve [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:45:56 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Service kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-with-ba4d693a-kserve-workload-svc [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:45:56 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.ServiceAccount kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-with-ba4d693a-epp-sa [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:45:56 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.ClusterRoleBinding /kserve-ci-e2e-test05addb65ba05195619f26ef266e8fc04-llmisvc-mode [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:45:56 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Role kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-with-ba4d693a-epp-role [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:45:56 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.RoleBinding kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-with-ba4d693a-epp-rb [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:45:56 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Deployment kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-with-ba4d693a-kserve-router-scheduler [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:45:56 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Service kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-with-ba4d693a-epp-service [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:46:17 kserve-ci-e2e-test LLMInferenceServiceController Normal Created (combined from similar events): Created v1.DestinationRule kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-with-ba4d693a-kserve-shadow-svc [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:46:16 kserve-ci-e2e-test LLMInferenceServiceController Normal Updated Updated v1.Secret kserve-ci-e2e-test/llmisv77ff2528d3e9b4972cd9335229fce9f0-kserve-self-signed-certs [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:46:16 kserve-ci-e2e-test LLMInferenceServiceController Normal Updated Updated v1.HTTPRoute kserve-ci-e2e-test/llmisvc-model-fb-opt-125m-with-ba4d693a-kserve-route [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:47:17 kserve-ci-e2e-test LLMInferenceServiceController Normal LLMInferenceServiceReady LLMInferenceService [llmisvc-model-fb-opt-125m-with-ba4d693a] is Ready [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 15:03:08 kserve-ci-e2e-test replicaset-controller Normal SuccessfulCreate Created pod: llmisvc-model-pvc-router-manage-59b9d263-kserve-585587bc9dnwscv [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 None kserve-ci-e2e-test Normal Scheduled Successfully assigned kserve-ci-e2e-test/llmisvc-model-pvc-router-manage-59b9d263-kserve-585587bc9dnwscv to ip-10-0-139-47.ec2.internal [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 15:03:10 kserve-ci-e2e-test attachdetach-controller Normal SuccessfulAttachVolume AttachVolume.Attach succeeded for volume "pvc-20b17380-3f46-45a4-a8e5-36f5cb935bca" [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 15:03:15 kserve-ci-e2e-test multus Normal AddedInterface Add eth0 [10.134.0.34/23] from ovn-kubernetes [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 15:03:15 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Pulled Container image "public.ecr.aws/q9t5s3a7/vllm-cpu-release-repo:v0.19.0" already present on machine [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 15:03:15 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Created Created container: main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 15:03:15 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Started Started container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 15:05:05 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Warning Unhealthy Startup probe failed: Get "https://10.134.0.34:8000/health": dial tcp 10.134.0.34:8000: connect: connection refused [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 15:03:08 kserve-ci-e2e-test deployment-controller Normal ScalingReplicaSet Scaled up replica set llmisvc-model-pvc-router-manage-59b9d263-kserve-585587bc9d from 0 to 1 [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 15:03:05 kserve-ci-e2e-test OpenDataHubModelController Warning ReconcileError Failed to reconcile LLMInferenceService: 1 error occurred: * failed to get HTTPRoute for AuthPolicy llmisvc-model-pvc-router-manage-59b9d263-kserve-route-authn: failed to get HTTPRoute kserve-ci-e2e-test/llmisvc-model-pvc-router-manage-59b9d263-kserve-route: HTTPRoute.gateway.networking.k8s.io "llmisvc-model-pvc-router-manage-59b9d263-kserve-route" not found [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 15:03:08 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Secret kserve-ci-e2e-test/llmisv8bf079eb6eda4debfb4ef5bb7817824c-kserve-self-signed-certs [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 15:03:08 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Deployment kserve-ci-e2e-test/llmisvc-model-pvc-router-manage-59b9d263-kserve [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 15:03:08 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Service kserve-ci-e2e-test/llmisvc-model-pvc-router-manage-59b9d263-kserve-workload-svc [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 15:03:08 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.ServiceAccount kserve-ci-e2e-test/llmisvc-model-pvc-router-manage-59b9d263-epp-sa [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 15:03:08 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.ClusterRoleBinding /kserve-ci-e2e-testc9569cf4801efc0ed27b2f25ffaee875-llmisvc-mode [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 15:03:08 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Role kserve-ci-e2e-test/llmisvc-model-pvc-router-manage-59b9d263-epp-role [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 15:03:08 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.RoleBinding kserve-ci-e2e-test/llmisvc-model-pvc-router-manage-59b9d263-epp-rb [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 15:03:08 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Deployment kserve-ci-e2e-test/llmisvc8bf079eb6eda4debfb4ef5bb7817824c-kserve-router-scheduler [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 15:03:08 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Service kserve-ci-e2e-test/llmisvc-model-pvc-router-manage-59b9d263-epp-service [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 15:03:26 kserve-ci-e2e-test LLMInferenceServiceController Normal Created (combined from similar events): Created v1.DestinationRule kserve-ci-e2e-test/llmisvc-model-pvc-router-manage-59b9d263-kserve-shadow-svc [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 15:03:25 kserve-ci-e2e-test LLMInferenceServiceController Normal Updated Updated v1.Secret kserve-ci-e2e-test/llmisv8bf079eb6eda4debfb4ef5bb7817824c-kserve-self-signed-certs [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 15:03:25 kserve-ci-e2e-test LLMInferenceServiceController Normal Updated Updated v1.HTTPRoute kserve-ci-e2e-test/llmisvc-model-pvc-router-manage-59b9d263-kserve-route [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 15:05:15 kserve-ci-e2e-test LLMInferenceServiceController Normal LLMInferenceServiceReady LLMInferenceService [llmisvc-model-pvc-router-manage-59b9d263] is Ready [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:54 kserve-ci-e2e-test replicaset-controller Normal SuccessfulCreate Created pod: llmisvc-model-qwen2-5-0-5b-rout-4f8c0978-kserve-5c54ddb98flf892 [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 None kserve-ci-e2e-test Normal Scheduled Successfully assigned kserve-ci-e2e-test/llmisvc-model-qwen2-5-0-5b-rout-4f8c0978-kserve-5c54ddb98flf892 to ip-10-0-135-220.ec2.internal [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:54 kserve-ci-e2e-test multus Normal AddedInterface Add eth0 [10.133.0.36/23] from ovn-kubernetes [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:54 kserve-ci-e2e-test kubelet/ip-10-0-135-220.ec2.internal Normal Pulled Container image "ghcr.io/llm-d/llm-d-inference-sim:v0.8.2" already present on machine [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:54 kserve-ci-e2e-test kubelet/ip-10-0-135-220.ec2.internal Normal Created Created container: main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:54 kserve-ci-e2e-test kubelet/ip-10-0-135-220.ec2.internal Normal Started Started container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:28:40 kserve-ci-e2e-test kubelet/ip-10-0-135-220.ec2.internal Normal Killing Stopping container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:28:56 kserve-ci-e2e-test kubelet/ip-10-0-135-220.ec2.internal Warning Unhealthy Readiness probe failed: Get "https://10.133.0.36:8000/health": context deadline exceeded [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:54 kserve-ci-e2e-test deployment-controller Normal ScalingReplicaSet Scaled up replica set llmisvc-model-qwen2-5-0-5b-rout-4f8c0978-kserve-5c54ddb98f from 0 to 1 [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:53 kserve-ci-e2e-test OpenDataHubModelController Warning ReconcileError Failed to reconcile LLMInferenceService: 1 error occurred: * failed to get HTTPRoute for AuthPolicy llmisvc-model-qwen2-5-0-5b-rout-4f8c0978-kserve-route-authn: failed to get HTTPRoute kserve-ci-e2e-test/llmisvc-model-qwen2-5-0-5b-rout-4f8c0978-kserve-route: HTTPRoute.gateway.networking.k8s.io "llmisvc-model-qwen2-5-0-5b-rout-4f8c0978-kserve-route" not found [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:54 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Secret kserve-ci-e2e-test/llmisv08544b88a8d963ffd553cc1f3ed82d16-kserve-self-signed-certs [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:54 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Deployment kserve-ci-e2e-test/llmisvc-model-qwen2-5-0-5b-rout-4f8c0978-kserve [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:54 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Service kserve-ci-e2e-test/llmisvc-model-qwen2-5-0-5b-rout-4f8c0978-kserve-workload-svc [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:54 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.ServiceAccount kserve-ci-e2e-test/llmisvc-model-qwen2-5-0-5b-rout-4f8c0978-epp-sa [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:54 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.ClusterRoleBinding /kserve-ci-e2e-test76d7f532acb694e4a7bcef75d32cd8a1-llmisvc-mode [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:54 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Role kserve-ci-e2e-test/llmisvc-model-qwen2-5-0-5b-rout-4f8c0978-epp-role [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:54 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.RoleBinding kserve-ci-e2e-test/llmisvc-model-qwen2-5-0-5b-rout-4f8c0978-epp-rb [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:54 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Deployment kserve-ci-e2e-test/llmisvc08544b88a8d963ffd553cc1f3ed82d16-kserve-router-scheduler [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:54 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Service kserve-ci-e2e-test/llmisvc-model-qwen2-5-0-5b-rout-4f8c0978-epp-service [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:28:06 kserve-ci-e2e-test LLMInferenceServiceController Normal Created (combined from similar events): Created v1.DestinationRule kserve-ci-e2e-test/llmisvc-model-qwen2-5-0-5b-rout-4f8c0978-kserve-shadow-svc [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:28:05 kserve-ci-e2e-test LLMInferenceServiceController Normal Updated Updated v1.Secret kserve-ci-e2e-test/llmisv08544b88a8d963ffd553cc1f3ed82d16-kserve-self-signed-certs [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:28:06 kserve-ci-e2e-test LLMInferenceServiceController Normal Updated Updated v1.HTTPRoute kserve-ci-e2e-test/llmisvc-model-qwen2-5-0-5b-rout-4f8c0978-kserve-route [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:28:31 kserve-ci-e2e-test LLMInferenceServiceController Normal LLMInferenceServiceReady LLMInferenceService [llmisvc-model-qwen2-5-0-5b-rout-4f8c0978] is Ready [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:28:40 kserve-ci-e2e-test LLMInferenceServiceController Normal Deleted Deleted v1.ClusterRoleBinding /kserve-ci-e2e-test76d7f532acb694e4a7bcef75d32cd8a1-llmisvc-mode [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:23 kserve-ci-e2e-test replicaset-controller Normal SuccessfulCreate Created pod: llmisvc-model-qwen2-5-0-5b-rout-a50492e9-kserve-5d8ffd58ddqnd8m [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 None kserve-ci-e2e-test Normal Scheduled Successfully assigned kserve-ci-e2e-test/llmisvc-model-qwen2-5-0-5b-rout-a50492e9-kserve-5d8ffd58ddqnd8m to ip-10-0-135-220.ec2.internal [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:23 kserve-ci-e2e-test multus Normal AddedInterface Add eth0 [10.133.0.34/23] from ovn-kubernetes [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:23 kserve-ci-e2e-test kubelet/ip-10-0-135-220.ec2.internal Normal Pulled Container image "ghcr.io/llm-d/llm-d-inference-sim:v0.8.2" already present on machine [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:23 kserve-ci-e2e-test kubelet/ip-10-0-135-220.ec2.internal Normal Created Created container: main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:23 kserve-ci-e2e-test kubelet/ip-10-0-135-220.ec2.internal Normal Started Started container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:26:13 kserve-ci-e2e-test kubelet/ip-10-0-135-220.ec2.internal Normal Killing Stopping container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:23 kserve-ci-e2e-test deployment-controller Normal ScalingReplicaSet Scaled up replica set llmisvc-model-qwen2-5-0-5b-rout-a50492e9-kserve-5d8ffd58dd from 0 to 1 [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:21 kserve-ci-e2e-test OpenDataHubModelController Warning ReconcileError Failed to reconcile LLMInferenceService: 1 error occurred: * failed to get HTTPRoute for AuthPolicy llmisvc-model-qwen2-5-0-5b-rout-a50492e9-kserve-route-authn: failed to get HTTPRoute kserve-ci-e2e-test/llmisvc-model-qwen2-5-0-5b-rout-a50492e9-kserve-route: HTTPRoute.gateway.networking.k8s.io "llmisvc-model-qwen2-5-0-5b-rout-a50492e9-kserve-route" not found [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:23 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Secret kserve-ci-e2e-test/llmisvb11a9c9832b99b016bc8f8e0ea095712-kserve-self-signed-certs [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:23 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Deployment kserve-ci-e2e-test/llmisvc-model-qwen2-5-0-5b-rout-a50492e9-kserve [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:23 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Service kserve-ci-e2e-test/llmisvc-model-qwen2-5-0-5b-rout-a50492e9-kserve-workload-svc [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:23 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.ServiceAccount kserve-ci-e2e-test/llmisvc-model-qwen2-5-0-5b-rout-a50492e9-epp-sa [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:23 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.ClusterRoleBinding /kserve-ci-e2e-testb7025ca4d8a6f8f5b2fd08b5581d2678-llmisvc-mode [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:23 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Role kserve-ci-e2e-test/llmisvc-model-qwen2-5-0-5b-rout-a50492e9-epp-role [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:23 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.RoleBinding kserve-ci-e2e-test/llmisvc-model-qwen2-5-0-5b-rout-a50492e9-epp-rb [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:23 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Deployment kserve-ci-e2e-test/llmisvcb11a9c9832b99b016bc8f8e0ea095712-kserve-router-scheduler [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:23 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Service kserve-ci-e2e-test/llmisvc-model-qwen2-5-0-5b-rout-a50492e9-epp-service [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:34 kserve-ci-e2e-test LLMInferenceServiceController Normal Created (combined from similar events): Created v1.DestinationRule kserve-ci-e2e-test/llmisvc-model-qwen2-5-0-5b-rout-a50492e9-kserve-shadow-svc [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:33 kserve-ci-e2e-test LLMInferenceServiceController Normal Updated Updated v1.Secret kserve-ci-e2e-test/llmisvb11a9c9832b99b016bc8f8e0ea095712-kserve-self-signed-certs [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:33 kserve-ci-e2e-test LLMInferenceServiceController Normal Updated Updated v1.HTTPRoute kserve-ci-e2e-test/llmisvc-model-qwen2-5-0-5b-rout-a50492e9-kserve-route [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:26:12 kserve-ci-e2e-test LLMInferenceServiceController Normal LLMInferenceServiceReady LLMInferenceService [llmisvc-model-qwen2-5-0-5b-rout-a50492e9] is Ready [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:26:13 kserve-ci-e2e-test LLMInferenceServiceController Normal Deleted Deleted v1.ClusterRoleBinding /kserve-ci-e2e-testb7025ca4d8a6f8f5b2fd08b5581d2678-llmisvc-mode [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 None kserve-ci-e2e-test Normal Scheduled Successfully assigned kserve-ci-e2e-test/llmisvc-router-managed-test-llm-4b931143-kserve-bd545d56-jcs4n to ip-10-0-135-220.ec2.internal [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:18 kserve-ci-e2e-test multus Normal AddedInterface Add eth0 [10.133.0.35/23] from ovn-kubernetes [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:18 kserve-ci-e2e-test kubelet/ip-10-0-135-220.ec2.internal Normal Pulled Container image "ghcr.io/llm-d/llm-d-inference-sim:v0.8.2" already present on machine [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:18 kserve-ci-e2e-test kubelet/ip-10-0-135-220.ec2.internal Normal Created Created container: main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:18 kserve-ci-e2e-test kubelet/ip-10-0-135-220.ec2.internal Normal Started Started container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:28:58 kserve-ci-e2e-test kubelet/ip-10-0-135-220.ec2.internal Normal Killing Stopping container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:17 kserve-ci-e2e-test replicaset-controller Normal SuccessfulCreate Created pod: llmisvc-router-managed-test-llm-4b931143-kserve-bd545d56-jcs4n [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:17 kserve-ci-e2e-test deployment-controller Normal ScalingReplicaSet Scaled up replica set llmisvc-router-managed-test-llm-4b931143-kserve-bd545d56 from 0 to 1 [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:14 kserve-ci-e2e-test OpenDataHubModelController Warning ReconcileError Failed to reconcile LLMInferenceService: 1 error occurred: * failed to get HTTPRoute for AuthPolicy llmisvc-router-managed-test-llm-4b931143-kserve-route-authn: failed to get HTTPRoute kserve-ci-e2e-test/llmisvc-router-managed-test-llm-4b931143-kserve-route: HTTPRoute.gateway.networking.k8s.io "llmisvc-router-managed-test-llm-4b931143-kserve-route" not found [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:17 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Secret kserve-ci-e2e-test/llmisvca2d2d7d499abb359505529ebe02c136-kserve-self-signed-certs [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:17 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Deployment kserve-ci-e2e-test/llmisvc-router-managed-test-llm-4b931143-kserve [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:17 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Service kserve-ci-e2e-test/llmisvc-router-managed-test-llm-4b931143-kserve-workload-svc [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:17 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.ServiceAccount kserve-ci-e2e-test/llmisvc-router-managed-test-llm-4b931143-epp-sa [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:17 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.ClusterRoleBinding /kserve-ci-e2e-test8ac8e3d2264ccb939eb021b0b835847c-llmisvc-rout [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:17 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Role kserve-ci-e2e-test/llmisvc-router-managed-test-llm-4b931143-epp-role [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:17 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.RoleBinding kserve-ci-e2e-test/llmisvc-router-managed-test-llm-4b931143-epp-rb [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:17 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Deployment kserve-ci-e2e-test/llmisvcca2d2d7d499abb359505529ebe02c136-kserve-router-scheduler [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:17 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Service kserve-ci-e2e-test/llmisvc-router-managed-test-llm-4b931143-epp-service [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:43 kserve-ci-e2e-test LLMInferenceServiceController Normal Created (combined from similar events): Created v1.DestinationRule kserve-ci-e2e-test/llmisvc-router-managed-test-llm-4b931143-kserve-shadow-svc [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:42 kserve-ci-e2e-test LLMInferenceServiceController Normal Updated Updated v1.Secret kserve-ci-e2e-test/llmisvca2d2d7d499abb359505529ebe02c136-kserve-self-signed-certs [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:43 kserve-ci-e2e-test LLMInferenceServiceController Normal Updated Updated v1.HTTPRoute kserve-ci-e2e-test/llmisvc-router-managed-test-llm-4b931143-kserve-route [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:50 kserve-ci-e2e-test LLMInferenceServiceController Normal LLMInferenceServiceReady LLMInferenceService [llmisvc-router-managed-test-llm-4b931143] is Ready [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:28:58 kserve-ci-e2e-test LLMInferenceServiceController Normal Deleted Deleted v1.ClusterRoleBinding /kserve-ci-e2e-test8ac8e3d2264ccb939eb021b0b835847c-llmisvc-rout [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:24:42 kserve-ci-e2e-test replicaset-controller Normal SuccessfulCreate Created pod: llmisvc-router-managed-test-llm-5b1e8f15-kserve-64df7bddb4xtf2m [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 None kserve-ci-e2e-test Normal Scheduled Successfully assigned kserve-ci-e2e-test/llmisvc-router-managed-test-llm-5b1e8f15-kserve-64df7bddb4xtf2m to ip-10-0-135-220.ec2.internal [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:24:43 kserve-ci-e2e-test multus Normal AddedInterface Add eth0 [10.133.0.33/23] from ovn-kubernetes [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:24:43 kserve-ci-e2e-test kubelet/ip-10-0-135-220.ec2.internal Normal Pulled Container image "ghcr.io/llm-d/llm-d-inference-sim:v0.8.2" already present on machine [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:24:43 kserve-ci-e2e-test kubelet/ip-10-0-135-220.ec2.internal Normal Created Created container: main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:24:43 kserve-ci-e2e-test kubelet/ip-10-0-135-220.ec2.internal Normal Started Started container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:26:33 kserve-ci-e2e-test kubelet/ip-10-0-135-220.ec2.internal Normal Killing Stopping container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:24:42 kserve-ci-e2e-test deployment-controller Normal ScalingReplicaSet Scaled up replica set llmisvc-router-managed-test-llm-5b1e8f15-kserve-64df7bddb4 from 0 to 1 [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:24:40 kserve-ci-e2e-test OpenDataHubModelController Warning ReconcileError Failed to reconcile LLMInferenceService: 1 error occurred: * failed to get HTTPRoute for AuthPolicy llmisvc-router-managed-test-llm-5b1e8f15-kserve-route-authn: failed to get HTTPRoute kserve-ci-e2e-test/llmisvc-router-managed-test-llm-5b1e8f15-kserve-route: HTTPRoute.gateway.networking.k8s.io "llmisvc-router-managed-test-llm-5b1e8f15-kserve-route" not found [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:24:42 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Secret kserve-ci-e2e-test/llmisve55ae740357a3a31a27cdb8b66ffe20f-kserve-self-signed-certs [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:24:42 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Deployment kserve-ci-e2e-test/llmisvc-router-managed-test-llm-5b1e8f15-kserve [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:24:42 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Service kserve-ci-e2e-test/llmisvc-router-managed-test-llm-5b1e8f15-kserve-workload-svc [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:24:42 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.ServiceAccount kserve-ci-e2e-test/llmisvc-router-managed-test-llm-5b1e8f15-epp-sa [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:24:42 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.ClusterRoleBinding /kserve-ci-e2e-test7f54e84970003a6e7372bdbcb574f7ed-llmisvc-rout [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:24:42 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Role kserve-ci-e2e-test/llmisvc-router-managed-test-llm-5b1e8f15-epp-role [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:24:43 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.RoleBinding kserve-ci-e2e-test/llmisvc-router-managed-test-llm-5b1e8f15-epp-rb [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:24:43 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Deployment kserve-ci-e2e-test/llmisvce55ae740357a3a31a27cdb8b66ffe20f-kserve-router-scheduler [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:24:43 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Service kserve-ci-e2e-test/llmisvc-router-managed-test-llm-5b1e8f15-epp-service [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:03 kserve-ci-e2e-test LLMInferenceServiceController Normal Created (combined from similar events): Created v1.DestinationRule kserve-ci-e2e-test/llmisvc-router-managed-test-llm-5b1e8f15-kserve-shadow-svc [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:03 kserve-ci-e2e-test LLMInferenceServiceController Normal Updated Updated v1.Secret kserve-ci-e2e-test/llmisve55ae740357a3a31a27cdb8b66ffe20f-kserve-self-signed-certs [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:03 kserve-ci-e2e-test LLMInferenceServiceController Normal Updated Updated v1.HTTPRoute kserve-ci-e2e-test/llmisvc-router-managed-test-llm-5b1e8f15-kserve-route [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:15 kserve-ci-e2e-test LLMInferenceServiceController Normal LLMInferenceServiceReady LLMInferenceService [llmisvc-router-managed-test-llm-5b1e8f15] is Ready [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:26:33 kserve-ci-e2e-test LLMInferenceServiceController Normal Deleted Deleted v1.ClusterRoleBinding /kserve-ci-e2e-test7f54e84970003a6e7372bdbcb574f7ed-llmisvc-rout [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:23:41 kserve-ci-e2e-test replicaset-controller Normal SuccessfulCreate Created pod: llmisvc-router-managed-test-llm-e45d1f79-kserve-7fdbbd4879ffz96 [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 None kserve-ci-e2e-test Normal Scheduled Successfully assigned kserve-ci-e2e-test/llmisvc-router-managed-test-llm-e45d1f79-kserve-7fdbbd4879ffz96 to ip-10-0-135-220.ec2.internal [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:23:41 kserve-ci-e2e-test multus Normal AddedInterface Add eth0 [10.133.0.32/23] from ovn-kubernetes [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:23:41 kserve-ci-e2e-test kubelet/ip-10-0-135-220.ec2.internal Normal Pulling Pulling image "ghcr.io/llm-d/llm-d-inference-sim:v0.8.2" [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:23:43 kserve-ci-e2e-test kubelet/ip-10-0-135-220.ec2.internal Normal Pulled Successfully pulled image "ghcr.io/llm-d/llm-d-inference-sim:v0.8.2" in 1.407s (1.407s including waiting). Image size: 98346788 bytes. [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:23:43 kserve-ci-e2e-test kubelet/ip-10-0-135-220.ec2.internal Normal Created Created container: main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:23:43 kserve-ci-e2e-test kubelet/ip-10-0-135-220.ec2.internal Normal Started Started container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:24:15 kserve-ci-e2e-test kubelet/ip-10-0-135-220.ec2.internal Normal Killing Stopping container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:23:41 kserve-ci-e2e-test deployment-controller Normal ScalingReplicaSet Scaled up replica set llmisvc-router-managed-test-llm-e45d1f79-kserve-7fdbbd4879 from 0 to 1 [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:23:36 kserve-ci-e2e-test OpenDataHubModelController Warning ReconcileError Failed to reconcile LLMInferenceService: 1 error occurred: * failed to get HTTPRoute for AuthPolicy llmisvc-router-managed-test-llm-e45d1f79-kserve-route-authn: failed to get HTTPRoute kserve-ci-e2e-test/llmisvc-router-managed-test-llm-e45d1f79-kserve-route: HTTPRoute.gateway.networking.k8s.io "llmisvc-router-managed-test-llm-e45d1f79-kserve-route" not found [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:23:41 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Secret kserve-ci-e2e-test/llmisv5c7e67b6c51568d1d6d13829a9337f2a-kserve-self-signed-certs [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:23:41 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Deployment kserve-ci-e2e-test/llmisvc-router-managed-test-llm-e45d1f79-kserve [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:23:41 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Service kserve-ci-e2e-test/llmisvc-router-managed-test-llm-e45d1f79-kserve-workload-svc [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:23:41 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.ServiceAccount kserve-ci-e2e-test/llmisvc-router-managed-test-llm-e45d1f79-epp-sa [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:23:41 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.ClusterRoleBinding /kserve-ci-e2e-testef4d2875be14b30dc1561ed84d0d4bde-llmisvc-rout [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:23:41 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Role kserve-ci-e2e-test/llmisvc-router-managed-test-llm-e45d1f79-epp-role [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:23:41 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.RoleBinding kserve-ci-e2e-test/llmisvc-router-managed-test-llm-e45d1f79-epp-rb [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:23:41 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Deployment kserve-ci-e2e-test/llmisvc5c7e67b6c51568d1d6d13829a9337f2a-kserve-router-scheduler [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:23:41 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Service kserve-ci-e2e-test/llmisvc-router-managed-test-llm-e45d1f79-epp-service [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:23:53 kserve-ci-e2e-test LLMInferenceServiceController Normal Created (combined from similar events): Created v1.DestinationRule kserve-ci-e2e-test/llmisvc-router-managed-test-llm-e45d1f79-kserve-shadow-svc [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:23:52 kserve-ci-e2e-test LLMInferenceServiceController Normal Updated Updated v1.Secret kserve-ci-e2e-test/llmisv5c7e67b6c51568d1d6d13829a9337f2a-kserve-self-signed-certs [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:23:52 kserve-ci-e2e-test LLMInferenceServiceController Normal Updated Updated v1.HTTPRoute kserve-ci-e2e-test/llmisvc-router-managed-test-llm-e45d1f79-kserve-route [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:24:14 kserve-ci-e2e-test LLMInferenceServiceController Normal LLMInferenceServiceReady LLMInferenceService [llmisvc-router-managed-test-llm-e45d1f79] is Ready [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:24:15 kserve-ci-e2e-test LLMInferenceServiceController Normal Deleted Deleted v1.ClusterRoleBinding /kserve-ci-e2e-testef4d2875be14b30dc1561ed84d0d4bde-llmisvc-rout [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:54 kserve-ci-e2e-test replicaset-controller Normal SuccessfulCreate Created pod: llmisvc08544b88a8d963ffd553cc1f3ed82d16-kserve-router-schefvtgh [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:54 kserve-ci-e2e-test deployment-controller Normal ScalingReplicaSet Scaled up replica set llmisvc08544b88a8d963ffd553cc1f3ed82d16-kserve-router-scheduler-6fcb489785 from 0 to 1 [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 None kserve-ci-e2e-test Normal Scheduled Successfully assigned kserve-ci-e2e-test/llmisvc08544b88a8d963ffd553cc1f3ed82d16-kserve-router-schefvtgh to ip-10-0-133-25.ec2.internal [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:55 kserve-ci-e2e-test multus Normal AddedInterface Add eth0 [10.132.0.50/23] from ovn-kubernetes [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:55 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Pulled Container image "ghcr.io/llm-d/llm-d-router-endpoint-picker:v0.9.0-rc.2" already present on machine [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:55 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Created Created container: main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:55 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Started Started container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:28:40 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Killing Stopping container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:28:40 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Warning FailedPreStopHook PreStopHook failed [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:23:41 kserve-ci-e2e-test replicaset-controller Normal SuccessfulCreate Created pod: llmisvc5c7e67b6c51568d1d6d13829a9337f2a-kserve-router-schexv898 [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:23:41 kserve-ci-e2e-test deployment-controller Normal ScalingReplicaSet Scaled up replica set llmisvc5c7e67b6c51568d1d6d13829a9337f2a-kserve-router-scheduler-5dd88bfbb7 from 0 to 1 [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 None kserve-ci-e2e-test Normal Scheduled Successfully assigned kserve-ci-e2e-test/llmisvc5c7e67b6c51568d1d6d13829a9337f2a-kserve-router-schexv898 to ip-10-0-139-47.ec2.internal [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:23:42 kserve-ci-e2e-test multus Normal AddedInterface Add eth0 [10.134.0.25/23] from ovn-kubernetes [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:23:42 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Pulled Container image "ghcr.io/llm-d/llm-d-router-endpoint-picker:v0.9.0-rc.2" already present on machine [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:23:42 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Created Created container: main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:23:42 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Started Started container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:24:15 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Killing Stopping container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:24:15 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Warning FailedPreStopHook PreStopHook failed [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 15:03:08 kserve-ci-e2e-test replicaset-controller Normal SuccessfulCreate Created pod: llmisvc8bf079eb6eda4debfb4ef5bb7817824c-kserve-router-scher4hs2 [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 15:03:08 kserve-ci-e2e-test deployment-controller Normal ScalingReplicaSet Scaled up replica set llmisvc8bf079eb6eda4debfb4ef5bb7817824c-kserve-router-scheduler-5f555d4d85 from 0 to 1 [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 None kserve-ci-e2e-test Normal Scheduled Successfully assigned kserve-ci-e2e-test/llmisvc8bf079eb6eda4debfb4ef5bb7817824c-kserve-router-scher4hs2 to ip-10-0-135-220.ec2.internal [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 15:03:09 kserve-ci-e2e-test multus Normal AddedInterface Add eth0 [10.133.0.39/23] from ovn-kubernetes [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 15:03:09 kserve-ci-e2e-test kubelet/ip-10-0-135-220.ec2.internal Normal Pulling Pulling image "ghcr.io/llm-d/llm-d-router-endpoint-picker:v0.9.0-rc.2" [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 15:03:10 kserve-ci-e2e-test attachdetach-controller Warning FailedAttachVolume Multi-Attach error for volume "pvc-20b17380-3f46-45a4-a8e5-36f5cb935bca" Volume is already used by pod(s) llmisvc-model-pvc-router-manage-59b9d263-kserve-585587bc9dnwscv [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 15:03:11 kserve-ci-e2e-test kubelet/ip-10-0-135-220.ec2.internal Normal Pulled Successfully pulled image "ghcr.io/llm-d/llm-d-router-endpoint-picker:v0.9.0-rc.2" in 2.885s (2.885s including waiting). Image size: 75073927 bytes. [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 15:03:12 kserve-ci-e2e-test kubelet/ip-10-0-135-220.ec2.internal Normal Created Created container: main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 15:03:12 kserve-ci-e2e-test kubelet/ip-10-0-135-220.ec2.internal Normal Started Started container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:31:40 kserve-ci-e2e-test replicaset-controller Normal SuccessfulCreate Created pod: llmisvca690bbc929faec8bc98c767f16c003c1-kserve-router-scheq44kv [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:31:40 kserve-ci-e2e-test deployment-controller Normal ScalingReplicaSet Scaled up replica set llmisvca690bbc929faec8bc98c767f16c003c1-kserve-router-scheduler-5d6998448f from 0 to 1 [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 None kserve-ci-e2e-test Normal Scheduled Successfully assigned kserve-ci-e2e-test/llmisvca690bbc929faec8bc98c767f16c003c1-kserve-router-scheq44kv to ip-10-0-133-25.ec2.internal [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:31:40 kserve-ci-e2e-test multus Normal AddedInterface Add eth0 [10.132.0.53/23] from ovn-kubernetes [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:31:41 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Pulled Container image "quay.io/opendatahub/kserve-storage-initializer@sha256:13ad765c0f323200e1db7b45c99661287e65c5158d6b54b16139829636d20be4" already present on machine [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:31:41 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Created Created container: storage-initializer [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:31:41 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Started Started container storage-initializer [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:31:42 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Pulled Container image "ghcr.io/llm-d/llm-d-router-endpoint-picker:v0.9.0-rc.2" already present on machine [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:31:42 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Created Created container: main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:31:42 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Started Started container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:23 kserve-ci-e2e-test replicaset-controller Normal SuccessfulCreate Created pod: llmisvcb11a9c9832b99b016bc8f8e0ea095712-kserve-router-scheknzg7 [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:23 kserve-ci-e2e-test deployment-controller Normal ScalingReplicaSet Scaled up replica set llmisvcb11a9c9832b99b016bc8f8e0ea095712-kserve-router-scheduler-6d86bd4d9d from 0 to 1 [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 None kserve-ci-e2e-test Normal Scheduled Successfully assigned kserve-ci-e2e-test/llmisvcb11a9c9832b99b016bc8f8e0ea095712-kserve-router-scheknzg7 to ip-10-0-133-25.ec2.internal [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:24 kserve-ci-e2e-test multus Normal AddedInterface Add eth0 [10.132.0.48/23] from ovn-kubernetes [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:24 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Pulled Container image "ghcr.io/llm-d/llm-d-router-endpoint-picker:v0.9.0-rc.2" already present on machine [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:24 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Created Created container: main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:25:24 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Started Started container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:26:13 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Killing Stopping container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:26:13 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Warning FailedPreStopHook PreStopHook failed [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 None kserve-ci-e2e-test Normal Scheduled Successfully assigned kserve-ci-e2e-test/llmisvcca2d2d7d499abb359505529ebe02c136-kserve-router-scheb2glb to ip-10-0-139-47.ec2.internal [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:18 kserve-ci-e2e-test multus Normal AddedInterface Add eth0 [10.134.0.28/23] from ovn-kubernetes [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:18 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Pulled Container image "ghcr.io/llm-d/llm-d-router-endpoint-picker:v0.9.0-rc.2" already present on machine [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:18 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Created Created container: main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:18 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Started Started container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:28:58 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Killing Stopping container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:28:58 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Warning FailedPreStopHook PreStopHook failed [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:17 kserve-ci-e2e-test replicaset-controller Normal SuccessfulCreate Created pod: llmisvcca2d2d7d499abb359505529ebe02c136-kserve-router-scheb2glb [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:27:17 kserve-ci-e2e-test deployment-controller Normal ScalingReplicaSet Scaled up replica set llmisvcca2d2d7d499abb359505529ebe02c136-kserve-router-scheduler-68cc9685d6 from 0 to 1 [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:24:43 kserve-ci-e2e-test replicaset-controller Normal SuccessfulCreate Created pod: llmisvce55ae740357a3a31a27cdb8b66ffe20f-kserve-router-schex4lqs [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:24:43 kserve-ci-e2e-test deployment-controller Normal ScalingReplicaSet Scaled up replica set llmisvce55ae740357a3a31a27cdb8b66ffe20f-kserve-router-scheduler-749449dbc8 from 0 to 1 [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 None kserve-ci-e2e-test Normal Scheduled Successfully assigned kserve-ci-e2e-test/llmisvce55ae740357a3a31a27cdb8b66ffe20f-kserve-router-schex4lqs to ip-10-0-139-47.ec2.internal [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:24:43 kserve-ci-e2e-test multus Normal AddedInterface Add eth0 [10.134.0.26/23] from ovn-kubernetes [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:24:43 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Pulled Container image "ghcr.io/llm-d/llm-d-router-endpoint-picker:v0.9.0-rc.2" already present on machine [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:24:43 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Created Created container: main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:24:43 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Started Started container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:26:33 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Killing Stopping container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:26:33 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Warning FailedPreStopHook PreStopHook failed [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 None kserve-ci-e2e-test Normal Scheduled Successfully assigned kserve-ci-e2e-test/precise-prefix-cache-test-kserve-5755bc98d9-f88w6 to ip-10-0-139-47.ec2.internal [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:55 kserve-ci-e2e-test multus Normal AddedInterface Add eth0 [10.134.0.22/23] from ovn-kubernetes [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:55 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Pulled Container image "quay.io/opendatahub/kserve-storage-initializer@sha256:13ad765c0f323200e1db7b45c99661287e65c5158d6b54b16139829636d20be4" already present on machine [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:55 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Created Created container: storage-initializer [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:55 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Started Started container storage-initializer [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:00 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Pulled Container image "ghcr.io/llm-d/llm-d-inference-sim:v0.8.2" already present on machine [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:01 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Created Created container: main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:01 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Started Started container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:23:10 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Killing Stopping container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 None kserve-ci-e2e-test Normal Scheduled Successfully assigned kserve-ci-e2e-test/precise-prefix-cache-test-kserve-5755bc98d9-hdpjr to ip-10-0-133-25.ec2.internal [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:55 kserve-ci-e2e-test multus Normal AddedInterface Add eth0 [10.132.0.45/23] from ovn-kubernetes [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:55 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Pulling Pulling image "quay.io/opendatahub/kserve-storage-initializer@sha256:13ad765c0f323200e1db7b45c99661287e65c5158d6b54b16139829636d20be4" [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:00 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Pulled Successfully pulled image "quay.io/opendatahub/kserve-storage-initializer@sha256:13ad765c0f323200e1db7b45c99661287e65c5158d6b54b16139829636d20be4" in 5.546s (5.546s including waiting). Image size: 301879255 bytes. [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:00 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Created Created container: storage-initializer [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:00 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Started Started container storage-initializer [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:07 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Pulling Pulling image "ghcr.io/llm-d/llm-d-inference-sim:v0.8.2" [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:08 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Pulled Successfully pulled image "ghcr.io/llm-d/llm-d-inference-sim:v0.8.2" in 1.434s (1.434s including waiting). Image size: 98346788 bytes. [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:09 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Created Created container: main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:09 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Started Started container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:23:10 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Killing Stopping container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:54 kserve-ci-e2e-test replicaset-controller Normal SuccessfulCreate Created pod: precise-prefix-cache-test-kserve-5755bc98d9-f88w6 [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:54 kserve-ci-e2e-test replicaset-controller Normal SuccessfulCreate Created pod: precise-prefix-cache-test-kserve-5755bc98d9-hdpjr [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:55 kserve-ci-e2e-test replicaset-controller Normal SuccessfulCreate Created pod: precise-prefix-cache-test-kserve-router-scheduler-647bfdf7jdncq [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 None kserve-ci-e2e-test Normal Scheduled Successfully assigned kserve-ci-e2e-test/precise-prefix-cache-test-kserve-router-scheduler-647bfdf7jdncq to ip-10-0-139-47.ec2.internal [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:55 kserve-ci-e2e-test multus Normal AddedInterface Add eth0 [10.134.0.23/23] from ovn-kubernetes [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:55 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Pulled Container image "quay.io/opendatahub/kserve-storage-initializer@sha256:13ad765c0f323200e1db7b45c99661287e65c5158d6b54b16139829636d20be4" already present on machine [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:55 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Created Created container: storage-initializer [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:55 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Started Started container storage-initializer [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:57 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Pulling Pulling image "ghcr.io/llm-d/llm-d-router-endpoint-picker:v0.9.0-rc.2" [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:58 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Pulled Successfully pulled image "ghcr.io/llm-d/llm-d-router-endpoint-picker:v0.9.0-rc.2" in 1.604s (1.604s including waiting). Image size: 75073927 bytes. [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:30 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Created Created container: main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:30 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Started Started container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:58 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Pulling Pulling image "ghcr.io/llm-d/llm-d-uds-tokenizer:vllm-v0.19.1" [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:29 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Pulled Successfully pulled image "ghcr.io/llm-d/llm-d-uds-tokenizer:vllm-v0.19.1" in 30.722s (30.722s including waiting). Image size: 2989890188 bytes. [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:29 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Created Created container: tokenizer [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:29 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Started Started container tokenizer [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:30 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Pulled Container image "ghcr.io/llm-d/llm-d-router-endpoint-picker:v0.9.0-rc.2" already present on machine [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:35 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Warning Unhealthy Startup probe failed: Get "http://10.134.0.23:8082/healthz": dial tcp 10.134.0.23:8082: connect: connection refused [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:46 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Warning Unhealthy Liveness probe failed: timeout: failed to connect service "10.134.0.23:9003" within 1s: context deadline exceeded [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:23:10 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Killing Stopping container tokenizer [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:23:10 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Normal Killing Stopping container main [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:23:10 kserve-ci-e2e-test kubelet/ip-10-0-139-47.ec2.internal Warning FailedPreStopHook PreStopHook failed [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:55 kserve-ci-e2e-test deployment-controller Normal ScalingReplicaSet Scaled up replica set precise-prefix-cache-test-kserve-router-scheduler-647bfdf79c from 0 to 1 [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:54 kserve-ci-e2e-test deployment-controller Normal ScalingReplicaSet Scaled up replica set precise-prefix-cache-test-kserve-5755bc98d9 from 0 to 2 [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:47 kserve-ci-e2e-test OpenDataHubModelController Warning ReconcileError Failed to reconcile LLMInferenceService: 1 error occurred: * failed to get HTTPRoute for AuthPolicy precise-prefix-cache-test-kserve-route-authn: failed to get HTTPRoute kserve-ci-e2e-test/precise-prefix-cache-test-kserve-route: HTTPRoute.gateway.networking.k8s.io "precise-prefix-cache-test-kserve-route" not found [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:54 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Secret kserve-ci-e2e-test/precise-prefix-cache-test-kserve-self-signed-certs [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:54 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Deployment kserve-ci-e2e-test/precise-prefix-cache-test-kserve [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:54 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Service kserve-ci-e2e-test/precise-prefix-cache-test-kserve-workload-svc [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:54 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.ServiceAccount kserve-ci-e2e-test/precise-prefix-cache-test-epp-sa [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:54 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.ClusterRoleBinding /kserve-ci-e2e-test-precise-prefix-cache-test-epp-auth-rb [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:54 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Role kserve-ci-e2e-test/precise-prefix-cache-test-epp-role [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:55 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.RoleBinding kserve-ci-e2e-test/precise-prefix-cache-test-epp-rb [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:55 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Deployment kserve-ci-e2e-test/precise-prefix-cache-test-kserve-router-scheduler [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:55 kserve-ci-e2e-test LLMInferenceServiceController Normal Created Created v1.Service kserve-ci-e2e-test/precise-prefix-cache-test-epp-service [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:26 kserve-ci-e2e-test LLMInferenceServiceController Normal Created (combined from similar events): Created v1.DestinationRule kserve-ci-e2e-test/precise-prefix-cache-test-kserve-shadow-svc [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:25 kserve-ci-e2e-test LLMInferenceServiceController Normal Updated Updated v1.Secret kserve-ci-e2e-test/precise-prefix-cache-test-kserve-self-signed-certs [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:26 kserve-ci-e2e-test LLMInferenceServiceController Normal Updated Updated v1.HTTPRoute kserve-ci-e2e-test/precise-prefix-cache-test-kserve-route [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:23:09 kserve-ci-e2e-test LLMInferenceServiceController Normal LLMInferenceServiceReady LLMInferenceService [precise-prefix-cache-test] is Ready [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:23:10 kserve-ci-e2e-test LLMInferenceServiceController Normal Deleted Deleted v1.ClusterRoleBinding /kserve-ci-e2e-test-precise-prefix-cache-test-epp-auth-rb [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 None kserve-ci-e2e-test Normal Scheduled Successfully assigned kserve-ci-e2e-test/router-gateway-1-openshift-default-799f46c59b-vt5f9 to ip-10-0-133-25.ec2.internal [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:36 kserve-ci-e2e-test multus Normal AddedInterface Add eth0 [10.132.0.44/23] from ovn-kubernetes [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:56 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Pulled Container image "registry.redhat.io/openshift-service-mesh/istio-proxyv2-rhel9@sha256:7d15cebf9b62f3f235c0eab5158ac8ff2fda86a1d193490dc94c301402c99da8" already present on machine [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:56 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Created Created container: istio-proxy [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:56 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Normal Started Started container istio-proxy [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:54 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Warning Unhealthy Readiness probe failed: Get "http://10.132.0.44:15021/healthz/ready": context deadline exceeded (Client.Timeout exceeded while awaiting headers) [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:22:02 kserve-ci-e2e-test kubelet/ip-10-0-133-25.ec2.internal Warning Unhealthy Startup probe failed: Get "http://10.132.0.44:15021/healthz/ready": dial tcp 10.132.0.44:15021: connect: connection refused [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:35 kserve-ci-e2e-test replicaset-controller Normal SuccessfulCreate Created pod: router-gateway-1-openshift-default-799f46c59b-vt5f9 [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:35 kserve-ci-e2e-test deployment-controller Normal ScalingReplicaSet Scaled up replica set router-gateway-1-openshift-default-799f46c59b from 0 to 1 [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:35 kserve-ci-e2e-test service-controller Normal EnsuringLoadBalancer Ensuring load balancer [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:21:40 kserve-ci-e2e-test service-controller Normal EnsuredLoadBalancer Ensured load balancer [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:56 2026-07-07 14:30:52 kserve-ci-e2e-test gateway_labeler_controller Normal AddedLabel Added label istio.io/rev=openshift-gateway to gateway router-gateway-1 [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:148 ### Pod custom-route-timeout-test-kserve-7f56b8c469-hkpbl (phase=Running) [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:188 #### init-container 'storage-initializer' (restarts=0) [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:201 # -- logs (current) -- [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:202 2026-07-07 14:48:58.209 1 storage.initializer INFO [initializer-entrypoint:():17] Initializing, args: (src_uri, dest_path): [('hf://facebook/opt-125m', '/mnt/models')] [e2e-llm-inference-service] 2026-07-07 14:48:58.209 1 storage.initializer INFO [kserve_storage.py:download():166] Copying contents of hf://facebook/opt-125m to local [e2e-llm-inference-service] Could not set the permissions on the file '/mnt/models/.cache/huggingface/download/wPaCkH-WbT7GsmxMKKrNZTV4nSM=.ac481c8eb05e4d2496fbe076a38a7b4835dd733d.incomplete'. Error: [Errno 13] Permission denied: '/mnt/tmp_fe89ed70-6992-49c0-99a1-087a606ea60b'. [e2e-llm-inference-service] Continuing without setting permissions. [e2e-llm-inference-service] Could not set the permissions on the file '/mnt/models/.cache/huggingface/download/5HHJ6px3_ZRDOG3OxNZMhuycwOk=.a591333512516f58bf2002045dece909a0ccdb8b.incomplete'. Error: [Errno 13] Permission denied: '/mnt/tmp_e54d9401-6cea-4c81-89ff-d72c8f8832ad'. [e2e-llm-inference-service] Continuing without setting permissions. [e2e-llm-inference-service] Could not set the permissions on the file '/mnt/models/.cache/huggingface/download/Xn7B-BWUGOee2Y6hCZtEhtFu4BE=.38c05904caf6e5b9f04ecda5c973d77e6c1da151.incomplete'. Error: [Errno 13] Permission denied: '/mnt/tmp_2af21743-bd5f-499b-b6f8-72669ae8c7dd'. [e2e-llm-inference-service] Continuing without setting permissions. [e2e-llm-inference-service] Could not set the permissions on the file '/mnt/models/.cache/huggingface/download/8_PA_wEVGiVa2goH2H4KQOQpvVY=.b3fb716a3024261980becb2382e31a3780985130.incomplete'. Error: [Errno 13] Permission denied: '/mnt/tmp_003b2096-da0a-4579-85de-ee0836524e52'. [e2e-llm-inference-service] Continuing without setting permissions. [e2e-llm-inference-service] Could not set the permissions on the file '/mnt/models/.cache/huggingface/download/gPcsVCQDYDHk-_n0G9uADl7PXIM=.61c60ec52ed43038fff0fbbd68b080c94b0d94b4c8458dbd65965f9b17631c89.incomplete'. Error: [Errno 13] Permission denied: '/mnt/tmp_2f2ad314-f891-461f-98a2-d59e603205a3'. [e2e-llm-inference-service] Continuing without setting permissions. [e2e-llm-inference-service] Could not set the permissions on the file '/mnt/models/.cache/huggingface/download/3EVKVggOldJcKSsGjSdoUCN1AyQ=.cf739e3ba86db7791ebab2828cc34b8a5acd3a86.incomplete'. Error: [Errno 13] Permission denied: '/mnt/tmp_f7cd5c1f-8f79-4dd9-ac99-aeeb595ede6f'. [e2e-llm-inference-service] Continuing without setting permissions. [e2e-llm-inference-service] Could not set the permissions on the file '/mnt/models/.cache/huggingface/download/PtHk0z_I45atnj23IIRhTExwT3w=.226b0752cac7789c48f0cb3ec53eda48b7be36cc.incomplete'. Error: [Errno 13] Permission denied: '/mnt/tmp_d7f1d6a6-be81-4ecb-99de-d7b829a27b2a'. [e2e-llm-inference-service] Continuing without setting permissions. [e2e-llm-inference-service] Could not set the permissions on the file '/mnt/models/.cache/huggingface/download/Q1p2l2BzM1m6P5jKvr8WTq1TUio=.2d74da6615135c58cf3cf9ad4cb11e7c613ff9e55fe658a47ab83b6c8d1174a9.incomplete'. Error: [Errno 13] Permission denied: '/mnt/tmp_0f410242-045c-42b0-978c-cfd3b6213f6a'. [e2e-llm-inference-service] Continuing without setting permissions. [e2e-llm-inference-service] Could not set the permissions on the file '/mnt/models/.cache/huggingface/download/ahkChHUJFxEmOdq5GDFEmerRzCY=.5dfa36546b8eddce0e04df3133c30df43fcc3828.incomplete'. Error: [Errno 13] Permission denied: '/mnt/tmp_df4fba8d-868a-4337-9fc4-2047fba3bb94'. [e2e-llm-inference-service] Continuing without setting permissions. [e2e-llm-inference-service] Could not set the permissions on the file '/mnt/models/.cache/huggingface/download/a7eHxRFT3OeMBIFg52k2nfj5m7w=.db7090b0c8b34dd957a7e0656c718f978f9203cc874018f37dda44108be5970a.incomplete'. Error: [Errno 13] Permission denied: '/mnt/tmp_983eb0b5-57f3-476c-b533-9ab9f16b1f93'. [e2e-llm-inference-service] Continuing without setting permissions. [e2e-llm-inference-service] Could not set the permissions on the file '/mnt/models/.cache/huggingface/download/vzaExXFZNBay89bvlQv-ZcI6BTg=.27c24ca9d908d0b678b20c698aeb9e950c44d865.incomplete'. Error: [Errno 13] Permission denied: '/mnt/tmp_2d023ff8-72ca-44c2-809a-306a989279f6'. [e2e-llm-inference-service] Continuing without setting permissions. [e2e-llm-inference-service] Could not set the permissions on the file '/mnt/models/.cache/huggingface/download/j3m-Hy6QvBddw8RXA1uSWl1AJ0c=.0a39732b2d8be8e493cab3da68b68cc3e28221de.incomplete'. Error: [Errno 13] Permission denied: '/mnt/tmp_18a01f94-ed24-423a-956a-e57e2f00db86'. [e2e-llm-inference-service] Continuing without setting permissions. [e2e-llm-inference-service] 2026-07-07 14:49:02.531 1 storage.initializer INFO [kserve_storage.py:download():234] Successfully copied hf://facebook/opt-125m to /mnt/models [e2e-llm-inference-service] 2026-07-07 14:49:02.531 1 storage.initializer INFO [kserve_storage.py:download():235] Model downloaded in 4.3219754369997645 seconds. [e2e-llm-inference-service] [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:188 #### container 'main' (restarts=0) [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:201 # -- logs (current) -- [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:202 (APIServer pid=1) DEBUG 07-07 15:03:16 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:03:17 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:03:17 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:03:18 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:03:19 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:03:20 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:03:21 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:03:22 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:03:23 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:03:24 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:03:25 [v1/metrics/loggers.py:259] Engine 000: Avg prompt throughput: 0.0 tokens/s, Avg generation throughput: 0.0 tokens/s, Running: 0 reqs, Waiting: 0 reqs, GPU KV cache usage: 0.0%, Prefix cache hit rate: 0.0% [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:03:25 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:03:26 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:03:27 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:03:27 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:03:28 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:03:29 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:03:30 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:03:31 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:03:32 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:03:33 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:03:34 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:03:35 [v1/metrics/loggers.py:259] Engine 000: Avg prompt throughput: 0.0 tokens/s, Avg generation throughput: 0.0 tokens/s, Running: 0 reqs, Waiting: 0 reqs, GPU KV cache usage: 0.0%, Prefix cache hit rate: 0.0% [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:03:35 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:03:36 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:03:37 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:03:37 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:03:38 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:03:39 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:03:40 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:03:41 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:03:42 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:03:43 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:03:44 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:03:45 [v1/metrics/loggers.py:259] Engine 000: Avg prompt throughput: 0.0 tokens/s, Avg generation throughput: 0.0 tokens/s, Running: 0 reqs, Waiting: 0 reqs, GPU KV cache usage: 0.0%, Prefix cache hit rate: 0.0% [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:03:45 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:03:46 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:03:47 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:03:47 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:03:48 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:03:49 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:03:50 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:03:51 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:03:52 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:03:53 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:03:54 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:03:55 [v1/metrics/loggers.py:259] Engine 000: Avg prompt throughput: 0.0 tokens/s, Avg generation throughput: 0.0 tokens/s, Running: 0 reqs, Waiting: 0 reqs, GPU KV cache usage: 0.0%, Prefix cache hit rate: 0.0% [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:03:55 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:03:56 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:03:57 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:03:57 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:03:58 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:03:59 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:04:00 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:04:01 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:04:02 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:04:03 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:04:04 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:04:05 [v1/metrics/loggers.py:259] Engine 000: Avg prompt throughput: 0.0 tokens/s, Avg generation throughput: 0.0 tokens/s, Running: 0 reqs, Waiting: 0 reqs, GPU KV cache usage: 0.0%, Prefix cache hit rate: 0.0% [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:04:05 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:04:06 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:04:07 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:04:07 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:04:08 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:04:09 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:04:10 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:04:11 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:04:12 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:04:13 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:04:14 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:04:15 [v1/metrics/loggers.py:259] Engine 000: Avg prompt throughput: 0.0 tokens/s, Avg generation throughput: 0.0 tokens/s, Running: 0 reqs, Waiting: 0 reqs, GPU KV cache usage: 0.0%, Prefix cache hit rate: 0.0% [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:04:15 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:04:16 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:04:17 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:04:17 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:04:18 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:04:19 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:04:20 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:04:21 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:04:22 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:04:23 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:04:24 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:04:25 [v1/metrics/loggers.py:259] Engine 000: Avg prompt throughput: 0.0 tokens/s, Avg generation throughput: 0.0 tokens/s, Running: 0 reqs, Waiting: 0 reqs, GPU KV cache usage: 0.0%, Prefix cache hit rate: 0.0% [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:04:25 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:04:26 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:04:27 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:04:27 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:04:28 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:04:29 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:04:30 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:04:31 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:04:32 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:04:33 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:04:34 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:04:35 [v1/metrics/loggers.py:259] Engine 000: Avg prompt throughput: 0.0 tokens/s, Avg generation throughput: 0.0 tokens/s, Running: 0 reqs, Waiting: 0 reqs, GPU KV cache usage: 0.0%, Prefix cache hit rate: 0.0% [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:04:35 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:04:36 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:04:37 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:04:37 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:04:38 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:04:39 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:04:40 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:04:41 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:04:42 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:04:43 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:04:44 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:04:45 [v1/metrics/loggers.py:259] Engine 000: Avg prompt throughput: 0.0 tokens/s, Avg generation throughput: 0.0 tokens/s, Running: 0 reqs, Waiting: 0 reqs, GPU KV cache usage: 0.0%, Prefix cache hit rate: 0.0% [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:04:45 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:04:46 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:04:47 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:04:47 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:04:48 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:04:49 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:04:50 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:04:51 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:04:52 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:04:53 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:04:54 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:04:55 [v1/metrics/loggers.py:259] Engine 000: Avg prompt throughput: 0.0 tokens/s, Avg generation throughput: 0.0 tokens/s, Running: 0 reqs, Waiting: 0 reqs, GPU KV cache usage: 0.0%, Prefix cache hit rate: 0.0% [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:04:55 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:04:56 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:04:57 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:04:57 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:04:58 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:04:59 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:05:00 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:05:01 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:05:02 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:05:03 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:05:04 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:05:05 [v1/metrics/loggers.py:259] Engine 000: Avg prompt throughput: 0.0 tokens/s, Avg generation throughput: 0.0 tokens/s, Running: 0 reqs, Waiting: 0 reqs, GPU KV cache usage: 0.0%, Prefix cache hit rate: 0.0% [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:05:05 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:05:06 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:05:07 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:05:07 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:05:08 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:05:09 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:05:10 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:05:11 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:05:12 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:05:13 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:05:14 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:05:15 [v1/metrics/loggers.py:259] Engine 000: Avg prompt throughput: 0.0 tokens/s, Avg generation throughput: 0.0 tokens/s, Running: 0 reqs, Waiting: 0 reqs, GPU KV cache usage: 0.0%, Prefix cache hit rate: 0.0% [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:05:15 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:05:16 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:05:17 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:05:17 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:05:18 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:05:19 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:05:20 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:05:21 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:05:22 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:05:23 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:05:24 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:05:25 [v1/metrics/loggers.py:259] Engine 000: Avg prompt throughput: 0.0 tokens/s, Avg generation throughput: 0.0 tokens/s, Running: 0 reqs, Waiting: 0 reqs, GPU KV cache usage: 0.0%, Prefix cache hit rate: 0.0% [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:05:25 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:05:26 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:05:27 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:05:27 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:05:28 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:05:29 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:05:30 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:05:31 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:05:32 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:05:33 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:05:34 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:05:35 [v1/metrics/loggers.py:259] Engine 000: Avg prompt throughput: 0.0 tokens/s, Avg generation throughput: 0.0 tokens/s, Running: 0 reqs, Waiting: 0 reqs, GPU KV cache usage: 0.0%, Prefix cache hit rate: 0.0% [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:05:35 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:05:36 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:05:37 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:05:37 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:05:38 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:05:39 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:05:40 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:05:41 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:05:42 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:05:43 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:05:44 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:05:45 [v1/metrics/loggers.py:259] Engine 000: Avg prompt throughput: 0.0 tokens/s, Avg generation throughput: 0.0 tokens/s, Running: 0 reqs, Waiting: 0 reqs, GPU KV cache usage: 0.0%, Prefix cache hit rate: 0.0% [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:05:45 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:05:46 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:05:47 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:05:47 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:05:48 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:05:49 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:05:50 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:05:51 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:05:52 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:05:53 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:05:54 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:05:55 [v1/metrics/loggers.py:259] Engine 000: Avg prompt throughput: 0.0 tokens/s, Avg generation throughput: 0.0 tokens/s, Running: 0 reqs, Waiting: 0 reqs, GPU KV cache usage: 0.0%, Prefix cache hit rate: 0.0% [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:05:55 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:05:56 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:05:57 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:05:57 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:05:58 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:05:59 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:06:00 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:06:01 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] (APIServer pid=1) DEBUG 07-07 15:06:02 [v1/engine/async_llm.py:875] Called check_health. [e2e-llm-inference-service] [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:148 ### Pod custom-route-timeout-test-kserve-router-scheduler-6578bddfnbxnb (phase=Running) [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:188 #### init-container 'storage-initializer' (restarts=0) [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:201 # -- logs (current) -- [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:202 2026-07-07 14:48:58.765 1 storage.initializer INFO [initializer-entrypoint:():17] Initializing, args: (src_uri, dest_path): [('hf://facebook/opt-125m', '/mnt/models')] [e2e-llm-inference-service] 2026-07-07 14:48:58.766 1 storage.initializer INFO [kserve_storage.py:download():166] Copying contents of hf://facebook/opt-125m to local [e2e-llm-inference-service] 2026-07-07 14:48:58.766 1 storage.initializer INFO [kserve_storage.py:download():169] Allow patterns: ['tokenizer.json', 'tokenizer_config.json', 'special_tokens_map.json', 'vocab.json', 'merges.txt', 'config.json', 'generation_config.json'] [e2e-llm-inference-service] Could not set the permissions on the file '/mnt/models/.cache/huggingface/download/8_PA_wEVGiVa2goH2H4KQOQpvVY=.b3fb716a3024261980becb2382e31a3780985130.incomplete'. Error: [Errno 13] Permission denied: '/mnt/tmp_82b8d195-1343-44af-ad6c-98f0508595eb'. [e2e-llm-inference-service] Continuing without setting permissions. [e2e-llm-inference-service] Could not set the permissions on the file '/mnt/models/.cache/huggingface/download/3EVKVggOldJcKSsGjSdoUCN1AyQ=.cf739e3ba86db7791ebab2828cc34b8a5acd3a86.incomplete'. Error: [Errno 13] Permission denied: '/mnt/tmp_6d8eda33-c00e-4ff7-bd49-ec03cdecb3ae'. [e2e-llm-inference-service] Continuing without setting permissions. [e2e-llm-inference-service] Could not set the permissions on the file '/mnt/models/.cache/huggingface/download/PtHk0z_I45atnj23IIRhTExwT3w=.226b0752cac7789c48f0cb3ec53eda48b7be36cc.incomplete'. Error: [Errno 13] Permission denied: '/mnt/tmp_9376046d-19c5-4019-bfdd-3bed40609458'. [e2e-llm-inference-service] Continuing without setting permissions. [e2e-llm-inference-service] Could not set the permissions on the file '/mnt/models/.cache/huggingface/download/ahkChHUJFxEmOdq5GDFEmerRzCY=.5dfa36546b8eddce0e04df3133c30df43fcc3828.incomplete'. Error: [Errno 13] Permission denied: '/mnt/tmp_a5ebd9df-1e9c-4a77-a68b-2855510aa2f6'. [e2e-llm-inference-service] Continuing without setting permissions. [e2e-llm-inference-service] Could not set the permissions on the file '/mnt/models/.cache/huggingface/download/vzaExXFZNBay89bvlQv-ZcI6BTg=.27c24ca9d908d0b678b20c698aeb9e950c44d865.incomplete'. Error: [Errno 13] Permission denied: '/mnt/tmp_9f16baed-882a-49af-9a0e-1ac024c2c9be'. [e2e-llm-inference-service] Continuing without setting permissions. [e2e-llm-inference-service] Could not set the permissions on the file '/mnt/models/.cache/huggingface/download/j3m-Hy6QvBddw8RXA1uSWl1AJ0c=.0a39732b2d8be8e493cab3da68b68cc3e28221de.incomplete'. Error: [Errno 13] Permission denied: '/mnt/tmp_ed5fd362-119a-4fcc-ac9d-548720bff411'. [e2e-llm-inference-service] Continuing without setting permissions. [e2e-llm-inference-service] 2026-07-07 14:48:59.253 1 storage.initializer INFO [kserve_storage.py:download():234] Successfully copied hf://facebook/opt-125m to /mnt/models [e2e-llm-inference-service] 2026-07-07 14:48:59.253 1 storage.initializer INFO [kserve_storage.py:download():235] Model downloaded in 0.48768534999999247 seconds. [e2e-llm-inference-service] [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:188 #### container 'main' (restarts=0) [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:201 # -- logs (current) -- [e2e-llm-inference-service] INFO e2e.llmisvc.logging:diagnostic.py:202 {"level":"info","ts":1783435739.9154088,"logger":"setup","caller":"runner/runner.go:196","msg":"GIE build","commit-sha":"181aa8358916e19b8844ccc752b2d6153d4b2ad6","build-ref":"v0.9.0-rc.2"} [e2e-llm-inference-service] Flag --model-server-metrics-scheme has been deprecated, This flag is deprecated. Configure via EndpointPickerConfig data layer plugin parameters instead. [e2e-llm-inference-service] {"level":"info","ts":1783435739.9158351,"logger":"setup","caller":"runner/runner.go:217","msg":"Flags processed","flags":{"cert-path":"/var/run/kserve/tls","config-file":"","config-text":"apiVersion: inference.networking.x-k8s.io/v1alpha1\nkind: EndpointPickerConfig\nplugins:\n- type: single-profile-handler\n- type: queue-scorer\n- type: prefix-cache-scorer\n- type: max-score-picker\nschedulingProfiles:\n- name: default\n plugins:\n - pluginRef: queue-scorer\n weight: 2\n - pluginRef: prefix-cache-scorer\n weight: 3\n - pluginRef: max-score-picker\n","disable-endpoint-subset-filter":false,"enable-cert-reload":true,"enable-grpc-stream-metrics":false,"enable-pprof":true,"endpoint-selector":"","endpoint-target-ports":{},"grpc-health-port":9003,"grpc-max-recv-msg-size":"","grpc-max-send-msg-size":"","grpc-port":9002,"ha-enable-leader-election":false,"health-checking":false,"metrics-endpoint-auth":true,"metrics-port":9090,"metrics-staleness-threshold":2000000000,"model-server-metrics-https-insecure-skip-verify":true,"model-server-metrics-path":"/metrics","model-server-metrics-port":0,"model-server-metrics-scheme":"https","pool-group":"inference.networking.k8s.io","pool-name":"custom-route-timeout-test-inference-pool","pool-namespace":"kserve-ci-e2e-test","refresh-metrics-interval":50000000,"refresh-prometheus-metrics-interval":5000000000,"secure-serving":true,"tracing":true,"v":2,"zap-devel":{},"zap-encoder":{},"zap-log-level":{},"zap-stacktrace-level":{},"zap-time-encoding":{}}} [e2e-llm-inference-service] {"level":"info","ts":1783435739.916115,"logger":"setup.trace","caller":"tracing/telemetry.go:123","msg":"init OTel trace exporter","type":"console"} [e2e-llm-inference-service] {"level":"info","ts":1783435739.9170027,"caller":"loader/configloader.go:89","msg":"DEPRECATION: apiVersion inference.networking.x-k8s.io/v1alpha1/EndpointPickerConfig is deprecated","replacement":"llm-d.ai/v1alpha1/EndpointPickerConfig"} [e2e-llm-inference-service] {"level":"info","ts":1783435739.917046,"caller":"loader/configloader.go:121","msg":"Loaded raw configuration","config":"{Plugins: [{Type: single-profile-handler} {Type: queue-scorer} {Type: prefix-cache-scorer} {Type: max-score-picker}], SchedulingProfiles: [{Name: default, Plugins: [{PluginRef: queue-scorer, Weight: 2.00} {PluginRef: prefix-cache-scorer, Weight: 3.00} {PluginRef: max-score-picker}]}]}"} [e2e-llm-inference-service] {"level":"info","ts":1783435739.9170556,"logger":"setup","caller":"runner/runner.go:622","msg":"Data layer: ENABLED"} [e2e-llm-inference-service] {"level":"info","ts":1783435739.917349,"logger":"setup","caller":"runner/runner.go:281","msg":"Raw config after phase one","config":{"apiVersion":"inference.networking.x-k8s.io/v1alpha1","dataLayer":null,"kind":"EndpointPickerConfig","plugins":[{"name":"single-profile-handler","parameters":null,"type":"single-profile-handler"},{"name":"queue-scorer","parameters":null,"type":"queue-scorer"},{"name":"prefix-cache-scorer","parameters":null,"type":"prefix-cache-scorer"},{"name":"max-score-picker","parameters":null,"type":"max-score-picker"}],"schedulingProfiles":[{"name":"default","plugins":[{"pluginRef":"queue-scorer","weight":2},{"pluginRef":"prefix-cache-scorer","weight":3},{"pluginRef":"max-score-picker","weight":null}]}]}} [e2e-llm-inference-service] {"level":"info","ts":1783435739.9370096,"logger":"utilization-detector/utilization-detector","caller":"utilization/detector.go:83","msg":"Creating new UtilizationDetector","queueDepthThreshold":5,"kvCacheUtilThreshold":0.8,"metricsStalenessThreshold":"200ms","headroom":0} [e2e-llm-inference-service] {"level":"info","ts":1783435739.9371223,"caller":"metrics/factories.go:230","msg":"Registered engine mapping","engine":"vllm","mapping":"Mapping{all specs enabled}"} [e2e-llm-inference-service] {"level":"info","ts":1783435739.9371636,"caller":"metrics/factories.go:230","msg":"Registered engine mapping","engine":"sglang","mapping":"Mapping{disabled: [lora]}"} [e2e-llm-inference-service] {"level":"info","ts":1783435739.9372008,"caller":"metrics/factories.go:230","msg":"Registered engine mapping","engine":"trtllm-serve","mapping":"Mapping{disabled: [lora, cacheInfo]}"} [e2e-llm-inference-service] {"level":"info","ts":1783435739.937265,"caller":"metrics/factories.go:230","msg":"Registered engine mapping","engine":"triton-tensorrt-llm","mapping":"Mapping{disabled: [lora, cacheInfo]}"} [e2e-llm-inference-service] {"level":"info","ts":1783435739.9372845,"caller":"metrics/factories.go:230","msg":"Registered engine mapping","engine":"triton","mapping":"Mapping{disabled: [kv, lora, cacheInfo]}"} [e2e-llm-inference-service] {"level":"info","ts":1783435739.9373307,"caller":"loader/configloader.go:154","msg":"Instantiated all plugins and applied system defaults. Effective raw configuration","config":"{Plugins: [{Name: single-profile-handler, Type: single-profile-handler} {Name: queue-scorer, Type: queue-scorer} {Name: prefix-cache-scorer, Type: prefix-cache-scorer} {Name: max-score-picker, Type: max-score-picker} {Name: fcfs-ordering-policy, Type: fcfs-ordering-policy} {Name: global-strict-fairness-policy, Type: global-strict-fairness-policy} {Name: static-usage-limit-policy, Type: static-usage-limit-policy} {Name: openai-parser, Type: openai-parser} {Name: anthropic-parser, Type: anthropic-parser} {Name: vllmhttp-parser, Type: vllmhttp-parser} {Name: utilization-detector, Type: utilization-detector} {Name: metrics-data-source, Type: metrics-data-source} {Name: core-metrics-extractor, Type: core-metrics-extractor}], SchedulingProfiles: [{Name: default, Plugins: [{PluginRef: queue-scorer, Weight: 2.00} {PluginRef: prefix-cache-scorer, Weight: 3.00} {PluginRef: max-score-picker}]}], DataLayer: {Sources: [{PluginRef: metrics-data-source, Extractors: [{PluginRef: core-metrics-extractor}]}], Discovery: }, FlowControl: {MaxBytes: unlimited, MaxRequests: unlimited, SaturationDetector: {PluginRef: utilization-detector}}, RequestHandler: {Parsers: [{PluginRef: openai-parser}, {PluginRef: anthropic-parser}, {PluginRef: vllmhttp-parser}]}}"} [e2e-llm-inference-service] {"level":"info","ts":1783435739.93738,"caller":"approximateprefix/plugin.go:88","msg":"Prefix DataProducer initialized","config":{"autoTune":true,"blockSizeTokens":16,"blockSize":0,"maxPrefixBlocksToMatch":2048,"maxPrefixTokensToMatch":131072,"lruCapacityPerServer":31250}} [e2e-llm-inference-service] {"level":"info","ts":1783435739.9374442,"caller":"approximateprefix/plugin.go:111","msg":"WARNING: configured blockSizeTokens is below the recommended minimum, overriding it.","blockSizeTokens":16,"minimum":64,"issue":"https://github.com/llm-d/llm-d-router/issues/1158"} [e2e-llm-inference-service] {"level":"info","ts":1783435739.9374657,"caller":"datalayer/data_graph.go:116","msg":"auto-created default producer","producer":"approx-prefix-cache-producer/approx-prefix-cache-producer","dataKey":"PrefixCacheMatchInfoDataKey/approx-prefix-cache-producer","consumer":"prefix-cache-scorer"} [e2e-llm-inference-service] {"level":"info","ts":1783435739.9374921,"caller":"datalayer/data_graph.go:116","msg":"auto-created default producer","producer":"token-producer/token-producer","dataKey":"TokenizedPrompt/token-producer","consumer":"approx-prefix-cache-producer"} [e2e-llm-inference-service] {"level":"info","ts":1783435739.9375703,"caller":"runner/runner.go:685","msg":"loaded configuration from file/text successfully"} [e2e-llm-inference-service] {"level":"info","ts":1783435739.9375947,"logger":"setup","caller":"runner/runner.go:308","msg":"EPP config after phase two","config":"{SchedulerConfig:{ProfileHandler: single-profile-handler/single-profile-handler, Profiles: map[default:{Filters: [], Scorers: [queue-scorer/queue-scorer: 2.000000, prefix-cache-scorer/prefix-cache-scorer: 3.000000], Picker: max-score-picker/max-score-picker}]} SaturationDetector:0xc0007decc0 DataConfig:{Sources:[{Plugin:0xc0003314d0 Extractors:[0xc0007deec0]}]} FlowControlConfig: ParserRegistry:0xc0007df340}"} [e2e-llm-inference-service] {"level":"info","ts":1783435739.954819,"logger":"setup","caller":"runner/runner.go:352","msg":"Setting pprof handlers"} [e2e-llm-inference-service] {"level":"info","ts":1783435739.954849,"caller":"manager/internal.go:201","msg":"Registering metrics http server extra handler","path":"/debug/pprof/cmdline"} [e2e-llm-inference-service] {"level":"info","ts":1783435739.9548626,"caller":"manager/internal.go:201","msg":"Registering metrics http server extra handler","path":"/debug/pprof/profile"} [e2e-llm-inference-service] {"level":"info","ts":1783435739.9548676,"caller":"manager/internal.go:201","msg":"Registering metrics http server extra handler","path":"/debug/pprof/heap"} [e2e-llm-inference-service] {"level":"info","ts":1783435739.9548724,"caller":"manager/internal.go:201","msg":"Registering metrics http server extra handler","path":"/debug/pprof/allocs"} [e2e-llm-inference-service] {"level":"info","ts":1783435739.9548771,"caller":"manager/internal.go:201","msg":"Registering metrics http server extra handler","path":"/debug/pprof/threadcreate"} [e2e-llm-inference-service] {"level":"info","ts":1783435739.9548814,"caller":"manager/internal.go:201","msg":"Registering metrics http server extra handler","path":"/debug/pprof/mutex"} [e2e-llm-inference-service] {"level":"info","ts":1783435739.9548857,"caller":"manager/internal.go:201","msg":"Registering metrics http server extra handler","path":"/debug/pprof/"} [e2e-llm-inference-service] {"level":"info","ts":1783435739.9548905,"caller":"manager/internal.go:201","msg":"Registering metrics http server extra handler","path":"/debug/pprof/symbol"} [e2e-llm-inference-service] {"level":"info","ts":1783435739.9548957,"caller":"manager/internal.go:201","msg":"Registering metrics http server extra handler","path":"/debug/pprof/trace"} [e2e-llm-inference-service] {"level":"info","ts":1783435739.9549003,"caller":"manager/internal.go:201","msg":"Registering metrics http server extra handler","path":"/debug/pprof/goroutine"} [e2e-llm-inference-service] {"level":"info","ts":1783435739.9549046,"caller":"manager/internal.go:201","msg":"Registering metrics http server extra handler","path":"/debug/pprof/block"} [e2e-llm-inference-service] {"level":"info","ts":1783435739.9549172,"caller":"manager/internal.go:201","msg":"Registering metrics http server extra handler","path":"/debug/plugins/state"} [e2e-llm-inference-service] {"level":"info","ts":1783435739.9549232,"logger":"setup","caller":"runner/runner.go:373","msg":"parsed config","scheduler-config":"{ProfileHandler: single-profile-handler/single-profile-handler, Profiles: map[default:{Filters: [], Scorers: [queue-scorer/queue-scorer: 2.000000, prefix-cache-scorer/prefix-cache-scorer: 3.000000], Picker: max-score-picker/max-score-picker}]}"} [e2e-llm-inference-service] {"level":"info","ts":1783435739.9549487,"logger":"setup","caller":"datalayer/runtime.go:99","msg":"Configuring datalayer runtime","numSources":1} [e2e-llm-inference-service] {"level":"info","ts":1783435739.9549575,"logger":"setup","caller":"datalayer/runtime.go:118","msg":"Processing source","source":"metrics-data-source","numExtractors":1} [e2e-llm-inference-service] {"level":"info","ts":1783435739.9549763,"logger":"setup","caller":"datalayer/runtime.go:147","msg":"Source configured","source":"metrics-data-source","extractors":["core-metrics-extractor/core-metrics-extractor"]} [e2e-llm-inference-service] {"level":"info","ts":1783435739.9549906,"logger":"setup","caller":"datalayer/runtime.go:206","msg":"Datalayer runtime configured","pollers":1,"notifiers":0,"endpointSources":0} [e2e-llm-inference-service] {"level":"info","ts":1783435739.9550002,"logger":"setup","caller":"runner/runner.go:833","msg":"Experimental Flow Control layer is disabled, using legacy admission control"} [e2e-llm-inference-service] {"level":"info","ts":1783435739.9550784,"logger":"setup","caller":"runner/runner.go:721","msg":"ExtProc server runner added to manager."} [e2e-llm-inference-service] {"level":"info","ts":1783435739.9550931,"logger":"setup","caller":"runner/runner.go:260","msg":"Controller manager starting"} [e2e-llm-inference-service] {"level":"info","ts":1783435739.9551284,"logger":"controller-runtime.metrics","caller":"server/server.go:208","msg":"Starting metrics server"} [e2e-llm-inference-service] {"level":"info","ts":1783435739.955351,"caller":"runnable/grpc.go:35","msg":"gRPC server starting","name":"health"} [e2e-llm-inference-service] {"level":"info","ts":1783435739.9554353,"caller":"runnable/grpc.go:43","msg":"gRPC server listening","name":"health","port":9003} [e2e-llm-inference-service] {"level":"info","ts":1783435739.9555283,"logger":"controller-runtime.metrics","caller":"server/server.go:247","msg":"Serving metrics server","bindAddress":":9090","secure":false} [e2e-llm-inference-service] {"level":"info","ts":1783435739.9556868,"caller":"controller/controller.go:370","msg":"Starting EventSource","controller":"inferenceobjective","controllerGroup":"inference.networking.x-k8s.io","controllerKind":"InferenceObjective","source":"kind source: *v1alpha2.InferenceObjective"} [e2e-llm-inference-service] {"level":"info","ts":1783435739.9557478,"caller":"controller/controller.go:370","msg":"Starting EventSource","controller":"pod","controllerGroup":"","controllerKind":"Pod","source":"kind source: *v1.Pod"} [e2e-llm-inference-service] {"level":"info","ts":1783435739.9556766,"caller":"controller/controller.go:370","msg":"Starting EventSource","controller":"inferencepool","controllerGroup":"inference.networking.k8s.io","controllerKind":"InferencePool","source":"kind source: *v1.InferencePool"} [e2e-llm-inference-service] {"level":"info","ts":1783435739.955684,"caller":"controller/controller.go:370","msg":"Starting EventSource","controller":"inferencemodelrewrite","controllerGroup":"inference.networking.x-k8s.io","controllerKind":"InferenceModelRewrite","source":"kind source: *v1alpha2.InferenceModelRewrite"} [e2e-llm-inference-service] {"level":"info","ts":1783435739.956464,"caller":"runnable/grpc.go:35","msg":"gRPC server starting","name":"ext-proc"} [e2e-llm-inference-service] {"level":"info","ts":1783435739.9565628,"caller":"runnable/grpc.go:43","msg":"gRPC server listening","name":"ext-proc","port":9002} [e2e-llm-inference-service] {"level":"info","ts":1783435739.9602182,"logger":"controller-runtime.cache","caller":"cache/reflector.go:446","msg":"Caches populated","type":"*v1alpha2.InferenceModelRewrite","reflector":"pkg/mod/k8s.io/client-go@v0.35.6/tools/cache/reflector.go:289"} [e2e-llm-inference-service] {"level":"info","ts":1783435739.9602242,"logger":"controller-runtime.cache","caller":"cache/reflector.go:446","msg":"Caches populated","type":"*v1alpha2.InferenceObjective","reflector":"pkg/mod/k8s.io/client-go@v0.35.6/tools/cache/reflector.go:289"} [e2e-llm-inference-service] {"level":"info","ts":1783435739.9604378,"logger":"controller-runtime.cache","caller":"cache/reflector.go:446","msg":"Caches populated","type":"*v1.InferencePool","reflector":"pkg/mod/k8s.io/client-go@v0.35.6/tools/cache/reflector.go:289"} [e2e-llm-inference-service] {"level":"info","ts":1783435739.9623933,"logger":"controller-runtime.cache","caller":"cache/reflector.go:446","msg":"Caches populated","type":"*v1.Pod","reflector":"pkg/mod/k8s.io/client-go@v0.35.6/tools/cache/reflector.go:289"} [e2e-llm-inference-service] {"level":"info","ts":1783435740.0572743,"caller":"controller/controller.go:303","msg":"Starting Controller","controller":"inferencemodelrewrite","controllerGroup":"inference.networking.x-k8s.io","controllerKind":"InferenceModelRewrite"} [e2e-llm-inference-service] {"level":"info","ts":1783435740.0573955,"caller":"controller/controller.go:306","msg":"Starting workers","controller":"inferencemodelrewrite","controllerGroup":"inference.networking.x-k8s.io","controllerKind":"InferenceModelRewrite","worker count":1} [e2e-llm-inference-service] {"level":"info","ts":1783435740.0572884,"caller":"controller/controller.go:303","msg":"Starting Controller","controller":"inferenceobjective","controllerGroup":"inference.networking.x-k8s.io","controllerKind":"InferenceObjective"} [e2e-llm-inference-service] {"level":"info","ts":1783435740.057559,"caller":"controller/controller.go:306","msg":"Starting workers","controller":"inferenceobjective","controllerGroup":"inference.networking.x-k8s.io","controllerKind":"InferenceObjective","worker count":1} [e2e-llm-inference-service] {"level":"info","ts":1783435740.0578568,"caller":"controller/controller.go:303","msg":"Starting Controller","controller":"inferencepool","controllerGroup":"inference.networking.k8s.io","controllerKind":"InferencePool"} [e2e-llm-inference-service] {"level":"info","ts":1783435740.0578845,"caller":"controller/controller.go:306","msg":"Starting workers","controller":"inferencepool","controllerGroup":"inference.networking.k8s.io","controllerKind":"InferencePool","worker count":1} [e2e-llm-inference-service] {"level":"info","ts":1783435740.0580494,"caller":"controller/inferencepool_reconciler.go:46","msg":"Reconciling InferencePool","controller":"inferencepool","controllerGroup":"inference.networking.k8s.io","controllerKind":"InferencePool","InferencePool":{"name":"custom-route-timeout-test-inference-pool","namespace":"kserve-ci-e2e-test"},"namespace":"kserve-ci-e2e-test","name":"custom-route-timeout-test-inference-pool","reconcileID":"a806e727-0634-4cb4-a52c-37c3ee6cfd2a"} [e2e-llm-inference-service] {"level":"info","ts":1783435740.1581156,"caller":"controller/controller.go:303","msg":"Starting Controller","controller":"pod","controllerGroup":"","controllerKind":"Pod"} [e2e-llm-inference-service] {"level":"info","ts":1783435740.1581774,"caller":"controller/controller.go:306","msg":"Starting workers","controller":"pod","controllerGroup":"","controllerKind":"Pod","worker count":1} [e2e-llm-inference-service] {"level":"info","ts":1783435751.8541691,"caller":"controller/inferencepool_reconciler.go:46","msg":"Reconciling InferencePool","controller":"inferencepool","controllerGroup":"inference.networking.k8s.io","controllerKind":"InferencePool","InferencePool":{"name":"custom-route-timeout-test-inference-pool","namespace":"kserve-ci-e2e-test"},"namespace":"kserve-ci-e2e-test","name":"custom-route-timeout-test-inference-pool","reconcileID":"0cdee3b8-73ec-47ea-be52-a87d0400f5c4"} [e2e-llm-inference-service] {"level":"info","ts":1783435857.7993805,"caller":"controller/pod_reconciler.go:99","msg":"Pod already exists","controller":"pod","controllerGroup":"","controllerKind":"Pod","Pod":{"name":"custom-route-timeout-test-kserve-7f56b8c469-hkpbl","namespace":"kserve-ci-e2e-test"},"namespace":"kserve-ci-e2e-test","name":"custom-route-timeout-test-kserve-7f56b8c469-hkpbl","reconcileID":"0f6ef02c-254f-443b-9249-e63b5a3b2670"} [e2e-llm-inference-service] [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1254 --- [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1255 metadata: [e2e-llm-inference-service] name: custom-route-timeout-test-epp-service [e2e-llm-inference-service] namespace: kserve-ci-e2e-test [e2e-llm-inference-service] uid: c610a373-ec20-4a90-bfc5-bfa87ad2e5c7 [e2e-llm-inference-service] resourceVersion: '49590' [e2e-llm-inference-service] creationTimestamp: '2026-07-07T14:48:57Z' [e2e-llm-inference-service] labels: [e2e-llm-inference-service] app.kubernetes.io/component: llminferenceservice-router-scheduler [e2e-llm-inference-service] app.kubernetes.io/name: custom-route-timeout-test [e2e-llm-inference-service] app.kubernetes.io/part-of: llminferenceservice [e2e-llm-inference-service] endpoints.kubernetes.io/managed-by: endpoint-controller [e2e-llm-inference-service] annotations: [e2e-llm-inference-service] endpoints.kubernetes.io/last-change-trigger-time: '2026-07-07T14:49:31Z' [e2e-llm-inference-service] managedFields: [e2e-llm-inference-service] - manager: kube-controller-manager [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] apiVersion: v1 [e2e-llm-inference-service] time: '2026-07-07T14:49:31Z' [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:metadata: [e2e-llm-inference-service] f:annotations: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:endpoints.kubernetes.io/last-change-trigger-time: {} [e2e-llm-inference-service] f:labels: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:app.kubernetes.io/component: {} [e2e-llm-inference-service] f:app.kubernetes.io/name: {} [e2e-llm-inference-service] f:app.kubernetes.io/part-of: {} [e2e-llm-inference-service] f:endpoints.kubernetes.io/managed-by: {} [e2e-llm-inference-service] f:subsets: {} [e2e-llm-inference-service] subsets: [e2e-llm-inference-service] - addresses: [e2e-llm-inference-service] - ip: 10.132.0.55 [e2e-llm-inference-service] nodeName: ip-10-0-133-25.ec2.internal [e2e-llm-inference-service] targetRef: [e2e-llm-inference-service] kind: Pod [e2e-llm-inference-service] namespace: kserve-ci-e2e-test [e2e-llm-inference-service] name: custom-route-timeout-test-kserve-router-scheduler-6578bddfnbxnb [e2e-llm-inference-service] uid: f8a7664e-b142-4621-b773-b5e45ceaa80f [e2e-llm-inference-service] ports: [e2e-llm-inference-service] - name: grpc-health [e2e-llm-inference-service] port: 9003 [e2e-llm-inference-service] protocol: TCP [e2e-llm-inference-service] - name: metrics [e2e-llm-inference-service] port: 9090 [e2e-llm-inference-service] protocol: TCP [e2e-llm-inference-service] - name: zmq [e2e-llm-inference-service] port: 5557 [e2e-llm-inference-service] protocol: TCP [e2e-llm-inference-service] - name: grpc [e2e-llm-inference-service] port: 9002 [e2e-llm-inference-service] protocol: TCP [e2e-llm-inference-service] apiVersion: v1 [e2e-llm-inference-service] kind: Endpoints [e2e-llm-inference-service] [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1254 --- [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1255 metadata: [e2e-llm-inference-service] name: custom-route-timeout-test-kserve-workload-svc [e2e-llm-inference-service] namespace: kserve-ci-e2e-test [e2e-llm-inference-service] uid: 7fdb8710-f832-4596-9e66-db9be0486f61 [e2e-llm-inference-service] resourceVersion: '50894' [e2e-llm-inference-service] creationTimestamp: '2026-07-07T14:48:57Z' [e2e-llm-inference-service] labels: [e2e-llm-inference-service] app.kubernetes.io/component: llminferenceservice-workload [e2e-llm-inference-service] app.kubernetes.io/name: custom-route-timeout-test [e2e-llm-inference-service] app.kubernetes.io/part-of: llminferenceservice [e2e-llm-inference-service] endpoints.kubernetes.io/managed-by: endpoint-controller [e2e-llm-inference-service] annotations: [e2e-llm-inference-service] endpoints.kubernetes.io/last-change-trigger-time: '2026-07-07T14:50:57Z' [e2e-llm-inference-service] managedFields: [e2e-llm-inference-service] - manager: kube-controller-manager [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] apiVersion: v1 [e2e-llm-inference-service] time: '2026-07-07T14:50:57Z' [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:metadata: [e2e-llm-inference-service] f:annotations: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:endpoints.kubernetes.io/last-change-trigger-time: {} [e2e-llm-inference-service] f:labels: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:app.kubernetes.io/component: {} [e2e-llm-inference-service] f:app.kubernetes.io/name: {} [e2e-llm-inference-service] f:app.kubernetes.io/part-of: {} [e2e-llm-inference-service] f:endpoints.kubernetes.io/managed-by: {} [e2e-llm-inference-service] f:subsets: {} [e2e-llm-inference-service] subsets: [e2e-llm-inference-service] - addresses: [e2e-llm-inference-service] - ip: 10.134.0.33 [e2e-llm-inference-service] nodeName: ip-10-0-139-47.ec2.internal [e2e-llm-inference-service] targetRef: [e2e-llm-inference-service] kind: Pod [e2e-llm-inference-service] namespace: kserve-ci-e2e-test [e2e-llm-inference-service] name: custom-route-timeout-test-kserve-7f56b8c469-hkpbl [e2e-llm-inference-service] uid: 79bc5e6e-61ea-4684-bc6f-a2fb38b92a73 [e2e-llm-inference-service] ports: [e2e-llm-inference-service] - name: https [e2e-llm-inference-service] port: 8000 [e2e-llm-inference-service] protocol: TCP [e2e-llm-inference-service] appProtocol: https [e2e-llm-inference-service] apiVersion: v1 [e2e-llm-inference-service] kind: Endpoints [e2e-llm-inference-service] [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1254 --- [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1255 metadata: [e2e-llm-inference-service] name: custom-route-timeout-test-kserve-7f56b8c469-hkpbl [e2e-llm-inference-service] generateName: custom-route-timeout-test-kserve-7f56b8c469- [e2e-llm-inference-service] namespace: kserve-ci-e2e-test [e2e-llm-inference-service] uid: 79bc5e6e-61ea-4684-bc6f-a2fb38b92a73 [e2e-llm-inference-service] resourceVersion: '50892' [e2e-llm-inference-service] generation: 1 [e2e-llm-inference-service] creationTimestamp: '2026-07-07T14:48:57Z' [e2e-llm-inference-service] labels: [e2e-llm-inference-service] app.kubernetes.io/component: llminferenceservice-workload [e2e-llm-inference-service] app.kubernetes.io/name: custom-route-timeout-test [e2e-llm-inference-service] app.kubernetes.io/part-of: llminferenceservice [e2e-llm-inference-service] kserve.io/component: workload [e2e-llm-inference-service] llm-d.ai/role: both [e2e-llm-inference-service] pod-template-hash: 7f56b8c469 [e2e-llm-inference-service] annotations: [e2e-llm-inference-service] k8s.ovn.org/pod-networks: '{"default":{"ip_addresses":["10.134.0.33/23"],"mac_address":"0a:58:0a:86:00:21","gateway_ips":["10.134.0.1"],"routes":[{"dest":"10.132.0.0/14","nextHop":"10.134.0.1"},{"dest":"172.31.0.0/16","nextHop":"10.134.0.1"},{"dest":"169.254.0.5/32","nextHop":"10.134.0.1"},{"dest":"100.64.0.0/16","nextHop":"10.134.0.1"}],"ip_address":"10.134.0.33/23","gateway_ip":"10.134.0.1","role":"primary"}}' [e2e-llm-inference-service] k8s.v1.cni.cncf.io/network-status: "[{\n \"name\": \"ovn-kubernetes\",\n \ [e2e-llm-inference-service] \ \"interface\": \"eth0\",\n \"ips\": [\n \"10.134.0.33\"\n ],\n\ [e2e-llm-inference-service] \ \"mac\": \"0a:58:0a:86:00:21\",\n \"default\": true,\n \"dns\": {}\n\ [e2e-llm-inference-service] }]" [e2e-llm-inference-service] openshift.io/scc: restricted-v2 [e2e-llm-inference-service] seccomp.security.alpha.kubernetes.io/pod: runtime/default [e2e-llm-inference-service] security.openshift.io/validated-scc-subject-type: user [e2e-llm-inference-service] ownerReferences: [e2e-llm-inference-service] - apiVersion: apps/v1 [e2e-llm-inference-service] kind: ReplicaSet [e2e-llm-inference-service] name: custom-route-timeout-test-kserve-7f56b8c469 [e2e-llm-inference-service] uid: 908c7ec2-616f-4fd2-a6a5-bea58b91399e [e2e-llm-inference-service] controller: true [e2e-llm-inference-service] blockOwnerDeletion: true [e2e-llm-inference-service] managedFields: [e2e-llm-inference-service] - manager: ip-10-0-139-47 [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] apiVersion: v1 [e2e-llm-inference-service] time: '2026-07-07T14:48:57Z' [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:metadata: [e2e-llm-inference-service] f:annotations: [e2e-llm-inference-service] f:k8s.ovn.org/pod-networks: {} [e2e-llm-inference-service] subresource: status [e2e-llm-inference-service] - manager: kube-controller-manager [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] apiVersion: v1 [e2e-llm-inference-service] time: '2026-07-07T14:48:57Z' [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:metadata: [e2e-llm-inference-service] f:generateName: {} [e2e-llm-inference-service] f:labels: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:app.kubernetes.io/component: {} [e2e-llm-inference-service] f:app.kubernetes.io/name: {} [e2e-llm-inference-service] f:app.kubernetes.io/part-of: {} [e2e-llm-inference-service] f:kserve.io/component: {} [e2e-llm-inference-service] f:llm-d.ai/role: {} [e2e-llm-inference-service] f:pod-template-hash: {} [e2e-llm-inference-service] f:ownerReferences: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"uid":"908c7ec2-616f-4fd2-a6a5-bea58b91399e"}: {} [e2e-llm-inference-service] f:spec: [e2e-llm-inference-service] f:containers: [e2e-llm-inference-service] k:{"name":"main"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:command: {} [e2e-llm-inference-service] f:env: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"name":"HF_HUB_CACHE"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"HOME"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"TORCHINDUCTOR_CACHE_DIR"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"USER"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"VLLM_CPU_KVCACHE_SPACE"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"VLLM_ENABLE_V1_MULTIPROCESSING"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"VLLM_LOGGING_LEVEL"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] f:image: {} [e2e-llm-inference-service] f:imagePullPolicy: {} [e2e-llm-inference-service] f:lifecycle: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:preStop: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:exec: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:command: {} [e2e-llm-inference-service] f:livenessProbe: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:failureThreshold: {} [e2e-llm-inference-service] f:httpGet: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:path: {} [e2e-llm-inference-service] f:port: {} [e2e-llm-inference-service] f:scheme: {} [e2e-llm-inference-service] f:periodSeconds: {} [e2e-llm-inference-service] f:successThreshold: {} [e2e-llm-inference-service] f:timeoutSeconds: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:ports: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"containerPort":8000,"protocol":"TCP"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:containerPort: {} [e2e-llm-inference-service] f:protocol: {} [e2e-llm-inference-service] f:readinessProbe: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:failureThreshold: {} [e2e-llm-inference-service] f:httpGet: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:path: {} [e2e-llm-inference-service] f:port: {} [e2e-llm-inference-service] f:scheme: {} [e2e-llm-inference-service] f:periodSeconds: {} [e2e-llm-inference-service] f:successThreshold: {} [e2e-llm-inference-service] f:timeoutSeconds: {} [e2e-llm-inference-service] f:resources: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:limits: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:cpu: {} [e2e-llm-inference-service] f:memory: {} [e2e-llm-inference-service] f:requests: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:cpu: {} [e2e-llm-inference-service] f:memory: {} [e2e-llm-inference-service] f:securityContext: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:allowPrivilegeEscalation: {} [e2e-llm-inference-service] f:capabilities: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:drop: {} [e2e-llm-inference-service] f:readOnlyRootFilesystem: {} [e2e-llm-inference-service] f:runAsNonRoot: {} [e2e-llm-inference-service] f:seccompProfile: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:type: {} [e2e-llm-inference-service] f:startupProbe: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:failureThreshold: {} [e2e-llm-inference-service] f:httpGet: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:path: {} [e2e-llm-inference-service] f:port: {} [e2e-llm-inference-service] f:scheme: {} [e2e-llm-inference-service] f:periodSeconds: {} [e2e-llm-inference-service] f:successThreshold: {} [e2e-llm-inference-service] f:timeoutSeconds: {} [e2e-llm-inference-service] f:terminationMessagePath: {} [e2e-llm-inference-service] f:terminationMessagePolicy: {} [e2e-llm-inference-service] f:volumeMounts: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"mountPath":"/dev/shm"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:mountPath: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] k:{"mountPath":"/home"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:mountPath: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] k:{"mountPath":"/mnt/models"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:mountPath: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:readOnly: {} [e2e-llm-inference-service] k:{"mountPath":"/models"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:mountPath: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] k:{"mountPath":"/tmp"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:mountPath: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] k:{"mountPath":"/var/run/kserve/tls"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:mountPath: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:readOnly: {} [e2e-llm-inference-service] f:dnsPolicy: {} [e2e-llm-inference-service] f:enableServiceLinks: {} [e2e-llm-inference-service] f:initContainers: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"name":"storage-initializer"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:args: {} [e2e-llm-inference-service] f:env: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"name":"AWS_ACCESS_KEY_ID"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:valueFrom: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:secretKeyRef: {} [e2e-llm-inference-service] k:{"name":"AWS_CA_BUNDLE"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"AWS_CA_BUNDLE_CONFIGMAP"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"AWS_ENDPOINT_URL"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"AWS_SECRET_ACCESS_KEY"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:valueFrom: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:secretKeyRef: {} [e2e-llm-inference-service] k:{"name":"HF_HUB_ENABLE_HF_TRANSFER"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"HF_XET_HIGH_PERFORMANCE"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"HF_XET_NUM_CONCURRENT_RANGE_GETS"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"S3_ENDPOINT"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"S3_USE_HTTPS"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"S3_VERIFY_SSL"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] f:image: {} [e2e-llm-inference-service] f:imagePullPolicy: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:resources: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:limits: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:cpu: {} [e2e-llm-inference-service] f:memory: {} [e2e-llm-inference-service] f:requests: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:cpu: {} [e2e-llm-inference-service] f:memory: {} [e2e-llm-inference-service] f:terminationMessagePath: {} [e2e-llm-inference-service] f:terminationMessagePolicy: {} [e2e-llm-inference-service] f:volumeMounts: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"mountPath":"/mnt/models"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:mountPath: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:restartPolicy: {} [e2e-llm-inference-service] f:schedulerName: {} [e2e-llm-inference-service] f:securityContext: {} [e2e-llm-inference-service] f:terminationGracePeriodSeconds: {} [e2e-llm-inference-service] f:volumes: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"name":"dshm"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:emptyDir: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:medium: {} [e2e-llm-inference-service] f:sizeLimit: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] k:{"name":"home"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:emptyDir: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] k:{"name":"kserve-provision-location"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:emptyDir: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] k:{"name":"model-cache"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:emptyDir: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] k:{"name":"tls-certs"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:secret: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:defaultMode: {} [e2e-llm-inference-service] f:secretName: {} [e2e-llm-inference-service] k:{"name":"tmp-dir"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:emptyDir: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] - manager: multus-daemon [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] apiVersion: v1 [e2e-llm-inference-service] time: '2026-07-07T14:48:57Z' [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:metadata: [e2e-llm-inference-service] f:annotations: [e2e-llm-inference-service] f:k8s.v1.cni.cncf.io/network-status: {} [e2e-llm-inference-service] subresource: status [e2e-llm-inference-service] - manager: kubelet [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] apiVersion: v1 [e2e-llm-inference-service] time: '2026-07-07T14:50:57Z' [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:status: [e2e-llm-inference-service] f:conditions: [e2e-llm-inference-service] k:{"type":"ContainersReady"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:lastProbeTime: {} [e2e-llm-inference-service] f:lastTransitionTime: {} [e2e-llm-inference-service] f:observedGeneration: {} [e2e-llm-inference-service] f:status: {} [e2e-llm-inference-service] f:type: {} [e2e-llm-inference-service] k:{"type":"Initialized"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:lastProbeTime: {} [e2e-llm-inference-service] f:lastTransitionTime: {} [e2e-llm-inference-service] f:observedGeneration: {} [e2e-llm-inference-service] f:status: {} [e2e-llm-inference-service] f:type: {} [e2e-llm-inference-service] k:{"type":"PodReadyToStartContainers"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:lastProbeTime: {} [e2e-llm-inference-service] f:lastTransitionTime: {} [e2e-llm-inference-service] f:observedGeneration: {} [e2e-llm-inference-service] f:status: {} [e2e-llm-inference-service] f:type: {} [e2e-llm-inference-service] k:{"type":"PodScheduled"}: [e2e-llm-inference-service] f:observedGeneration: {} [e2e-llm-inference-service] k:{"type":"Ready"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:lastProbeTime: {} [e2e-llm-inference-service] f:lastTransitionTime: {} [e2e-llm-inference-service] f:observedGeneration: {} [e2e-llm-inference-service] f:status: {} [e2e-llm-inference-service] f:type: {} [e2e-llm-inference-service] f:containerStatuses: {} [e2e-llm-inference-service] f:hostIP: {} [e2e-llm-inference-service] f:hostIPs: {} [e2e-llm-inference-service] f:initContainerStatuses: {} [e2e-llm-inference-service] f:observedGeneration: {} [e2e-llm-inference-service] f:phase: {} [e2e-llm-inference-service] f:podIP: {} [e2e-llm-inference-service] f:podIPs: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"ip":"10.134.0.33"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:ip: {} [e2e-llm-inference-service] f:startTime: {} [e2e-llm-inference-service] subresource: status [e2e-llm-inference-service] spec: [e2e-llm-inference-service] volumes: [e2e-llm-inference-service] - name: home [e2e-llm-inference-service] emptyDir: {} [e2e-llm-inference-service] - name: dshm [e2e-llm-inference-service] emptyDir: [e2e-llm-inference-service] medium: Memory [e2e-llm-inference-service] sizeLimit: 1Gi [e2e-llm-inference-service] - name: model-cache [e2e-llm-inference-service] emptyDir: {} [e2e-llm-inference-service] - name: tmp-dir [e2e-llm-inference-service] emptyDir: {} [e2e-llm-inference-service] - name: tls-certs [e2e-llm-inference-service] secret: [e2e-llm-inference-service] secretName: custom-route-timeout-test-kserve-self-signed-certs [e2e-llm-inference-service] defaultMode: 420 [e2e-llm-inference-service] - name: kserve-provision-location [e2e-llm-inference-service] emptyDir: {} [e2e-llm-inference-service] - name: kube-api-access-992bp [e2e-llm-inference-service] projected: [e2e-llm-inference-service] sources: [e2e-llm-inference-service] - serviceAccountToken: [e2e-llm-inference-service] expirationSeconds: 3607 [e2e-llm-inference-service] path: token [e2e-llm-inference-service] - configMap: [e2e-llm-inference-service] name: kube-root-ca.crt [e2e-llm-inference-service] items: [e2e-llm-inference-service] - key: ca.crt [e2e-llm-inference-service] path: ca.crt [e2e-llm-inference-service] - downwardAPI: [e2e-llm-inference-service] items: [e2e-llm-inference-service] - path: namespace [e2e-llm-inference-service] fieldRef: [e2e-llm-inference-service] apiVersion: v1 [e2e-llm-inference-service] fieldPath: metadata.namespace [e2e-llm-inference-service] - configMap: [e2e-llm-inference-service] name: openshift-service-ca.crt [e2e-llm-inference-service] items: [e2e-llm-inference-service] - key: service-ca.crt [e2e-llm-inference-service] path: service-ca.crt [e2e-llm-inference-service] defaultMode: 420 [e2e-llm-inference-service] initContainers: [e2e-llm-inference-service] - name: storage-initializer [e2e-llm-inference-service] image: quay.io/opendatahub/kserve-storage-initializer@sha256:13ad765c0f323200e1db7b45c99661287e65c5158d6b54b16139829636d20be4 [e2e-llm-inference-service] args: [e2e-llm-inference-service] - hf://facebook/opt-125m [e2e-llm-inference-service] - /mnt/models [e2e-llm-inference-service] env: [e2e-llm-inference-service] - name: AWS_ACCESS_KEY_ID [e2e-llm-inference-service] valueFrom: [e2e-llm-inference-service] secretKeyRef: [e2e-llm-inference-service] name: seaweedfs-s3-creds [e2e-llm-inference-service] key: AWS_ACCESS_KEY_ID [e2e-llm-inference-service] - name: AWS_SECRET_ACCESS_KEY [e2e-llm-inference-service] valueFrom: [e2e-llm-inference-service] secretKeyRef: [e2e-llm-inference-service] name: seaweedfs-s3-creds [e2e-llm-inference-service] key: AWS_SECRET_ACCESS_KEY [e2e-llm-inference-service] - name: S3_USE_HTTPS [e2e-llm-inference-service] value: '0' [e2e-llm-inference-service] - name: S3_ENDPOINT [e2e-llm-inference-service] value: s3-service.kserve:8333 [e2e-llm-inference-service] - name: AWS_ENDPOINT_URL [e2e-llm-inference-service] value: http://s3-service.kserve:8333 [e2e-llm-inference-service] - name: S3_VERIFY_SSL [e2e-llm-inference-service] value: '0' [e2e-llm-inference-service] - name: AWS_CA_BUNDLE [e2e-llm-inference-service] value: /etc/ssl/custom-certs/cabundle.crt [e2e-llm-inference-service] - name: AWS_CA_BUNDLE_CONFIGMAP [e2e-llm-inference-service] value: odh-kserve-custom-ca-bundle [e2e-llm-inference-service] - name: HF_HUB_ENABLE_HF_TRANSFER [e2e-llm-inference-service] value: '1' [e2e-llm-inference-service] - name: HF_XET_HIGH_PERFORMANCE [e2e-llm-inference-service] value: '1' [e2e-llm-inference-service] - name: HF_XET_NUM_CONCURRENT_RANGE_GETS [e2e-llm-inference-service] value: '8' [e2e-llm-inference-service] resources: [e2e-llm-inference-service] limits: [e2e-llm-inference-service] cpu: '1' [e2e-llm-inference-service] memory: 24Gi [e2e-llm-inference-service] requests: [e2e-llm-inference-service] cpu: 100m [e2e-llm-inference-service] memory: 100Mi [e2e-llm-inference-service] volumeMounts: [e2e-llm-inference-service] - name: kserve-provision-location [e2e-llm-inference-service] mountPath: /mnt/models [e2e-llm-inference-service] - name: kube-api-access-992bp [e2e-llm-inference-service] readOnly: true [e2e-llm-inference-service] mountPath: /var/run/secrets/kubernetes.io/serviceaccount [e2e-llm-inference-service] terminationMessagePath: /dev/termination-log [e2e-llm-inference-service] terminationMessagePolicy: FallbackToLogsOnError [e2e-llm-inference-service] imagePullPolicy: IfNotPresent [e2e-llm-inference-service] securityContext: [e2e-llm-inference-service] capabilities: [e2e-llm-inference-service] drop: [e2e-llm-inference-service] - ALL [e2e-llm-inference-service] runAsUser: 1000690000 [e2e-llm-inference-service] runAsNonRoot: true [e2e-llm-inference-service] allowPrivilegeEscalation: false [e2e-llm-inference-service] containers: [e2e-llm-inference-service] - name: main [e2e-llm-inference-service] image: public.ecr.aws/q9t5s3a7/vllm-cpu-release-repo:v0.19.0 [e2e-llm-inference-service] command: [e2e-llm-inference-service] - /bin/bash [e2e-llm-inference-service] - -c [e2e-llm-inference-service] - "if [ -f /etc/profile.d/ibm-aiu-setup.sh ]; then\n source /etc/profile.d/ibm-aiu-setup.sh\n\ [e2e-llm-inference-service] fi\n\nif [ \"$KSERVE_INFER_ROCE\" = \"true\" ]; then\n echo \"Trying to infer\ [e2e-llm-inference-service] \ RoCE configs ... \"\n grep -H . /sys/class/infiniband/*/ports/*/gids/* 2>/dev/null\n\ [e2e-llm-inference-service] \ grep -H . /sys/class/infiniband/*/ports/*/gid_attrs/types/* 2>/dev/null\n\ [e2e-llm-inference-service] \n cat /proc/driver/nvidia/params\n\n KSERVE_INFER_IB_GID_INDEX_GREP=${KSERVE_INFER_IB_GID_INDEX_GREP:-\"\ [e2e-llm-inference-service] RoCE v2\"}\n\n echo \"[Infer RoCE] Discovering active HCAs ...\"\n active_hcas=()\n\ [e2e-llm-inference-service] \ # Loop through all mlx5 devices found in sysfs\n for hca_dir in /sys/class/infiniband/mlx5_*;\ [e2e-llm-inference-service] \ do\n # Ensure it's a directory before proceeding\n if [ -d \"$hca_dir\"\ [e2e-llm-inference-service] \ ]; then\n hca_name=$(basename \"$hca_dir\")\n port_state_file=\"\ [e2e-llm-inference-service] $hca_dir/ports/1/state\" # Assume port 1\n type_file=\"$hca_dir/ports/1/gid_attrs/types/*\"\ [e2e-llm-inference-service] \n\n echo \"[Infer RoCE] Check if the port state file ${port_state_file}\ [e2e-llm-inference-service] \ exists and contains 'ACTIVE'\"\n if [ -f \"$port_state_file\" ] &&\ [e2e-llm-inference-service] \ grep -q \"ACTIVE\" \"$port_state_file\" && grep -q \"${KSERVE_INFER_IB_GID_INDEX_GREP}\"\ [e2e-llm-inference-service] \ ${type_file} 2>/dev/null; then\n echo \"[Infer RoCE] Found active\ [e2e-llm-inference-service] \ HCA: $hca_name\"\n active_hcas+=(\"$hca_name\")\n else\n\ [e2e-llm-inference-service] \ echo \"[Infer RoCE] Skipping inactive or down HCA: $hca_name\"\ [e2e-llm-inference-service] \n fi\n fi\n done\n\n # Check if we found any active HCAs\n\ [e2e-llm-inference-service] \ if [ ${#active_hcas[@]} -gt 0 ]; then\n # Join the array elements with\ [e2e-llm-inference-service] \ a comma\n hca_port_pairs=()\n for hca in \"${active_hcas[@]}\";\ [e2e-llm-inference-service] \ do\n hca_port_pairs+=(\"${hca}:1\")\n done\n\n active_hca_list=$(IFS=,;\ [e2e-llm-inference-service] \ echo \"${active_hcas[*]}\")\n hca_port_pairs_list=$(IFS=,; echo \"${hca_port_pairs[*]}\"\ [e2e-llm-inference-service] )\n echo \"[Infer RoCE] Setting active HCAs: ${active_hca_list}\"\n \ [e2e-llm-inference-service] \ export NCCL_IB_HCA=${NCCL_IB_HCA:-${active_hca_list}}\n export NVSHMEM_HCA_LIST=${NVSHMEM_HCA_LIST:-${hca_port_pairs_list}}\n\ [e2e-llm-inference-service] \ export UCX_NET_DEVICES=${UCX_NET_DEVICES:-${hca_port_pairs_list}}\n\n\ [e2e-llm-inference-service] \ echo \"[Infer RoCE] NCCL_IB_HCA=${NCCL_IB_HCA}\"\n echo \"[Infer\ [e2e-llm-inference-service] \ RoCE] NVSHMEM_HCA_LIST=${NVSHMEM_HCA_LIST}\"\n echo \"[Infer RoCE] UCX_NET_DEVICES=${UCX_NET_DEVICES}\"\ [e2e-llm-inference-service] \n else\n echo \"[Infer RoCE] WARNING: No active RoCE HCAs found. NCCL_IB_HCA\ [e2e-llm-inference-service] \ will not be set.\"\n fi\n\n if [ ${#active_hcas[@]} -gt 0 ]; then\n \ [e2e-llm-inference-service] \ echo \"[Infer RoCE] Finding GID_INDEX for each active HCA (SR-IOV compatible)...\"\ [e2e-llm-inference-service] \n\n # For SR-IOV environments, find the most common IPv4 RoCE v2 GID index\ [e2e-llm-inference-service] \ across all HCAs\n declare -A gid_index_count\n declare -A hca_gid_index\n\ [e2e-llm-inference-service] \n for hca_name in \"${active_hcas[@]}\"; do\n echo \"[Infer RoCE]\ [e2e-llm-inference-service] \ Processing HCA: ${hca_name}\"\n\n # Find all RoCE v2 IPv4 GIDs for\ [e2e-llm-inference-service] \ this HCA and count by index\n for tpath in /sys/class/infiniband/${hca_name}/ports/1/gid_attrs/types/*;\ [e2e-llm-inference-service] \ do\n if grep -q \"${KSERVE_INFER_IB_GID_INDEX_GREP}\" \"$tpath\"\ [e2e-llm-inference-service] \ 2>/dev/null; then\n idx=$(basename \"$tpath\")\n \ [e2e-llm-inference-service] \ gid_file=\"/sys/class/infiniband/${hca_name}/ports/1/gids/${idx}\"\ [e2e-llm-inference-service] \n # Check for IPv4 GID (contains ffff:)\n \ [e2e-llm-inference-service] \ if [ -f \"$gid_file\" ] && grep -q \"ffff:\" \"$gid_file\"; then\n \ [e2e-llm-inference-service] \ gid_value=$(cat \"$gid_file\" 2>/dev/null || echo \"\")\n \ [e2e-llm-inference-service] \ echo \"[Infer RoCE] Found IPv4 RoCE v2 GID for ${hca_name}:\ [e2e-llm-inference-service] \ index=${idx}, gid=${gid_value}\"\n hca_gid_index[\"${hca_name}\"\ [e2e-llm-inference-service] ]=\"${idx}\"\n gid_index_count[\"${idx}\"]=$((${gid_index_count[\"\ [e2e-llm-inference-service] ${idx}\"]} + 1))\n break # Use first found IPv4 GID per\ [e2e-llm-inference-service] \ HCA\n fi\n fi\n done\n done\n\n\ [e2e-llm-inference-service] \ # Find the most common GID index (most likely to be consistent across\ [e2e-llm-inference-service] \ nodes)\n best_gid_index=\"\"\n max_count=0\n for idx in \"\ [e2e-llm-inference-service] ${!gid_index_count[@]}\"; do\n count=${gid_index_count[\"${idx}\"]}\n\ [e2e-llm-inference-service] \ echo \"[Infer RoCE] GID_INDEX ${idx} found on ${count} HCAs\"\n \ [e2e-llm-inference-service] \ if [ $count -gt $max_count ]; then\n max_count=$count\n\ [e2e-llm-inference-service] \ best_gid_index=\"$idx\"\n fi\n done\n\n #\ [e2e-llm-inference-service] \ Use deterministic fallback if tied - prefer index 3 (SR-IOV standard)\n \ [e2e-llm-inference-service] \ if [ ${#gid_index_count[@]} -gt 1 ]; then\n echo \"[Infer RoCE]\ [e2e-llm-inference-service] \ Multiple GID indices found, selecting most common: ${best_gid_index}\"\n \ [e2e-llm-inference-service] \ # If there's a tie, prefer index 3 as it's most common in SR-IOV setups\n\ [e2e-llm-inference-service] \ if [ -n \"${gid_index_count['3']}\" ] && [ \"${gid_index_count['3']}\"\ [e2e-llm-inference-service] \ -eq \"$max_count\" ]; then\n best_gid_index=\"3\"\n \ [e2e-llm-inference-service] \ echo \"[Infer RoCE] Using deterministic fallback: GID_INDEX=3 (SR-IOV\ [e2e-llm-inference-service] \ standard)\"\n fi\n fi\n\n # Check if GID_INDEX is already\ [e2e-llm-inference-service] \ set via environment variables\n if [ -n \"${NCCL_IB_GID_INDEX}\" ]; then\n\ [e2e-llm-inference-service] \ echo \"[Infer RoCE] Using pre-configured NCCL_IB_GID_INDEX=${NCCL_IB_GID_INDEX}\ [e2e-llm-inference-service] \ from environment\"\n export NVSHMEM_IB_GID_INDEX=${NVSHMEM_IB_GID_INDEX:-$NCCL_IB_GID_INDEX}\n\ [e2e-llm-inference-service] \ export UCX_IB_GID_INDEX=${UCX_IB_GID_INDEX:-$NCCL_IB_GID_INDEX}\n\ [e2e-llm-inference-service] \ echo \"[Infer RoCE] Using pre-configured GID_INDEX=${NCCL_IB_GID_INDEX}\ [e2e-llm-inference-service] \ for NCCL, NVSHMEM, and UCX\"\n elif [ -n \"$best_gid_index\" ]; then\n\ [e2e-llm-inference-service] \ echo \"[Infer RoCE] Selected GID_INDEX: ${best_gid_index} (found\ [e2e-llm-inference-service] \ on ${max_count} HCAs)\"\n\n export NCCL_IB_GID_INDEX=${NCCL_IB_GID_INDEX:-$best_gid_index}\n\ [e2e-llm-inference-service] \ export NVSHMEM_IB_GID_INDEX=${NVSHMEM_IB_GID_INDEX:-$best_gid_index}\n\ [e2e-llm-inference-service] \ export UCX_IB_GID_INDEX=${UCX_IB_GID_INDEX:-$best_gid_index}\n\n\ [e2e-llm-inference-service] \ echo \"[Infer RoCE] Exported GID_INDEX=${best_gid_index} for NCCL,\ [e2e-llm-inference-service] \ NVSHMEM, and UCX\"\n else\n echo \"[Infer RoCE] ERROR: No valid\ [e2e-llm-inference-service] \ IPv4 ${KSERVE_INFER_IB_GID_INDEX_GREP} GID_INDEX found on any HCA.\"\n \ [e2e-llm-inference-service] \ fi\n else\n echo \"[Infer RoCE] No active HCAs found, skipping GID_INDEX\ [e2e-llm-inference-service] \ inference.\"\n fi\nfi\n\n# --disable-access-log-for-endpoints landed in vLLM\ [e2e-llm-inference-service] \ 0.16.0 (vllm-project/vllm#30011).\n# Older versions still need the blanket\ [e2e-llm-inference-service] \ --disable-uvicorn-access-log.\nACCESS_LOG_ARGS=\"--disable-uvicorn-access-log\"\ [e2e-llm-inference-service] \nVLLM_VERSION=$(vllm --version 2>/dev/null | tail -1 | awk '{print $NF}')\n\ [e2e-llm-inference-service] echo \"[access-log-detect] vllm version='${VLLM_VERSION}'\"\nif [[ \"$VLLM_VERSION\"\ [e2e-llm-inference-service] \ =~ ^[0-9]+\\.[0-9]+ ]] && [ \"$(printf '%s\\n%s\\n' \"0.16.0\" \"${VLLM_VERSION}\"\ [e2e-llm-inference-service] \ | sort -V | head -1)\" = \"0.16.0\" ]; then\n ACCESS_LOG_ARGS=\"--disable-access-log-for-endpoints\ [e2e-llm-inference-service] \ /health,/metrics,/ping\"\nfi\necho \"[access-log-detect] selected ACCESS_LOG_ARGS='${ACCESS_LOG_ARGS}'\"\ [e2e-llm-inference-service] \n\n# --shutdown-timeout landed in vLLM 0.18.0 (vllm-project/vllm#36666).\n\ [e2e-llm-inference-service] SHUTDOWN_TIMEOUT_ARGS=\"\"\nif [[ \"$VLLM_VERSION\" =~ ^[0-9]+\\.[0-9]+ ]] &&\ [e2e-llm-inference-service] \ [ \"$(printf '%s\\n%s\\n' \"0.18.0\" \"${VLLM_VERSION}\" | sort -V | head\ [e2e-llm-inference-service] \ -1)\" = \"0.18.0\" ]; then\n SHUTDOWN_TIMEOUT_ARGS=\"--shutdown-timeout 40\"\ [e2e-llm-inference-service] \nfi\n\neval \"exec vllm serve /mnt/models \\\n --served-model-name \"facebook/opt-125m\"\ [e2e-llm-inference-service] \ \"publishers/kserve-ci-e2e-test/models/facebook/opt-125m\" \\\n --port 8000\ [e2e-llm-inference-service] \ \\\n ${ACCESS_LOG_ARGS} \\\n ${SHUTDOWN_TIMEOUT_ARGS} \\\n --enable-ssl-refresh\ [e2e-llm-inference-service] \ \\\n --ssl-certfile /var/run/kserve/tls/tls.crt \\\n --ssl-keyfile /var/run/kserve/tls/tls.key\ [e2e-llm-inference-service] \ \\\n ${VLLM_ADDITIONAL_ARGS} \\\n $@\"" [e2e-llm-inference-service] - -- [e2e-llm-inference-service] ports: [e2e-llm-inference-service] - containerPort: 8000 [e2e-llm-inference-service] protocol: TCP [e2e-llm-inference-service] env: [e2e-llm-inference-service] - name: HOME [e2e-llm-inference-service] value: /home [e2e-llm-inference-service] - name: VLLM_LOGGING_LEVEL [e2e-llm-inference-service] value: DEBUG [e2e-llm-inference-service] - name: VLLM_CPU_KVCACHE_SPACE [e2e-llm-inference-service] value: '1' [e2e-llm-inference-service] - name: VLLM_ENABLE_V1_MULTIPROCESSING [e2e-llm-inference-service] value: '0' [e2e-llm-inference-service] - name: USER [e2e-llm-inference-service] value: nonroot [e2e-llm-inference-service] - name: TORCHINDUCTOR_CACHE_DIR [e2e-llm-inference-service] value: /tmp/torchinductor-cache [e2e-llm-inference-service] - name: HF_HUB_CACHE [e2e-llm-inference-service] value: /models [e2e-llm-inference-service] resources: [e2e-llm-inference-service] limits: [e2e-llm-inference-service] cpu: '2' [e2e-llm-inference-service] memory: 7Gi [e2e-llm-inference-service] requests: [e2e-llm-inference-service] cpu: 200m [e2e-llm-inference-service] memory: 2Gi [e2e-llm-inference-service] volumeMounts: [e2e-llm-inference-service] - name: home [e2e-llm-inference-service] mountPath: /home [e2e-llm-inference-service] - name: tmp-dir [e2e-llm-inference-service] mountPath: /tmp [e2e-llm-inference-service] - name: dshm [e2e-llm-inference-service] mountPath: /dev/shm [e2e-llm-inference-service] - name: model-cache [e2e-llm-inference-service] mountPath: /models [e2e-llm-inference-service] - name: tls-certs [e2e-llm-inference-service] readOnly: true [e2e-llm-inference-service] mountPath: /var/run/kserve/tls [e2e-llm-inference-service] - name: kserve-provision-location [e2e-llm-inference-service] readOnly: true [e2e-llm-inference-service] mountPath: /mnt/models [e2e-llm-inference-service] - name: kube-api-access-992bp [e2e-llm-inference-service] readOnly: true [e2e-llm-inference-service] mountPath: /var/run/secrets/kubernetes.io/serviceaccount [e2e-llm-inference-service] livenessProbe: [e2e-llm-inference-service] httpGet: [e2e-llm-inference-service] path: /health [e2e-llm-inference-service] port: 8000 [e2e-llm-inference-service] scheme: HTTPS [e2e-llm-inference-service] timeoutSeconds: 1 [e2e-llm-inference-service] periodSeconds: 10 [e2e-llm-inference-service] successThreshold: 1 [e2e-llm-inference-service] failureThreshold: 10 [e2e-llm-inference-service] readinessProbe: [e2e-llm-inference-service] httpGet: [e2e-llm-inference-service] path: /health [e2e-llm-inference-service] port: 8000 [e2e-llm-inference-service] scheme: HTTPS [e2e-llm-inference-service] timeoutSeconds: 1 [e2e-llm-inference-service] periodSeconds: 1 [e2e-llm-inference-service] successThreshold: 1 [e2e-llm-inference-service] failureThreshold: 2 [e2e-llm-inference-service] startupProbe: [e2e-llm-inference-service] httpGet: [e2e-llm-inference-service] path: /health [e2e-llm-inference-service] port: 8000 [e2e-llm-inference-service] scheme: HTTPS [e2e-llm-inference-service] timeoutSeconds: 1 [e2e-llm-inference-service] periodSeconds: 10 [e2e-llm-inference-service] successThreshold: 1 [e2e-llm-inference-service] failureThreshold: 60 [e2e-llm-inference-service] lifecycle: [e2e-llm-inference-service] preStop: [e2e-llm-inference-service] exec: [e2e-llm-inference-service] command: [e2e-llm-inference-service] - /bin/sleep [e2e-llm-inference-service] - '15' [e2e-llm-inference-service] terminationMessagePath: /dev/termination-log [e2e-llm-inference-service] terminationMessagePolicy: FallbackToLogsOnError [e2e-llm-inference-service] imagePullPolicy: IfNotPresent [e2e-llm-inference-service] securityContext: [e2e-llm-inference-service] capabilities: [e2e-llm-inference-service] drop: [e2e-llm-inference-service] - ALL [e2e-llm-inference-service] runAsUser: 1000690000 [e2e-llm-inference-service] runAsNonRoot: true [e2e-llm-inference-service] readOnlyRootFilesystem: true [e2e-llm-inference-service] allowPrivilegeEscalation: false [e2e-llm-inference-service] seccompProfile: [e2e-llm-inference-service] type: RuntimeDefault [e2e-llm-inference-service] restartPolicy: Always [e2e-llm-inference-service] terminationGracePeriodSeconds: 60 [e2e-llm-inference-service] dnsPolicy: ClusterFirst [e2e-llm-inference-service] serviceAccountName: default [e2e-llm-inference-service] serviceAccount: default [e2e-llm-inference-service] nodeName: ip-10-0-139-47.ec2.internal [e2e-llm-inference-service] securityContext: [e2e-llm-inference-service] seLinuxOptions: [e2e-llm-inference-service] level: s0:c26,c20 [e2e-llm-inference-service] fsGroup: 1000690000 [e2e-llm-inference-service] seccompProfile: [e2e-llm-inference-service] type: RuntimeDefault [e2e-llm-inference-service] imagePullSecrets: [e2e-llm-inference-service] - name: default-dockercfg-tbbgr [e2e-llm-inference-service] schedulerName: default-scheduler [e2e-llm-inference-service] tolerations: [e2e-llm-inference-service] - key: node.kubernetes.io/not-ready [e2e-llm-inference-service] operator: Exists [e2e-llm-inference-service] effect: NoExecute [e2e-llm-inference-service] tolerationSeconds: 300 [e2e-llm-inference-service] - key: node.kubernetes.io/unreachable [e2e-llm-inference-service] operator: Exists [e2e-llm-inference-service] effect: NoExecute [e2e-llm-inference-service] tolerationSeconds: 300 [e2e-llm-inference-service] - key: node.kubernetes.io/memory-pressure [e2e-llm-inference-service] operator: Exists [e2e-llm-inference-service] effect: NoSchedule [e2e-llm-inference-service] priority: 0 [e2e-llm-inference-service] enableServiceLinks: true [e2e-llm-inference-service] preemptionPolicy: PreemptLowerPriority [e2e-llm-inference-service] status: [e2e-llm-inference-service] observedGeneration: 1 [e2e-llm-inference-service] phase: Running [e2e-llm-inference-service] conditions: [e2e-llm-inference-service] - type: PodReadyToStartContainers [e2e-llm-inference-service] observedGeneration: 1 [e2e-llm-inference-service] status: 'True' [e2e-llm-inference-service] lastProbeTime: null [e2e-llm-inference-service] lastTransitionTime: '2026-07-07T14:48:58Z' [e2e-llm-inference-service] - type: Initialized [e2e-llm-inference-service] observedGeneration: 1 [e2e-llm-inference-service] status: 'True' [e2e-llm-inference-service] lastProbeTime: null [e2e-llm-inference-service] lastTransitionTime: '2026-07-07T14:49:02Z' [e2e-llm-inference-service] - type: Ready [e2e-llm-inference-service] observedGeneration: 1 [e2e-llm-inference-service] status: 'True' [e2e-llm-inference-service] lastProbeTime: null [e2e-llm-inference-service] lastTransitionTime: '2026-07-07T14:50:57Z' [e2e-llm-inference-service] - type: ContainersReady [e2e-llm-inference-service] observedGeneration: 1 [e2e-llm-inference-service] status: 'True' [e2e-llm-inference-service] lastProbeTime: null [e2e-llm-inference-service] lastTransitionTime: '2026-07-07T14:50:57Z' [e2e-llm-inference-service] - type: PodScheduled [e2e-llm-inference-service] observedGeneration: 1 [e2e-llm-inference-service] status: 'True' [e2e-llm-inference-service] lastProbeTime: null [e2e-llm-inference-service] lastTransitionTime: '2026-07-07T14:48:57Z' [e2e-llm-inference-service] hostIP: 10.0.139.47 [e2e-llm-inference-service] hostIPs: [e2e-llm-inference-service] - ip: 10.0.139.47 [e2e-llm-inference-service] podIP: 10.134.0.33 [e2e-llm-inference-service] podIPs: [e2e-llm-inference-service] - ip: 10.134.0.33 [e2e-llm-inference-service] startTime: '2026-07-07T14:48:57Z' [e2e-llm-inference-service] initContainerStatuses: [e2e-llm-inference-service] - name: storage-initializer [e2e-llm-inference-service] state: [e2e-llm-inference-service] terminated: [e2e-llm-inference-service] exitCode: 0 [e2e-llm-inference-service] reason: Completed [e2e-llm-inference-service] startedAt: '2026-07-07T14:48:57Z' [e2e-llm-inference-service] finishedAt: '2026-07-07T14:49:02Z' [e2e-llm-inference-service] containerID: cri-o://1bd82d0cbb8da9c9edb20b966e4d80b69f7d52c97fa6c0eb4c9fd003e464bd1b [e2e-llm-inference-service] lastState: {} [e2e-llm-inference-service] ready: true [e2e-llm-inference-service] restartCount: 0 [e2e-llm-inference-service] image: quay.io/opendatahub/kserve-storage-initializer@sha256:13ad765c0f323200e1db7b45c99661287e65c5158d6b54b16139829636d20be4 [e2e-llm-inference-service] imageID: quay.io/opendatahub/kserve-storage-initializer@sha256:13ad765c0f323200e1db7b45c99661287e65c5158d6b54b16139829636d20be4 [e2e-llm-inference-service] containerID: cri-o://1bd82d0cbb8da9c9edb20b966e4d80b69f7d52c97fa6c0eb4c9fd003e464bd1b [e2e-llm-inference-service] started: false [e2e-llm-inference-service] allocatedResources: [e2e-llm-inference-service] cpu: 100m [e2e-llm-inference-service] memory: 100Mi [e2e-llm-inference-service] resources: [e2e-llm-inference-service] limits: [e2e-llm-inference-service] cpu: '1' [e2e-llm-inference-service] memory: 24Gi [e2e-llm-inference-service] requests: [e2e-llm-inference-service] cpu: 100m [e2e-llm-inference-service] memory: 100Mi [e2e-llm-inference-service] volumeMounts: [e2e-llm-inference-service] - name: kserve-provision-location [e2e-llm-inference-service] mountPath: /mnt/models [e2e-llm-inference-service] - name: kube-api-access-992bp [e2e-llm-inference-service] mountPath: /var/run/secrets/kubernetes.io/serviceaccount [e2e-llm-inference-service] readOnly: true [e2e-llm-inference-service] recursiveReadOnly: Disabled [e2e-llm-inference-service] user: [e2e-llm-inference-service] linux: [e2e-llm-inference-service] uid: 1000690000 [e2e-llm-inference-service] gid: 0 [e2e-llm-inference-service] supplementalGroups: [e2e-llm-inference-service] - 0 [e2e-llm-inference-service] - 1000690000 [e2e-llm-inference-service] containerStatuses: [e2e-llm-inference-service] - name: main [e2e-llm-inference-service] state: [e2e-llm-inference-service] running: [e2e-llm-inference-service] startedAt: '2026-07-07T14:49:02Z' [e2e-llm-inference-service] lastState: {} [e2e-llm-inference-service] ready: true [e2e-llm-inference-service] restartCount: 0 [e2e-llm-inference-service] image: public.ecr.aws/q9t5s3a7/vllm-cpu-release-repo:v0.19.0 [e2e-llm-inference-service] imageID: public.ecr.aws/q9t5s3a7/vllm-cpu-release-repo@sha256:afb39fca138b51d019d986229d546531b45a2a3deb73bcf59bd42406e13fbba0 [e2e-llm-inference-service] containerID: cri-o://e2e13998b4e2594ec3fff7b075685b98273c313a102daf1204562350db36302a [e2e-llm-inference-service] started: true [e2e-llm-inference-service] allocatedResources: [e2e-llm-inference-service] cpu: 200m [e2e-llm-inference-service] memory: 2Gi [e2e-llm-inference-service] resources: [e2e-llm-inference-service] limits: [e2e-llm-inference-service] cpu: '2' [e2e-llm-inference-service] memory: 7Gi [e2e-llm-inference-service] requests: [e2e-llm-inference-service] cpu: 200m [e2e-llm-inference-service] memory: 2Gi [e2e-llm-inference-service] volumeMounts: [e2e-llm-inference-service] - name: home [e2e-llm-inference-service] mountPath: /home [e2e-llm-inference-service] - name: tmp-dir [e2e-llm-inference-service] mountPath: /tmp [e2e-llm-inference-service] - name: dshm [e2e-llm-inference-service] mountPath: /dev/shm [e2e-llm-inference-service] - name: model-cache [e2e-llm-inference-service] mountPath: /models [e2e-llm-inference-service] - name: tls-certs [e2e-llm-inference-service] mountPath: /var/run/kserve/tls [e2e-llm-inference-service] readOnly: true [e2e-llm-inference-service] recursiveReadOnly: Disabled [e2e-llm-inference-service] - name: kserve-provision-location [e2e-llm-inference-service] mountPath: /mnt/models [e2e-llm-inference-service] readOnly: true [e2e-llm-inference-service] recursiveReadOnly: Disabled [e2e-llm-inference-service] - name: kube-api-access-992bp [e2e-llm-inference-service] mountPath: /var/run/secrets/kubernetes.io/serviceaccount [e2e-llm-inference-service] readOnly: true [e2e-llm-inference-service] recursiveReadOnly: Disabled [e2e-llm-inference-service] user: [e2e-llm-inference-service] linux: [e2e-llm-inference-service] uid: 1000690000 [e2e-llm-inference-service] gid: 0 [e2e-llm-inference-service] supplementalGroups: [e2e-llm-inference-service] - 0 [e2e-llm-inference-service] - 1000690000 [e2e-llm-inference-service] qosClass: Burstable [e2e-llm-inference-service] apiVersion: v1 [e2e-llm-inference-service] kind: Pod [e2e-llm-inference-service] [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1254 --- [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1255 metadata: [e2e-llm-inference-service] name: custom-route-timeout-test-kserve-router-scheduler-6578bddfnbxnb [e2e-llm-inference-service] generateName: custom-route-timeout-test-kserve-router-scheduler-6578bddf5b- [e2e-llm-inference-service] namespace: kserve-ci-e2e-test [e2e-llm-inference-service] uid: f8a7664e-b142-4621-b773-b5e45ceaa80f [e2e-llm-inference-service] resourceVersion: '49588' [e2e-llm-inference-service] generation: 1 [e2e-llm-inference-service] creationTimestamp: '2026-07-07T14:48:57Z' [e2e-llm-inference-service] labels: [e2e-llm-inference-service] app.kubernetes.io/component: llminferenceservice-router-scheduler [e2e-llm-inference-service] app.kubernetes.io/name: custom-route-timeout-test [e2e-llm-inference-service] app.kubernetes.io/part-of: llminferenceservice [e2e-llm-inference-service] pod-template-hash: 6578bddf5b [e2e-llm-inference-service] annotations: [e2e-llm-inference-service] app.kubernetes.io/version: 0.9.0 [e2e-llm-inference-service] certificates.kserve.io/expiration-v2: 'true' [e2e-llm-inference-service] k8s.ovn.org/pod-networks: '{"default":{"ip_addresses":["10.132.0.55/23"],"mac_address":"0a:58:0a:84:00:37","gateway_ips":["10.132.0.1"],"routes":[{"dest":"10.132.0.0/14","nextHop":"10.132.0.1"},{"dest":"172.31.0.0/16","nextHop":"10.132.0.1"},{"dest":"169.254.0.5/32","nextHop":"10.132.0.1"},{"dest":"100.64.0.0/16","nextHop":"10.132.0.1"}],"ip_address":"10.132.0.55/23","gateway_ip":"10.132.0.1","role":"primary"}}' [e2e-llm-inference-service] k8s.v1.cni.cncf.io/network-status: "[{\n \"name\": \"ovn-kubernetes\",\n \ [e2e-llm-inference-service] \ \"interface\": \"eth0\",\n \"ips\": [\n \"10.132.0.55\"\n ],\n\ [e2e-llm-inference-service] \ \"mac\": \"0a:58:0a:84:00:37\",\n \"default\": true,\n \"dns\": {}\n\ [e2e-llm-inference-service] }]" [e2e-llm-inference-service] openshift.io/scc: restricted-v2 [e2e-llm-inference-service] seccomp.security.alpha.kubernetes.io/pod: runtime/default [e2e-llm-inference-service] security.openshift.io/validated-scc-subject-type: user [e2e-llm-inference-service] ownerReferences: [e2e-llm-inference-service] - apiVersion: apps/v1 [e2e-llm-inference-service] kind: ReplicaSet [e2e-llm-inference-service] name: custom-route-timeout-test-kserve-router-scheduler-6578bddf5b [e2e-llm-inference-service] uid: 92292d9f-7d6e-42e1-b2be-abdf1d622bd7 [e2e-llm-inference-service] controller: true [e2e-llm-inference-service] blockOwnerDeletion: true [e2e-llm-inference-service] managedFields: [e2e-llm-inference-service] - manager: ip-10-0-133-25 [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] apiVersion: v1 [e2e-llm-inference-service] time: '2026-07-07T14:48:57Z' [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:metadata: [e2e-llm-inference-service] f:annotations: [e2e-llm-inference-service] f:k8s.ovn.org/pod-networks: {} [e2e-llm-inference-service] subresource: status [e2e-llm-inference-service] - manager: kube-controller-manager [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] apiVersion: v1 [e2e-llm-inference-service] time: '2026-07-07T14:48:57Z' [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:metadata: [e2e-llm-inference-service] f:annotations: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:app.kubernetes.io/version: {} [e2e-llm-inference-service] f:certificates.kserve.io/expiration-v2: {} [e2e-llm-inference-service] f:generateName: {} [e2e-llm-inference-service] f:labels: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:app.kubernetes.io/component: {} [e2e-llm-inference-service] f:app.kubernetes.io/name: {} [e2e-llm-inference-service] f:app.kubernetes.io/part-of: {} [e2e-llm-inference-service] f:pod-template-hash: {} [e2e-llm-inference-service] f:ownerReferences: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"uid":"92292d9f-7d6e-42e1-b2be-abdf1d622bd7"}: {} [e2e-llm-inference-service] f:spec: [e2e-llm-inference-service] f:containers: [e2e-llm-inference-service] k:{"name":"main"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:args: {} [e2e-llm-inference-service] f:command: {} [e2e-llm-inference-service] f:env: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"name":"SSL_CERT_DIR"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] f:image: {} [e2e-llm-inference-service] f:imagePullPolicy: {} [e2e-llm-inference-service] f:lifecycle: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:preStop: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:exec: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:command: {} [e2e-llm-inference-service] f:livenessProbe: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:failureThreshold: {} [e2e-llm-inference-service] f:grpc: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:port: {} [e2e-llm-inference-service] f:service: {} [e2e-llm-inference-service] f:initialDelaySeconds: {} [e2e-llm-inference-service] f:periodSeconds: {} [e2e-llm-inference-service] f:successThreshold: {} [e2e-llm-inference-service] f:timeoutSeconds: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:ports: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"containerPort":5557,"protocol":"TCP"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:containerPort: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:protocol: {} [e2e-llm-inference-service] k:{"containerPort":9002,"protocol":"TCP"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:containerPort: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:protocol: {} [e2e-llm-inference-service] k:{"containerPort":9003,"protocol":"TCP"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:containerPort: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:protocol: {} [e2e-llm-inference-service] k:{"containerPort":9090,"protocol":"TCP"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:containerPort: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:protocol: {} [e2e-llm-inference-service] f:readinessProbe: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:failureThreshold: {} [e2e-llm-inference-service] f:grpc: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:port: {} [e2e-llm-inference-service] f:service: {} [e2e-llm-inference-service] f:initialDelaySeconds: {} [e2e-llm-inference-service] f:periodSeconds: {} [e2e-llm-inference-service] f:successThreshold: {} [e2e-llm-inference-service] f:timeoutSeconds: {} [e2e-llm-inference-service] f:resources: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:requests: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:cpu: {} [e2e-llm-inference-service] f:memory: {} [e2e-llm-inference-service] f:securityContext: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:allowPrivilegeEscalation: {} [e2e-llm-inference-service] f:capabilities: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:drop: {} [e2e-llm-inference-service] f:readOnlyRootFilesystem: {} [e2e-llm-inference-service] f:runAsNonRoot: {} [e2e-llm-inference-service] f:seccompProfile: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:type: {} [e2e-llm-inference-service] f:terminationMessagePath: {} [e2e-llm-inference-service] f:terminationMessagePolicy: {} [e2e-llm-inference-service] f:volumeMounts: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"mountPath":"/tmp/tokenizer"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:mountPath: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] k:{"mountPath":"/var/run/kserve/tls"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:mountPath: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:readOnly: {} [e2e-llm-inference-service] f:dnsPolicy: {} [e2e-llm-inference-service] f:enableServiceLinks: {} [e2e-llm-inference-service] f:initContainers: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"name":"storage-initializer"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:args: {} [e2e-llm-inference-service] f:env: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"name":"AWS_ACCESS_KEY_ID"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:valueFrom: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:secretKeyRef: {} [e2e-llm-inference-service] k:{"name":"AWS_CA_BUNDLE"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"AWS_CA_BUNDLE_CONFIGMAP"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"AWS_ENDPOINT_URL"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"AWS_SECRET_ACCESS_KEY"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:valueFrom: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:secretKeyRef: {} [e2e-llm-inference-service] k:{"name":"HF_HUB_ENABLE_HF_TRANSFER"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"HF_XET_HIGH_PERFORMANCE"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"HF_XET_NUM_CONCURRENT_RANGE_GETS"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"S3_ENDPOINT"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"S3_USE_HTTPS"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"S3_VERIFY_SSL"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"STORAGE_ALLOW_PATTERNS"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] f:image: {} [e2e-llm-inference-service] f:imagePullPolicy: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:resources: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:limits: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:cpu: {} [e2e-llm-inference-service] f:memory: {} [e2e-llm-inference-service] f:requests: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:cpu: {} [e2e-llm-inference-service] f:memory: {} [e2e-llm-inference-service] f:terminationMessagePath: {} [e2e-llm-inference-service] f:terminationMessagePolicy: {} [e2e-llm-inference-service] f:volumeMounts: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"mountPath":"/mnt/models"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:mountPath: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:restartPolicy: {} [e2e-llm-inference-service] f:schedulerName: {} [e2e-llm-inference-service] f:securityContext: {} [e2e-llm-inference-service] f:serviceAccount: {} [e2e-llm-inference-service] f:serviceAccountName: {} [e2e-llm-inference-service] f:terminationGracePeriodSeconds: {} [e2e-llm-inference-service] f:volumes: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"name":"kserve-provision-location"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:emptyDir: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] k:{"name":"tls-certs"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:secret: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:defaultMode: {} [e2e-llm-inference-service] f:secretName: {} [e2e-llm-inference-service] k:{"name":"tokenizer-cache"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:emptyDir: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] k:{"name":"tokenizer-tmp"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:emptyDir: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] k:{"name":"tokenizer-uds"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:emptyDir: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] - manager: multus-daemon [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] apiVersion: v1 [e2e-llm-inference-service] time: '2026-07-07T14:48:58Z' [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:metadata: [e2e-llm-inference-service] f:annotations: [e2e-llm-inference-service] f:k8s.v1.cni.cncf.io/network-status: {} [e2e-llm-inference-service] subresource: status [e2e-llm-inference-service] - manager: kubelet [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] apiVersion: v1 [e2e-llm-inference-service] time: '2026-07-07T14:49:31Z' [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:status: [e2e-llm-inference-service] f:conditions: [e2e-llm-inference-service] k:{"type":"ContainersReady"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:lastProbeTime: {} [e2e-llm-inference-service] f:lastTransitionTime: {} [e2e-llm-inference-service] f:observedGeneration: {} [e2e-llm-inference-service] f:status: {} [e2e-llm-inference-service] f:type: {} [e2e-llm-inference-service] k:{"type":"Initialized"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:lastProbeTime: {} [e2e-llm-inference-service] f:lastTransitionTime: {} [e2e-llm-inference-service] f:observedGeneration: {} [e2e-llm-inference-service] f:status: {} [e2e-llm-inference-service] f:type: {} [e2e-llm-inference-service] k:{"type":"PodReadyToStartContainers"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:lastProbeTime: {} [e2e-llm-inference-service] f:lastTransitionTime: {} [e2e-llm-inference-service] f:observedGeneration: {} [e2e-llm-inference-service] f:status: {} [e2e-llm-inference-service] f:type: {} [e2e-llm-inference-service] k:{"type":"PodScheduled"}: [e2e-llm-inference-service] f:observedGeneration: {} [e2e-llm-inference-service] k:{"type":"Ready"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:lastProbeTime: {} [e2e-llm-inference-service] f:lastTransitionTime: {} [e2e-llm-inference-service] f:observedGeneration: {} [e2e-llm-inference-service] f:status: {} [e2e-llm-inference-service] f:type: {} [e2e-llm-inference-service] f:containerStatuses: {} [e2e-llm-inference-service] f:hostIP: {} [e2e-llm-inference-service] f:hostIPs: {} [e2e-llm-inference-service] f:initContainerStatuses: {} [e2e-llm-inference-service] f:observedGeneration: {} [e2e-llm-inference-service] f:phase: {} [e2e-llm-inference-service] f:podIP: {} [e2e-llm-inference-service] f:podIPs: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"ip":"10.132.0.55"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:ip: {} [e2e-llm-inference-service] f:startTime: {} [e2e-llm-inference-service] subresource: status [e2e-llm-inference-service] spec: [e2e-llm-inference-service] volumes: [e2e-llm-inference-service] - name: tls-certs [e2e-llm-inference-service] secret: [e2e-llm-inference-service] secretName: custom-route-timeout-test-kserve-self-signed-certs [e2e-llm-inference-service] defaultMode: 420 [e2e-llm-inference-service] - name: tokenizer-uds [e2e-llm-inference-service] emptyDir: {} [e2e-llm-inference-service] - name: tokenizer-tmp [e2e-llm-inference-service] emptyDir: {} [e2e-llm-inference-service] - name: tokenizer-cache [e2e-llm-inference-service] emptyDir: {} [e2e-llm-inference-service] - name: kserve-provision-location [e2e-llm-inference-service] emptyDir: {} [e2e-llm-inference-service] - name: kube-api-access-qh2vb [e2e-llm-inference-service] projected: [e2e-llm-inference-service] sources: [e2e-llm-inference-service] - serviceAccountToken: [e2e-llm-inference-service] expirationSeconds: 3607 [e2e-llm-inference-service] path: token [e2e-llm-inference-service] - configMap: [e2e-llm-inference-service] name: kube-root-ca.crt [e2e-llm-inference-service] items: [e2e-llm-inference-service] - key: ca.crt [e2e-llm-inference-service] path: ca.crt [e2e-llm-inference-service] - downwardAPI: [e2e-llm-inference-service] items: [e2e-llm-inference-service] - path: namespace [e2e-llm-inference-service] fieldRef: [e2e-llm-inference-service] apiVersion: v1 [e2e-llm-inference-service] fieldPath: metadata.namespace [e2e-llm-inference-service] - configMap: [e2e-llm-inference-service] name: openshift-service-ca.crt [e2e-llm-inference-service] items: [e2e-llm-inference-service] - key: service-ca.crt [e2e-llm-inference-service] path: service-ca.crt [e2e-llm-inference-service] defaultMode: 420 [e2e-llm-inference-service] initContainers: [e2e-llm-inference-service] - name: storage-initializer [e2e-llm-inference-service] image: quay.io/opendatahub/kserve-storage-initializer@sha256:13ad765c0f323200e1db7b45c99661287e65c5158d6b54b16139829636d20be4 [e2e-llm-inference-service] args: [e2e-llm-inference-service] - hf://facebook/opt-125m [e2e-llm-inference-service] - /mnt/models [e2e-llm-inference-service] env: [e2e-llm-inference-service] - name: AWS_ACCESS_KEY_ID [e2e-llm-inference-service] valueFrom: [e2e-llm-inference-service] secretKeyRef: [e2e-llm-inference-service] name: seaweedfs-s3-creds [e2e-llm-inference-service] key: AWS_ACCESS_KEY_ID [e2e-llm-inference-service] - name: AWS_SECRET_ACCESS_KEY [e2e-llm-inference-service] valueFrom: [e2e-llm-inference-service] secretKeyRef: [e2e-llm-inference-service] name: seaweedfs-s3-creds [e2e-llm-inference-service] key: AWS_SECRET_ACCESS_KEY [e2e-llm-inference-service] - name: S3_USE_HTTPS [e2e-llm-inference-service] value: '0' [e2e-llm-inference-service] - name: S3_ENDPOINT [e2e-llm-inference-service] value: s3-service.kserve:8333 [e2e-llm-inference-service] - name: AWS_ENDPOINT_URL [e2e-llm-inference-service] value: http://s3-service.kserve:8333 [e2e-llm-inference-service] - name: S3_VERIFY_SSL [e2e-llm-inference-service] value: '0' [e2e-llm-inference-service] - name: AWS_CA_BUNDLE [e2e-llm-inference-service] value: /etc/ssl/custom-certs/cabundle.crt [e2e-llm-inference-service] - name: AWS_CA_BUNDLE_CONFIGMAP [e2e-llm-inference-service] value: odh-kserve-custom-ca-bundle [e2e-llm-inference-service] - name: HF_HUB_ENABLE_HF_TRANSFER [e2e-llm-inference-service] value: '1' [e2e-llm-inference-service] - name: HF_XET_HIGH_PERFORMANCE [e2e-llm-inference-service] value: '1' [e2e-llm-inference-service] - name: HF_XET_NUM_CONCURRENT_RANGE_GETS [e2e-llm-inference-service] value: '8' [e2e-llm-inference-service] - name: STORAGE_ALLOW_PATTERNS [e2e-llm-inference-service] value: '["tokenizer.json", "tokenizer_config.json", "special_tokens_map.json", [e2e-llm-inference-service] "vocab.json", "merges.txt", "config.json", "generation_config.json"]' [e2e-llm-inference-service] resources: [e2e-llm-inference-service] limits: [e2e-llm-inference-service] cpu: '1' [e2e-llm-inference-service] memory: 24Gi [e2e-llm-inference-service] requests: [e2e-llm-inference-service] cpu: 100m [e2e-llm-inference-service] memory: 100Mi [e2e-llm-inference-service] volumeMounts: [e2e-llm-inference-service] - name: kserve-provision-location [e2e-llm-inference-service] mountPath: /mnt/models [e2e-llm-inference-service] - name: kube-api-access-qh2vb [e2e-llm-inference-service] readOnly: true [e2e-llm-inference-service] mountPath: /var/run/secrets/kubernetes.io/serviceaccount [e2e-llm-inference-service] terminationMessagePath: /dev/termination-log [e2e-llm-inference-service] terminationMessagePolicy: FallbackToLogsOnError [e2e-llm-inference-service] imagePullPolicy: IfNotPresent [e2e-llm-inference-service] securityContext: [e2e-llm-inference-service] capabilities: [e2e-llm-inference-service] drop: [e2e-llm-inference-service] - ALL [e2e-llm-inference-service] runAsUser: 1000690000 [e2e-llm-inference-service] runAsNonRoot: true [e2e-llm-inference-service] allowPrivilegeEscalation: false [e2e-llm-inference-service] containers: [e2e-llm-inference-service] - name: main [e2e-llm-inference-service] image: ghcr.io/llm-d/llm-d-router-endpoint-picker:v0.9.0-rc.2 [e2e-llm-inference-service] command: [e2e-llm-inference-service] - /app/epp [e2e-llm-inference-service] - --pool-name [e2e-llm-inference-service] - custom-route-timeout-test-inference-pool [e2e-llm-inference-service] - --pool-namespace [e2e-llm-inference-service] - kserve-ci-e2e-test [e2e-llm-inference-service] - --zap-encoder [e2e-llm-inference-service] - json [e2e-llm-inference-service] - --grpc-port [e2e-llm-inference-service] - '9002' [e2e-llm-inference-service] - --grpc-health-port [e2e-llm-inference-service] - '9003' [e2e-llm-inference-service] - --enable-cert-reload=true [e2e-llm-inference-service] - --secure-serving=true [e2e-llm-inference-service] - --model-server-metrics-scheme=https [e2e-llm-inference-service] - --cert-path=/var/run/kserve/tls [e2e-llm-inference-service] args: [e2e-llm-inference-service] - --config-text [e2e-llm-inference-service] - "apiVersion: inference.networking.x-k8s.io/v1alpha1\nkind: EndpointPickerConfig\n\ [e2e-llm-inference-service] plugins:\n- type: single-profile-handler\n- type: queue-scorer\n- type: prefix-cache-scorer\n\ [e2e-llm-inference-service] - type: max-score-picker\nschedulingProfiles:\n- name: default\n plugins:\n\ [e2e-llm-inference-service] \ - pluginRef: queue-scorer\n weight: 2\n - pluginRef: prefix-cache-scorer\n\ [e2e-llm-inference-service] \ weight: 3\n - pluginRef: max-score-picker\n" [e2e-llm-inference-service] ports: [e2e-llm-inference-service] - name: grpc [e2e-llm-inference-service] containerPort: 9002 [e2e-llm-inference-service] protocol: TCP [e2e-llm-inference-service] - name: grpc-health [e2e-llm-inference-service] containerPort: 9003 [e2e-llm-inference-service] protocol: TCP [e2e-llm-inference-service] - name: metrics [e2e-llm-inference-service] containerPort: 9090 [e2e-llm-inference-service] protocol: TCP [e2e-llm-inference-service] - name: zmq [e2e-llm-inference-service] containerPort: 5557 [e2e-llm-inference-service] protocol: TCP [e2e-llm-inference-service] env: [e2e-llm-inference-service] - name: SSL_CERT_DIR [e2e-llm-inference-service] value: /var/run/kserve/tls:/var/run/secrets/kubernetes.io/serviceaccount:/etc/pki/tls/certs [e2e-llm-inference-service] resources: [e2e-llm-inference-service] requests: [e2e-llm-inference-service] cpu: 256m [e2e-llm-inference-service] memory: 500Mi [e2e-llm-inference-service] volumeMounts: [e2e-llm-inference-service] - name: tls-certs [e2e-llm-inference-service] readOnly: true [e2e-llm-inference-service] mountPath: /var/run/kserve/tls [e2e-llm-inference-service] - name: tokenizer-uds [e2e-llm-inference-service] mountPath: /tmp/tokenizer [e2e-llm-inference-service] - name: kube-api-access-qh2vb [e2e-llm-inference-service] readOnly: true [e2e-llm-inference-service] mountPath: /var/run/secrets/kubernetes.io/serviceaccount [e2e-llm-inference-service] livenessProbe: [e2e-llm-inference-service] grpc: [e2e-llm-inference-service] port: 9003 [e2e-llm-inference-service] service: liveness [e2e-llm-inference-service] initialDelaySeconds: 5 [e2e-llm-inference-service] timeoutSeconds: 1 [e2e-llm-inference-service] periodSeconds: 10 [e2e-llm-inference-service] successThreshold: 1 [e2e-llm-inference-service] failureThreshold: 3 [e2e-llm-inference-service] readinessProbe: [e2e-llm-inference-service] grpc: [e2e-llm-inference-service] port: 9003 [e2e-llm-inference-service] service: readiness [e2e-llm-inference-service] initialDelaySeconds: 30 [e2e-llm-inference-service] timeoutSeconds: 1 [e2e-llm-inference-service] periodSeconds: 10 [e2e-llm-inference-service] successThreshold: 1 [e2e-llm-inference-service] failureThreshold: 3 [e2e-llm-inference-service] lifecycle: [e2e-llm-inference-service] preStop: [e2e-llm-inference-service] exec: [e2e-llm-inference-service] command: [e2e-llm-inference-service] - /bin/sleep [e2e-llm-inference-service] - '15' [e2e-llm-inference-service] terminationMessagePath: /dev/termination-log [e2e-llm-inference-service] terminationMessagePolicy: FallbackToLogsOnError [e2e-llm-inference-service] imagePullPolicy: IfNotPresent [e2e-llm-inference-service] securityContext: [e2e-llm-inference-service] capabilities: [e2e-llm-inference-service] drop: [e2e-llm-inference-service] - ALL [e2e-llm-inference-service] runAsUser: 1000690000 [e2e-llm-inference-service] runAsNonRoot: true [e2e-llm-inference-service] readOnlyRootFilesystem: true [e2e-llm-inference-service] allowPrivilegeEscalation: false [e2e-llm-inference-service] seccompProfile: [e2e-llm-inference-service] type: RuntimeDefault [e2e-llm-inference-service] restartPolicy: Always [e2e-llm-inference-service] terminationGracePeriodSeconds: 60 [e2e-llm-inference-service] dnsPolicy: ClusterFirst [e2e-llm-inference-service] serviceAccountName: custom-route-timeout-test-epp-sa [e2e-llm-inference-service] serviceAccount: custom-route-timeout-test-epp-sa [e2e-llm-inference-service] nodeName: ip-10-0-133-25.ec2.internal [e2e-llm-inference-service] securityContext: [e2e-llm-inference-service] seLinuxOptions: [e2e-llm-inference-service] level: s0:c26,c20 [e2e-llm-inference-service] fsGroup: 1000690000 [e2e-llm-inference-service] seccompProfile: [e2e-llm-inference-service] type: RuntimeDefault [e2e-llm-inference-service] imagePullSecrets: [e2e-llm-inference-service] - name: default-dockercfg-tbbgr [e2e-llm-inference-service] - name: custom-route-timeout-test-epp-sa-dockercfg-pk5jw [e2e-llm-inference-service] schedulerName: default-scheduler [e2e-llm-inference-service] tolerations: [e2e-llm-inference-service] - key: node.kubernetes.io/not-ready [e2e-llm-inference-service] operator: Exists [e2e-llm-inference-service] effect: NoExecute [e2e-llm-inference-service] tolerationSeconds: 300 [e2e-llm-inference-service] - key: node.kubernetes.io/unreachable [e2e-llm-inference-service] operator: Exists [e2e-llm-inference-service] effect: NoExecute [e2e-llm-inference-service] tolerationSeconds: 300 [e2e-llm-inference-service] - key: node.kubernetes.io/memory-pressure [e2e-llm-inference-service] operator: Exists [e2e-llm-inference-service] effect: NoSchedule [e2e-llm-inference-service] priority: 0 [e2e-llm-inference-service] enableServiceLinks: true [e2e-llm-inference-service] preemptionPolicy: PreemptLowerPriority [e2e-llm-inference-service] status: [e2e-llm-inference-service] observedGeneration: 1 [e2e-llm-inference-service] phase: Running [e2e-llm-inference-service] conditions: [e2e-llm-inference-service] - type: PodReadyToStartContainers [e2e-llm-inference-service] observedGeneration: 1 [e2e-llm-inference-service] status: 'True' [e2e-llm-inference-service] lastProbeTime: null [e2e-llm-inference-service] lastTransitionTime: '2026-07-07T14:48:58Z' [e2e-llm-inference-service] - type: Initialized [e2e-llm-inference-service] observedGeneration: 1 [e2e-llm-inference-service] status: 'True' [e2e-llm-inference-service] lastProbeTime: null [e2e-llm-inference-service] lastTransitionTime: '2026-07-07T14:48:59Z' [e2e-llm-inference-service] - type: Ready [e2e-llm-inference-service] observedGeneration: 1 [e2e-llm-inference-service] status: 'True' [e2e-llm-inference-service] lastProbeTime: null [e2e-llm-inference-service] lastTransitionTime: '2026-07-07T14:49:31Z' [e2e-llm-inference-service] - type: ContainersReady [e2e-llm-inference-service] observedGeneration: 1 [e2e-llm-inference-service] status: 'True' [e2e-llm-inference-service] lastProbeTime: null [e2e-llm-inference-service] lastTransitionTime: '2026-07-07T14:49:31Z' [e2e-llm-inference-service] - type: PodScheduled [e2e-llm-inference-service] observedGeneration: 1 [e2e-llm-inference-service] status: 'True' [e2e-llm-inference-service] lastProbeTime: null [e2e-llm-inference-service] lastTransitionTime: '2026-07-07T14:48:57Z' [e2e-llm-inference-service] hostIP: 10.0.133.25 [e2e-llm-inference-service] hostIPs: [e2e-llm-inference-service] - ip: 10.0.133.25 [e2e-llm-inference-service] podIP: 10.132.0.55 [e2e-llm-inference-service] podIPs: [e2e-llm-inference-service] - ip: 10.132.0.55 [e2e-llm-inference-service] startTime: '2026-07-07T14:48:57Z' [e2e-llm-inference-service] initContainerStatuses: [e2e-llm-inference-service] - name: storage-initializer [e2e-llm-inference-service] state: [e2e-llm-inference-service] terminated: [e2e-llm-inference-service] exitCode: 0 [e2e-llm-inference-service] reason: Completed [e2e-llm-inference-service] startedAt: '2026-07-07T14:48:58Z' [e2e-llm-inference-service] finishedAt: '2026-07-07T14:48:59Z' [e2e-llm-inference-service] containerID: cri-o://cae24cc14faecee342f0e220ba341d79d3fcae8e9c3f4702391497aafdb50958 [e2e-llm-inference-service] lastState: {} [e2e-llm-inference-service] ready: true [e2e-llm-inference-service] restartCount: 0 [e2e-llm-inference-service] image: quay.io/opendatahub/kserve-storage-initializer@sha256:13ad765c0f323200e1db7b45c99661287e65c5158d6b54b16139829636d20be4 [e2e-llm-inference-service] imageID: quay.io/opendatahub/kserve-storage-initializer@sha256:13ad765c0f323200e1db7b45c99661287e65c5158d6b54b16139829636d20be4 [e2e-llm-inference-service] containerID: cri-o://cae24cc14faecee342f0e220ba341d79d3fcae8e9c3f4702391497aafdb50958 [e2e-llm-inference-service] started: false [e2e-llm-inference-service] allocatedResources: [e2e-llm-inference-service] cpu: 100m [e2e-llm-inference-service] memory: 100Mi [e2e-llm-inference-service] resources: [e2e-llm-inference-service] limits: [e2e-llm-inference-service] cpu: '1' [e2e-llm-inference-service] memory: 24Gi [e2e-llm-inference-service] requests: [e2e-llm-inference-service] cpu: 100m [e2e-llm-inference-service] memory: 100Mi [e2e-llm-inference-service] volumeMounts: [e2e-llm-inference-service] - name: kserve-provision-location [e2e-llm-inference-service] mountPath: /mnt/models [e2e-llm-inference-service] - name: kube-api-access-qh2vb [e2e-llm-inference-service] mountPath: /var/run/secrets/kubernetes.io/serviceaccount [e2e-llm-inference-service] readOnly: true [e2e-llm-inference-service] recursiveReadOnly: Disabled [e2e-llm-inference-service] user: [e2e-llm-inference-service] linux: [e2e-llm-inference-service] uid: 1000690000 [e2e-llm-inference-service] gid: 0 [e2e-llm-inference-service] supplementalGroups: [e2e-llm-inference-service] - 0 [e2e-llm-inference-service] - 1000690000 [e2e-llm-inference-service] containerStatuses: [e2e-llm-inference-service] - name: main [e2e-llm-inference-service] state: [e2e-llm-inference-service] running: [e2e-llm-inference-service] startedAt: '2026-07-07T14:48:59Z' [e2e-llm-inference-service] lastState: {} [e2e-llm-inference-service] ready: true [e2e-llm-inference-service] restartCount: 0 [e2e-llm-inference-service] image: ghcr.io/llm-d/llm-d-router-endpoint-picker:v0.9.0-rc.2 [e2e-llm-inference-service] imageID: ghcr.io/llm-d/llm-d-router-endpoint-picker@sha256:06b6c75d77afd0e07053402752a9736c2dfbc12a306d0d37d963aac4c1d4e6a6 [e2e-llm-inference-service] containerID: cri-o://b24f3dc48b9a05fe5b9513300d60dc3ecc59dcc576bbfc326324cb6ea9f3733e [e2e-llm-inference-service] started: true [e2e-llm-inference-service] allocatedResources: [e2e-llm-inference-service] cpu: 256m [e2e-llm-inference-service] memory: 500Mi [e2e-llm-inference-service] resources: [e2e-llm-inference-service] requests: [e2e-llm-inference-service] cpu: 256m [e2e-llm-inference-service] memory: 500Mi [e2e-llm-inference-service] volumeMounts: [e2e-llm-inference-service] - name: tls-certs [e2e-llm-inference-service] mountPath: /var/run/kserve/tls [e2e-llm-inference-service] readOnly: true [e2e-llm-inference-service] recursiveReadOnly: Disabled [e2e-llm-inference-service] - name: tokenizer-uds [e2e-llm-inference-service] mountPath: /tmp/tokenizer [e2e-llm-inference-service] - name: kube-api-access-qh2vb [e2e-llm-inference-service] mountPath: /var/run/secrets/kubernetes.io/serviceaccount [e2e-llm-inference-service] readOnly: true [e2e-llm-inference-service] recursiveReadOnly: Disabled [e2e-llm-inference-service] user: [e2e-llm-inference-service] linux: [e2e-llm-inference-service] uid: 1000690000 [e2e-llm-inference-service] gid: 0 [e2e-llm-inference-service] supplementalGroups: [e2e-llm-inference-service] - 0 [e2e-llm-inference-service] - 1000690000 [e2e-llm-inference-service] qosClass: Burstable [e2e-llm-inference-service] apiVersion: v1 [e2e-llm-inference-service] kind: Pod [e2e-llm-inference-service] [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1254 --- [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1255 metadata: [e2e-llm-inference-service] name: custom-route-timeout-test-epp-sa [e2e-llm-inference-service] namespace: kserve-ci-e2e-test [e2e-llm-inference-service] uid: 47b7ebae-2bcf-40b5-992f-769a5d5390a2 [e2e-llm-inference-service] resourceVersion: '48972' [e2e-llm-inference-service] creationTimestamp: '2026-07-07T14:48:57Z' [e2e-llm-inference-service] labels: [e2e-llm-inference-service] app.kubernetes.io/component: llminferenceservice-router-scheduler [e2e-llm-inference-service] app.kubernetes.io/name: custom-route-timeout-test [e2e-llm-inference-service] app.kubernetes.io/part-of: llminferenceservice [e2e-llm-inference-service] annotations: [e2e-llm-inference-service] openshift.io/internal-registry-pull-secret-ref: custom-route-timeout-test-epp-sa-dockercfg-pk5jw [e2e-llm-inference-service] ownerReferences: [e2e-llm-inference-service] - apiVersion: serving.kserve.io/v1alpha2 [e2e-llm-inference-service] kind: LLMInferenceService [e2e-llm-inference-service] name: custom-route-timeout-test [e2e-llm-inference-service] uid: 986f2d2c-a6d1-463d-9850-3aad6459eace [e2e-llm-inference-service] controller: true [e2e-llm-inference-service] blockOwnerDeletion: true [e2e-llm-inference-service] managedFields: [e2e-llm-inference-service] - manager: openshift.io/image-registry-pull-secrets_service-account-controller [e2e-llm-inference-service] operation: Apply [e2e-llm-inference-service] apiVersion: v1 [e2e-llm-inference-service] time: '2026-07-07T14:48:57Z' [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:imagePullSecrets: {} [e2e-llm-inference-service] f:metadata: [e2e-llm-inference-service] f:annotations: [e2e-llm-inference-service] f:openshift.io/internal-registry-pull-secret-ref: {} [e2e-llm-inference-service] f:secrets: [e2e-llm-inference-service] k:{"name":"custom-route-timeout-test-epp-sa-dockercfg-pk5jw"}: {} [e2e-llm-inference-service] - manager: manager [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] apiVersion: v1 [e2e-llm-inference-service] time: '2026-07-07T14:48:57Z' [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:metadata: [e2e-llm-inference-service] f:labels: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:app.kubernetes.io/component: {} [e2e-llm-inference-service] f:app.kubernetes.io/name: {} [e2e-llm-inference-service] f:app.kubernetes.io/part-of: {} [e2e-llm-inference-service] f:ownerReferences: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"uid":"986f2d2c-a6d1-463d-9850-3aad6459eace"}: {} [e2e-llm-inference-service] f:secrets: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"name":"default-dockercfg-tbbgr"}: {} [e2e-llm-inference-service] k:{"name":"seaweedfs-s3-creds"}: {} [e2e-llm-inference-service] secrets: [e2e-llm-inference-service] - name: default-dockercfg-tbbgr [e2e-llm-inference-service] - name: seaweedfs-s3-creds [e2e-llm-inference-service] - name: custom-route-timeout-test-epp-sa-dockercfg-pk5jw [e2e-llm-inference-service] imagePullSecrets: [e2e-llm-inference-service] - name: default-dockercfg-tbbgr [e2e-llm-inference-service] - name: custom-route-timeout-test-epp-sa-dockercfg-pk5jw [e2e-llm-inference-service] apiVersion: v1 [e2e-llm-inference-service] kind: ServiceAccount [e2e-llm-inference-service] [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1254 --- [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1255 metadata: [e2e-llm-inference-service] name: custom-route-timeout-test-epp-service [e2e-llm-inference-service] namespace: kserve-ci-e2e-test [e2e-llm-inference-service] uid: 38f4b9e5-98c2-4938-a86e-9aec7ba9c9ea [e2e-llm-inference-service] resourceVersion: '48995' [e2e-llm-inference-service] creationTimestamp: '2026-07-07T14:48:57Z' [e2e-llm-inference-service] labels: [e2e-llm-inference-service] app.kubernetes.io/component: llminferenceservice-router-scheduler [e2e-llm-inference-service] app.kubernetes.io/name: custom-route-timeout-test [e2e-llm-inference-service] app.kubernetes.io/part-of: llminferenceservice [e2e-llm-inference-service] ownerReferences: [e2e-llm-inference-service] - apiVersion: serving.kserve.io/v1alpha2 [e2e-llm-inference-service] kind: LLMInferenceService [e2e-llm-inference-service] name: custom-route-timeout-test [e2e-llm-inference-service] uid: 986f2d2c-a6d1-463d-9850-3aad6459eace [e2e-llm-inference-service] controller: true [e2e-llm-inference-service] blockOwnerDeletion: true [e2e-llm-inference-service] managedFields: [e2e-llm-inference-service] - manager: manager [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] apiVersion: v1 [e2e-llm-inference-service] time: '2026-07-07T14:48:57Z' [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:metadata: [e2e-llm-inference-service] f:labels: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:app.kubernetes.io/component: {} [e2e-llm-inference-service] f:app.kubernetes.io/name: {} [e2e-llm-inference-service] f:app.kubernetes.io/part-of: {} [e2e-llm-inference-service] f:ownerReferences: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"uid":"986f2d2c-a6d1-463d-9850-3aad6459eace"}: {} [e2e-llm-inference-service] f:spec: [e2e-llm-inference-service] f:internalTrafficPolicy: {} [e2e-llm-inference-service] f:ports: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"port":5557,"protocol":"TCP"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:port: {} [e2e-llm-inference-service] f:protocol: {} [e2e-llm-inference-service] f:targetPort: {} [e2e-llm-inference-service] k:{"port":9002,"protocol":"TCP"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:port: {} [e2e-llm-inference-service] f:protocol: {} [e2e-llm-inference-service] f:targetPort: {} [e2e-llm-inference-service] k:{"port":9003,"protocol":"TCP"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:port: {} [e2e-llm-inference-service] f:protocol: {} [e2e-llm-inference-service] f:targetPort: {} [e2e-llm-inference-service] k:{"port":9090,"protocol":"TCP"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:port: {} [e2e-llm-inference-service] f:protocol: {} [e2e-llm-inference-service] f:targetPort: {} [e2e-llm-inference-service] f:selector: {} [e2e-llm-inference-service] f:sessionAffinity: {} [e2e-llm-inference-service] f:type: {} [e2e-llm-inference-service] spec: [e2e-llm-inference-service] ports: [e2e-llm-inference-service] - name: grpc [e2e-llm-inference-service] protocol: TCP [e2e-llm-inference-service] port: 9002 [e2e-llm-inference-service] targetPort: grpc [e2e-llm-inference-service] - name: grpc-health [e2e-llm-inference-service] protocol: TCP [e2e-llm-inference-service] port: 9003 [e2e-llm-inference-service] targetPort: grpc-health [e2e-llm-inference-service] - name: metrics [e2e-llm-inference-service] protocol: TCP [e2e-llm-inference-service] port: 9090 [e2e-llm-inference-service] targetPort: metrics [e2e-llm-inference-service] - name: zmq [e2e-llm-inference-service] protocol: TCP [e2e-llm-inference-service] port: 5557 [e2e-llm-inference-service] targetPort: zmq [e2e-llm-inference-service] selector: [e2e-llm-inference-service] app.kubernetes.io/component: llminferenceservice-router-scheduler [e2e-llm-inference-service] app.kubernetes.io/name: custom-route-timeout-test [e2e-llm-inference-service] app.kubernetes.io/part-of: llminferenceservice [e2e-llm-inference-service] clusterIP: 172.31.210.241 [e2e-llm-inference-service] clusterIPs: [e2e-llm-inference-service] - 172.31.210.241 [e2e-llm-inference-service] type: ClusterIP [e2e-llm-inference-service] sessionAffinity: None [e2e-llm-inference-service] ipFamilies: [e2e-llm-inference-service] - IPv4 [e2e-llm-inference-service] ipFamilyPolicy: SingleStack [e2e-llm-inference-service] internalTrafficPolicy: Cluster [e2e-llm-inference-service] status: [e2e-llm-inference-service] loadBalancer: {} [e2e-llm-inference-service] apiVersion: v1 [e2e-llm-inference-service] kind: Service [e2e-llm-inference-service] [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1254 --- [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1255 metadata: [e2e-llm-inference-service] name: custom-route-timeout-test-kserve-workload-svc [e2e-llm-inference-service] namespace: kserve-ci-e2e-test [e2e-llm-inference-service] uid: 8a311167-fdf9-4de8-b36a-8990fdd59c71 [e2e-llm-inference-service] resourceVersion: '48953' [e2e-llm-inference-service] creationTimestamp: '2026-07-07T14:48:57Z' [e2e-llm-inference-service] labels: [e2e-llm-inference-service] app.kubernetes.io/component: llminferenceservice-workload [e2e-llm-inference-service] app.kubernetes.io/name: custom-route-timeout-test [e2e-llm-inference-service] app.kubernetes.io/part-of: llminferenceservice [e2e-llm-inference-service] ownerReferences: [e2e-llm-inference-service] - apiVersion: serving.kserve.io/v1alpha2 [e2e-llm-inference-service] kind: LLMInferenceService [e2e-llm-inference-service] name: custom-route-timeout-test [e2e-llm-inference-service] uid: 986f2d2c-a6d1-463d-9850-3aad6459eace [e2e-llm-inference-service] controller: true [e2e-llm-inference-service] blockOwnerDeletion: true [e2e-llm-inference-service] managedFields: [e2e-llm-inference-service] - manager: manager [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] apiVersion: v1 [e2e-llm-inference-service] time: '2026-07-07T14:48:57Z' [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:metadata: [e2e-llm-inference-service] f:labels: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:app.kubernetes.io/component: {} [e2e-llm-inference-service] f:app.kubernetes.io/name: {} [e2e-llm-inference-service] f:app.kubernetes.io/part-of: {} [e2e-llm-inference-service] f:ownerReferences: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"uid":"986f2d2c-a6d1-463d-9850-3aad6459eace"}: {} [e2e-llm-inference-service] f:spec: [e2e-llm-inference-service] f:internalTrafficPolicy: {} [e2e-llm-inference-service] f:ports: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"port":8000,"protocol":"TCP"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:appProtocol: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:port: {} [e2e-llm-inference-service] f:protocol: {} [e2e-llm-inference-service] f:targetPort: {} [e2e-llm-inference-service] f:selector: {} [e2e-llm-inference-service] f:sessionAffinity: {} [e2e-llm-inference-service] f:type: {} [e2e-llm-inference-service] spec: [e2e-llm-inference-service] ports: [e2e-llm-inference-service] - name: https [e2e-llm-inference-service] protocol: TCP [e2e-llm-inference-service] appProtocol: https [e2e-llm-inference-service] port: 8000 [e2e-llm-inference-service] targetPort: 8000 [e2e-llm-inference-service] selector: [e2e-llm-inference-service] app.kubernetes.io/name: custom-route-timeout-test [e2e-llm-inference-service] app.kubernetes.io/part-of: llminferenceservice [e2e-llm-inference-service] kserve.io/component: workload [e2e-llm-inference-service] clusterIP: 172.31.108.176 [e2e-llm-inference-service] clusterIPs: [e2e-llm-inference-service] - 172.31.108.176 [e2e-llm-inference-service] type: ClusterIP [e2e-llm-inference-service] sessionAffinity: None [e2e-llm-inference-service] ipFamilies: [e2e-llm-inference-service] - IPv4 [e2e-llm-inference-service] ipFamilyPolicy: SingleStack [e2e-llm-inference-service] internalTrafficPolicy: Cluster [e2e-llm-inference-service] status: [e2e-llm-inference-service] loadBalancer: {} [e2e-llm-inference-service] apiVersion: v1 [e2e-llm-inference-service] kind: Service [e2e-llm-inference-service] [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1254 --- [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1255 metadata: [e2e-llm-inference-service] name: custom-route-timeout-test-kserve [e2e-llm-inference-service] namespace: kserve-ci-e2e-test [e2e-llm-inference-service] uid: e9876375-48ff-4fa2-88dd-23a9c4ed5714 [e2e-llm-inference-service] resourceVersion: '50899' [e2e-llm-inference-service] generation: 1 [e2e-llm-inference-service] creationTimestamp: '2026-07-07T14:48:57Z' [e2e-llm-inference-service] labels: [e2e-llm-inference-service] app.kubernetes.io/component: llminferenceservice-workload [e2e-llm-inference-service] app.kubernetes.io/name: custom-route-timeout-test [e2e-llm-inference-service] app.kubernetes.io/part-of: llminferenceservice [e2e-llm-inference-service] kserve.io/component: workload [e2e-llm-inference-service] llm-d.ai/role: both [e2e-llm-inference-service] annotations: [e2e-llm-inference-service] deployment.kubernetes.io/revision: '1' [e2e-llm-inference-service] ownerReferences: [e2e-llm-inference-service] - apiVersion: serving.kserve.io/v1alpha2 [e2e-llm-inference-service] kind: LLMInferenceService [e2e-llm-inference-service] name: custom-route-timeout-test [e2e-llm-inference-service] uid: 986f2d2c-a6d1-463d-9850-3aad6459eace [e2e-llm-inference-service] controller: true [e2e-llm-inference-service] blockOwnerDeletion: true [e2e-llm-inference-service] managedFields: [e2e-llm-inference-service] - manager: manager [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] apiVersion: apps/v1 [e2e-llm-inference-service] time: '2026-07-07T14:48:57Z' [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:metadata: [e2e-llm-inference-service] f:labels: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:app.kubernetes.io/component: {} [e2e-llm-inference-service] f:app.kubernetes.io/name: {} [e2e-llm-inference-service] f:app.kubernetes.io/part-of: {} [e2e-llm-inference-service] f:kserve.io/component: {} [e2e-llm-inference-service] f:llm-d.ai/role: {} [e2e-llm-inference-service] f:ownerReferences: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"uid":"986f2d2c-a6d1-463d-9850-3aad6459eace"}: {} [e2e-llm-inference-service] f:spec: [e2e-llm-inference-service] f:progressDeadlineSeconds: {} [e2e-llm-inference-service] f:replicas: {} [e2e-llm-inference-service] f:revisionHistoryLimit: {} [e2e-llm-inference-service] f:selector: {} [e2e-llm-inference-service] f:strategy: [e2e-llm-inference-service] f:rollingUpdate: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:maxSurge: {} [e2e-llm-inference-service] f:maxUnavailable: {} [e2e-llm-inference-service] f:type: {} [e2e-llm-inference-service] f:template: [e2e-llm-inference-service] f:metadata: [e2e-llm-inference-service] f:labels: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:app.kubernetes.io/component: {} [e2e-llm-inference-service] f:app.kubernetes.io/name: {} [e2e-llm-inference-service] f:app.kubernetes.io/part-of: {} [e2e-llm-inference-service] f:kserve.io/component: {} [e2e-llm-inference-service] f:llm-d.ai/role: {} [e2e-llm-inference-service] f:spec: [e2e-llm-inference-service] f:containers: [e2e-llm-inference-service] k:{"name":"main"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:command: {} [e2e-llm-inference-service] f:env: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"name":"HF_HUB_CACHE"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"HOME"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"TORCHINDUCTOR_CACHE_DIR"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"USER"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"VLLM_CPU_KVCACHE_SPACE"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"VLLM_ENABLE_V1_MULTIPROCESSING"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"VLLM_LOGGING_LEVEL"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] f:image: {} [e2e-llm-inference-service] f:imagePullPolicy: {} [e2e-llm-inference-service] f:lifecycle: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:preStop: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:exec: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:command: {} [e2e-llm-inference-service] f:livenessProbe: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:failureThreshold: {} [e2e-llm-inference-service] f:httpGet: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:path: {} [e2e-llm-inference-service] f:port: {} [e2e-llm-inference-service] f:scheme: {} [e2e-llm-inference-service] f:periodSeconds: {} [e2e-llm-inference-service] f:successThreshold: {} [e2e-llm-inference-service] f:timeoutSeconds: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:ports: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"containerPort":8000,"protocol":"TCP"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:containerPort: {} [e2e-llm-inference-service] f:protocol: {} [e2e-llm-inference-service] f:readinessProbe: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:failureThreshold: {} [e2e-llm-inference-service] f:httpGet: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:path: {} [e2e-llm-inference-service] f:port: {} [e2e-llm-inference-service] f:scheme: {} [e2e-llm-inference-service] f:periodSeconds: {} [e2e-llm-inference-service] f:successThreshold: {} [e2e-llm-inference-service] f:timeoutSeconds: {} [e2e-llm-inference-service] f:resources: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:limits: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:cpu: {} [e2e-llm-inference-service] f:memory: {} [e2e-llm-inference-service] f:requests: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:cpu: {} [e2e-llm-inference-service] f:memory: {} [e2e-llm-inference-service] f:securityContext: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:allowPrivilegeEscalation: {} [e2e-llm-inference-service] f:capabilities: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:drop: {} [e2e-llm-inference-service] f:readOnlyRootFilesystem: {} [e2e-llm-inference-service] f:runAsNonRoot: {} [e2e-llm-inference-service] f:seccompProfile: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:type: {} [e2e-llm-inference-service] f:startupProbe: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:failureThreshold: {} [e2e-llm-inference-service] f:httpGet: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:path: {} [e2e-llm-inference-service] f:port: {} [e2e-llm-inference-service] f:scheme: {} [e2e-llm-inference-service] f:periodSeconds: {} [e2e-llm-inference-service] f:successThreshold: {} [e2e-llm-inference-service] f:timeoutSeconds: {} [e2e-llm-inference-service] f:terminationMessagePath: {} [e2e-llm-inference-service] f:terminationMessagePolicy: {} [e2e-llm-inference-service] f:volumeMounts: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"mountPath":"/dev/shm"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:mountPath: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] k:{"mountPath":"/home"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:mountPath: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] k:{"mountPath":"/mnt/models"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:mountPath: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:readOnly: {} [e2e-llm-inference-service] k:{"mountPath":"/models"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:mountPath: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] k:{"mountPath":"/tmp"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:mountPath: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] k:{"mountPath":"/var/run/kserve/tls"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:mountPath: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:readOnly: {} [e2e-llm-inference-service] f:dnsPolicy: {} [e2e-llm-inference-service] f:initContainers: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"name":"storage-initializer"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:args: {} [e2e-llm-inference-service] f:env: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"name":"AWS_ACCESS_KEY_ID"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:valueFrom: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:secretKeyRef: {} [e2e-llm-inference-service] k:{"name":"AWS_CA_BUNDLE"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"AWS_CA_BUNDLE_CONFIGMAP"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"AWS_ENDPOINT_URL"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"AWS_SECRET_ACCESS_KEY"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:valueFrom: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:secretKeyRef: {} [e2e-llm-inference-service] k:{"name":"HF_HUB_ENABLE_HF_TRANSFER"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"HF_XET_HIGH_PERFORMANCE"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"HF_XET_NUM_CONCURRENT_RANGE_GETS"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"S3_ENDPOINT"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"S3_USE_HTTPS"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"S3_VERIFY_SSL"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] f:image: {} [e2e-llm-inference-service] f:imagePullPolicy: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:resources: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:limits: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:cpu: {} [e2e-llm-inference-service] f:memory: {} [e2e-llm-inference-service] f:requests: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:cpu: {} [e2e-llm-inference-service] f:memory: {} [e2e-llm-inference-service] f:terminationMessagePath: {} [e2e-llm-inference-service] f:terminationMessagePolicy: {} [e2e-llm-inference-service] f:volumeMounts: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"mountPath":"/mnt/models"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:mountPath: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:restartPolicy: {} [e2e-llm-inference-service] f:schedulerName: {} [e2e-llm-inference-service] f:securityContext: {} [e2e-llm-inference-service] f:terminationGracePeriodSeconds: {} [e2e-llm-inference-service] f:volumes: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"name":"dshm"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:emptyDir: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:medium: {} [e2e-llm-inference-service] f:sizeLimit: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] k:{"name":"home"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:emptyDir: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] k:{"name":"kserve-provision-location"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:emptyDir: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] k:{"name":"model-cache"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:emptyDir: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] k:{"name":"tls-certs"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:secret: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:defaultMode: {} [e2e-llm-inference-service] f:secretName: {} [e2e-llm-inference-service] k:{"name":"tmp-dir"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:emptyDir: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] - manager: kube-controller-manager [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] apiVersion: apps/v1 [e2e-llm-inference-service] time: '2026-07-07T14:50:57Z' [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:metadata: [e2e-llm-inference-service] f:annotations: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:deployment.kubernetes.io/revision: {} [e2e-llm-inference-service] f:status: [e2e-llm-inference-service] f:availableReplicas: {} [e2e-llm-inference-service] f:conditions: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"type":"Available"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:lastTransitionTime: {} [e2e-llm-inference-service] f:lastUpdateTime: {} [e2e-llm-inference-service] f:message: {} [e2e-llm-inference-service] f:reason: {} [e2e-llm-inference-service] f:status: {} [e2e-llm-inference-service] f:type: {} [e2e-llm-inference-service] k:{"type":"Progressing"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:lastTransitionTime: {} [e2e-llm-inference-service] f:lastUpdateTime: {} [e2e-llm-inference-service] f:message: {} [e2e-llm-inference-service] f:reason: {} [e2e-llm-inference-service] f:status: {} [e2e-llm-inference-service] f:type: {} [e2e-llm-inference-service] f:observedGeneration: {} [e2e-llm-inference-service] f:readyReplicas: {} [e2e-llm-inference-service] f:replicas: {} [e2e-llm-inference-service] f:updatedReplicas: {} [e2e-llm-inference-service] subresource: status [e2e-llm-inference-service] spec: [e2e-llm-inference-service] replicas: 1 [e2e-llm-inference-service] selector: [e2e-llm-inference-service] matchLabels: [e2e-llm-inference-service] app.kubernetes.io/component: llminferenceservice-workload [e2e-llm-inference-service] app.kubernetes.io/name: custom-route-timeout-test [e2e-llm-inference-service] app.kubernetes.io/part-of: llminferenceservice [e2e-llm-inference-service] kserve.io/component: workload [e2e-llm-inference-service] llm-d.ai/role: both [e2e-llm-inference-service] template: [e2e-llm-inference-service] metadata: [e2e-llm-inference-service] labels: [e2e-llm-inference-service] app.kubernetes.io/component: llminferenceservice-workload [e2e-llm-inference-service] app.kubernetes.io/name: custom-route-timeout-test [e2e-llm-inference-service] app.kubernetes.io/part-of: llminferenceservice [e2e-llm-inference-service] kserve.io/component: workload [e2e-llm-inference-service] llm-d.ai/role: both [e2e-llm-inference-service] spec: [e2e-llm-inference-service] volumes: [e2e-llm-inference-service] - name: home [e2e-llm-inference-service] emptyDir: {} [e2e-llm-inference-service] - name: dshm [e2e-llm-inference-service] emptyDir: [e2e-llm-inference-service] medium: Memory [e2e-llm-inference-service] sizeLimit: 1Gi [e2e-llm-inference-service] - name: model-cache [e2e-llm-inference-service] emptyDir: {} [e2e-llm-inference-service] - name: tmp-dir [e2e-llm-inference-service] emptyDir: {} [e2e-llm-inference-service] - name: tls-certs [e2e-llm-inference-service] secret: [e2e-llm-inference-service] secretName: custom-route-timeout-test-kserve-self-signed-certs [e2e-llm-inference-service] defaultMode: 420 [e2e-llm-inference-service] - name: kserve-provision-location [e2e-llm-inference-service] emptyDir: {} [e2e-llm-inference-service] initContainers: [e2e-llm-inference-service] - name: storage-initializer [e2e-llm-inference-service] image: quay.io/opendatahub/kserve-storage-initializer@sha256:13ad765c0f323200e1db7b45c99661287e65c5158d6b54b16139829636d20be4 [e2e-llm-inference-service] args: [e2e-llm-inference-service] - hf://facebook/opt-125m [e2e-llm-inference-service] - /mnt/models [e2e-llm-inference-service] env: [e2e-llm-inference-service] - name: AWS_ACCESS_KEY_ID [e2e-llm-inference-service] valueFrom: [e2e-llm-inference-service] secretKeyRef: [e2e-llm-inference-service] name: seaweedfs-s3-creds [e2e-llm-inference-service] key: AWS_ACCESS_KEY_ID [e2e-llm-inference-service] - name: AWS_SECRET_ACCESS_KEY [e2e-llm-inference-service] valueFrom: [e2e-llm-inference-service] secretKeyRef: [e2e-llm-inference-service] name: seaweedfs-s3-creds [e2e-llm-inference-service] key: AWS_SECRET_ACCESS_KEY [e2e-llm-inference-service] - name: S3_USE_HTTPS [e2e-llm-inference-service] value: '0' [e2e-llm-inference-service] - name: S3_ENDPOINT [e2e-llm-inference-service] value: s3-service.kserve:8333 [e2e-llm-inference-service] - name: AWS_ENDPOINT_URL [e2e-llm-inference-service] value: http://s3-service.kserve:8333 [e2e-llm-inference-service] - name: S3_VERIFY_SSL [e2e-llm-inference-service] value: '0' [e2e-llm-inference-service] - name: AWS_CA_BUNDLE [e2e-llm-inference-service] value: /etc/ssl/custom-certs/cabundle.crt [e2e-llm-inference-service] - name: AWS_CA_BUNDLE_CONFIGMAP [e2e-llm-inference-service] value: odh-kserve-custom-ca-bundle [e2e-llm-inference-service] - name: HF_HUB_ENABLE_HF_TRANSFER [e2e-llm-inference-service] value: '1' [e2e-llm-inference-service] - name: HF_XET_HIGH_PERFORMANCE [e2e-llm-inference-service] value: '1' [e2e-llm-inference-service] - name: HF_XET_NUM_CONCURRENT_RANGE_GETS [e2e-llm-inference-service] value: '8' [e2e-llm-inference-service] resources: [e2e-llm-inference-service] limits: [e2e-llm-inference-service] cpu: '1' [e2e-llm-inference-service] memory: 24Gi [e2e-llm-inference-service] requests: [e2e-llm-inference-service] cpu: 100m [e2e-llm-inference-service] memory: 100Mi [e2e-llm-inference-service] volumeMounts: [e2e-llm-inference-service] - name: kserve-provision-location [e2e-llm-inference-service] mountPath: /mnt/models [e2e-llm-inference-service] terminationMessagePath: /dev/termination-log [e2e-llm-inference-service] terminationMessagePolicy: FallbackToLogsOnError [e2e-llm-inference-service] imagePullPolicy: IfNotPresent [e2e-llm-inference-service] containers: [e2e-llm-inference-service] - name: main [e2e-llm-inference-service] image: public.ecr.aws/q9t5s3a7/vllm-cpu-release-repo:v0.19.0 [e2e-llm-inference-service] command: [e2e-llm-inference-service] - /bin/bash [e2e-llm-inference-service] - -c [e2e-llm-inference-service] - "if [ -f /etc/profile.d/ibm-aiu-setup.sh ]; then\n source /etc/profile.d/ibm-aiu-setup.sh\n\ [e2e-llm-inference-service] fi\n\nif [ \"$KSERVE_INFER_ROCE\" = \"true\" ]; then\n echo \"Trying to\ [e2e-llm-inference-service] \ infer RoCE configs ... \"\n grep -H . /sys/class/infiniband/*/ports/*/gids/*\ [e2e-llm-inference-service] \ 2>/dev/null\n grep -H . /sys/class/infiniband/*/ports/*/gid_attrs/types/*\ [e2e-llm-inference-service] \ 2>/dev/null\n\n cat /proc/driver/nvidia/params\n\n KSERVE_INFER_IB_GID_INDEX_GREP=${KSERVE_INFER_IB_GID_INDEX_GREP:-\"\ [e2e-llm-inference-service] RoCE v2\"}\n\n echo \"[Infer RoCE] Discovering active HCAs ...\"\n active_hcas=()\n\ [e2e-llm-inference-service] \ # Loop through all mlx5 devices found in sysfs\n for hca_dir in /sys/class/infiniband/mlx5_*;\ [e2e-llm-inference-service] \ do\n # Ensure it's a directory before proceeding\n if [ -d \"\ [e2e-llm-inference-service] $hca_dir\" ]; then\n hca_name=$(basename \"$hca_dir\")\n \ [e2e-llm-inference-service] \ port_state_file=\"$hca_dir/ports/1/state\" # Assume port 1\n \ [e2e-llm-inference-service] \ type_file=\"$hca_dir/ports/1/gid_attrs/types/*\"\n\n echo\ [e2e-llm-inference-service] \ \"[Infer RoCE] Check if the port state file ${port_state_file} exists\ [e2e-llm-inference-service] \ and contains 'ACTIVE'\"\n if [ -f \"$port_state_file\" ] && grep\ [e2e-llm-inference-service] \ -q \"ACTIVE\" \"$port_state_file\" && grep -q \"${KSERVE_INFER_IB_GID_INDEX_GREP}\"\ [e2e-llm-inference-service] \ ${type_file} 2>/dev/null; then\n echo \"[Infer RoCE] Found\ [e2e-llm-inference-service] \ active HCA: $hca_name\"\n active_hcas+=(\"$hca_name\")\n\ [e2e-llm-inference-service] \ else\n echo \"[Infer RoCE] Skipping inactive or\ [e2e-llm-inference-service] \ down HCA: $hca_name\"\n fi\n fi\n done\n\n # Check if\ [e2e-llm-inference-service] \ we found any active HCAs\n if [ ${#active_hcas[@]} -gt 0 ]; then\n \ [e2e-llm-inference-service] \ # Join the array elements with a comma\n hca_port_pairs=()\n \ [e2e-llm-inference-service] \ for hca in \"${active_hcas[@]}\"; do\n hca_port_pairs+=(\"\ [e2e-llm-inference-service] ${hca}:1\")\n done\n\n active_hca_list=$(IFS=,; echo \"${active_hcas[*]}\"\ [e2e-llm-inference-service] )\n hca_port_pairs_list=$(IFS=,; echo \"${hca_port_pairs[*]}\")\n \ [e2e-llm-inference-service] \ echo \"[Infer RoCE] Setting active HCAs: ${active_hca_list}\"\n \ [e2e-llm-inference-service] \ export NCCL_IB_HCA=${NCCL_IB_HCA:-${active_hca_list}}\n export\ [e2e-llm-inference-service] \ NVSHMEM_HCA_LIST=${NVSHMEM_HCA_LIST:-${hca_port_pairs_list}}\n export\ [e2e-llm-inference-service] \ UCX_NET_DEVICES=${UCX_NET_DEVICES:-${hca_port_pairs_list}}\n\n echo\ [e2e-llm-inference-service] \ \"[Infer RoCE] NCCL_IB_HCA=${NCCL_IB_HCA}\"\n echo \"[Infer RoCE]\ [e2e-llm-inference-service] \ NVSHMEM_HCA_LIST=${NVSHMEM_HCA_LIST}\"\n echo \"[Infer RoCE] UCX_NET_DEVICES=${UCX_NET_DEVICES}\"\ [e2e-llm-inference-service] \n else\n echo \"[Infer RoCE] WARNING: No active RoCE HCAs found.\ [e2e-llm-inference-service] \ NCCL_IB_HCA will not be set.\"\n fi\n\n if [ ${#active_hcas[@]} -gt\ [e2e-llm-inference-service] \ 0 ]; then\n echo \"[Infer RoCE] Finding GID_INDEX for each active\ [e2e-llm-inference-service] \ HCA (SR-IOV compatible)...\"\n\n # For SR-IOV environments, find\ [e2e-llm-inference-service] \ the most common IPv4 RoCE v2 GID index across all HCAs\n declare\ [e2e-llm-inference-service] \ -A gid_index_count\n declare -A hca_gid_index\n\n for hca_name\ [e2e-llm-inference-service] \ in \"${active_hcas[@]}\"; do\n echo \"[Infer RoCE] Processing\ [e2e-llm-inference-service] \ HCA: ${hca_name}\"\n\n # Find all RoCE v2 IPv4 GIDs for this\ [e2e-llm-inference-service] \ HCA and count by index\n for tpath in /sys/class/infiniband/${hca_name}/ports/1/gid_attrs/types/*;\ [e2e-llm-inference-service] \ do\n if grep -q \"${KSERVE_INFER_IB_GID_INDEX_GREP}\" \"\ [e2e-llm-inference-service] $tpath\" 2>/dev/null; then\n idx=$(basename \"$tpath\"\ [e2e-llm-inference-service] )\n gid_file=\"/sys/class/infiniband/${hca_name}/ports/1/gids/${idx}\"\ [e2e-llm-inference-service] \n # Check for IPv4 GID (contains ffff:)\n \ [e2e-llm-inference-service] \ if [ -f \"$gid_file\" ] && grep -q \"ffff:\" \"$gid_file\"; then\n\ [e2e-llm-inference-service] \ gid_value=$(cat \"$gid_file\" 2>/dev/null || echo\ [e2e-llm-inference-service] \ \"\")\n echo \"[Infer RoCE] Found IPv4 RoCE v2 GID\ [e2e-llm-inference-service] \ for ${hca_name}: index=${idx}, gid=${gid_value}\"\n \ [e2e-llm-inference-service] \ hca_gid_index[\"${hca_name}\"]=\"${idx}\"\n gid_index_count[\"\ [e2e-llm-inference-service] ${idx}\"]=$((${gid_index_count[\"${idx}\"]} + 1))\n \ [e2e-llm-inference-service] \ break # Use first found IPv4 GID per HCA\n fi\n \ [e2e-llm-inference-service] \ fi\n done\n done\n\n # Find the most common\ [e2e-llm-inference-service] \ GID index (most likely to be consistent across nodes)\n best_gid_index=\"\ [e2e-llm-inference-service] \"\n max_count=0\n for idx in \"${!gid_index_count[@]}\"; do\n\ [e2e-llm-inference-service] \ count=${gid_index_count[\"${idx}\"]}\n echo \"[Infer\ [e2e-llm-inference-service] \ RoCE] GID_INDEX ${idx} found on ${count} HCAs\"\n if [ $count\ [e2e-llm-inference-service] \ -gt $max_count ]; then\n max_count=$count\n \ [e2e-llm-inference-service] \ best_gid_index=\"$idx\"\n fi\n done\n\n # Use deterministic\ [e2e-llm-inference-service] \ fallback if tied - prefer index 3 (SR-IOV standard)\n if [ ${#gid_index_count[@]}\ [e2e-llm-inference-service] \ -gt 1 ]; then\n echo \"[Infer RoCE] Multiple GID indices found,\ [e2e-llm-inference-service] \ selecting most common: ${best_gid_index}\"\n # If there's a tie,\ [e2e-llm-inference-service] \ prefer index 3 as it's most common in SR-IOV setups\n if [ -n\ [e2e-llm-inference-service] \ \"${gid_index_count['3']}\" ] && [ \"${gid_index_count['3']}\" -eq \"\ [e2e-llm-inference-service] $max_count\" ]; then\n best_gid_index=\"3\"\n \ [e2e-llm-inference-service] \ echo \"[Infer RoCE] Using deterministic fallback: GID_INDEX=3 (SR-IOV\ [e2e-llm-inference-service] \ standard)\"\n fi\n fi\n\n # Check if GID_INDEX is already\ [e2e-llm-inference-service] \ set via environment variables\n if [ -n \"${NCCL_IB_GID_INDEX}\"\ [e2e-llm-inference-service] \ ]; then\n echo \"[Infer RoCE] Using pre-configured NCCL_IB_GID_INDEX=${NCCL_IB_GID_INDEX}\ [e2e-llm-inference-service] \ from environment\"\n export NVSHMEM_IB_GID_INDEX=${NVSHMEM_IB_GID_INDEX:-$NCCL_IB_GID_INDEX}\n\ [e2e-llm-inference-service] \ export UCX_IB_GID_INDEX=${UCX_IB_GID_INDEX:-$NCCL_IB_GID_INDEX}\n\ [e2e-llm-inference-service] \ echo \"[Infer RoCE] Using pre-configured GID_INDEX=${NCCL_IB_GID_INDEX}\ [e2e-llm-inference-service] \ for NCCL, NVSHMEM, and UCX\"\n elif [ -n \"$best_gid_index\" ]; then\n\ [e2e-llm-inference-service] \ echo \"[Infer RoCE] Selected GID_INDEX: ${best_gid_index} (found\ [e2e-llm-inference-service] \ on ${max_count} HCAs)\"\n\n export NCCL_IB_GID_INDEX=${NCCL_IB_GID_INDEX:-$best_gid_index}\n\ [e2e-llm-inference-service] \ export NVSHMEM_IB_GID_INDEX=${NVSHMEM_IB_GID_INDEX:-$best_gid_index}\n\ [e2e-llm-inference-service] \ export UCX_IB_GID_INDEX=${UCX_IB_GID_INDEX:-$best_gid_index}\n\ [e2e-llm-inference-service] \n echo \"[Infer RoCE] Exported GID_INDEX=${best_gid_index} for\ [e2e-llm-inference-service] \ NCCL, NVSHMEM, and UCX\"\n else\n echo \"[Infer RoCE] ERROR:\ [e2e-llm-inference-service] \ No valid IPv4 ${KSERVE_INFER_IB_GID_INDEX_GREP} GID_INDEX found on any\ [e2e-llm-inference-service] \ HCA.\"\n fi\n else\n echo \"[Infer RoCE] No active HCAs found,\ [e2e-llm-inference-service] \ skipping GID_INDEX inference.\"\n fi\nfi\n\n# --disable-access-log-for-endpoints\ [e2e-llm-inference-service] \ landed in vLLM 0.16.0 (vllm-project/vllm#30011).\n# Older versions still\ [e2e-llm-inference-service] \ need the blanket --disable-uvicorn-access-log.\nACCESS_LOG_ARGS=\"--disable-uvicorn-access-log\"\ [e2e-llm-inference-service] \nVLLM_VERSION=$(vllm --version 2>/dev/null | tail -1 | awk '{print $NF}')\n\ [e2e-llm-inference-service] echo \"[access-log-detect] vllm version='${VLLM_VERSION}'\"\nif [[ \"$VLLM_VERSION\"\ [e2e-llm-inference-service] \ =~ ^[0-9]+\\.[0-9]+ ]] && [ \"$(printf '%s\\n%s\\n' \"0.16.0\" \"${VLLM_VERSION}\"\ [e2e-llm-inference-service] \ | sort -V | head -1)\" = \"0.16.0\" ]; then\n ACCESS_LOG_ARGS=\"--disable-access-log-for-endpoints\ [e2e-llm-inference-service] \ /health,/metrics,/ping\"\nfi\necho \"[access-log-detect] selected ACCESS_LOG_ARGS='${ACCESS_LOG_ARGS}'\"\ [e2e-llm-inference-service] \n\n# --shutdown-timeout landed in vLLM 0.18.0 (vllm-project/vllm#36666).\n\ [e2e-llm-inference-service] SHUTDOWN_TIMEOUT_ARGS=\"\"\nif [[ \"$VLLM_VERSION\" =~ ^[0-9]+\\.[0-9]+\ [e2e-llm-inference-service] \ ]] && [ \"$(printf '%s\\n%s\\n' \"0.18.0\" \"${VLLM_VERSION}\" | sort\ [e2e-llm-inference-service] \ -V | head -1)\" = \"0.18.0\" ]; then\n SHUTDOWN_TIMEOUT_ARGS=\"--shutdown-timeout\ [e2e-llm-inference-service] \ 40\"\nfi\n\neval \"exec vllm serve /mnt/models \\\n --served-model-name\ [e2e-llm-inference-service] \ \"facebook/opt-125m\" \"publishers/kserve-ci-e2e-test/models/facebook/opt-125m\"\ [e2e-llm-inference-service] \ \\\n --port 8000 \\\n ${ACCESS_LOG_ARGS} \\\n ${SHUTDOWN_TIMEOUT_ARGS}\ [e2e-llm-inference-service] \ \\\n --enable-ssl-refresh \\\n --ssl-certfile /var/run/kserve/tls/tls.crt\ [e2e-llm-inference-service] \ \\\n --ssl-keyfile /var/run/kserve/tls/tls.key \\\n ${VLLM_ADDITIONAL_ARGS}\ [e2e-llm-inference-service] \ \\\n $@\"" [e2e-llm-inference-service] - -- [e2e-llm-inference-service] ports: [e2e-llm-inference-service] - containerPort: 8000 [e2e-llm-inference-service] protocol: TCP [e2e-llm-inference-service] env: [e2e-llm-inference-service] - name: HOME [e2e-llm-inference-service] value: /home [e2e-llm-inference-service] - name: VLLM_LOGGING_LEVEL [e2e-llm-inference-service] value: DEBUG [e2e-llm-inference-service] - name: VLLM_CPU_KVCACHE_SPACE [e2e-llm-inference-service] value: '1' [e2e-llm-inference-service] - name: VLLM_ENABLE_V1_MULTIPROCESSING [e2e-llm-inference-service] value: '0' [e2e-llm-inference-service] - name: USER [e2e-llm-inference-service] value: nonroot [e2e-llm-inference-service] - name: TORCHINDUCTOR_CACHE_DIR [e2e-llm-inference-service] value: /tmp/torchinductor-cache [e2e-llm-inference-service] - name: HF_HUB_CACHE [e2e-llm-inference-service] value: /models [e2e-llm-inference-service] resources: [e2e-llm-inference-service] limits: [e2e-llm-inference-service] cpu: '2' [e2e-llm-inference-service] memory: 7Gi [e2e-llm-inference-service] requests: [e2e-llm-inference-service] cpu: 200m [e2e-llm-inference-service] memory: 2Gi [e2e-llm-inference-service] volumeMounts: [e2e-llm-inference-service] - name: home [e2e-llm-inference-service] mountPath: /home [e2e-llm-inference-service] - name: tmp-dir [e2e-llm-inference-service] mountPath: /tmp [e2e-llm-inference-service] - name: dshm [e2e-llm-inference-service] mountPath: /dev/shm [e2e-llm-inference-service] - name: model-cache [e2e-llm-inference-service] mountPath: /models [e2e-llm-inference-service] - name: tls-certs [e2e-llm-inference-service] readOnly: true [e2e-llm-inference-service] mountPath: /var/run/kserve/tls [e2e-llm-inference-service] - name: kserve-provision-location [e2e-llm-inference-service] readOnly: true [e2e-llm-inference-service] mountPath: /mnt/models [e2e-llm-inference-service] livenessProbe: [e2e-llm-inference-service] httpGet: [e2e-llm-inference-service] path: /health [e2e-llm-inference-service] port: 8000 [e2e-llm-inference-service] scheme: HTTPS [e2e-llm-inference-service] timeoutSeconds: 1 [e2e-llm-inference-service] periodSeconds: 10 [e2e-llm-inference-service] successThreshold: 1 [e2e-llm-inference-service] failureThreshold: 10 [e2e-llm-inference-service] readinessProbe: [e2e-llm-inference-service] httpGet: [e2e-llm-inference-service] path: /health [e2e-llm-inference-service] port: 8000 [e2e-llm-inference-service] scheme: HTTPS [e2e-llm-inference-service] timeoutSeconds: 1 [e2e-llm-inference-service] periodSeconds: 1 [e2e-llm-inference-service] successThreshold: 1 [e2e-llm-inference-service] failureThreshold: 2 [e2e-llm-inference-service] startupProbe: [e2e-llm-inference-service] httpGet: [e2e-llm-inference-service] path: /health [e2e-llm-inference-service] port: 8000 [e2e-llm-inference-service] scheme: HTTPS [e2e-llm-inference-service] timeoutSeconds: 1 [e2e-llm-inference-service] periodSeconds: 10 [e2e-llm-inference-service] successThreshold: 1 [e2e-llm-inference-service] failureThreshold: 60 [e2e-llm-inference-service] lifecycle: [e2e-llm-inference-service] preStop: [e2e-llm-inference-service] exec: [e2e-llm-inference-service] command: [e2e-llm-inference-service] - /bin/sleep [e2e-llm-inference-service] - '15' [e2e-llm-inference-service] terminationMessagePath: /dev/termination-log [e2e-llm-inference-service] terminationMessagePolicy: FallbackToLogsOnError [e2e-llm-inference-service] imagePullPolicy: IfNotPresent [e2e-llm-inference-service] securityContext: [e2e-llm-inference-service] capabilities: [e2e-llm-inference-service] drop: [e2e-llm-inference-service] - ALL [e2e-llm-inference-service] runAsNonRoot: true [e2e-llm-inference-service] readOnlyRootFilesystem: true [e2e-llm-inference-service] allowPrivilegeEscalation: false [e2e-llm-inference-service] seccompProfile: [e2e-llm-inference-service] type: RuntimeDefault [e2e-llm-inference-service] restartPolicy: Always [e2e-llm-inference-service] terminationGracePeriodSeconds: 60 [e2e-llm-inference-service] dnsPolicy: ClusterFirst [e2e-llm-inference-service] securityContext: {} [e2e-llm-inference-service] schedulerName: default-scheduler [e2e-llm-inference-service] strategy: [e2e-llm-inference-service] type: RollingUpdate [e2e-llm-inference-service] rollingUpdate: [e2e-llm-inference-service] maxUnavailable: 25% [e2e-llm-inference-service] maxSurge: 25% [e2e-llm-inference-service] revisionHistoryLimit: 10 [e2e-llm-inference-service] progressDeadlineSeconds: 600 [e2e-llm-inference-service] status: [e2e-llm-inference-service] observedGeneration: 1 [e2e-llm-inference-service] replicas: 1 [e2e-llm-inference-service] updatedReplicas: 1 [e2e-llm-inference-service] readyReplicas: 1 [e2e-llm-inference-service] availableReplicas: 1 [e2e-llm-inference-service] conditions: [e2e-llm-inference-service] - type: Available [e2e-llm-inference-service] status: 'True' [e2e-llm-inference-service] lastUpdateTime: '2026-07-07T14:50:57Z' [e2e-llm-inference-service] lastTransitionTime: '2026-07-07T14:50:57Z' [e2e-llm-inference-service] reason: MinimumReplicasAvailable [e2e-llm-inference-service] message: Deployment has minimum availability. [e2e-llm-inference-service] - type: Progressing [e2e-llm-inference-service] status: 'True' [e2e-llm-inference-service] lastUpdateTime: '2026-07-07T14:50:57Z' [e2e-llm-inference-service] lastTransitionTime: '2026-07-07T14:48:57Z' [e2e-llm-inference-service] reason: NewReplicaSetAvailable [e2e-llm-inference-service] message: ReplicaSet "custom-route-timeout-test-kserve-7f56b8c469" has successfully [e2e-llm-inference-service] progressed. [e2e-llm-inference-service] apiVersion: apps/v1 [e2e-llm-inference-service] kind: Deployment [e2e-llm-inference-service] [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1254 --- [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1255 metadata: [e2e-llm-inference-service] name: custom-route-timeout-test-kserve-router-scheduler [e2e-llm-inference-service] namespace: kserve-ci-e2e-test [e2e-llm-inference-service] uid: eca54665-59f5-44bb-a135-818e809de51c [e2e-llm-inference-service] resourceVersion: '49592' [e2e-llm-inference-service] generation: 1 [e2e-llm-inference-service] creationTimestamp: '2026-07-07T14:48:57Z' [e2e-llm-inference-service] labels: [e2e-llm-inference-service] app.kubernetes.io/component: llminferenceservice-router-scheduler [e2e-llm-inference-service] app.kubernetes.io/name: custom-route-timeout-test [e2e-llm-inference-service] app.kubernetes.io/part-of: llminferenceservice [e2e-llm-inference-service] annotations: [e2e-llm-inference-service] deployment.kubernetes.io/revision: '1' [e2e-llm-inference-service] ownerReferences: [e2e-llm-inference-service] - apiVersion: serving.kserve.io/v1alpha2 [e2e-llm-inference-service] kind: LLMInferenceService [e2e-llm-inference-service] name: custom-route-timeout-test [e2e-llm-inference-service] uid: 986f2d2c-a6d1-463d-9850-3aad6459eace [e2e-llm-inference-service] controller: true [e2e-llm-inference-service] blockOwnerDeletion: true [e2e-llm-inference-service] managedFields: [e2e-llm-inference-service] - manager: manager [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] apiVersion: apps/v1 [e2e-llm-inference-service] time: '2026-07-07T14:48:57Z' [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:metadata: [e2e-llm-inference-service] f:labels: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:app.kubernetes.io/component: {} [e2e-llm-inference-service] f:app.kubernetes.io/name: {} [e2e-llm-inference-service] f:app.kubernetes.io/part-of: {} [e2e-llm-inference-service] f:ownerReferences: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"uid":"986f2d2c-a6d1-463d-9850-3aad6459eace"}: {} [e2e-llm-inference-service] f:spec: [e2e-llm-inference-service] f:progressDeadlineSeconds: {} [e2e-llm-inference-service] f:replicas: {} [e2e-llm-inference-service] f:revisionHistoryLimit: {} [e2e-llm-inference-service] f:selector: {} [e2e-llm-inference-service] f:strategy: [e2e-llm-inference-service] f:type: {} [e2e-llm-inference-service] f:template: [e2e-llm-inference-service] f:metadata: [e2e-llm-inference-service] f:annotations: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:app.kubernetes.io/version: {} [e2e-llm-inference-service] f:certificates.kserve.io/expiration-v2: {} [e2e-llm-inference-service] f:labels: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:app.kubernetes.io/component: {} [e2e-llm-inference-service] f:app.kubernetes.io/name: {} [e2e-llm-inference-service] f:app.kubernetes.io/part-of: {} [e2e-llm-inference-service] f:spec: [e2e-llm-inference-service] f:containers: [e2e-llm-inference-service] k:{"name":"main"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:args: {} [e2e-llm-inference-service] f:command: {} [e2e-llm-inference-service] f:env: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"name":"SSL_CERT_DIR"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] f:image: {} [e2e-llm-inference-service] f:imagePullPolicy: {} [e2e-llm-inference-service] f:lifecycle: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:preStop: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:exec: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:command: {} [e2e-llm-inference-service] f:livenessProbe: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:failureThreshold: {} [e2e-llm-inference-service] f:grpc: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:port: {} [e2e-llm-inference-service] f:service: {} [e2e-llm-inference-service] f:initialDelaySeconds: {} [e2e-llm-inference-service] f:periodSeconds: {} [e2e-llm-inference-service] f:successThreshold: {} [e2e-llm-inference-service] f:timeoutSeconds: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:ports: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"containerPort":5557,"protocol":"TCP"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:containerPort: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:protocol: {} [e2e-llm-inference-service] k:{"containerPort":9002,"protocol":"TCP"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:containerPort: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:protocol: {} [e2e-llm-inference-service] k:{"containerPort":9003,"protocol":"TCP"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:containerPort: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:protocol: {} [e2e-llm-inference-service] k:{"containerPort":9090,"protocol":"TCP"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:containerPort: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:protocol: {} [e2e-llm-inference-service] f:readinessProbe: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:failureThreshold: {} [e2e-llm-inference-service] f:grpc: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:port: {} [e2e-llm-inference-service] f:service: {} [e2e-llm-inference-service] f:initialDelaySeconds: {} [e2e-llm-inference-service] f:periodSeconds: {} [e2e-llm-inference-service] f:successThreshold: {} [e2e-llm-inference-service] f:timeoutSeconds: {} [e2e-llm-inference-service] f:resources: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:requests: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:cpu: {} [e2e-llm-inference-service] f:memory: {} [e2e-llm-inference-service] f:securityContext: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:allowPrivilegeEscalation: {} [e2e-llm-inference-service] f:capabilities: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:drop: {} [e2e-llm-inference-service] f:readOnlyRootFilesystem: {} [e2e-llm-inference-service] f:runAsNonRoot: {} [e2e-llm-inference-service] f:seccompProfile: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:type: {} [e2e-llm-inference-service] f:terminationMessagePath: {} [e2e-llm-inference-service] f:terminationMessagePolicy: {} [e2e-llm-inference-service] f:volumeMounts: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"mountPath":"/tmp/tokenizer"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:mountPath: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] k:{"mountPath":"/var/run/kserve/tls"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:mountPath: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:readOnly: {} [e2e-llm-inference-service] f:dnsPolicy: {} [e2e-llm-inference-service] f:initContainers: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"name":"storage-initializer"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:args: {} [e2e-llm-inference-service] f:env: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"name":"AWS_ACCESS_KEY_ID"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:valueFrom: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:secretKeyRef: {} [e2e-llm-inference-service] k:{"name":"AWS_CA_BUNDLE"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"AWS_CA_BUNDLE_CONFIGMAP"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"AWS_ENDPOINT_URL"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"AWS_SECRET_ACCESS_KEY"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:valueFrom: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:secretKeyRef: {} [e2e-llm-inference-service] k:{"name":"HF_HUB_ENABLE_HF_TRANSFER"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"HF_XET_HIGH_PERFORMANCE"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"HF_XET_NUM_CONCURRENT_RANGE_GETS"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"S3_ENDPOINT"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"S3_USE_HTTPS"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"S3_VERIFY_SSL"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"STORAGE_ALLOW_PATTERNS"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] f:image: {} [e2e-llm-inference-service] f:imagePullPolicy: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:resources: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:limits: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:cpu: {} [e2e-llm-inference-service] f:memory: {} [e2e-llm-inference-service] f:requests: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:cpu: {} [e2e-llm-inference-service] f:memory: {} [e2e-llm-inference-service] f:terminationMessagePath: {} [e2e-llm-inference-service] f:terminationMessagePolicy: {} [e2e-llm-inference-service] f:volumeMounts: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"mountPath":"/mnt/models"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:mountPath: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:restartPolicy: {} [e2e-llm-inference-service] f:schedulerName: {} [e2e-llm-inference-service] f:securityContext: {} [e2e-llm-inference-service] f:serviceAccount: {} [e2e-llm-inference-service] f:serviceAccountName: {} [e2e-llm-inference-service] f:terminationGracePeriodSeconds: {} [e2e-llm-inference-service] f:volumes: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"name":"kserve-provision-location"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:emptyDir: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] k:{"name":"tls-certs"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:secret: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:defaultMode: {} [e2e-llm-inference-service] f:secretName: {} [e2e-llm-inference-service] k:{"name":"tokenizer-cache"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:emptyDir: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] k:{"name":"tokenizer-tmp"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:emptyDir: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] k:{"name":"tokenizer-uds"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:emptyDir: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] - manager: kube-controller-manager [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] apiVersion: apps/v1 [e2e-llm-inference-service] time: '2026-07-07T14:49:31Z' [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:metadata: [e2e-llm-inference-service] f:annotations: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:deployment.kubernetes.io/revision: {} [e2e-llm-inference-service] f:status: [e2e-llm-inference-service] f:availableReplicas: {} [e2e-llm-inference-service] f:conditions: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"type":"Available"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:lastTransitionTime: {} [e2e-llm-inference-service] f:lastUpdateTime: {} [e2e-llm-inference-service] f:message: {} [e2e-llm-inference-service] f:reason: {} [e2e-llm-inference-service] f:status: {} [e2e-llm-inference-service] f:type: {} [e2e-llm-inference-service] k:{"type":"Progressing"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:lastTransitionTime: {} [e2e-llm-inference-service] f:lastUpdateTime: {} [e2e-llm-inference-service] f:message: {} [e2e-llm-inference-service] f:reason: {} [e2e-llm-inference-service] f:status: {} [e2e-llm-inference-service] f:type: {} [e2e-llm-inference-service] f:observedGeneration: {} [e2e-llm-inference-service] f:readyReplicas: {} [e2e-llm-inference-service] f:replicas: {} [e2e-llm-inference-service] f:updatedReplicas: {} [e2e-llm-inference-service] subresource: status [e2e-llm-inference-service] spec: [e2e-llm-inference-service] replicas: 1 [e2e-llm-inference-service] selector: [e2e-llm-inference-service] matchLabels: [e2e-llm-inference-service] app.kubernetes.io/component: llminferenceservice-router-scheduler [e2e-llm-inference-service] app.kubernetes.io/name: custom-route-timeout-test [e2e-llm-inference-service] app.kubernetes.io/part-of: llminferenceservice [e2e-llm-inference-service] template: [e2e-llm-inference-service] metadata: [e2e-llm-inference-service] labels: [e2e-llm-inference-service] app.kubernetes.io/component: llminferenceservice-router-scheduler [e2e-llm-inference-service] app.kubernetes.io/name: custom-route-timeout-test [e2e-llm-inference-service] app.kubernetes.io/part-of: llminferenceservice [e2e-llm-inference-service] annotations: [e2e-llm-inference-service] app.kubernetes.io/version: 0.9.0 [e2e-llm-inference-service] certificates.kserve.io/expiration-v2: 'true' [e2e-llm-inference-service] spec: [e2e-llm-inference-service] volumes: [e2e-llm-inference-service] - name: tls-certs [e2e-llm-inference-service] secret: [e2e-llm-inference-service] secretName: custom-route-timeout-test-kserve-self-signed-certs [e2e-llm-inference-service] defaultMode: 420 [e2e-llm-inference-service] - name: tokenizer-uds [e2e-llm-inference-service] emptyDir: {} [e2e-llm-inference-service] - name: tokenizer-tmp [e2e-llm-inference-service] emptyDir: {} [e2e-llm-inference-service] - name: tokenizer-cache [e2e-llm-inference-service] emptyDir: {} [e2e-llm-inference-service] - name: kserve-provision-location [e2e-llm-inference-service] emptyDir: {} [e2e-llm-inference-service] initContainers: [e2e-llm-inference-service] - name: storage-initializer [e2e-llm-inference-service] image: quay.io/opendatahub/kserve-storage-initializer@sha256:13ad765c0f323200e1db7b45c99661287e65c5158d6b54b16139829636d20be4 [e2e-llm-inference-service] args: [e2e-llm-inference-service] - hf://facebook/opt-125m [e2e-llm-inference-service] - /mnt/models [e2e-llm-inference-service] env: [e2e-llm-inference-service] - name: AWS_ACCESS_KEY_ID [e2e-llm-inference-service] valueFrom: [e2e-llm-inference-service] secretKeyRef: [e2e-llm-inference-service] name: seaweedfs-s3-creds [e2e-llm-inference-service] key: AWS_ACCESS_KEY_ID [e2e-llm-inference-service] - name: AWS_SECRET_ACCESS_KEY [e2e-llm-inference-service] valueFrom: [e2e-llm-inference-service] secretKeyRef: [e2e-llm-inference-service] name: seaweedfs-s3-creds [e2e-llm-inference-service] key: AWS_SECRET_ACCESS_KEY [e2e-llm-inference-service] - name: S3_USE_HTTPS [e2e-llm-inference-service] value: '0' [e2e-llm-inference-service] - name: S3_ENDPOINT [e2e-llm-inference-service] value: s3-service.kserve:8333 [e2e-llm-inference-service] - name: AWS_ENDPOINT_URL [e2e-llm-inference-service] value: http://s3-service.kserve:8333 [e2e-llm-inference-service] - name: S3_VERIFY_SSL [e2e-llm-inference-service] value: '0' [e2e-llm-inference-service] - name: AWS_CA_BUNDLE [e2e-llm-inference-service] value: /etc/ssl/custom-certs/cabundle.crt [e2e-llm-inference-service] - name: AWS_CA_BUNDLE_CONFIGMAP [e2e-llm-inference-service] value: odh-kserve-custom-ca-bundle [e2e-llm-inference-service] - name: HF_HUB_ENABLE_HF_TRANSFER [e2e-llm-inference-service] value: '1' [e2e-llm-inference-service] - name: HF_XET_HIGH_PERFORMANCE [e2e-llm-inference-service] value: '1' [e2e-llm-inference-service] - name: HF_XET_NUM_CONCURRENT_RANGE_GETS [e2e-llm-inference-service] value: '8' [e2e-llm-inference-service] - name: STORAGE_ALLOW_PATTERNS [e2e-llm-inference-service] value: '["tokenizer.json", "tokenizer_config.json", "special_tokens_map.json", [e2e-llm-inference-service] "vocab.json", "merges.txt", "config.json", "generation_config.json"]' [e2e-llm-inference-service] resources: [e2e-llm-inference-service] limits: [e2e-llm-inference-service] cpu: '1' [e2e-llm-inference-service] memory: 24Gi [e2e-llm-inference-service] requests: [e2e-llm-inference-service] cpu: 100m [e2e-llm-inference-service] memory: 100Mi [e2e-llm-inference-service] volumeMounts: [e2e-llm-inference-service] - name: kserve-provision-location [e2e-llm-inference-service] mountPath: /mnt/models [e2e-llm-inference-service] terminationMessagePath: /dev/termination-log [e2e-llm-inference-service] terminationMessagePolicy: FallbackToLogsOnError [e2e-llm-inference-service] imagePullPolicy: IfNotPresent [e2e-llm-inference-service] containers: [e2e-llm-inference-service] - name: main [e2e-llm-inference-service] image: ghcr.io/llm-d/llm-d-router-endpoint-picker:v0.9.0-rc.2 [e2e-llm-inference-service] command: [e2e-llm-inference-service] - /app/epp [e2e-llm-inference-service] - --pool-name [e2e-llm-inference-service] - custom-route-timeout-test-inference-pool [e2e-llm-inference-service] - --pool-namespace [e2e-llm-inference-service] - kserve-ci-e2e-test [e2e-llm-inference-service] - --zap-encoder [e2e-llm-inference-service] - json [e2e-llm-inference-service] - --grpc-port [e2e-llm-inference-service] - '9002' [e2e-llm-inference-service] - --grpc-health-port [e2e-llm-inference-service] - '9003' [e2e-llm-inference-service] - --enable-cert-reload=true [e2e-llm-inference-service] - --secure-serving=true [e2e-llm-inference-service] - --model-server-metrics-scheme=https [e2e-llm-inference-service] - --cert-path=/var/run/kserve/tls [e2e-llm-inference-service] args: [e2e-llm-inference-service] - --config-text [e2e-llm-inference-service] - "apiVersion: inference.networking.x-k8s.io/v1alpha1\nkind: EndpointPickerConfig\n\ [e2e-llm-inference-service] plugins:\n- type: single-profile-handler\n- type: queue-scorer\n- type:\ [e2e-llm-inference-service] \ prefix-cache-scorer\n- type: max-score-picker\nschedulingProfiles:\n-\ [e2e-llm-inference-service] \ name: default\n plugins:\n - pluginRef: queue-scorer\n weight: 2\n\ [e2e-llm-inference-service] \ - pluginRef: prefix-cache-scorer\n weight: 3\n - pluginRef: max-score-picker\n" [e2e-llm-inference-service] ports: [e2e-llm-inference-service] - name: grpc [e2e-llm-inference-service] containerPort: 9002 [e2e-llm-inference-service] protocol: TCP [e2e-llm-inference-service] - name: grpc-health [e2e-llm-inference-service] containerPort: 9003 [e2e-llm-inference-service] protocol: TCP [e2e-llm-inference-service] - name: metrics [e2e-llm-inference-service] containerPort: 9090 [e2e-llm-inference-service] protocol: TCP [e2e-llm-inference-service] - name: zmq [e2e-llm-inference-service] containerPort: 5557 [e2e-llm-inference-service] protocol: TCP [e2e-llm-inference-service] env: [e2e-llm-inference-service] - name: SSL_CERT_DIR [e2e-llm-inference-service] value: /var/run/kserve/tls:/var/run/secrets/kubernetes.io/serviceaccount:/etc/pki/tls/certs [e2e-llm-inference-service] resources: [e2e-llm-inference-service] requests: [e2e-llm-inference-service] cpu: 256m [e2e-llm-inference-service] memory: 500Mi [e2e-llm-inference-service] volumeMounts: [e2e-llm-inference-service] - name: tls-certs [e2e-llm-inference-service] readOnly: true [e2e-llm-inference-service] mountPath: /var/run/kserve/tls [e2e-llm-inference-service] - name: tokenizer-uds [e2e-llm-inference-service] mountPath: /tmp/tokenizer [e2e-llm-inference-service] livenessProbe: [e2e-llm-inference-service] grpc: [e2e-llm-inference-service] port: 9003 [e2e-llm-inference-service] service: liveness [e2e-llm-inference-service] initialDelaySeconds: 5 [e2e-llm-inference-service] timeoutSeconds: 1 [e2e-llm-inference-service] periodSeconds: 10 [e2e-llm-inference-service] successThreshold: 1 [e2e-llm-inference-service] failureThreshold: 3 [e2e-llm-inference-service] readinessProbe: [e2e-llm-inference-service] grpc: [e2e-llm-inference-service] port: 9003 [e2e-llm-inference-service] service: readiness [e2e-llm-inference-service] initialDelaySeconds: 30 [e2e-llm-inference-service] timeoutSeconds: 1 [e2e-llm-inference-service] periodSeconds: 10 [e2e-llm-inference-service] successThreshold: 1 [e2e-llm-inference-service] failureThreshold: 3 [e2e-llm-inference-service] lifecycle: [e2e-llm-inference-service] preStop: [e2e-llm-inference-service] exec: [e2e-llm-inference-service] command: [e2e-llm-inference-service] - /bin/sleep [e2e-llm-inference-service] - '15' [e2e-llm-inference-service] terminationMessagePath: /dev/termination-log [e2e-llm-inference-service] terminationMessagePolicy: FallbackToLogsOnError [e2e-llm-inference-service] imagePullPolicy: IfNotPresent [e2e-llm-inference-service] securityContext: [e2e-llm-inference-service] capabilities: [e2e-llm-inference-service] drop: [e2e-llm-inference-service] - ALL [e2e-llm-inference-service] runAsNonRoot: true [e2e-llm-inference-service] readOnlyRootFilesystem: true [e2e-llm-inference-service] allowPrivilegeEscalation: false [e2e-llm-inference-service] seccompProfile: [e2e-llm-inference-service] type: RuntimeDefault [e2e-llm-inference-service] restartPolicy: Always [e2e-llm-inference-service] terminationGracePeriodSeconds: 60 [e2e-llm-inference-service] dnsPolicy: ClusterFirst [e2e-llm-inference-service] serviceAccountName: custom-route-timeout-test-epp-sa [e2e-llm-inference-service] serviceAccount: custom-route-timeout-test-epp-sa [e2e-llm-inference-service] securityContext: {} [e2e-llm-inference-service] schedulerName: default-scheduler [e2e-llm-inference-service] strategy: [e2e-llm-inference-service] type: Recreate [e2e-llm-inference-service] revisionHistoryLimit: 10 [e2e-llm-inference-service] progressDeadlineSeconds: 600 [e2e-llm-inference-service] status: [e2e-llm-inference-service] observedGeneration: 1 [e2e-llm-inference-service] replicas: 1 [e2e-llm-inference-service] updatedReplicas: 1 [e2e-llm-inference-service] readyReplicas: 1 [e2e-llm-inference-service] availableReplicas: 1 [e2e-llm-inference-service] conditions: [e2e-llm-inference-service] - type: Available [e2e-llm-inference-service] status: 'True' [e2e-llm-inference-service] lastUpdateTime: '2026-07-07T14:49:31Z' [e2e-llm-inference-service] lastTransitionTime: '2026-07-07T14:49:31Z' [e2e-llm-inference-service] reason: MinimumReplicasAvailable [e2e-llm-inference-service] message: Deployment has minimum availability. [e2e-llm-inference-service] - type: Progressing [e2e-llm-inference-service] status: 'True' [e2e-llm-inference-service] lastUpdateTime: '2026-07-07T14:49:31Z' [e2e-llm-inference-service] lastTransitionTime: '2026-07-07T14:48:57Z' [e2e-llm-inference-service] reason: NewReplicaSetAvailable [e2e-llm-inference-service] message: ReplicaSet "custom-route-timeout-test-kserve-router-scheduler-6578bddf5b" [e2e-llm-inference-service] has successfully progressed. [e2e-llm-inference-service] apiVersion: apps/v1 [e2e-llm-inference-service] kind: Deployment [e2e-llm-inference-service] [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1254 --- [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1255 metadata: [e2e-llm-inference-service] name: custom-route-timeout-test-kserve-7f56b8c469 [e2e-llm-inference-service] namespace: kserve-ci-e2e-test [e2e-llm-inference-service] uid: 908c7ec2-616f-4fd2-a6a5-bea58b91399e [e2e-llm-inference-service] resourceVersion: '50897' [e2e-llm-inference-service] generation: 1 [e2e-llm-inference-service] creationTimestamp: '2026-07-07T14:48:57Z' [e2e-llm-inference-service] labels: [e2e-llm-inference-service] app.kubernetes.io/component: llminferenceservice-workload [e2e-llm-inference-service] app.kubernetes.io/name: custom-route-timeout-test [e2e-llm-inference-service] app.kubernetes.io/part-of: llminferenceservice [e2e-llm-inference-service] kserve.io/component: workload [e2e-llm-inference-service] llm-d.ai/role: both [e2e-llm-inference-service] pod-template-hash: 7f56b8c469 [e2e-llm-inference-service] annotations: [e2e-llm-inference-service] deployment.kubernetes.io/desired-replicas: '1' [e2e-llm-inference-service] deployment.kubernetes.io/max-replicas: '2' [e2e-llm-inference-service] deployment.kubernetes.io/revision: '1' [e2e-llm-inference-service] ownerReferences: [e2e-llm-inference-service] - apiVersion: apps/v1 [e2e-llm-inference-service] kind: Deployment [e2e-llm-inference-service] name: custom-route-timeout-test-kserve [e2e-llm-inference-service] uid: e9876375-48ff-4fa2-88dd-23a9c4ed5714 [e2e-llm-inference-service] controller: true [e2e-llm-inference-service] blockOwnerDeletion: true [e2e-llm-inference-service] managedFields: [e2e-llm-inference-service] - manager: kube-controller-manager [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] apiVersion: apps/v1 [e2e-llm-inference-service] time: '2026-07-07T14:48:57Z' [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:metadata: [e2e-llm-inference-service] f:annotations: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:deployment.kubernetes.io/desired-replicas: {} [e2e-llm-inference-service] f:deployment.kubernetes.io/max-replicas: {} [e2e-llm-inference-service] f:deployment.kubernetes.io/revision: {} [e2e-llm-inference-service] f:labels: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:app.kubernetes.io/component: {} [e2e-llm-inference-service] f:app.kubernetes.io/name: {} [e2e-llm-inference-service] f:app.kubernetes.io/part-of: {} [e2e-llm-inference-service] f:kserve.io/component: {} [e2e-llm-inference-service] f:llm-d.ai/role: {} [e2e-llm-inference-service] f:pod-template-hash: {} [e2e-llm-inference-service] f:ownerReferences: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"uid":"e9876375-48ff-4fa2-88dd-23a9c4ed5714"}: {} [e2e-llm-inference-service] f:spec: [e2e-llm-inference-service] f:replicas: {} [e2e-llm-inference-service] f:selector: {} [e2e-llm-inference-service] f:template: [e2e-llm-inference-service] f:metadata: [e2e-llm-inference-service] f:labels: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:app.kubernetes.io/component: {} [e2e-llm-inference-service] f:app.kubernetes.io/name: {} [e2e-llm-inference-service] f:app.kubernetes.io/part-of: {} [e2e-llm-inference-service] f:kserve.io/component: {} [e2e-llm-inference-service] f:llm-d.ai/role: {} [e2e-llm-inference-service] f:pod-template-hash: {} [e2e-llm-inference-service] f:spec: [e2e-llm-inference-service] f:containers: [e2e-llm-inference-service] k:{"name":"main"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:command: {} [e2e-llm-inference-service] f:env: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"name":"HF_HUB_CACHE"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"HOME"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"TORCHINDUCTOR_CACHE_DIR"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"USER"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"VLLM_CPU_KVCACHE_SPACE"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"VLLM_ENABLE_V1_MULTIPROCESSING"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"VLLM_LOGGING_LEVEL"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] f:image: {} [e2e-llm-inference-service] f:imagePullPolicy: {} [e2e-llm-inference-service] f:lifecycle: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:preStop: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:exec: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:command: {} [e2e-llm-inference-service] f:livenessProbe: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:failureThreshold: {} [e2e-llm-inference-service] f:httpGet: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:path: {} [e2e-llm-inference-service] f:port: {} [e2e-llm-inference-service] f:scheme: {} [e2e-llm-inference-service] f:periodSeconds: {} [e2e-llm-inference-service] f:successThreshold: {} [e2e-llm-inference-service] f:timeoutSeconds: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:ports: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"containerPort":8000,"protocol":"TCP"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:containerPort: {} [e2e-llm-inference-service] f:protocol: {} [e2e-llm-inference-service] f:readinessProbe: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:failureThreshold: {} [e2e-llm-inference-service] f:httpGet: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:path: {} [e2e-llm-inference-service] f:port: {} [e2e-llm-inference-service] f:scheme: {} [e2e-llm-inference-service] f:periodSeconds: {} [e2e-llm-inference-service] f:successThreshold: {} [e2e-llm-inference-service] f:timeoutSeconds: {} [e2e-llm-inference-service] f:resources: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:limits: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:cpu: {} [e2e-llm-inference-service] f:memory: {} [e2e-llm-inference-service] f:requests: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:cpu: {} [e2e-llm-inference-service] f:memory: {} [e2e-llm-inference-service] f:securityContext: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:allowPrivilegeEscalation: {} [e2e-llm-inference-service] f:capabilities: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:drop: {} [e2e-llm-inference-service] f:readOnlyRootFilesystem: {} [e2e-llm-inference-service] f:runAsNonRoot: {} [e2e-llm-inference-service] f:seccompProfile: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:type: {} [e2e-llm-inference-service] f:startupProbe: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:failureThreshold: {} [e2e-llm-inference-service] f:httpGet: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:path: {} [e2e-llm-inference-service] f:port: {} [e2e-llm-inference-service] f:scheme: {} [e2e-llm-inference-service] f:periodSeconds: {} [e2e-llm-inference-service] f:successThreshold: {} [e2e-llm-inference-service] f:timeoutSeconds: {} [e2e-llm-inference-service] f:terminationMessagePath: {} [e2e-llm-inference-service] f:terminationMessagePolicy: {} [e2e-llm-inference-service] f:volumeMounts: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"mountPath":"/dev/shm"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:mountPath: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] k:{"mountPath":"/home"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:mountPath: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] k:{"mountPath":"/mnt/models"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:mountPath: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:readOnly: {} [e2e-llm-inference-service] k:{"mountPath":"/models"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:mountPath: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] k:{"mountPath":"/tmp"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:mountPath: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] k:{"mountPath":"/var/run/kserve/tls"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:mountPath: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:readOnly: {} [e2e-llm-inference-service] f:dnsPolicy: {} [e2e-llm-inference-service] f:initContainers: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"name":"storage-initializer"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:args: {} [e2e-llm-inference-service] f:env: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"name":"AWS_ACCESS_KEY_ID"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:valueFrom: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:secretKeyRef: {} [e2e-llm-inference-service] k:{"name":"AWS_CA_BUNDLE"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"AWS_CA_BUNDLE_CONFIGMAP"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"AWS_ENDPOINT_URL"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"AWS_SECRET_ACCESS_KEY"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:valueFrom: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:secretKeyRef: {} [e2e-llm-inference-service] k:{"name":"HF_HUB_ENABLE_HF_TRANSFER"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"HF_XET_HIGH_PERFORMANCE"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"HF_XET_NUM_CONCURRENT_RANGE_GETS"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"S3_ENDPOINT"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"S3_USE_HTTPS"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"S3_VERIFY_SSL"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] f:image: {} [e2e-llm-inference-service] f:imagePullPolicy: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:resources: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:limits: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:cpu: {} [e2e-llm-inference-service] f:memory: {} [e2e-llm-inference-service] f:requests: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:cpu: {} [e2e-llm-inference-service] f:memory: {} [e2e-llm-inference-service] f:terminationMessagePath: {} [e2e-llm-inference-service] f:terminationMessagePolicy: {} [e2e-llm-inference-service] f:volumeMounts: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"mountPath":"/mnt/models"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:mountPath: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:restartPolicy: {} [e2e-llm-inference-service] f:schedulerName: {} [e2e-llm-inference-service] f:securityContext: {} [e2e-llm-inference-service] f:terminationGracePeriodSeconds: {} [e2e-llm-inference-service] f:volumes: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"name":"dshm"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:emptyDir: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:medium: {} [e2e-llm-inference-service] f:sizeLimit: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] k:{"name":"home"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:emptyDir: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] k:{"name":"kserve-provision-location"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:emptyDir: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] k:{"name":"model-cache"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:emptyDir: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] k:{"name":"tls-certs"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:secret: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:defaultMode: {} [e2e-llm-inference-service] f:secretName: {} [e2e-llm-inference-service] k:{"name":"tmp-dir"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:emptyDir: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] - manager: kube-controller-manager [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] apiVersion: apps/v1 [e2e-llm-inference-service] time: '2026-07-07T14:50:57Z' [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:status: [e2e-llm-inference-service] f:availableReplicas: {} [e2e-llm-inference-service] f:fullyLabeledReplicas: {} [e2e-llm-inference-service] f:observedGeneration: {} [e2e-llm-inference-service] f:readyReplicas: {} [e2e-llm-inference-service] f:replicas: {} [e2e-llm-inference-service] subresource: status [e2e-llm-inference-service] spec: [e2e-llm-inference-service] replicas: 1 [e2e-llm-inference-service] selector: [e2e-llm-inference-service] matchLabels: [e2e-llm-inference-service] app.kubernetes.io/component: llminferenceservice-workload [e2e-llm-inference-service] app.kubernetes.io/name: custom-route-timeout-test [e2e-llm-inference-service] app.kubernetes.io/part-of: llminferenceservice [e2e-llm-inference-service] kserve.io/component: workload [e2e-llm-inference-service] llm-d.ai/role: both [e2e-llm-inference-service] pod-template-hash: 7f56b8c469 [e2e-llm-inference-service] template: [e2e-llm-inference-service] metadata: [e2e-llm-inference-service] labels: [e2e-llm-inference-service] app.kubernetes.io/component: llminferenceservice-workload [e2e-llm-inference-service] app.kubernetes.io/name: custom-route-timeout-test [e2e-llm-inference-service] app.kubernetes.io/part-of: llminferenceservice [e2e-llm-inference-service] kserve.io/component: workload [e2e-llm-inference-service] llm-d.ai/role: both [e2e-llm-inference-service] pod-template-hash: 7f56b8c469 [e2e-llm-inference-service] spec: [e2e-llm-inference-service] volumes: [e2e-llm-inference-service] - name: home [e2e-llm-inference-service] emptyDir: {} [e2e-llm-inference-service] - name: dshm [e2e-llm-inference-service] emptyDir: [e2e-llm-inference-service] medium: Memory [e2e-llm-inference-service] sizeLimit: 1Gi [e2e-llm-inference-service] - name: model-cache [e2e-llm-inference-service] emptyDir: {} [e2e-llm-inference-service] - name: tmp-dir [e2e-llm-inference-service] emptyDir: {} [e2e-llm-inference-service] - name: tls-certs [e2e-llm-inference-service] secret: [e2e-llm-inference-service] secretName: custom-route-timeout-test-kserve-self-signed-certs [e2e-llm-inference-service] defaultMode: 420 [e2e-llm-inference-service] - name: kserve-provision-location [e2e-llm-inference-service] emptyDir: {} [e2e-llm-inference-service] initContainers: [e2e-llm-inference-service] - name: storage-initializer [e2e-llm-inference-service] image: quay.io/opendatahub/kserve-storage-initializer@sha256:13ad765c0f323200e1db7b45c99661287e65c5158d6b54b16139829636d20be4 [e2e-llm-inference-service] args: [e2e-llm-inference-service] - hf://facebook/opt-125m [e2e-llm-inference-service] - /mnt/models [e2e-llm-inference-service] env: [e2e-llm-inference-service] - name: AWS_ACCESS_KEY_ID [e2e-llm-inference-service] valueFrom: [e2e-llm-inference-service] secretKeyRef: [e2e-llm-inference-service] name: seaweedfs-s3-creds [e2e-llm-inference-service] key: AWS_ACCESS_KEY_ID [e2e-llm-inference-service] - name: AWS_SECRET_ACCESS_KEY [e2e-llm-inference-service] valueFrom: [e2e-llm-inference-service] secretKeyRef: [e2e-llm-inference-service] name: seaweedfs-s3-creds [e2e-llm-inference-service] key: AWS_SECRET_ACCESS_KEY [e2e-llm-inference-service] - name: S3_USE_HTTPS [e2e-llm-inference-service] value: '0' [e2e-llm-inference-service] - name: S3_ENDPOINT [e2e-llm-inference-service] value: s3-service.kserve:8333 [e2e-llm-inference-service] - name: AWS_ENDPOINT_URL [e2e-llm-inference-service] value: http://s3-service.kserve:8333 [e2e-llm-inference-service] - name: S3_VERIFY_SSL [e2e-llm-inference-service] value: '0' [e2e-llm-inference-service] - name: AWS_CA_BUNDLE [e2e-llm-inference-service] value: /etc/ssl/custom-certs/cabundle.crt [e2e-llm-inference-service] - name: AWS_CA_BUNDLE_CONFIGMAP [e2e-llm-inference-service] value: odh-kserve-custom-ca-bundle [e2e-llm-inference-service] - name: HF_HUB_ENABLE_HF_TRANSFER [e2e-llm-inference-service] value: '1' [e2e-llm-inference-service] - name: HF_XET_HIGH_PERFORMANCE [e2e-llm-inference-service] value: '1' [e2e-llm-inference-service] - name: HF_XET_NUM_CONCURRENT_RANGE_GETS [e2e-llm-inference-service] value: '8' [e2e-llm-inference-service] resources: [e2e-llm-inference-service] limits: [e2e-llm-inference-service] cpu: '1' [e2e-llm-inference-service] memory: 24Gi [e2e-llm-inference-service] requests: [e2e-llm-inference-service] cpu: 100m [e2e-llm-inference-service] memory: 100Mi [e2e-llm-inference-service] volumeMounts: [e2e-llm-inference-service] - name: kserve-provision-location [e2e-llm-inference-service] mountPath: /mnt/models [e2e-llm-inference-service] terminationMessagePath: /dev/termination-log [e2e-llm-inference-service] terminationMessagePolicy: FallbackToLogsOnError [e2e-llm-inference-service] imagePullPolicy: IfNotPresent [e2e-llm-inference-service] containers: [e2e-llm-inference-service] - name: main [e2e-llm-inference-service] image: public.ecr.aws/q9t5s3a7/vllm-cpu-release-repo:v0.19.0 [e2e-llm-inference-service] command: [e2e-llm-inference-service] - /bin/bash [e2e-llm-inference-service] - -c [e2e-llm-inference-service] - "if [ -f /etc/profile.d/ibm-aiu-setup.sh ]; then\n source /etc/profile.d/ibm-aiu-setup.sh\n\ [e2e-llm-inference-service] fi\n\nif [ \"$KSERVE_INFER_ROCE\" = \"true\" ]; then\n echo \"Trying to\ [e2e-llm-inference-service] \ infer RoCE configs ... \"\n grep -H . /sys/class/infiniband/*/ports/*/gids/*\ [e2e-llm-inference-service] \ 2>/dev/null\n grep -H . /sys/class/infiniband/*/ports/*/gid_attrs/types/*\ [e2e-llm-inference-service] \ 2>/dev/null\n\n cat /proc/driver/nvidia/params\n\n KSERVE_INFER_IB_GID_INDEX_GREP=${KSERVE_INFER_IB_GID_INDEX_GREP:-\"\ [e2e-llm-inference-service] RoCE v2\"}\n\n echo \"[Infer RoCE] Discovering active HCAs ...\"\n active_hcas=()\n\ [e2e-llm-inference-service] \ # Loop through all mlx5 devices found in sysfs\n for hca_dir in /sys/class/infiniband/mlx5_*;\ [e2e-llm-inference-service] \ do\n # Ensure it's a directory before proceeding\n if [ -d \"\ [e2e-llm-inference-service] $hca_dir\" ]; then\n hca_name=$(basename \"$hca_dir\")\n \ [e2e-llm-inference-service] \ port_state_file=\"$hca_dir/ports/1/state\" # Assume port 1\n \ [e2e-llm-inference-service] \ type_file=\"$hca_dir/ports/1/gid_attrs/types/*\"\n\n echo\ [e2e-llm-inference-service] \ \"[Infer RoCE] Check if the port state file ${port_state_file} exists\ [e2e-llm-inference-service] \ and contains 'ACTIVE'\"\n if [ -f \"$port_state_file\" ] && grep\ [e2e-llm-inference-service] \ -q \"ACTIVE\" \"$port_state_file\" && grep -q \"${KSERVE_INFER_IB_GID_INDEX_GREP}\"\ [e2e-llm-inference-service] \ ${type_file} 2>/dev/null; then\n echo \"[Infer RoCE] Found\ [e2e-llm-inference-service] \ active HCA: $hca_name\"\n active_hcas+=(\"$hca_name\")\n\ [e2e-llm-inference-service] \ else\n echo \"[Infer RoCE] Skipping inactive or\ [e2e-llm-inference-service] \ down HCA: $hca_name\"\n fi\n fi\n done\n\n # Check if\ [e2e-llm-inference-service] \ we found any active HCAs\n if [ ${#active_hcas[@]} -gt 0 ]; then\n \ [e2e-llm-inference-service] \ # Join the array elements with a comma\n hca_port_pairs=()\n \ [e2e-llm-inference-service] \ for hca in \"${active_hcas[@]}\"; do\n hca_port_pairs+=(\"\ [e2e-llm-inference-service] ${hca}:1\")\n done\n\n active_hca_list=$(IFS=,; echo \"${active_hcas[*]}\"\ [e2e-llm-inference-service] )\n hca_port_pairs_list=$(IFS=,; echo \"${hca_port_pairs[*]}\")\n \ [e2e-llm-inference-service] \ echo \"[Infer RoCE] Setting active HCAs: ${active_hca_list}\"\n \ [e2e-llm-inference-service] \ export NCCL_IB_HCA=${NCCL_IB_HCA:-${active_hca_list}}\n export\ [e2e-llm-inference-service] \ NVSHMEM_HCA_LIST=${NVSHMEM_HCA_LIST:-${hca_port_pairs_list}}\n export\ [e2e-llm-inference-service] \ UCX_NET_DEVICES=${UCX_NET_DEVICES:-${hca_port_pairs_list}}\n\n echo\ [e2e-llm-inference-service] \ \"[Infer RoCE] NCCL_IB_HCA=${NCCL_IB_HCA}\"\n echo \"[Infer RoCE]\ [e2e-llm-inference-service] \ NVSHMEM_HCA_LIST=${NVSHMEM_HCA_LIST}\"\n echo \"[Infer RoCE] UCX_NET_DEVICES=${UCX_NET_DEVICES}\"\ [e2e-llm-inference-service] \n else\n echo \"[Infer RoCE] WARNING: No active RoCE HCAs found.\ [e2e-llm-inference-service] \ NCCL_IB_HCA will not be set.\"\n fi\n\n if [ ${#active_hcas[@]} -gt\ [e2e-llm-inference-service] \ 0 ]; then\n echo \"[Infer RoCE] Finding GID_INDEX for each active\ [e2e-llm-inference-service] \ HCA (SR-IOV compatible)...\"\n\n # For SR-IOV environments, find\ [e2e-llm-inference-service] \ the most common IPv4 RoCE v2 GID index across all HCAs\n declare\ [e2e-llm-inference-service] \ -A gid_index_count\n declare -A hca_gid_index\n\n for hca_name\ [e2e-llm-inference-service] \ in \"${active_hcas[@]}\"; do\n echo \"[Infer RoCE] Processing\ [e2e-llm-inference-service] \ HCA: ${hca_name}\"\n\n # Find all RoCE v2 IPv4 GIDs for this\ [e2e-llm-inference-service] \ HCA and count by index\n for tpath in /sys/class/infiniband/${hca_name}/ports/1/gid_attrs/types/*;\ [e2e-llm-inference-service] \ do\n if grep -q \"${KSERVE_INFER_IB_GID_INDEX_GREP}\" \"\ [e2e-llm-inference-service] $tpath\" 2>/dev/null; then\n idx=$(basename \"$tpath\"\ [e2e-llm-inference-service] )\n gid_file=\"/sys/class/infiniband/${hca_name}/ports/1/gids/${idx}\"\ [e2e-llm-inference-service] \n # Check for IPv4 GID (contains ffff:)\n \ [e2e-llm-inference-service] \ if [ -f \"$gid_file\" ] && grep -q \"ffff:\" \"$gid_file\"; then\n\ [e2e-llm-inference-service] \ gid_value=$(cat \"$gid_file\" 2>/dev/null || echo\ [e2e-llm-inference-service] \ \"\")\n echo \"[Infer RoCE] Found IPv4 RoCE v2 GID\ [e2e-llm-inference-service] \ for ${hca_name}: index=${idx}, gid=${gid_value}\"\n \ [e2e-llm-inference-service] \ hca_gid_index[\"${hca_name}\"]=\"${idx}\"\n gid_index_count[\"\ [e2e-llm-inference-service] ${idx}\"]=$((${gid_index_count[\"${idx}\"]} + 1))\n \ [e2e-llm-inference-service] \ break # Use first found IPv4 GID per HCA\n fi\n \ [e2e-llm-inference-service] \ fi\n done\n done\n\n # Find the most common\ [e2e-llm-inference-service] \ GID index (most likely to be consistent across nodes)\n best_gid_index=\"\ [e2e-llm-inference-service] \"\n max_count=0\n for idx in \"${!gid_index_count[@]}\"; do\n\ [e2e-llm-inference-service] \ count=${gid_index_count[\"${idx}\"]}\n echo \"[Infer\ [e2e-llm-inference-service] \ RoCE] GID_INDEX ${idx} found on ${count} HCAs\"\n if [ $count\ [e2e-llm-inference-service] \ -gt $max_count ]; then\n max_count=$count\n \ [e2e-llm-inference-service] \ best_gid_index=\"$idx\"\n fi\n done\n\n # Use deterministic\ [e2e-llm-inference-service] \ fallback if tied - prefer index 3 (SR-IOV standard)\n if [ ${#gid_index_count[@]}\ [e2e-llm-inference-service] \ -gt 1 ]; then\n echo \"[Infer RoCE] Multiple GID indices found,\ [e2e-llm-inference-service] \ selecting most common: ${best_gid_index}\"\n # If there's a tie,\ [e2e-llm-inference-service] \ prefer index 3 as it's most common in SR-IOV setups\n if [ -n\ [e2e-llm-inference-service] \ \"${gid_index_count['3']}\" ] && [ \"${gid_index_count['3']}\" -eq \"\ [e2e-llm-inference-service] $max_count\" ]; then\n best_gid_index=\"3\"\n \ [e2e-llm-inference-service] \ echo \"[Infer RoCE] Using deterministic fallback: GID_INDEX=3 (SR-IOV\ [e2e-llm-inference-service] \ standard)\"\n fi\n fi\n\n # Check if GID_INDEX is already\ [e2e-llm-inference-service] \ set via environment variables\n if [ -n \"${NCCL_IB_GID_INDEX}\"\ [e2e-llm-inference-service] \ ]; then\n echo \"[Infer RoCE] Using pre-configured NCCL_IB_GID_INDEX=${NCCL_IB_GID_INDEX}\ [e2e-llm-inference-service] \ from environment\"\n export NVSHMEM_IB_GID_INDEX=${NVSHMEM_IB_GID_INDEX:-$NCCL_IB_GID_INDEX}\n\ [e2e-llm-inference-service] \ export UCX_IB_GID_INDEX=${UCX_IB_GID_INDEX:-$NCCL_IB_GID_INDEX}\n\ [e2e-llm-inference-service] \ echo \"[Infer RoCE] Using pre-configured GID_INDEX=${NCCL_IB_GID_INDEX}\ [e2e-llm-inference-service] \ for NCCL, NVSHMEM, and UCX\"\n elif [ -n \"$best_gid_index\" ]; then\n\ [e2e-llm-inference-service] \ echo \"[Infer RoCE] Selected GID_INDEX: ${best_gid_index} (found\ [e2e-llm-inference-service] \ on ${max_count} HCAs)\"\n\n export NCCL_IB_GID_INDEX=${NCCL_IB_GID_INDEX:-$best_gid_index}\n\ [e2e-llm-inference-service] \ export NVSHMEM_IB_GID_INDEX=${NVSHMEM_IB_GID_INDEX:-$best_gid_index}\n\ [e2e-llm-inference-service] \ export UCX_IB_GID_INDEX=${UCX_IB_GID_INDEX:-$best_gid_index}\n\ [e2e-llm-inference-service] \n echo \"[Infer RoCE] Exported GID_INDEX=${best_gid_index} for\ [e2e-llm-inference-service] \ NCCL, NVSHMEM, and UCX\"\n else\n echo \"[Infer RoCE] ERROR:\ [e2e-llm-inference-service] \ No valid IPv4 ${KSERVE_INFER_IB_GID_INDEX_GREP} GID_INDEX found on any\ [e2e-llm-inference-service] \ HCA.\"\n fi\n else\n echo \"[Infer RoCE] No active HCAs found,\ [e2e-llm-inference-service] \ skipping GID_INDEX inference.\"\n fi\nfi\n\n# --disable-access-log-for-endpoints\ [e2e-llm-inference-service] \ landed in vLLM 0.16.0 (vllm-project/vllm#30011).\n# Older versions still\ [e2e-llm-inference-service] \ need the blanket --disable-uvicorn-access-log.\nACCESS_LOG_ARGS=\"--disable-uvicorn-access-log\"\ [e2e-llm-inference-service] \nVLLM_VERSION=$(vllm --version 2>/dev/null | tail -1 | awk '{print $NF}')\n\ [e2e-llm-inference-service] echo \"[access-log-detect] vllm version='${VLLM_VERSION}'\"\nif [[ \"$VLLM_VERSION\"\ [e2e-llm-inference-service] \ =~ ^[0-9]+\\.[0-9]+ ]] && [ \"$(printf '%s\\n%s\\n' \"0.16.0\" \"${VLLM_VERSION}\"\ [e2e-llm-inference-service] \ | sort -V | head -1)\" = \"0.16.0\" ]; then\n ACCESS_LOG_ARGS=\"--disable-access-log-for-endpoints\ [e2e-llm-inference-service] \ /health,/metrics,/ping\"\nfi\necho \"[access-log-detect] selected ACCESS_LOG_ARGS='${ACCESS_LOG_ARGS}'\"\ [e2e-llm-inference-service] \n\n# --shutdown-timeout landed in vLLM 0.18.0 (vllm-project/vllm#36666).\n\ [e2e-llm-inference-service] SHUTDOWN_TIMEOUT_ARGS=\"\"\nif [[ \"$VLLM_VERSION\" =~ ^[0-9]+\\.[0-9]+\ [e2e-llm-inference-service] \ ]] && [ \"$(printf '%s\\n%s\\n' \"0.18.0\" \"${VLLM_VERSION}\" | sort\ [e2e-llm-inference-service] \ -V | head -1)\" = \"0.18.0\" ]; then\n SHUTDOWN_TIMEOUT_ARGS=\"--shutdown-timeout\ [e2e-llm-inference-service] \ 40\"\nfi\n\neval \"exec vllm serve /mnt/models \\\n --served-model-name\ [e2e-llm-inference-service] \ \"facebook/opt-125m\" \"publishers/kserve-ci-e2e-test/models/facebook/opt-125m\"\ [e2e-llm-inference-service] \ \\\n --port 8000 \\\n ${ACCESS_LOG_ARGS} \\\n ${SHUTDOWN_TIMEOUT_ARGS}\ [e2e-llm-inference-service] \ \\\n --enable-ssl-refresh \\\n --ssl-certfile /var/run/kserve/tls/tls.crt\ [e2e-llm-inference-service] \ \\\n --ssl-keyfile /var/run/kserve/tls/tls.key \\\n ${VLLM_ADDITIONAL_ARGS}\ [e2e-llm-inference-service] \ \\\n $@\"" [e2e-llm-inference-service] - -- [e2e-llm-inference-service] ports: [e2e-llm-inference-service] - containerPort: 8000 [e2e-llm-inference-service] protocol: TCP [e2e-llm-inference-service] env: [e2e-llm-inference-service] - name: HOME [e2e-llm-inference-service] value: /home [e2e-llm-inference-service] - name: VLLM_LOGGING_LEVEL [e2e-llm-inference-service] value: DEBUG [e2e-llm-inference-service] - name: VLLM_CPU_KVCACHE_SPACE [e2e-llm-inference-service] value: '1' [e2e-llm-inference-service] - name: VLLM_ENABLE_V1_MULTIPROCESSING [e2e-llm-inference-service] value: '0' [e2e-llm-inference-service] - name: USER [e2e-llm-inference-service] value: nonroot [e2e-llm-inference-service] - name: TORCHINDUCTOR_CACHE_DIR [e2e-llm-inference-service] value: /tmp/torchinductor-cache [e2e-llm-inference-service] - name: HF_HUB_CACHE [e2e-llm-inference-service] value: /models [e2e-llm-inference-service] resources: [e2e-llm-inference-service] limits: [e2e-llm-inference-service] cpu: '2' [e2e-llm-inference-service] memory: 7Gi [e2e-llm-inference-service] requests: [e2e-llm-inference-service] cpu: 200m [e2e-llm-inference-service] memory: 2Gi [e2e-llm-inference-service] volumeMounts: [e2e-llm-inference-service] - name: home [e2e-llm-inference-service] mountPath: /home [e2e-llm-inference-service] - name: tmp-dir [e2e-llm-inference-service] mountPath: /tmp [e2e-llm-inference-service] - name: dshm [e2e-llm-inference-service] mountPath: /dev/shm [e2e-llm-inference-service] - name: model-cache [e2e-llm-inference-service] mountPath: /models [e2e-llm-inference-service] - name: tls-certs [e2e-llm-inference-service] readOnly: true [e2e-llm-inference-service] mountPath: /var/run/kserve/tls [e2e-llm-inference-service] - name: kserve-provision-location [e2e-llm-inference-service] readOnly: true [e2e-llm-inference-service] mountPath: /mnt/models [e2e-llm-inference-service] livenessProbe: [e2e-llm-inference-service] httpGet: [e2e-llm-inference-service] path: /health [e2e-llm-inference-service] port: 8000 [e2e-llm-inference-service] scheme: HTTPS [e2e-llm-inference-service] timeoutSeconds: 1 [e2e-llm-inference-service] periodSeconds: 10 [e2e-llm-inference-service] successThreshold: 1 [e2e-llm-inference-service] failureThreshold: 10 [e2e-llm-inference-service] readinessProbe: [e2e-llm-inference-service] httpGet: [e2e-llm-inference-service] path: /health [e2e-llm-inference-service] port: 8000 [e2e-llm-inference-service] scheme: HTTPS [e2e-llm-inference-service] timeoutSeconds: 1 [e2e-llm-inference-service] periodSeconds: 1 [e2e-llm-inference-service] successThreshold: 1 [e2e-llm-inference-service] failureThreshold: 2 [e2e-llm-inference-service] startupProbe: [e2e-llm-inference-service] httpGet: [e2e-llm-inference-service] path: /health [e2e-llm-inference-service] port: 8000 [e2e-llm-inference-service] scheme: HTTPS [e2e-llm-inference-service] timeoutSeconds: 1 [e2e-llm-inference-service] periodSeconds: 10 [e2e-llm-inference-service] successThreshold: 1 [e2e-llm-inference-service] failureThreshold: 60 [e2e-llm-inference-service] lifecycle: [e2e-llm-inference-service] preStop: [e2e-llm-inference-service] exec: [e2e-llm-inference-service] command: [e2e-llm-inference-service] - /bin/sleep [e2e-llm-inference-service] - '15' [e2e-llm-inference-service] terminationMessagePath: /dev/termination-log [e2e-llm-inference-service] terminationMessagePolicy: FallbackToLogsOnError [e2e-llm-inference-service] imagePullPolicy: IfNotPresent [e2e-llm-inference-service] securityContext: [e2e-llm-inference-service] capabilities: [e2e-llm-inference-service] drop: [e2e-llm-inference-service] - ALL [e2e-llm-inference-service] runAsNonRoot: true [e2e-llm-inference-service] readOnlyRootFilesystem: true [e2e-llm-inference-service] allowPrivilegeEscalation: false [e2e-llm-inference-service] seccompProfile: [e2e-llm-inference-service] type: RuntimeDefault [e2e-llm-inference-service] restartPolicy: Always [e2e-llm-inference-service] terminationGracePeriodSeconds: 60 [e2e-llm-inference-service] dnsPolicy: ClusterFirst [e2e-llm-inference-service] securityContext: {} [e2e-llm-inference-service] schedulerName: default-scheduler [e2e-llm-inference-service] status: [e2e-llm-inference-service] replicas: 1 [e2e-llm-inference-service] fullyLabeledReplicas: 1 [e2e-llm-inference-service] readyReplicas: 1 [e2e-llm-inference-service] availableReplicas: 1 [e2e-llm-inference-service] observedGeneration: 1 [e2e-llm-inference-service] apiVersion: apps/v1 [e2e-llm-inference-service] kind: ReplicaSet [e2e-llm-inference-service] [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1254 --- [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1255 metadata: [e2e-llm-inference-service] name: custom-route-timeout-test-kserve-router-scheduler-6578bddf5b [e2e-llm-inference-service] namespace: kserve-ci-e2e-test [e2e-llm-inference-service] uid: 92292d9f-7d6e-42e1-b2be-abdf1d622bd7 [e2e-llm-inference-service] resourceVersion: '49591' [e2e-llm-inference-service] generation: 1 [e2e-llm-inference-service] creationTimestamp: '2026-07-07T14:48:57Z' [e2e-llm-inference-service] labels: [e2e-llm-inference-service] app.kubernetes.io/component: llminferenceservice-router-scheduler [e2e-llm-inference-service] app.kubernetes.io/name: custom-route-timeout-test [e2e-llm-inference-service] app.kubernetes.io/part-of: llminferenceservice [e2e-llm-inference-service] pod-template-hash: 6578bddf5b [e2e-llm-inference-service] annotations: [e2e-llm-inference-service] deployment.kubernetes.io/desired-replicas: '1' [e2e-llm-inference-service] deployment.kubernetes.io/max-replicas: '1' [e2e-llm-inference-service] deployment.kubernetes.io/revision: '1' [e2e-llm-inference-service] ownerReferences: [e2e-llm-inference-service] - apiVersion: apps/v1 [e2e-llm-inference-service] kind: Deployment [e2e-llm-inference-service] name: custom-route-timeout-test-kserve-router-scheduler [e2e-llm-inference-service] uid: eca54665-59f5-44bb-a135-818e809de51c [e2e-llm-inference-service] controller: true [e2e-llm-inference-service] blockOwnerDeletion: true [e2e-llm-inference-service] managedFields: [e2e-llm-inference-service] - manager: kube-controller-manager [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] apiVersion: apps/v1 [e2e-llm-inference-service] time: '2026-07-07T14:48:57Z' [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:metadata: [e2e-llm-inference-service] f:annotations: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:deployment.kubernetes.io/desired-replicas: {} [e2e-llm-inference-service] f:deployment.kubernetes.io/max-replicas: {} [e2e-llm-inference-service] f:deployment.kubernetes.io/revision: {} [e2e-llm-inference-service] f:labels: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:app.kubernetes.io/component: {} [e2e-llm-inference-service] f:app.kubernetes.io/name: {} [e2e-llm-inference-service] f:app.kubernetes.io/part-of: {} [e2e-llm-inference-service] f:pod-template-hash: {} [e2e-llm-inference-service] f:ownerReferences: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"uid":"eca54665-59f5-44bb-a135-818e809de51c"}: {} [e2e-llm-inference-service] f:spec: [e2e-llm-inference-service] f:replicas: {} [e2e-llm-inference-service] f:selector: {} [e2e-llm-inference-service] f:template: [e2e-llm-inference-service] f:metadata: [e2e-llm-inference-service] f:annotations: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:app.kubernetes.io/version: {} [e2e-llm-inference-service] f:certificates.kserve.io/expiration-v2: {} [e2e-llm-inference-service] f:labels: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:app.kubernetes.io/component: {} [e2e-llm-inference-service] f:app.kubernetes.io/name: {} [e2e-llm-inference-service] f:app.kubernetes.io/part-of: {} [e2e-llm-inference-service] f:pod-template-hash: {} [e2e-llm-inference-service] f:spec: [e2e-llm-inference-service] f:containers: [e2e-llm-inference-service] k:{"name":"main"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:args: {} [e2e-llm-inference-service] f:command: {} [e2e-llm-inference-service] f:env: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"name":"SSL_CERT_DIR"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] f:image: {} [e2e-llm-inference-service] f:imagePullPolicy: {} [e2e-llm-inference-service] f:lifecycle: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:preStop: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:exec: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:command: {} [e2e-llm-inference-service] f:livenessProbe: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:failureThreshold: {} [e2e-llm-inference-service] f:grpc: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:port: {} [e2e-llm-inference-service] f:service: {} [e2e-llm-inference-service] f:initialDelaySeconds: {} [e2e-llm-inference-service] f:periodSeconds: {} [e2e-llm-inference-service] f:successThreshold: {} [e2e-llm-inference-service] f:timeoutSeconds: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:ports: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"containerPort":5557,"protocol":"TCP"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:containerPort: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:protocol: {} [e2e-llm-inference-service] k:{"containerPort":9002,"protocol":"TCP"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:containerPort: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:protocol: {} [e2e-llm-inference-service] k:{"containerPort":9003,"protocol":"TCP"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:containerPort: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:protocol: {} [e2e-llm-inference-service] k:{"containerPort":9090,"protocol":"TCP"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:containerPort: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:protocol: {} [e2e-llm-inference-service] f:readinessProbe: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:failureThreshold: {} [e2e-llm-inference-service] f:grpc: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:port: {} [e2e-llm-inference-service] f:service: {} [e2e-llm-inference-service] f:initialDelaySeconds: {} [e2e-llm-inference-service] f:periodSeconds: {} [e2e-llm-inference-service] f:successThreshold: {} [e2e-llm-inference-service] f:timeoutSeconds: {} [e2e-llm-inference-service] f:resources: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:requests: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:cpu: {} [e2e-llm-inference-service] f:memory: {} [e2e-llm-inference-service] f:securityContext: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:allowPrivilegeEscalation: {} [e2e-llm-inference-service] f:capabilities: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:drop: {} [e2e-llm-inference-service] f:readOnlyRootFilesystem: {} [e2e-llm-inference-service] f:runAsNonRoot: {} [e2e-llm-inference-service] f:seccompProfile: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:type: {} [e2e-llm-inference-service] f:terminationMessagePath: {} [e2e-llm-inference-service] f:terminationMessagePolicy: {} [e2e-llm-inference-service] f:volumeMounts: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"mountPath":"/tmp/tokenizer"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:mountPath: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] k:{"mountPath":"/var/run/kserve/tls"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:mountPath: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:readOnly: {} [e2e-llm-inference-service] f:dnsPolicy: {} [e2e-llm-inference-service] f:initContainers: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"name":"storage-initializer"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:args: {} [e2e-llm-inference-service] f:env: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"name":"AWS_ACCESS_KEY_ID"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:valueFrom: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:secretKeyRef: {} [e2e-llm-inference-service] k:{"name":"AWS_CA_BUNDLE"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"AWS_CA_BUNDLE_CONFIGMAP"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"AWS_ENDPOINT_URL"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"AWS_SECRET_ACCESS_KEY"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:valueFrom: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:secretKeyRef: {} [e2e-llm-inference-service] k:{"name":"HF_HUB_ENABLE_HF_TRANSFER"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"HF_XET_HIGH_PERFORMANCE"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"HF_XET_NUM_CONCURRENT_RANGE_GETS"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"S3_ENDPOINT"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"S3_USE_HTTPS"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"S3_VERIFY_SSL"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] k:{"name":"STORAGE_ALLOW_PATTERNS"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] f:image: {} [e2e-llm-inference-service] f:imagePullPolicy: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:resources: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:limits: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:cpu: {} [e2e-llm-inference-service] f:memory: {} [e2e-llm-inference-service] f:requests: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:cpu: {} [e2e-llm-inference-service] f:memory: {} [e2e-llm-inference-service] f:terminationMessagePath: {} [e2e-llm-inference-service] f:terminationMessagePolicy: {} [e2e-llm-inference-service] f:volumeMounts: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"mountPath":"/mnt/models"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:mountPath: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:restartPolicy: {} [e2e-llm-inference-service] f:schedulerName: {} [e2e-llm-inference-service] f:securityContext: {} [e2e-llm-inference-service] f:serviceAccount: {} [e2e-llm-inference-service] f:serviceAccountName: {} [e2e-llm-inference-service] f:terminationGracePeriodSeconds: {} [e2e-llm-inference-service] f:volumes: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"name":"kserve-provision-location"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:emptyDir: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] k:{"name":"tls-certs"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:secret: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:defaultMode: {} [e2e-llm-inference-service] f:secretName: {} [e2e-llm-inference-service] k:{"name":"tokenizer-cache"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:emptyDir: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] k:{"name":"tokenizer-tmp"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:emptyDir: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] k:{"name":"tokenizer-uds"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:emptyDir: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] - manager: kube-controller-manager [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] apiVersion: apps/v1 [e2e-llm-inference-service] time: '2026-07-07T14:49:31Z' [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:status: [e2e-llm-inference-service] f:availableReplicas: {} [e2e-llm-inference-service] f:fullyLabeledReplicas: {} [e2e-llm-inference-service] f:observedGeneration: {} [e2e-llm-inference-service] f:readyReplicas: {} [e2e-llm-inference-service] f:replicas: {} [e2e-llm-inference-service] subresource: status [e2e-llm-inference-service] spec: [e2e-llm-inference-service] replicas: 1 [e2e-llm-inference-service] selector: [e2e-llm-inference-service] matchLabels: [e2e-llm-inference-service] app.kubernetes.io/component: llminferenceservice-router-scheduler [e2e-llm-inference-service] app.kubernetes.io/name: custom-route-timeout-test [e2e-llm-inference-service] app.kubernetes.io/part-of: llminferenceservice [e2e-llm-inference-service] pod-template-hash: 6578bddf5b [e2e-llm-inference-service] template: [e2e-llm-inference-service] metadata: [e2e-llm-inference-service] labels: [e2e-llm-inference-service] app.kubernetes.io/component: llminferenceservice-router-scheduler [e2e-llm-inference-service] app.kubernetes.io/name: custom-route-timeout-test [e2e-llm-inference-service] app.kubernetes.io/part-of: llminferenceservice [e2e-llm-inference-service] pod-template-hash: 6578bddf5b [e2e-llm-inference-service] annotations: [e2e-llm-inference-service] app.kubernetes.io/version: 0.9.0 [e2e-llm-inference-service] certificates.kserve.io/expiration-v2: 'true' [e2e-llm-inference-service] spec: [e2e-llm-inference-service] volumes: [e2e-llm-inference-service] - name: tls-certs [e2e-llm-inference-service] secret: [e2e-llm-inference-service] secretName: custom-route-timeout-test-kserve-self-signed-certs [e2e-llm-inference-service] defaultMode: 420 [e2e-llm-inference-service] - name: tokenizer-uds [e2e-llm-inference-service] emptyDir: {} [e2e-llm-inference-service] - name: tokenizer-tmp [e2e-llm-inference-service] emptyDir: {} [e2e-llm-inference-service] - name: tokenizer-cache [e2e-llm-inference-service] emptyDir: {} [e2e-llm-inference-service] - name: kserve-provision-location [e2e-llm-inference-service] emptyDir: {} [e2e-llm-inference-service] initContainers: [e2e-llm-inference-service] - name: storage-initializer [e2e-llm-inference-service] image: quay.io/opendatahub/kserve-storage-initializer@sha256:13ad765c0f323200e1db7b45c99661287e65c5158d6b54b16139829636d20be4 [e2e-llm-inference-service] args: [e2e-llm-inference-service] - hf://facebook/opt-125m [e2e-llm-inference-service] - /mnt/models [e2e-llm-inference-service] env: [e2e-llm-inference-service] - name: AWS_ACCESS_KEY_ID [e2e-llm-inference-service] valueFrom: [e2e-llm-inference-service] secretKeyRef: [e2e-llm-inference-service] name: seaweedfs-s3-creds [e2e-llm-inference-service] key: AWS_ACCESS_KEY_ID [e2e-llm-inference-service] - name: AWS_SECRET_ACCESS_KEY [e2e-llm-inference-service] valueFrom: [e2e-llm-inference-service] secretKeyRef: [e2e-llm-inference-service] name: seaweedfs-s3-creds [e2e-llm-inference-service] key: AWS_SECRET_ACCESS_KEY [e2e-llm-inference-service] - name: S3_USE_HTTPS [e2e-llm-inference-service] value: '0' [e2e-llm-inference-service] - name: S3_ENDPOINT [e2e-llm-inference-service] value: s3-service.kserve:8333 [e2e-llm-inference-service] - name: AWS_ENDPOINT_URL [e2e-llm-inference-service] value: http://s3-service.kserve:8333 [e2e-llm-inference-service] - name: S3_VERIFY_SSL [e2e-llm-inference-service] value: '0' [e2e-llm-inference-service] - name: AWS_CA_BUNDLE [e2e-llm-inference-service] value: /etc/ssl/custom-certs/cabundle.crt [e2e-llm-inference-service] - name: AWS_CA_BUNDLE_CONFIGMAP [e2e-llm-inference-service] value: odh-kserve-custom-ca-bundle [e2e-llm-inference-service] - name: HF_HUB_ENABLE_HF_TRANSFER [e2e-llm-inference-service] value: '1' [e2e-llm-inference-service] - name: HF_XET_HIGH_PERFORMANCE [e2e-llm-inference-service] value: '1' [e2e-llm-inference-service] - name: HF_XET_NUM_CONCURRENT_RANGE_GETS [e2e-llm-inference-service] value: '8' [e2e-llm-inference-service] - name: STORAGE_ALLOW_PATTERNS [e2e-llm-inference-service] value: '["tokenizer.json", "tokenizer_config.json", "special_tokens_map.json", [e2e-llm-inference-service] "vocab.json", "merges.txt", "config.json", "generation_config.json"]' [e2e-llm-inference-service] resources: [e2e-llm-inference-service] limits: [e2e-llm-inference-service] cpu: '1' [e2e-llm-inference-service] memory: 24Gi [e2e-llm-inference-service] requests: [e2e-llm-inference-service] cpu: 100m [e2e-llm-inference-service] memory: 100Mi [e2e-llm-inference-service] volumeMounts: [e2e-llm-inference-service] - name: kserve-provision-location [e2e-llm-inference-service] mountPath: /mnt/models [e2e-llm-inference-service] terminationMessagePath: /dev/termination-log [e2e-llm-inference-service] terminationMessagePolicy: FallbackToLogsOnError [e2e-llm-inference-service] imagePullPolicy: IfNotPresent [e2e-llm-inference-service] containers: [e2e-llm-inference-service] - name: main [e2e-llm-inference-service] image: ghcr.io/llm-d/llm-d-router-endpoint-picker:v0.9.0-rc.2 [e2e-llm-inference-service] command: [e2e-llm-inference-service] - /app/epp [e2e-llm-inference-service] - --pool-name [e2e-llm-inference-service] - custom-route-timeout-test-inference-pool [e2e-llm-inference-service] - --pool-namespace [e2e-llm-inference-service] - kserve-ci-e2e-test [e2e-llm-inference-service] - --zap-encoder [e2e-llm-inference-service] - json [e2e-llm-inference-service] - --grpc-port [e2e-llm-inference-service] - '9002' [e2e-llm-inference-service] - --grpc-health-port [e2e-llm-inference-service] - '9003' [e2e-llm-inference-service] - --enable-cert-reload=true [e2e-llm-inference-service] - --secure-serving=true [e2e-llm-inference-service] - --model-server-metrics-scheme=https [e2e-llm-inference-service] - --cert-path=/var/run/kserve/tls [e2e-llm-inference-service] args: [e2e-llm-inference-service] - --config-text [e2e-llm-inference-service] - "apiVersion: inference.networking.x-k8s.io/v1alpha1\nkind: EndpointPickerConfig\n\ [e2e-llm-inference-service] plugins:\n- type: single-profile-handler\n- type: queue-scorer\n- type:\ [e2e-llm-inference-service] \ prefix-cache-scorer\n- type: max-score-picker\nschedulingProfiles:\n-\ [e2e-llm-inference-service] \ name: default\n plugins:\n - pluginRef: queue-scorer\n weight: 2\n\ [e2e-llm-inference-service] \ - pluginRef: prefix-cache-scorer\n weight: 3\n - pluginRef: max-score-picker\n" [e2e-llm-inference-service] ports: [e2e-llm-inference-service] - name: grpc [e2e-llm-inference-service] containerPort: 9002 [e2e-llm-inference-service] protocol: TCP [e2e-llm-inference-service] - name: grpc-health [e2e-llm-inference-service] containerPort: 9003 [e2e-llm-inference-service] protocol: TCP [e2e-llm-inference-service] - name: metrics [e2e-llm-inference-service] containerPort: 9090 [e2e-llm-inference-service] protocol: TCP [e2e-llm-inference-service] - name: zmq [e2e-llm-inference-service] containerPort: 5557 [e2e-llm-inference-service] protocol: TCP [e2e-llm-inference-service] env: [e2e-llm-inference-service] - name: SSL_CERT_DIR [e2e-llm-inference-service] value: /var/run/kserve/tls:/var/run/secrets/kubernetes.io/serviceaccount:/etc/pki/tls/certs [e2e-llm-inference-service] resources: [e2e-llm-inference-service] requests: [e2e-llm-inference-service] cpu: 256m [e2e-llm-inference-service] memory: 500Mi [e2e-llm-inference-service] volumeMounts: [e2e-llm-inference-service] - name: tls-certs [e2e-llm-inference-service] readOnly: true [e2e-llm-inference-service] mountPath: /var/run/kserve/tls [e2e-llm-inference-service] - name: tokenizer-uds [e2e-llm-inference-service] mountPath: /tmp/tokenizer [e2e-llm-inference-service] livenessProbe: [e2e-llm-inference-service] grpc: [e2e-llm-inference-service] port: 9003 [e2e-llm-inference-service] service: liveness [e2e-llm-inference-service] initialDelaySeconds: 5 [e2e-llm-inference-service] timeoutSeconds: 1 [e2e-llm-inference-service] periodSeconds: 10 [e2e-llm-inference-service] successThreshold: 1 [e2e-llm-inference-service] failureThreshold: 3 [e2e-llm-inference-service] readinessProbe: [e2e-llm-inference-service] grpc: [e2e-llm-inference-service] port: 9003 [e2e-llm-inference-service] service: readiness [e2e-llm-inference-service] initialDelaySeconds: 30 [e2e-llm-inference-service] timeoutSeconds: 1 [e2e-llm-inference-service] periodSeconds: 10 [e2e-llm-inference-service] successThreshold: 1 [e2e-llm-inference-service] failureThreshold: 3 [e2e-llm-inference-service] lifecycle: [e2e-llm-inference-service] preStop: [e2e-llm-inference-service] exec: [e2e-llm-inference-service] command: [e2e-llm-inference-service] - /bin/sleep [e2e-llm-inference-service] - '15' [e2e-llm-inference-service] terminationMessagePath: /dev/termination-log [e2e-llm-inference-service] terminationMessagePolicy: FallbackToLogsOnError [e2e-llm-inference-service] imagePullPolicy: IfNotPresent [e2e-llm-inference-service] securityContext: [e2e-llm-inference-service] capabilities: [e2e-llm-inference-service] drop: [e2e-llm-inference-service] - ALL [e2e-llm-inference-service] runAsNonRoot: true [e2e-llm-inference-service] readOnlyRootFilesystem: true [e2e-llm-inference-service] allowPrivilegeEscalation: false [e2e-llm-inference-service] seccompProfile: [e2e-llm-inference-service] type: RuntimeDefault [e2e-llm-inference-service] restartPolicy: Always [e2e-llm-inference-service] terminationGracePeriodSeconds: 60 [e2e-llm-inference-service] dnsPolicy: ClusterFirst [e2e-llm-inference-service] serviceAccountName: custom-route-timeout-test-epp-sa [e2e-llm-inference-service] serviceAccount: custom-route-timeout-test-epp-sa [e2e-llm-inference-service] securityContext: {} [e2e-llm-inference-service] schedulerName: default-scheduler [e2e-llm-inference-service] status: [e2e-llm-inference-service] replicas: 1 [e2e-llm-inference-service] fullyLabeledReplicas: 1 [e2e-llm-inference-service] readyReplicas: 1 [e2e-llm-inference-service] availableReplicas: 1 [e2e-llm-inference-service] observedGeneration: 1 [e2e-llm-inference-service] apiVersion: apps/v1 [e2e-llm-inference-service] kind: ReplicaSet [e2e-llm-inference-service] [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1254 --- [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1255 metadata: [e2e-llm-inference-service] name: custom-route-timeout-test-epp-rb [e2e-llm-inference-service] namespace: kserve-ci-e2e-test [e2e-llm-inference-service] uid: ce20d641-d03b-4916-9799-6413f17c84d2 [e2e-llm-inference-service] resourceVersion: '48986' [e2e-llm-inference-service] creationTimestamp: '2026-07-07T14:48:57Z' [e2e-llm-inference-service] labels: [e2e-llm-inference-service] app.kubernetes.io/component: llminferenceservice-router-scheduler [e2e-llm-inference-service] app.kubernetes.io/name: custom-route-timeout-test [e2e-llm-inference-service] app.kubernetes.io/part-of: llminferenceservice [e2e-llm-inference-service] ownerReferences: [e2e-llm-inference-service] - apiVersion: serving.kserve.io/v1alpha2 [e2e-llm-inference-service] kind: LLMInferenceService [e2e-llm-inference-service] name: custom-route-timeout-test [e2e-llm-inference-service] uid: 986f2d2c-a6d1-463d-9850-3aad6459eace [e2e-llm-inference-service] controller: true [e2e-llm-inference-service] blockOwnerDeletion: true [e2e-llm-inference-service] managedFields: [e2e-llm-inference-service] - manager: manager [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] apiVersion: rbac.authorization.k8s.io/v1 [e2e-llm-inference-service] time: '2026-07-07T14:48:57Z' [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:metadata: [e2e-llm-inference-service] f:labels: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:app.kubernetes.io/component: {} [e2e-llm-inference-service] f:app.kubernetes.io/name: {} [e2e-llm-inference-service] f:app.kubernetes.io/part-of: {} [e2e-llm-inference-service] f:ownerReferences: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"uid":"986f2d2c-a6d1-463d-9850-3aad6459eace"}: {} [e2e-llm-inference-service] f:roleRef: {} [e2e-llm-inference-service] f:subjects: {} [e2e-llm-inference-service] subjects: [e2e-llm-inference-service] - kind: ServiceAccount [e2e-llm-inference-service] name: custom-route-timeout-test-epp-sa [e2e-llm-inference-service] namespace: kserve-ci-e2e-test [e2e-llm-inference-service] roleRef: [e2e-llm-inference-service] apiGroup: rbac.authorization.k8s.io [e2e-llm-inference-service] kind: Role [e2e-llm-inference-service] name: custom-route-timeout-test-epp-role [e2e-llm-inference-service] apiVersion: rbac.authorization.k8s.io/v1 [e2e-llm-inference-service] kind: RoleBinding [e2e-llm-inference-service] [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1254 --- [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1255 metadata: [e2e-llm-inference-service] name: custom-route-timeout-test-epp-role [e2e-llm-inference-service] namespace: kserve-ci-e2e-test [e2e-llm-inference-service] uid: 849b4bc3-0d16-4af4-acf7-dca0f6b0e70e [e2e-llm-inference-service] resourceVersion: '48981' [e2e-llm-inference-service] creationTimestamp: '2026-07-07T14:48:57Z' [e2e-llm-inference-service] labels: [e2e-llm-inference-service] app.kubernetes.io/component: llminferenceservice-router-scheduler [e2e-llm-inference-service] app.kubernetes.io/name: custom-route-timeout-test [e2e-llm-inference-service] app.kubernetes.io/part-of: llminferenceservice [e2e-llm-inference-service] ownerReferences: [e2e-llm-inference-service] - apiVersion: serving.kserve.io/v1alpha2 [e2e-llm-inference-service] kind: LLMInferenceService [e2e-llm-inference-service] name: custom-route-timeout-test [e2e-llm-inference-service] uid: 986f2d2c-a6d1-463d-9850-3aad6459eace [e2e-llm-inference-service] controller: true [e2e-llm-inference-service] blockOwnerDeletion: true [e2e-llm-inference-service] managedFields: [e2e-llm-inference-service] - manager: manager [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] apiVersion: rbac.authorization.k8s.io/v1 [e2e-llm-inference-service] time: '2026-07-07T14:48:57Z' [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:metadata: [e2e-llm-inference-service] f:labels: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:app.kubernetes.io/component: {} [e2e-llm-inference-service] f:app.kubernetes.io/name: {} [e2e-llm-inference-service] f:app.kubernetes.io/part-of: {} [e2e-llm-inference-service] f:ownerReferences: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"uid":"986f2d2c-a6d1-463d-9850-3aad6459eace"}: {} [e2e-llm-inference-service] f:rules: {} [e2e-llm-inference-service] rules: [e2e-llm-inference-service] - verbs: [e2e-llm-inference-service] - get [e2e-llm-inference-service] - list [e2e-llm-inference-service] - watch [e2e-llm-inference-service] apiGroups: [e2e-llm-inference-service] - '' [e2e-llm-inference-service] resources: [e2e-llm-inference-service] - pods [e2e-llm-inference-service] - verbs: [e2e-llm-inference-service] - get [e2e-llm-inference-service] - list [e2e-llm-inference-service] - watch [e2e-llm-inference-service] apiGroups: [e2e-llm-inference-service] - inference.networking.k8s.io [e2e-llm-inference-service] - inference.networking.x-k8s.io [e2e-llm-inference-service] resources: [e2e-llm-inference-service] - inferencepools [e2e-llm-inference-service] - inferenceobjectives [e2e-llm-inference-service] - inferencemodels [e2e-llm-inference-service] - verbs: [e2e-llm-inference-service] - get [e2e-llm-inference-service] - list [e2e-llm-inference-service] - watch [e2e-llm-inference-service] apiGroups: [e2e-llm-inference-service] - inference.networking.x-k8s.io [e2e-llm-inference-service] resources: [e2e-llm-inference-service] - inferencemodelrewrites [e2e-llm-inference-service] - inferencepoolimports [e2e-llm-inference-service] - verbs: [e2e-llm-inference-service] - get [e2e-llm-inference-service] - list [e2e-llm-inference-service] - watch [e2e-llm-inference-service] apiGroups: [e2e-llm-inference-service] - discovery.k8s.io [e2e-llm-inference-service] resources: [e2e-llm-inference-service] - endpointslices [e2e-llm-inference-service] - verbs: [e2e-llm-inference-service] - get [e2e-llm-inference-service] - list [e2e-llm-inference-service] - watch [e2e-llm-inference-service] - create [e2e-llm-inference-service] - update [e2e-llm-inference-service] - patch [e2e-llm-inference-service] - delete [e2e-llm-inference-service] apiGroups: [e2e-llm-inference-service] - coordination.k8s.io [e2e-llm-inference-service] resources: [e2e-llm-inference-service] - leases [e2e-llm-inference-service] apiVersion: rbac.authorization.k8s.io/v1 [e2e-llm-inference-service] kind: Role [e2e-llm-inference-service] [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1254 --- [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1255 metadata: [e2e-llm-inference-service] name: custom-route-timeout-test-epp-service-sw2db [e2e-llm-inference-service] generateName: custom-route-timeout-test-epp-service- [e2e-llm-inference-service] namespace: kserve-ci-e2e-test [e2e-llm-inference-service] uid: 6545ee00-28ef-48ca-83cc-d69b45aa8017 [e2e-llm-inference-service] resourceVersion: '49589' [e2e-llm-inference-service] generation: 3 [e2e-llm-inference-service] creationTimestamp: '2026-07-07T14:48:57Z' [e2e-llm-inference-service] labels: [e2e-llm-inference-service] app.kubernetes.io/component: llminferenceservice-router-scheduler [e2e-llm-inference-service] app.kubernetes.io/name: custom-route-timeout-test [e2e-llm-inference-service] app.kubernetes.io/part-of: llminferenceservice [e2e-llm-inference-service] endpointslice.kubernetes.io/managed-by: endpointslice-controller.k8s.io [e2e-llm-inference-service] kubernetes.io/service-name: custom-route-timeout-test-epp-service [e2e-llm-inference-service] annotations: [e2e-llm-inference-service] endpoints.kubernetes.io/last-change-trigger-time: '2026-07-07T14:49:31Z' [e2e-llm-inference-service] ownerReferences: [e2e-llm-inference-service] - apiVersion: v1 [e2e-llm-inference-service] kind: Service [e2e-llm-inference-service] name: custom-route-timeout-test-epp-service [e2e-llm-inference-service] uid: 38f4b9e5-98c2-4938-a86e-9aec7ba9c9ea [e2e-llm-inference-service] controller: true [e2e-llm-inference-service] blockOwnerDeletion: true [e2e-llm-inference-service] managedFields: [e2e-llm-inference-service] - manager: kube-controller-manager [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] apiVersion: discovery.k8s.io/v1 [e2e-llm-inference-service] time: '2026-07-07T14:49:31Z' [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:addressType: {} [e2e-llm-inference-service] f:endpoints: {} [e2e-llm-inference-service] f:metadata: [e2e-llm-inference-service] f:annotations: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:endpoints.kubernetes.io/last-change-trigger-time: {} [e2e-llm-inference-service] f:generateName: {} [e2e-llm-inference-service] f:labels: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:app.kubernetes.io/component: {} [e2e-llm-inference-service] f:app.kubernetes.io/name: {} [e2e-llm-inference-service] f:app.kubernetes.io/part-of: {} [e2e-llm-inference-service] f:endpointslice.kubernetes.io/managed-by: {} [e2e-llm-inference-service] f:kubernetes.io/service-name: {} [e2e-llm-inference-service] f:ownerReferences: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"uid":"38f4b9e5-98c2-4938-a86e-9aec7ba9c9ea"}: {} [e2e-llm-inference-service] f:ports: {} [e2e-llm-inference-service] addressType: IPv4 [e2e-llm-inference-service] endpoints: [e2e-llm-inference-service] - addresses: [e2e-llm-inference-service] - 10.132.0.55 [e2e-llm-inference-service] conditions: [e2e-llm-inference-service] ready: true [e2e-llm-inference-service] serving: true [e2e-llm-inference-service] terminating: false [e2e-llm-inference-service] targetRef: [e2e-llm-inference-service] kind: Pod [e2e-llm-inference-service] namespace: kserve-ci-e2e-test [e2e-llm-inference-service] name: custom-route-timeout-test-kserve-router-scheduler-6578bddfnbxnb [e2e-llm-inference-service] uid: f8a7664e-b142-4621-b773-b5e45ceaa80f [e2e-llm-inference-service] nodeName: ip-10-0-133-25.ec2.internal [e2e-llm-inference-service] zone: us-east-1a [e2e-llm-inference-service] ports: [e2e-llm-inference-service] - name: grpc [e2e-llm-inference-service] protocol: TCP [e2e-llm-inference-service] port: 9002 [e2e-llm-inference-service] - name: grpc-health [e2e-llm-inference-service] protocol: TCP [e2e-llm-inference-service] port: 9003 [e2e-llm-inference-service] - name: metrics [e2e-llm-inference-service] protocol: TCP [e2e-llm-inference-service] port: 9090 [e2e-llm-inference-service] - name: zmq [e2e-llm-inference-service] protocol: TCP [e2e-llm-inference-service] port: 5557 [e2e-llm-inference-service] apiVersion: discovery.k8s.io/v1 [e2e-llm-inference-service] kind: EndpointSlice [e2e-llm-inference-service] [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1254 --- [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1255 metadata: [e2e-llm-inference-service] name: custom-route-timeout-test-kserve-workload-svc-jzbv8 [e2e-llm-inference-service] generateName: custom-route-timeout-test-kserve-workload-svc- [e2e-llm-inference-service] namespace: kserve-ci-e2e-test [e2e-llm-inference-service] uid: 2294ef15-2b34-4d37-9961-609622360754 [e2e-llm-inference-service] resourceVersion: '50895' [e2e-llm-inference-service] generation: 3 [e2e-llm-inference-service] creationTimestamp: '2026-07-07T14:48:57Z' [e2e-llm-inference-service] labels: [e2e-llm-inference-service] app.kubernetes.io/component: llminferenceservice-workload [e2e-llm-inference-service] app.kubernetes.io/name: custom-route-timeout-test [e2e-llm-inference-service] app.kubernetes.io/part-of: llminferenceservice [e2e-llm-inference-service] endpointslice.kubernetes.io/managed-by: endpointslice-controller.k8s.io [e2e-llm-inference-service] kubernetes.io/service-name: custom-route-timeout-test-kserve-workload-svc [e2e-llm-inference-service] annotations: [e2e-llm-inference-service] endpoints.kubernetes.io/last-change-trigger-time: '2026-07-07T14:50:57Z' [e2e-llm-inference-service] ownerReferences: [e2e-llm-inference-service] - apiVersion: v1 [e2e-llm-inference-service] kind: Service [e2e-llm-inference-service] name: custom-route-timeout-test-kserve-workload-svc [e2e-llm-inference-service] uid: 8a311167-fdf9-4de8-b36a-8990fdd59c71 [e2e-llm-inference-service] controller: true [e2e-llm-inference-service] blockOwnerDeletion: true [e2e-llm-inference-service] managedFields: [e2e-llm-inference-service] - manager: kube-controller-manager [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] apiVersion: discovery.k8s.io/v1 [e2e-llm-inference-service] time: '2026-07-07T14:50:57Z' [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:addressType: {} [e2e-llm-inference-service] f:endpoints: {} [e2e-llm-inference-service] f:metadata: [e2e-llm-inference-service] f:annotations: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:endpoints.kubernetes.io/last-change-trigger-time: {} [e2e-llm-inference-service] f:generateName: {} [e2e-llm-inference-service] f:labels: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:app.kubernetes.io/component: {} [e2e-llm-inference-service] f:app.kubernetes.io/name: {} [e2e-llm-inference-service] f:app.kubernetes.io/part-of: {} [e2e-llm-inference-service] f:endpointslice.kubernetes.io/managed-by: {} [e2e-llm-inference-service] f:kubernetes.io/service-name: {} [e2e-llm-inference-service] f:ownerReferences: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"uid":"8a311167-fdf9-4de8-b36a-8990fdd59c71"}: {} [e2e-llm-inference-service] f:ports: {} [e2e-llm-inference-service] addressType: IPv4 [e2e-llm-inference-service] endpoints: [e2e-llm-inference-service] - addresses: [e2e-llm-inference-service] - 10.134.0.33 [e2e-llm-inference-service] conditions: [e2e-llm-inference-service] ready: true [e2e-llm-inference-service] serving: true [e2e-llm-inference-service] terminating: false [e2e-llm-inference-service] targetRef: [e2e-llm-inference-service] kind: Pod [e2e-llm-inference-service] namespace: kserve-ci-e2e-test [e2e-llm-inference-service] name: custom-route-timeout-test-kserve-7f56b8c469-hkpbl [e2e-llm-inference-service] uid: 79bc5e6e-61ea-4684-bc6f-a2fb38b92a73 [e2e-llm-inference-service] nodeName: ip-10-0-139-47.ec2.internal [e2e-llm-inference-service] zone: us-east-1a [e2e-llm-inference-service] ports: [e2e-llm-inference-service] - name: https [e2e-llm-inference-service] protocol: TCP [e2e-llm-inference-service] port: 8000 [e2e-llm-inference-service] appProtocol: https [e2e-llm-inference-service] apiVersion: discovery.k8s.io/v1 [e2e-llm-inference-service] kind: EndpointSlice [e2e-llm-inference-service] [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1254 --- [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1255 metadata: [e2e-llm-inference-service] name: custom-route-timeout-test-epp-rb [e2e-llm-inference-service] namespace: kserve-ci-e2e-test [e2e-llm-inference-service] uid: ce20d641-d03b-4916-9799-6413f17c84d2 [e2e-llm-inference-service] resourceVersion: '48986' [e2e-llm-inference-service] creationTimestamp: '2026-07-07T14:48:57Z' [e2e-llm-inference-service] labels: [e2e-llm-inference-service] app.kubernetes.io/component: llminferenceservice-router-scheduler [e2e-llm-inference-service] app.kubernetes.io/name: custom-route-timeout-test [e2e-llm-inference-service] app.kubernetes.io/part-of: llminferenceservice [e2e-llm-inference-service] ownerReferences: [e2e-llm-inference-service] - apiVersion: serving.kserve.io/v1alpha2 [e2e-llm-inference-service] kind: LLMInferenceService [e2e-llm-inference-service] name: custom-route-timeout-test [e2e-llm-inference-service] uid: 986f2d2c-a6d1-463d-9850-3aad6459eace [e2e-llm-inference-service] controller: true [e2e-llm-inference-service] blockOwnerDeletion: true [e2e-llm-inference-service] managedFields: [e2e-llm-inference-service] - manager: manager [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] apiVersion: rbac.authorization.k8s.io/v1 [e2e-llm-inference-service] time: '2026-07-07T14:48:57Z' [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:metadata: [e2e-llm-inference-service] f:labels: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:app.kubernetes.io/component: {} [e2e-llm-inference-service] f:app.kubernetes.io/name: {} [e2e-llm-inference-service] f:app.kubernetes.io/part-of: {} [e2e-llm-inference-service] f:ownerReferences: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"uid":"986f2d2c-a6d1-463d-9850-3aad6459eace"}: {} [e2e-llm-inference-service] f:roleRef: {} [e2e-llm-inference-service] f:subjects: {} [e2e-llm-inference-service] userNames: [e2e-llm-inference-service] - system:serviceaccount:kserve-ci-e2e-test:custom-route-timeout-test-epp-sa [e2e-llm-inference-service] groupNames: null [e2e-llm-inference-service] subjects: [e2e-llm-inference-service] - kind: ServiceAccount [e2e-llm-inference-service] namespace: kserve-ci-e2e-test [e2e-llm-inference-service] name: custom-route-timeout-test-epp-sa [e2e-llm-inference-service] roleRef: [e2e-llm-inference-service] namespace: kserve-ci-e2e-test [e2e-llm-inference-service] name: custom-route-timeout-test-epp-role [e2e-llm-inference-service] apiVersion: authorization.openshift.io/v1 [e2e-llm-inference-service] kind: RoleBinding [e2e-llm-inference-service] [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1254 --- [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1255 metadata: [e2e-llm-inference-service] name: custom-route-timeout-test-epp-role [e2e-llm-inference-service] namespace: kserve-ci-e2e-test [e2e-llm-inference-service] uid: 849b4bc3-0d16-4af4-acf7-dca0f6b0e70e [e2e-llm-inference-service] resourceVersion: '48981' [e2e-llm-inference-service] creationTimestamp: '2026-07-07T14:48:57Z' [e2e-llm-inference-service] labels: [e2e-llm-inference-service] app.kubernetes.io/component: llminferenceservice-router-scheduler [e2e-llm-inference-service] app.kubernetes.io/name: custom-route-timeout-test [e2e-llm-inference-service] app.kubernetes.io/part-of: llminferenceservice [e2e-llm-inference-service] ownerReferences: [e2e-llm-inference-service] - apiVersion: serving.kserve.io/v1alpha2 [e2e-llm-inference-service] kind: LLMInferenceService [e2e-llm-inference-service] name: custom-route-timeout-test [e2e-llm-inference-service] uid: 986f2d2c-a6d1-463d-9850-3aad6459eace [e2e-llm-inference-service] controller: true [e2e-llm-inference-service] blockOwnerDeletion: true [e2e-llm-inference-service] managedFields: [e2e-llm-inference-service] - manager: manager [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] apiVersion: rbac.authorization.k8s.io/v1 [e2e-llm-inference-service] time: '2026-07-07T14:48:57Z' [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:metadata: [e2e-llm-inference-service] f:labels: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:app.kubernetes.io/component: {} [e2e-llm-inference-service] f:app.kubernetes.io/name: {} [e2e-llm-inference-service] f:app.kubernetes.io/part-of: {} [e2e-llm-inference-service] f:ownerReferences: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"uid":"986f2d2c-a6d1-463d-9850-3aad6459eace"}: {} [e2e-llm-inference-service] f:rules: {} [e2e-llm-inference-service] rules: [e2e-llm-inference-service] - verbs: [e2e-llm-inference-service] - get [e2e-llm-inference-service] - list [e2e-llm-inference-service] - watch [e2e-llm-inference-service] attributeRestrictions: null [e2e-llm-inference-service] apiGroups: [e2e-llm-inference-service] - '' [e2e-llm-inference-service] resources: [e2e-llm-inference-service] - pods [e2e-llm-inference-service] - verbs: [e2e-llm-inference-service] - get [e2e-llm-inference-service] - list [e2e-llm-inference-service] - watch [e2e-llm-inference-service] attributeRestrictions: null [e2e-llm-inference-service] apiGroups: [e2e-llm-inference-service] - inference.networking.k8s.io [e2e-llm-inference-service] - inference.networking.x-k8s.io [e2e-llm-inference-service] resources: [e2e-llm-inference-service] - inferencemodels [e2e-llm-inference-service] - inferenceobjectives [e2e-llm-inference-service] - inferencepools [e2e-llm-inference-service] - verbs: [e2e-llm-inference-service] - get [e2e-llm-inference-service] - list [e2e-llm-inference-service] - watch [e2e-llm-inference-service] attributeRestrictions: null [e2e-llm-inference-service] apiGroups: [e2e-llm-inference-service] - inference.networking.x-k8s.io [e2e-llm-inference-service] resources: [e2e-llm-inference-service] - inferencemodelrewrites [e2e-llm-inference-service] - inferencepoolimports [e2e-llm-inference-service] - verbs: [e2e-llm-inference-service] - get [e2e-llm-inference-service] - list [e2e-llm-inference-service] - watch [e2e-llm-inference-service] attributeRestrictions: null [e2e-llm-inference-service] apiGroups: [e2e-llm-inference-service] - discovery.k8s.io [e2e-llm-inference-service] resources: [e2e-llm-inference-service] - endpointslices [e2e-llm-inference-service] - verbs: [e2e-llm-inference-service] - create [e2e-llm-inference-service] - delete [e2e-llm-inference-service] - get [e2e-llm-inference-service] - list [e2e-llm-inference-service] - patch [e2e-llm-inference-service] - update [e2e-llm-inference-service] - watch [e2e-llm-inference-service] attributeRestrictions: null [e2e-llm-inference-service] apiGroups: [e2e-llm-inference-service] - coordination.k8s.io [e2e-llm-inference-service] resources: [e2e-llm-inference-service] - leases [e2e-llm-inference-service] apiVersion: authorization.openshift.io/v1 [e2e-llm-inference-service] kind: Role [e2e-llm-inference-service] [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1254 --- [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1255 apiVersion: gateway.networking.k8s.io/v1 [e2e-llm-inference-service] kind: HTTPRoute [e2e-llm-inference-service] metadata: [e2e-llm-inference-service] annotations: [e2e-llm-inference-service] serving.kserve.io/inference-pool-migrated: v1 [e2e-llm-inference-service] creationTimestamp: '2026-07-07T14:48:58Z' [e2e-llm-inference-service] generation: 2 [e2e-llm-inference-service] labels: [e2e-llm-inference-service] app.kubernetes.io/component: llminferenceservice-router [e2e-llm-inference-service] app.kubernetes.io/name: custom-route-timeout-test [e2e-llm-inference-service] app.kubernetes.io/part-of: llminferenceservice [e2e-llm-inference-service] managedFields: [e2e-llm-inference-service] - apiVersion: gateway.networking.k8s.io/v1 [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:metadata: [e2e-llm-inference-service] f:annotations: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:serving.kserve.io/inference-pool-migrated: {} [e2e-llm-inference-service] f:labels: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:app.kubernetes.io/component: {} [e2e-llm-inference-service] f:app.kubernetes.io/name: {} [e2e-llm-inference-service] f:app.kubernetes.io/part-of: {} [e2e-llm-inference-service] f:ownerReferences: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"uid":"986f2d2c-a6d1-463d-9850-3aad6459eace"}: {} [e2e-llm-inference-service] f:spec: {} [e2e-llm-inference-service] manager: manager [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] time: '2026-07-07T14:49:11Z' [e2e-llm-inference-service] - apiVersion: gateway.networking.k8s.io/v1 [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:status: [e2e-llm-inference-service] f:parents: {} [e2e-llm-inference-service] manager: manager [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] subresource: status [e2e-llm-inference-service] time: '2026-07-07T14:49:12Z' [e2e-llm-inference-service] - apiVersion: gateway.networking.k8s.io/v1beta1 [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:status: {} [e2e-llm-inference-service] manager: pilot-discovery [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] subresource: status [e2e-llm-inference-service] time: '2026-07-07T14:49:12Z' [e2e-llm-inference-service] name: custom-route-timeout-test-kserve-route [e2e-llm-inference-service] namespace: kserve-ci-e2e-test [e2e-llm-inference-service] ownerReferences: [e2e-llm-inference-service] - apiVersion: serving.kserve.io/v1alpha2 [e2e-llm-inference-service] blockOwnerDeletion: true [e2e-llm-inference-service] controller: true [e2e-llm-inference-service] kind: LLMInferenceService [e2e-llm-inference-service] name: custom-route-timeout-test [e2e-llm-inference-service] uid: 986f2d2c-a6d1-463d-9850-3aad6459eace [e2e-llm-inference-service] resourceVersion: '49305' [e2e-llm-inference-service] uid: 860dd025-8366-4bbd-bfa8-a5334ef9574c [e2e-llm-inference-service] spec: [e2e-llm-inference-service] parentRefs: [e2e-llm-inference-service] - group: gateway.networking.k8s.io [e2e-llm-inference-service] kind: Gateway [e2e-llm-inference-service] name: openshift-ai-inference [e2e-llm-inference-service] namespace: openshift-ingress [e2e-llm-inference-service] rules: [e2e-llm-inference-service] - backendRefs: [e2e-llm-inference-service] - group: inference.networking.k8s.io [e2e-llm-inference-service] kind: InferencePool [e2e-llm-inference-service] name: custom-route-timeout-test-inference-pool [e2e-llm-inference-service] namespace: kserve-ci-e2e-test [e2e-llm-inference-service] port: 8000 [e2e-llm-inference-service] weight: 1 [e2e-llm-inference-service] filters: [e2e-llm-inference-service] - type: URLRewrite [e2e-llm-inference-service] urlRewrite: [e2e-llm-inference-service] path: [e2e-llm-inference-service] replacePrefixMatch: /v1/completions [e2e-llm-inference-service] type: ReplacePrefixMatch [e2e-llm-inference-service] matches: [e2e-llm-inference-service] - path: [e2e-llm-inference-service] type: PathPrefix [e2e-llm-inference-service] value: /kserve-ci-e2e-test/custom-route-timeout-test/v1/completions [e2e-llm-inference-service] timeouts: [e2e-llm-inference-service] backendRequest: 30s [e2e-llm-inference-service] request: 30s [e2e-llm-inference-service] - backendRefs: [e2e-llm-inference-service] - group: inference.networking.k8s.io [e2e-llm-inference-service] kind: InferencePool [e2e-llm-inference-service] name: custom-route-timeout-test-inference-pool [e2e-llm-inference-service] namespace: kserve-ci-e2e-test [e2e-llm-inference-service] port: 8000 [e2e-llm-inference-service] weight: 1 [e2e-llm-inference-service] filters: [e2e-llm-inference-service] - type: URLRewrite [e2e-llm-inference-service] urlRewrite: [e2e-llm-inference-service] path: [e2e-llm-inference-service] replacePrefixMatch: /v1/chat/completions [e2e-llm-inference-service] type: ReplacePrefixMatch [e2e-llm-inference-service] matches: [e2e-llm-inference-service] - path: [e2e-llm-inference-service] type: PathPrefix [e2e-llm-inference-service] value: /kserve-ci-e2e-test/custom-route-timeout-test/v1/chat/completions [e2e-llm-inference-service] timeouts: [e2e-llm-inference-service] backendRequest: 30s [e2e-llm-inference-service] request: 30s [e2e-llm-inference-service] - backendRefs: [e2e-llm-inference-service] - group: '' [e2e-llm-inference-service] kind: Service [e2e-llm-inference-service] name: custom-route-timeout-test-kserve-workload-svc [e2e-llm-inference-service] namespace: kserve-ci-e2e-test [e2e-llm-inference-service] port: 8000 [e2e-llm-inference-service] weight: 1 [e2e-llm-inference-service] filters: [e2e-llm-inference-service] - type: URLRewrite [e2e-llm-inference-service] urlRewrite: [e2e-llm-inference-service] path: [e2e-llm-inference-service] replacePrefixMatch: / [e2e-llm-inference-service] type: ReplacePrefixMatch [e2e-llm-inference-service] matches: [e2e-llm-inference-service] - path: [e2e-llm-inference-service] type: PathPrefix [e2e-llm-inference-service] value: /kserve-ci-e2e-test/custom-route-timeout-test [e2e-llm-inference-service] timeouts: [e2e-llm-inference-service] backendRequest: 30s [e2e-llm-inference-service] request: 30s [e2e-llm-inference-service] status: [e2e-llm-inference-service] parents: [e2e-llm-inference-service] - conditions: [e2e-llm-inference-service] - lastTransitionTime: '2026-07-07T14:48:58Z' [e2e-llm-inference-service] message: Route was valid [e2e-llm-inference-service] observedGeneration: 2 [e2e-llm-inference-service] reason: Accepted [e2e-llm-inference-service] status: 'True' [e2e-llm-inference-service] type: Accepted [e2e-llm-inference-service] - lastTransitionTime: '2026-07-07T14:49:12Z' [e2e-llm-inference-service] message: All references resolved [e2e-llm-inference-service] observedGeneration: 2 [e2e-llm-inference-service] reason: ResolvedRefs [e2e-llm-inference-service] status: 'True' [e2e-llm-inference-service] type: ResolvedRefs [e2e-llm-inference-service] controllerName: openshift.io/gateway-controller/v1 [e2e-llm-inference-service] parentRef: [e2e-llm-inference-service] group: gateway.networking.k8s.io [e2e-llm-inference-service] kind: Gateway [e2e-llm-inference-service] name: openshift-ai-inference [e2e-llm-inference-service] namespace: openshift-ingress [e2e-llm-inference-service] - conditions: [e2e-llm-inference-service] - lastTransitionTime: '2026-07-07T14:48:59Z' [e2e-llm-inference-service] message: Object affected by AuthPolicy [kserve-ci-e2e-test/custom-route-timeout-test-kserve-route-authn [e2e-llm-inference-service] openshift-ingress/openshift-ai-inference-authn] [e2e-llm-inference-service] observedGeneration: 2 [e2e-llm-inference-service] reason: Accepted [e2e-llm-inference-service] status: 'True' [e2e-llm-inference-service] type: kuadrant.io/AuthPolicyAffected [e2e-llm-inference-service] controllerName: kuadrant.io/policy-controller [e2e-llm-inference-service] parentRef: [e2e-llm-inference-service] group: gateway.networking.k8s.io [e2e-llm-inference-service] kind: Gateway [e2e-llm-inference-service] name: openshift-ai-inference [e2e-llm-inference-service] namespace: openshift-ingress [e2e-llm-inference-service] [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1254 --- [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1255 apiVersion: gateway.networking.k8s.io/v1beta1 [e2e-llm-inference-service] kind: HTTPRoute [e2e-llm-inference-service] metadata: [e2e-llm-inference-service] annotations: [e2e-llm-inference-service] serving.kserve.io/inference-pool-migrated: v1 [e2e-llm-inference-service] creationTimestamp: '2026-07-07T14:48:58Z' [e2e-llm-inference-service] generation: 2 [e2e-llm-inference-service] labels: [e2e-llm-inference-service] app.kubernetes.io/component: llminferenceservice-router [e2e-llm-inference-service] app.kubernetes.io/name: custom-route-timeout-test [e2e-llm-inference-service] app.kubernetes.io/part-of: llminferenceservice [e2e-llm-inference-service] managedFields: [e2e-llm-inference-service] - apiVersion: gateway.networking.k8s.io/v1 [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:metadata: [e2e-llm-inference-service] f:annotations: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:serving.kserve.io/inference-pool-migrated: {} [e2e-llm-inference-service] f:labels: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:app.kubernetes.io/component: {} [e2e-llm-inference-service] f:app.kubernetes.io/name: {} [e2e-llm-inference-service] f:app.kubernetes.io/part-of: {} [e2e-llm-inference-service] f:ownerReferences: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"uid":"986f2d2c-a6d1-463d-9850-3aad6459eace"}: {} [e2e-llm-inference-service] f:spec: {} [e2e-llm-inference-service] manager: manager [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] time: '2026-07-07T14:49:11Z' [e2e-llm-inference-service] - apiVersion: gateway.networking.k8s.io/v1 [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:status: [e2e-llm-inference-service] f:parents: {} [e2e-llm-inference-service] manager: manager [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] subresource: status [e2e-llm-inference-service] time: '2026-07-07T14:49:12Z' [e2e-llm-inference-service] - apiVersion: gateway.networking.k8s.io/v1beta1 [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:status: {} [e2e-llm-inference-service] manager: pilot-discovery [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] subresource: status [e2e-llm-inference-service] time: '2026-07-07T14:49:12Z' [e2e-llm-inference-service] name: custom-route-timeout-test-kserve-route [e2e-llm-inference-service] namespace: kserve-ci-e2e-test [e2e-llm-inference-service] ownerReferences: [e2e-llm-inference-service] - apiVersion: serving.kserve.io/v1alpha2 [e2e-llm-inference-service] blockOwnerDeletion: true [e2e-llm-inference-service] controller: true [e2e-llm-inference-service] kind: LLMInferenceService [e2e-llm-inference-service] name: custom-route-timeout-test [e2e-llm-inference-service] uid: 986f2d2c-a6d1-463d-9850-3aad6459eace [e2e-llm-inference-service] resourceVersion: '49305' [e2e-llm-inference-service] uid: 860dd025-8366-4bbd-bfa8-a5334ef9574c [e2e-llm-inference-service] spec: [e2e-llm-inference-service] parentRefs: [e2e-llm-inference-service] - group: gateway.networking.k8s.io [e2e-llm-inference-service] kind: Gateway [e2e-llm-inference-service] name: openshift-ai-inference [e2e-llm-inference-service] namespace: openshift-ingress [e2e-llm-inference-service] rules: [e2e-llm-inference-service] - backendRefs: [e2e-llm-inference-service] - group: inference.networking.k8s.io [e2e-llm-inference-service] kind: InferencePool [e2e-llm-inference-service] name: custom-route-timeout-test-inference-pool [e2e-llm-inference-service] namespace: kserve-ci-e2e-test [e2e-llm-inference-service] port: 8000 [e2e-llm-inference-service] weight: 1 [e2e-llm-inference-service] filters: [e2e-llm-inference-service] - type: URLRewrite [e2e-llm-inference-service] urlRewrite: [e2e-llm-inference-service] path: [e2e-llm-inference-service] replacePrefixMatch: /v1/completions [e2e-llm-inference-service] type: ReplacePrefixMatch [e2e-llm-inference-service] matches: [e2e-llm-inference-service] - path: [e2e-llm-inference-service] type: PathPrefix [e2e-llm-inference-service] value: /kserve-ci-e2e-test/custom-route-timeout-test/v1/completions [e2e-llm-inference-service] timeouts: [e2e-llm-inference-service] backendRequest: 30s [e2e-llm-inference-service] request: 30s [e2e-llm-inference-service] - backendRefs: [e2e-llm-inference-service] - group: inference.networking.k8s.io [e2e-llm-inference-service] kind: InferencePool [e2e-llm-inference-service] name: custom-route-timeout-test-inference-pool [e2e-llm-inference-service] namespace: kserve-ci-e2e-test [e2e-llm-inference-service] port: 8000 [e2e-llm-inference-service] weight: 1 [e2e-llm-inference-service] filters: [e2e-llm-inference-service] - type: URLRewrite [e2e-llm-inference-service] urlRewrite: [e2e-llm-inference-service] path: [e2e-llm-inference-service] replacePrefixMatch: /v1/chat/completions [e2e-llm-inference-service] type: ReplacePrefixMatch [e2e-llm-inference-service] matches: [e2e-llm-inference-service] - path: [e2e-llm-inference-service] type: PathPrefix [e2e-llm-inference-service] value: /kserve-ci-e2e-test/custom-route-timeout-test/v1/chat/completions [e2e-llm-inference-service] timeouts: [e2e-llm-inference-service] backendRequest: 30s [e2e-llm-inference-service] request: 30s [e2e-llm-inference-service] - backendRefs: [e2e-llm-inference-service] - group: '' [e2e-llm-inference-service] kind: Service [e2e-llm-inference-service] name: custom-route-timeout-test-kserve-workload-svc [e2e-llm-inference-service] namespace: kserve-ci-e2e-test [e2e-llm-inference-service] port: 8000 [e2e-llm-inference-service] weight: 1 [e2e-llm-inference-service] filters: [e2e-llm-inference-service] - type: URLRewrite [e2e-llm-inference-service] urlRewrite: [e2e-llm-inference-service] path: [e2e-llm-inference-service] replacePrefixMatch: / [e2e-llm-inference-service] type: ReplacePrefixMatch [e2e-llm-inference-service] matches: [e2e-llm-inference-service] - path: [e2e-llm-inference-service] type: PathPrefix [e2e-llm-inference-service] value: /kserve-ci-e2e-test/custom-route-timeout-test [e2e-llm-inference-service] timeouts: [e2e-llm-inference-service] backendRequest: 30s [e2e-llm-inference-service] request: 30s [e2e-llm-inference-service] status: [e2e-llm-inference-service] parents: [e2e-llm-inference-service] - conditions: [e2e-llm-inference-service] - lastTransitionTime: '2026-07-07T14:48:58Z' [e2e-llm-inference-service] message: Route was valid [e2e-llm-inference-service] observedGeneration: 2 [e2e-llm-inference-service] reason: Accepted [e2e-llm-inference-service] status: 'True' [e2e-llm-inference-service] type: Accepted [e2e-llm-inference-service] - lastTransitionTime: '2026-07-07T14:49:12Z' [e2e-llm-inference-service] message: All references resolved [e2e-llm-inference-service] observedGeneration: 2 [e2e-llm-inference-service] reason: ResolvedRefs [e2e-llm-inference-service] status: 'True' [e2e-llm-inference-service] type: ResolvedRefs [e2e-llm-inference-service] controllerName: openshift.io/gateway-controller/v1 [e2e-llm-inference-service] parentRef: [e2e-llm-inference-service] group: gateway.networking.k8s.io [e2e-llm-inference-service] kind: Gateway [e2e-llm-inference-service] name: openshift-ai-inference [e2e-llm-inference-service] namespace: openshift-ingress [e2e-llm-inference-service] - conditions: [e2e-llm-inference-service] - lastTransitionTime: '2026-07-07T14:48:59Z' [e2e-llm-inference-service] message: Object affected by AuthPolicy [kserve-ci-e2e-test/custom-route-timeout-test-kserve-route-authn [e2e-llm-inference-service] openshift-ingress/openshift-ai-inference-authn] [e2e-llm-inference-service] observedGeneration: 2 [e2e-llm-inference-service] reason: Accepted [e2e-llm-inference-service] status: 'True' [e2e-llm-inference-service] type: kuadrant.io/AuthPolicyAffected [e2e-llm-inference-service] controllerName: kuadrant.io/policy-controller [e2e-llm-inference-service] parentRef: [e2e-llm-inference-service] group: gateway.networking.k8s.io [e2e-llm-inference-service] kind: Gateway [e2e-llm-inference-service] name: openshift-ai-inference [e2e-llm-inference-service] namespace: openshift-ingress [e2e-llm-inference-service] [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1254 --- [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1255 apiVersion: inference.networking.k8s.io/v1 [e2e-llm-inference-service] kind: InferencePool [e2e-llm-inference-service] metadata: [e2e-llm-inference-service] creationTimestamp: '2026-07-07T14:48:57Z' [e2e-llm-inference-service] generation: 1 [e2e-llm-inference-service] labels: [e2e-llm-inference-service] app.kubernetes.io/component: llminferenceservice-router-scheduler [e2e-llm-inference-service] app.kubernetes.io/name: custom-route-timeout-test [e2e-llm-inference-service] app.kubernetes.io/part-of: llminferenceservice [e2e-llm-inference-service] managedFields: [e2e-llm-inference-service] - apiVersion: inference.networking.k8s.io/v1 [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:metadata: [e2e-llm-inference-service] f:labels: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:app.kubernetes.io/component: {} [e2e-llm-inference-service] f:app.kubernetes.io/name: {} [e2e-llm-inference-service] f:app.kubernetes.io/part-of: {} [e2e-llm-inference-service] f:ownerReferences: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"uid":"986f2d2c-a6d1-463d-9850-3aad6459eace"}: {} [e2e-llm-inference-service] f:spec: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:appProtocol: {} [e2e-llm-inference-service] f:endpointPickerRef: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:failureMode: {} [e2e-llm-inference-service] f:group: {} [e2e-llm-inference-service] f:kind: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:port: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:number: {} [e2e-llm-inference-service] f:selector: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:matchLabels: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:app.kubernetes.io/name: {} [e2e-llm-inference-service] f:app.kubernetes.io/part-of: {} [e2e-llm-inference-service] f:kserve.io/component: {} [e2e-llm-inference-service] f:targetPorts: {} [e2e-llm-inference-service] manager: manager [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] time: '2026-07-07T14:48:57Z' [e2e-llm-inference-service] - apiVersion: inference.networking.k8s.io/v1 [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:status: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:parents: {} [e2e-llm-inference-service] manager: pilot-discovery [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] subresource: status [e2e-llm-inference-service] time: '2026-07-07T14:49:11Z' [e2e-llm-inference-service] name: custom-route-timeout-test-inference-pool [e2e-llm-inference-service] namespace: kserve-ci-e2e-test [e2e-llm-inference-service] ownerReferences: [e2e-llm-inference-service] - apiVersion: serving.kserve.io/v1alpha2 [e2e-llm-inference-service] blockOwnerDeletion: true [e2e-llm-inference-service] controller: true [e2e-llm-inference-service] kind: LLMInferenceService [e2e-llm-inference-service] name: custom-route-timeout-test [e2e-llm-inference-service] uid: 986f2d2c-a6d1-463d-9850-3aad6459eace [e2e-llm-inference-service] resourceVersion: '49280' [e2e-llm-inference-service] uid: 959f2d42-e448-47e1-891b-ff50823639a6 [e2e-llm-inference-service] spec: [e2e-llm-inference-service] appProtocol: http [e2e-llm-inference-service] endpointPickerRef: [e2e-llm-inference-service] failureMode: FailOpen [e2e-llm-inference-service] group: '' [e2e-llm-inference-service] kind: Service [e2e-llm-inference-service] name: custom-route-timeout-test-epp-service [e2e-llm-inference-service] port: [e2e-llm-inference-service] number: 9002 [e2e-llm-inference-service] selector: [e2e-llm-inference-service] matchLabels: [e2e-llm-inference-service] app.kubernetes.io/name: custom-route-timeout-test [e2e-llm-inference-service] app.kubernetes.io/part-of: llminferenceservice [e2e-llm-inference-service] kserve.io/component: workload [e2e-llm-inference-service] targetPorts: [e2e-llm-inference-service] - number: 8000 [e2e-llm-inference-service] status: [e2e-llm-inference-service] parents: [e2e-llm-inference-service] - conditions: [e2e-llm-inference-service] - lastTransitionTime: '2026-07-07T14:49:11Z' [e2e-llm-inference-service] message: Referenced by an HTTPRoute accepted by the parentRef Gateway [e2e-llm-inference-service] observedGeneration: 1 [e2e-llm-inference-service] reason: Accepted [e2e-llm-inference-service] status: 'True' [e2e-llm-inference-service] type: Accepted [e2e-llm-inference-service] - lastTransitionTime: '2026-07-07T14:49:11Z' [e2e-llm-inference-service] message: Referenced ExtensionRef resolved successfully [e2e-llm-inference-service] observedGeneration: 1 [e2e-llm-inference-service] reason: ResolvedRefs [e2e-llm-inference-service] status: 'True' [e2e-llm-inference-service] type: ResolvedRefs [e2e-llm-inference-service] parentRef: [e2e-llm-inference-service] group: networking.istio.io [e2e-llm-inference-service] kind: Gateway [e2e-llm-inference-service] name: openshift-ai-inference [e2e-llm-inference-service] namespace: openshift-ingress [e2e-llm-inference-service] [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1254 --- [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1255 apiVersion: kuadrant.io/v1 [e2e-llm-inference-service] kind: AuthPolicy [e2e-llm-inference-service] metadata: [e2e-llm-inference-service] creationTimestamp: '2026-07-07T14:48:58Z' [e2e-llm-inference-service] generation: 1 [e2e-llm-inference-service] labels: [e2e-llm-inference-service] app.kubernetes.io/component: llminferenceservice-policies [e2e-llm-inference-service] app.kubernetes.io/managed-by: odh-model-controller [e2e-llm-inference-service] app.kubernetes.io/name: custom-route-timeout-test [e2e-llm-inference-service] app.kubernetes.io/part-of: llminferenceservice [e2e-llm-inference-service] managedFields: [e2e-llm-inference-service] - apiVersion: kuadrant.io/v1 [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:metadata: [e2e-llm-inference-service] f:labels: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:app.kubernetes.io/component: {} [e2e-llm-inference-service] f:app.kubernetes.io/managed-by: {} [e2e-llm-inference-service] f:app.kubernetes.io/name: {} [e2e-llm-inference-service] f:app.kubernetes.io/part-of: {} [e2e-llm-inference-service] f:ownerReferences: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"uid":"986f2d2c-a6d1-463d-9850-3aad6459eace"}: {} [e2e-llm-inference-service] f:spec: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:rules: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:authentication: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:public: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:anonymous: {} [e2e-llm-inference-service] f:credentials: {} [e2e-llm-inference-service] f:metrics: {} [e2e-llm-inference-service] f:overrides: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:fairness: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] f:objective: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:value: {} [e2e-llm-inference-service] f:priority: {} [e2e-llm-inference-service] f:response: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:success: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:headers: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:x-gateway-inference-fairness-id: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:metrics: {} [e2e-llm-inference-service] f:plain: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:expression: {} [e2e-llm-inference-service] f:priority: {} [e2e-llm-inference-service] f:x-gateway-inference-objective: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:metrics: {} [e2e-llm-inference-service] f:plain: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:expression: {} [e2e-llm-inference-service] f:priority: {} [e2e-llm-inference-service] f:targetRef: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:group: {} [e2e-llm-inference-service] f:kind: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] manager: manager [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] time: '2026-07-07T14:48:58Z' [e2e-llm-inference-service] - apiVersion: kuadrant.io/v1 [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:status: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:conditions: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"type":"Accepted"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:lastTransitionTime: {} [e2e-llm-inference-service] f:message: {} [e2e-llm-inference-service] f:reason: {} [e2e-llm-inference-service] f:status: {} [e2e-llm-inference-service] f:type: {} [e2e-llm-inference-service] k:{"type":"Enforced"}: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:lastTransitionTime: {} [e2e-llm-inference-service] f:message: {} [e2e-llm-inference-service] f:reason: {} [e2e-llm-inference-service] f:status: {} [e2e-llm-inference-service] f:type: {} [e2e-llm-inference-service] f:observedGeneration: {} [e2e-llm-inference-service] manager: manager [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] subresource: status [e2e-llm-inference-service] time: '2026-07-07T14:49:00Z' [e2e-llm-inference-service] name: custom-route-timeout-test-kserve-route-authn [e2e-llm-inference-service] namespace: kserve-ci-e2e-test [e2e-llm-inference-service] ownerReferences: [e2e-llm-inference-service] - apiVersion: serving.kserve.io/v1alpha2 [e2e-llm-inference-service] blockOwnerDeletion: true [e2e-llm-inference-service] controller: true [e2e-llm-inference-service] kind: LLMInferenceService [e2e-llm-inference-service] name: custom-route-timeout-test [e2e-llm-inference-service] uid: 986f2d2c-a6d1-463d-9850-3aad6459eace [e2e-llm-inference-service] resourceVersion: '49118' [e2e-llm-inference-service] uid: a0b2843f-1e05-4a87-ab16-2b365c502962 [e2e-llm-inference-service] spec: [e2e-llm-inference-service] rules: [e2e-llm-inference-service] authentication: [e2e-llm-inference-service] public: [e2e-llm-inference-service] anonymous: {} [e2e-llm-inference-service] credentials: {} [e2e-llm-inference-service] metrics: false [e2e-llm-inference-service] overrides: [e2e-llm-inference-service] fairness: [e2e-llm-inference-service] value: unauthenticated [e2e-llm-inference-service] objective: [e2e-llm-inference-service] value: unauthenticated [e2e-llm-inference-service] priority: 0 [e2e-llm-inference-service] response: [e2e-llm-inference-service] success: [e2e-llm-inference-service] headers: [e2e-llm-inference-service] x-gateway-inference-fairness-id: [e2e-llm-inference-service] metrics: false [e2e-llm-inference-service] plain: [e2e-llm-inference-service] expression: auth.identity.fairness [e2e-llm-inference-service] priority: 0 [e2e-llm-inference-service] x-gateway-inference-objective: [e2e-llm-inference-service] metrics: false [e2e-llm-inference-service] plain: [e2e-llm-inference-service] expression: auth.identity.objective [e2e-llm-inference-service] priority: 0 [e2e-llm-inference-service] targetRef: [e2e-llm-inference-service] group: gateway.networking.k8s.io [e2e-llm-inference-service] kind: HTTPRoute [e2e-llm-inference-service] name: custom-route-timeout-test-kserve-route [e2e-llm-inference-service] status: [e2e-llm-inference-service] conditions: [e2e-llm-inference-service] - lastTransitionTime: '2026-07-07T14:49:00Z' [e2e-llm-inference-service] message: AuthPolicy has been accepted [e2e-llm-inference-service] reason: Accepted [e2e-llm-inference-service] status: 'True' [e2e-llm-inference-service] type: Accepted [e2e-llm-inference-service] - lastTransitionTime: '2026-07-07T14:49:00Z' [e2e-llm-inference-service] message: AuthPolicy has been successfully enforced [e2e-llm-inference-service] reason: Enforced [e2e-llm-inference-service] status: 'True' [e2e-llm-inference-service] type: Enforced [e2e-llm-inference-service] observedGeneration: 1 [e2e-llm-inference-service] [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1254 --- [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1255 apiVersion: networking.istio.io/v1 [e2e-llm-inference-service] kind: DestinationRule [e2e-llm-inference-service] metadata: [e2e-llm-inference-service] creationTimestamp: '2026-07-07T14:48:58Z' [e2e-llm-inference-service] generation: 1 [e2e-llm-inference-service] labels: [e2e-llm-inference-service] app.kubernetes.io/component: llminferenceservice-router-scheduler [e2e-llm-inference-service] app.kubernetes.io/name: custom-route-timeout-test [e2e-llm-inference-service] app.kubernetes.io/part-of: llminferenceservice [e2e-llm-inference-service] llm-d.ai/managed: 'true' [e2e-llm-inference-service] managedFields: [e2e-llm-inference-service] - apiVersion: networking.istio.io/v1 [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:metadata: [e2e-llm-inference-service] f:labels: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:app.kubernetes.io/component: {} [e2e-llm-inference-service] f:app.kubernetes.io/name: {} [e2e-llm-inference-service] f:app.kubernetes.io/part-of: {} [e2e-llm-inference-service] f:llm-d.ai/managed: {} [e2e-llm-inference-service] f:ownerReferences: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"uid":"986f2d2c-a6d1-463d-9850-3aad6459eace"}: {} [e2e-llm-inference-service] f:spec: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:exportTo: {} [e2e-llm-inference-service] f:host: {} [e2e-llm-inference-service] f:trafficPolicy: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:tls: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:insecureSkipVerify: {} [e2e-llm-inference-service] f:mode: {} [e2e-llm-inference-service] f:sni: {} [e2e-llm-inference-service] manager: manager [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] time: '2026-07-07T14:48:58Z' [e2e-llm-inference-service] name: custom-route-timeout-test-kserve-scheduler [e2e-llm-inference-service] namespace: kserve-ci-e2e-test [e2e-llm-inference-service] ownerReferences: [e2e-llm-inference-service] - apiVersion: serving.kserve.io/v1alpha2 [e2e-llm-inference-service] blockOwnerDeletion: true [e2e-llm-inference-service] controller: true [e2e-llm-inference-service] kind: LLMInferenceService [e2e-llm-inference-service] name: custom-route-timeout-test [e2e-llm-inference-service] uid: 986f2d2c-a6d1-463d-9850-3aad6459eace [e2e-llm-inference-service] resourceVersion: '49032' [e2e-llm-inference-service] uid: bf45848f-ac6d-43ea-b8ca-25c81e3490bf [e2e-llm-inference-service] spec: [e2e-llm-inference-service] exportTo: [e2e-llm-inference-service] - '*' [e2e-llm-inference-service] host: custom-route-timeout-test-epp-service.kserve-ci-e2e-test.svc.cluster.local [e2e-llm-inference-service] trafficPolicy: [e2e-llm-inference-service] tls: [e2e-llm-inference-service] insecureSkipVerify: true [e2e-llm-inference-service] mode: SIMPLE [e2e-llm-inference-service] sni: custom-route-timeout-test-epp-service.kserve-ci-e2e-test.svc.cluster.local [e2e-llm-inference-service] [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1254 --- [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1255 apiVersion: networking.istio.io/v1 [e2e-llm-inference-service] kind: DestinationRule [e2e-llm-inference-service] metadata: [e2e-llm-inference-service] creationTimestamp: '2026-07-07T14:49:11Z' [e2e-llm-inference-service] generation: 1 [e2e-llm-inference-service] labels: [e2e-llm-inference-service] app.kubernetes.io/component: llminferenceservice-shadow-service [e2e-llm-inference-service] app.kubernetes.io/name: custom-route-timeout-test [e2e-llm-inference-service] app.kubernetes.io/part-of: llminferenceservice [e2e-llm-inference-service] llm-d.ai/managed: 'true' [e2e-llm-inference-service] managedFields: [e2e-llm-inference-service] - apiVersion: networking.istio.io/v1 [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:metadata: [e2e-llm-inference-service] f:labels: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:app.kubernetes.io/component: {} [e2e-llm-inference-service] f:app.kubernetes.io/name: {} [e2e-llm-inference-service] f:app.kubernetes.io/part-of: {} [e2e-llm-inference-service] f:llm-d.ai/managed: {} [e2e-llm-inference-service] f:ownerReferences: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"uid":"986f2d2c-a6d1-463d-9850-3aad6459eace"}: {} [e2e-llm-inference-service] f:spec: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:exportTo: {} [e2e-llm-inference-service] f:host: {} [e2e-llm-inference-service] f:trafficPolicy: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:tls: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:insecureSkipVerify: {} [e2e-llm-inference-service] f:mode: {} [e2e-llm-inference-service] f:sni: {} [e2e-llm-inference-service] manager: manager [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] time: '2026-07-07T14:49:11Z' [e2e-llm-inference-service] name: custom-route-timeout-test-kserve-shadow-svc [e2e-llm-inference-service] namespace: kserve-ci-e2e-test [e2e-llm-inference-service] ownerReferences: [e2e-llm-inference-service] - apiVersion: serving.kserve.io/v1alpha2 [e2e-llm-inference-service] blockOwnerDeletion: true [e2e-llm-inference-service] controller: true [e2e-llm-inference-service] kind: LLMInferenceService [e2e-llm-inference-service] name: custom-route-timeout-test [e2e-llm-inference-service] uid: 986f2d2c-a6d1-463d-9850-3aad6459eace [e2e-llm-inference-service] resourceVersion: '49285' [e2e-llm-inference-service] uid: f48d9a3f-0453-4611-beca-3363d0a4e3ac [e2e-llm-inference-service] spec: [e2e-llm-inference-service] exportTo: [e2e-llm-inference-service] - '*' [e2e-llm-inference-service] host: custom-route-timeout-test-inference-pool-ip-45342e47.kserve-ci-e2e-test.svc.cluster.local [e2e-llm-inference-service] trafficPolicy: [e2e-llm-inference-service] tls: [e2e-llm-inference-service] insecureSkipVerify: true [e2e-llm-inference-service] mode: SIMPLE [e2e-llm-inference-service] sni: custom-route-timeout-test-kserve-workload-svc.kserve-ci-e2e-test.svc.cluster.local [e2e-llm-inference-service] [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1254 --- [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1255 apiVersion: networking.istio.io/v1 [e2e-llm-inference-service] kind: DestinationRule [e2e-llm-inference-service] metadata: [e2e-llm-inference-service] creationTimestamp: '2026-07-07T14:48:58Z' [e2e-llm-inference-service] generation: 1 [e2e-llm-inference-service] labels: [e2e-llm-inference-service] app.kubernetes.io/component: llminferenceservice-workload [e2e-llm-inference-service] app.kubernetes.io/name: custom-route-timeout-test [e2e-llm-inference-service] app.kubernetes.io/part-of: llminferenceservice [e2e-llm-inference-service] llm-d.ai/managed: 'true' [e2e-llm-inference-service] managedFields: [e2e-llm-inference-service] - apiVersion: networking.istio.io/v1 [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:metadata: [e2e-llm-inference-service] f:labels: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:app.kubernetes.io/component: {} [e2e-llm-inference-service] f:app.kubernetes.io/name: {} [e2e-llm-inference-service] f:app.kubernetes.io/part-of: {} [e2e-llm-inference-service] f:llm-d.ai/managed: {} [e2e-llm-inference-service] f:ownerReferences: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"uid":"986f2d2c-a6d1-463d-9850-3aad6459eace"}: {} [e2e-llm-inference-service] f:spec: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:exportTo: {} [e2e-llm-inference-service] f:host: {} [e2e-llm-inference-service] f:trafficPolicy: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:tls: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:caCertificates: {} [e2e-llm-inference-service] f:insecureSkipVerify: {} [e2e-llm-inference-service] f:mode: {} [e2e-llm-inference-service] f:sni: {} [e2e-llm-inference-service] manager: manager [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] time: '2026-07-07T14:48:58Z' [e2e-llm-inference-service] name: custom-route-timeout-test-kserve-workload-svc [e2e-llm-inference-service] namespace: kserve-ci-e2e-test [e2e-llm-inference-service] ownerReferences: [e2e-llm-inference-service] - apiVersion: serving.kserve.io/v1alpha2 [e2e-llm-inference-service] blockOwnerDeletion: true [e2e-llm-inference-service] controller: true [e2e-llm-inference-service] kind: LLMInferenceService [e2e-llm-inference-service] name: custom-route-timeout-test [e2e-llm-inference-service] uid: 986f2d2c-a6d1-463d-9850-3aad6459eace [e2e-llm-inference-service] resourceVersion: '49055' [e2e-llm-inference-service] uid: c3658218-7bc8-4350-8172-4da512823fec [e2e-llm-inference-service] spec: [e2e-llm-inference-service] exportTo: [e2e-llm-inference-service] - '*' [e2e-llm-inference-service] host: custom-route-timeout-test-kserve-workload-svc.kserve-ci-e2e-test.svc.cluster.local [e2e-llm-inference-service] trafficPolicy: [e2e-llm-inference-service] tls: [e2e-llm-inference-service] caCertificates: /var/run/secrets/kubernetes.io/serviceaccount/service-ca.crt [e2e-llm-inference-service] insecureSkipVerify: false [e2e-llm-inference-service] mode: SIMPLE [e2e-llm-inference-service] sni: custom-route-timeout-test-kserve-workload-svc.kserve-ci-e2e-test.svc.cluster.local [e2e-llm-inference-service] [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1254 --- [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1255 apiVersion: networking.istio.io/v1beta1 [e2e-llm-inference-service] kind: DestinationRule [e2e-llm-inference-service] metadata: [e2e-llm-inference-service] creationTimestamp: '2026-07-07T14:48:58Z' [e2e-llm-inference-service] generation: 1 [e2e-llm-inference-service] labels: [e2e-llm-inference-service] app.kubernetes.io/component: llminferenceservice-router-scheduler [e2e-llm-inference-service] app.kubernetes.io/name: custom-route-timeout-test [e2e-llm-inference-service] app.kubernetes.io/part-of: llminferenceservice [e2e-llm-inference-service] llm-d.ai/managed: 'true' [e2e-llm-inference-service] managedFields: [e2e-llm-inference-service] - apiVersion: networking.istio.io/v1 [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:metadata: [e2e-llm-inference-service] f:labels: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:app.kubernetes.io/component: {} [e2e-llm-inference-service] f:app.kubernetes.io/name: {} [e2e-llm-inference-service] f:app.kubernetes.io/part-of: {} [e2e-llm-inference-service] f:llm-d.ai/managed: {} [e2e-llm-inference-service] f:ownerReferences: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"uid":"986f2d2c-a6d1-463d-9850-3aad6459eace"}: {} [e2e-llm-inference-service] f:spec: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:exportTo: {} [e2e-llm-inference-service] f:host: {} [e2e-llm-inference-service] f:trafficPolicy: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:tls: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:insecureSkipVerify: {} [e2e-llm-inference-service] f:mode: {} [e2e-llm-inference-service] f:sni: {} [e2e-llm-inference-service] manager: manager [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] time: '2026-07-07T14:48:58Z' [e2e-llm-inference-service] name: custom-route-timeout-test-kserve-scheduler [e2e-llm-inference-service] namespace: kserve-ci-e2e-test [e2e-llm-inference-service] ownerReferences: [e2e-llm-inference-service] - apiVersion: serving.kserve.io/v1alpha2 [e2e-llm-inference-service] blockOwnerDeletion: true [e2e-llm-inference-service] controller: true [e2e-llm-inference-service] kind: LLMInferenceService [e2e-llm-inference-service] name: custom-route-timeout-test [e2e-llm-inference-service] uid: 986f2d2c-a6d1-463d-9850-3aad6459eace [e2e-llm-inference-service] resourceVersion: '49032' [e2e-llm-inference-service] uid: bf45848f-ac6d-43ea-b8ca-25c81e3490bf [e2e-llm-inference-service] spec: [e2e-llm-inference-service] exportTo: [e2e-llm-inference-service] - '*' [e2e-llm-inference-service] host: custom-route-timeout-test-epp-service.kserve-ci-e2e-test.svc.cluster.local [e2e-llm-inference-service] trafficPolicy: [e2e-llm-inference-service] tls: [e2e-llm-inference-service] insecureSkipVerify: true [e2e-llm-inference-service] mode: SIMPLE [e2e-llm-inference-service] sni: custom-route-timeout-test-epp-service.kserve-ci-e2e-test.svc.cluster.local [e2e-llm-inference-service] [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1254 --- [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1255 apiVersion: networking.istio.io/v1beta1 [e2e-llm-inference-service] kind: DestinationRule [e2e-llm-inference-service] metadata: [e2e-llm-inference-service] creationTimestamp: '2026-07-07T14:49:11Z' [e2e-llm-inference-service] generation: 1 [e2e-llm-inference-service] labels: [e2e-llm-inference-service] app.kubernetes.io/component: llminferenceservice-shadow-service [e2e-llm-inference-service] app.kubernetes.io/name: custom-route-timeout-test [e2e-llm-inference-service] app.kubernetes.io/part-of: llminferenceservice [e2e-llm-inference-service] llm-d.ai/managed: 'true' [e2e-llm-inference-service] managedFields: [e2e-llm-inference-service] - apiVersion: networking.istio.io/v1 [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:metadata: [e2e-llm-inference-service] f:labels: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:app.kubernetes.io/component: {} [e2e-llm-inference-service] f:app.kubernetes.io/name: {} [e2e-llm-inference-service] f:app.kubernetes.io/part-of: {} [e2e-llm-inference-service] f:llm-d.ai/managed: {} [e2e-llm-inference-service] f:ownerReferences: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"uid":"986f2d2c-a6d1-463d-9850-3aad6459eace"}: {} [e2e-llm-inference-service] f:spec: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:exportTo: {} [e2e-llm-inference-service] f:host: {} [e2e-llm-inference-service] f:trafficPolicy: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:tls: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:insecureSkipVerify: {} [e2e-llm-inference-service] f:mode: {} [e2e-llm-inference-service] f:sni: {} [e2e-llm-inference-service] manager: manager [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] time: '2026-07-07T14:49:11Z' [e2e-llm-inference-service] name: custom-route-timeout-test-kserve-shadow-svc [e2e-llm-inference-service] namespace: kserve-ci-e2e-test [e2e-llm-inference-service] ownerReferences: [e2e-llm-inference-service] - apiVersion: serving.kserve.io/v1alpha2 [e2e-llm-inference-service] blockOwnerDeletion: true [e2e-llm-inference-service] controller: true [e2e-llm-inference-service] kind: LLMInferenceService [e2e-llm-inference-service] name: custom-route-timeout-test [e2e-llm-inference-service] uid: 986f2d2c-a6d1-463d-9850-3aad6459eace [e2e-llm-inference-service] resourceVersion: '49285' [e2e-llm-inference-service] uid: f48d9a3f-0453-4611-beca-3363d0a4e3ac [e2e-llm-inference-service] spec: [e2e-llm-inference-service] exportTo: [e2e-llm-inference-service] - '*' [e2e-llm-inference-service] host: custom-route-timeout-test-inference-pool-ip-45342e47.kserve-ci-e2e-test.svc.cluster.local [e2e-llm-inference-service] trafficPolicy: [e2e-llm-inference-service] tls: [e2e-llm-inference-service] insecureSkipVerify: true [e2e-llm-inference-service] mode: SIMPLE [e2e-llm-inference-service] sni: custom-route-timeout-test-kserve-workload-svc.kserve-ci-e2e-test.svc.cluster.local [e2e-llm-inference-service] [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1254 --- [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1255 apiVersion: networking.istio.io/v1beta1 [e2e-llm-inference-service] kind: DestinationRule [e2e-llm-inference-service] metadata: [e2e-llm-inference-service] creationTimestamp: '2026-07-07T14:48:58Z' [e2e-llm-inference-service] generation: 1 [e2e-llm-inference-service] labels: [e2e-llm-inference-service] app.kubernetes.io/component: llminferenceservice-workload [e2e-llm-inference-service] app.kubernetes.io/name: custom-route-timeout-test [e2e-llm-inference-service] app.kubernetes.io/part-of: llminferenceservice [e2e-llm-inference-service] llm-d.ai/managed: 'true' [e2e-llm-inference-service] managedFields: [e2e-llm-inference-service] - apiVersion: networking.istio.io/v1 [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:metadata: [e2e-llm-inference-service] f:labels: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:app.kubernetes.io/component: {} [e2e-llm-inference-service] f:app.kubernetes.io/name: {} [e2e-llm-inference-service] f:app.kubernetes.io/part-of: {} [e2e-llm-inference-service] f:llm-d.ai/managed: {} [e2e-llm-inference-service] f:ownerReferences: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"uid":"986f2d2c-a6d1-463d-9850-3aad6459eace"}: {} [e2e-llm-inference-service] f:spec: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:exportTo: {} [e2e-llm-inference-service] f:host: {} [e2e-llm-inference-service] f:trafficPolicy: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:tls: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:caCertificates: {} [e2e-llm-inference-service] f:insecureSkipVerify: {} [e2e-llm-inference-service] f:mode: {} [e2e-llm-inference-service] f:sni: {} [e2e-llm-inference-service] manager: manager [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] time: '2026-07-07T14:48:58Z' [e2e-llm-inference-service] name: custom-route-timeout-test-kserve-workload-svc [e2e-llm-inference-service] namespace: kserve-ci-e2e-test [e2e-llm-inference-service] ownerReferences: [e2e-llm-inference-service] - apiVersion: serving.kserve.io/v1alpha2 [e2e-llm-inference-service] blockOwnerDeletion: true [e2e-llm-inference-service] controller: true [e2e-llm-inference-service] kind: LLMInferenceService [e2e-llm-inference-service] name: custom-route-timeout-test [e2e-llm-inference-service] uid: 986f2d2c-a6d1-463d-9850-3aad6459eace [e2e-llm-inference-service] resourceVersion: '49055' [e2e-llm-inference-service] uid: c3658218-7bc8-4350-8172-4da512823fec [e2e-llm-inference-service] spec: [e2e-llm-inference-service] exportTo: [e2e-llm-inference-service] - '*' [e2e-llm-inference-service] host: custom-route-timeout-test-kserve-workload-svc.kserve-ci-e2e-test.svc.cluster.local [e2e-llm-inference-service] trafficPolicy: [e2e-llm-inference-service] tls: [e2e-llm-inference-service] caCertificates: /var/run/secrets/kubernetes.io/serviceaccount/service-ca.crt [e2e-llm-inference-service] insecureSkipVerify: false [e2e-llm-inference-service] mode: SIMPLE [e2e-llm-inference-service] sni: custom-route-timeout-test-kserve-workload-svc.kserve-ci-e2e-test.svc.cluster.local [e2e-llm-inference-service] [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1254 --- [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1255 apiVersion: networking.istio.io/v1alpha3 [e2e-llm-inference-service] kind: DestinationRule [e2e-llm-inference-service] metadata: [e2e-llm-inference-service] creationTimestamp: '2026-07-07T14:48:58Z' [e2e-llm-inference-service] generation: 1 [e2e-llm-inference-service] labels: [e2e-llm-inference-service] app.kubernetes.io/component: llminferenceservice-router-scheduler [e2e-llm-inference-service] app.kubernetes.io/name: custom-route-timeout-test [e2e-llm-inference-service] app.kubernetes.io/part-of: llminferenceservice [e2e-llm-inference-service] llm-d.ai/managed: 'true' [e2e-llm-inference-service] managedFields: [e2e-llm-inference-service] - apiVersion: networking.istio.io/v1 [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:metadata: [e2e-llm-inference-service] f:labels: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:app.kubernetes.io/component: {} [e2e-llm-inference-service] f:app.kubernetes.io/name: {} [e2e-llm-inference-service] f:app.kubernetes.io/part-of: {} [e2e-llm-inference-service] f:llm-d.ai/managed: {} [e2e-llm-inference-service] f:ownerReferences: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"uid":"986f2d2c-a6d1-463d-9850-3aad6459eace"}: {} [e2e-llm-inference-service] f:spec: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:exportTo: {} [e2e-llm-inference-service] f:host: {} [e2e-llm-inference-service] f:trafficPolicy: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:tls: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:insecureSkipVerify: {} [e2e-llm-inference-service] f:mode: {} [e2e-llm-inference-service] f:sni: {} [e2e-llm-inference-service] manager: manager [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] time: '2026-07-07T14:48:58Z' [e2e-llm-inference-service] name: custom-route-timeout-test-kserve-scheduler [e2e-llm-inference-service] namespace: kserve-ci-e2e-test [e2e-llm-inference-service] ownerReferences: [e2e-llm-inference-service] - apiVersion: serving.kserve.io/v1alpha2 [e2e-llm-inference-service] blockOwnerDeletion: true [e2e-llm-inference-service] controller: true [e2e-llm-inference-service] kind: LLMInferenceService [e2e-llm-inference-service] name: custom-route-timeout-test [e2e-llm-inference-service] uid: 986f2d2c-a6d1-463d-9850-3aad6459eace [e2e-llm-inference-service] resourceVersion: '49032' [e2e-llm-inference-service] uid: bf45848f-ac6d-43ea-b8ca-25c81e3490bf [e2e-llm-inference-service] spec: [e2e-llm-inference-service] exportTo: [e2e-llm-inference-service] - '*' [e2e-llm-inference-service] host: custom-route-timeout-test-epp-service.kserve-ci-e2e-test.svc.cluster.local [e2e-llm-inference-service] trafficPolicy: [e2e-llm-inference-service] tls: [e2e-llm-inference-service] insecureSkipVerify: true [e2e-llm-inference-service] mode: SIMPLE [e2e-llm-inference-service] sni: custom-route-timeout-test-epp-service.kserve-ci-e2e-test.svc.cluster.local [e2e-llm-inference-service] [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1254 --- [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1255 apiVersion: networking.istio.io/v1alpha3 [e2e-llm-inference-service] kind: DestinationRule [e2e-llm-inference-service] metadata: [e2e-llm-inference-service] creationTimestamp: '2026-07-07T14:49:11Z' [e2e-llm-inference-service] generation: 1 [e2e-llm-inference-service] labels: [e2e-llm-inference-service] app.kubernetes.io/component: llminferenceservice-shadow-service [e2e-llm-inference-service] app.kubernetes.io/name: custom-route-timeout-test [e2e-llm-inference-service] app.kubernetes.io/part-of: llminferenceservice [e2e-llm-inference-service] llm-d.ai/managed: 'true' [e2e-llm-inference-service] managedFields: [e2e-llm-inference-service] - apiVersion: networking.istio.io/v1 [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:metadata: [e2e-llm-inference-service] f:labels: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:app.kubernetes.io/component: {} [e2e-llm-inference-service] f:app.kubernetes.io/name: {} [e2e-llm-inference-service] f:app.kubernetes.io/part-of: {} [e2e-llm-inference-service] f:llm-d.ai/managed: {} [e2e-llm-inference-service] f:ownerReferences: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"uid":"986f2d2c-a6d1-463d-9850-3aad6459eace"}: {} [e2e-llm-inference-service] f:spec: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:exportTo: {} [e2e-llm-inference-service] f:host: {} [e2e-llm-inference-service] f:trafficPolicy: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:tls: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:insecureSkipVerify: {} [e2e-llm-inference-service] f:mode: {} [e2e-llm-inference-service] f:sni: {} [e2e-llm-inference-service] manager: manager [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] time: '2026-07-07T14:49:11Z' [e2e-llm-inference-service] name: custom-route-timeout-test-kserve-shadow-svc [e2e-llm-inference-service] namespace: kserve-ci-e2e-test [e2e-llm-inference-service] ownerReferences: [e2e-llm-inference-service] - apiVersion: serving.kserve.io/v1alpha2 [e2e-llm-inference-service] blockOwnerDeletion: true [e2e-llm-inference-service] controller: true [e2e-llm-inference-service] kind: LLMInferenceService [e2e-llm-inference-service] name: custom-route-timeout-test [e2e-llm-inference-service] uid: 986f2d2c-a6d1-463d-9850-3aad6459eace [e2e-llm-inference-service] resourceVersion: '49285' [e2e-llm-inference-service] uid: f48d9a3f-0453-4611-beca-3363d0a4e3ac [e2e-llm-inference-service] spec: [e2e-llm-inference-service] exportTo: [e2e-llm-inference-service] - '*' [e2e-llm-inference-service] host: custom-route-timeout-test-inference-pool-ip-45342e47.kserve-ci-e2e-test.svc.cluster.local [e2e-llm-inference-service] trafficPolicy: [e2e-llm-inference-service] tls: [e2e-llm-inference-service] insecureSkipVerify: true [e2e-llm-inference-service] mode: SIMPLE [e2e-llm-inference-service] sni: custom-route-timeout-test-kserve-workload-svc.kserve-ci-e2e-test.svc.cluster.local [e2e-llm-inference-service] [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1254 --- [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1255 apiVersion: networking.istio.io/v1alpha3 [e2e-llm-inference-service] kind: DestinationRule [e2e-llm-inference-service] metadata: [e2e-llm-inference-service] creationTimestamp: '2026-07-07T14:48:58Z' [e2e-llm-inference-service] generation: 1 [e2e-llm-inference-service] labels: [e2e-llm-inference-service] app.kubernetes.io/component: llminferenceservice-workload [e2e-llm-inference-service] app.kubernetes.io/name: custom-route-timeout-test [e2e-llm-inference-service] app.kubernetes.io/part-of: llminferenceservice [e2e-llm-inference-service] llm-d.ai/managed: 'true' [e2e-llm-inference-service] managedFields: [e2e-llm-inference-service] - apiVersion: networking.istio.io/v1 [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:metadata: [e2e-llm-inference-service] f:labels: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:app.kubernetes.io/component: {} [e2e-llm-inference-service] f:app.kubernetes.io/name: {} [e2e-llm-inference-service] f:app.kubernetes.io/part-of: {} [e2e-llm-inference-service] f:llm-d.ai/managed: {} [e2e-llm-inference-service] f:ownerReferences: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"uid":"986f2d2c-a6d1-463d-9850-3aad6459eace"}: {} [e2e-llm-inference-service] f:spec: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:exportTo: {} [e2e-llm-inference-service] f:host: {} [e2e-llm-inference-service] f:trafficPolicy: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:tls: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:caCertificates: {} [e2e-llm-inference-service] f:insecureSkipVerify: {} [e2e-llm-inference-service] f:mode: {} [e2e-llm-inference-service] f:sni: {} [e2e-llm-inference-service] manager: manager [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] time: '2026-07-07T14:48:58Z' [e2e-llm-inference-service] name: custom-route-timeout-test-kserve-workload-svc [e2e-llm-inference-service] namespace: kserve-ci-e2e-test [e2e-llm-inference-service] ownerReferences: [e2e-llm-inference-service] - apiVersion: serving.kserve.io/v1alpha2 [e2e-llm-inference-service] blockOwnerDeletion: true [e2e-llm-inference-service] controller: true [e2e-llm-inference-service] kind: LLMInferenceService [e2e-llm-inference-service] name: custom-route-timeout-test [e2e-llm-inference-service] uid: 986f2d2c-a6d1-463d-9850-3aad6459eace [e2e-llm-inference-service] resourceVersion: '49055' [e2e-llm-inference-service] uid: c3658218-7bc8-4350-8172-4da512823fec [e2e-llm-inference-service] spec: [e2e-llm-inference-service] exportTo: [e2e-llm-inference-service] - '*' [e2e-llm-inference-service] host: custom-route-timeout-test-kserve-workload-svc.kserve-ci-e2e-test.svc.cluster.local [e2e-llm-inference-service] trafficPolicy: [e2e-llm-inference-service] tls: [e2e-llm-inference-service] caCertificates: /var/run/secrets/kubernetes.io/serviceaccount/service-ca.crt [e2e-llm-inference-service] insecureSkipVerify: false [e2e-llm-inference-service] mode: SIMPLE [e2e-llm-inference-service] sni: custom-route-timeout-test-kserve-workload-svc.kserve-ci-e2e-test.svc.cluster.local [e2e-llm-inference-service] [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1254 --- [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1255 apiVersion: inference.networking.x-k8s.io/v1alpha2 [e2e-llm-inference-service] kind: InferencePool [e2e-llm-inference-service] metadata: [e2e-llm-inference-service] creationTimestamp: '2026-07-07T14:48:57Z' [e2e-llm-inference-service] generation: 1 [e2e-llm-inference-service] labels: [e2e-llm-inference-service] app.kubernetes.io/component: llminferenceservice-router-scheduler [e2e-llm-inference-service] app.kubernetes.io/name: custom-route-timeout-test [e2e-llm-inference-service] app.kubernetes.io/part-of: llminferenceservice [e2e-llm-inference-service] managedFields: [e2e-llm-inference-service] - apiVersion: inference.networking.x-k8s.io/v1alpha2 [e2e-llm-inference-service] fieldsType: FieldsV1 [e2e-llm-inference-service] fieldsV1: [e2e-llm-inference-service] f:metadata: [e2e-llm-inference-service] f:labels: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:app.kubernetes.io/component: {} [e2e-llm-inference-service] f:app.kubernetes.io/name: {} [e2e-llm-inference-service] f:app.kubernetes.io/part-of: {} [e2e-llm-inference-service] f:ownerReferences: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] k:{"uid":"986f2d2c-a6d1-463d-9850-3aad6459eace"}: {} [e2e-llm-inference-service] f:spec: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:extensionRef: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:failureMode: {} [e2e-llm-inference-service] f:group: {} [e2e-llm-inference-service] f:kind: {} [e2e-llm-inference-service] f:name: {} [e2e-llm-inference-service] f:portNumber: {} [e2e-llm-inference-service] f:selector: [e2e-llm-inference-service] .: {} [e2e-llm-inference-service] f:app.kubernetes.io/name: {} [e2e-llm-inference-service] f:app.kubernetes.io/part-of: {} [e2e-llm-inference-service] f:kserve.io/component: {} [e2e-llm-inference-service] f:targetPortNumber: {} [e2e-llm-inference-service] manager: manager [e2e-llm-inference-service] operation: Update [e2e-llm-inference-service] time: '2026-07-07T14:48:57Z' [e2e-llm-inference-service] name: custom-route-timeout-test-inference-pool [e2e-llm-inference-service] namespace: kserve-ci-e2e-test [e2e-llm-inference-service] ownerReferences: [e2e-llm-inference-service] - apiVersion: serving.kserve.io/v1alpha2 [e2e-llm-inference-service] blockOwnerDeletion: true [e2e-llm-inference-service] controller: true [e2e-llm-inference-service] kind: LLMInferenceService [e2e-llm-inference-service] name: custom-route-timeout-test [e2e-llm-inference-service] uid: 986f2d2c-a6d1-463d-9850-3aad6459eace [e2e-llm-inference-service] resourceVersion: '49010' [e2e-llm-inference-service] uid: 9229a052-2e49-43b5-b199-461339295f5a [e2e-llm-inference-service] spec: [e2e-llm-inference-service] extensionRef: [e2e-llm-inference-service] failureMode: FailOpen [e2e-llm-inference-service] group: '' [e2e-llm-inference-service] kind: Service [e2e-llm-inference-service] name: custom-route-timeout-test-epp-service [e2e-llm-inference-service] portNumber: 9002 [e2e-llm-inference-service] selector: [e2e-llm-inference-service] app.kubernetes.io/name: custom-route-timeout-test [e2e-llm-inference-service] app.kubernetes.io/part-of: llminferenceservice [e2e-llm-inference-service] kserve.io/component: workload [e2e-llm-inference-service] targetPortNumber: 8000 [e2e-llm-inference-service] status: [e2e-llm-inference-service] parent: [e2e-llm-inference-service] - conditions: [e2e-llm-inference-service] - lastTransitionTime: '1970-01-01T00:00:00Z' [e2e-llm-inference-service] message: Waiting for controller [e2e-llm-inference-service] reason: Pending [e2e-llm-inference-service] status: Unknown [e2e-llm-inference-service] type: Accepted [e2e-llm-inference-service] parentRef: [e2e-llm-inference-service] group: gateway.networking.k8s.io [e2e-llm-inference-service] kind: Status [e2e-llm-inference-service] name: default [e2e-llm-inference-service] [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1254 --- [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1255 metadata: [e2e-llm-inference-service] name: custom-route-timeout-test-kserve-7f56b8c469-hkpbl [e2e-llm-inference-service] namespace: kserve-ci-e2e-test [e2e-llm-inference-service] creationTimestamp: '2026-07-07T15:06:04Z' [e2e-llm-inference-service] labels: [e2e-llm-inference-service] app.kubernetes.io/component: llminferenceservice-workload [e2e-llm-inference-service] app.kubernetes.io/name: custom-route-timeout-test [e2e-llm-inference-service] app.kubernetes.io/part-of: llminferenceservice [e2e-llm-inference-service] kserve.io/component: workload [e2e-llm-inference-service] llm-d.ai/role: both [e2e-llm-inference-service] pod-template-hash: 7f56b8c469 [e2e-llm-inference-service] timestamp: '2026-07-07T15:05:54Z' [e2e-llm-inference-service] window: 20.768s [e2e-llm-inference-service] containers: [e2e-llm-inference-service] - name: main [e2e-llm-inference-service] usage: [e2e-llm-inference-service] cpu: 99853669n [e2e-llm-inference-service] memory: 2414156Ki [e2e-llm-inference-service] apiVersion: metrics.k8s.io/v1beta1 [e2e-llm-inference-service] kind: PodMetrics [e2e-llm-inference-service] [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1254 --- [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1255 metadata: [e2e-llm-inference-service] name: custom-route-timeout-test-kserve-router-scheduler-6578bddfnbxnb [e2e-llm-inference-service] namespace: kserve-ci-e2e-test [e2e-llm-inference-service] creationTimestamp: '2026-07-07T15:06:04Z' [e2e-llm-inference-service] labels: [e2e-llm-inference-service] app.kubernetes.io/component: llminferenceservice-router-scheduler [e2e-llm-inference-service] app.kubernetes.io/name: custom-route-timeout-test [e2e-llm-inference-service] app.kubernetes.io/part-of: llminferenceservice [e2e-llm-inference-service] pod-template-hash: 6578bddf5b [e2e-llm-inference-service] timestamp: '2026-07-07T15:05:49Z' [e2e-llm-inference-service] window: 14.133s [e2e-llm-inference-service] containers: [e2e-llm-inference-service] - name: main [e2e-llm-inference-service] usage: [e2e-llm-inference-service] cpu: 58725748n [e2e-llm-inference-service] memory: 31024Ki [e2e-llm-inference-service] apiVersion: metrics.k8s.io/v1beta1 [e2e-llm-inference-service] kind: PodMetrics [e2e-llm-inference-service] [e2e-llm-inference-service] ERROR e2e.llmisvc.logging:logging.py:48 [test_llm_inference_service] [2026-07-07T15:06:05.127617] end - ❌ 1036.467s: ❌ Failed to call model: HTTPConnectionPool(host='ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com', port=80): Max retries exceeded with url: /kserve-ci-e2e-test/custom-route-timeout-test/v1/completions (Caused by ReadTimeoutError("HTTPConnectionPool(host='ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com', port=80): Read timed out. (read timeout=60)")) [e2e-llm-inference-service] _ test_rolling_upgrade_coordination[router-managed-workload-llmd-simulator-model-fb-opt-125m] _ [e2e-llm-inference-service] [gw1] linux -- Python 3.11.13 /workspace/source/python/kserve/.venv/bin/python [e2e-llm-inference-service] [e2e-llm-inference-service] self = [e2e-llm-inference-service] conn = [e2e-llm-inference-service] method = 'POST', url = '/kserve-ci-e2e-test/rolling-upgrade-test/v1/completions' [e2e-llm-inference-service] body = b'{"model": "facebook/opt-125m", "prompt": "KServe is a", "max_tokens": 20}' [e2e-llm-inference-service] headers = {'User-Agent': 'python-requests/2.32.3', 'Accept-Encoding': 'gzip, deflate', 'Accept': '*/*', 'Connection': 'keep-alive', 'Content-Type': 'application/json', 'Content-Length': '73'} [e2e-llm-inference-service] retries = Retry(total=0, connect=None, read=None, redirect=None, status=None) [e2e-llm-inference-service] timeout = Timeout(connect=60, read=60, total=None), chunked = False [e2e-llm-inference-service] response_conn = [e2e-llm-inference-service] preload_content = False, decode_content = False, enforce_content_length = True [e2e-llm-inference-service] [e2e-llm-inference-service] def _make_request( [e2e-llm-inference-service] self, [e2e-llm-inference-service] conn: BaseHTTPConnection, [e2e-llm-inference-service] method: str, [e2e-llm-inference-service] url: str, [e2e-llm-inference-service] body: _TYPE_BODY | None = None, [e2e-llm-inference-service] headers: typing.Mapping[str, str] | None = None, [e2e-llm-inference-service] retries: Retry | None = None, [e2e-llm-inference-service] timeout: _TYPE_TIMEOUT = _DEFAULT_TIMEOUT, [e2e-llm-inference-service] chunked: bool = False, [e2e-llm-inference-service] response_conn: BaseHTTPConnection | None = None, [e2e-llm-inference-service] preload_content: bool = True, [e2e-llm-inference-service] decode_content: bool = True, [e2e-llm-inference-service] enforce_content_length: bool = True, [e2e-llm-inference-service] ) -> BaseHTTPResponse: [e2e-llm-inference-service] """ [e2e-llm-inference-service] Perform a request on a given urllib connection object taken from our [e2e-llm-inference-service] pool. [e2e-llm-inference-service] [e2e-llm-inference-service] :param conn: [e2e-llm-inference-service] a connection from one of our connection pools [e2e-llm-inference-service] [e2e-llm-inference-service] :param method: [e2e-llm-inference-service] HTTP request method (such as GET, POST, PUT, etc.) [e2e-llm-inference-service] [e2e-llm-inference-service] :param url: [e2e-llm-inference-service] The URL to perform the request on. [e2e-llm-inference-service] [e2e-llm-inference-service] :param body: [e2e-llm-inference-service] Data to send in the request body, either :class:`str`, :class:`bytes`, [e2e-llm-inference-service] an iterable of :class:`str`/:class:`bytes`, or a file-like object. [e2e-llm-inference-service] [e2e-llm-inference-service] :param headers: [e2e-llm-inference-service] Dictionary of custom headers to send, such as User-Agent, [e2e-llm-inference-service] If-None-Match, etc. If None, pool headers are used. If provided, [e2e-llm-inference-service] these headers completely replace any pool-specific headers. [e2e-llm-inference-service] [e2e-llm-inference-service] :param retries: [e2e-llm-inference-service] Configure the number of retries to allow before raising a [e2e-llm-inference-service] :class:`~urllib3.exceptions.MaxRetryError` exception. [e2e-llm-inference-service] [e2e-llm-inference-service] Pass ``None`` to retry until you receive a response. Pass a [e2e-llm-inference-service] :class:`~urllib3.util.retry.Retry` object for fine-grained control [e2e-llm-inference-service] over different types of retries. [e2e-llm-inference-service] Pass an integer number to retry connection errors that many times, [e2e-llm-inference-service] but no other types of errors. Pass zero to never retry. [e2e-llm-inference-service] [e2e-llm-inference-service] If ``False``, then retries are disabled and any exception is raised [e2e-llm-inference-service] immediately. Also, instead of raising a MaxRetryError on redirects, [e2e-llm-inference-service] the redirect response will be returned. [e2e-llm-inference-service] [e2e-llm-inference-service] :type retries: :class:`~urllib3.util.retry.Retry`, False, or an int. [e2e-llm-inference-service] [e2e-llm-inference-service] :param timeout: [e2e-llm-inference-service] If specified, overrides the default timeout for this one [e2e-llm-inference-service] request. It may be a float (in seconds) or an instance of [e2e-llm-inference-service] :class:`urllib3.util.Timeout`. [e2e-llm-inference-service] [e2e-llm-inference-service] :param chunked: [e2e-llm-inference-service] If True, urllib3 will send the body using chunked transfer [e2e-llm-inference-service] encoding. Otherwise, urllib3 will send the body using the standard [e2e-llm-inference-service] content-length form. Defaults to False. [e2e-llm-inference-service] [e2e-llm-inference-service] :param response_conn: [e2e-llm-inference-service] Set this to ``None`` if you will handle releasing the connection or [e2e-llm-inference-service] set the connection to have the response release it. [e2e-llm-inference-service] [e2e-llm-inference-service] :param preload_content: [e2e-llm-inference-service] If True, the response's body will be preloaded during construction. [e2e-llm-inference-service] [e2e-llm-inference-service] :param decode_content: [e2e-llm-inference-service] If True, will attempt to decode the body based on the [e2e-llm-inference-service] 'content-encoding' header. [e2e-llm-inference-service] [e2e-llm-inference-service] :param enforce_content_length: [e2e-llm-inference-service] Enforce content length checking. Body returned by server must match [e2e-llm-inference-service] value of Content-Length header, if present. Otherwise, raise error. [e2e-llm-inference-service] """ [e2e-llm-inference-service] self.num_requests += 1 [e2e-llm-inference-service] [e2e-llm-inference-service] timeout_obj = self._get_timeout(timeout) [e2e-llm-inference-service] timeout_obj.start_connect() [e2e-llm-inference-service] conn.timeout = Timeout.resolve_default_timeout(timeout_obj.connect_timeout) [e2e-llm-inference-service] [e2e-llm-inference-service] try: [e2e-llm-inference-service] # Trigger any extra validation we need to do. [e2e-llm-inference-service] try: [e2e-llm-inference-service] self._validate_conn(conn) [e2e-llm-inference-service] except (SocketTimeout, BaseSSLError) as e: [e2e-llm-inference-service] self._raise_timeout(err=e, url=url, timeout_value=conn.timeout) [e2e-llm-inference-service] raise [e2e-llm-inference-service] [e2e-llm-inference-service] # _validate_conn() starts the connection to an HTTPS proxy [e2e-llm-inference-service] # so we need to wrap errors with 'ProxyError' here too. [e2e-llm-inference-service] except ( [e2e-llm-inference-service] OSError, [e2e-llm-inference-service] NewConnectionError, [e2e-llm-inference-service] TimeoutError, [e2e-llm-inference-service] BaseSSLError, [e2e-llm-inference-service] CertificateError, [e2e-llm-inference-service] SSLError, [e2e-llm-inference-service] ) as e: [e2e-llm-inference-service] new_e: Exception = e [e2e-llm-inference-service] if isinstance(e, (BaseSSLError, CertificateError)): [e2e-llm-inference-service] new_e = SSLError(e) [e2e-llm-inference-service] # If the connection didn't successfully connect to it's proxy [e2e-llm-inference-service] # then there [e2e-llm-inference-service] if isinstance( [e2e-llm-inference-service] new_e, (OSError, NewConnectionError, TimeoutError, SSLError) [e2e-llm-inference-service] ) and (conn and conn.proxy and not conn.has_connected_to_proxy): [e2e-llm-inference-service] new_e = _wrap_proxy_error(new_e, conn.proxy.scheme) [e2e-llm-inference-service] raise new_e [e2e-llm-inference-service] [e2e-llm-inference-service] # conn.request() calls http.client.*.request, not the method in [e2e-llm-inference-service] # urllib3.request. It also calls makefile (recv) on the socket. [e2e-llm-inference-service] try: [e2e-llm-inference-service] conn.request( [e2e-llm-inference-service] method, [e2e-llm-inference-service] url, [e2e-llm-inference-service] body=body, [e2e-llm-inference-service] headers=headers, [e2e-llm-inference-service] chunked=chunked, [e2e-llm-inference-service] preload_content=preload_content, [e2e-llm-inference-service] decode_content=decode_content, [e2e-llm-inference-service] enforce_content_length=enforce_content_length, [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] # We are swallowing BrokenPipeError (errno.EPIPE) since the server is [e2e-llm-inference-service] # legitimately able to close the connection after sending a valid response. [e2e-llm-inference-service] # With this behaviour, the received response is still readable. [e2e-llm-inference-service] except BrokenPipeError: [e2e-llm-inference-service] pass [e2e-llm-inference-service] except OSError as e: [e2e-llm-inference-service] # MacOS/Linux [e2e-llm-inference-service] # EPROTOTYPE and ECONNRESET are needed on macOS [e2e-llm-inference-service] # https://erickt.github.io/blog/2014/11/19/adventures-in-debugging-a-potential-osx-kernel-bug/ [e2e-llm-inference-service] # Condition changed later to emit ECONNRESET instead of only EPROTOTYPE. [e2e-llm-inference-service] if e.errno != errno.EPROTOTYPE and e.errno != errno.ECONNRESET: [e2e-llm-inference-service] raise [e2e-llm-inference-service] [e2e-llm-inference-service] # Reset the timeout for the recv() on the socket [e2e-llm-inference-service] read_timeout = timeout_obj.read_timeout [e2e-llm-inference-service] [e2e-llm-inference-service] if not conn.is_closed: [e2e-llm-inference-service] # In Python 3 socket.py will catch EAGAIN and return None when you [e2e-llm-inference-service] # try and read into the file pointer created by http.client, which [e2e-llm-inference-service] # instead raises a BadStatusLine exception. Instead of catching [e2e-llm-inference-service] # the exception and assuming all BadStatusLine exceptions are read [e2e-llm-inference-service] # timeouts, check for a zero timeout before making the request. [e2e-llm-inference-service] if read_timeout == 0: [e2e-llm-inference-service] raise ReadTimeoutError( [e2e-llm-inference-service] self, url, f"Read timed out. (read timeout={read_timeout})" [e2e-llm-inference-service] ) [e2e-llm-inference-service] conn.timeout = read_timeout [e2e-llm-inference-service] [e2e-llm-inference-service] # Receive the response from the server [e2e-llm-inference-service] try: [e2e-llm-inference-service] > response = conn.getresponse() [e2e-llm-inference-service] [e2e-llm-inference-service] ../../python/kserve/.venv/lib64/python3.11/site-packages/urllib3/connectionpool.py:534: [e2e-llm-inference-service] _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ [e2e-llm-inference-service] [e2e-llm-inference-service] self = [e2e-llm-inference-service] [e2e-llm-inference-service] def getresponse( # type: ignore[override] [e2e-llm-inference-service] self, [e2e-llm-inference-service] ) -> HTTPResponse: [e2e-llm-inference-service] """ [e2e-llm-inference-service] Get the response from the server. [e2e-llm-inference-service] [e2e-llm-inference-service] If the HTTPConnection is in the correct state, returns an instance of HTTPResponse or of whatever object is returned by the response_class variable. [e2e-llm-inference-service] [e2e-llm-inference-service] If a request has not been sent or if a previous response has not be handled, ResponseNotReady is raised. If the HTTP response indicates that the connection should be closed, then it will be closed before the response is returned. When the connection is closed, the underlying socket is closed. [e2e-llm-inference-service] """ [e2e-llm-inference-service] # Raise the same error as http.client.HTTPConnection [e2e-llm-inference-service] if self._response_options is None: [e2e-llm-inference-service] raise ResponseNotReady() [e2e-llm-inference-service] [e2e-llm-inference-service] # Reset this attribute for being used again. [e2e-llm-inference-service] resp_options = self._response_options [e2e-llm-inference-service] self._response_options = None [e2e-llm-inference-service] [e2e-llm-inference-service] # Since the connection's timeout value may have been updated [e2e-llm-inference-service] # we need to set the timeout on the socket. [e2e-llm-inference-service] self.sock.settimeout(self.timeout) [e2e-llm-inference-service] [e2e-llm-inference-service] # This is needed here to avoid circular import errors [e2e-llm-inference-service] from .response import HTTPResponse [e2e-llm-inference-service] [e2e-llm-inference-service] # Save a reference to the shutdown function before ownership is passed [e2e-llm-inference-service] # to httplib_response [e2e-llm-inference-service] # TODO should we implement it everywhere? [e2e-llm-inference-service] _shutdown = getattr(self.sock, "shutdown", None) [e2e-llm-inference-service] [e2e-llm-inference-service] # Get the response from http.client.HTTPConnection [e2e-llm-inference-service] > httplib_response = super().getresponse() [e2e-llm-inference-service] [e2e-llm-inference-service] ../../python/kserve/.venv/lib64/python3.11/site-packages/urllib3/connection.py:571: [e2e-llm-inference-service] _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ [e2e-llm-inference-service] [e2e-llm-inference-service] self = [e2e-llm-inference-service] [e2e-llm-inference-service] def getresponse(self): [e2e-llm-inference-service] """Get the response from the server. [e2e-llm-inference-service] [e2e-llm-inference-service] If the HTTPConnection is in the correct state, returns an [e2e-llm-inference-service] instance of HTTPResponse or of whatever object is returned by [e2e-llm-inference-service] the response_class variable. [e2e-llm-inference-service] [e2e-llm-inference-service] If a request has not been sent or if a previous response has [e2e-llm-inference-service] not be handled, ResponseNotReady is raised. If the HTTP [e2e-llm-inference-service] response indicates that the connection should be closed, then [e2e-llm-inference-service] it will be closed before the response is returned. When the [e2e-llm-inference-service] connection is closed, the underlying socket is closed. [e2e-llm-inference-service] """ [e2e-llm-inference-service] [e2e-llm-inference-service] # if a prior response has been completed, then forget about it. [e2e-llm-inference-service] if self.__response and self.__response.isclosed(): [e2e-llm-inference-service] self.__response = None [e2e-llm-inference-service] [e2e-llm-inference-service] # if a prior response exists, then it must be completed (otherwise, we [e2e-llm-inference-service] # cannot read this response's header to determine the connection-close [e2e-llm-inference-service] # behavior) [e2e-llm-inference-service] # [e2e-llm-inference-service] # note: if a prior response existed, but was connection-close, then the [e2e-llm-inference-service] # socket and response were made independent of this HTTPConnection [e2e-llm-inference-service] # object since a new request requires that we open a whole new [e2e-llm-inference-service] # connection [e2e-llm-inference-service] # [e2e-llm-inference-service] # this means the prior response had one of two states: [e2e-llm-inference-service] # 1) will_close: this connection was reset and the prior socket and [e2e-llm-inference-service] # response operate independently [e2e-llm-inference-service] # 2) persistent: the response was retained and we await its [e2e-llm-inference-service] # isclosed() status to become true. [e2e-llm-inference-service] # [e2e-llm-inference-service] if self.__state != _CS_REQ_SENT or self.__response: [e2e-llm-inference-service] raise ResponseNotReady(self.__state) [e2e-llm-inference-service] [e2e-llm-inference-service] if self.debuglevel > 0: [e2e-llm-inference-service] response = self.response_class(self.sock, self.debuglevel, [e2e-llm-inference-service] method=self._method) [e2e-llm-inference-service] else: [e2e-llm-inference-service] response = self.response_class(self.sock, method=self._method) [e2e-llm-inference-service] [e2e-llm-inference-service] try: [e2e-llm-inference-service] try: [e2e-llm-inference-service] > response.begin() [e2e-llm-inference-service] [e2e-llm-inference-service] /usr/lib64/python3.11/http/client.py:1395: [e2e-llm-inference-service] _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ [e2e-llm-inference-service] [e2e-llm-inference-service] self = [e2e-llm-inference-service] [e2e-llm-inference-service] def begin(self): [e2e-llm-inference-service] if self.headers is not None: [e2e-llm-inference-service] # we've already started reading the response [e2e-llm-inference-service] return [e2e-llm-inference-service] [e2e-llm-inference-service] # read until we get a non-100 response [e2e-llm-inference-service] while True: [e2e-llm-inference-service] > version, status, reason = self._read_status() [e2e-llm-inference-service] [e2e-llm-inference-service] /usr/lib64/python3.11/http/client.py:325: [e2e-llm-inference-service] _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ [e2e-llm-inference-service] [e2e-llm-inference-service] self = [e2e-llm-inference-service] [e2e-llm-inference-service] def _read_status(self): [e2e-llm-inference-service] > line = str(self.fp.readline(_MAXLINE + 1), "iso-8859-1") [e2e-llm-inference-service] [e2e-llm-inference-service] /usr/lib64/python3.11/http/client.py:286: [e2e-llm-inference-service] _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ [e2e-llm-inference-service] [e2e-llm-inference-service] self = [e2e-llm-inference-service] b = [e2e-llm-inference-service] [e2e-llm-inference-service] def readinto(self, b): [e2e-llm-inference-service] """Read up to len(b) bytes into the writable buffer *b* and return [e2e-llm-inference-service] the number of bytes read. If the socket is non-blocking and no bytes [e2e-llm-inference-service] are available, None is returned. [e2e-llm-inference-service] [e2e-llm-inference-service] If *b* is non-empty, a 0 return value indicates that the connection [e2e-llm-inference-service] was shutdown at the other end. [e2e-llm-inference-service] """ [e2e-llm-inference-service] self._checkClosed() [e2e-llm-inference-service] self._checkReadable() [e2e-llm-inference-service] if self._timeout_occurred: [e2e-llm-inference-service] raise OSError("cannot read from timed out object") [e2e-llm-inference-service] while True: [e2e-llm-inference-service] try: [e2e-llm-inference-service] > return self._sock.recv_into(b) [e2e-llm-inference-service] E TimeoutError: timed out [e2e-llm-inference-service] [e2e-llm-inference-service] /usr/lib64/python3.11/socket.py:718: TimeoutError [e2e-llm-inference-service] [e2e-llm-inference-service] The above exception was the direct cause of the following exception: [e2e-llm-inference-service] [e2e-llm-inference-service] self = [e2e-llm-inference-service] method = 'POST', url = '/kserve-ci-e2e-test/rolling-upgrade-test/v1/completions' [e2e-llm-inference-service] body = b'{"model": "facebook/opt-125m", "prompt": "KServe is a", "max_tokens": 20}' [e2e-llm-inference-service] headers = {'User-Agent': 'python-requests/2.32.3', 'Accept-Encoding': 'gzip, deflate', 'Accept': '*/*', 'Connection': 'keep-alive', 'Content-Type': 'application/json', 'Content-Length': '73'} [e2e-llm-inference-service] retries = Retry(total=0, connect=None, read=None, redirect=None, status=None) [e2e-llm-inference-service] redirect = False, assert_same_host = False [e2e-llm-inference-service] timeout = Timeout(connect=60, read=60, total=None), pool_timeout = None [e2e-llm-inference-service] release_conn = False, chunked = False, body_pos = None, preload_content = False [e2e-llm-inference-service] decode_content = False, response_kw = {} [e2e-llm-inference-service] parsed_url = Url(scheme=None, auth=None, host=None, port=None, path='/kserve-ci-e2e-test/rolling-upgrade-test/v1/completions', query=None, fragment=None) [e2e-llm-inference-service] destination_scheme = None, conn = None, release_this_conn = True [e2e-llm-inference-service] http_tunnel_required = False, err = None, clean_exit = False [e2e-llm-inference-service] [e2e-llm-inference-service] def urlopen( # type: ignore[override] [e2e-llm-inference-service] self, [e2e-llm-inference-service] method: str, [e2e-llm-inference-service] url: str, [e2e-llm-inference-service] body: _TYPE_BODY | None = None, [e2e-llm-inference-service] headers: typing.Mapping[str, str] | None = None, [e2e-llm-inference-service] retries: Retry | bool | int | None = None, [e2e-llm-inference-service] redirect: bool = True, [e2e-llm-inference-service] assert_same_host: bool = True, [e2e-llm-inference-service] timeout: _TYPE_TIMEOUT = _DEFAULT_TIMEOUT, [e2e-llm-inference-service] pool_timeout: int | None = None, [e2e-llm-inference-service] release_conn: bool | None = None, [e2e-llm-inference-service] chunked: bool = False, [e2e-llm-inference-service] body_pos: _TYPE_BODY_POSITION | None = None, [e2e-llm-inference-service] preload_content: bool = True, [e2e-llm-inference-service] decode_content: bool = True, [e2e-llm-inference-service] **response_kw: typing.Any, [e2e-llm-inference-service] ) -> BaseHTTPResponse: [e2e-llm-inference-service] """ [e2e-llm-inference-service] Get a connection from the pool and perform an HTTP request. This is the [e2e-llm-inference-service] lowest level call for making a request, so you'll need to specify all [e2e-llm-inference-service] the raw details. [e2e-llm-inference-service] [e2e-llm-inference-service] .. note:: [e2e-llm-inference-service] [e2e-llm-inference-service] More commonly, it's appropriate to use a convenience method [e2e-llm-inference-service] such as :meth:`request`. [e2e-llm-inference-service] [e2e-llm-inference-service] .. note:: [e2e-llm-inference-service] [e2e-llm-inference-service] `release_conn` will only behave as expected if [e2e-llm-inference-service] `preload_content=False` because we want to make [e2e-llm-inference-service] `preload_content=False` the default behaviour someday soon without [e2e-llm-inference-service] breaking backwards compatibility. [e2e-llm-inference-service] [e2e-llm-inference-service] :param method: [e2e-llm-inference-service] HTTP request method (such as GET, POST, PUT, etc.) [e2e-llm-inference-service] [e2e-llm-inference-service] :param url: [e2e-llm-inference-service] The URL to perform the request on. [e2e-llm-inference-service] [e2e-llm-inference-service] :param body: [e2e-llm-inference-service] Data to send in the request body, either :class:`str`, :class:`bytes`, [e2e-llm-inference-service] an iterable of :class:`str`/:class:`bytes`, or a file-like object. [e2e-llm-inference-service] [e2e-llm-inference-service] :param headers: [e2e-llm-inference-service] Dictionary of custom headers to send, such as User-Agent, [e2e-llm-inference-service] If-None-Match, etc. If None, pool headers are used. If provided, [e2e-llm-inference-service] these headers completely replace any pool-specific headers. [e2e-llm-inference-service] [e2e-llm-inference-service] :param retries: [e2e-llm-inference-service] Configure the number of retries to allow before raising a [e2e-llm-inference-service] :class:`~urllib3.exceptions.MaxRetryError` exception. [e2e-llm-inference-service] [e2e-llm-inference-service] If ``None`` (default) will retry 3 times, see ``Retry.DEFAULT``. Pass a [e2e-llm-inference-service] :class:`~urllib3.util.retry.Retry` object for fine-grained control [e2e-llm-inference-service] over different types of retries. [e2e-llm-inference-service] Pass an integer number to retry connection errors that many times, [e2e-llm-inference-service] but no other types of errors. Pass zero to never retry. [e2e-llm-inference-service] [e2e-llm-inference-service] If ``False``, then retries are disabled and any exception is raised [e2e-llm-inference-service] immediately. Also, instead of raising a MaxRetryError on redirects, [e2e-llm-inference-service] the redirect response will be returned. [e2e-llm-inference-service] [e2e-llm-inference-service] :type retries: :class:`~urllib3.util.retry.Retry`, False, or an int. [e2e-llm-inference-service] [e2e-llm-inference-service] :param redirect: [e2e-llm-inference-service] If True, automatically handle redirects (status codes 301, 302, [e2e-llm-inference-service] 303, 307, 308). Each redirect counts as a retry. Disabling retries [e2e-llm-inference-service] will disable redirect, too. [e2e-llm-inference-service] [e2e-llm-inference-service] :param assert_same_host: [e2e-llm-inference-service] If ``True``, will make sure that the host of the pool requests is [e2e-llm-inference-service] consistent else will raise HostChangedError. When ``False``, you can [e2e-llm-inference-service] use the pool on an HTTP proxy and request foreign hosts. [e2e-llm-inference-service] [e2e-llm-inference-service] :param timeout: [e2e-llm-inference-service] If specified, overrides the default timeout for this one [e2e-llm-inference-service] request. It may be a float (in seconds) or an instance of [e2e-llm-inference-service] :class:`urllib3.util.Timeout`. [e2e-llm-inference-service] [e2e-llm-inference-service] :param pool_timeout: [e2e-llm-inference-service] If set and the pool is set to block=True, then this method will [e2e-llm-inference-service] block for ``pool_timeout`` seconds and raise EmptyPoolError if no [e2e-llm-inference-service] connection is available within the time period. [e2e-llm-inference-service] [e2e-llm-inference-service] :param bool preload_content: [e2e-llm-inference-service] If True, the response's body will be preloaded into memory. [e2e-llm-inference-service] [e2e-llm-inference-service] :param bool decode_content: [e2e-llm-inference-service] If True, will attempt to decode the body based on the [e2e-llm-inference-service] 'content-encoding' header. [e2e-llm-inference-service] [e2e-llm-inference-service] :param release_conn: [e2e-llm-inference-service] If False, then the urlopen call will not release the connection [e2e-llm-inference-service] back into the pool once a response is received (but will release if [e2e-llm-inference-service] you read the entire contents of the response such as when [e2e-llm-inference-service] `preload_content=True`). This is useful if you're not preloading [e2e-llm-inference-service] the response's content immediately. You will need to call [e2e-llm-inference-service] ``r.release_conn()`` on the response ``r`` to return the connection [e2e-llm-inference-service] back into the pool. If None, it takes the value of ``preload_content`` [e2e-llm-inference-service] which defaults to ``True``. [e2e-llm-inference-service] [e2e-llm-inference-service] :param bool chunked: [e2e-llm-inference-service] If True, urllib3 will send the body using chunked transfer [e2e-llm-inference-service] encoding. Otherwise, urllib3 will send the body using the standard [e2e-llm-inference-service] content-length form. Defaults to False. [e2e-llm-inference-service] [e2e-llm-inference-service] :param int body_pos: [e2e-llm-inference-service] Position to seek to in file-like body in the event of a retry or [e2e-llm-inference-service] redirect. Typically this won't need to be set because urllib3 will [e2e-llm-inference-service] auto-populate the value when needed. [e2e-llm-inference-service] """ [e2e-llm-inference-service] parsed_url = parse_url(url) [e2e-llm-inference-service] destination_scheme = parsed_url.scheme [e2e-llm-inference-service] [e2e-llm-inference-service] if headers is None: [e2e-llm-inference-service] headers = self.headers [e2e-llm-inference-service] [e2e-llm-inference-service] if not isinstance(retries, Retry): [e2e-llm-inference-service] retries = Retry.from_int(retries, redirect=redirect, default=self.retries) [e2e-llm-inference-service] [e2e-llm-inference-service] if release_conn is None: [e2e-llm-inference-service] release_conn = preload_content [e2e-llm-inference-service] [e2e-llm-inference-service] # Check host [e2e-llm-inference-service] if assert_same_host and not self.is_same_host(url): [e2e-llm-inference-service] raise HostChangedError(self, url, retries) [e2e-llm-inference-service] [e2e-llm-inference-service] # Ensure that the URL we're connecting to is properly encoded [e2e-llm-inference-service] if url.startswith("/"): [e2e-llm-inference-service] url = to_str(_encode_target(url)) [e2e-llm-inference-service] else: [e2e-llm-inference-service] url = to_str(parsed_url.url) [e2e-llm-inference-service] [e2e-llm-inference-service] conn = None [e2e-llm-inference-service] [e2e-llm-inference-service] # Track whether `conn` needs to be released before [e2e-llm-inference-service] # returning/raising/recursing. Update this variable if necessary, and [e2e-llm-inference-service] # leave `release_conn` constant throughout the function. That way, if [e2e-llm-inference-service] # the function recurses, the original value of `release_conn` will be [e2e-llm-inference-service] # passed down into the recursive call, and its value will be respected. [e2e-llm-inference-service] # [e2e-llm-inference-service] # See issue #651 [1] for details. [e2e-llm-inference-service] # [e2e-llm-inference-service] # [1] [e2e-llm-inference-service] release_this_conn = release_conn [e2e-llm-inference-service] [e2e-llm-inference-service] http_tunnel_required = connection_requires_http_tunnel( [e2e-llm-inference-service] self.proxy, self.proxy_config, destination_scheme [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] # Merge the proxy headers. Only done when not using HTTP CONNECT. We [e2e-llm-inference-service] # have to copy the headers dict so we can safely change it without those [e2e-llm-inference-service] # changes being reflected in anyone else's copy. [e2e-llm-inference-service] if not http_tunnel_required: [e2e-llm-inference-service] headers = headers.copy() # type: ignore[attr-defined] [e2e-llm-inference-service] headers.update(self.proxy_headers) # type: ignore[union-attr] [e2e-llm-inference-service] [e2e-llm-inference-service] # Must keep the exception bound to a separate variable or else Python 3 [e2e-llm-inference-service] # complains about UnboundLocalError. [e2e-llm-inference-service] err = None [e2e-llm-inference-service] [e2e-llm-inference-service] # Keep track of whether we cleanly exited the except block. This [e2e-llm-inference-service] # ensures we do proper cleanup in finally. [e2e-llm-inference-service] clean_exit = False [e2e-llm-inference-service] [e2e-llm-inference-service] # Rewind body position, if needed. Record current position [e2e-llm-inference-service] # for future rewinds in the event of a redirect/retry. [e2e-llm-inference-service] body_pos = set_file_position(body, body_pos) [e2e-llm-inference-service] [e2e-llm-inference-service] try: [e2e-llm-inference-service] # Request a connection from the queue. [e2e-llm-inference-service] timeout_obj = self._get_timeout(timeout) [e2e-llm-inference-service] conn = self._get_conn(timeout=pool_timeout) [e2e-llm-inference-service] [e2e-llm-inference-service] conn.timeout = timeout_obj.connect_timeout # type: ignore[assignment] [e2e-llm-inference-service] [e2e-llm-inference-service] # Is this a closed/new connection that requires CONNECT tunnelling? [e2e-llm-inference-service] if self.proxy is not None and http_tunnel_required and conn.is_closed: [e2e-llm-inference-service] try: [e2e-llm-inference-service] self._prepare_proxy(conn) [e2e-llm-inference-service] except (BaseSSLError, OSError, SocketTimeout) as e: [e2e-llm-inference-service] self._raise_timeout( [e2e-llm-inference-service] err=e, url=self.proxy.url, timeout_value=conn.timeout [e2e-llm-inference-service] ) [e2e-llm-inference-service] raise [e2e-llm-inference-service] [e2e-llm-inference-service] # If we're going to release the connection in ``finally:``, then [e2e-llm-inference-service] # the response doesn't need to know about the connection. Otherwise [e2e-llm-inference-service] # it will also try to release it and we'll have a double-release [e2e-llm-inference-service] # mess. [e2e-llm-inference-service] response_conn = conn if not release_conn else None [e2e-llm-inference-service] [e2e-llm-inference-service] # Make the request on the HTTPConnection object [e2e-llm-inference-service] > response = self._make_request( [e2e-llm-inference-service] conn, [e2e-llm-inference-service] method, [e2e-llm-inference-service] url, [e2e-llm-inference-service] timeout=timeout_obj, [e2e-llm-inference-service] body=body, [e2e-llm-inference-service] headers=headers, [e2e-llm-inference-service] chunked=chunked, [e2e-llm-inference-service] retries=retries, [e2e-llm-inference-service] response_conn=response_conn, [e2e-llm-inference-service] preload_content=preload_content, [e2e-llm-inference-service] decode_content=decode_content, [e2e-llm-inference-service] **response_kw, [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] ../../python/kserve/.venv/lib64/python3.11/site-packages/urllib3/connectionpool.py:787: [e2e-llm-inference-service] _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ [e2e-llm-inference-service] [e2e-llm-inference-service] self = [e2e-llm-inference-service] conn = [e2e-llm-inference-service] method = 'POST', url = '/kserve-ci-e2e-test/rolling-upgrade-test/v1/completions' [e2e-llm-inference-service] body = b'{"model": "facebook/opt-125m", "prompt": "KServe is a", "max_tokens": 20}' [e2e-llm-inference-service] headers = {'User-Agent': 'python-requests/2.32.3', 'Accept-Encoding': 'gzip, deflate', 'Accept': '*/*', 'Connection': 'keep-alive', 'Content-Type': 'application/json', 'Content-Length': '73'} [e2e-llm-inference-service] retries = Retry(total=0, connect=None, read=None, redirect=None, status=None) [e2e-llm-inference-service] timeout = Timeout(connect=60, read=60, total=None), chunked = False [e2e-llm-inference-service] response_conn = [e2e-llm-inference-service] preload_content = False, decode_content = False, enforce_content_length = True [e2e-llm-inference-service] [e2e-llm-inference-service] def _make_request( [e2e-llm-inference-service] self, [e2e-llm-inference-service] conn: BaseHTTPConnection, [e2e-llm-inference-service] method: str, [e2e-llm-inference-service] url: str, [e2e-llm-inference-service] body: _TYPE_BODY | None = None, [e2e-llm-inference-service] headers: typing.Mapping[str, str] | None = None, [e2e-llm-inference-service] retries: Retry | None = None, [e2e-llm-inference-service] timeout: _TYPE_TIMEOUT = _DEFAULT_TIMEOUT, [e2e-llm-inference-service] chunked: bool = False, [e2e-llm-inference-service] response_conn: BaseHTTPConnection | None = None, [e2e-llm-inference-service] preload_content: bool = True, [e2e-llm-inference-service] decode_content: bool = True, [e2e-llm-inference-service] enforce_content_length: bool = True, [e2e-llm-inference-service] ) -> BaseHTTPResponse: [e2e-llm-inference-service] """ [e2e-llm-inference-service] Perform a request on a given urllib connection object taken from our [e2e-llm-inference-service] pool. [e2e-llm-inference-service] [e2e-llm-inference-service] :param conn: [e2e-llm-inference-service] a connection from one of our connection pools [e2e-llm-inference-service] [e2e-llm-inference-service] :param method: [e2e-llm-inference-service] HTTP request method (such as GET, POST, PUT, etc.) [e2e-llm-inference-service] [e2e-llm-inference-service] :param url: [e2e-llm-inference-service] The URL to perform the request on. [e2e-llm-inference-service] [e2e-llm-inference-service] :param body: [e2e-llm-inference-service] Data to send in the request body, either :class:`str`, :class:`bytes`, [e2e-llm-inference-service] an iterable of :class:`str`/:class:`bytes`, or a file-like object. [e2e-llm-inference-service] [e2e-llm-inference-service] :param headers: [e2e-llm-inference-service] Dictionary of custom headers to send, such as User-Agent, [e2e-llm-inference-service] If-None-Match, etc. If None, pool headers are used. If provided, [e2e-llm-inference-service] these headers completely replace any pool-specific headers. [e2e-llm-inference-service] [e2e-llm-inference-service] :param retries: [e2e-llm-inference-service] Configure the number of retries to allow before raising a [e2e-llm-inference-service] :class:`~urllib3.exceptions.MaxRetryError` exception. [e2e-llm-inference-service] [e2e-llm-inference-service] Pass ``None`` to retry until you receive a response. Pass a [e2e-llm-inference-service] :class:`~urllib3.util.retry.Retry` object for fine-grained control [e2e-llm-inference-service] over different types of retries. [e2e-llm-inference-service] Pass an integer number to retry connection errors that many times, [e2e-llm-inference-service] but no other types of errors. Pass zero to never retry. [e2e-llm-inference-service] [e2e-llm-inference-service] If ``False``, then retries are disabled and any exception is raised [e2e-llm-inference-service] immediately. Also, instead of raising a MaxRetryError on redirects, [e2e-llm-inference-service] the redirect response will be returned. [e2e-llm-inference-service] [e2e-llm-inference-service] :type retries: :class:`~urllib3.util.retry.Retry`, False, or an int. [e2e-llm-inference-service] [e2e-llm-inference-service] :param timeout: [e2e-llm-inference-service] If specified, overrides the default timeout for this one [e2e-llm-inference-service] request. It may be a float (in seconds) or an instance of [e2e-llm-inference-service] :class:`urllib3.util.Timeout`. [e2e-llm-inference-service] [e2e-llm-inference-service] :param chunked: [e2e-llm-inference-service] If True, urllib3 will send the body using chunked transfer [e2e-llm-inference-service] encoding. Otherwise, urllib3 will send the body using the standard [e2e-llm-inference-service] content-length form. Defaults to False. [e2e-llm-inference-service] [e2e-llm-inference-service] :param response_conn: [e2e-llm-inference-service] Set this to ``None`` if you will handle releasing the connection or [e2e-llm-inference-service] set the connection to have the response release it. [e2e-llm-inference-service] [e2e-llm-inference-service] :param preload_content: [e2e-llm-inference-service] If True, the response's body will be preloaded during construction. [e2e-llm-inference-service] [e2e-llm-inference-service] :param decode_content: [e2e-llm-inference-service] If True, will attempt to decode the body based on the [e2e-llm-inference-service] 'content-encoding' header. [e2e-llm-inference-service] [e2e-llm-inference-service] :param enforce_content_length: [e2e-llm-inference-service] Enforce content length checking. Body returned by server must match [e2e-llm-inference-service] value of Content-Length header, if present. Otherwise, raise error. [e2e-llm-inference-service] """ [e2e-llm-inference-service] self.num_requests += 1 [e2e-llm-inference-service] [e2e-llm-inference-service] timeout_obj = self._get_timeout(timeout) [e2e-llm-inference-service] timeout_obj.start_connect() [e2e-llm-inference-service] conn.timeout = Timeout.resolve_default_timeout(timeout_obj.connect_timeout) [e2e-llm-inference-service] [e2e-llm-inference-service] try: [e2e-llm-inference-service] # Trigger any extra validation we need to do. [e2e-llm-inference-service] try: [e2e-llm-inference-service] self._validate_conn(conn) [e2e-llm-inference-service] except (SocketTimeout, BaseSSLError) as e: [e2e-llm-inference-service] self._raise_timeout(err=e, url=url, timeout_value=conn.timeout) [e2e-llm-inference-service] raise [e2e-llm-inference-service] [e2e-llm-inference-service] # _validate_conn() starts the connection to an HTTPS proxy [e2e-llm-inference-service] # so we need to wrap errors with 'ProxyError' here too. [e2e-llm-inference-service] except ( [e2e-llm-inference-service] OSError, [e2e-llm-inference-service] NewConnectionError, [e2e-llm-inference-service] TimeoutError, [e2e-llm-inference-service] BaseSSLError, [e2e-llm-inference-service] CertificateError, [e2e-llm-inference-service] SSLError, [e2e-llm-inference-service] ) as e: [e2e-llm-inference-service] new_e: Exception = e [e2e-llm-inference-service] if isinstance(e, (BaseSSLError, CertificateError)): [e2e-llm-inference-service] new_e = SSLError(e) [e2e-llm-inference-service] # If the connection didn't successfully connect to it's proxy [e2e-llm-inference-service] # then there [e2e-llm-inference-service] if isinstance( [e2e-llm-inference-service] new_e, (OSError, NewConnectionError, TimeoutError, SSLError) [e2e-llm-inference-service] ) and (conn and conn.proxy and not conn.has_connected_to_proxy): [e2e-llm-inference-service] new_e = _wrap_proxy_error(new_e, conn.proxy.scheme) [e2e-llm-inference-service] raise new_e [e2e-llm-inference-service] [e2e-llm-inference-service] # conn.request() calls http.client.*.request, not the method in [e2e-llm-inference-service] # urllib3.request. It also calls makefile (recv) on the socket. [e2e-llm-inference-service] try: [e2e-llm-inference-service] conn.request( [e2e-llm-inference-service] method, [e2e-llm-inference-service] url, [e2e-llm-inference-service] body=body, [e2e-llm-inference-service] headers=headers, [e2e-llm-inference-service] chunked=chunked, [e2e-llm-inference-service] preload_content=preload_content, [e2e-llm-inference-service] decode_content=decode_content, [e2e-llm-inference-service] enforce_content_length=enforce_content_length, [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] # We are swallowing BrokenPipeError (errno.EPIPE) since the server is [e2e-llm-inference-service] # legitimately able to close the connection after sending a valid response. [e2e-llm-inference-service] # With this behaviour, the received response is still readable. [e2e-llm-inference-service] except BrokenPipeError: [e2e-llm-inference-service] pass [e2e-llm-inference-service] except OSError as e: [e2e-llm-inference-service] # MacOS/Linux [e2e-llm-inference-service] # EPROTOTYPE and ECONNRESET are needed on macOS [e2e-llm-inference-service] # https://erickt.github.io/blog/2014/11/19/adventures-in-debugging-a-potential-osx-kernel-bug/ [e2e-llm-inference-service] # Condition changed later to emit ECONNRESET instead of only EPROTOTYPE. [e2e-llm-inference-service] if e.errno != errno.EPROTOTYPE and e.errno != errno.ECONNRESET: [e2e-llm-inference-service] raise [e2e-llm-inference-service] [e2e-llm-inference-service] # Reset the timeout for the recv() on the socket [e2e-llm-inference-service] read_timeout = timeout_obj.read_timeout [e2e-llm-inference-service] [e2e-llm-inference-service] if not conn.is_closed: [e2e-llm-inference-service] # In Python 3 socket.py will catch EAGAIN and return None when you [e2e-llm-inference-service] # try and read into the file pointer created by http.client, which [e2e-llm-inference-service] # instead raises a BadStatusLine exception. Instead of catching [e2e-llm-inference-service] # the exception and assuming all BadStatusLine exceptions are read [e2e-llm-inference-service] # timeouts, check for a zero timeout before making the request. [e2e-llm-inference-service] if read_timeout == 0: [e2e-llm-inference-service] raise ReadTimeoutError( [e2e-llm-inference-service] self, url, f"Read timed out. (read timeout={read_timeout})" [e2e-llm-inference-service] ) [e2e-llm-inference-service] conn.timeout = read_timeout [e2e-llm-inference-service] [e2e-llm-inference-service] # Receive the response from the server [e2e-llm-inference-service] try: [e2e-llm-inference-service] response = conn.getresponse() [e2e-llm-inference-service] except (BaseSSLError, OSError) as e: [e2e-llm-inference-service] > self._raise_timeout(err=e, url=url, timeout_value=read_timeout) [e2e-llm-inference-service] [e2e-llm-inference-service] ../../python/kserve/.venv/lib64/python3.11/site-packages/urllib3/connectionpool.py:536: [e2e-llm-inference-service] _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ [e2e-llm-inference-service] [e2e-llm-inference-service] self = [e2e-llm-inference-service] err = TimeoutError('timed out') [e2e-llm-inference-service] url = '/kserve-ci-e2e-test/rolling-upgrade-test/v1/completions' [e2e-llm-inference-service] timeout_value = 60 [e2e-llm-inference-service] [e2e-llm-inference-service] def _raise_timeout( [e2e-llm-inference-service] self, [e2e-llm-inference-service] err: BaseSSLError | OSError | SocketTimeout, [e2e-llm-inference-service] url: str, [e2e-llm-inference-service] timeout_value: _TYPE_TIMEOUT | None, [e2e-llm-inference-service] ) -> None: [e2e-llm-inference-service] """Is the error actually a timeout? Will raise a ReadTimeout or pass""" [e2e-llm-inference-service] [e2e-llm-inference-service] if isinstance(err, SocketTimeout): [e2e-llm-inference-service] > raise ReadTimeoutError( [e2e-llm-inference-service] self, url, f"Read timed out. (read timeout={timeout_value})" [e2e-llm-inference-service] ) from err [e2e-llm-inference-service] E urllib3.exceptions.ReadTimeoutError: HTTPConnectionPool(host='ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com', port=80): Read timed out. (read timeout=60) [e2e-llm-inference-service] [e2e-llm-inference-service] ../../python/kserve/.venv/lib64/python3.11/site-packages/urllib3/connectionpool.py:367: ReadTimeoutError [e2e-llm-inference-service] [e2e-llm-inference-service] The above exception was the direct cause of the following exception: [e2e-llm-inference-service] [e2e-llm-inference-service] self = [e2e-llm-inference-service] request = , stream = False [e2e-llm-inference-service] timeout = Timeout(connect=60, read=60, total=None), verify = '/tmp/ca.crt' [e2e-llm-inference-service] cert = None, proxies = OrderedDict() [e2e-llm-inference-service] [e2e-llm-inference-service] def send( [e2e-llm-inference-service] self, request, stream=False, timeout=None, verify=True, cert=None, proxies=None [e2e-llm-inference-service] ): [e2e-llm-inference-service] """Sends PreparedRequest object. Returns Response object. [e2e-llm-inference-service] [e2e-llm-inference-service] :param request: The :class:`PreparedRequest ` being sent. [e2e-llm-inference-service] :param stream: (optional) Whether to stream the request content. [e2e-llm-inference-service] :param timeout: (optional) How long to wait for the server to send [e2e-llm-inference-service] data before giving up, as a float, or a :ref:`(connect timeout, [e2e-llm-inference-service] read timeout) ` tuple. [e2e-llm-inference-service] :type timeout: float or tuple or urllib3 Timeout object [e2e-llm-inference-service] :param verify: (optional) Either a boolean, in which case it controls whether [e2e-llm-inference-service] we verify the server's TLS certificate, or a string, in which case it [e2e-llm-inference-service] must be a path to a CA bundle to use [e2e-llm-inference-service] :param cert: (optional) Any user-provided SSL certificate to be trusted. [e2e-llm-inference-service] :param proxies: (optional) The proxies dictionary to apply to the request. [e2e-llm-inference-service] :rtype: requests.Response [e2e-llm-inference-service] """ [e2e-llm-inference-service] [e2e-llm-inference-service] try: [e2e-llm-inference-service] conn = self.get_connection_with_tls_context( [e2e-llm-inference-service] request, verify, proxies=proxies, cert=cert [e2e-llm-inference-service] ) [e2e-llm-inference-service] except LocationValueError as e: [e2e-llm-inference-service] raise InvalidURL(e, request=request) [e2e-llm-inference-service] [e2e-llm-inference-service] self.cert_verify(conn, request.url, verify, cert) [e2e-llm-inference-service] url = self.request_url(request, proxies) [e2e-llm-inference-service] self.add_headers( [e2e-llm-inference-service] request, [e2e-llm-inference-service] stream=stream, [e2e-llm-inference-service] timeout=timeout, [e2e-llm-inference-service] verify=verify, [e2e-llm-inference-service] cert=cert, [e2e-llm-inference-service] proxies=proxies, [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] chunked = not (request.body is None or "Content-Length" in request.headers) [e2e-llm-inference-service] [e2e-llm-inference-service] if isinstance(timeout, tuple): [e2e-llm-inference-service] try: [e2e-llm-inference-service] connect, read = timeout [e2e-llm-inference-service] timeout = TimeoutSauce(connect=connect, read=read) [e2e-llm-inference-service] except ValueError: [e2e-llm-inference-service] raise ValueError( [e2e-llm-inference-service] f"Invalid timeout {timeout}. Pass a (connect, read) timeout tuple, " [e2e-llm-inference-service] f"or a single float to set both timeouts to the same value." [e2e-llm-inference-service] ) [e2e-llm-inference-service] elif isinstance(timeout, TimeoutSauce): [e2e-llm-inference-service] pass [e2e-llm-inference-service] else: [e2e-llm-inference-service] timeout = TimeoutSauce(connect=timeout, read=timeout) [e2e-llm-inference-service] [e2e-llm-inference-service] try: [e2e-llm-inference-service] > resp = conn.urlopen( [e2e-llm-inference-service] method=request.method, [e2e-llm-inference-service] url=url, [e2e-llm-inference-service] body=request.body, [e2e-llm-inference-service] headers=request.headers, [e2e-llm-inference-service] redirect=False, [e2e-llm-inference-service] assert_same_host=False, [e2e-llm-inference-service] preload_content=False, [e2e-llm-inference-service] decode_content=False, [e2e-llm-inference-service] retries=self.max_retries, [e2e-llm-inference-service] timeout=timeout, [e2e-llm-inference-service] chunked=chunked, [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] ../../python/kserve/.venv/lib64/python3.11/site-packages/requests/adapters.py:667: [e2e-llm-inference-service] _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ [e2e-llm-inference-service] [e2e-llm-inference-service] self = [e2e-llm-inference-service] method = 'POST', url = '/kserve-ci-e2e-test/rolling-upgrade-test/v1/completions' [e2e-llm-inference-service] body = b'{"model": "facebook/opt-125m", "prompt": "KServe is a", "max_tokens": 20}' [e2e-llm-inference-service] headers = {'User-Agent': 'python-requests/2.32.3', 'Accept-Encoding': 'gzip, deflate', 'Accept': '*/*', 'Connection': 'keep-alive', 'Content-Type': 'application/json', 'Content-Length': '73'} [e2e-llm-inference-service] retries = Retry(total=7, connect=None, read=None, redirect=None, status=None) [e2e-llm-inference-service] redirect = False, assert_same_host = False [e2e-llm-inference-service] timeout = Timeout(connect=60, read=60, total=None), pool_timeout = None [e2e-llm-inference-service] release_conn = False, chunked = False, body_pos = None, preload_content = False [e2e-llm-inference-service] decode_content = False, response_kw = {} [e2e-llm-inference-service] parsed_url = Url(scheme=None, auth=None, host=None, port=None, path='/kserve-ci-e2e-test/rolling-upgrade-test/v1/completions', query=None, fragment=None) [e2e-llm-inference-service] destination_scheme = None, conn = None, release_this_conn = True [e2e-llm-inference-service] http_tunnel_required = False [e2e-llm-inference-service] err = ReadTimeoutError("HTTPConnectionPool(host='ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com', port=80): Read timed out. (read timeout=60)") [e2e-llm-inference-service] clean_exit = False [e2e-llm-inference-service] [e2e-llm-inference-service] def urlopen( # type: ignore[override] [e2e-llm-inference-service] self, [e2e-llm-inference-service] method: str, [e2e-llm-inference-service] url: str, [e2e-llm-inference-service] body: _TYPE_BODY | None = None, [e2e-llm-inference-service] headers: typing.Mapping[str, str] | None = None, [e2e-llm-inference-service] retries: Retry | bool | int | None = None, [e2e-llm-inference-service] redirect: bool = True, [e2e-llm-inference-service] assert_same_host: bool = True, [e2e-llm-inference-service] timeout: _TYPE_TIMEOUT = _DEFAULT_TIMEOUT, [e2e-llm-inference-service] pool_timeout: int | None = None, [e2e-llm-inference-service] release_conn: bool | None = None, [e2e-llm-inference-service] chunked: bool = False, [e2e-llm-inference-service] body_pos: _TYPE_BODY_POSITION | None = None, [e2e-llm-inference-service] preload_content: bool = True, [e2e-llm-inference-service] decode_content: bool = True, [e2e-llm-inference-service] **response_kw: typing.Any, [e2e-llm-inference-service] ) -> BaseHTTPResponse: [e2e-llm-inference-service] """ [e2e-llm-inference-service] Get a connection from the pool and perform an HTTP request. This is the [e2e-llm-inference-service] lowest level call for making a request, so you'll need to specify all [e2e-llm-inference-service] the raw details. [e2e-llm-inference-service] [e2e-llm-inference-service] .. note:: [e2e-llm-inference-service] [e2e-llm-inference-service] More commonly, it's appropriate to use a convenience method [e2e-llm-inference-service] such as :meth:`request`. [e2e-llm-inference-service] [e2e-llm-inference-service] .. note:: [e2e-llm-inference-service] [e2e-llm-inference-service] `release_conn` will only behave as expected if [e2e-llm-inference-service] `preload_content=False` because we want to make [e2e-llm-inference-service] `preload_content=False` the default behaviour someday soon without [e2e-llm-inference-service] breaking backwards compatibility. [e2e-llm-inference-service] [e2e-llm-inference-service] :param method: [e2e-llm-inference-service] HTTP request method (such as GET, POST, PUT, etc.) [e2e-llm-inference-service] [e2e-llm-inference-service] :param url: [e2e-llm-inference-service] The URL to perform the request on. [e2e-llm-inference-service] [e2e-llm-inference-service] :param body: [e2e-llm-inference-service] Data to send in the request body, either :class:`str`, :class:`bytes`, [e2e-llm-inference-service] an iterable of :class:`str`/:class:`bytes`, or a file-like object. [e2e-llm-inference-service] [e2e-llm-inference-service] :param headers: [e2e-llm-inference-service] Dictionary of custom headers to send, such as User-Agent, [e2e-llm-inference-service] If-None-Match, etc. If None, pool headers are used. If provided, [e2e-llm-inference-service] these headers completely replace any pool-specific headers. [e2e-llm-inference-service] [e2e-llm-inference-service] :param retries: [e2e-llm-inference-service] Configure the number of retries to allow before raising a [e2e-llm-inference-service] :class:`~urllib3.exceptions.MaxRetryError` exception. [e2e-llm-inference-service] [e2e-llm-inference-service] If ``None`` (default) will retry 3 times, see ``Retry.DEFAULT``. Pass a [e2e-llm-inference-service] :class:`~urllib3.util.retry.Retry` object for fine-grained control [e2e-llm-inference-service] over different types of retries. [e2e-llm-inference-service] Pass an integer number to retry connection errors that many times, [e2e-llm-inference-service] but no other types of errors. Pass zero to never retry. [e2e-llm-inference-service] [e2e-llm-inference-service] If ``False``, then retries are disabled and any exception is raised [e2e-llm-inference-service] immediately. Also, instead of raising a MaxRetryError on redirects, [e2e-llm-inference-service] the redirect response will be returned. [e2e-llm-inference-service] [e2e-llm-inference-service] :type retries: :class:`~urllib3.util.retry.Retry`, False, or an int. [e2e-llm-inference-service] [e2e-llm-inference-service] :param redirect: [e2e-llm-inference-service] If True, automatically handle redirects (status codes 301, 302, [e2e-llm-inference-service] 303, 307, 308). Each redirect counts as a retry. Disabling retries [e2e-llm-inference-service] will disable redirect, too. [e2e-llm-inference-service] [e2e-llm-inference-service] :param assert_same_host: [e2e-llm-inference-service] If ``True``, will make sure that the host of the pool requests is [e2e-llm-inference-service] consistent else will raise HostChangedError. When ``False``, you can [e2e-llm-inference-service] use the pool on an HTTP proxy and request foreign hosts. [e2e-llm-inference-service] [e2e-llm-inference-service] :param timeout: [e2e-llm-inference-service] If specified, overrides the default timeout for this one [e2e-llm-inference-service] request. It may be a float (in seconds) or an instance of [e2e-llm-inference-service] :class:`urllib3.util.Timeout`. [e2e-llm-inference-service] [e2e-llm-inference-service] :param pool_timeout: [e2e-llm-inference-service] If set and the pool is set to block=True, then this method will [e2e-llm-inference-service] block for ``pool_timeout`` seconds and raise EmptyPoolError if no [e2e-llm-inference-service] connection is available within the time period. [e2e-llm-inference-service] [e2e-llm-inference-service] :param bool preload_content: [e2e-llm-inference-service] If True, the response's body will be preloaded into memory. [e2e-llm-inference-service] [e2e-llm-inference-service] :param bool decode_content: [e2e-llm-inference-service] If True, will attempt to decode the body based on the [e2e-llm-inference-service] 'content-encoding' header. [e2e-llm-inference-service] [e2e-llm-inference-service] :param release_conn: [e2e-llm-inference-service] If False, then the urlopen call will not release the connection [e2e-llm-inference-service] back into the pool once a response is received (but will release if [e2e-llm-inference-service] you read the entire contents of the response such as when [e2e-llm-inference-service] `preload_content=True`). This is useful if you're not preloading [e2e-llm-inference-service] the response's content immediately. You will need to call [e2e-llm-inference-service] ``r.release_conn()`` on the response ``r`` to return the connection [e2e-llm-inference-service] back into the pool. If None, it takes the value of ``preload_content`` [e2e-llm-inference-service] which defaults to ``True``. [e2e-llm-inference-service] [e2e-llm-inference-service] :param bool chunked: [e2e-llm-inference-service] If True, urllib3 will send the body using chunked transfer [e2e-llm-inference-service] encoding. Otherwise, urllib3 will send the body using the standard [e2e-llm-inference-service] content-length form. Defaults to False. [e2e-llm-inference-service] [e2e-llm-inference-service] :param int body_pos: [e2e-llm-inference-service] Position to seek to in file-like body in the event of a retry or [e2e-llm-inference-service] redirect. Typically this won't need to be set because urllib3 will [e2e-llm-inference-service] auto-populate the value when needed. [e2e-llm-inference-service] """ [e2e-llm-inference-service] parsed_url = parse_url(url) [e2e-llm-inference-service] destination_scheme = parsed_url.scheme [e2e-llm-inference-service] [e2e-llm-inference-service] if headers is None: [e2e-llm-inference-service] headers = self.headers [e2e-llm-inference-service] [e2e-llm-inference-service] if not isinstance(retries, Retry): [e2e-llm-inference-service] retries = Retry.from_int(retries, redirect=redirect, default=self.retries) [e2e-llm-inference-service] [e2e-llm-inference-service] if release_conn is None: [e2e-llm-inference-service] release_conn = preload_content [e2e-llm-inference-service] [e2e-llm-inference-service] # Check host [e2e-llm-inference-service] if assert_same_host and not self.is_same_host(url): [e2e-llm-inference-service] raise HostChangedError(self, url, retries) [e2e-llm-inference-service] [e2e-llm-inference-service] # Ensure that the URL we're connecting to is properly encoded [e2e-llm-inference-service] if url.startswith("/"): [e2e-llm-inference-service] url = to_str(_encode_target(url)) [e2e-llm-inference-service] else: [e2e-llm-inference-service] url = to_str(parsed_url.url) [e2e-llm-inference-service] [e2e-llm-inference-service] conn = None [e2e-llm-inference-service] [e2e-llm-inference-service] # Track whether `conn` needs to be released before [e2e-llm-inference-service] # returning/raising/recursing. Update this variable if necessary, and [e2e-llm-inference-service] # leave `release_conn` constant throughout the function. That way, if [e2e-llm-inference-service] # the function recurses, the original value of `release_conn` will be [e2e-llm-inference-service] # passed down into the recursive call, and its value will be respected. [e2e-llm-inference-service] # [e2e-llm-inference-service] # See issue #651 [1] for details. [e2e-llm-inference-service] # [e2e-llm-inference-service] # [1] [e2e-llm-inference-service] release_this_conn = release_conn [e2e-llm-inference-service] [e2e-llm-inference-service] http_tunnel_required = connection_requires_http_tunnel( [e2e-llm-inference-service] self.proxy, self.proxy_config, destination_scheme [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] # Merge the proxy headers. Only done when not using HTTP CONNECT. We [e2e-llm-inference-service] # have to copy the headers dict so we can safely change it without those [e2e-llm-inference-service] # changes being reflected in anyone else's copy. [e2e-llm-inference-service] if not http_tunnel_required: [e2e-llm-inference-service] headers = headers.copy() # type: ignore[attr-defined] [e2e-llm-inference-service] headers.update(self.proxy_headers) # type: ignore[union-attr] [e2e-llm-inference-service] [e2e-llm-inference-service] # Must keep the exception bound to a separate variable or else Python 3 [e2e-llm-inference-service] # complains about UnboundLocalError. [e2e-llm-inference-service] err = None [e2e-llm-inference-service] [e2e-llm-inference-service] # Keep track of whether we cleanly exited the except block. This [e2e-llm-inference-service] # ensures we do proper cleanup in finally. [e2e-llm-inference-service] clean_exit = False [e2e-llm-inference-service] [e2e-llm-inference-service] # Rewind body position, if needed. Record current position [e2e-llm-inference-service] # for future rewinds in the event of a redirect/retry. [e2e-llm-inference-service] body_pos = set_file_position(body, body_pos) [e2e-llm-inference-service] [e2e-llm-inference-service] try: [e2e-llm-inference-service] # Request a connection from the queue. [e2e-llm-inference-service] timeout_obj = self._get_timeout(timeout) [e2e-llm-inference-service] conn = self._get_conn(timeout=pool_timeout) [e2e-llm-inference-service] [e2e-llm-inference-service] conn.timeout = timeout_obj.connect_timeout # type: ignore[assignment] [e2e-llm-inference-service] [e2e-llm-inference-service] # Is this a closed/new connection that requires CONNECT tunnelling? [e2e-llm-inference-service] if self.proxy is not None and http_tunnel_required and conn.is_closed: [e2e-llm-inference-service] try: [e2e-llm-inference-service] self._prepare_proxy(conn) [e2e-llm-inference-service] except (BaseSSLError, OSError, SocketTimeout) as e: [e2e-llm-inference-service] self._raise_timeout( [e2e-llm-inference-service] err=e, url=self.proxy.url, timeout_value=conn.timeout [e2e-llm-inference-service] ) [e2e-llm-inference-service] raise [e2e-llm-inference-service] [e2e-llm-inference-service] # If we're going to release the connection in ``finally:``, then [e2e-llm-inference-service] # the response doesn't need to know about the connection. Otherwise [e2e-llm-inference-service] # it will also try to release it and we'll have a double-release [e2e-llm-inference-service] # mess. [e2e-llm-inference-service] response_conn = conn if not release_conn else None [e2e-llm-inference-service] [e2e-llm-inference-service] # Make the request on the HTTPConnection object [e2e-llm-inference-service] response = self._make_request( [e2e-llm-inference-service] conn, [e2e-llm-inference-service] method, [e2e-llm-inference-service] url, [e2e-llm-inference-service] timeout=timeout_obj, [e2e-llm-inference-service] body=body, [e2e-llm-inference-service] headers=headers, [e2e-llm-inference-service] chunked=chunked, [e2e-llm-inference-service] retries=retries, [e2e-llm-inference-service] response_conn=response_conn, [e2e-llm-inference-service] preload_content=preload_content, [e2e-llm-inference-service] decode_content=decode_content, [e2e-llm-inference-service] **response_kw, [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] # Everything went great! [e2e-llm-inference-service] clean_exit = True [e2e-llm-inference-service] [e2e-llm-inference-service] except EmptyPoolError: [e2e-llm-inference-service] # Didn't get a connection from the pool, no need to clean up [e2e-llm-inference-service] clean_exit = True [e2e-llm-inference-service] release_this_conn = False [e2e-llm-inference-service] raise [e2e-llm-inference-service] [e2e-llm-inference-service] except ( [e2e-llm-inference-service] TimeoutError, [e2e-llm-inference-service] HTTPException, [e2e-llm-inference-service] OSError, [e2e-llm-inference-service] ProtocolError, [e2e-llm-inference-service] BaseSSLError, [e2e-llm-inference-service] SSLError, [e2e-llm-inference-service] CertificateError, [e2e-llm-inference-service] ProxyError, [e2e-llm-inference-service] ) as e: [e2e-llm-inference-service] # Discard the connection for these exceptions. It will be [e2e-llm-inference-service] # replaced during the next _get_conn() call. [e2e-llm-inference-service] clean_exit = False [e2e-llm-inference-service] new_e: Exception = e [e2e-llm-inference-service] if isinstance(e, (BaseSSLError, CertificateError)): [e2e-llm-inference-service] new_e = SSLError(e) [e2e-llm-inference-service] if isinstance( [e2e-llm-inference-service] new_e, [e2e-llm-inference-service] ( [e2e-llm-inference-service] OSError, [e2e-llm-inference-service] NewConnectionError, [e2e-llm-inference-service] TimeoutError, [e2e-llm-inference-service] SSLError, [e2e-llm-inference-service] HTTPException, [e2e-llm-inference-service] ), [e2e-llm-inference-service] ) and (conn and conn.proxy and not conn.has_connected_to_proxy): [e2e-llm-inference-service] new_e = _wrap_proxy_error(new_e, conn.proxy.scheme) [e2e-llm-inference-service] elif isinstance(new_e, (OSError, HTTPException)): [e2e-llm-inference-service] new_e = ProtocolError("Connection aborted.", new_e) [e2e-llm-inference-service] [e2e-llm-inference-service] retries = retries.increment( [e2e-llm-inference-service] method, url, error=new_e, _pool=self, _stacktrace=sys.exc_info()[2] [e2e-llm-inference-service] ) [e2e-llm-inference-service] retries.sleep() [e2e-llm-inference-service] [e2e-llm-inference-service] # Keep track of the error for the retry warning. [e2e-llm-inference-service] err = e [e2e-llm-inference-service] [e2e-llm-inference-service] finally: [e2e-llm-inference-service] if not clean_exit: [e2e-llm-inference-service] # We hit some kind of exception, handled or otherwise. We need [e2e-llm-inference-service] # to throw the connection away unless explicitly told not to. [e2e-llm-inference-service] # Close the connection, set the variable to None, and make sure [e2e-llm-inference-service] # we put the None back in the pool to avoid leaking it. [e2e-llm-inference-service] if conn: [e2e-llm-inference-service] conn.close() [e2e-llm-inference-service] conn = None [e2e-llm-inference-service] release_this_conn = True [e2e-llm-inference-service] [e2e-llm-inference-service] if release_this_conn: [e2e-llm-inference-service] # Put the connection back to be reused. If the connection is [e2e-llm-inference-service] # expired then it will be None, which will get replaced with a [e2e-llm-inference-service] # fresh connection during _get_conn. [e2e-llm-inference-service] self._put_conn(conn) [e2e-llm-inference-service] [e2e-llm-inference-service] if not conn: [e2e-llm-inference-service] # Try again [e2e-llm-inference-service] log.warning( [e2e-llm-inference-service] "Retrying (%r) after connection broken by '%r': %s", retries, err, url [e2e-llm-inference-service] ) [e2e-llm-inference-service] > return self.urlopen( [e2e-llm-inference-service] method, [e2e-llm-inference-service] url, [e2e-llm-inference-service] body, [e2e-llm-inference-service] headers, [e2e-llm-inference-service] retries, [e2e-llm-inference-service] redirect, [e2e-llm-inference-service] assert_same_host, [e2e-llm-inference-service] timeout=timeout, [e2e-llm-inference-service] pool_timeout=pool_timeout, [e2e-llm-inference-service] release_conn=release_conn, [e2e-llm-inference-service] chunked=chunked, [e2e-llm-inference-service] body_pos=body_pos, [e2e-llm-inference-service] preload_content=preload_content, [e2e-llm-inference-service] decode_content=decode_content, [e2e-llm-inference-service] **response_kw, [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] ../../python/kserve/.venv/lib64/python3.11/site-packages/urllib3/connectionpool.py:871: [e2e-llm-inference-service] _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ [e2e-llm-inference-service] [e2e-llm-inference-service] self = [e2e-llm-inference-service] method = 'POST', url = '/kserve-ci-e2e-test/rolling-upgrade-test/v1/completions' [e2e-llm-inference-service] body = b'{"model": "facebook/opt-125m", "prompt": "KServe is a", "max_tokens": 20}' [e2e-llm-inference-service] headers = {'User-Agent': 'python-requests/2.32.3', 'Accept-Encoding': 'gzip, deflate', 'Accept': '*/*', 'Connection': 'keep-alive', 'Content-Type': 'application/json', 'Content-Length': '73'} [e2e-llm-inference-service] retries = Retry(total=6, connect=None, read=None, redirect=None, status=None) [e2e-llm-inference-service] redirect = False, assert_same_host = False [e2e-llm-inference-service] timeout = Timeout(connect=60, read=60, total=None), pool_timeout = None [e2e-llm-inference-service] release_conn = False, chunked = False, body_pos = None, preload_content = False [e2e-llm-inference-service] decode_content = False, response_kw = {} [e2e-llm-inference-service] parsed_url = Url(scheme=None, auth=None, host=None, port=None, path='/kserve-ci-e2e-test/rolling-upgrade-test/v1/completions', query=None, fragment=None) [e2e-llm-inference-service] destination_scheme = None, conn = None, release_this_conn = True [e2e-llm-inference-service] http_tunnel_required = False [e2e-llm-inference-service] err = ReadTimeoutError("HTTPConnectionPool(host='ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com', port=80): Read timed out. (read timeout=60)") [e2e-llm-inference-service] clean_exit = False [e2e-llm-inference-service] [e2e-llm-inference-service] def urlopen( # type: ignore[override] [e2e-llm-inference-service] self, [e2e-llm-inference-service] method: str, [e2e-llm-inference-service] url: str, [e2e-llm-inference-service] body: _TYPE_BODY | None = None, [e2e-llm-inference-service] headers: typing.Mapping[str, str] | None = None, [e2e-llm-inference-service] retries: Retry | bool | int | None = None, [e2e-llm-inference-service] redirect: bool = True, [e2e-llm-inference-service] assert_same_host: bool = True, [e2e-llm-inference-service] timeout: _TYPE_TIMEOUT = _DEFAULT_TIMEOUT, [e2e-llm-inference-service] pool_timeout: int | None = None, [e2e-llm-inference-service] release_conn: bool | None = None, [e2e-llm-inference-service] chunked: bool = False, [e2e-llm-inference-service] body_pos: _TYPE_BODY_POSITION | None = None, [e2e-llm-inference-service] preload_content: bool = True, [e2e-llm-inference-service] decode_content: bool = True, [e2e-llm-inference-service] **response_kw: typing.Any, [e2e-llm-inference-service] ) -> BaseHTTPResponse: [e2e-llm-inference-service] """ [e2e-llm-inference-service] Get a connection from the pool and perform an HTTP request. This is the [e2e-llm-inference-service] lowest level call for making a request, so you'll need to specify all [e2e-llm-inference-service] the raw details. [e2e-llm-inference-service] [e2e-llm-inference-service] .. note:: [e2e-llm-inference-service] [e2e-llm-inference-service] More commonly, it's appropriate to use a convenience method [e2e-llm-inference-service] such as :meth:`request`. [e2e-llm-inference-service] [e2e-llm-inference-service] .. note:: [e2e-llm-inference-service] [e2e-llm-inference-service] `release_conn` will only behave as expected if [e2e-llm-inference-service] `preload_content=False` because we want to make [e2e-llm-inference-service] `preload_content=False` the default behaviour someday soon without [e2e-llm-inference-service] breaking backwards compatibility. [e2e-llm-inference-service] [e2e-llm-inference-service] :param method: [e2e-llm-inference-service] HTTP request method (such as GET, POST, PUT, etc.) [e2e-llm-inference-service] [e2e-llm-inference-service] :param url: [e2e-llm-inference-service] The URL to perform the request on. [e2e-llm-inference-service] [e2e-llm-inference-service] :param body: [e2e-llm-inference-service] Data to send in the request body, either :class:`str`, :class:`bytes`, [e2e-llm-inference-service] an iterable of :class:`str`/:class:`bytes`, or a file-like object. [e2e-llm-inference-service] [e2e-llm-inference-service] :param headers: [e2e-llm-inference-service] Dictionary of custom headers to send, such as User-Agent, [e2e-llm-inference-service] If-None-Match, etc. If None, pool headers are used. If provided, [e2e-llm-inference-service] these headers completely replace any pool-specific headers. [e2e-llm-inference-service] [e2e-llm-inference-service] :param retries: [e2e-llm-inference-service] Configure the number of retries to allow before raising a [e2e-llm-inference-service] :class:`~urllib3.exceptions.MaxRetryError` exception. [e2e-llm-inference-service] [e2e-llm-inference-service] If ``None`` (default) will retry 3 times, see ``Retry.DEFAULT``. Pass a [e2e-llm-inference-service] :class:`~urllib3.util.retry.Retry` object for fine-grained control [e2e-llm-inference-service] over different types of retries. [e2e-llm-inference-service] Pass an integer number to retry connection errors that many times, [e2e-llm-inference-service] but no other types of errors. Pass zero to never retry. [e2e-llm-inference-service] [e2e-llm-inference-service] If ``False``, then retries are disabled and any exception is raised [e2e-llm-inference-service] immediately. Also, instead of raising a MaxRetryError on redirects, [e2e-llm-inference-service] the redirect response will be returned. [e2e-llm-inference-service] [e2e-llm-inference-service] :type retries: :class:`~urllib3.util.retry.Retry`, False, or an int. [e2e-llm-inference-service] [e2e-llm-inference-service] :param redirect: [e2e-llm-inference-service] If True, automatically handle redirects (status codes 301, 302, [e2e-llm-inference-service] 303, 307, 308). Each redirect counts as a retry. Disabling retries [e2e-llm-inference-service] will disable redirect, too. [e2e-llm-inference-service] [e2e-llm-inference-service] :param assert_same_host: [e2e-llm-inference-service] If ``True``, will make sure that the host of the pool requests is [e2e-llm-inference-service] consistent else will raise HostChangedError. When ``False``, you can [e2e-llm-inference-service] use the pool on an HTTP proxy and request foreign hosts. [e2e-llm-inference-service] [e2e-llm-inference-service] :param timeout: [e2e-llm-inference-service] If specified, overrides the default timeout for this one [e2e-llm-inference-service] request. It may be a float (in seconds) or an instance of [e2e-llm-inference-service] :class:`urllib3.util.Timeout`. [e2e-llm-inference-service] [e2e-llm-inference-service] :param pool_timeout: [e2e-llm-inference-service] If set and the pool is set to block=True, then this method will [e2e-llm-inference-service] block for ``pool_timeout`` seconds and raise EmptyPoolError if no [e2e-llm-inference-service] connection is available within the time period. [e2e-llm-inference-service] [e2e-llm-inference-service] :param bool preload_content: [e2e-llm-inference-service] If True, the response's body will be preloaded into memory. [e2e-llm-inference-service] [e2e-llm-inference-service] :param bool decode_content: [e2e-llm-inference-service] If True, will attempt to decode the body based on the [e2e-llm-inference-service] 'content-encoding' header. [e2e-llm-inference-service] [e2e-llm-inference-service] :param release_conn: [e2e-llm-inference-service] If False, then the urlopen call will not release the connection [e2e-llm-inference-service] back into the pool once a response is received (but will release if [e2e-llm-inference-service] you read the entire contents of the response such as when [e2e-llm-inference-service] `preload_content=True`). This is useful if you're not preloading [e2e-llm-inference-service] the response's content immediately. You will need to call [e2e-llm-inference-service] ``r.release_conn()`` on the response ``r`` to return the connection [e2e-llm-inference-service] back into the pool. If None, it takes the value of ``preload_content`` [e2e-llm-inference-service] which defaults to ``True``. [e2e-llm-inference-service] [e2e-llm-inference-service] :param bool chunked: [e2e-llm-inference-service] If True, urllib3 will send the body using chunked transfer [e2e-llm-inference-service] encoding. Otherwise, urllib3 will send the body using the standard [e2e-llm-inference-service] content-length form. Defaults to False. [e2e-llm-inference-service] [e2e-llm-inference-service] :param int body_pos: [e2e-llm-inference-service] Position to seek to in file-like body in the event of a retry or [e2e-llm-inference-service] redirect. Typically this won't need to be set because urllib3 will [e2e-llm-inference-service] auto-populate the value when needed. [e2e-llm-inference-service] """ [e2e-llm-inference-service] parsed_url = parse_url(url) [e2e-llm-inference-service] destination_scheme = parsed_url.scheme [e2e-llm-inference-service] [e2e-llm-inference-service] if headers is None: [e2e-llm-inference-service] headers = self.headers [e2e-llm-inference-service] [e2e-llm-inference-service] if not isinstance(retries, Retry): [e2e-llm-inference-service] retries = Retry.from_int(retries, redirect=redirect, default=self.retries) [e2e-llm-inference-service] [e2e-llm-inference-service] if release_conn is None: [e2e-llm-inference-service] release_conn = preload_content [e2e-llm-inference-service] [e2e-llm-inference-service] # Check host [e2e-llm-inference-service] if assert_same_host and not self.is_same_host(url): [e2e-llm-inference-service] raise HostChangedError(self, url, retries) [e2e-llm-inference-service] [e2e-llm-inference-service] # Ensure that the URL we're connecting to is properly encoded [e2e-llm-inference-service] if url.startswith("/"): [e2e-llm-inference-service] url = to_str(_encode_target(url)) [e2e-llm-inference-service] else: [e2e-llm-inference-service] url = to_str(parsed_url.url) [e2e-llm-inference-service] [e2e-llm-inference-service] conn = None [e2e-llm-inference-service] [e2e-llm-inference-service] # Track whether `conn` needs to be released before [e2e-llm-inference-service] # returning/raising/recursing. Update this variable if necessary, and [e2e-llm-inference-service] # leave `release_conn` constant throughout the function. That way, if [e2e-llm-inference-service] # the function recurses, the original value of `release_conn` will be [e2e-llm-inference-service] # passed down into the recursive call, and its value will be respected. [e2e-llm-inference-service] # [e2e-llm-inference-service] # See issue #651 [1] for details. [e2e-llm-inference-service] # [e2e-llm-inference-service] # [1] [e2e-llm-inference-service] release_this_conn = release_conn [e2e-llm-inference-service] [e2e-llm-inference-service] http_tunnel_required = connection_requires_http_tunnel( [e2e-llm-inference-service] self.proxy, self.proxy_config, destination_scheme [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] # Merge the proxy headers. Only done when not using HTTP CONNECT. We [e2e-llm-inference-service] # have to copy the headers dict so we can safely change it without those [e2e-llm-inference-service] # changes being reflected in anyone else's copy. [e2e-llm-inference-service] if not http_tunnel_required: [e2e-llm-inference-service] headers = headers.copy() # type: ignore[attr-defined] [e2e-llm-inference-service] headers.update(self.proxy_headers) # type: ignore[union-attr] [e2e-llm-inference-service] [e2e-llm-inference-service] # Must keep the exception bound to a separate variable or else Python 3 [e2e-llm-inference-service] # complains about UnboundLocalError. [e2e-llm-inference-service] err = None [e2e-llm-inference-service] [e2e-llm-inference-service] # Keep track of whether we cleanly exited the except block. This [e2e-llm-inference-service] # ensures we do proper cleanup in finally. [e2e-llm-inference-service] clean_exit = False [e2e-llm-inference-service] [e2e-llm-inference-service] # Rewind body position, if needed. Record current position [e2e-llm-inference-service] # for future rewinds in the event of a redirect/retry. [e2e-llm-inference-service] body_pos = set_file_position(body, body_pos) [e2e-llm-inference-service] [e2e-llm-inference-service] try: [e2e-llm-inference-service] # Request a connection from the queue. [e2e-llm-inference-service] timeout_obj = self._get_timeout(timeout) [e2e-llm-inference-service] conn = self._get_conn(timeout=pool_timeout) [e2e-llm-inference-service] [e2e-llm-inference-service] conn.timeout = timeout_obj.connect_timeout # type: ignore[assignment] [e2e-llm-inference-service] [e2e-llm-inference-service] # Is this a closed/new connection that requires CONNECT tunnelling? [e2e-llm-inference-service] if self.proxy is not None and http_tunnel_required and conn.is_closed: [e2e-llm-inference-service] try: [e2e-llm-inference-service] self._prepare_proxy(conn) [e2e-llm-inference-service] except (BaseSSLError, OSError, SocketTimeout) as e: [e2e-llm-inference-service] self._raise_timeout( [e2e-llm-inference-service] err=e, url=self.proxy.url, timeout_value=conn.timeout [e2e-llm-inference-service] ) [e2e-llm-inference-service] raise [e2e-llm-inference-service] [e2e-llm-inference-service] # If we're going to release the connection in ``finally:``, then [e2e-llm-inference-service] # the response doesn't need to know about the connection. Otherwise [e2e-llm-inference-service] # it will also try to release it and we'll have a double-release [e2e-llm-inference-service] # mess. [e2e-llm-inference-service] response_conn = conn if not release_conn else None [e2e-llm-inference-service] [e2e-llm-inference-service] # Make the request on the HTTPConnection object [e2e-llm-inference-service] response = self._make_request( [e2e-llm-inference-service] conn, [e2e-llm-inference-service] method, [e2e-llm-inference-service] url, [e2e-llm-inference-service] timeout=timeout_obj, [e2e-llm-inference-service] body=body, [e2e-llm-inference-service] headers=headers, [e2e-llm-inference-service] chunked=chunked, [e2e-llm-inference-service] retries=retries, [e2e-llm-inference-service] response_conn=response_conn, [e2e-llm-inference-service] preload_content=preload_content, [e2e-llm-inference-service] decode_content=decode_content, [e2e-llm-inference-service] **response_kw, [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] # Everything went great! [e2e-llm-inference-service] clean_exit = True [e2e-llm-inference-service] [e2e-llm-inference-service] except EmptyPoolError: [e2e-llm-inference-service] # Didn't get a connection from the pool, no need to clean up [e2e-llm-inference-service] clean_exit = True [e2e-llm-inference-service] release_this_conn = False [e2e-llm-inference-service] raise [e2e-llm-inference-service] [e2e-llm-inference-service] except ( [e2e-llm-inference-service] TimeoutError, [e2e-llm-inference-service] HTTPException, [e2e-llm-inference-service] OSError, [e2e-llm-inference-service] ProtocolError, [e2e-llm-inference-service] BaseSSLError, [e2e-llm-inference-service] SSLError, [e2e-llm-inference-service] CertificateError, [e2e-llm-inference-service] ProxyError, [e2e-llm-inference-service] ) as e: [e2e-llm-inference-service] # Discard the connection for these exceptions. It will be [e2e-llm-inference-service] # replaced during the next _get_conn() call. [e2e-llm-inference-service] clean_exit = False [e2e-llm-inference-service] new_e: Exception = e [e2e-llm-inference-service] if isinstance(e, (BaseSSLError, CertificateError)): [e2e-llm-inference-service] new_e = SSLError(e) [e2e-llm-inference-service] if isinstance( [e2e-llm-inference-service] new_e, [e2e-llm-inference-service] ( [e2e-llm-inference-service] OSError, [e2e-llm-inference-service] NewConnectionError, [e2e-llm-inference-service] TimeoutError, [e2e-llm-inference-service] SSLError, [e2e-llm-inference-service] HTTPException, [e2e-llm-inference-service] ), [e2e-llm-inference-service] ) and (conn and conn.proxy and not conn.has_connected_to_proxy): [e2e-llm-inference-service] new_e = _wrap_proxy_error(new_e, conn.proxy.scheme) [e2e-llm-inference-service] elif isinstance(new_e, (OSError, HTTPException)): [e2e-llm-inference-service] new_e = ProtocolError("Connection aborted.", new_e) [e2e-llm-inference-service] [e2e-llm-inference-service] retries = retries.increment( [e2e-llm-inference-service] method, url, error=new_e, _pool=self, _stacktrace=sys.exc_info()[2] [e2e-llm-inference-service] ) [e2e-llm-inference-service] retries.sleep() [e2e-llm-inference-service] [e2e-llm-inference-service] # Keep track of the error for the retry warning. [e2e-llm-inference-service] err = e [e2e-llm-inference-service] [e2e-llm-inference-service] finally: [e2e-llm-inference-service] if not clean_exit: [e2e-llm-inference-service] # We hit some kind of exception, handled or otherwise. We need [e2e-llm-inference-service] # to throw the connection away unless explicitly told not to. [e2e-llm-inference-service] # Close the connection, set the variable to None, and make sure [e2e-llm-inference-service] # we put the None back in the pool to avoid leaking it. [e2e-llm-inference-service] if conn: [e2e-llm-inference-service] conn.close() [e2e-llm-inference-service] conn = None [e2e-llm-inference-service] release_this_conn = True [e2e-llm-inference-service] [e2e-llm-inference-service] if release_this_conn: [e2e-llm-inference-service] # Put the connection back to be reused. If the connection is [e2e-llm-inference-service] # expired then it will be None, which will get replaced with a [e2e-llm-inference-service] # fresh connection during _get_conn. [e2e-llm-inference-service] self._put_conn(conn) [e2e-llm-inference-service] [e2e-llm-inference-service] if not conn: [e2e-llm-inference-service] # Try again [e2e-llm-inference-service] log.warning( [e2e-llm-inference-service] "Retrying (%r) after connection broken by '%r': %s", retries, err, url [e2e-llm-inference-service] ) [e2e-llm-inference-service] > return self.urlopen( [e2e-llm-inference-service] method, [e2e-llm-inference-service] url, [e2e-llm-inference-service] body, [e2e-llm-inference-service] headers, [e2e-llm-inference-service] retries, [e2e-llm-inference-service] redirect, [e2e-llm-inference-service] assert_same_host, [e2e-llm-inference-service] timeout=timeout, [e2e-llm-inference-service] pool_timeout=pool_timeout, [e2e-llm-inference-service] release_conn=release_conn, [e2e-llm-inference-service] chunked=chunked, [e2e-llm-inference-service] body_pos=body_pos, [e2e-llm-inference-service] preload_content=preload_content, [e2e-llm-inference-service] decode_content=decode_content, [e2e-llm-inference-service] **response_kw, [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] ../../python/kserve/.venv/lib64/python3.11/site-packages/urllib3/connectionpool.py:871: [e2e-llm-inference-service] _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ [e2e-llm-inference-service] [e2e-llm-inference-service] self = [e2e-llm-inference-service] method = 'POST', url = '/kserve-ci-e2e-test/rolling-upgrade-test/v1/completions' [e2e-llm-inference-service] body = b'{"model": "facebook/opt-125m", "prompt": "KServe is a", "max_tokens": 20}' [e2e-llm-inference-service] headers = {'User-Agent': 'python-requests/2.32.3', 'Accept-Encoding': 'gzip, deflate', 'Accept': '*/*', 'Connection': 'keep-alive', 'Content-Type': 'application/json', 'Content-Length': '73'} [e2e-llm-inference-service] retries = Retry(total=5, connect=None, read=None, redirect=None, status=None) [e2e-llm-inference-service] redirect = False, assert_same_host = False [e2e-llm-inference-service] timeout = Timeout(connect=60, read=60, total=None), pool_timeout = None [e2e-llm-inference-service] release_conn = False, chunked = False, body_pos = None, preload_content = False [e2e-llm-inference-service] decode_content = False, response_kw = {} [e2e-llm-inference-service] parsed_url = Url(scheme=None, auth=None, host=None, port=None, path='/kserve-ci-e2e-test/rolling-upgrade-test/v1/completions', query=None, fragment=None) [e2e-llm-inference-service] destination_scheme = None, conn = None, release_this_conn = True [e2e-llm-inference-service] http_tunnel_required = False [e2e-llm-inference-service] err = ReadTimeoutError("HTTPConnectionPool(host='ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com', port=80): Read timed out. (read timeout=60)") [e2e-llm-inference-service] clean_exit = False [e2e-llm-inference-service] [e2e-llm-inference-service] def urlopen( # type: ignore[override] [e2e-llm-inference-service] self, [e2e-llm-inference-service] method: str, [e2e-llm-inference-service] url: str, [e2e-llm-inference-service] body: _TYPE_BODY | None = None, [e2e-llm-inference-service] headers: typing.Mapping[str, str] | None = None, [e2e-llm-inference-service] retries: Retry | bool | int | None = None, [e2e-llm-inference-service] redirect: bool = True, [e2e-llm-inference-service] assert_same_host: bool = True, [e2e-llm-inference-service] timeout: _TYPE_TIMEOUT = _DEFAULT_TIMEOUT, [e2e-llm-inference-service] pool_timeout: int | None = None, [e2e-llm-inference-service] release_conn: bool | None = None, [e2e-llm-inference-service] chunked: bool = False, [e2e-llm-inference-service] body_pos: _TYPE_BODY_POSITION | None = None, [e2e-llm-inference-service] preload_content: bool = True, [e2e-llm-inference-service] decode_content: bool = True, [e2e-llm-inference-service] **response_kw: typing.Any, [e2e-llm-inference-service] ) -> BaseHTTPResponse: [e2e-llm-inference-service] """ [e2e-llm-inference-service] Get a connection from the pool and perform an HTTP request. This is the [e2e-llm-inference-service] lowest level call for making a request, so you'll need to specify all [e2e-llm-inference-service] the raw details. [e2e-llm-inference-service] [e2e-llm-inference-service] .. note:: [e2e-llm-inference-service] [e2e-llm-inference-service] More commonly, it's appropriate to use a convenience method [e2e-llm-inference-service] such as :meth:`request`. [e2e-llm-inference-service] [e2e-llm-inference-service] .. note:: [e2e-llm-inference-service] [e2e-llm-inference-service] `release_conn` will only behave as expected if [e2e-llm-inference-service] `preload_content=False` because we want to make [e2e-llm-inference-service] `preload_content=False` the default behaviour someday soon without [e2e-llm-inference-service] breaking backwards compatibility. [e2e-llm-inference-service] [e2e-llm-inference-service] :param method: [e2e-llm-inference-service] HTTP request method (such as GET, POST, PUT, etc.) [e2e-llm-inference-service] [e2e-llm-inference-service] :param url: [e2e-llm-inference-service] The URL to perform the request on. [e2e-llm-inference-service] [e2e-llm-inference-service] :param body: [e2e-llm-inference-service] Data to send in the request body, either :class:`str`, :class:`bytes`, [e2e-llm-inference-service] an iterable of :class:`str`/:class:`bytes`, or a file-like object. [e2e-llm-inference-service] [e2e-llm-inference-service] :param headers: [e2e-llm-inference-service] Dictionary of custom headers to send, such as User-Agent, [e2e-llm-inference-service] If-None-Match, etc. If None, pool headers are used. If provided, [e2e-llm-inference-service] these headers completely replace any pool-specific headers. [e2e-llm-inference-service] [e2e-llm-inference-service] :param retries: [e2e-llm-inference-service] Configure the number of retries to allow before raising a [e2e-llm-inference-service] :class:`~urllib3.exceptions.MaxRetryError` exception. [e2e-llm-inference-service] [e2e-llm-inference-service] If ``None`` (default) will retry 3 times, see ``Retry.DEFAULT``. Pass a [e2e-llm-inference-service] :class:`~urllib3.util.retry.Retry` object for fine-grained control [e2e-llm-inference-service] over different types of retries. [e2e-llm-inference-service] Pass an integer number to retry connection errors that many times, [e2e-llm-inference-service] but no other types of errors. Pass zero to never retry. [e2e-llm-inference-service] [e2e-llm-inference-service] If ``False``, then retries are disabled and any exception is raised [e2e-llm-inference-service] immediately. Also, instead of raising a MaxRetryError on redirects, [e2e-llm-inference-service] the redirect response will be returned. [e2e-llm-inference-service] [e2e-llm-inference-service] :type retries: :class:`~urllib3.util.retry.Retry`, False, or an int. [e2e-llm-inference-service] [e2e-llm-inference-service] :param redirect: [e2e-llm-inference-service] If True, automatically handle redirects (status codes 301, 302, [e2e-llm-inference-service] 303, 307, 308). Each redirect counts as a retry. Disabling retries [e2e-llm-inference-service] will disable redirect, too. [e2e-llm-inference-service] [e2e-llm-inference-service] :param assert_same_host: [e2e-llm-inference-service] If ``True``, will make sure that the host of the pool requests is [e2e-llm-inference-service] consistent else will raise HostChangedError. When ``False``, you can [e2e-llm-inference-service] use the pool on an HTTP proxy and request foreign hosts. [e2e-llm-inference-service] [e2e-llm-inference-service] :param timeout: [e2e-llm-inference-service] If specified, overrides the default timeout for this one [e2e-llm-inference-service] request. It may be a float (in seconds) or an instance of [e2e-llm-inference-service] :class:`urllib3.util.Timeout`. [e2e-llm-inference-service] [e2e-llm-inference-service] :param pool_timeout: [e2e-llm-inference-service] If set and the pool is set to block=True, then this method will [e2e-llm-inference-service] block for ``pool_timeout`` seconds and raise EmptyPoolError if no [e2e-llm-inference-service] connection is available within the time period. [e2e-llm-inference-service] [e2e-llm-inference-service] :param bool preload_content: [e2e-llm-inference-service] If True, the response's body will be preloaded into memory. [e2e-llm-inference-service] [e2e-llm-inference-service] :param bool decode_content: [e2e-llm-inference-service] If True, will attempt to decode the body based on the [e2e-llm-inference-service] 'content-encoding' header. [e2e-llm-inference-service] [e2e-llm-inference-service] :param release_conn: [e2e-llm-inference-service] If False, then the urlopen call will not release the connection [e2e-llm-inference-service] back into the pool once a response is received (but will release if [e2e-llm-inference-service] you read the entire contents of the response such as when [e2e-llm-inference-service] `preload_content=True`). This is useful if you're not preloading [e2e-llm-inference-service] the response's content immediately. You will need to call [e2e-llm-inference-service] ``r.release_conn()`` on the response ``r`` to return the connection [e2e-llm-inference-service] back into the pool. If None, it takes the value of ``preload_content`` [e2e-llm-inference-service] which defaults to ``True``. [e2e-llm-inference-service] [e2e-llm-inference-service] :param bool chunked: [e2e-llm-inference-service] If True, urllib3 will send the body using chunked transfer [e2e-llm-inference-service] encoding. Otherwise, urllib3 will send the body using the standard [e2e-llm-inference-service] content-length form. Defaults to False. [e2e-llm-inference-service] [e2e-llm-inference-service] :param int body_pos: [e2e-llm-inference-service] Position to seek to in file-like body in the event of a retry or [e2e-llm-inference-service] redirect. Typically this won't need to be set because urllib3 will [e2e-llm-inference-service] auto-populate the value when needed. [e2e-llm-inference-service] """ [e2e-llm-inference-service] parsed_url = parse_url(url) [e2e-llm-inference-service] destination_scheme = parsed_url.scheme [e2e-llm-inference-service] [e2e-llm-inference-service] if headers is None: [e2e-llm-inference-service] headers = self.headers [e2e-llm-inference-service] [e2e-llm-inference-service] if not isinstance(retries, Retry): [e2e-llm-inference-service] retries = Retry.from_int(retries, redirect=redirect, default=self.retries) [e2e-llm-inference-service] [e2e-llm-inference-service] if release_conn is None: [e2e-llm-inference-service] release_conn = preload_content [e2e-llm-inference-service] [e2e-llm-inference-service] # Check host [e2e-llm-inference-service] if assert_same_host and not self.is_same_host(url): [e2e-llm-inference-service] raise HostChangedError(self, url, retries) [e2e-llm-inference-service] [e2e-llm-inference-service] # Ensure that the URL we're connecting to is properly encoded [e2e-llm-inference-service] if url.startswith("/"): [e2e-llm-inference-service] url = to_str(_encode_target(url)) [e2e-llm-inference-service] else: [e2e-llm-inference-service] url = to_str(parsed_url.url) [e2e-llm-inference-service] [e2e-llm-inference-service] conn = None [e2e-llm-inference-service] [e2e-llm-inference-service] # Track whether `conn` needs to be released before [e2e-llm-inference-service] # returning/raising/recursing. Update this variable if necessary, and [e2e-llm-inference-service] # leave `release_conn` constant throughout the function. That way, if [e2e-llm-inference-service] # the function recurses, the original value of `release_conn` will be [e2e-llm-inference-service] # passed down into the recursive call, and its value will be respected. [e2e-llm-inference-service] # [e2e-llm-inference-service] # See issue #651 [1] for details. [e2e-llm-inference-service] # [e2e-llm-inference-service] # [1] [e2e-llm-inference-service] release_this_conn = release_conn [e2e-llm-inference-service] [e2e-llm-inference-service] http_tunnel_required = connection_requires_http_tunnel( [e2e-llm-inference-service] self.proxy, self.proxy_config, destination_scheme [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] # Merge the proxy headers. Only done when not using HTTP CONNECT. We [e2e-llm-inference-service] # have to copy the headers dict so we can safely change it without those [e2e-llm-inference-service] # changes being reflected in anyone else's copy. [e2e-llm-inference-service] if not http_tunnel_required: [e2e-llm-inference-service] headers = headers.copy() # type: ignore[attr-defined] [e2e-llm-inference-service] headers.update(self.proxy_headers) # type: ignore[union-attr] [e2e-llm-inference-service] [e2e-llm-inference-service] # Must keep the exception bound to a separate variable or else Python 3 [e2e-llm-inference-service] # complains about UnboundLocalError. [e2e-llm-inference-service] err = None [e2e-llm-inference-service] [e2e-llm-inference-service] # Keep track of whether we cleanly exited the except block. This [e2e-llm-inference-service] # ensures we do proper cleanup in finally. [e2e-llm-inference-service] clean_exit = False [e2e-llm-inference-service] [e2e-llm-inference-service] # Rewind body position, if needed. Record current position [e2e-llm-inference-service] # for future rewinds in the event of a redirect/retry. [e2e-llm-inference-service] body_pos = set_file_position(body, body_pos) [e2e-llm-inference-service] [e2e-llm-inference-service] try: [e2e-llm-inference-service] # Request a connection from the queue. [e2e-llm-inference-service] timeout_obj = self._get_timeout(timeout) [e2e-llm-inference-service] conn = self._get_conn(timeout=pool_timeout) [e2e-llm-inference-service] [e2e-llm-inference-service] conn.timeout = timeout_obj.connect_timeout # type: ignore[assignment] [e2e-llm-inference-service] [e2e-llm-inference-service] # Is this a closed/new connection that requires CONNECT tunnelling? [e2e-llm-inference-service] if self.proxy is not None and http_tunnel_required and conn.is_closed: [e2e-llm-inference-service] try: [e2e-llm-inference-service] self._prepare_proxy(conn) [e2e-llm-inference-service] except (BaseSSLError, OSError, SocketTimeout) as e: [e2e-llm-inference-service] self._raise_timeout( [e2e-llm-inference-service] err=e, url=self.proxy.url, timeout_value=conn.timeout [e2e-llm-inference-service] ) [e2e-llm-inference-service] raise [e2e-llm-inference-service] [e2e-llm-inference-service] # If we're going to release the connection in ``finally:``, then [e2e-llm-inference-service] # the response doesn't need to know about the connection. Otherwise [e2e-llm-inference-service] # it will also try to release it and we'll have a double-release [e2e-llm-inference-service] # mess. [e2e-llm-inference-service] response_conn = conn if not release_conn else None [e2e-llm-inference-service] [e2e-llm-inference-service] # Make the request on the HTTPConnection object [e2e-llm-inference-service] response = self._make_request( [e2e-llm-inference-service] conn, [e2e-llm-inference-service] method, [e2e-llm-inference-service] url, [e2e-llm-inference-service] timeout=timeout_obj, [e2e-llm-inference-service] body=body, [e2e-llm-inference-service] headers=headers, [e2e-llm-inference-service] chunked=chunked, [e2e-llm-inference-service] retries=retries, [e2e-llm-inference-service] response_conn=response_conn, [e2e-llm-inference-service] preload_content=preload_content, [e2e-llm-inference-service] decode_content=decode_content, [e2e-llm-inference-service] **response_kw, [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] # Everything went great! [e2e-llm-inference-service] clean_exit = True [e2e-llm-inference-service] [e2e-llm-inference-service] except EmptyPoolError: [e2e-llm-inference-service] # Didn't get a connection from the pool, no need to clean up [e2e-llm-inference-service] clean_exit = True [e2e-llm-inference-service] release_this_conn = False [e2e-llm-inference-service] raise [e2e-llm-inference-service] [e2e-llm-inference-service] except ( [e2e-llm-inference-service] TimeoutError, [e2e-llm-inference-service] HTTPException, [e2e-llm-inference-service] OSError, [e2e-llm-inference-service] ProtocolError, [e2e-llm-inference-service] BaseSSLError, [e2e-llm-inference-service] SSLError, [e2e-llm-inference-service] CertificateError, [e2e-llm-inference-service] ProxyError, [e2e-llm-inference-service] ) as e: [e2e-llm-inference-service] # Discard the connection for these exceptions. It will be [e2e-llm-inference-service] # replaced during the next _get_conn() call. [e2e-llm-inference-service] clean_exit = False [e2e-llm-inference-service] new_e: Exception = e [e2e-llm-inference-service] if isinstance(e, (BaseSSLError, CertificateError)): [e2e-llm-inference-service] new_e = SSLError(e) [e2e-llm-inference-service] if isinstance( [e2e-llm-inference-service] new_e, [e2e-llm-inference-service] ( [e2e-llm-inference-service] OSError, [e2e-llm-inference-service] NewConnectionError, [e2e-llm-inference-service] TimeoutError, [e2e-llm-inference-service] SSLError, [e2e-llm-inference-service] HTTPException, [e2e-llm-inference-service] ), [e2e-llm-inference-service] ) and (conn and conn.proxy and not conn.has_connected_to_proxy): [e2e-llm-inference-service] new_e = _wrap_proxy_error(new_e, conn.proxy.scheme) [e2e-llm-inference-service] elif isinstance(new_e, (OSError, HTTPException)): [e2e-llm-inference-service] new_e = ProtocolError("Connection aborted.", new_e) [e2e-llm-inference-service] [e2e-llm-inference-service] retries = retries.increment( [e2e-llm-inference-service] method, url, error=new_e, _pool=self, _stacktrace=sys.exc_info()[2] [e2e-llm-inference-service] ) [e2e-llm-inference-service] retries.sleep() [e2e-llm-inference-service] [e2e-llm-inference-service] # Keep track of the error for the retry warning. [e2e-llm-inference-service] err = e [e2e-llm-inference-service] [e2e-llm-inference-service] finally: [e2e-llm-inference-service] if not clean_exit: [e2e-llm-inference-service] # We hit some kind of exception, handled or otherwise. We need [e2e-llm-inference-service] # to throw the connection away unless explicitly told not to. [e2e-llm-inference-service] # Close the connection, set the variable to None, and make sure [e2e-llm-inference-service] # we put the None back in the pool to avoid leaking it. [e2e-llm-inference-service] if conn: [e2e-llm-inference-service] conn.close() [e2e-llm-inference-service] conn = None [e2e-llm-inference-service] release_this_conn = True [e2e-llm-inference-service] [e2e-llm-inference-service] if release_this_conn: [e2e-llm-inference-service] # Put the connection back to be reused. If the connection is [e2e-llm-inference-service] # expired then it will be None, which will get replaced with a [e2e-llm-inference-service] # fresh connection during _get_conn. [e2e-llm-inference-service] self._put_conn(conn) [e2e-llm-inference-service] [e2e-llm-inference-service] if not conn: [e2e-llm-inference-service] # Try again [e2e-llm-inference-service] log.warning( [e2e-llm-inference-service] "Retrying (%r) after connection broken by '%r': %s", retries, err, url [e2e-llm-inference-service] ) [e2e-llm-inference-service] > return self.urlopen( [e2e-llm-inference-service] method, [e2e-llm-inference-service] url, [e2e-llm-inference-service] body, [e2e-llm-inference-service] headers, [e2e-llm-inference-service] retries, [e2e-llm-inference-service] redirect, [e2e-llm-inference-service] assert_same_host, [e2e-llm-inference-service] timeout=timeout, [e2e-llm-inference-service] pool_timeout=pool_timeout, [e2e-llm-inference-service] release_conn=release_conn, [e2e-llm-inference-service] chunked=chunked, [e2e-llm-inference-service] body_pos=body_pos, [e2e-llm-inference-service] preload_content=preload_content, [e2e-llm-inference-service] decode_content=decode_content, [e2e-llm-inference-service] **response_kw, [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] ../../python/kserve/.venv/lib64/python3.11/site-packages/urllib3/connectionpool.py:871: [e2e-llm-inference-service] _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ [e2e-llm-inference-service] [e2e-llm-inference-service] self = [e2e-llm-inference-service] method = 'POST', url = '/kserve-ci-e2e-test/rolling-upgrade-test/v1/completions' [e2e-llm-inference-service] body = b'{"model": "facebook/opt-125m", "prompt": "KServe is a", "max_tokens": 20}' [e2e-llm-inference-service] headers = {'User-Agent': 'python-requests/2.32.3', 'Accept-Encoding': 'gzip, deflate', 'Accept': '*/*', 'Connection': 'keep-alive', 'Content-Type': 'application/json', 'Content-Length': '73'} [e2e-llm-inference-service] retries = Retry(total=4, connect=None, read=None, redirect=None, status=None) [e2e-llm-inference-service] redirect = False, assert_same_host = False [e2e-llm-inference-service] timeout = Timeout(connect=60, read=60, total=None), pool_timeout = None [e2e-llm-inference-service] release_conn = False, chunked = False, body_pos = None, preload_content = False [e2e-llm-inference-service] decode_content = False, response_kw = {} [e2e-llm-inference-service] parsed_url = Url(scheme=None, auth=None, host=None, port=None, path='/kserve-ci-e2e-test/rolling-upgrade-test/v1/completions', query=None, fragment=None) [e2e-llm-inference-service] destination_scheme = None, conn = None, release_this_conn = True [e2e-llm-inference-service] http_tunnel_required = False [e2e-llm-inference-service] err = ReadTimeoutError("HTTPConnectionPool(host='ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com', port=80): Read timed out. (read timeout=60)") [e2e-llm-inference-service] clean_exit = False [e2e-llm-inference-service] [e2e-llm-inference-service] def urlopen( # type: ignore[override] [e2e-llm-inference-service] self, [e2e-llm-inference-service] method: str, [e2e-llm-inference-service] url: str, [e2e-llm-inference-service] body: _TYPE_BODY | None = None, [e2e-llm-inference-service] headers: typing.Mapping[str, str] | None = None, [e2e-llm-inference-service] retries: Retry | bool | int | None = None, [e2e-llm-inference-service] redirect: bool = True, [e2e-llm-inference-service] assert_same_host: bool = True, [e2e-llm-inference-service] timeout: _TYPE_TIMEOUT = _DEFAULT_TIMEOUT, [e2e-llm-inference-service] pool_timeout: int | None = None, [e2e-llm-inference-service] release_conn: bool | None = None, [e2e-llm-inference-service] chunked: bool = False, [e2e-llm-inference-service] body_pos: _TYPE_BODY_POSITION | None = None, [e2e-llm-inference-service] preload_content: bool = True, [e2e-llm-inference-service] decode_content: bool = True, [e2e-llm-inference-service] **response_kw: typing.Any, [e2e-llm-inference-service] ) -> BaseHTTPResponse: [e2e-llm-inference-service] """ [e2e-llm-inference-service] Get a connection from the pool and perform an HTTP request. This is the [e2e-llm-inference-service] lowest level call for making a request, so you'll need to specify all [e2e-llm-inference-service] the raw details. [e2e-llm-inference-service] [e2e-llm-inference-service] .. note:: [e2e-llm-inference-service] [e2e-llm-inference-service] More commonly, it's appropriate to use a convenience method [e2e-llm-inference-service] such as :meth:`request`. [e2e-llm-inference-service] [e2e-llm-inference-service] .. note:: [e2e-llm-inference-service] [e2e-llm-inference-service] `release_conn` will only behave as expected if [e2e-llm-inference-service] `preload_content=False` because we want to make [e2e-llm-inference-service] `preload_content=False` the default behaviour someday soon without [e2e-llm-inference-service] breaking backwards compatibility. [e2e-llm-inference-service] [e2e-llm-inference-service] :param method: [e2e-llm-inference-service] HTTP request method (such as GET, POST, PUT, etc.) [e2e-llm-inference-service] [e2e-llm-inference-service] :param url: [e2e-llm-inference-service] The URL to perform the request on. [e2e-llm-inference-service] [e2e-llm-inference-service] :param body: [e2e-llm-inference-service] Data to send in the request body, either :class:`str`, :class:`bytes`, [e2e-llm-inference-service] an iterable of :class:`str`/:class:`bytes`, or a file-like object. [e2e-llm-inference-service] [e2e-llm-inference-service] :param headers: [e2e-llm-inference-service] Dictionary of custom headers to send, such as User-Agent, [e2e-llm-inference-service] If-None-Match, etc. If None, pool headers are used. If provided, [e2e-llm-inference-service] these headers completely replace any pool-specific headers. [e2e-llm-inference-service] [e2e-llm-inference-service] :param retries: [e2e-llm-inference-service] Configure the number of retries to allow before raising a [e2e-llm-inference-service] :class:`~urllib3.exceptions.MaxRetryError` exception. [e2e-llm-inference-service] [e2e-llm-inference-service] If ``None`` (default) will retry 3 times, see ``Retry.DEFAULT``. Pass a [e2e-llm-inference-service] :class:`~urllib3.util.retry.Retry` object for fine-grained control [e2e-llm-inference-service] over different types of retries. [e2e-llm-inference-service] Pass an integer number to retry connection errors that many times, [e2e-llm-inference-service] but no other types of errors. Pass zero to never retry. [e2e-llm-inference-service] [e2e-llm-inference-service] If ``False``, then retries are disabled and any exception is raised [e2e-llm-inference-service] immediately. Also, instead of raising a MaxRetryError on redirects, [e2e-llm-inference-service] the redirect response will be returned. [e2e-llm-inference-service] [e2e-llm-inference-service] :type retries: :class:`~urllib3.util.retry.Retry`, False, or an int. [e2e-llm-inference-service] [e2e-llm-inference-service] :param redirect: [e2e-llm-inference-service] If True, automatically handle redirects (status codes 301, 302, [e2e-llm-inference-service] 303, 307, 308). Each redirect counts as a retry. Disabling retries [e2e-llm-inference-service] will disable redirect, too. [e2e-llm-inference-service] [e2e-llm-inference-service] :param assert_same_host: [e2e-llm-inference-service] If ``True``, will make sure that the host of the pool requests is [e2e-llm-inference-service] consistent else will raise HostChangedError. When ``False``, you can [e2e-llm-inference-service] use the pool on an HTTP proxy and request foreign hosts. [e2e-llm-inference-service] [e2e-llm-inference-service] :param timeout: [e2e-llm-inference-service] If specified, overrides the default timeout for this one [e2e-llm-inference-service] request. It may be a float (in seconds) or an instance of [e2e-llm-inference-service] :class:`urllib3.util.Timeout`. [e2e-llm-inference-service] [e2e-llm-inference-service] :param pool_timeout: [e2e-llm-inference-service] If set and the pool is set to block=True, then this method will [e2e-llm-inference-service] block for ``pool_timeout`` seconds and raise EmptyPoolError if no [e2e-llm-inference-service] connection is available within the time period. [e2e-llm-inference-service] [e2e-llm-inference-service] :param bool preload_content: [e2e-llm-inference-service] If True, the response's body will be preloaded into memory. [e2e-llm-inference-service] [e2e-llm-inference-service] :param bool decode_content: [e2e-llm-inference-service] If True, will attempt to decode the body based on the [e2e-llm-inference-service] 'content-encoding' header. [e2e-llm-inference-service] [e2e-llm-inference-service] :param release_conn: [e2e-llm-inference-service] If False, then the urlopen call will not release the connection [e2e-llm-inference-service] back into the pool once a response is received (but will release if [e2e-llm-inference-service] you read the entire contents of the response such as when [e2e-llm-inference-service] `preload_content=True`). This is useful if you're not preloading [e2e-llm-inference-service] the response's content immediately. You will need to call [e2e-llm-inference-service] ``r.release_conn()`` on the response ``r`` to return the connection [e2e-llm-inference-service] back into the pool. If None, it takes the value of ``preload_content`` [e2e-llm-inference-service] which defaults to ``True``. [e2e-llm-inference-service] [e2e-llm-inference-service] :param bool chunked: [e2e-llm-inference-service] If True, urllib3 will send the body using chunked transfer [e2e-llm-inference-service] encoding. Otherwise, urllib3 will send the body using the standard [e2e-llm-inference-service] content-length form. Defaults to False. [e2e-llm-inference-service] [e2e-llm-inference-service] :param int body_pos: [e2e-llm-inference-service] Position to seek to in file-like body in the event of a retry or [e2e-llm-inference-service] redirect. Typically this won't need to be set because urllib3 will [e2e-llm-inference-service] auto-populate the value when needed. [e2e-llm-inference-service] """ [e2e-llm-inference-service] parsed_url = parse_url(url) [e2e-llm-inference-service] destination_scheme = parsed_url.scheme [e2e-llm-inference-service] [e2e-llm-inference-service] if headers is None: [e2e-llm-inference-service] headers = self.headers [e2e-llm-inference-service] [e2e-llm-inference-service] if not isinstance(retries, Retry): [e2e-llm-inference-service] retries = Retry.from_int(retries, redirect=redirect, default=self.retries) [e2e-llm-inference-service] [e2e-llm-inference-service] if release_conn is None: [e2e-llm-inference-service] release_conn = preload_content [e2e-llm-inference-service] [e2e-llm-inference-service] # Check host [e2e-llm-inference-service] if assert_same_host and not self.is_same_host(url): [e2e-llm-inference-service] raise HostChangedError(self, url, retries) [e2e-llm-inference-service] [e2e-llm-inference-service] # Ensure that the URL we're connecting to is properly encoded [e2e-llm-inference-service] if url.startswith("/"): [e2e-llm-inference-service] url = to_str(_encode_target(url)) [e2e-llm-inference-service] else: [e2e-llm-inference-service] url = to_str(parsed_url.url) [e2e-llm-inference-service] [e2e-llm-inference-service] conn = None [e2e-llm-inference-service] [e2e-llm-inference-service] # Track whether `conn` needs to be released before [e2e-llm-inference-service] # returning/raising/recursing. Update this variable if necessary, and [e2e-llm-inference-service] # leave `release_conn` constant throughout the function. That way, if [e2e-llm-inference-service] # the function recurses, the original value of `release_conn` will be [e2e-llm-inference-service] # passed down into the recursive call, and its value will be respected. [e2e-llm-inference-service] # [e2e-llm-inference-service] # See issue #651 [1] for details. [e2e-llm-inference-service] # [e2e-llm-inference-service] # [1] [e2e-llm-inference-service] release_this_conn = release_conn [e2e-llm-inference-service] [e2e-llm-inference-service] http_tunnel_required = connection_requires_http_tunnel( [e2e-llm-inference-service] self.proxy, self.proxy_config, destination_scheme [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] # Merge the proxy headers. Only done when not using HTTP CONNECT. We [e2e-llm-inference-service] # have to copy the headers dict so we can safely change it without those [e2e-llm-inference-service] # changes being reflected in anyone else's copy. [e2e-llm-inference-service] if not http_tunnel_required: [e2e-llm-inference-service] headers = headers.copy() # type: ignore[attr-defined] [e2e-llm-inference-service] headers.update(self.proxy_headers) # type: ignore[union-attr] [e2e-llm-inference-service] [e2e-llm-inference-service] # Must keep the exception bound to a separate variable or else Python 3 [e2e-llm-inference-service] # complains about UnboundLocalError. [e2e-llm-inference-service] err = None [e2e-llm-inference-service] [e2e-llm-inference-service] # Keep track of whether we cleanly exited the except block. This [e2e-llm-inference-service] # ensures we do proper cleanup in finally. [e2e-llm-inference-service] clean_exit = False [e2e-llm-inference-service] [e2e-llm-inference-service] # Rewind body position, if needed. Record current position [e2e-llm-inference-service] # for future rewinds in the event of a redirect/retry. [e2e-llm-inference-service] body_pos = set_file_position(body, body_pos) [e2e-llm-inference-service] [e2e-llm-inference-service] try: [e2e-llm-inference-service] # Request a connection from the queue. [e2e-llm-inference-service] timeout_obj = self._get_timeout(timeout) [e2e-llm-inference-service] conn = self._get_conn(timeout=pool_timeout) [e2e-llm-inference-service] [e2e-llm-inference-service] conn.timeout = timeout_obj.connect_timeout # type: ignore[assignment] [e2e-llm-inference-service] [e2e-llm-inference-service] # Is this a closed/new connection that requires CONNECT tunnelling? [e2e-llm-inference-service] if self.proxy is not None and http_tunnel_required and conn.is_closed: [e2e-llm-inference-service] try: [e2e-llm-inference-service] self._prepare_proxy(conn) [e2e-llm-inference-service] except (BaseSSLError, OSError, SocketTimeout) as e: [e2e-llm-inference-service] self._raise_timeout( [e2e-llm-inference-service] err=e, url=self.proxy.url, timeout_value=conn.timeout [e2e-llm-inference-service] ) [e2e-llm-inference-service] raise [e2e-llm-inference-service] [e2e-llm-inference-service] # If we're going to release the connection in ``finally:``, then [e2e-llm-inference-service] # the response doesn't need to know about the connection. Otherwise [e2e-llm-inference-service] # it will also try to release it and we'll have a double-release [e2e-llm-inference-service] # mess. [e2e-llm-inference-service] response_conn = conn if not release_conn else None [e2e-llm-inference-service] [e2e-llm-inference-service] # Make the request on the HTTPConnection object [e2e-llm-inference-service] response = self._make_request( [e2e-llm-inference-service] conn, [e2e-llm-inference-service] method, [e2e-llm-inference-service] url, [e2e-llm-inference-service] timeout=timeout_obj, [e2e-llm-inference-service] body=body, [e2e-llm-inference-service] headers=headers, [e2e-llm-inference-service] chunked=chunked, [e2e-llm-inference-service] retries=retries, [e2e-llm-inference-service] response_conn=response_conn, [e2e-llm-inference-service] preload_content=preload_content, [e2e-llm-inference-service] decode_content=decode_content, [e2e-llm-inference-service] **response_kw, [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] # Everything went great! [e2e-llm-inference-service] clean_exit = True [e2e-llm-inference-service] [e2e-llm-inference-service] except EmptyPoolError: [e2e-llm-inference-service] # Didn't get a connection from the pool, no need to clean up [e2e-llm-inference-service] clean_exit = True [e2e-llm-inference-service] release_this_conn = False [e2e-llm-inference-service] raise [e2e-llm-inference-service] [e2e-llm-inference-service] except ( [e2e-llm-inference-service] TimeoutError, [e2e-llm-inference-service] HTTPException, [e2e-llm-inference-service] OSError, [e2e-llm-inference-service] ProtocolError, [e2e-llm-inference-service] BaseSSLError, [e2e-llm-inference-service] SSLError, [e2e-llm-inference-service] CertificateError, [e2e-llm-inference-service] ProxyError, [e2e-llm-inference-service] ) as e: [e2e-llm-inference-service] # Discard the connection for these exceptions. It will be [e2e-llm-inference-service] # replaced during the next _get_conn() call. [e2e-llm-inference-service] clean_exit = False [e2e-llm-inference-service] new_e: Exception = e [e2e-llm-inference-service] if isinstance(e, (BaseSSLError, CertificateError)): [e2e-llm-inference-service] new_e = SSLError(e) [e2e-llm-inference-service] if isinstance( [e2e-llm-inference-service] new_e, [e2e-llm-inference-service] ( [e2e-llm-inference-service] OSError, [e2e-llm-inference-service] NewConnectionError, [e2e-llm-inference-service] TimeoutError, [e2e-llm-inference-service] SSLError, [e2e-llm-inference-service] HTTPException, [e2e-llm-inference-service] ), [e2e-llm-inference-service] ) and (conn and conn.proxy and not conn.has_connected_to_proxy): [e2e-llm-inference-service] new_e = _wrap_proxy_error(new_e, conn.proxy.scheme) [e2e-llm-inference-service] elif isinstance(new_e, (OSError, HTTPException)): [e2e-llm-inference-service] new_e = ProtocolError("Connection aborted.", new_e) [e2e-llm-inference-service] [e2e-llm-inference-service] retries = retries.increment( [e2e-llm-inference-service] method, url, error=new_e, _pool=self, _stacktrace=sys.exc_info()[2] [e2e-llm-inference-service] ) [e2e-llm-inference-service] retries.sleep() [e2e-llm-inference-service] [e2e-llm-inference-service] # Keep track of the error for the retry warning. [e2e-llm-inference-service] err = e [e2e-llm-inference-service] [e2e-llm-inference-service] finally: [e2e-llm-inference-service] if not clean_exit: [e2e-llm-inference-service] # We hit some kind of exception, handled or otherwise. We need [e2e-llm-inference-service] # to throw the connection away unless explicitly told not to. [e2e-llm-inference-service] # Close the connection, set the variable to None, and make sure [e2e-llm-inference-service] # we put the None back in the pool to avoid leaking it. [e2e-llm-inference-service] if conn: [e2e-llm-inference-service] conn.close() [e2e-llm-inference-service] conn = None [e2e-llm-inference-service] release_this_conn = True [e2e-llm-inference-service] [e2e-llm-inference-service] if release_this_conn: [e2e-llm-inference-service] # Put the connection back to be reused. If the connection is [e2e-llm-inference-service] # expired then it will be None, which will get replaced with a [e2e-llm-inference-service] # fresh connection during _get_conn. [e2e-llm-inference-service] self._put_conn(conn) [e2e-llm-inference-service] [e2e-llm-inference-service] if not conn: [e2e-llm-inference-service] # Try again [e2e-llm-inference-service] log.warning( [e2e-llm-inference-service] "Retrying (%r) after connection broken by '%r': %s", retries, err, url [e2e-llm-inference-service] ) [e2e-llm-inference-service] > return self.urlopen( [e2e-llm-inference-service] method, [e2e-llm-inference-service] url, [e2e-llm-inference-service] body, [e2e-llm-inference-service] headers, [e2e-llm-inference-service] retries, [e2e-llm-inference-service] redirect, [e2e-llm-inference-service] assert_same_host, [e2e-llm-inference-service] timeout=timeout, [e2e-llm-inference-service] pool_timeout=pool_timeout, [e2e-llm-inference-service] release_conn=release_conn, [e2e-llm-inference-service] chunked=chunked, [e2e-llm-inference-service] body_pos=body_pos, [e2e-llm-inference-service] preload_content=preload_content, [e2e-llm-inference-service] decode_content=decode_content, [e2e-llm-inference-service] **response_kw, [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] ../../python/kserve/.venv/lib64/python3.11/site-packages/urllib3/connectionpool.py:871: [e2e-llm-inference-service] _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ [e2e-llm-inference-service] [e2e-llm-inference-service] self = [e2e-llm-inference-service] method = 'POST', url = '/kserve-ci-e2e-test/rolling-upgrade-test/v1/completions' [e2e-llm-inference-service] body = b'{"model": "facebook/opt-125m", "prompt": "KServe is a", "max_tokens": 20}' [e2e-llm-inference-service] headers = {'User-Agent': 'python-requests/2.32.3', 'Accept-Encoding': 'gzip, deflate', 'Accept': '*/*', 'Connection': 'keep-alive', 'Content-Type': 'application/json', 'Content-Length': '73'} [e2e-llm-inference-service] retries = Retry(total=3, connect=None, read=None, redirect=None, status=None) [e2e-llm-inference-service] redirect = False, assert_same_host = False [e2e-llm-inference-service] timeout = Timeout(connect=60, read=60, total=None), pool_timeout = None [e2e-llm-inference-service] release_conn = False, chunked = False, body_pos = None, preload_content = False [e2e-llm-inference-service] decode_content = False, response_kw = {} [e2e-llm-inference-service] parsed_url = Url(scheme=None, auth=None, host=None, port=None, path='/kserve-ci-e2e-test/rolling-upgrade-test/v1/completions', query=None, fragment=None) [e2e-llm-inference-service] destination_scheme = None, conn = None, release_this_conn = True [e2e-llm-inference-service] http_tunnel_required = False [e2e-llm-inference-service] err = ReadTimeoutError("HTTPConnectionPool(host='ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com', port=80): Read timed out. (read timeout=60)") [e2e-llm-inference-service] clean_exit = False [e2e-llm-inference-service] [e2e-llm-inference-service] def urlopen( # type: ignore[override] [e2e-llm-inference-service] self, [e2e-llm-inference-service] method: str, [e2e-llm-inference-service] url: str, [e2e-llm-inference-service] body: _TYPE_BODY | None = None, [e2e-llm-inference-service] headers: typing.Mapping[str, str] | None = None, [e2e-llm-inference-service] retries: Retry | bool | int | None = None, [e2e-llm-inference-service] redirect: bool = True, [e2e-llm-inference-service] assert_same_host: bool = True, [e2e-llm-inference-service] timeout: _TYPE_TIMEOUT = _DEFAULT_TIMEOUT, [e2e-llm-inference-service] pool_timeout: int | None = None, [e2e-llm-inference-service] release_conn: bool | None = None, [e2e-llm-inference-service] chunked: bool = False, [e2e-llm-inference-service] body_pos: _TYPE_BODY_POSITION | None = None, [e2e-llm-inference-service] preload_content: bool = True, [e2e-llm-inference-service] decode_content: bool = True, [e2e-llm-inference-service] **response_kw: typing.Any, [e2e-llm-inference-service] ) -> BaseHTTPResponse: [e2e-llm-inference-service] """ [e2e-llm-inference-service] Get a connection from the pool and perform an HTTP request. This is the [e2e-llm-inference-service] lowest level call for making a request, so you'll need to specify all [e2e-llm-inference-service] the raw details. [e2e-llm-inference-service] [e2e-llm-inference-service] .. note:: [e2e-llm-inference-service] [e2e-llm-inference-service] More commonly, it's appropriate to use a convenience method [e2e-llm-inference-service] such as :meth:`request`. [e2e-llm-inference-service] [e2e-llm-inference-service] .. note:: [e2e-llm-inference-service] [e2e-llm-inference-service] `release_conn` will only behave as expected if [e2e-llm-inference-service] `preload_content=False` because we want to make [e2e-llm-inference-service] `preload_content=False` the default behaviour someday soon without [e2e-llm-inference-service] breaking backwards compatibility. [e2e-llm-inference-service] [e2e-llm-inference-service] :param method: [e2e-llm-inference-service] HTTP request method (such as GET, POST, PUT, etc.) [e2e-llm-inference-service] [e2e-llm-inference-service] :param url: [e2e-llm-inference-service] The URL to perform the request on. [e2e-llm-inference-service] [e2e-llm-inference-service] :param body: [e2e-llm-inference-service] Data to send in the request body, either :class:`str`, :class:`bytes`, [e2e-llm-inference-service] an iterable of :class:`str`/:class:`bytes`, or a file-like object. [e2e-llm-inference-service] [e2e-llm-inference-service] :param headers: [e2e-llm-inference-service] Dictionary of custom headers to send, such as User-Agent, [e2e-llm-inference-service] If-None-Match, etc. If None, pool headers are used. If provided, [e2e-llm-inference-service] these headers completely replace any pool-specific headers. [e2e-llm-inference-service] [e2e-llm-inference-service] :param retries: [e2e-llm-inference-service] Configure the number of retries to allow before raising a [e2e-llm-inference-service] :class:`~urllib3.exceptions.MaxRetryError` exception. [e2e-llm-inference-service] [e2e-llm-inference-service] If ``None`` (default) will retry 3 times, see ``Retry.DEFAULT``. Pass a [e2e-llm-inference-service] :class:`~urllib3.util.retry.Retry` object for fine-grained control [e2e-llm-inference-service] over different types of retries. [e2e-llm-inference-service] Pass an integer number to retry connection errors that many times, [e2e-llm-inference-service] but no other types of errors. Pass zero to never retry. [e2e-llm-inference-service] [e2e-llm-inference-service] If ``False``, then retries are disabled and any exception is raised [e2e-llm-inference-service] immediately. Also, instead of raising a MaxRetryError on redirects, [e2e-llm-inference-service] the redirect response will be returned. [e2e-llm-inference-service] [e2e-llm-inference-service] :type retries: :class:`~urllib3.util.retry.Retry`, False, or an int. [e2e-llm-inference-service] [e2e-llm-inference-service] :param redirect: [e2e-llm-inference-service] If True, automatically handle redirects (status codes 301, 302, [e2e-llm-inference-service] 303, 307, 308). Each redirect counts as a retry. Disabling retries [e2e-llm-inference-service] will disable redirect, too. [e2e-llm-inference-service] [e2e-llm-inference-service] :param assert_same_host: [e2e-llm-inference-service] If ``True``, will make sure that the host of the pool requests is [e2e-llm-inference-service] consistent else will raise HostChangedError. When ``False``, you can [e2e-llm-inference-service] use the pool on an HTTP proxy and request foreign hosts. [e2e-llm-inference-service] [e2e-llm-inference-service] :param timeout: [e2e-llm-inference-service] If specified, overrides the default timeout for this one [e2e-llm-inference-service] request. It may be a float (in seconds) or an instance of [e2e-llm-inference-service] :class:`urllib3.util.Timeout`. [e2e-llm-inference-service] [e2e-llm-inference-service] :param pool_timeout: [e2e-llm-inference-service] If set and the pool is set to block=True, then this method will [e2e-llm-inference-service] block for ``pool_timeout`` seconds and raise EmptyPoolError if no [e2e-llm-inference-service] connection is available within the time period. [e2e-llm-inference-service] [e2e-llm-inference-service] :param bool preload_content: [e2e-llm-inference-service] If True, the response's body will be preloaded into memory. [e2e-llm-inference-service] [e2e-llm-inference-service] :param bool decode_content: [e2e-llm-inference-service] If True, will attempt to decode the body based on the [e2e-llm-inference-service] 'content-encoding' header. [e2e-llm-inference-service] [e2e-llm-inference-service] :param release_conn: [e2e-llm-inference-service] If False, then the urlopen call will not release the connection [e2e-llm-inference-service] back into the pool once a response is received (but will release if [e2e-llm-inference-service] you read the entire contents of the response such as when [e2e-llm-inference-service] `preload_content=True`). This is useful if you're not preloading [e2e-llm-inference-service] the response's content immediately. You will need to call [e2e-llm-inference-service] ``r.release_conn()`` on the response ``r`` to return the connection [e2e-llm-inference-service] back into the pool. If None, it takes the value of ``preload_content`` [e2e-llm-inference-service] which defaults to ``True``. [e2e-llm-inference-service] [e2e-llm-inference-service] :param bool chunked: [e2e-llm-inference-service] If True, urllib3 will send the body using chunked transfer [e2e-llm-inference-service] encoding. Otherwise, urllib3 will send the body using the standard [e2e-llm-inference-service] content-length form. Defaults to False. [e2e-llm-inference-service] [e2e-llm-inference-service] :param int body_pos: [e2e-llm-inference-service] Position to seek to in file-like body in the event of a retry or [e2e-llm-inference-service] redirect. Typically this won't need to be set because urllib3 will [e2e-llm-inference-service] auto-populate the value when needed. [e2e-llm-inference-service] """ [e2e-llm-inference-service] parsed_url = parse_url(url) [e2e-llm-inference-service] destination_scheme = parsed_url.scheme [e2e-llm-inference-service] [e2e-llm-inference-service] if headers is None: [e2e-llm-inference-service] headers = self.headers [e2e-llm-inference-service] [e2e-llm-inference-service] if not isinstance(retries, Retry): [e2e-llm-inference-service] retries = Retry.from_int(retries, redirect=redirect, default=self.retries) [e2e-llm-inference-service] [e2e-llm-inference-service] if release_conn is None: [e2e-llm-inference-service] release_conn = preload_content [e2e-llm-inference-service] [e2e-llm-inference-service] # Check host [e2e-llm-inference-service] if assert_same_host and not self.is_same_host(url): [e2e-llm-inference-service] raise HostChangedError(self, url, retries) [e2e-llm-inference-service] [e2e-llm-inference-service] # Ensure that the URL we're connecting to is properly encoded [e2e-llm-inference-service] if url.startswith("/"): [e2e-llm-inference-service] url = to_str(_encode_target(url)) [e2e-llm-inference-service] else: [e2e-llm-inference-service] url = to_str(parsed_url.url) [e2e-llm-inference-service] [e2e-llm-inference-service] conn = None [e2e-llm-inference-service] [e2e-llm-inference-service] # Track whether `conn` needs to be released before [e2e-llm-inference-service] # returning/raising/recursing. Update this variable if necessary, and [e2e-llm-inference-service] # leave `release_conn` constant throughout the function. That way, if [e2e-llm-inference-service] # the function recurses, the original value of `release_conn` will be [e2e-llm-inference-service] # passed down into the recursive call, and its value will be respected. [e2e-llm-inference-service] # [e2e-llm-inference-service] # See issue #651 [1] for details. [e2e-llm-inference-service] # [e2e-llm-inference-service] # [1] [e2e-llm-inference-service] release_this_conn = release_conn [e2e-llm-inference-service] [e2e-llm-inference-service] http_tunnel_required = connection_requires_http_tunnel( [e2e-llm-inference-service] self.proxy, self.proxy_config, destination_scheme [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] # Merge the proxy headers. Only done when not using HTTP CONNECT. We [e2e-llm-inference-service] # have to copy the headers dict so we can safely change it without those [e2e-llm-inference-service] # changes being reflected in anyone else's copy. [e2e-llm-inference-service] if not http_tunnel_required: [e2e-llm-inference-service] headers = headers.copy() # type: ignore[attr-defined] [e2e-llm-inference-service] headers.update(self.proxy_headers) # type: ignore[union-attr] [e2e-llm-inference-service] [e2e-llm-inference-service] # Must keep the exception bound to a separate variable or else Python 3 [e2e-llm-inference-service] # complains about UnboundLocalError. [e2e-llm-inference-service] err = None [e2e-llm-inference-service] [e2e-llm-inference-service] # Keep track of whether we cleanly exited the except block. This [e2e-llm-inference-service] # ensures we do proper cleanup in finally. [e2e-llm-inference-service] clean_exit = False [e2e-llm-inference-service] [e2e-llm-inference-service] # Rewind body position, if needed. Record current position [e2e-llm-inference-service] # for future rewinds in the event of a redirect/retry. [e2e-llm-inference-service] body_pos = set_file_position(body, body_pos) [e2e-llm-inference-service] [e2e-llm-inference-service] try: [e2e-llm-inference-service] # Request a connection from the queue. [e2e-llm-inference-service] timeout_obj = self._get_timeout(timeout) [e2e-llm-inference-service] conn = self._get_conn(timeout=pool_timeout) [e2e-llm-inference-service] [e2e-llm-inference-service] conn.timeout = timeout_obj.connect_timeout # type: ignore[assignment] [e2e-llm-inference-service] [e2e-llm-inference-service] # Is this a closed/new connection that requires CONNECT tunnelling? [e2e-llm-inference-service] if self.proxy is not None and http_tunnel_required and conn.is_closed: [e2e-llm-inference-service] try: [e2e-llm-inference-service] self._prepare_proxy(conn) [e2e-llm-inference-service] except (BaseSSLError, OSError, SocketTimeout) as e: [e2e-llm-inference-service] self._raise_timeout( [e2e-llm-inference-service] err=e, url=self.proxy.url, timeout_value=conn.timeout [e2e-llm-inference-service] ) [e2e-llm-inference-service] raise [e2e-llm-inference-service] [e2e-llm-inference-service] # If we're going to release the connection in ``finally:``, then [e2e-llm-inference-service] # the response doesn't need to know about the connection. Otherwise [e2e-llm-inference-service] # it will also try to release it and we'll have a double-release [e2e-llm-inference-service] # mess. [e2e-llm-inference-service] response_conn = conn if not release_conn else None [e2e-llm-inference-service] [e2e-llm-inference-service] # Make the request on the HTTPConnection object [e2e-llm-inference-service] response = self._make_request( [e2e-llm-inference-service] conn, [e2e-llm-inference-service] method, [e2e-llm-inference-service] url, [e2e-llm-inference-service] timeout=timeout_obj, [e2e-llm-inference-service] body=body, [e2e-llm-inference-service] headers=headers, [e2e-llm-inference-service] chunked=chunked, [e2e-llm-inference-service] retries=retries, [e2e-llm-inference-service] response_conn=response_conn, [e2e-llm-inference-service] preload_content=preload_content, [e2e-llm-inference-service] decode_content=decode_content, [e2e-llm-inference-service] **response_kw, [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] # Everything went great! [e2e-llm-inference-service] clean_exit = True [e2e-llm-inference-service] [e2e-llm-inference-service] except EmptyPoolError: [e2e-llm-inference-service] # Didn't get a connection from the pool, no need to clean up [e2e-llm-inference-service] clean_exit = True [e2e-llm-inference-service] release_this_conn = False [e2e-llm-inference-service] raise [e2e-llm-inference-service] [e2e-llm-inference-service] except ( [e2e-llm-inference-service] TimeoutError, [e2e-llm-inference-service] HTTPException, [e2e-llm-inference-service] OSError, [e2e-llm-inference-service] ProtocolError, [e2e-llm-inference-service] BaseSSLError, [e2e-llm-inference-service] SSLError, [e2e-llm-inference-service] CertificateError, [e2e-llm-inference-service] ProxyError, [e2e-llm-inference-service] ) as e: [e2e-llm-inference-service] # Discard the connection for these exceptions. It will be [e2e-llm-inference-service] # replaced during the next _get_conn() call. [e2e-llm-inference-service] clean_exit = False [e2e-llm-inference-service] new_e: Exception = e [e2e-llm-inference-service] if isinstance(e, (BaseSSLError, CertificateError)): [e2e-llm-inference-service] new_e = SSLError(e) [e2e-llm-inference-service] if isinstance( [e2e-llm-inference-service] new_e, [e2e-llm-inference-service] ( [e2e-llm-inference-service] OSError, [e2e-llm-inference-service] NewConnectionError, [e2e-llm-inference-service] TimeoutError, [e2e-llm-inference-service] SSLError, [e2e-llm-inference-service] HTTPException, [e2e-llm-inference-service] ), [e2e-llm-inference-service] ) and (conn and conn.proxy and not conn.has_connected_to_proxy): [e2e-llm-inference-service] new_e = _wrap_proxy_error(new_e, conn.proxy.scheme) [e2e-llm-inference-service] elif isinstance(new_e, (OSError, HTTPException)): [e2e-llm-inference-service] new_e = ProtocolError("Connection aborted.", new_e) [e2e-llm-inference-service] [e2e-llm-inference-service] retries = retries.increment( [e2e-llm-inference-service] method, url, error=new_e, _pool=self, _stacktrace=sys.exc_info()[2] [e2e-llm-inference-service] ) [e2e-llm-inference-service] retries.sleep() [e2e-llm-inference-service] [e2e-llm-inference-service] # Keep track of the error for the retry warning. [e2e-llm-inference-service] err = e [e2e-llm-inference-service] [e2e-llm-inference-service] finally: [e2e-llm-inference-service] if not clean_exit: [e2e-llm-inference-service] # We hit some kind of exception, handled or otherwise. We need [e2e-llm-inference-service] # to throw the connection away unless explicitly told not to. [e2e-llm-inference-service] # Close the connection, set the variable to None, and make sure [e2e-llm-inference-service] # we put the None back in the pool to avoid leaking it. [e2e-llm-inference-service] if conn: [e2e-llm-inference-service] conn.close() [e2e-llm-inference-service] conn = None [e2e-llm-inference-service] release_this_conn = True [e2e-llm-inference-service] [e2e-llm-inference-service] if release_this_conn: [e2e-llm-inference-service] # Put the connection back to be reused. If the connection is [e2e-llm-inference-service] # expired then it will be None, which will get replaced with a [e2e-llm-inference-service] # fresh connection during _get_conn. [e2e-llm-inference-service] self._put_conn(conn) [e2e-llm-inference-service] [e2e-llm-inference-service] if not conn: [e2e-llm-inference-service] # Try again [e2e-llm-inference-service] log.warning( [e2e-llm-inference-service] "Retrying (%r) after connection broken by '%r': %s", retries, err, url [e2e-llm-inference-service] ) [e2e-llm-inference-service] > return self.urlopen( [e2e-llm-inference-service] method, [e2e-llm-inference-service] url, [e2e-llm-inference-service] body, [e2e-llm-inference-service] headers, [e2e-llm-inference-service] retries, [e2e-llm-inference-service] redirect, [e2e-llm-inference-service] assert_same_host, [e2e-llm-inference-service] timeout=timeout, [e2e-llm-inference-service] pool_timeout=pool_timeout, [e2e-llm-inference-service] release_conn=release_conn, [e2e-llm-inference-service] chunked=chunked, [e2e-llm-inference-service] body_pos=body_pos, [e2e-llm-inference-service] preload_content=preload_content, [e2e-llm-inference-service] decode_content=decode_content, [e2e-llm-inference-service] **response_kw, [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] ../../python/kserve/.venv/lib64/python3.11/site-packages/urllib3/connectionpool.py:871: [e2e-llm-inference-service] _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ [e2e-llm-inference-service] [e2e-llm-inference-service] self = [e2e-llm-inference-service] method = 'POST', url = '/kserve-ci-e2e-test/rolling-upgrade-test/v1/completions' [e2e-llm-inference-service] body = b'{"model": "facebook/opt-125m", "prompt": "KServe is a", "max_tokens": 20}' [e2e-llm-inference-service] headers = {'User-Agent': 'python-requests/2.32.3', 'Accept-Encoding': 'gzip, deflate', 'Accept': '*/*', 'Connection': 'keep-alive', 'Content-Type': 'application/json', 'Content-Length': '73'} [e2e-llm-inference-service] retries = Retry(total=2, connect=None, read=None, redirect=None, status=None) [e2e-llm-inference-service] redirect = False, assert_same_host = False [e2e-llm-inference-service] timeout = Timeout(connect=60, read=60, total=None), pool_timeout = None [e2e-llm-inference-service] release_conn = False, chunked = False, body_pos = None, preload_content = False [e2e-llm-inference-service] decode_content = False, response_kw = {} [e2e-llm-inference-service] parsed_url = Url(scheme=None, auth=None, host=None, port=None, path='/kserve-ci-e2e-test/rolling-upgrade-test/v1/completions', query=None, fragment=None) [e2e-llm-inference-service] destination_scheme = None, conn = None, release_this_conn = True [e2e-llm-inference-service] http_tunnel_required = False [e2e-llm-inference-service] err = ReadTimeoutError("HTTPConnectionPool(host='ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com', port=80): Read timed out. (read timeout=60)") [e2e-llm-inference-service] clean_exit = False [e2e-llm-inference-service] [e2e-llm-inference-service] def urlopen( # type: ignore[override] [e2e-llm-inference-service] self, [e2e-llm-inference-service] method: str, [e2e-llm-inference-service] url: str, [e2e-llm-inference-service] body: _TYPE_BODY | None = None, [e2e-llm-inference-service] headers: typing.Mapping[str, str] | None = None, [e2e-llm-inference-service] retries: Retry | bool | int | None = None, [e2e-llm-inference-service] redirect: bool = True, [e2e-llm-inference-service] assert_same_host: bool = True, [e2e-llm-inference-service] timeout: _TYPE_TIMEOUT = _DEFAULT_TIMEOUT, [e2e-llm-inference-service] pool_timeout: int | None = None, [e2e-llm-inference-service] release_conn: bool | None = None, [e2e-llm-inference-service] chunked: bool = False, [e2e-llm-inference-service] body_pos: _TYPE_BODY_POSITION | None = None, [e2e-llm-inference-service] preload_content: bool = True, [e2e-llm-inference-service] decode_content: bool = True, [e2e-llm-inference-service] **response_kw: typing.Any, [e2e-llm-inference-service] ) -> BaseHTTPResponse: [e2e-llm-inference-service] """ [e2e-llm-inference-service] Get a connection from the pool and perform an HTTP request. This is the [e2e-llm-inference-service] lowest level call for making a request, so you'll need to specify all [e2e-llm-inference-service] the raw details. [e2e-llm-inference-service] [e2e-llm-inference-service] .. note:: [e2e-llm-inference-service] [e2e-llm-inference-service] More commonly, it's appropriate to use a convenience method [e2e-llm-inference-service] such as :meth:`request`. [e2e-llm-inference-service] [e2e-llm-inference-service] .. note:: [e2e-llm-inference-service] [e2e-llm-inference-service] `release_conn` will only behave as expected if [e2e-llm-inference-service] `preload_content=False` because we want to make [e2e-llm-inference-service] `preload_content=False` the default behaviour someday soon without [e2e-llm-inference-service] breaking backwards compatibility. [e2e-llm-inference-service] [e2e-llm-inference-service] :param method: [e2e-llm-inference-service] HTTP request method (such as GET, POST, PUT, etc.) [e2e-llm-inference-service] [e2e-llm-inference-service] :param url: [e2e-llm-inference-service] The URL to perform the request on. [e2e-llm-inference-service] [e2e-llm-inference-service] :param body: [e2e-llm-inference-service] Data to send in the request body, either :class:`str`, :class:`bytes`, [e2e-llm-inference-service] an iterable of :class:`str`/:class:`bytes`, or a file-like object. [e2e-llm-inference-service] [e2e-llm-inference-service] :param headers: [e2e-llm-inference-service] Dictionary of custom headers to send, such as User-Agent, [e2e-llm-inference-service] If-None-Match, etc. If None, pool headers are used. If provided, [e2e-llm-inference-service] these headers completely replace any pool-specific headers. [e2e-llm-inference-service] [e2e-llm-inference-service] :param retries: [e2e-llm-inference-service] Configure the number of retries to allow before raising a [e2e-llm-inference-service] :class:`~urllib3.exceptions.MaxRetryError` exception. [e2e-llm-inference-service] [e2e-llm-inference-service] If ``None`` (default) will retry 3 times, see ``Retry.DEFAULT``. Pass a [e2e-llm-inference-service] :class:`~urllib3.util.retry.Retry` object for fine-grained control [e2e-llm-inference-service] over different types of retries. [e2e-llm-inference-service] Pass an integer number to retry connection errors that many times, [e2e-llm-inference-service] but no other types of errors. Pass zero to never retry. [e2e-llm-inference-service] [e2e-llm-inference-service] If ``False``, then retries are disabled and any exception is raised [e2e-llm-inference-service] immediately. Also, instead of raising a MaxRetryError on redirects, [e2e-llm-inference-service] the redirect response will be returned. [e2e-llm-inference-service] [e2e-llm-inference-service] :type retries: :class:`~urllib3.util.retry.Retry`, False, or an int. [e2e-llm-inference-service] [e2e-llm-inference-service] :param redirect: [e2e-llm-inference-service] If True, automatically handle redirects (status codes 301, 302, [e2e-llm-inference-service] 303, 307, 308). Each redirect counts as a retry. Disabling retries [e2e-llm-inference-service] will disable redirect, too. [e2e-llm-inference-service] [e2e-llm-inference-service] :param assert_same_host: [e2e-llm-inference-service] If ``True``, will make sure that the host of the pool requests is [e2e-llm-inference-service] consistent else will raise HostChangedError. When ``False``, you can [e2e-llm-inference-service] use the pool on an HTTP proxy and request foreign hosts. [e2e-llm-inference-service] [e2e-llm-inference-service] :param timeout: [e2e-llm-inference-service] If specified, overrides the default timeout for this one [e2e-llm-inference-service] request. It may be a float (in seconds) or an instance of [e2e-llm-inference-service] :class:`urllib3.util.Timeout`. [e2e-llm-inference-service] [e2e-llm-inference-service] :param pool_timeout: [e2e-llm-inference-service] If set and the pool is set to block=True, then this method will [e2e-llm-inference-service] block for ``pool_timeout`` seconds and raise EmptyPoolError if no [e2e-llm-inference-service] connection is available within the time period. [e2e-llm-inference-service] [e2e-llm-inference-service] :param bool preload_content: [e2e-llm-inference-service] If True, the response's body will be preloaded into memory. [e2e-llm-inference-service] [e2e-llm-inference-service] :param bool decode_content: [e2e-llm-inference-service] If True, will attempt to decode the body based on the [e2e-llm-inference-service] 'content-encoding' header. [e2e-llm-inference-service] [e2e-llm-inference-service] :param release_conn: [e2e-llm-inference-service] If False, then the urlopen call will not release the connection [e2e-llm-inference-service] back into the pool once a response is received (but will release if [e2e-llm-inference-service] you read the entire contents of the response such as when [e2e-llm-inference-service] `preload_content=True`). This is useful if you're not preloading [e2e-llm-inference-service] the response's content immediately. You will need to call [e2e-llm-inference-service] ``r.release_conn()`` on the response ``r`` to return the connection [e2e-llm-inference-service] back into the pool. If None, it takes the value of ``preload_content`` [e2e-llm-inference-service] which defaults to ``True``. [e2e-llm-inference-service] [e2e-llm-inference-service] :param bool chunked: [e2e-llm-inference-service] If True, urllib3 will send the body using chunked transfer [e2e-llm-inference-service] encoding. Otherwise, urllib3 will send the body using the standard [e2e-llm-inference-service] content-length form. Defaults to False. [e2e-llm-inference-service] [e2e-llm-inference-service] :param int body_pos: [e2e-llm-inference-service] Position to seek to in file-like body in the event of a retry or [e2e-llm-inference-service] redirect. Typically this won't need to be set because urllib3 will [e2e-llm-inference-service] auto-populate the value when needed. [e2e-llm-inference-service] """ [e2e-llm-inference-service] parsed_url = parse_url(url) [e2e-llm-inference-service] destination_scheme = parsed_url.scheme [e2e-llm-inference-service] [e2e-llm-inference-service] if headers is None: [e2e-llm-inference-service] headers = self.headers [e2e-llm-inference-service] [e2e-llm-inference-service] if not isinstance(retries, Retry): [e2e-llm-inference-service] retries = Retry.from_int(retries, redirect=redirect, default=self.retries) [e2e-llm-inference-service] [e2e-llm-inference-service] if release_conn is None: [e2e-llm-inference-service] release_conn = preload_content [e2e-llm-inference-service] [e2e-llm-inference-service] # Check host [e2e-llm-inference-service] if assert_same_host and not self.is_same_host(url): [e2e-llm-inference-service] raise HostChangedError(self, url, retries) [e2e-llm-inference-service] [e2e-llm-inference-service] # Ensure that the URL we're connecting to is properly encoded [e2e-llm-inference-service] if url.startswith("/"): [e2e-llm-inference-service] url = to_str(_encode_target(url)) [e2e-llm-inference-service] else: [e2e-llm-inference-service] url = to_str(parsed_url.url) [e2e-llm-inference-service] [e2e-llm-inference-service] conn = None [e2e-llm-inference-service] [e2e-llm-inference-service] # Track whether `conn` needs to be released before [e2e-llm-inference-service] # returning/raising/recursing. Update this variable if necessary, and [e2e-llm-inference-service] # leave `release_conn` constant throughout the function. That way, if [e2e-llm-inference-service] # the function recurses, the original value of `release_conn` will be [e2e-llm-inference-service] # passed down into the recursive call, and its value will be respected. [e2e-llm-inference-service] # [e2e-llm-inference-service] # See issue #651 [1] for details. [e2e-llm-inference-service] # [e2e-llm-inference-service] # [1] [e2e-llm-inference-service] release_this_conn = release_conn [e2e-llm-inference-service] [e2e-llm-inference-service] http_tunnel_required = connection_requires_http_tunnel( [e2e-llm-inference-service] self.proxy, self.proxy_config, destination_scheme [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] # Merge the proxy headers. Only done when not using HTTP CONNECT. We [e2e-llm-inference-service] # have to copy the headers dict so we can safely change it without those [e2e-llm-inference-service] # changes being reflected in anyone else's copy. [e2e-llm-inference-service] if not http_tunnel_required: [e2e-llm-inference-service] headers = headers.copy() # type: ignore[attr-defined] [e2e-llm-inference-service] headers.update(self.proxy_headers) # type: ignore[union-attr] [e2e-llm-inference-service] [e2e-llm-inference-service] # Must keep the exception bound to a separate variable or else Python 3 [e2e-llm-inference-service] # complains about UnboundLocalError. [e2e-llm-inference-service] err = None [e2e-llm-inference-service] [e2e-llm-inference-service] # Keep track of whether we cleanly exited the except block. This [e2e-llm-inference-service] # ensures we do proper cleanup in finally. [e2e-llm-inference-service] clean_exit = False [e2e-llm-inference-service] [e2e-llm-inference-service] # Rewind body position, if needed. Record current position [e2e-llm-inference-service] # for future rewinds in the event of a redirect/retry. [e2e-llm-inference-service] body_pos = set_file_position(body, body_pos) [e2e-llm-inference-service] [e2e-llm-inference-service] try: [e2e-llm-inference-service] # Request a connection from the queue. [e2e-llm-inference-service] timeout_obj = self._get_timeout(timeout) [e2e-llm-inference-service] conn = self._get_conn(timeout=pool_timeout) [e2e-llm-inference-service] [e2e-llm-inference-service] conn.timeout = timeout_obj.connect_timeout # type: ignore[assignment] [e2e-llm-inference-service] [e2e-llm-inference-service] # Is this a closed/new connection that requires CONNECT tunnelling? [e2e-llm-inference-service] if self.proxy is not None and http_tunnel_required and conn.is_closed: [e2e-llm-inference-service] try: [e2e-llm-inference-service] self._prepare_proxy(conn) [e2e-llm-inference-service] except (BaseSSLError, OSError, SocketTimeout) as e: [e2e-llm-inference-service] self._raise_timeout( [e2e-llm-inference-service] err=e, url=self.proxy.url, timeout_value=conn.timeout [e2e-llm-inference-service] ) [e2e-llm-inference-service] raise [e2e-llm-inference-service] [e2e-llm-inference-service] # If we're going to release the connection in ``finally:``, then [e2e-llm-inference-service] # the response doesn't need to know about the connection. Otherwise [e2e-llm-inference-service] # it will also try to release it and we'll have a double-release [e2e-llm-inference-service] # mess. [e2e-llm-inference-service] response_conn = conn if not release_conn else None [e2e-llm-inference-service] [e2e-llm-inference-service] # Make the request on the HTTPConnection object [e2e-llm-inference-service] response = self._make_request( [e2e-llm-inference-service] conn, [e2e-llm-inference-service] method, [e2e-llm-inference-service] url, [e2e-llm-inference-service] timeout=timeout_obj, [e2e-llm-inference-service] body=body, [e2e-llm-inference-service] headers=headers, [e2e-llm-inference-service] chunked=chunked, [e2e-llm-inference-service] retries=retries, [e2e-llm-inference-service] response_conn=response_conn, [e2e-llm-inference-service] preload_content=preload_content, [e2e-llm-inference-service] decode_content=decode_content, [e2e-llm-inference-service] **response_kw, [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] # Everything went great! [e2e-llm-inference-service] clean_exit = True [e2e-llm-inference-service] [e2e-llm-inference-service] except EmptyPoolError: [e2e-llm-inference-service] # Didn't get a connection from the pool, no need to clean up [e2e-llm-inference-service] clean_exit = True [e2e-llm-inference-service] release_this_conn = False [e2e-llm-inference-service] raise [e2e-llm-inference-service] [e2e-llm-inference-service] except ( [e2e-llm-inference-service] TimeoutError, [e2e-llm-inference-service] HTTPException, [e2e-llm-inference-service] OSError, [e2e-llm-inference-service] ProtocolError, [e2e-llm-inference-service] BaseSSLError, [e2e-llm-inference-service] SSLError, [e2e-llm-inference-service] CertificateError, [e2e-llm-inference-service] ProxyError, [e2e-llm-inference-service] ) as e: [e2e-llm-inference-service] # Discard the connection for these exceptions. It will be [e2e-llm-inference-service] # replaced during the next _get_conn() call. [e2e-llm-inference-service] clean_exit = False [e2e-llm-inference-service] new_e: Exception = e [e2e-llm-inference-service] if isinstance(e, (BaseSSLError, CertificateError)): [e2e-llm-inference-service] new_e = SSLError(e) [e2e-llm-inference-service] if isinstance( [e2e-llm-inference-service] new_e, [e2e-llm-inference-service] ( [e2e-llm-inference-service] OSError, [e2e-llm-inference-service] NewConnectionError, [e2e-llm-inference-service] TimeoutError, [e2e-llm-inference-service] SSLError, [e2e-llm-inference-service] HTTPException, [e2e-llm-inference-service] ), [e2e-llm-inference-service] ) and (conn and conn.proxy and not conn.has_connected_to_proxy): [e2e-llm-inference-service] new_e = _wrap_proxy_error(new_e, conn.proxy.scheme) [e2e-llm-inference-service] elif isinstance(new_e, (OSError, HTTPException)): [e2e-llm-inference-service] new_e = ProtocolError("Connection aborted.", new_e) [e2e-llm-inference-service] [e2e-llm-inference-service] retries = retries.increment( [e2e-llm-inference-service] method, url, error=new_e, _pool=self, _stacktrace=sys.exc_info()[2] [e2e-llm-inference-service] ) [e2e-llm-inference-service] retries.sleep() [e2e-llm-inference-service] [e2e-llm-inference-service] # Keep track of the error for the retry warning. [e2e-llm-inference-service] err = e [e2e-llm-inference-service] [e2e-llm-inference-service] finally: [e2e-llm-inference-service] if not clean_exit: [e2e-llm-inference-service] # We hit some kind of exception, handled or otherwise. We need [e2e-llm-inference-service] # to throw the connection away unless explicitly told not to. [e2e-llm-inference-service] # Close the connection, set the variable to None, and make sure [e2e-llm-inference-service] # we put the None back in the pool to avoid leaking it. [e2e-llm-inference-service] if conn: [e2e-llm-inference-service] conn.close() [e2e-llm-inference-service] conn = None [e2e-llm-inference-service] release_this_conn = True [e2e-llm-inference-service] [e2e-llm-inference-service] if release_this_conn: [e2e-llm-inference-service] # Put the connection back to be reused. If the connection is [e2e-llm-inference-service] # expired then it will be None, which will get replaced with a [e2e-llm-inference-service] # fresh connection during _get_conn. [e2e-llm-inference-service] self._put_conn(conn) [e2e-llm-inference-service] [e2e-llm-inference-service] if not conn: [e2e-llm-inference-service] # Try again [e2e-llm-inference-service] log.warning( [e2e-llm-inference-service] "Retrying (%r) after connection broken by '%r': %s", retries, err, url [e2e-llm-inference-service] ) [e2e-llm-inference-service] > return self.urlopen( [e2e-llm-inference-service] method, [e2e-llm-inference-service] url, [e2e-llm-inference-service] body, [e2e-llm-inference-service] headers, [e2e-llm-inference-service] retries, [e2e-llm-inference-service] redirect, [e2e-llm-inference-service] assert_same_host, [e2e-llm-inference-service] timeout=timeout, [e2e-llm-inference-service] pool_timeout=pool_timeout, [e2e-llm-inference-service] release_conn=release_conn, [e2e-llm-inference-service] chunked=chunked, [e2e-llm-inference-service] body_pos=body_pos, [e2e-llm-inference-service] preload_content=preload_content, [e2e-llm-inference-service] decode_content=decode_content, [e2e-llm-inference-service] **response_kw, [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] ../../python/kserve/.venv/lib64/python3.11/site-packages/urllib3/connectionpool.py:871: [e2e-llm-inference-service] _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ [e2e-llm-inference-service] [e2e-llm-inference-service] self = [e2e-llm-inference-service] method = 'POST', url = '/kserve-ci-e2e-test/rolling-upgrade-test/v1/completions' [e2e-llm-inference-service] body = b'{"model": "facebook/opt-125m", "prompt": "KServe is a", "max_tokens": 20}' [e2e-llm-inference-service] headers = {'User-Agent': 'python-requests/2.32.3', 'Accept-Encoding': 'gzip, deflate', 'Accept': '*/*', 'Connection': 'keep-alive', 'Content-Type': 'application/json', 'Content-Length': '73'} [e2e-llm-inference-service] retries = Retry(total=1, connect=None, read=None, redirect=None, status=None) [e2e-llm-inference-service] redirect = False, assert_same_host = False [e2e-llm-inference-service] timeout = Timeout(connect=60, read=60, total=None), pool_timeout = None [e2e-llm-inference-service] release_conn = False, chunked = False, body_pos = None, preload_content = False [e2e-llm-inference-service] decode_content = False, response_kw = {} [e2e-llm-inference-service] parsed_url = Url(scheme=None, auth=None, host=None, port=None, path='/kserve-ci-e2e-test/rolling-upgrade-test/v1/completions', query=None, fragment=None) [e2e-llm-inference-service] destination_scheme = None, conn = None, release_this_conn = True [e2e-llm-inference-service] http_tunnel_required = False [e2e-llm-inference-service] err = ReadTimeoutError("HTTPConnectionPool(host='ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com', port=80): Read timed out. (read timeout=60)") [e2e-llm-inference-service] clean_exit = False [e2e-llm-inference-service] [e2e-llm-inference-service] def urlopen( # type: ignore[override] [e2e-llm-inference-service] self, [e2e-llm-inference-service] method: str, [e2e-llm-inference-service] url: str, [e2e-llm-inference-service] body: _TYPE_BODY | None = None, [e2e-llm-inference-service] headers: typing.Mapping[str, str] | None = None, [e2e-llm-inference-service] retries: Retry | bool | int | None = None, [e2e-llm-inference-service] redirect: bool = True, [e2e-llm-inference-service] assert_same_host: bool = True, [e2e-llm-inference-service] timeout: _TYPE_TIMEOUT = _DEFAULT_TIMEOUT, [e2e-llm-inference-service] pool_timeout: int | None = None, [e2e-llm-inference-service] release_conn: bool | None = None, [e2e-llm-inference-service] chunked: bool = False, [e2e-llm-inference-service] body_pos: _TYPE_BODY_POSITION | None = None, [e2e-llm-inference-service] preload_content: bool = True, [e2e-llm-inference-service] decode_content: bool = True, [e2e-llm-inference-service] **response_kw: typing.Any, [e2e-llm-inference-service] ) -> BaseHTTPResponse: [e2e-llm-inference-service] """ [e2e-llm-inference-service] Get a connection from the pool and perform an HTTP request. This is the [e2e-llm-inference-service] lowest level call for making a request, so you'll need to specify all [e2e-llm-inference-service] the raw details. [e2e-llm-inference-service] [e2e-llm-inference-service] .. note:: [e2e-llm-inference-service] [e2e-llm-inference-service] More commonly, it's appropriate to use a convenience method [e2e-llm-inference-service] such as :meth:`request`. [e2e-llm-inference-service] [e2e-llm-inference-service] .. note:: [e2e-llm-inference-service] [e2e-llm-inference-service] `release_conn` will only behave as expected if [e2e-llm-inference-service] `preload_content=False` because we want to make [e2e-llm-inference-service] `preload_content=False` the default behaviour someday soon without [e2e-llm-inference-service] breaking backwards compatibility. [e2e-llm-inference-service] [e2e-llm-inference-service] :param method: [e2e-llm-inference-service] HTTP request method (such as GET, POST, PUT, etc.) [e2e-llm-inference-service] [e2e-llm-inference-service] :param url: [e2e-llm-inference-service] The URL to perform the request on. [e2e-llm-inference-service] [e2e-llm-inference-service] :param body: [e2e-llm-inference-service] Data to send in the request body, either :class:`str`, :class:`bytes`, [e2e-llm-inference-service] an iterable of :class:`str`/:class:`bytes`, or a file-like object. [e2e-llm-inference-service] [e2e-llm-inference-service] :param headers: [e2e-llm-inference-service] Dictionary of custom headers to send, such as User-Agent, [e2e-llm-inference-service] If-None-Match, etc. If None, pool headers are used. If provided, [e2e-llm-inference-service] these headers completely replace any pool-specific headers. [e2e-llm-inference-service] [e2e-llm-inference-service] :param retries: [e2e-llm-inference-service] Configure the number of retries to allow before raising a [e2e-llm-inference-service] :class:`~urllib3.exceptions.MaxRetryError` exception. [e2e-llm-inference-service] [e2e-llm-inference-service] If ``None`` (default) will retry 3 times, see ``Retry.DEFAULT``. Pass a [e2e-llm-inference-service] :class:`~urllib3.util.retry.Retry` object for fine-grained control [e2e-llm-inference-service] over different types of retries. [e2e-llm-inference-service] Pass an integer number to retry connection errors that many times, [e2e-llm-inference-service] but no other types of errors. Pass zero to never retry. [e2e-llm-inference-service] [e2e-llm-inference-service] If ``False``, then retries are disabled and any exception is raised [e2e-llm-inference-service] immediately. Also, instead of raising a MaxRetryError on redirects, [e2e-llm-inference-service] the redirect response will be returned. [e2e-llm-inference-service] [e2e-llm-inference-service] :type retries: :class:`~urllib3.util.retry.Retry`, False, or an int. [e2e-llm-inference-service] [e2e-llm-inference-service] :param redirect: [e2e-llm-inference-service] If True, automatically handle redirects (status codes 301, 302, [e2e-llm-inference-service] 303, 307, 308). Each redirect counts as a retry. Disabling retries [e2e-llm-inference-service] will disable redirect, too. [e2e-llm-inference-service] [e2e-llm-inference-service] :param assert_same_host: [e2e-llm-inference-service] If ``True``, will make sure that the host of the pool requests is [e2e-llm-inference-service] consistent else will raise HostChangedError. When ``False``, you can [e2e-llm-inference-service] use the pool on an HTTP proxy and request foreign hosts. [e2e-llm-inference-service] [e2e-llm-inference-service] :param timeout: [e2e-llm-inference-service] If specified, overrides the default timeout for this one [e2e-llm-inference-service] request. It may be a float (in seconds) or an instance of [e2e-llm-inference-service] :class:`urllib3.util.Timeout`. [e2e-llm-inference-service] [e2e-llm-inference-service] :param pool_timeout: [e2e-llm-inference-service] If set and the pool is set to block=True, then this method will [e2e-llm-inference-service] block for ``pool_timeout`` seconds and raise EmptyPoolError if no [e2e-llm-inference-service] connection is available within the time period. [e2e-llm-inference-service] [e2e-llm-inference-service] :param bool preload_content: [e2e-llm-inference-service] If True, the response's body will be preloaded into memory. [e2e-llm-inference-service] [e2e-llm-inference-service] :param bool decode_content: [e2e-llm-inference-service] If True, will attempt to decode the body based on the [e2e-llm-inference-service] 'content-encoding' header. [e2e-llm-inference-service] [e2e-llm-inference-service] :param release_conn: [e2e-llm-inference-service] If False, then the urlopen call will not release the connection [e2e-llm-inference-service] back into the pool once a response is received (but will release if [e2e-llm-inference-service] you read the entire contents of the response such as when [e2e-llm-inference-service] `preload_content=True`). This is useful if you're not preloading [e2e-llm-inference-service] the response's content immediately. You will need to call [e2e-llm-inference-service] ``r.release_conn()`` on the response ``r`` to return the connection [e2e-llm-inference-service] back into the pool. If None, it takes the value of ``preload_content`` [e2e-llm-inference-service] which defaults to ``True``. [e2e-llm-inference-service] [e2e-llm-inference-service] :param bool chunked: [e2e-llm-inference-service] If True, urllib3 will send the body using chunked transfer [e2e-llm-inference-service] encoding. Otherwise, urllib3 will send the body using the standard [e2e-llm-inference-service] content-length form. Defaults to False. [e2e-llm-inference-service] [e2e-llm-inference-service] :param int body_pos: [e2e-llm-inference-service] Position to seek to in file-like body in the event of a retry or [e2e-llm-inference-service] redirect. Typically this won't need to be set because urllib3 will [e2e-llm-inference-service] auto-populate the value when needed. [e2e-llm-inference-service] """ [e2e-llm-inference-service] parsed_url = parse_url(url) [e2e-llm-inference-service] destination_scheme = parsed_url.scheme [e2e-llm-inference-service] [e2e-llm-inference-service] if headers is None: [e2e-llm-inference-service] headers = self.headers [e2e-llm-inference-service] [e2e-llm-inference-service] if not isinstance(retries, Retry): [e2e-llm-inference-service] retries = Retry.from_int(retries, redirect=redirect, default=self.retries) [e2e-llm-inference-service] [e2e-llm-inference-service] if release_conn is None: [e2e-llm-inference-service] release_conn = preload_content [e2e-llm-inference-service] [e2e-llm-inference-service] # Check host [e2e-llm-inference-service] if assert_same_host and not self.is_same_host(url): [e2e-llm-inference-service] raise HostChangedError(self, url, retries) [e2e-llm-inference-service] [e2e-llm-inference-service] # Ensure that the URL we're connecting to is properly encoded [e2e-llm-inference-service] if url.startswith("/"): [e2e-llm-inference-service] url = to_str(_encode_target(url)) [e2e-llm-inference-service] else: [e2e-llm-inference-service] url = to_str(parsed_url.url) [e2e-llm-inference-service] [e2e-llm-inference-service] conn = None [e2e-llm-inference-service] [e2e-llm-inference-service] # Track whether `conn` needs to be released before [e2e-llm-inference-service] # returning/raising/recursing. Update this variable if necessary, and [e2e-llm-inference-service] # leave `release_conn` constant throughout the function. That way, if [e2e-llm-inference-service] # the function recurses, the original value of `release_conn` will be [e2e-llm-inference-service] # passed down into the recursive call, and its value will be respected. [e2e-llm-inference-service] # [e2e-llm-inference-service] # See issue #651 [1] for details. [e2e-llm-inference-service] # [e2e-llm-inference-service] # [1] [e2e-llm-inference-service] release_this_conn = release_conn [e2e-llm-inference-service] [e2e-llm-inference-service] http_tunnel_required = connection_requires_http_tunnel( [e2e-llm-inference-service] self.proxy, self.proxy_config, destination_scheme [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] # Merge the proxy headers. Only done when not using HTTP CONNECT. We [e2e-llm-inference-service] # have to copy the headers dict so we can safely change it without those [e2e-llm-inference-service] # changes being reflected in anyone else's copy. [e2e-llm-inference-service] if not http_tunnel_required: [e2e-llm-inference-service] headers = headers.copy() # type: ignore[attr-defined] [e2e-llm-inference-service] headers.update(self.proxy_headers) # type: ignore[union-attr] [e2e-llm-inference-service] [e2e-llm-inference-service] # Must keep the exception bound to a separate variable or else Python 3 [e2e-llm-inference-service] # complains about UnboundLocalError. [e2e-llm-inference-service] err = None [e2e-llm-inference-service] [e2e-llm-inference-service] # Keep track of whether we cleanly exited the except block. This [e2e-llm-inference-service] # ensures we do proper cleanup in finally. [e2e-llm-inference-service] clean_exit = False [e2e-llm-inference-service] [e2e-llm-inference-service] # Rewind body position, if needed. Record current position [e2e-llm-inference-service] # for future rewinds in the event of a redirect/retry. [e2e-llm-inference-service] body_pos = set_file_position(body, body_pos) [e2e-llm-inference-service] [e2e-llm-inference-service] try: [e2e-llm-inference-service] # Request a connection from the queue. [e2e-llm-inference-service] timeout_obj = self._get_timeout(timeout) [e2e-llm-inference-service] conn = self._get_conn(timeout=pool_timeout) [e2e-llm-inference-service] [e2e-llm-inference-service] conn.timeout = timeout_obj.connect_timeout # type: ignore[assignment] [e2e-llm-inference-service] [e2e-llm-inference-service] # Is this a closed/new connection that requires CONNECT tunnelling? [e2e-llm-inference-service] if self.proxy is not None and http_tunnel_required and conn.is_closed: [e2e-llm-inference-service] try: [e2e-llm-inference-service] self._prepare_proxy(conn) [e2e-llm-inference-service] except (BaseSSLError, OSError, SocketTimeout) as e: [e2e-llm-inference-service] self._raise_timeout( [e2e-llm-inference-service] err=e, url=self.proxy.url, timeout_value=conn.timeout [e2e-llm-inference-service] ) [e2e-llm-inference-service] raise [e2e-llm-inference-service] [e2e-llm-inference-service] # If we're going to release the connection in ``finally:``, then [e2e-llm-inference-service] # the response doesn't need to know about the connection. Otherwise [e2e-llm-inference-service] # it will also try to release it and we'll have a double-release [e2e-llm-inference-service] # mess. [e2e-llm-inference-service] response_conn = conn if not release_conn else None [e2e-llm-inference-service] [e2e-llm-inference-service] # Make the request on the HTTPConnection object [e2e-llm-inference-service] response = self._make_request( [e2e-llm-inference-service] conn, [e2e-llm-inference-service] method, [e2e-llm-inference-service] url, [e2e-llm-inference-service] timeout=timeout_obj, [e2e-llm-inference-service] body=body, [e2e-llm-inference-service] headers=headers, [e2e-llm-inference-service] chunked=chunked, [e2e-llm-inference-service] retries=retries, [e2e-llm-inference-service] response_conn=response_conn, [e2e-llm-inference-service] preload_content=preload_content, [e2e-llm-inference-service] decode_content=decode_content, [e2e-llm-inference-service] **response_kw, [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] # Everything went great! [e2e-llm-inference-service] clean_exit = True [e2e-llm-inference-service] [e2e-llm-inference-service] except EmptyPoolError: [e2e-llm-inference-service] # Didn't get a connection from the pool, no need to clean up [e2e-llm-inference-service] clean_exit = True [e2e-llm-inference-service] release_this_conn = False [e2e-llm-inference-service] raise [e2e-llm-inference-service] [e2e-llm-inference-service] except ( [e2e-llm-inference-service] TimeoutError, [e2e-llm-inference-service] HTTPException, [e2e-llm-inference-service] OSError, [e2e-llm-inference-service] ProtocolError, [e2e-llm-inference-service] BaseSSLError, [e2e-llm-inference-service] SSLError, [e2e-llm-inference-service] CertificateError, [e2e-llm-inference-service] ProxyError, [e2e-llm-inference-service] ) as e: [e2e-llm-inference-service] # Discard the connection for these exceptions. It will be [e2e-llm-inference-service] # replaced during the next _get_conn() call. [e2e-llm-inference-service] clean_exit = False [e2e-llm-inference-service] new_e: Exception = e [e2e-llm-inference-service] if isinstance(e, (BaseSSLError, CertificateError)): [e2e-llm-inference-service] new_e = SSLError(e) [e2e-llm-inference-service] if isinstance( [e2e-llm-inference-service] new_e, [e2e-llm-inference-service] ( [e2e-llm-inference-service] OSError, [e2e-llm-inference-service] NewConnectionError, [e2e-llm-inference-service] TimeoutError, [e2e-llm-inference-service] SSLError, [e2e-llm-inference-service] HTTPException, [e2e-llm-inference-service] ), [e2e-llm-inference-service] ) and (conn and conn.proxy and not conn.has_connected_to_proxy): [e2e-llm-inference-service] new_e = _wrap_proxy_error(new_e, conn.proxy.scheme) [e2e-llm-inference-service] elif isinstance(new_e, (OSError, HTTPException)): [e2e-llm-inference-service] new_e = ProtocolError("Connection aborted.", new_e) [e2e-llm-inference-service] [e2e-llm-inference-service] retries = retries.increment( [e2e-llm-inference-service] method, url, error=new_e, _pool=self, _stacktrace=sys.exc_info()[2] [e2e-llm-inference-service] ) [e2e-llm-inference-service] retries.sleep() [e2e-llm-inference-service] [e2e-llm-inference-service] # Keep track of the error for the retry warning. [e2e-llm-inference-service] err = e [e2e-llm-inference-service] [e2e-llm-inference-service] finally: [e2e-llm-inference-service] if not clean_exit: [e2e-llm-inference-service] # We hit some kind of exception, handled or otherwise. We need [e2e-llm-inference-service] # to throw the connection away unless explicitly told not to. [e2e-llm-inference-service] # Close the connection, set the variable to None, and make sure [e2e-llm-inference-service] # we put the None back in the pool to avoid leaking it. [e2e-llm-inference-service] if conn: [e2e-llm-inference-service] conn.close() [e2e-llm-inference-service] conn = None [e2e-llm-inference-service] release_this_conn = True [e2e-llm-inference-service] [e2e-llm-inference-service] if release_this_conn: [e2e-llm-inference-service] # Put the connection back to be reused. If the connection is [e2e-llm-inference-service] # expired then it will be None, which will get replaced with a [e2e-llm-inference-service] # fresh connection during _get_conn. [e2e-llm-inference-service] self._put_conn(conn) [e2e-llm-inference-service] [e2e-llm-inference-service] if not conn: [e2e-llm-inference-service] # Try again [e2e-llm-inference-service] log.warning( [e2e-llm-inference-service] "Retrying (%r) after connection broken by '%r': %s", retries, err, url [e2e-llm-inference-service] ) [e2e-llm-inference-service] > return self.urlopen( [e2e-llm-inference-service] method, [e2e-llm-inference-service] url, [e2e-llm-inference-service] body, [e2e-llm-inference-service] headers, [e2e-llm-inference-service] retries, [e2e-llm-inference-service] redirect, [e2e-llm-inference-service] assert_same_host, [e2e-llm-inference-service] timeout=timeout, [e2e-llm-inference-service] pool_timeout=pool_timeout, [e2e-llm-inference-service] release_conn=release_conn, [e2e-llm-inference-service] chunked=chunked, [e2e-llm-inference-service] body_pos=body_pos, [e2e-llm-inference-service] preload_content=preload_content, [e2e-llm-inference-service] decode_content=decode_content, [e2e-llm-inference-service] **response_kw, [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] ../../python/kserve/.venv/lib64/python3.11/site-packages/urllib3/connectionpool.py:871: [e2e-llm-inference-service] _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ [e2e-llm-inference-service] [e2e-llm-inference-service] self = [e2e-llm-inference-service] method = 'POST', url = '/kserve-ci-e2e-test/rolling-upgrade-test/v1/completions' [e2e-llm-inference-service] body = b'{"model": "facebook/opt-125m", "prompt": "KServe is a", "max_tokens": 20}' [e2e-llm-inference-service] headers = {'User-Agent': 'python-requests/2.32.3', 'Accept-Encoding': 'gzip, deflate', 'Accept': '*/*', 'Connection': 'keep-alive', 'Content-Type': 'application/json', 'Content-Length': '73'} [e2e-llm-inference-service] retries = Retry(total=0, connect=None, read=None, redirect=None, status=None) [e2e-llm-inference-service] redirect = False, assert_same_host = False [e2e-llm-inference-service] timeout = Timeout(connect=60, read=60, total=None), pool_timeout = None [e2e-llm-inference-service] release_conn = False, chunked = False, body_pos = None, preload_content = False [e2e-llm-inference-service] decode_content = False, response_kw = {} [e2e-llm-inference-service] parsed_url = Url(scheme=None, auth=None, host=None, port=None, path='/kserve-ci-e2e-test/rolling-upgrade-test/v1/completions', query=None, fragment=None) [e2e-llm-inference-service] destination_scheme = None, conn = None, release_this_conn = True [e2e-llm-inference-service] http_tunnel_required = False [e2e-llm-inference-service] err = ReadTimeoutError("HTTPConnectionPool(host='ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com', port=80): Read timed out. (read timeout=60)") [e2e-llm-inference-service] clean_exit = False [e2e-llm-inference-service] [e2e-llm-inference-service] def urlopen( # type: ignore[override] [e2e-llm-inference-service] self, [e2e-llm-inference-service] method: str, [e2e-llm-inference-service] url: str, [e2e-llm-inference-service] body: _TYPE_BODY | None = None, [e2e-llm-inference-service] headers: typing.Mapping[str, str] | None = None, [e2e-llm-inference-service] retries: Retry | bool | int | None = None, [e2e-llm-inference-service] redirect: bool = True, [e2e-llm-inference-service] assert_same_host: bool = True, [e2e-llm-inference-service] timeout: _TYPE_TIMEOUT = _DEFAULT_TIMEOUT, [e2e-llm-inference-service] pool_timeout: int | None = None, [e2e-llm-inference-service] release_conn: bool | None = None, [e2e-llm-inference-service] chunked: bool = False, [e2e-llm-inference-service] body_pos: _TYPE_BODY_POSITION | None = None, [e2e-llm-inference-service] preload_content: bool = True, [e2e-llm-inference-service] decode_content: bool = True, [e2e-llm-inference-service] **response_kw: typing.Any, [e2e-llm-inference-service] ) -> BaseHTTPResponse: [e2e-llm-inference-service] """ [e2e-llm-inference-service] Get a connection from the pool and perform an HTTP request. This is the [e2e-llm-inference-service] lowest level call for making a request, so you'll need to specify all [e2e-llm-inference-service] the raw details. [e2e-llm-inference-service] [e2e-llm-inference-service] .. note:: [e2e-llm-inference-service] [e2e-llm-inference-service] More commonly, it's appropriate to use a convenience method [e2e-llm-inference-service] such as :meth:`request`. [e2e-llm-inference-service] [e2e-llm-inference-service] .. note:: [e2e-llm-inference-service] [e2e-llm-inference-service] `release_conn` will only behave as expected if [e2e-llm-inference-service] `preload_content=False` because we want to make [e2e-llm-inference-service] `preload_content=False` the default behaviour someday soon without [e2e-llm-inference-service] breaking backwards compatibility. [e2e-llm-inference-service] [e2e-llm-inference-service] :param method: [e2e-llm-inference-service] HTTP request method (such as GET, POST, PUT, etc.) [e2e-llm-inference-service] [e2e-llm-inference-service] :param url: [e2e-llm-inference-service] The URL to perform the request on. [e2e-llm-inference-service] [e2e-llm-inference-service] :param body: [e2e-llm-inference-service] Data to send in the request body, either :class:`str`, :class:`bytes`, [e2e-llm-inference-service] an iterable of :class:`str`/:class:`bytes`, or a file-like object. [e2e-llm-inference-service] [e2e-llm-inference-service] :param headers: [e2e-llm-inference-service] Dictionary of custom headers to send, such as User-Agent, [e2e-llm-inference-service] If-None-Match, etc. If None, pool headers are used. If provided, [e2e-llm-inference-service] these headers completely replace any pool-specific headers. [e2e-llm-inference-service] [e2e-llm-inference-service] :param retries: [e2e-llm-inference-service] Configure the number of retries to allow before raising a [e2e-llm-inference-service] :class:`~urllib3.exceptions.MaxRetryError` exception. [e2e-llm-inference-service] [e2e-llm-inference-service] If ``None`` (default) will retry 3 times, see ``Retry.DEFAULT``. Pass a [e2e-llm-inference-service] :class:`~urllib3.util.retry.Retry` object for fine-grained control [e2e-llm-inference-service] over different types of retries. [e2e-llm-inference-service] Pass an integer number to retry connection errors that many times, [e2e-llm-inference-service] but no other types of errors. Pass zero to never retry. [e2e-llm-inference-service] [e2e-llm-inference-service] If ``False``, then retries are disabled and any exception is raised [e2e-llm-inference-service] immediately. Also, instead of raising a MaxRetryError on redirects, [e2e-llm-inference-service] the redirect response will be returned. [e2e-llm-inference-service] [e2e-llm-inference-service] :type retries: :class:`~urllib3.util.retry.Retry`, False, or an int. [e2e-llm-inference-service] [e2e-llm-inference-service] :param redirect: [e2e-llm-inference-service] If True, automatically handle redirects (status codes 301, 302, [e2e-llm-inference-service] 303, 307, 308). Each redirect counts as a retry. Disabling retries [e2e-llm-inference-service] will disable redirect, too. [e2e-llm-inference-service] [e2e-llm-inference-service] :param assert_same_host: [e2e-llm-inference-service] If ``True``, will make sure that the host of the pool requests is [e2e-llm-inference-service] consistent else will raise HostChangedError. When ``False``, you can [e2e-llm-inference-service] use the pool on an HTTP proxy and request foreign hosts. [e2e-llm-inference-service] [e2e-llm-inference-service] :param timeout: [e2e-llm-inference-service] If specified, overrides the default timeout for this one [e2e-llm-inference-service] request. It may be a float (in seconds) or an instance of [e2e-llm-inference-service] :class:`urllib3.util.Timeout`. [e2e-llm-inference-service] [e2e-llm-inference-service] :param pool_timeout: [e2e-llm-inference-service] If set and the pool is set to block=True, then this method will [e2e-llm-inference-service] block for ``pool_timeout`` seconds and raise EmptyPoolError if no [e2e-llm-inference-service] connection is available within the time period. [e2e-llm-inference-service] [e2e-llm-inference-service] :param bool preload_content: [e2e-llm-inference-service] If True, the response's body will be preloaded into memory. [e2e-llm-inference-service] [e2e-llm-inference-service] :param bool decode_content: [e2e-llm-inference-service] If True, will attempt to decode the body based on the [e2e-llm-inference-service] 'content-encoding' header. [e2e-llm-inference-service] [e2e-llm-inference-service] :param release_conn: [e2e-llm-inference-service] If False, then the urlopen call will not release the connection [e2e-llm-inference-service] back into the pool once a response is received (but will release if [e2e-llm-inference-service] you read the entire contents of the response such as when [e2e-llm-inference-service] `preload_content=True`). This is useful if you're not preloading [e2e-llm-inference-service] the response's content immediately. You will need to call [e2e-llm-inference-service] ``r.release_conn()`` on the response ``r`` to return the connection [e2e-llm-inference-service] back into the pool. If None, it takes the value of ``preload_content`` [e2e-llm-inference-service] which defaults to ``True``. [e2e-llm-inference-service] [e2e-llm-inference-service] :param bool chunked: [e2e-llm-inference-service] If True, urllib3 will send the body using chunked transfer [e2e-llm-inference-service] encoding. Otherwise, urllib3 will send the body using the standard [e2e-llm-inference-service] content-length form. Defaults to False. [e2e-llm-inference-service] [e2e-llm-inference-service] :param int body_pos: [e2e-llm-inference-service] Position to seek to in file-like body in the event of a retry or [e2e-llm-inference-service] redirect. Typically this won't need to be set because urllib3 will [e2e-llm-inference-service] auto-populate the value when needed. [e2e-llm-inference-service] """ [e2e-llm-inference-service] parsed_url = parse_url(url) [e2e-llm-inference-service] destination_scheme = parsed_url.scheme [e2e-llm-inference-service] [e2e-llm-inference-service] if headers is None: [e2e-llm-inference-service] headers = self.headers [e2e-llm-inference-service] [e2e-llm-inference-service] if not isinstance(retries, Retry): [e2e-llm-inference-service] retries = Retry.from_int(retries, redirect=redirect, default=self.retries) [e2e-llm-inference-service] [e2e-llm-inference-service] if release_conn is None: [e2e-llm-inference-service] release_conn = preload_content [e2e-llm-inference-service] [e2e-llm-inference-service] # Check host [e2e-llm-inference-service] if assert_same_host and not self.is_same_host(url): [e2e-llm-inference-service] raise HostChangedError(self, url, retries) [e2e-llm-inference-service] [e2e-llm-inference-service] # Ensure that the URL we're connecting to is properly encoded [e2e-llm-inference-service] if url.startswith("/"): [e2e-llm-inference-service] url = to_str(_encode_target(url)) [e2e-llm-inference-service] else: [e2e-llm-inference-service] url = to_str(parsed_url.url) [e2e-llm-inference-service] [e2e-llm-inference-service] conn = None [e2e-llm-inference-service] [e2e-llm-inference-service] # Track whether `conn` needs to be released before [e2e-llm-inference-service] # returning/raising/recursing. Update this variable if necessary, and [e2e-llm-inference-service] # leave `release_conn` constant throughout the function. That way, if [e2e-llm-inference-service] # the function recurses, the original value of `release_conn` will be [e2e-llm-inference-service] # passed down into the recursive call, and its value will be respected. [e2e-llm-inference-service] # [e2e-llm-inference-service] # See issue #651 [1] for details. [e2e-llm-inference-service] # [e2e-llm-inference-service] # [1] [e2e-llm-inference-service] release_this_conn = release_conn [e2e-llm-inference-service] [e2e-llm-inference-service] http_tunnel_required = connection_requires_http_tunnel( [e2e-llm-inference-service] self.proxy, self.proxy_config, destination_scheme [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] # Merge the proxy headers. Only done when not using HTTP CONNECT. We [e2e-llm-inference-service] # have to copy the headers dict so we can safely change it without those [e2e-llm-inference-service] # changes being reflected in anyone else's copy. [e2e-llm-inference-service] if not http_tunnel_required: [e2e-llm-inference-service] headers = headers.copy() # type: ignore[attr-defined] [e2e-llm-inference-service] headers.update(self.proxy_headers) # type: ignore[union-attr] [e2e-llm-inference-service] [e2e-llm-inference-service] # Must keep the exception bound to a separate variable or else Python 3 [e2e-llm-inference-service] # complains about UnboundLocalError. [e2e-llm-inference-service] err = None [e2e-llm-inference-service] [e2e-llm-inference-service] # Keep track of whether we cleanly exited the except block. This [e2e-llm-inference-service] # ensures we do proper cleanup in finally. [e2e-llm-inference-service] clean_exit = False [e2e-llm-inference-service] [e2e-llm-inference-service] # Rewind body position, if needed. Record current position [e2e-llm-inference-service] # for future rewinds in the event of a redirect/retry. [e2e-llm-inference-service] body_pos = set_file_position(body, body_pos) [e2e-llm-inference-service] [e2e-llm-inference-service] try: [e2e-llm-inference-service] # Request a connection from the queue. [e2e-llm-inference-service] timeout_obj = self._get_timeout(timeout) [e2e-llm-inference-service] conn = self._get_conn(timeout=pool_timeout) [e2e-llm-inference-service] [e2e-llm-inference-service] conn.timeout = timeout_obj.connect_timeout # type: ignore[assignment] [e2e-llm-inference-service] [e2e-llm-inference-service] # Is this a closed/new connection that requires CONNECT tunnelling? [e2e-llm-inference-service] if self.proxy is not None and http_tunnel_required and conn.is_closed: [e2e-llm-inference-service] try: [e2e-llm-inference-service] self._prepare_proxy(conn) [e2e-llm-inference-service] except (BaseSSLError, OSError, SocketTimeout) as e: [e2e-llm-inference-service] self._raise_timeout( [e2e-llm-inference-service] err=e, url=self.proxy.url, timeout_value=conn.timeout [e2e-llm-inference-service] ) [e2e-llm-inference-service] raise [e2e-llm-inference-service] [e2e-llm-inference-service] # If we're going to release the connection in ``finally:``, then [e2e-llm-inference-service] # the response doesn't need to know about the connection. Otherwise [e2e-llm-inference-service] # it will also try to release it and we'll have a double-release [e2e-llm-inference-service] # mess. [e2e-llm-inference-service] response_conn = conn if not release_conn else None [e2e-llm-inference-service] [e2e-llm-inference-service] # Make the request on the HTTPConnection object [e2e-llm-inference-service] response = self._make_request( [e2e-llm-inference-service] conn, [e2e-llm-inference-service] method, [e2e-llm-inference-service] url, [e2e-llm-inference-service] timeout=timeout_obj, [e2e-llm-inference-service] body=body, [e2e-llm-inference-service] headers=headers, [e2e-llm-inference-service] chunked=chunked, [e2e-llm-inference-service] retries=retries, [e2e-llm-inference-service] response_conn=response_conn, [e2e-llm-inference-service] preload_content=preload_content, [e2e-llm-inference-service] decode_content=decode_content, [e2e-llm-inference-service] **response_kw, [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] # Everything went great! [e2e-llm-inference-service] clean_exit = True [e2e-llm-inference-service] [e2e-llm-inference-service] except EmptyPoolError: [e2e-llm-inference-service] # Didn't get a connection from the pool, no need to clean up [e2e-llm-inference-service] clean_exit = True [e2e-llm-inference-service] release_this_conn = False [e2e-llm-inference-service] raise [e2e-llm-inference-service] [e2e-llm-inference-service] except ( [e2e-llm-inference-service] TimeoutError, [e2e-llm-inference-service] HTTPException, [e2e-llm-inference-service] OSError, [e2e-llm-inference-service] ProtocolError, [e2e-llm-inference-service] BaseSSLError, [e2e-llm-inference-service] SSLError, [e2e-llm-inference-service] CertificateError, [e2e-llm-inference-service] ProxyError, [e2e-llm-inference-service] ) as e: [e2e-llm-inference-service] # Discard the connection for these exceptions. It will be [e2e-llm-inference-service] # replaced during the next _get_conn() call. [e2e-llm-inference-service] clean_exit = False [e2e-llm-inference-service] new_e: Exception = e [e2e-llm-inference-service] if isinstance(e, (BaseSSLError, CertificateError)): [e2e-llm-inference-service] new_e = SSLError(e) [e2e-llm-inference-service] if isinstance( [e2e-llm-inference-service] new_e, [e2e-llm-inference-service] ( [e2e-llm-inference-service] OSError, [e2e-llm-inference-service] NewConnectionError, [e2e-llm-inference-service] TimeoutError, [e2e-llm-inference-service] SSLError, [e2e-llm-inference-service] HTTPException, [e2e-llm-inference-service] ), [e2e-llm-inference-service] ) and (conn and conn.proxy and not conn.has_connected_to_proxy): [e2e-llm-inference-service] new_e = _wrap_proxy_error(new_e, conn.proxy.scheme) [e2e-llm-inference-service] elif isinstance(new_e, (OSError, HTTPException)): [e2e-llm-inference-service] new_e = ProtocolError("Connection aborted.", new_e) [e2e-llm-inference-service] [e2e-llm-inference-service] retries = retries.increment( [e2e-llm-inference-service] method, url, error=new_e, _pool=self, _stacktrace=sys.exc_info()[2] [e2e-llm-inference-service] ) [e2e-llm-inference-service] retries.sleep() [e2e-llm-inference-service] [e2e-llm-inference-service] # Keep track of the error for the retry warning. [e2e-llm-inference-service] err = e [e2e-llm-inference-service] [e2e-llm-inference-service] finally: [e2e-llm-inference-service] if not clean_exit: [e2e-llm-inference-service] # We hit some kind of exception, handled or otherwise. We need [e2e-llm-inference-service] # to throw the connection away unless explicitly told not to. [e2e-llm-inference-service] # Close the connection, set the variable to None, and make sure [e2e-llm-inference-service] # we put the None back in the pool to avoid leaking it. [e2e-llm-inference-service] if conn: [e2e-llm-inference-service] conn.close() [e2e-llm-inference-service] conn = None [e2e-llm-inference-service] release_this_conn = True [e2e-llm-inference-service] [e2e-llm-inference-service] if release_this_conn: [e2e-llm-inference-service] # Put the connection back to be reused. If the connection is [e2e-llm-inference-service] # expired then it will be None, which will get replaced with a [e2e-llm-inference-service] # fresh connection during _get_conn. [e2e-llm-inference-service] self._put_conn(conn) [e2e-llm-inference-service] [e2e-llm-inference-service] if not conn: [e2e-llm-inference-service] # Try again [e2e-llm-inference-service] log.warning( [e2e-llm-inference-service] "Retrying (%r) after connection broken by '%r': %s", retries, err, url [e2e-llm-inference-service] ) [e2e-llm-inference-service] > return self.urlopen( [e2e-llm-inference-service] method, [e2e-llm-inference-service] url, [e2e-llm-inference-service] body, [e2e-llm-inference-service] headers, [e2e-llm-inference-service] retries, [e2e-llm-inference-service] redirect, [e2e-llm-inference-service] assert_same_host, [e2e-llm-inference-service] timeout=timeout, [e2e-llm-inference-service] pool_timeout=pool_timeout, [e2e-llm-inference-service] release_conn=release_conn, [e2e-llm-inference-service] chunked=chunked, [e2e-llm-inference-service] body_pos=body_pos, [e2e-llm-inference-service] preload_content=preload_content, [e2e-llm-inference-service] decode_content=decode_content, [e2e-llm-inference-service] **response_kw, [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] ../../python/kserve/.venv/lib64/python3.11/site-packages/urllib3/connectionpool.py:871: [e2e-llm-inference-service] _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ [e2e-llm-inference-service] [e2e-llm-inference-service] self = [e2e-llm-inference-service] method = 'POST', url = '/kserve-ci-e2e-test/rolling-upgrade-test/v1/completions' [e2e-llm-inference-service] body = b'{"model": "facebook/opt-125m", "prompt": "KServe is a", "max_tokens": 20}' [e2e-llm-inference-service] headers = {'User-Agent': 'python-requests/2.32.3', 'Accept-Encoding': 'gzip, deflate', 'Accept': '*/*', 'Connection': 'keep-alive', 'Content-Type': 'application/json', 'Content-Length': '73'} [e2e-llm-inference-service] retries = Retry(total=0, connect=None, read=None, redirect=None, status=None) [e2e-llm-inference-service] redirect = False, assert_same_host = False [e2e-llm-inference-service] timeout = Timeout(connect=60, read=60, total=None), pool_timeout = None [e2e-llm-inference-service] release_conn = False, chunked = False, body_pos = None, preload_content = False [e2e-llm-inference-service] decode_content = False, response_kw = {} [e2e-llm-inference-service] parsed_url = Url(scheme=None, auth=None, host=None, port=None, path='/kserve-ci-e2e-test/rolling-upgrade-test/v1/completions', query=None, fragment=None) [e2e-llm-inference-service] destination_scheme = None, conn = None, release_this_conn = True [e2e-llm-inference-service] http_tunnel_required = False, err = None, clean_exit = False [e2e-llm-inference-service] [e2e-llm-inference-service] def urlopen( # type: ignore[override] [e2e-llm-inference-service] self, [e2e-llm-inference-service] method: str, [e2e-llm-inference-service] url: str, [e2e-llm-inference-service] body: _TYPE_BODY | None = None, [e2e-llm-inference-service] headers: typing.Mapping[str, str] | None = None, [e2e-llm-inference-service] retries: Retry | bool | int | None = None, [e2e-llm-inference-service] redirect: bool = True, [e2e-llm-inference-service] assert_same_host: bool = True, [e2e-llm-inference-service] timeout: _TYPE_TIMEOUT = _DEFAULT_TIMEOUT, [e2e-llm-inference-service] pool_timeout: int | None = None, [e2e-llm-inference-service] release_conn: bool | None = None, [e2e-llm-inference-service] chunked: bool = False, [e2e-llm-inference-service] body_pos: _TYPE_BODY_POSITION | None = None, [e2e-llm-inference-service] preload_content: bool = True, [e2e-llm-inference-service] decode_content: bool = True, [e2e-llm-inference-service] **response_kw: typing.Any, [e2e-llm-inference-service] ) -> BaseHTTPResponse: [e2e-llm-inference-service] """ [e2e-llm-inference-service] Get a connection from the pool and perform an HTTP request. This is the [e2e-llm-inference-service] lowest level call for making a request, so you'll need to specify all [e2e-llm-inference-service] the raw details. [e2e-llm-inference-service] [e2e-llm-inference-service] .. note:: [e2e-llm-inference-service] [e2e-llm-inference-service] More commonly, it's appropriate to use a convenience method [e2e-llm-inference-service] such as :meth:`request`. [e2e-llm-inference-service] [e2e-llm-inference-service] .. note:: [e2e-llm-inference-service] [e2e-llm-inference-service] `release_conn` will only behave as expected if [e2e-llm-inference-service] `preload_content=False` because we want to make [e2e-llm-inference-service] `preload_content=False` the default behaviour someday soon without [e2e-llm-inference-service] breaking backwards compatibility. [e2e-llm-inference-service] [e2e-llm-inference-service] :param method: [e2e-llm-inference-service] HTTP request method (such as GET, POST, PUT, etc.) [e2e-llm-inference-service] [e2e-llm-inference-service] :param url: [e2e-llm-inference-service] The URL to perform the request on. [e2e-llm-inference-service] [e2e-llm-inference-service] :param body: [e2e-llm-inference-service] Data to send in the request body, either :class:`str`, :class:`bytes`, [e2e-llm-inference-service] an iterable of :class:`str`/:class:`bytes`, or a file-like object. [e2e-llm-inference-service] [e2e-llm-inference-service] :param headers: [e2e-llm-inference-service] Dictionary of custom headers to send, such as User-Agent, [e2e-llm-inference-service] If-None-Match, etc. If None, pool headers are used. If provided, [e2e-llm-inference-service] these headers completely replace any pool-specific headers. [e2e-llm-inference-service] [e2e-llm-inference-service] :param retries: [e2e-llm-inference-service] Configure the number of retries to allow before raising a [e2e-llm-inference-service] :class:`~urllib3.exceptions.MaxRetryError` exception. [e2e-llm-inference-service] [e2e-llm-inference-service] If ``None`` (default) will retry 3 times, see ``Retry.DEFAULT``. Pass a [e2e-llm-inference-service] :class:`~urllib3.util.retry.Retry` object for fine-grained control [e2e-llm-inference-service] over different types of retries. [e2e-llm-inference-service] Pass an integer number to retry connection errors that many times, [e2e-llm-inference-service] but no other types of errors. Pass zero to never retry. [e2e-llm-inference-service] [e2e-llm-inference-service] If ``False``, then retries are disabled and any exception is raised [e2e-llm-inference-service] immediately. Also, instead of raising a MaxRetryError on redirects, [e2e-llm-inference-service] the redirect response will be returned. [e2e-llm-inference-service] [e2e-llm-inference-service] :type retries: :class:`~urllib3.util.retry.Retry`, False, or an int. [e2e-llm-inference-service] [e2e-llm-inference-service] :param redirect: [e2e-llm-inference-service] If True, automatically handle redirects (status codes 301, 302, [e2e-llm-inference-service] 303, 307, 308). Each redirect counts as a retry. Disabling retries [e2e-llm-inference-service] will disable redirect, too. [e2e-llm-inference-service] [e2e-llm-inference-service] :param assert_same_host: [e2e-llm-inference-service] If ``True``, will make sure that the host of the pool requests is [e2e-llm-inference-service] consistent else will raise HostChangedError. When ``False``, you can [e2e-llm-inference-service] use the pool on an HTTP proxy and request foreign hosts. [e2e-llm-inference-service] [e2e-llm-inference-service] :param timeout: [e2e-llm-inference-service] If specified, overrides the default timeout for this one [e2e-llm-inference-service] request. It may be a float (in seconds) or an instance of [e2e-llm-inference-service] :class:`urllib3.util.Timeout`. [e2e-llm-inference-service] [e2e-llm-inference-service] :param pool_timeout: [e2e-llm-inference-service] If set and the pool is set to block=True, then this method will [e2e-llm-inference-service] block for ``pool_timeout`` seconds and raise EmptyPoolError if no [e2e-llm-inference-service] connection is available within the time period. [e2e-llm-inference-service] [e2e-llm-inference-service] :param bool preload_content: [e2e-llm-inference-service] If True, the response's body will be preloaded into memory. [e2e-llm-inference-service] [e2e-llm-inference-service] :param bool decode_content: [e2e-llm-inference-service] If True, will attempt to decode the body based on the [e2e-llm-inference-service] 'content-encoding' header. [e2e-llm-inference-service] [e2e-llm-inference-service] :param release_conn: [e2e-llm-inference-service] If False, then the urlopen call will not release the connection [e2e-llm-inference-service] back into the pool once a response is received (but will release if [e2e-llm-inference-service] you read the entire contents of the response such as when [e2e-llm-inference-service] `preload_content=True`). This is useful if you're not preloading [e2e-llm-inference-service] the response's content immediately. You will need to call [e2e-llm-inference-service] ``r.release_conn()`` on the response ``r`` to return the connection [e2e-llm-inference-service] back into the pool. If None, it takes the value of ``preload_content`` [e2e-llm-inference-service] which defaults to ``True``. [e2e-llm-inference-service] [e2e-llm-inference-service] :param bool chunked: [e2e-llm-inference-service] If True, urllib3 will send the body using chunked transfer [e2e-llm-inference-service] encoding. Otherwise, urllib3 will send the body using the standard [e2e-llm-inference-service] content-length form. Defaults to False. [e2e-llm-inference-service] [e2e-llm-inference-service] :param int body_pos: [e2e-llm-inference-service] Position to seek to in file-like body in the event of a retry or [e2e-llm-inference-service] redirect. Typically this won't need to be set because urllib3 will [e2e-llm-inference-service] auto-populate the value when needed. [e2e-llm-inference-service] """ [e2e-llm-inference-service] parsed_url = parse_url(url) [e2e-llm-inference-service] destination_scheme = parsed_url.scheme [e2e-llm-inference-service] [e2e-llm-inference-service] if headers is None: [e2e-llm-inference-service] headers = self.headers [e2e-llm-inference-service] [e2e-llm-inference-service] if not isinstance(retries, Retry): [e2e-llm-inference-service] retries = Retry.from_int(retries, redirect=redirect, default=self.retries) [e2e-llm-inference-service] [e2e-llm-inference-service] if release_conn is None: [e2e-llm-inference-service] release_conn = preload_content [e2e-llm-inference-service] [e2e-llm-inference-service] # Check host [e2e-llm-inference-service] if assert_same_host and not self.is_same_host(url): [e2e-llm-inference-service] raise HostChangedError(self, url, retries) [e2e-llm-inference-service] [e2e-llm-inference-service] # Ensure that the URL we're connecting to is properly encoded [e2e-llm-inference-service] if url.startswith("/"): [e2e-llm-inference-service] url = to_str(_encode_target(url)) [e2e-llm-inference-service] else: [e2e-llm-inference-service] url = to_str(parsed_url.url) [e2e-llm-inference-service] [e2e-llm-inference-service] conn = None [e2e-llm-inference-service] [e2e-llm-inference-service] # Track whether `conn` needs to be released before [e2e-llm-inference-service] # returning/raising/recursing. Update this variable if necessary, and [e2e-llm-inference-service] # leave `release_conn` constant throughout the function. That way, if [e2e-llm-inference-service] # the function recurses, the original value of `release_conn` will be [e2e-llm-inference-service] # passed down into the recursive call, and its value will be respected. [e2e-llm-inference-service] # [e2e-llm-inference-service] # See issue #651 [1] for details. [e2e-llm-inference-service] # [e2e-llm-inference-service] # [1] [e2e-llm-inference-service] release_this_conn = release_conn [e2e-llm-inference-service] [e2e-llm-inference-service] http_tunnel_required = connection_requires_http_tunnel( [e2e-llm-inference-service] self.proxy, self.proxy_config, destination_scheme [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] # Merge the proxy headers. Only done when not using HTTP CONNECT. We [e2e-llm-inference-service] # have to copy the headers dict so we can safely change it without those [e2e-llm-inference-service] # changes being reflected in anyone else's copy. [e2e-llm-inference-service] if not http_tunnel_required: [e2e-llm-inference-service] headers = headers.copy() # type: ignore[attr-defined] [e2e-llm-inference-service] headers.update(self.proxy_headers) # type: ignore[union-attr] [e2e-llm-inference-service] [e2e-llm-inference-service] # Must keep the exception bound to a separate variable or else Python 3 [e2e-llm-inference-service] # complains about UnboundLocalError. [e2e-llm-inference-service] err = None [e2e-llm-inference-service] [e2e-llm-inference-service] # Keep track of whether we cleanly exited the except block. This [e2e-llm-inference-service] # ensures we do proper cleanup in finally. [e2e-llm-inference-service] clean_exit = False [e2e-llm-inference-service] [e2e-llm-inference-service] # Rewind body position, if needed. Record current position [e2e-llm-inference-service] # for future rewinds in the event of a redirect/retry. [e2e-llm-inference-service] body_pos = set_file_position(body, body_pos) [e2e-llm-inference-service] [e2e-llm-inference-service] try: [e2e-llm-inference-service] # Request a connection from the queue. [e2e-llm-inference-service] timeout_obj = self._get_timeout(timeout) [e2e-llm-inference-service] conn = self._get_conn(timeout=pool_timeout) [e2e-llm-inference-service] [e2e-llm-inference-service] conn.timeout = timeout_obj.connect_timeout # type: ignore[assignment] [e2e-llm-inference-service] [e2e-llm-inference-service] # Is this a closed/new connection that requires CONNECT tunnelling? [e2e-llm-inference-service] if self.proxy is not None and http_tunnel_required and conn.is_closed: [e2e-llm-inference-service] try: [e2e-llm-inference-service] self._prepare_proxy(conn) [e2e-llm-inference-service] except (BaseSSLError, OSError, SocketTimeout) as e: [e2e-llm-inference-service] self._raise_timeout( [e2e-llm-inference-service] err=e, url=self.proxy.url, timeout_value=conn.timeout [e2e-llm-inference-service] ) [e2e-llm-inference-service] raise [e2e-llm-inference-service] [e2e-llm-inference-service] # If we're going to release the connection in ``finally:``, then [e2e-llm-inference-service] # the response doesn't need to know about the connection. Otherwise [e2e-llm-inference-service] # it will also try to release it and we'll have a double-release [e2e-llm-inference-service] # mess. [e2e-llm-inference-service] response_conn = conn if not release_conn else None [e2e-llm-inference-service] [e2e-llm-inference-service] # Make the request on the HTTPConnection object [e2e-llm-inference-service] response = self._make_request( [e2e-llm-inference-service] conn, [e2e-llm-inference-service] method, [e2e-llm-inference-service] url, [e2e-llm-inference-service] timeout=timeout_obj, [e2e-llm-inference-service] body=body, [e2e-llm-inference-service] headers=headers, [e2e-llm-inference-service] chunked=chunked, [e2e-llm-inference-service] retries=retries, [e2e-llm-inference-service] response_conn=response_conn, [e2e-llm-inference-service] preload_content=preload_content, [e2e-llm-inference-service] decode_content=decode_content, [e2e-llm-inference-service] **response_kw, [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] # Everything went great! [e2e-llm-inference-service] clean_exit = True [e2e-llm-inference-service] [e2e-llm-inference-service] except EmptyPoolError: [e2e-llm-inference-service] # Didn't get a connection from the pool, no need to clean up [e2e-llm-inference-service] clean_exit = True [e2e-llm-inference-service] release_this_conn = False [e2e-llm-inference-service] raise [e2e-llm-inference-service] [e2e-llm-inference-service] except ( [e2e-llm-inference-service] TimeoutError, [e2e-llm-inference-service] HTTPException, [e2e-llm-inference-service] OSError, [e2e-llm-inference-service] ProtocolError, [e2e-llm-inference-service] BaseSSLError, [e2e-llm-inference-service] SSLError, [e2e-llm-inference-service] CertificateError, [e2e-llm-inference-service] ProxyError, [e2e-llm-inference-service] ) as e: [e2e-llm-inference-service] # Discard the connection for these exceptions. It will be [e2e-llm-inference-service] # replaced during the next _get_conn() call. [e2e-llm-inference-service] clean_exit = False [e2e-llm-inference-service] new_e: Exception = e [e2e-llm-inference-service] if isinstance(e, (BaseSSLError, CertificateError)): [e2e-llm-inference-service] new_e = SSLError(e) [e2e-llm-inference-service] if isinstance( [e2e-llm-inference-service] new_e, [e2e-llm-inference-service] ( [e2e-llm-inference-service] OSError, [e2e-llm-inference-service] NewConnectionError, [e2e-llm-inference-service] TimeoutError, [e2e-llm-inference-service] SSLError, [e2e-llm-inference-service] HTTPException, [e2e-llm-inference-service] ), [e2e-llm-inference-service] ) and (conn and conn.proxy and not conn.has_connected_to_proxy): [e2e-llm-inference-service] new_e = _wrap_proxy_error(new_e, conn.proxy.scheme) [e2e-llm-inference-service] elif isinstance(new_e, (OSError, HTTPException)): [e2e-llm-inference-service] new_e = ProtocolError("Connection aborted.", new_e) [e2e-llm-inference-service] [e2e-llm-inference-service] > retries = retries.increment( [e2e-llm-inference-service] method, url, error=new_e, _pool=self, _stacktrace=sys.exc_info()[2] [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] ../../python/kserve/.venv/lib64/python3.11/site-packages/urllib3/connectionpool.py:841: [e2e-llm-inference-service] _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ [e2e-llm-inference-service] [e2e-llm-inference-service] self = Retry(total=0, connect=None, read=None, redirect=None, status=None) [e2e-llm-inference-service] method = 'POST', url = '/kserve-ci-e2e-test/rolling-upgrade-test/v1/completions' [e2e-llm-inference-service] response = None [e2e-llm-inference-service] error = ReadTimeoutError("HTTPConnectionPool(host='ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com', port=80): Read timed out. (read timeout=60)") [e2e-llm-inference-service] _pool = [e2e-llm-inference-service] _stacktrace = [e2e-llm-inference-service] [e2e-llm-inference-service] def increment( [e2e-llm-inference-service] self, [e2e-llm-inference-service] method: str | None = None, [e2e-llm-inference-service] url: str | None = None, [e2e-llm-inference-service] response: BaseHTTPResponse | None = None, [e2e-llm-inference-service] error: Exception | None = None, [e2e-llm-inference-service] _pool: ConnectionPool | None = None, [e2e-llm-inference-service] _stacktrace: TracebackType | None = None, [e2e-llm-inference-service] ) -> Self: [e2e-llm-inference-service] """Return a new Retry object with incremented retry counters. [e2e-llm-inference-service] [e2e-llm-inference-service] :param response: A response object, or None, if the server did not [e2e-llm-inference-service] return a response. [e2e-llm-inference-service] :type response: :class:`~urllib3.response.BaseHTTPResponse` [e2e-llm-inference-service] :param Exception error: An error encountered during the request, or [e2e-llm-inference-service] None if the response was received successfully. [e2e-llm-inference-service] [e2e-llm-inference-service] :return: A new ``Retry`` object. [e2e-llm-inference-service] """ [e2e-llm-inference-service] if self.total is False and error: [e2e-llm-inference-service] # Disabled, indicate to re-raise the error. [e2e-llm-inference-service] raise reraise(type(error), error, _stacktrace) [e2e-llm-inference-service] [e2e-llm-inference-service] total = self.total [e2e-llm-inference-service] if total is not None: [e2e-llm-inference-service] total -= 1 [e2e-llm-inference-service] [e2e-llm-inference-service] connect = self.connect [e2e-llm-inference-service] read = self.read [e2e-llm-inference-service] redirect = self.redirect [e2e-llm-inference-service] status_count = self.status [e2e-llm-inference-service] other = self.other [e2e-llm-inference-service] cause = "unknown" [e2e-llm-inference-service] status = None [e2e-llm-inference-service] redirect_location = None [e2e-llm-inference-service] [e2e-llm-inference-service] if error and self._is_connection_error(error): [e2e-llm-inference-service] # Connect retry? [e2e-llm-inference-service] if connect is False: [e2e-llm-inference-service] raise reraise(type(error), error, _stacktrace) [e2e-llm-inference-service] elif connect is not None: [e2e-llm-inference-service] connect -= 1 [e2e-llm-inference-service] [e2e-llm-inference-service] elif error and self._is_read_error(error): [e2e-llm-inference-service] # Read retry? [e2e-llm-inference-service] if read is False or method is None or not self._is_method_retryable(method): [e2e-llm-inference-service] raise reraise(type(error), error, _stacktrace) [e2e-llm-inference-service] elif read is not None: [e2e-llm-inference-service] read -= 1 [e2e-llm-inference-service] [e2e-llm-inference-service] elif error: [e2e-llm-inference-service] # Other retry? [e2e-llm-inference-service] if other is not None: [e2e-llm-inference-service] other -= 1 [e2e-llm-inference-service] [e2e-llm-inference-service] elif response and response.get_redirect_location(): [e2e-llm-inference-service] # Redirect retry? [e2e-llm-inference-service] if redirect is not None: [e2e-llm-inference-service] redirect -= 1 [e2e-llm-inference-service] cause = "too many redirects" [e2e-llm-inference-service] response_redirect_location = response.get_redirect_location() [e2e-llm-inference-service] if response_redirect_location: [e2e-llm-inference-service] redirect_location = response_redirect_location [e2e-llm-inference-service] status = response.status [e2e-llm-inference-service] [e2e-llm-inference-service] else: [e2e-llm-inference-service] # Incrementing because of a server error like a 500 in [e2e-llm-inference-service] # status_forcelist and the given method is in the allowed_methods [e2e-llm-inference-service] cause = ResponseError.GENERIC_ERROR [e2e-llm-inference-service] if response and response.status: [e2e-llm-inference-service] if status_count is not None: [e2e-llm-inference-service] status_count -= 1 [e2e-llm-inference-service] cause = ResponseError.SPECIFIC_ERROR.format(status_code=response.status) [e2e-llm-inference-service] status = response.status [e2e-llm-inference-service] [e2e-llm-inference-service] history = self.history + ( [e2e-llm-inference-service] RequestHistory(method, url, error, status, redirect_location), [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] new_retry = self.new( [e2e-llm-inference-service] total=total, [e2e-llm-inference-service] connect=connect, [e2e-llm-inference-service] read=read, [e2e-llm-inference-service] redirect=redirect, [e2e-llm-inference-service] status=status_count, [e2e-llm-inference-service] other=other, [e2e-llm-inference-service] history=history, [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] if new_retry.is_exhausted(): [e2e-llm-inference-service] reason = error or ResponseError(cause) [e2e-llm-inference-service] > raise MaxRetryError(_pool, url, reason) from reason # type: ignore[arg-type] [e2e-llm-inference-service] E urllib3.exceptions.MaxRetryError: HTTPConnectionPool(host='ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com', port=80): Max retries exceeded with url: /kserve-ci-e2e-test/rolling-upgrade-test/v1/completions (Caused by ReadTimeoutError("HTTPConnectionPool(host='ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com', port=80): Read timed out. (read timeout=60)")) [e2e-llm-inference-service] [e2e-llm-inference-service] ../../python/kserve/.venv/lib64/python3.11/site-packages/urllib3/util/retry.py:519: MaxRetryError [e2e-llm-inference-service] [e2e-llm-inference-service] During handling of the above exception, another exception occurred: [e2e-llm-inference-service] [e2e-llm-inference-service] def get_successful_response(): [e2e-llm-inference-service] try: [e2e-llm-inference-service] if test_case.url_getter: [e2e-llm-inference-service] service_url = test_case.url_getter(kserve_client, test_case.llm_service) [e2e-llm-inference-service] else: [e2e-llm-inference-service] service_url = get_llm_service_url(kserve_client, test_case.llm_service) [e2e-llm-inference-service] except Exception as e: [e2e-llm-inference-service] raise AssertionError(f"❌ Failed to get service URL: {e}") from e [e2e-llm-inference-service] [e2e-llm-inference-service] model_url = service_url + test_case.endpoint [e2e-llm-inference-service] [e2e-llm-inference-service] headers = {"Content-Type": "application/json"} [e2e-llm-inference-service] if extra_headers: [e2e-llm-inference-service] headers.update(extra_headers) [e2e-llm-inference-service] [e2e-llm-inference-service] if test_case.payload_formatter is not None: [e2e-llm-inference-service] test_payload = test_case.payload_formatter(test_case) [e2e-llm-inference-service] elif test_case.prompt is not None: [e2e-llm-inference-service] test_payload = { [e2e-llm-inference-service] "model": test_case.model_name [e2e-llm-inference-service] if not extra_headers or MODEL_ROUTING_HEADER not in extra_headers [e2e-llm-inference-service] else extra_headers[MODEL_ROUTING_HEADER], [e2e-llm-inference-service] "prompt": test_case.prompt, [e2e-llm-inference-service] "max_tokens": test_case.max_tokens, [e2e-llm-inference-service] } [e2e-llm-inference-service] else: [e2e-llm-inference-service] test_payload = None [e2e-llm-inference-service] [e2e-llm-inference-service] logger.info(f"Calling LLM service at {model_url} with payload {test_payload}") [e2e-llm-inference-service] try: [e2e-llm-inference-service] if test_payload is not None: [e2e-llm-inference-service] > response = post_with_retry( [e2e-llm-inference-service] model_url, [e2e-llm-inference-service] headers=headers, [e2e-llm-inference-service] json_data=test_payload, [e2e-llm-inference-service] timeout=test_case.response_timeout, [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] llmisvc/test_llm_inference_service.py:1095: [e2e-llm-inference-service] _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ [e2e-llm-inference-service] [e2e-llm-inference-service] url = 'http://ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com/kserve-ci-e2e-test/rolling-upgrade-test/v1/completions' [e2e-llm-inference-service] [e2e-llm-inference-service] def post_with_retry( [e2e-llm-inference-service] url: str, [e2e-llm-inference-service] *, [e2e-llm-inference-service] headers: Dict = None, [e2e-llm-inference-service] json_data: Union[Dict, List] = None, [e2e-llm-inference-service] data: Union[str, bytes] = None, [e2e-llm-inference-service] stream: bool = False, [e2e-llm-inference-service] timeout: float = None, [e2e-llm-inference-service] total_retries: int = DEFAULT_RETRY_TOTAL, [e2e-llm-inference-service] backoff_factor: float = DEFAULT_RETRY_BACKOFF_FACTOR, [e2e-llm-inference-service] retry_status_codes=DEFAULT_RETRY_STATUS_CODES, [e2e-llm-inference-service] ) -> requests.Response: [e2e-llm-inference-service] """ [e2e-llm-inference-service] Send POST request with retries for transient HTTP and network failures. [e2e-llm-inference-service] """ [e2e-llm-inference-service] if json_data is not None and data is not None: [e2e-llm-inference-service] raise ValueError("Only one of json_data or data can be provided.") [e2e-llm-inference-service] [e2e-llm-inference-service] with _retry_session( [e2e-llm-inference-service] ["POST"], total_retries, backoff_factor, retry_status_codes [e2e-llm-inference-service] ) as session: [e2e-llm-inference-service] > return session.post( [e2e-llm-inference-service] url, [e2e-llm-inference-service] json=json_data, [e2e-llm-inference-service] data=data, [e2e-llm-inference-service] headers=headers, [e2e-llm-inference-service] stream=stream, [e2e-llm-inference-service] timeout=timeout, [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] common/http_retry.py:70: [e2e-llm-inference-service] _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ [e2e-llm-inference-service] [e2e-llm-inference-service] self = [e2e-llm-inference-service] url = 'http://ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com/kserve-ci-e2e-test/rolling-upgrade-test/v1/completions' [e2e-llm-inference-service] data = None [e2e-llm-inference-service] json = {'max_tokens': 20, 'model': 'facebook/opt-125m', 'prompt': 'KServe is a'} [e2e-llm-inference-service] kwargs = {'headers': {'Content-Type': 'application/json'}, 'stream': False, 'timeout': 60} [e2e-llm-inference-service] [e2e-llm-inference-service] def post(self, url, data=None, json=None, **kwargs): [e2e-llm-inference-service] r"""Sends a POST request. Returns :class:`Response` object. [e2e-llm-inference-service] [e2e-llm-inference-service] :param url: URL for the new :class:`Request` object. [e2e-llm-inference-service] :param data: (optional) Dictionary, list of tuples, bytes, or file-like [e2e-llm-inference-service] object to send in the body of the :class:`Request`. [e2e-llm-inference-service] :param json: (optional) json to send in the body of the :class:`Request`. [e2e-llm-inference-service] :param \*\*kwargs: Optional arguments that ``request`` takes. [e2e-llm-inference-service] :rtype: requests.Response [e2e-llm-inference-service] """ [e2e-llm-inference-service] [e2e-llm-inference-service] > return self.request("POST", url, data=data, json=json, **kwargs) [e2e-llm-inference-service] [e2e-llm-inference-service] ../../python/kserve/.venv/lib64/python3.11/site-packages/requests/sessions.py:637: [e2e-llm-inference-service] _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ [e2e-llm-inference-service] [e2e-llm-inference-service] self = , method = 'POST' [e2e-llm-inference-service] url = 'http://ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com/kserve-ci-e2e-test/rolling-upgrade-test/v1/completions' [e2e-llm-inference-service] params = None, data = None, headers = {'Content-Type': 'application/json'} [e2e-llm-inference-service] cookies = None, files = None, auth = None, timeout = 60, allow_redirects = True [e2e-llm-inference-service] proxies = {}, hooks = None, stream = False, verify = None, cert = None [e2e-llm-inference-service] json = {'max_tokens': 20, 'model': 'facebook/opt-125m', 'prompt': 'KServe is a'} [e2e-llm-inference-service] [e2e-llm-inference-service] def request( [e2e-llm-inference-service] self, [e2e-llm-inference-service] method, [e2e-llm-inference-service] url, [e2e-llm-inference-service] params=None, [e2e-llm-inference-service] data=None, [e2e-llm-inference-service] headers=None, [e2e-llm-inference-service] cookies=None, [e2e-llm-inference-service] files=None, [e2e-llm-inference-service] auth=None, [e2e-llm-inference-service] timeout=None, [e2e-llm-inference-service] allow_redirects=True, [e2e-llm-inference-service] proxies=None, [e2e-llm-inference-service] hooks=None, [e2e-llm-inference-service] stream=None, [e2e-llm-inference-service] verify=None, [e2e-llm-inference-service] cert=None, [e2e-llm-inference-service] json=None, [e2e-llm-inference-service] ): [e2e-llm-inference-service] """Constructs a :class:`Request `, prepares it and sends it. [e2e-llm-inference-service] Returns :class:`Response ` object. [e2e-llm-inference-service] [e2e-llm-inference-service] :param method: method for the new :class:`Request` object. [e2e-llm-inference-service] :param url: URL for the new :class:`Request` object. [e2e-llm-inference-service] :param params: (optional) Dictionary or bytes to be sent in the query [e2e-llm-inference-service] string for the :class:`Request`. [e2e-llm-inference-service] :param data: (optional) Dictionary, list of tuples, bytes, or file-like [e2e-llm-inference-service] object to send in the body of the :class:`Request`. [e2e-llm-inference-service] :param json: (optional) json to send in the body of the [e2e-llm-inference-service] :class:`Request`. [e2e-llm-inference-service] :param headers: (optional) Dictionary of HTTP Headers to send with the [e2e-llm-inference-service] :class:`Request`. [e2e-llm-inference-service] :param cookies: (optional) Dict or CookieJar object to send with the [e2e-llm-inference-service] :class:`Request`. [e2e-llm-inference-service] :param files: (optional) Dictionary of ``'filename': file-like-objects`` [e2e-llm-inference-service] for multipart encoding upload. [e2e-llm-inference-service] :param auth: (optional) Auth tuple or callable to enable [e2e-llm-inference-service] Basic/Digest/Custom HTTP Auth. [e2e-llm-inference-service] :param timeout: (optional) How long to wait for the server to send [e2e-llm-inference-service] data before giving up, as a float, or a :ref:`(connect timeout, [e2e-llm-inference-service] read timeout) ` tuple. [e2e-llm-inference-service] :type timeout: float or tuple [e2e-llm-inference-service] :param allow_redirects: (optional) Set to True by default. [e2e-llm-inference-service] :type allow_redirects: bool [e2e-llm-inference-service] :param proxies: (optional) Dictionary mapping protocol or protocol and [e2e-llm-inference-service] hostname to the URL of the proxy. [e2e-llm-inference-service] :param hooks: (optional) Dictionary mapping hook name to one event or [e2e-llm-inference-service] list of events, event must be callable. [e2e-llm-inference-service] :param stream: (optional) whether to immediately download the response [e2e-llm-inference-service] content. Defaults to ``False``. [e2e-llm-inference-service] :param verify: (optional) Either a boolean, in which case it controls whether we verify [e2e-llm-inference-service] the server's TLS certificate, or a string, in which case it must be a path [e2e-llm-inference-service] to a CA bundle to use. Defaults to ``True``. When set to [e2e-llm-inference-service] ``False``, requests will accept any TLS certificate presented by [e2e-llm-inference-service] the server, and will ignore hostname mismatches and/or expired [e2e-llm-inference-service] certificates, which will make your application vulnerable to [e2e-llm-inference-service] man-in-the-middle (MitM) attacks. Setting verify to ``False`` [e2e-llm-inference-service] may be useful during local development or testing. [e2e-llm-inference-service] :param cert: (optional) if String, path to ssl client cert file (.pem). [e2e-llm-inference-service] If Tuple, ('cert', 'key') pair. [e2e-llm-inference-service] :rtype: requests.Response [e2e-llm-inference-service] """ [e2e-llm-inference-service] # Create the Request. [e2e-llm-inference-service] req = Request( [e2e-llm-inference-service] method=method.upper(), [e2e-llm-inference-service] url=url, [e2e-llm-inference-service] headers=headers, [e2e-llm-inference-service] files=files, [e2e-llm-inference-service] data=data or {}, [e2e-llm-inference-service] json=json, [e2e-llm-inference-service] params=params or {}, [e2e-llm-inference-service] auth=auth, [e2e-llm-inference-service] cookies=cookies, [e2e-llm-inference-service] hooks=hooks, [e2e-llm-inference-service] ) [e2e-llm-inference-service] prep = self.prepare_request(req) [e2e-llm-inference-service] [e2e-llm-inference-service] proxies = proxies or {} [e2e-llm-inference-service] [e2e-llm-inference-service] settings = self.merge_environment_settings( [e2e-llm-inference-service] prep.url, proxies, stream, verify, cert [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] # Send the request. [e2e-llm-inference-service] send_kwargs = { [e2e-llm-inference-service] "timeout": timeout, [e2e-llm-inference-service] "allow_redirects": allow_redirects, [e2e-llm-inference-service] } [e2e-llm-inference-service] send_kwargs.update(settings) [e2e-llm-inference-service] > resp = self.send(prep, **send_kwargs) [e2e-llm-inference-service] [e2e-llm-inference-service] ../../python/kserve/.venv/lib64/python3.11/site-packages/requests/sessions.py:589: [e2e-llm-inference-service] _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ [e2e-llm-inference-service] [e2e-llm-inference-service] self = [e2e-llm-inference-service] request = [e2e-llm-inference-service] kwargs = {'cert': None, 'proxies': OrderedDict(), 'stream': False, 'timeout': 60, ...} [e2e-llm-inference-service] allow_redirects = True, stream = False, hooks = {'response': []} [e2e-llm-inference-service] adapter = [e2e-llm-inference-service] start = 1783439108.7332075 [e2e-llm-inference-service] [e2e-llm-inference-service] def send(self, request, **kwargs): [e2e-llm-inference-service] """Send a given PreparedRequest. [e2e-llm-inference-service] [e2e-llm-inference-service] :rtype: requests.Response [e2e-llm-inference-service] """ [e2e-llm-inference-service] # Set defaults that the hooks can utilize to ensure they always have [e2e-llm-inference-service] # the correct parameters to reproduce the previous request. [e2e-llm-inference-service] kwargs.setdefault("stream", self.stream) [e2e-llm-inference-service] kwargs.setdefault("verify", self.verify) [e2e-llm-inference-service] kwargs.setdefault("cert", self.cert) [e2e-llm-inference-service] if "proxies" not in kwargs: [e2e-llm-inference-service] kwargs["proxies"] = resolve_proxies(request, self.proxies, self.trust_env) [e2e-llm-inference-service] [e2e-llm-inference-service] # It's possible that users might accidentally send a Request object. [e2e-llm-inference-service] # Guard against that specific failure case. [e2e-llm-inference-service] if isinstance(request, Request): [e2e-llm-inference-service] raise ValueError("You can only send PreparedRequests.") [e2e-llm-inference-service] [e2e-llm-inference-service] # Set up variables needed for resolve_redirects and dispatching of hooks [e2e-llm-inference-service] allow_redirects = kwargs.pop("allow_redirects", True) [e2e-llm-inference-service] stream = kwargs.get("stream") [e2e-llm-inference-service] hooks = request.hooks [e2e-llm-inference-service] [e2e-llm-inference-service] # Get the appropriate adapter to use [e2e-llm-inference-service] adapter = self.get_adapter(url=request.url) [e2e-llm-inference-service] [e2e-llm-inference-service] # Start time (approximately) of the request [e2e-llm-inference-service] start = preferred_clock() [e2e-llm-inference-service] [e2e-llm-inference-service] # Send the request [e2e-llm-inference-service] > r = adapter.send(request, **kwargs) [e2e-llm-inference-service] [e2e-llm-inference-service] ../../python/kserve/.venv/lib64/python3.11/site-packages/requests/sessions.py:703: [e2e-llm-inference-service] _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ [e2e-llm-inference-service] [e2e-llm-inference-service] self = [e2e-llm-inference-service] request = , stream = False [e2e-llm-inference-service] timeout = Timeout(connect=60, read=60, total=None), verify = '/tmp/ca.crt' [e2e-llm-inference-service] cert = None, proxies = OrderedDict() [e2e-llm-inference-service] [e2e-llm-inference-service] def send( [e2e-llm-inference-service] self, request, stream=False, timeout=None, verify=True, cert=None, proxies=None [e2e-llm-inference-service] ): [e2e-llm-inference-service] """Sends PreparedRequest object. Returns Response object. [e2e-llm-inference-service] [e2e-llm-inference-service] :param request: The :class:`PreparedRequest ` being sent. [e2e-llm-inference-service] :param stream: (optional) Whether to stream the request content. [e2e-llm-inference-service] :param timeout: (optional) How long to wait for the server to send [e2e-llm-inference-service] data before giving up, as a float, or a :ref:`(connect timeout, [e2e-llm-inference-service] read timeout) ` tuple. [e2e-llm-inference-service] :type timeout: float or tuple or urllib3 Timeout object [e2e-llm-inference-service] :param verify: (optional) Either a boolean, in which case it controls whether [e2e-llm-inference-service] we verify the server's TLS certificate, or a string, in which case it [e2e-llm-inference-service] must be a path to a CA bundle to use [e2e-llm-inference-service] :param cert: (optional) Any user-provided SSL certificate to be trusted. [e2e-llm-inference-service] :param proxies: (optional) The proxies dictionary to apply to the request. [e2e-llm-inference-service] :rtype: requests.Response [e2e-llm-inference-service] """ [e2e-llm-inference-service] [e2e-llm-inference-service] try: [e2e-llm-inference-service] conn = self.get_connection_with_tls_context( [e2e-llm-inference-service] request, verify, proxies=proxies, cert=cert [e2e-llm-inference-service] ) [e2e-llm-inference-service] except LocationValueError as e: [e2e-llm-inference-service] raise InvalidURL(e, request=request) [e2e-llm-inference-service] [e2e-llm-inference-service] self.cert_verify(conn, request.url, verify, cert) [e2e-llm-inference-service] url = self.request_url(request, proxies) [e2e-llm-inference-service] self.add_headers( [e2e-llm-inference-service] request, [e2e-llm-inference-service] stream=stream, [e2e-llm-inference-service] timeout=timeout, [e2e-llm-inference-service] verify=verify, [e2e-llm-inference-service] cert=cert, [e2e-llm-inference-service] proxies=proxies, [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] chunked = not (request.body is None or "Content-Length" in request.headers) [e2e-llm-inference-service] [e2e-llm-inference-service] if isinstance(timeout, tuple): [e2e-llm-inference-service] try: [e2e-llm-inference-service] connect, read = timeout [e2e-llm-inference-service] timeout = TimeoutSauce(connect=connect, read=read) [e2e-llm-inference-service] except ValueError: [e2e-llm-inference-service] raise ValueError( [e2e-llm-inference-service] f"Invalid timeout {timeout}. Pass a (connect, read) timeout tuple, " [e2e-llm-inference-service] f"or a single float to set both timeouts to the same value." [e2e-llm-inference-service] ) [e2e-llm-inference-service] elif isinstance(timeout, TimeoutSauce): [e2e-llm-inference-service] pass [e2e-llm-inference-service] else: [e2e-llm-inference-service] timeout = TimeoutSauce(connect=timeout, read=timeout) [e2e-llm-inference-service] [e2e-llm-inference-service] try: [e2e-llm-inference-service] resp = conn.urlopen( [e2e-llm-inference-service] method=request.method, [e2e-llm-inference-service] url=url, [e2e-llm-inference-service] body=request.body, [e2e-llm-inference-service] headers=request.headers, [e2e-llm-inference-service] redirect=False, [e2e-llm-inference-service] assert_same_host=False, [e2e-llm-inference-service] preload_content=False, [e2e-llm-inference-service] decode_content=False, [e2e-llm-inference-service] retries=self.max_retries, [e2e-llm-inference-service] timeout=timeout, [e2e-llm-inference-service] chunked=chunked, [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] except (ProtocolError, OSError) as err: [e2e-llm-inference-service] raise ConnectionError(err, request=request) [e2e-llm-inference-service] [e2e-llm-inference-service] except MaxRetryError as e: [e2e-llm-inference-service] if isinstance(e.reason, ConnectTimeoutError): [e2e-llm-inference-service] # TODO: Remove this in 3.0.0: see #2811 [e2e-llm-inference-service] if not isinstance(e.reason, NewConnectionError): [e2e-llm-inference-service] raise ConnectTimeout(e, request=request) [e2e-llm-inference-service] [e2e-llm-inference-service] if isinstance(e.reason, ResponseError): [e2e-llm-inference-service] raise RetryError(e, request=request) [e2e-llm-inference-service] [e2e-llm-inference-service] if isinstance(e.reason, _ProxyError): [e2e-llm-inference-service] raise ProxyError(e, request=request) [e2e-llm-inference-service] [e2e-llm-inference-service] if isinstance(e.reason, _SSLError): [e2e-llm-inference-service] # This branch is for urllib3 v1.22 and later. [e2e-llm-inference-service] raise SSLError(e, request=request) [e2e-llm-inference-service] [e2e-llm-inference-service] > raise ConnectionError(e, request=request) [e2e-llm-inference-service] E requests.exceptions.ConnectionError: HTTPConnectionPool(host='ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com', port=80): Max retries exceeded with url: /kserve-ci-e2e-test/rolling-upgrade-test/v1/completions (Caused by ReadTimeoutError("HTTPConnectionPool(host='ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com', port=80): Read timed out. (read timeout=60)")) [e2e-llm-inference-service] [e2e-llm-inference-service] ../../python/kserve/.venv/lib64/python3.11/site-packages/requests/adapters.py:700: ConnectionError [e2e-llm-inference-service] [e2e-llm-inference-service] The above exception was the direct cause of the following exception: [e2e-llm-inference-service] [e2e-llm-inference-service] test_case = TestCase(base_refs=['router-managed', 'workload-llmd-simulator', 'model-fb-opt-125m'], prompt='KServe is a', service_n... {'name': 'model-fb-opt-125m-rolling-upgra-30e65219'}]}, [e2e-llm-inference-service] 'status': None}, model_name='facebook/opt-125m') [e2e-llm-inference-service] [e2e-llm-inference-service] @pytest.mark.llminferenceservice [e2e-llm-inference-service] @pytest.mark.parametrize( [e2e-llm-inference-service] "test_case", [e2e-llm-inference-service] [ [e2e-llm-inference-service] pytest.param( [e2e-llm-inference-service] TestCase( [e2e-llm-inference-service] base_refs=[ [e2e-llm-inference-service] "router-managed", [e2e-llm-inference-service] "workload-llmd-simulator", [e2e-llm-inference-service] "model-fb-opt-125m", [e2e-llm-inference-service] ], [e2e-llm-inference-service] prompt="KServe is a", [e2e-llm-inference-service] service_name="rolling-upgrade-test", [e2e-llm-inference-service] ), [e2e-llm-inference-service] marks=[pytest.mark.cluster_cpu, pytest.mark.cluster_single_node], [e2e-llm-inference-service] ), [e2e-llm-inference-service] ], [e2e-llm-inference-service] indirect=["test_case"], [e2e-llm-inference-service] ids=generate_test_id, [e2e-llm-inference-service] ) [e2e-llm-inference-service] @log_execution [e2e-llm-inference-service] def test_rolling_upgrade_coordination(test_case: TestCase): [e2e-llm-inference-service] """ [e2e-llm-inference-service] Verify the service recovers cleanly after a workload rolling update. [e2e-llm-inference-service] [e2e-llm-inference-service] Triggers a rolling update by patching a pod-template annotation on the [e2e-llm-inference-service] LLMInferenceService. After the rollout completes the workload Deployment [e2e-llm-inference-service] must be Available and the overall service must be Ready and able to serve [e2e-llm-inference-service] inference requests — confirming the workload-before-EPP update ordering [e2e-llm-inference-service] left the system in a healthy state. [e2e-llm-inference-service] """ [e2e-llm-inference-service] inject_k8s_proxy() [e2e-llm-inference-service] [e2e-llm-inference-service] kserve_client = KServeClient( [e2e-llm-inference-service] config_file=os.environ.get("KUBECONFIG", "~/.kube/config"), [e2e-llm-inference-service] client_configuration=client.Configuration(), [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] service_name = test_case.llm_service.metadata.name [e2e-llm-inference-service] namespace = test_case.llm_service.metadata.namespace [e2e-llm-inference-service] test_failed = False [e2e-llm-inference-service] [e2e-llm-inference-service] if not test_case.llm_service.metadata.annotations: [e2e-llm-inference-service] test_case.llm_service.metadata.annotations = {} [e2e-llm-inference-service] test_case.llm_service.metadata.annotations[ [e2e-llm-inference-service] "security.opendatahub.io/enable-auth" [e2e-llm-inference-service] ] = "false" [e2e-llm-inference-service] [e2e-llm-inference-service] try: [e2e-llm-inference-service] print(f"Creating LLMInferenceService {service_name}") [e2e-llm-inference-service] create_llmisvc(kserve_client, test_case.llm_service) [e2e-llm-inference-service] [e2e-llm-inference-service] print(f"Waiting for {service_name} to be ready") [e2e-llm-inference-service] wait_for_llm_isvc_ready( [e2e-llm-inference-service] kserve_client, test_case.llm_service, test_case.wait_timeout [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] apps_v1 = client.AppsV1Api() [e2e-llm-inference-service] workload_name = f"{service_name}-kserve" [e2e-llm-inference-service] [e2e-llm-inference-service] initial = apps_v1.read_namespaced_deployment(workload_name, namespace) [e2e-llm-inference-service] initial_generation = initial.metadata.generation [e2e-llm-inference-service] print(f"Workload Deployment initial generation: {initial_generation}") [e2e-llm-inference-service] [e2e-llm-inference-service] print("Triggering workload rolling update via pod-template annotation patch") [e2e-llm-inference-service] _patch_rolling_update_trigger(kserve_client, test_case.llm_service) [e2e-llm-inference-service] [e2e-llm-inference-service] print("Waiting for workload rollout to start (generation increment)") [e2e-llm-inference-service] wait_for_rollout_started( [e2e-llm-inference-service] apps_v1, workload_name, namespace, initial_generation, timeout_seconds=60 [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] print("Waiting for workload rollout to complete (Deployment Available)") [e2e-llm-inference-service] wait_for_deployment_available( [e2e-llm-inference-service] apps_v1, workload_name, namespace, timeout_seconds=120 [e2e-llm-inference-service] ) [e2e-llm-inference-service] print("Workload rollout complete") [e2e-llm-inference-service] [e2e-llm-inference-service] print(f"Waiting for {service_name} to be Ready after rollout") [e2e-llm-inference-service] wait_for_llm_isvc_ready( [e2e-llm-inference-service] kserve_client, test_case.llm_service, test_case.wait_timeout [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] print("Verifying inference still works after rolling update") [e2e-llm-inference-service] > wait_for_model_response(kserve_client, test_case, test_case.wait_timeout) [e2e-llm-inference-service] [e2e-llm-inference-service] llmisvc/test_rolling_upgrade.py:125: [e2e-llm-inference-service] _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ [e2e-llm-inference-service] [e2e-llm-inference-service] args = (, TestCase(base_refs=['router-managed', 'workload-llm... {'name': 'model-fb-opt-125m-rolling-upgra-30e65219'}]}, [e2e-llm-inference-service] 'status': None}, model_name='facebook/opt-125m'), 900) [e2e-llm-inference-service] kwargs = {}, func_name = 'wait_for_model_response' [e2e-llm-inference-service] timestamp_start = '2026-07-07T15:45:08.690808', start_time = 1783439108.691052 [e2e-llm-inference-service] duration = 904.6563992500305, timestamp_end = '2026-07-07T16:00:13.347454' [e2e-llm-inference-service] [e2e-llm-inference-service] @functools.wraps(func) [e2e-llm-inference-service] def wrapper(*args, **kwargs): [e2e-llm-inference-service] func_name = func.__name__ [e2e-llm-inference-service] [e2e-llm-inference-service] timestamp_start = datetime.now().isoformat() [e2e-llm-inference-service] logger.info( [e2e-llm-inference-service] f"[{func_name}] [{timestamp_start}] start - args={args}, kwargs={kwargs}" [e2e-llm-inference-service] ) [e2e-llm-inference-service] start_time = time.time() [e2e-llm-inference-service] [e2e-llm-inference-service] try: [e2e-llm-inference-service] > result = func(*args, **kwargs) [e2e-llm-inference-service] [e2e-llm-inference-service] llmisvc/logging.py:40: [e2e-llm-inference-service] _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ [e2e-llm-inference-service] [e2e-llm-inference-service] kserve_client = [e2e-llm-inference-service] test_case = TestCase(base_refs=['router-managed', 'workload-llmd-simulator', 'model-fb-opt-125m'], prompt='KServe is a', service_n... {'name': 'model-fb-opt-125m-rolling-upgra-30e65219'}]}, [e2e-llm-inference-service] 'status': None}, model_name='facebook/opt-125m') [e2e-llm-inference-service] timeout_seconds = 900, extra_headers = None [e2e-llm-inference-service] [e2e-llm-inference-service] @log_execution [e2e-llm-inference-service] def wait_for_model_response( [e2e-llm-inference-service] kserve_client: KServeClient, [e2e-llm-inference-service] test_case: TestCase, # noqa: F811 [e2e-llm-inference-service] timeout_seconds: int = 900, [e2e-llm-inference-service] extra_headers: Optional[Dict[str, str]] = None, [e2e-llm-inference-service] ) -> str: [e2e-llm-inference-service] def get_successful_response(): [e2e-llm-inference-service] try: [e2e-llm-inference-service] if test_case.url_getter: [e2e-llm-inference-service] service_url = test_case.url_getter(kserve_client, test_case.llm_service) [e2e-llm-inference-service] else: [e2e-llm-inference-service] service_url = get_llm_service_url(kserve_client, test_case.llm_service) [e2e-llm-inference-service] except Exception as e: [e2e-llm-inference-service] raise AssertionError(f"❌ Failed to get service URL: {e}") from e [e2e-llm-inference-service] [e2e-llm-inference-service] model_url = service_url + test_case.endpoint [e2e-llm-inference-service] [e2e-llm-inference-service] headers = {"Content-Type": "application/json"} [e2e-llm-inference-service] if extra_headers: [e2e-llm-inference-service] headers.update(extra_headers) [e2e-llm-inference-service] [e2e-llm-inference-service] if test_case.payload_formatter is not None: [e2e-llm-inference-service] test_payload = test_case.payload_formatter(test_case) [e2e-llm-inference-service] elif test_case.prompt is not None: [e2e-llm-inference-service] test_payload = { [e2e-llm-inference-service] "model": test_case.model_name [e2e-llm-inference-service] if not extra_headers or MODEL_ROUTING_HEADER not in extra_headers [e2e-llm-inference-service] else extra_headers[MODEL_ROUTING_HEADER], [e2e-llm-inference-service] "prompt": test_case.prompt, [e2e-llm-inference-service] "max_tokens": test_case.max_tokens, [e2e-llm-inference-service] } [e2e-llm-inference-service] else: [e2e-llm-inference-service] test_payload = None [e2e-llm-inference-service] [e2e-llm-inference-service] logger.info(f"Calling LLM service at {model_url} with payload {test_payload}") [e2e-llm-inference-service] try: [e2e-llm-inference-service] if test_payload is not None: [e2e-llm-inference-service] response = post_with_retry( [e2e-llm-inference-service] model_url, [e2e-llm-inference-service] headers=headers, [e2e-llm-inference-service] json_data=test_payload, [e2e-llm-inference-service] timeout=test_case.response_timeout, [e2e-llm-inference-service] ) [e2e-llm-inference-service] else: [e2e-llm-inference-service] response = get_with_retry( [e2e-llm-inference-service] model_url, [e2e-llm-inference-service] headers=headers, [e2e-llm-inference-service] timeout=test_case.response_timeout, [e2e-llm-inference-service] ) [e2e-llm-inference-service] except Exception as e: [e2e-llm-inference-service] logger.error(f"❌ Failed to call model: {e}") [e2e-llm-inference-service] raise AssertionError(f"❌ Failed to call model: {e}") from e [e2e-llm-inference-service] [e2e-llm-inference-service] logger.info(f"Model response is {response.status_code}: {response.text[:500]}") [e2e-llm-inference-service] [e2e-llm-inference-service] if 200 <= response.status_code < 300: [e2e-llm-inference-service] return response [e2e-llm-inference-service] raise AssertionError( [e2e-llm-inference-service] f"Service returned {response.status_code}: {response.text}" [e2e-llm-inference-service] ) [e2e-llm-inference-service] [e2e-llm-inference-service] > response = wait_for(get_successful_response, timeout=timeout_seconds, interval=5.0) [e2e-llm-inference-service] [e2e-llm-inference-service] llmisvc/test_llm_inference_service.py:1119: [e2e-llm-inference-service] _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ [e2e-llm-inference-service] [e2e-llm-inference-service] assertion_fn = .get_successful_response at 0x7f1d17b70040> [e2e-llm-inference-service] timeout = 900, interval = 5.0 [e2e-llm-inference-service] [e2e-llm-inference-service] def wait_for( [e2e-llm-inference-service] assertion_fn: Callable[[], Any], timeout: float = 5.0, interval: float = 0.1 [e2e-llm-inference-service] ) -> Any: [e2e-llm-inference-service] """Wait for the assertion to succeed within timeout.""" [e2e-llm-inference-service] deadline = time.time() + timeout [e2e-llm-inference-service] last_msg = None [e2e-llm-inference-service] while True: [e2e-llm-inference-service] try: [e2e-llm-inference-service] > return assertion_fn() [e2e-llm-inference-service] [e2e-llm-inference-service] llmisvc/test_llm_inference_service.py:1215: [e2e-llm-inference-service] _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ [e2e-llm-inference-service] [e2e-llm-inference-service] def get_successful_response(): [e2e-llm-inference-service] try: [e2e-llm-inference-service] if test_case.url_getter: [e2e-llm-inference-service] service_url = test_case.url_getter(kserve_client, test_case.llm_service) [e2e-llm-inference-service] else: [e2e-llm-inference-service] service_url = get_llm_service_url(kserve_client, test_case.llm_service) [e2e-llm-inference-service] except Exception as e: [e2e-llm-inference-service] raise AssertionError(f"❌ Failed to get service URL: {e}") from e [e2e-llm-inference-service] [e2e-llm-inference-service] model_url = service_url + test_case.endpoint [e2e-llm-inference-service] [e2e-llm-inference-service] headers = {"Content-Type": "application/json"} [e2e-llm-inference-service] if extra_headers: [e2e-llm-inference-service] headers.update(extra_headers) [e2e-llm-inference-service] [e2e-llm-inference-service] if test_case.payload_formatter is not None: [e2e-llm-inference-service] test_payload = test_case.payload_formatter(test_case) [e2e-llm-inference-service] elif test_case.prompt is not None: [e2e-llm-inference-service] test_payload = { [e2e-llm-inference-service] "model": test_case.model_name [e2e-llm-inference-service] if not extra_headers or MODEL_ROUTING_HEADER not in extra_headers [e2e-llm-inference-service] else extra_headers[MODEL_ROUTING_HEADER], [e2e-llm-inference-service] "prompt": test_case.prompt, [e2e-llm-inference-service] "max_tokens": test_case.max_tokens, [e2e-llm-inference-service] } [e2e-llm-inference-service] else: [e2e-llm-inference-service] test_payload = None [e2e-llm-inference-service] [e2e-llm-inference-service] logger.info(f"Calling LLM service at {model_url} with payload {test_payload}") [e2e-llm-inference-service] try: [e2e-llm-inference-service] if test_payload is not None: [e2e-llm-inference-service] response = post_with_retry( [e2e-llm-inference-service] model_url, [e2e-llm-inference-service] headers=headers, [e2e-llm-inference-service] json_data=test_payload, [e2e-llm-inference-service] timeout=test_case.response_timeout, [e2e-llm-inference-service] ) [e2e-llm-inference-service] else: [e2e-llm-inference-service] response = get_with_retry( [e2e-llm-inference-service] model_url, [e2e-llm-inference-service] headers=headers, [e2e-llm-inference-service] timeout=test_case.response_timeout, [e2e-llm-inference-service] ) [e2e-llm-inference-service] except Exception as e: [e2e-llm-inference-service] logger.error(f"❌ Failed to call model: {e}") [e2e-llm-inference-service] > raise AssertionError(f"❌ Failed to call model: {e}") from e [e2e-llm-inference-service] E AssertionError: ❌ Failed to call model: HTTPConnectionPool(host='ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com', port=80): Max retries exceeded with url: /kserve-ci-e2e-test/rolling-upgrade-test/v1/completions (Caused by ReadTimeoutError("HTTPConnectionPool(host='ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com', port=80): Read timed out. (read timeout=60)")) [e2e-llm-inference-service] [e2e-llm-inference-service] llmisvc/test_llm_inference_service.py:1109: AssertionError [e2e-llm-inference-service] ------------------------------ Captured log setup ------------------------------ [e2e-llm-inference-service] INFO e2e.llmisvc.logging:fixtures.py:1637 No HTTP proxy configured for k8s client [e2e-llm-inference-service] INFO e2e.llmisvc.logging:fixtures.py:1586 Checking LLMInferenceServiceConfig router-managed-rolling-upgrade-64e041bd in namespace kserve-ci-e2e-test [e2e-llm-inference-service] INFO e2e.llmisvc.logging:fixtures.py:1612 Resource not found, creating LLMInferenceServiceConfig router-managed-rolling-upgrade-64e041bd [e2e-llm-inference-service] INFO e2e.llmisvc.logging:fixtures.py:1622 ✓ Successfully created LLMInferenceServiceConfig router-managed-rolling-upgrade-64e041bd [e2e-llm-inference-service] INFO e2e.llmisvc.logging:fixtures.py:1586 Checking LLMInferenceServiceConfig workload-llmd-simulator-rolling-9660aa23 in namespace kserve-ci-e2e-test [e2e-llm-inference-service] INFO e2e.llmisvc.logging:fixtures.py:1612 Resource not found, creating LLMInferenceServiceConfig workload-llmd-simulator-rolling-9660aa23 [e2e-llm-inference-service] INFO e2e.llmisvc.logging:fixtures.py:1622 ✓ Successfully created LLMInferenceServiceConfig workload-llmd-simulator-rolling-9660aa23 [e2e-llm-inference-service] INFO e2e.llmisvc.logging:fixtures.py:1586 Checking LLMInferenceServiceConfig model-fb-opt-125m-rolling-upgra-30e65219 in namespace kserve-ci-e2e-test [e2e-llm-inference-service] INFO e2e.llmisvc.logging:fixtures.py:1612 Resource not found, creating LLMInferenceServiceConfig model-fb-opt-125m-rolling-upgra-30e65219 [e2e-llm-inference-service] INFO e2e.llmisvc.logging:fixtures.py:1622 ✓ Successfully created LLMInferenceServiceConfig model-fb-opt-125m-rolling-upgra-30e65219 [e2e-llm-inference-service] ------------------------------ Captured log call ------------------------------- [e2e-llm-inference-service] INFO e2e.llmisvc.logging:logging.py:34 [test_rolling_upgrade_coordination] [2026-07-07T15:44:24.351232] start - args=(), kwargs={'test_case': TestCase(base_refs=['router-managed', 'workload-llmd-simulator', 'model-fb-opt-125m'], prompt='KServe is a', service_name='rolling-upgrade-test', endpoint='/v1/completions', max_tokens=20, payload_formatter=None, response_assertion=, wait_timeout=900, response_timeout=60, extra_headers=None, url_getter=None, expected_gateway=None, before_test=[], after_test=[], peers=[], llm_service={'api_version': 'serving.kserve.io/v1alpha1', [e2e-llm-inference-service] 'kind': 'LLMInferenceService', [e2e-llm-inference-service] 'metadata': {'annotations': None, [e2e-llm-inference-service] 'creation_timestamp': None, [e2e-llm-inference-service] 'deletion_grace_period_seconds': None, [e2e-llm-inference-service] 'deletion_timestamp': None, [e2e-llm-inference-service] 'finalizers': None, [e2e-llm-inference-service] 'generate_name': None, [e2e-llm-inference-service] 'generation': None, [e2e-llm-inference-service] 'labels': None, [e2e-llm-inference-service] 'managed_fields': None, [e2e-llm-inference-service] 'name': 'rolling-upgrade-test', [e2e-llm-inference-service] 'namespace': 'kserve-ci-e2e-test', [e2e-llm-inference-service] 'owner_references': None, [e2e-llm-inference-service] 'resource_version': None, [e2e-llm-inference-service] 'self_link': None, [e2e-llm-inference-service] 'uid': None}, [e2e-llm-inference-service] 'spec': {'baseRefs': [{'name': 'router-managed-rolling-upgrade-64e041bd'}, [e2e-llm-inference-service] {'name': 'workload-llmd-simulator-rolling-9660aa23'}, [e2e-llm-inference-service] {'name': 'model-fb-opt-125m-rolling-upgra-30e65219'}]}, [e2e-llm-inference-service] 'status': None}, model_name='facebook/opt-125m')} [e2e-llm-inference-service] INFO e2e.llmisvc.logging:fixtures.py:1637 No HTTP proxy configured for k8s client [e2e-llm-inference-service] INFO e2e.llmisvc.logging:logging.py:34 [create_llmisvc] [2026-07-07T15:44:24.363846] start - args=(, {'api_version': 'serving.kserve.io/v1alpha1', [e2e-llm-inference-service] 'kind': 'LLMInferenceService', [e2e-llm-inference-service] 'metadata': {'annotations': {'security.opendatahub.io/enable-auth': 'false'}, [e2e-llm-inference-service] 'creation_timestamp': None, [e2e-llm-inference-service] 'deletion_grace_period_seconds': None, [e2e-llm-inference-service] 'deletion_timestamp': None, [e2e-llm-inference-service] 'finalizers': None, [e2e-llm-inference-service] 'generate_name': None, [e2e-llm-inference-service] 'generation': None, [e2e-llm-inference-service] 'labels': None, [e2e-llm-inference-service] 'managed_fields': None, [e2e-llm-inference-service] 'name': 'rolling-upgrade-test', [e2e-llm-inference-service] 'namespace': 'kserve-ci-e2e-test', [e2e-llm-inference-service] 'owner_references': None, [e2e-llm-inference-service] 'resource_version': None, [e2e-llm-inference-service] 'self_link': None, [e2e-llm-inference-service] 'uid': None}, [e2e-llm-inference-service] 'spec': {'baseRefs': [{'name': 'router-managed-rolling-upgrade-64e041bd'}, [e2e-llm-inference-service] {'name': 'workload-llmd-simulator-rolling-9660aa23'}, [e2e-llm-inference-service] {'name': 'model-fb-opt-125m-rolling-upgra-30e65219'}]}, [e2e-llm-inference-service] 'status': None}), kwargs={} [e2e-llm-inference-service] INFO e2e.llmisvc.logging:logging.py:43 [create_llmisvc] [2026-07-07T15:44:24.437561] end - ✅ in 0.073s [e2e-llm-inference-service] INFO e2e.llmisvc.logging:logging.py:34 [wait_for_llm_isvc_ready] [2026-07-07T15:44:24.437739] start - args=(, {'api_version': 'serving.kserve.io/v1alpha1', [e2e-llm-inference-service] 'kind': 'LLMInferenceService', [e2e-llm-inference-service] 'metadata': {'annotations': {'security.opendatahub.io/enable-auth': 'false'}, [e2e-llm-inference-service] 'creation_timestamp': None, [e2e-llm-inference-service] 'deletion_grace_period_seconds': None, [e2e-llm-inference-service] 'deletion_timestamp': None, [e2e-llm-inference-service] 'finalizers': None, [e2e-llm-inference-service] 'generate_name': None, [e2e-llm-inference-service] 'generation': None, [e2e-llm-inference-service] 'labels': None, [e2e-llm-inference-service] 'managed_fields': None, [e2e-llm-inference-service] 'name': 'rolling-upgrade-test', [e2e-llm-inference-service] 'namespace': 'kserve-ci-e2e-test', [e2e-llm-inference-service] 'owner_references': None, [e2e-llm-inference-service] 'resource_version': None, [e2e-llm-inference-service] 'self_link': None, [e2e-llm-inference-service] 'uid': None}, [e2e-llm-inference-service] 'spec': {'baseRefs': [{'name': 'router-managed-rolling-upgrade-64e041bd'}, [e2e-llm-inference-service] {'name': 'workload-llmd-simulator-rolling-9660aa23'}, [e2e-llm-inference-service] {'name': 'model-fb-opt-125m-rolling-upgra-30e65219'}]}, [e2e-llm-inference-service] 'status': None}, 900), kwargs={} [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1222 Waiting: No conditions found in status [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1222 Waiting: Missing true conditions: {'RouterReady', 'Ready', 'WorkloadsReady'}, expected {'RouterReady', 'Ready', 'WorkloadsReady'}, got [{'lastTransitionTime': '2026-07-07T15:44:32Z', 'message': 'The following HTTPRoutes are not ready: [kserve-ci-e2e-test/rolling-upgrade-test-kserve-route: "False" (reason "InvalidKind", message "referencing unsupported backendRef: group \\"inference.networking.x-k8s.io\\" kind \\"InferencePool\\"")]', 'reason': 'HTTPRoutesNotReady', 'severity': 'Info', 'status': 'False', 'type': 'HTTPRoutesReady'}, {'lastTransitionTime': '2026-07-07T15:44:32Z', 'message': 'Inference Pool kserve-ci-e2e-test/rolling-upgrade-test-inference-pool exists but no Gateway controller has accepted it yet', 'reason': 'WaitingForGateway', 'severity': 'Info', 'status': 'False', 'type': 'InferencePoolReady'}, {'lastTransitionTime': '2026-07-07T15:44:32Z', 'reason': 'Progressing', 'severity': 'Info', 'status': 'False', 'type': 'MainWorkloadReady'}, {'lastTransitionTime': '2026-07-07T15:44:32Z', 'severity': 'Info', 'status': 'True', 'type': 'PresetsCombined'}, {'lastTransitionTime': '2026-07-07T15:44:32Z', 'message': 'The following HTTPRoutes are not ready: [kserve-ci-e2e-test/rolling-upgrade-test-kserve-route: "False" (reason "InvalidKind", message "referencing unsupported backendRef: group \\"inference.networking.x-k8s.io\\" kind \\"InferencePool\\"")]', 'reason': 'HTTPRoutesNotReady', 'status': 'False', 'type': 'Ready'}, {'lastTransitionTime': '2026-07-07T15:44:32Z', 'message': 'The following HTTPRoutes are not ready: [kserve-ci-e2e-test/rolling-upgrade-test-kserve-route: "False" (reason "InvalidKind", message "referencing unsupported backendRef: group \\"inference.networking.x-k8s.io\\" kind \\"InferencePool\\"")]', 'reason': 'HTTPRoutesNotReady', 'status': 'False', 'type': 'RouterReady'}, {'lastTransitionTime': '2026-07-07T15:44:32Z', 'message': 'Deployment does not have minimum availability.', 'reason': 'MinimumReplicasUnavailable', 'severity': 'Info', 'status': 'False', 'type': 'SchedulerWorkloadReady'}, {'lastTransitionTime': '2026-07-07T15:44:32Z', 'reason': 'Progressing', 'status': 'False', 'type': 'WorkloadsReady'}] [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1222 Waiting: Missing true conditions: {'RouterReady', 'Ready'}, expected {'RouterReady', 'Ready', 'WorkloadsReady'}, got [{'lastTransitionTime': '2026-07-07T15:44:56Z', 'severity': 'Info', 'status': 'True', 'type': 'HTTPRoutesReady'}, {'lastTransitionTime': '2026-07-07T15:44:56Z', 'severity': 'Info', 'status': 'True', 'type': 'InferencePoolReady'}, {'lastTransitionTime': '2026-07-07T15:44:56Z', 'severity': 'Info', 'status': 'True', 'type': 'MainWorkloadReady'}, {'lastTransitionTime': '2026-07-07T15:44:32Z', 'severity': 'Info', 'status': 'True', 'type': 'PresetsCombined'}, {'lastTransitionTime': '2026-07-07T15:44:56Z', 'message': 'Deployment does not have minimum availability.', 'reason': 'MinimumReplicasUnavailable', 'status': 'False', 'type': 'Ready'}, {'lastTransitionTime': '2026-07-07T15:44:56Z', 'message': 'Deployment does not have minimum availability.', 'reason': 'MinimumReplicasUnavailable', 'status': 'False', 'type': 'RouterReady'}, {'lastTransitionTime': '2026-07-07T15:44:32Z', 'message': 'Deployment does not have minimum availability.', 'reason': 'MinimumReplicasUnavailable', 'severity': 'Info', 'status': 'False', 'type': 'SchedulerWorkloadReady'}, {'lastTransitionTime': '2026-07-07T15:44:56Z', 'status': 'True', 'type': 'WorkloadsReady'}] [e2e-llm-inference-service] INFO e2e.llmisvc.logging:logging.py:43 [wait_for_llm_isvc_ready] [2026-07-07T15:45:06.241261] end - ✅ in 41.803s [e2e-llm-inference-service] INFO e2e.llmisvc.logging:logging.py:34 [_patch_rolling_update_trigger] [2026-07-07T15:45:06.286148] start - args=(, {'api_version': 'serving.kserve.io/v1alpha1', [e2e-llm-inference-service] 'kind': 'LLMInferenceService', [e2e-llm-inference-service] 'metadata': {'annotations': {'security.opendatahub.io/enable-auth': 'false'}, [e2e-llm-inference-service] 'creation_timestamp': None, [e2e-llm-inference-service] 'deletion_grace_period_seconds': None, [e2e-llm-inference-service] 'deletion_timestamp': None, [e2e-llm-inference-service] 'finalizers': None, [e2e-llm-inference-service] 'generate_name': None, [e2e-llm-inference-service] 'generation': None, [e2e-llm-inference-service] 'labels': None, [e2e-llm-inference-service] 'managed_fields': None, [e2e-llm-inference-service] 'name': 'rolling-upgrade-test', [e2e-llm-inference-service] 'namespace': 'kserve-ci-e2e-test', [e2e-llm-inference-service] 'owner_references': None, [e2e-llm-inference-service] 'resource_version': None, [e2e-llm-inference-service] 'self_link': None, [e2e-llm-inference-service] 'uid': None}, [e2e-llm-inference-service] 'spec': {'baseRefs': [{'name': 'router-managed-rolling-upgrade-64e041bd'}, [e2e-llm-inference-service] {'name': 'workload-llmd-simulator-rolling-9660aa23'}, [e2e-llm-inference-service] {'name': 'model-fb-opt-125m-rolling-upgra-30e65219'}]}, [e2e-llm-inference-service] 'status': None}), kwargs={} [e2e-llm-inference-service] INFO e2e.llmisvc.logging:logging.py:43 [_patch_rolling_update_trigger] [2026-07-07T15:45:06.633601] end - ✅ in 0.347s [e2e-llm-inference-service] INFO e2e.llmisvc.logging:logging.py:34 [wait_for_rollout_started] [2026-07-07T15:45:06.633803] start - args=(, 'rolling-upgrade-test-kserve', 'kserve-ci-e2e-test', 1), kwargs={'timeout_seconds': 60} [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1222 Waiting: Deployment rolling-upgrade-test-kserve generation still at 1, expected > 1 [e2e-llm-inference-service] assert 1 > 1 [e2e-llm-inference-service] + where 1 = {'annotations': {'deployment.kubernetes.io/revision': '1'},\n 'creation_timestamp': datetime.datetime(2026, 7, 7, 15, 44, 29, tzinfo=tzlocal()),\n 'deletion_grace_period_seconds': None,\n 'deletion_timestamp': None,\n 'finalizers': None,\n 'generate_name': None,\n 'generation': 1,\n 'labels': {'app.kubernetes.io/component': 'llminferenceservice-workload',\n 'app.kubernetes.io/name': 'rolling-upgrade-test',\n 'app.kubernetes.io/part-of': 'llminferenceservice',\n 'kserve.io/component': 'workload',\n 'llm-d.ai/role': 'both'},\n 'managed_fields': [{'api_version': 'apps/v1',\n 'fields_type': 'FieldsV1',\n 'fields_v1': {'f:metadata': {'f:labels': {'.': {},\n 'f:app.kubernetes.io/component': {},\n 'f:app.kubernetes.io/name': {},\n 'f:app.kubernetes.io/part-of': {},\n 'f:kserve.io/component': {},\n 'f:llm-d.ai/role': {}},\n 'f:ownerReferences': {'.': {},\n 'k:{"uid":"ec474624-8b23-4879-bd20-332992667000"}': {}}},\n 'f:spec': {'f:progressDeadlineSeconds': {},\n 'f:replicas': {},\n 'f:revisionHistoryLimit': {},\n 'f:selector': {},\n 'f:strategy': {'f:rollingUpdate': {'.': {},\n 'f:maxSurge': {},\n 'f:maxUnavailable': {}},\n 'f:type': {}},\n 'f:template': {'f:metadata': {'f:labels': {'.': {},\n 'f:app.kubernetes.io/component': {},\n 'f:app.kubernetes.io/name': {},\n 'f:app.kubernetes.io/part-of': {},\n 'f:kserve.io/component': {},\n 'f:llm-d.ai/role': {}}},\n 'f:spec': {'f:containers': {'k:{"name":"main"}': {'.': {},\n 'f:args': {},\n 'f:command': {},\n 'f:env': {'.': {},\n 'k:{"name":"HF_HUB_CACHE"}': {'.': {},\n 'f:name': {},\n 'f:value': {}},\n 'k:{"name":"HOME"}': {'.': {},\n [e2e-llm-inference-service] 'f:name': {},\n 'f:value': {}},\n 'k:{"name":"VLLM_LOGGING_LEVEL"}': {'.': {},\n 'f:name': {},\n 'f:value': {}}},\n 'f:image': {},\n 'f:imagePullPolicy': {},\n 'f:lifecycle': {'.': {},\n 'f:preStop': {'.': {},\n 'f:exec': {'.': {},\n 'f:command': {}}}},\n 'f:livenessProbe': {'.': {},\n 'f:failureThreshold': {},\n 'f:httpGet': {'.': {},\n 'f:path': {},\n 'f:port': {},\n 'f:scheme': {}},\n 'f:periodSeconds': {},\n 'f:successThreshold': {},\n 'f:timeoutSeconds': {}},\n 'f:name': {},\n 'f:ports': {'.': {},\n 'k:{"containerPort":8000,"protocol":"TCP"}': {'.': {},\n 'f:containerPort': {},\n 'f:protocol': {}}},\n 'f:readinessProbe': {'.': {},\n [e2e-llm-inference-service] 'f:failureThreshold': {},\n 'f:httpGet': {'.': {},\n 'f:path': {},\n 'f:port': {},\n 'f:scheme': {}},\n 'f:periodSeconds': {},\n 'f:successThreshold': {},\n 'f:timeoutSeconds': {}},\n 'f:resources': {'.': {},\n 'f:limits': {'.': {},\n 'f:cpu': {},\n 'f:memory': {}},\n 'f:requests': {'.': {},\n 'f:cpu': {},\n 'f:memory': {}}},\n 'f:securityContext': {'.': {},\n 'f:allowPrivilegeEscalation': {},\n 'f:capabilities': {'.': {},\n 'f:drop': {}},\n 'f:readOnlyRootFilesystem': {},\n 'f:runAsNonRoot': {},\n 'f:seccompProfile': {'.': {},\n 'f:type': {}}},\n 'f:startupProbe': {'.': {},\n 'f:failureThreshold': {},\n 'f:httpGet': {'.': {},\n [e2e-llm-inference-service] 'f:path': {},\n 'f:port': {},\n 'f:scheme': {}},\n 'f:periodSeconds': {},\n 'f:successThreshold': {},\n 'f:timeoutSeconds': {}},\n 'f:terminationMessagePath': {},\n 'f:terminationMessagePolicy': {},\n 'f:volumeMounts': {'.': {},\n 'k:{"mountPath":"/dev/shm"}': {'.': {},\n 'f:mountPath': {},\n 'f:name': {}},\n 'k:{"mountPath":"/home"}': {'.': {},\n 'f:mountPath': {},\n 'f:name': {}},\n 'k:{"mountPath":"/models"}': {'.': {},\n 'f:mountPath': {},\n 'f:name': {}},\n 'k:{"mountPath":"/tmp"}': {'.': {},\n 'f:mountPath': {},\n 'f:name': {}},\n 'k:{"mountPath":"/var/run/kserve/tls"}': {'.': {},\n 'f:mountPath': {},\n 'f:name': {},\n 'f:readOnly': {}}}}},\n [e2e-llm-inference-service] 'f:dnsPolicy': {},\n 'f:restartPolicy': {},\n 'f:schedulerName': {},\n 'f:securityContext': {},\n 'f:terminationGracePeriodSeconds': {},\n 'f:volumes': {'.': {},\n 'k:{"name":"dshm"}': {'.': {},\n 'f:emptyDir': {'.': {},\n 'f:medium': {},\n 'f:sizeLimit': {}},\n 'f:name': {}},\n 'k:{"name":"home"}': {'.': {},\n 'f:emptyDir': {},\n 'f:name': {}},\n 'k:{"name":"model-cache"}': {'.': {},\n 'f:emptyDir': {},\n 'f:name': {}},\n 'k:{"name":"tls-certs"}': {'.': {},\n 'f:name': {},\n 'f:secret': {'.': {},\n 'f:defaultMode': {},\n 'f:secretName': {}}},\n 'k:{"name":"tmp-dir"}': {'.': {},\n 'f:emptyDir': {},\n 'f:name': {}}}}}}},\n 'manager': 'manager',\n 'operation': 'Update',\n 'subresource': None,\n 'time': datetime.datetime(2026, 7, 7, 15, 44, 29, tzinfo=tzlocal())},\n {'api_version': 'apps/v1',\n 'fields_type': 'FieldsV1',\n 'fields_v1': {'f:metadata': {'f:annotations': {'.': {},\n 'f:deployment.kubernetes.io/revision': {}}},\n 'f:status': {'f:availableReplicas': {},\n 'f:conditions': {'.': {},\n 'k:{"type":"Available"}': {'.': {},\n 'f:lastTransitionTime': {},\n 'f:lastUpdateTime': {},\n [e2e-llm-inference-service] 'f:message': {},\n 'f:reason': {},\n 'f:status': {},\n 'f:type': {}},\n 'k:{"type":"Progressing"}': {'.': {},\n 'f:lastTransitionTime': {},\n 'f:lastUpdateTime': {},\n 'f:message': {},\n 'f:reason': {},\n 'f:status': {},\n 'f:type': {}}},\n 'f:observedGeneration': {},\n 'f:readyReplicas': {},\n 'f:replicas': {},\n 'f:updatedReplicas': {}}},\n 'manager': 'kube-controller-manager',\n 'operation': 'Update',\n 'subresource': 'status',\n 'time': datetime.datetime(2026, 7, 7, 15, 44, 39, tzinfo=tzlocal())}],\n 'name': 'rolling-upgrade-test-kserve',\n 'namespace': 'kserve-ci-e2e-test',\n 'owner_references': [{'api_version': 'serving.kserve.io/v1alpha2',\n 'block_owner_deletion': True,\n 'controller': True,\n 'kind': 'LLMInferenceService',\n 'name': 'rolling-upgrade-test',\n 'uid': 'ec474624-8b23-4879-bd20-332992667000'}],\n 'resource_version': '100692',\n 'self_link': None,\n 'uid': 'c05c7a31-033f-4eb3-a160-b1c45e37ed78'}.generation [e2e-llm-inference-service] + where {'annotations': {'deployment.kubernetes.io/revision': '1'},\n 'creation_timestamp': datetime.datetime(2026, 7, 7, 15, 44, 29, tzinfo=tzlocal()),\n 'deletion_grace_period_seconds': None,\n 'deletion_timestamp': None,\n 'finalizers': None,\n 'generate_name': None,\n 'generation': 1,\n 'labels': {'app.kubernetes.io/component': 'llminferenceservice-workload',\n 'app.kubernetes.io/name': 'rolling-upgrade-test',\n 'app.kubernetes.io/part-of': 'llminferenceservice',\n 'kserve.io/component': 'workload',\n 'llm-d.ai/role': 'both'},\n 'managed_fields': [{'api_version': 'apps/v1',\n 'fields_type': 'FieldsV1',\n 'fields_v1': {'f:metadata': {'f:labels': {'.': {},\n 'f:app.kubernetes.io/component': {},\n 'f:app.kubernetes.io/name': {},\n 'f:app.kubernetes.io/part-of': {},\n 'f:kserve.io/component': {},\n 'f:llm-d.ai/role': {}},\n 'f:ownerReferences': {'.': {},\n 'k:{"uid":"ec474624-8b23-4879-bd20-332992667000"}': {}}},\n 'f:spec': {'f:progressDeadlineSeconds': {},\n 'f:replicas': {},\n 'f:revisionHistoryLimit': {},\n 'f:selector': {},\n 'f:strategy': {'f:rollingUpdate': {'.': {},\n 'f:maxSurge': {},\n 'f:maxUnavailable': {}},\n 'f:type': {}},\n 'f:template': {'f:metadata': {'f:labels': {'.': {},\n 'f:app.kubernetes.io/component': {},\n 'f:app.kubernetes.io/name': {},\n 'f:app.kubernetes.io/part-of': {},\n 'f:kserve.io/component': {},\n 'f:llm-d.ai/role': {}}},\n 'f:spec': {'f:containers': {'k:{"name":"main"}': {'.': {},\n 'f:args': {},\n 'f:command': {},\n 'f:env': {'.': {},\n 'k:{"name":"HF_HUB_CACHE"}': {'.': {},\n 'f:name': {},\n 'f:value': {}},\n 'k:{"name":"HOME"}': {'.': {},\n [e2e-llm-inference-service] 'f:name': {},\n 'f:value': {}},\n 'k:{"name":"VLLM_LOGGING_LEVEL"}': {'.': {},\n 'f:name': {},\n 'f:value': {}}},\n 'f:image': {},\n 'f:imagePullPolicy': {},\n 'f:lifecycle': {'.': {},\n 'f:preStop': {'.': {},\n 'f:exec': {'.': {},\n 'f:command': {}}}},\n 'f:livenessProbe': {'.': {},\n 'f:failureThreshold': {},\n 'f:httpGet': {'.': {},\n 'f:path': {},\n 'f:port': {},\n 'f:scheme': {}},\n 'f:periodSeconds': {},\n 'f:successThreshold': {},\n 'f:timeoutSeconds': {}},\n 'f:name': {},\n 'f:ports': {'.': {},\n 'k:{"containerPort":8000,"protocol":"TCP"}': {'.': {},\n 'f:containerPort': {},\n 'f:protocol': {}}},\n 'f:readinessProbe': {'.': {},\n ' [e2e-llm-inference-service] f:failureThreshold': {},\n 'f:httpGet': {'.': {},\n 'f:path': {},\n 'f:port': {},\n 'f:scheme': {}},\n 'f:periodSeconds': {},\n 'f:successThreshold': {},\n 'f:timeoutSeconds': {}},\n 'f:resources': {'.': {},\n 'f:limits': {'.': {},\n 'f:cpu': {},\n 'f:memory': {}},\n 'f:requests': {'.': {},\n 'f:cpu': {},\n 'f:memory': {}}},\n 'f:securityContext': {'.': {},\n 'f:allowPrivilegeEscalation': {},\n 'f:capabilities': {'.': {},\n 'f:drop': {}},\n 'f:readOnlyRootFilesystem': {},\n 'f:runAsNonRoot': {},\n 'f:seccompProfile': {'.': {},\n 'f:type': {}}},\n 'f:startupProbe': {'.': {},\n 'f:failureThreshold': {},\n 'f:httpGet': {'.': {},\n [e2e-llm-inference-service] 'f:path': {},\n 'f:port': {},\n 'f:scheme': {}},\n 'f:periodSeconds': {},\n 'f:successThreshold': {},\n 'f:timeoutSeconds': {}},\n 'f:terminationMessagePath': {},\n 'f:terminationMessagePolicy': {},\n 'f:volumeMounts': {'.': {},\n 'k:{"mountPath":"/dev/shm"}': {'.': {},\n 'f:mountPath': {},\n 'f:name': {}},\n 'k:{"mountPath":"/home"}': {'.': {},\n 'f:mountPath': {},\n 'f:name': {}},\n 'k:{"mountPath":"/models"}': {'.': {},\n 'f:mountPath': {},\n 'f:name': {}},\n 'k:{"mountPath":"/tmp"}': {'.': {},\n 'f:mountPath': {},\n 'f:name': {}},\n 'k:{"mountPath":"/var/run/kserve/tls"}': {'.': {},\n 'f:mountPath': {},\n 'f:name': {},\n 'f:readOnly': {}}}}},\n [e2e-llm-inference-service] 'f:dnsPolicy': {},\n 'f:restartPolicy': {},\n 'f:schedulerName': {},\n 'f:securityContext': {},\n 'f:terminationGracePeriodSeconds': {},\n 'f:volumes': {'.': {},\n 'k:{"name":"dshm"}': {'.': {},\n 'f:emptyDir': {'.': {},\n 'f:medium': {},\n 'f:sizeLimit': {}},\n 'f:name': {}},\n 'k:{"name":"home"}': {'.': {},\n 'f:emptyDir': {},\n 'f:name': {}},\n 'k:{"name":"model-cache"}': {'.': {},\n 'f:emptyDir': {},\n 'f:name': {}},\n 'k:{"name":"tls-certs"}': {'.': {},\n 'f:name': {},\n 'f:secret': {'.': {},\n 'f:defaultMode': {},\n 'f:secretName': {}}},\n 'k:{"name":"tmp-dir"}': {'.': {},\n 'f:emptyDir': {},\n 'f:name': {}}}}}}},\n 'manager': 'manager',\n 'operation': 'Update',\n 'subresource': None,\n 'time': datetime.datetime(2026, 7, 7, 15, 44, 29, tzinfo=tzlocal())},\n {'api_version': 'apps/v1',\n 'fields_type': 'FieldsV1',\n 'fields_v1': {'f:metadata': {'f:annotations': {'.': {},\n 'f:deployment.kubernetes.io/revision': {}}},\n 'f:status': {'f:availableReplicas': {},\n 'f:conditions': {'.': {},\n 'k:{"type":"Available"}': {'.': {},\n 'f:lastTransitionTime': {},\n 'f:lastUpdateTime': {},\n [e2e-llm-inference-service] 'f:message': {},\n 'f:reason': {},\n 'f:status': {},\n 'f:type': {}},\n 'k:{"type":"Progressing"}': {'.': {},\n 'f:lastTransitionTime': {},\n 'f:lastUpdateTime': {},\n 'f:message': {},\n 'f:reason': {},\n 'f:status': {},\n 'f:type': {}}},\n 'f:observedGeneration': {},\n 'f:readyReplicas': {},\n 'f:replicas': {},\n 'f:updatedReplicas': {}}},\n 'manager': 'kube-controller-manager',\n 'operation': 'Update',\n 'subresource': 'status',\n 'time': datetime.datetime(2026, 7, 7, 15, 44, 39, tzinfo=tzlocal())}],\n 'name': 'rolling-upgrade-test-kserve',\n 'namespace': 'kserve-ci-e2e-test',\n 'owner_references': [{'api_version': 'serving.kserve.io/v1alpha2',\n 'block_owner_deletion': True,\n 'controller': True,\n 'kind': 'LLMInferenceService',\n 'name': 'rolling-upgrade-test',\n 'uid': 'ec474624-8b23-4879-bd20-332992667000'}],\n 'resource_version': '100692',\n 'self_link': None,\n 'uid': 'c05c7a31-033f-4eb3-a160-b1c45e37ed78'} = {'api_version': 'apps/v1',\n 'kind': 'Deployment',\n 'metadata': {'annotations': {'deployment.kubernetes.io/revision': '1'},\n 'creation_timestamp': datetime.datetime(2026, 7, 7, 15, 44, 29, tzinfo=tzlocal()),\n 'deletion_grace_period_seconds': None,\n 'deletion_timestamp': None,\n 'finalizers': None,\n 'generate_name': None,\n 'generation': 1,\n 'labels': {'app.kubernetes.io/component': 'llminferenceservice-workload',\n 'app.kubernetes.io/name': 'rolling-upgrade-test',\n 'app.kubernetes.io/part-of': 'llminferenceservice',\n 'kserve.io/component': 'workload',\n 'llm-d.ai/role': 'both'},\n 'managed_fields': [{'api_version': 'apps/v1',\n 'fields_type': 'FieldsV1',\n 'fields_v1': {'f:metadata': {'f:labels': {'.': {},\n 'f:app.kubernetes.io/component': {},\n 'f:app.kubernetes.io/name': {},\n 'f:app.kubernetes.io/part-of': {},\n 'f:kserve.io/component': {},\n 'f:llm-d.ai/role': {}},\n 'f:ownerReferences': {'.': {},\n 'k:{"uid":"ec474624-8b23-4879-bd20-332992667000"}': {}}},\n 'f:spe [e2e-llm-inference-service] c': {'f:progressDeadlineSeconds': {},\n 'f:replicas': {},\n 'f:revisionHistoryLimit': {},\n 'f:selector': {},\n 'f:strategy': {'f:rollingUpdate': {'.': {},\n 'f:maxSurge': {},\n 'f:maxUnavailable': {}},\n 'f:type': {}},\n 'f:template': {'f:metadata': {'f:labels': {'.': {},\n 'f:app.kubernetes.io/component': {},\n 'f:app.kubernetes.io/name': {},\n 'f:app.kubernetes.io/part-of': {},\n 'f:kserve.io/component': {},\n 'f:llm-d.ai/role': {}}},\n 'f:spec': {'f:containers': {'k:{"name":"main"}': {'.': {},\n 'f:args': {},\n 'f:command': {},\n 'f:env': {'.': {},\n 'k:{"name":"HF_HUB_CACHE"}': {'.': {},\n 'f:name': {},\n 'f:value': {}},\n 'k:{"name":"HOME"}': {'.': {},\n 'f:name': {},\n 'f:value': {}},\n 'k:{"name":"VLLM_LOGGING_LEVEL"}': {'.': {},\n 'f:name': {},\n 'f:value': {}}},\n 'f:image': {},\n 'f:imagePullPolicy': {},\n 'f:lifecycle': {'.': {},\n [e2e-llm-inference-service] 'f:preStop': {'.': {},\n 'f:exec': {'.': {},\n 'f:command': {}}}},\n 'f:livenessProbe': {'.': {},\n 'f:failureThreshold': {},\n 'f:httpGet': {'.': {},\n 'f:path': {},\n 'f:port': {},\n 'f:scheme': {}},\n 'f:periodSeconds': {},\n 'f:successThreshold': {},\n 'f:timeoutSeconds': {}},\n 'f:name': {},\n 'f:ports': {'.': {},\n 'k:{"containerPort":8000,"protocol":"TCP"}': {'.': {},\n 'f:containerPort': {},\n 'f:protocol': {}}},\n 'f:readinessProbe': {'.': {},\n 'f:failureThreshold': {},\n 'f:httpGet': {'.': {},\n 'f:path': {},\n 'f:port': {},\n 'f:scheme': {}},\n [e2e-llm-inference-service] 'f:periodSeconds': {},\n 'f:successThreshold': {},\n 'f:timeoutSeconds': {}},\n 'f:resources': {'.': {},\n 'f:limits': {'.': {},\n 'f:cpu': {},\n 'f:memory': {}},\n 'f:requests': {'.': {},\n 'f:cpu': {},\n 'f:memory': {}}},\n 'f:securityContext': {'.': {},\n 'f:allowPrivilegeEscalation': {},\n 'f:capabilities': {'.': {},\n 'f:drop': {}},\n 'f:readOnlyRootFilesystem': {},\n 'f:runAsNonRoot': {},\n 'f:seccompProfile': {'.': {},\n 'f:type': {}}},\n 'f:startupProbe': {'.': {},\n 'f:failureThreshold': {},\n 'f:httpGet': {'.': {},\n 'f:path': {},\n 'f:port': {},\n 'f:scheme': {}},\n 'f:periodSecon [e2e-llm-inference-service] ds': {},\n 'f:successThreshold': {},\n 'f:timeoutSeconds': {}},\n 'f:terminationMessagePath': {},\n 'f:terminationMessagePolicy': {},\n 'f:volumeMounts': {'.': {},\n 'k:{"mountPath":"/dev/shm"}': {'.': {},\n 'f:mountPath': {},\n 'f:name': {}},\n 'k:{"mountPath":"/home"}': {'.': {},\n 'f:mountPath': {},\n 'f:name': {}},\n 'k:{"mountPath":"/models"}': {'.': {},\n 'f:mountPath': {},\n 'f:name': {}},\n 'k:{"mountPath":"/tmp"}': {'.': {},\n 'f:mountPath': {},\n 'f:name': {}},\n 'k:{"mountPath":"/var/run/kserve/tls"}': {'.': {},\n 'f:mountPath': {},\n 'f:name': {},\n 'f:readOnly': {}}}}},\n 'f:dnsPolicy': {},\n 'f:restartPolicy': {},\n [e2e-llm-inference-service] 'f:schedulerName': {},\n 'f:securityContext': {},\n 'f:terminationGracePeriodSeconds': {},\n 'f:volumes': {'.': {},\n 'k:{"name":"dshm"}': {'.': {},\n 'f:emptyDir': {'.': {},\n 'f:medium': {},\n 'f:sizeLimit': {}},\n 'f:name': {}},\n 'k:{"name":"home"}': {'.': {},\n 'f:emptyDir': {},\n 'f:name': {}},\n 'k:{"name":"model-cache"}': {'.': {},\n 'f:emptyDir': {},\n 'f:name': {}},\n 'k:{"name":"tls-certs"}': {'.': {},\n 'f:name': {},\n 'f:secret': {'.': {},\n 'f:defaultMode': {},\n 'f:secretName': {}}},\n 'k:{"name":"tmp-dir"}': {'.': {},\n 'f:emptyDir': {},\n 'f:name': {}}}}}}},\n 'manager': 'manager',\n 'operation': 'Update',\n 'subresource': None,\n 'time': datetime.datetime(2026, 7, 7, 15, 44, 29, tzinfo=tzlocal())},\n {'api_version': 'apps/v1',\n 'fields_type': 'FieldsV1',\n 'fields_v1': {'f:metadata': {'f:annotations': {'.': {},\n 'f:deployment.kubernetes.io/revision': {}}},\n 'f:status': {'f:availableReplicas': {},\n 'f:conditions': {'.': {},\n 'k:{"type":"Available"}': {'.': {},\n [e2e-llm-inference-service] 'f:lastTransitionTime': {},\n 'f:lastUpdateTime': {},\n 'f:message': {},\n 'f:reason': {},\n 'f:status': {},\n 'f:type': {}},\n 'k:{"type":"Progressing"}': {'.': {},\n 'f:lastTransitionTime': {},\n 'f:lastUpdateTime': {},\n 'f:message': {},\n 'f:reason': {},\n 'f:status': {},\n 'f:type': {}}},\n 'f:observedGeneration': {},\n 'f:readyReplicas': {},\n 'f:replicas': {},\n 'f:updatedReplicas': {}}},\n 'manager': 'kube-controller-manager',\n 'operation': 'Update',\n 'subresource': 'status',\n 'time': datetime.datetime(2026, 7, 7, 15, 44, 39, tzinfo=tzlocal())}],\n 'name': 'rolling-upgrade-test-kserve',\n 'namespace': 'kserve-ci-e2e-test',\n 'owner_references': [{'api_version': 'serving.kserve.io/v1alpha2',\n 'block_owner_deletion': True,\n 'controller': True,\n 'kind': 'LLMInferenceService',\n 'name': 'rolling-upgrade-test',\n 'uid': 'ec474624-8b23-4879-bd20-332992667000'}],\n 'resource_version': '100692',\n 'self_link': None,\n 'uid': 'c05c7a31-033f-4eb3-a160-b1c45e37ed78'},\n 'spec': {'min_ready_seconds': None,\n 'paused': None,\n 'progress_deadline_seconds': 600,\n 'replicas': 1,\n 'revision_history_limit': 10,\n 'selector': {'match_expressions': None,\n 'match_labels': {'app.kubernetes.io/component': 'llminferenceservice-workload',\n 'app.kubernetes.io/name': 'rolling-upgrade-test',\n 'app.kubernetes.io/part-of': 'llminferenceservice',\n 'kserve.io/component': 'workload',\n 'llm-d.ai/role': 'both'}},\n 'strategy': {'rolling_update': {'max_surge': '25%',\n 'max_unavailable': '25%'},\n 'type': 'RollingUpdate'},\n 'template': {'metadata': {'annotations': None,\n 'creation_timestamp': None,\n 'deletion_grace_period_seconds': None,\n 'deletion_timestamp': None,\n 'finalizers' [e2e-llm-inference-service] : None,\n 'generate_name': None,\n 'generation': None,\n 'labels': {'app.kubernetes.io/component': 'llminferenceservice-workload',\n 'app.kubernetes.io/name': 'rolling-upgrade-test',\n 'app.kubernetes.io/part-of': 'llminferenceservice',\n 'kserve.io/component': 'workload',\n 'llm-d.ai/role': 'both'},\n 'managed_fields': None,\n 'name': None,\n 'namespace': None,\n 'owner_references': None,\n 'resource_version': None,\n 'self_link': None,\n 'uid': None},\n 'spec': {'active_deadline_seconds': None,\n 'affinity': None,\n 'automount_service_account_token': None,\n 'containers': [{'args': ['--port',\n '8000',\n '--model',\n 'facebook/opt-125m',\n '--mode',\n 'random',\n '--ssl-certfile',\n '/var/run/kserve/tls/tls.crt',\n '--ssl-keyfile',\n '/var/run/kserve/tls/tls.key'],\n 'command': ['/app/llm-d-inference-sim'],\n 'env': [{'name': 'HOME',\n 'value': '/home',\n 'value_from': None},\n {'name': 'VLLM_LOGGING_LEVEL',\n 'value': 'INFO',\n 'value_from': None},\n {'name': 'HF_HUB_CACHE',\n 'value': '/models',\n 'value_from': None}],\n 'env_from': None,\n 'image': 'ghcr.io/llm-d/llm-d-inference-sim:v0.8.2',\n 'image_pull_policy': 'IfNotPresent',\n 'lifecycle': {'post_start': None,\n 'pre_stop': {'_exec': {'command': ['/bin/sleep',\n '15']},\n 'http_get': None,\n 'sleep': None,\n 'tcp_socket': None}},\n 'liveness_probe': {'_exec': None,\n 'failure_threshold': 10,\n 'grpc': None,\n 'http_get': {'host': None,\n 'http_headers': None,\n [e2e-llm-inference-service] 'path': '/health',\n 'port': 8000,\n 'scheme': 'HTTPS'},\n 'initial_delay_seconds': None,\n 'period_seconds': 10,\n 'success_threshold': 1,\n 'tcp_socket': None,\n 'termination_grace_period_seconds': None,\n 'timeout_seconds': 1},\n 'name': 'main',\n 'ports': [{'container_port': 8000,\n 'host_ip': None,\n 'host_port': None,\n 'name': None,\n 'protocol': 'TCP'}],\n 'readiness_probe': {'_exec': None,\n 'failure_threshold': 2,\n 'grpc': None,\n 'http_get': {'host': None,\n 'http_headers': None,\n 'path': '/health',\n 'port': 8000,\n 'scheme': 'HTTPS'},\n 'initial_delay_seconds': None,\n 'period_seconds': 1,\n 'success_threshold': 1,\n 'tcp_socket': None,\n 'termination_grace_period_seconds': None,\n 'timeout_seconds': 1},\n 'resize_policy': None,\n 'resources': {'claims': None,\n 'limits': {'cpu': '1',\n 'memory': '2Gi'},\n 'requests': {'cpu': '200m',\n 'memory': '2Gi'}},\n 'restart_policy': None,\n 'security_context': {'allow_privilege_escalation': False,\n 'app_armor_profile': None,\n 'capabilities': {'add': None,\n 'drop': ['ALL']},\n 'privileged': None,\n 'proc_mount': None,\n 'read_only_root_filesystem': True,\n 'run_as_group': None,\n 'run_as_non_root': True,\n [e2e-llm-inference-service] 'run_as_user': None,\n 'se_linux_options': None,\n 'seccomp_profile': {'localhost_profile': None,\n 'type': 'RuntimeDefault'},\n 'windows_options': None},\n 'startup_probe': {'_exec': None,\n 'failure_threshold': 60,\n 'grpc': None,\n 'http_get': {'host': None,\n 'http_headers': None,\n 'path': '/health',\n 'port': 8000,\n 'scheme': 'HTTPS'},\n 'initial_delay_seconds': None,\n 'period_seconds': 10,\n 'success_threshold': 1,\n 'tcp_socket': None,\n 'termination_grace_period_seconds': None,\n 'timeout_seconds': 1},\n 'stdin': None,\n 'stdin_once': None,\n 'termination_message_path': '/dev/termination-log',\n 'termination_message_policy': 'FallbackToLogsOnError',\n 'tty': None,\n 'volume_devices': None,\n 'volume_mounts': [{'mount_path': '/home',\n 'mount_propagation': None,\n 'name': 'home',\n 'read_only': None,\n 'recursive_read_only': None,\n 'sub_path': None,\n 'sub_path_expr': None},\n {'mount_path': '/tmp',\n 'mount_propagation': None,\n 'name': 'tmp-dir',\n 'read_only': None,\n 'recursive_read_only': None,\n 'sub_path': None,\n 'sub_path_expr': None},\n {'mount_path': '/dev/shm',\n 'mount_propagation': None,\n 'name': 'dshm',\n 'read_only': None,\n 'recursive_read_only': None,\n 'sub_path': None,\n [e2e-llm-inference-service] 'sub_path_expr': None},\n {'mount_path': '/models',\n 'mount_propagation': None,\n 'name': 'model-cache',\n 'read_only': None,\n 'recursive_read_only': None,\n 'sub_path': None,\n 'sub_path_expr': None},\n {'mount_path': '/var/run/kserve/tls',\n 'mount_propagation': None,\n 'name': 'tls-certs',\n 'read_only': True,\n 'recursive_read_only': None,\n 'sub_path': None,\n 'sub_path_expr': None}],\n 'working_dir': None}],\n 'dns_config': None,\n 'dns_policy': 'ClusterFirst',\n 'enable_service_links': None,\n 'ephemeral_containers': None,\n 'host_aliases': None,\n 'host_ipc': None,\n 'host_network': None,\n 'host_pid': None,\n 'host_users': None,\n 'hostname': None,\n 'image_pull_secrets': None,\n 'init_containers': None,\n 'node_name': None,\n 'node_selector': None,\n 'os': None,\n 'overhead': None,\n 'preemption_policy': None,\n 'priority': None,\n 'priority_class_name': None,\n 'readiness_gates': None,\n 'resource_claims': None,\n 'resources': None,\n 'restart_policy': 'Always',\n 'runtime_class_name': None,\n 'scheduler_name': 'default-scheduler',\n 'scheduling_gates': None,\n 'security_context': {'app_armor_profile': None,\n 'fs_group': None,\n 'fs_group_change_policy': None,\n 'run_as_group': None,\n 'run_as_non_root': None,\n 'run_as_user': None,\n 'se_linux_change_policy': None,\n 'se_linux_options': None,\n 'seccomp_profile': None,\n 'supplemental_groups': None,\n 'supplemental_groups_policy': None,\n 'sysctls': None,\n 'windows_options': None},\n 'service_account': None,\n 'service_account_name': None,\n [e2e-llm-inference-service] 'set_hostname_as_fqdn': None,\n 'share_process_namespace': None,\n 'subdomain': None,\n 'termination_grace_period_seconds': 60,\n 'tolerations': None,\n 'topology_spread_constraints': None,\n 'volumes': [{'aws_elastic_block_store': None,\n 'azure_disk': None,\n 'azure_file': None,\n 'cephfs': None,\n 'cinder': None,\n 'config_map': None,\n 'csi': None,\n 'downward_api': None,\n 'empty_dir': {'medium': None,\n 'size_limit': None},\n 'ephemeral': None,\n 'fc': None,\n 'flex_volume': None,\n 'flocker': None,\n 'gce_persistent_disk': None,\n 'git_repo': None,\n 'glusterfs': None,\n 'host_path': None,\n 'image': None,\n 'iscsi': None,\n 'name': 'home',\n 'nfs': None,\n 'persistent_volume_claim': None,\n 'photon_persistent_disk': None,\n 'portworx_volume': None,\n 'projected': None,\n 'quobyte': None,\n 'rbd': None,\n 'scale_io': None,\n 'secret': None,\n 'storageos': None,\n 'vsphere_volume': None},\n {'aws_elastic_block_store': None,\n 'azure_disk': None,\n 'azure_file': None,\n 'cephfs': None,\n 'cinder': None,\n 'config_map': None,\n 'csi': None,\n 'downward_api': None,\n 'empty_dir': {'medium': 'Memory',\n 'size_limit': '1Gi'},\n 'ephemeral': None,\n 'fc': None,\n 'flex_volume': None,\n 'flocker': None,\n 'gce_persistent_disk': None,\n 'git_repo': None,\n 'glusterfs': None,\n 'host_path': None,\n 'image': None,\n 'iscsi': None,\n 'name': 'dshm',\n 'nfs': None,\n 'persistent_volume_claim': None,\n 'photon_persistent_disk': None,\n [e2e-llm-inference-service] 'portworx_volume': None,\n 'projected': None,\n 'quobyte': None,\n 'rbd': None,\n 'scale_io': None,\n 'secret': None,\n 'storageos': None,\n 'vsphere_volume': None},\n {'aws_elastic_block_store': None,\n 'azure_disk': None,\n 'azure_file': None,\n 'cephfs': None,\n 'cinder': None,\n 'config_map': None,\n 'csi': None,\n 'downward_api': None,\n 'empty_dir': {'medium': None,\n 'size_limit': None},\n 'ephemeral': None,\n 'fc': None,\n 'flex_volume': None,\n 'flocker': None,\n 'gce_persistent_disk': None,\n 'git_repo': None,\n 'glusterfs': None,\n 'host_path': None,\n 'image': None,\n 'iscsi': None,\n 'name': 'model-cache',\n 'nfs': None,\n 'persistent_volume_claim': None,\n 'photon_persistent_disk': None,\n 'portworx_volume': None,\n 'projected': None,\n 'quobyte': None,\n 'rbd': None,\n 'scale_io': None,\n 'secret': None,\n 'storageos': None,\n 'vsphere_volume': None},\n {'aws_elastic_block_store': None,\n 'azure_disk': None,\n 'azure_file': None,\n 'cephfs': None,\n 'cinder': None,\n 'config_map': None,\n 'csi': None,\n 'downward_api': None,\n 'empty_dir': {'medium': None,\n 'size_limit': None},\n 'ephemeral': None,\n 'fc': None,\n 'flex_volume': None,\n 'flocker': None,\n 'gce_persistent_disk': None,\n 'git_repo': None,\n 'glusterfs': None,\n 'host_path': None,\n 'image': None,\n 'iscsi': None,\n 'name': 'tmp-dir',\n 'nfs': None [e2e-llm-inference-service] ,\n 'persistent_volume_claim': None,\n 'photon_persistent_disk': None,\n 'portworx_volume': None,\n 'projected': None,\n 'quobyte': None,\n 'rbd': None,\n 'scale_io': None,\n 'secret': None,\n 'storageos': None,\n 'vsphere_volume': None},\n {'aws_elastic_block_store': None,\n 'azure_disk': None,\n 'azure_file': None,\n 'cephfs': None,\n 'cinder': None,\n 'config_map': None,\n 'csi': None,\n 'downward_api': None,\n 'empty_dir': None,\n 'ephemeral': None,\n 'fc': None,\n 'flex_volume': None,\n 'flocker': None,\n 'gce_persistent_disk': None,\n 'git_repo': None,\n 'glusterfs': None,\n 'host_path': None,\n 'image': None,\n 'iscsi': None,\n 'name': 'tls-certs',\n 'nfs': None,\n 'persistent_volume_claim': None,\n 'photon_persistent_disk': None,\n 'portworx_volume': None,\n 'projected': None,\n 'quobyte': None,\n 'rbd': None,\n 'scale_io': None,\n 'secret': {'default_mode': 420,\n 'items': None,\n 'optional': None,\n 'secret_name': 'rolling-upgrade-test-kserve-self-signed-certs'},\n 'storageos': None,\n 'vsphere_volume': None}]}}},\n 'status': {'available_replicas': 1,\n 'collision_count': None,\n 'conditions': [{'last_transition_time': datetime.datetime(2026, 7, 7, 15, 44, 39, tzinfo=tzlocal()),\n 'last_update_time': datetime.datetime(2026, 7, 7, 15, 44, 39, tzinfo=tzlocal()),\n 'message': 'Deployment has minimum availability.',\n 'reason': 'MinimumReplicasAvailable',\n 'status': 'True',\n 'type': 'Available'},\n {'last_transition_time': datetime.datetime(2026, 7, 7, 15, 44, 29, tzinfo=tzlocal()),\n 'last_update_time': datetime.datetime(2026, 7, 7, 15, 44, 39, tzinfo=tzlocal()),\n 'message': 'ReplicaSet '\n '"rolling-upgrade-test-kserve-54bc76f55b" '\n 'has successfully progressed.',\n 'reason': 'NewReplicaSetAvailable',\n 'status': 'True [e2e-llm-inference-service] ',\n 'type': 'Progressing'}],\n 'observed_generation': 1,\n 'ready_replicas': 1,\n 'replicas': 1,\n 'unavailable_replicas': None,\n 'updated_replicas': 1}}.metadata [e2e-llm-inference-service] INFO e2e.llmisvc.logging:logging.py:43 [wait_for_rollout_started] [2026-07-07T15:45:08.676674] end - ✅ in 2.043s [e2e-llm-inference-service] INFO e2e.llmisvc.logging:logging.py:34 [wait_for_deployment_available] [2026-07-07T15:45:08.676790] start - args=(, 'rolling-upgrade-test-kserve', 'kserve-ci-e2e-test'), kwargs={'timeout_seconds': 120} [e2e-llm-inference-service] INFO e2e.llmisvc.logging:logging.py:43 [wait_for_deployment_available] [2026-07-07T15:45:08.683132] end - ✅ in 0.006s [e2e-llm-inference-service] INFO e2e.llmisvc.logging:logging.py:34 [wait_for_llm_isvc_ready] [2026-07-07T15:45:08.683205] start - args=(, {'api_version': 'serving.kserve.io/v1alpha1', [e2e-llm-inference-service] 'kind': 'LLMInferenceService', [e2e-llm-inference-service] 'metadata': {'annotations': {'security.opendatahub.io/enable-auth': 'false'}, [e2e-llm-inference-service] 'creation_timestamp': None, [e2e-llm-inference-service] 'deletion_grace_period_seconds': None, [e2e-llm-inference-service] 'deletion_timestamp': None, [e2e-llm-inference-service] 'finalizers': None, [e2e-llm-inference-service] 'generate_name': None, [e2e-llm-inference-service] 'generation': None, [e2e-llm-inference-service] 'labels': None, [e2e-llm-inference-service] 'managed_fields': None, [e2e-llm-inference-service] 'name': 'rolling-upgrade-test', [e2e-llm-inference-service] 'namespace': 'kserve-ci-e2e-test', [e2e-llm-inference-service] 'owner_references': None, [e2e-llm-inference-service] 'resource_version': None, [e2e-llm-inference-service] 'self_link': None, [e2e-llm-inference-service] 'uid': None}, [e2e-llm-inference-service] 'spec': {'baseRefs': [{'name': 'router-managed-rolling-upgrade-64e041bd'}, [e2e-llm-inference-service] {'name': 'workload-llmd-simulator-rolling-9660aa23'}, [e2e-llm-inference-service] {'name': 'model-fb-opt-125m-rolling-upgra-30e65219'}]}, [e2e-llm-inference-service] 'status': None}, 900), kwargs={} [e2e-llm-inference-service] INFO e2e.llmisvc.logging:logging.py:43 [wait_for_llm_isvc_ready] [2026-07-07T15:45:08.690745] end - ✅ in 0.007s [e2e-llm-inference-service] INFO e2e.llmisvc.logging:logging.py:34 [wait_for_model_response] [2026-07-07T15:45:08.690808] start - args=(, TestCase(base_refs=['router-managed', 'workload-llmd-simulator', 'model-fb-opt-125m'], prompt='KServe is a', service_name='rolling-upgrade-test', endpoint='/v1/completions', max_tokens=20, payload_formatter=None, response_assertion=, wait_timeout=900, response_timeout=60, extra_headers=None, url_getter=None, expected_gateway=None, before_test=[], after_test=[], peers=[], llm_service={'api_version': 'serving.kserve.io/v1alpha1', [e2e-llm-inference-service] 'kind': 'LLMInferenceService', [e2e-llm-inference-service] 'metadata': {'annotations': {'security.opendatahub.io/enable-auth': 'false'}, [e2e-llm-inference-service] 'creation_timestamp': None, [e2e-llm-inference-service] 'deletion_grace_period_seconds': None, [e2e-llm-inference-service] 'deletion_timestamp': None, [e2e-llm-inference-service] 'finalizers': None, [e2e-llm-inference-service] 'generate_name': None, [e2e-llm-inference-service] 'generation': None, [e2e-llm-inference-service] 'labels': None, [e2e-llm-inference-service] 'managed_fields': None, [e2e-llm-inference-service] 'name': 'rolling-upgrade-test', [e2e-llm-inference-service] 'namespace': 'kserve-ci-e2e-test', [e2e-llm-inference-service] 'owner_references': None, [e2e-llm-inference-service] 'resource_version': None, [e2e-llm-inference-service] 'self_link': None, [e2e-llm-inference-service] 'uid': None}, [e2e-llm-inference-service] 'spec': {'baseRefs': [{'name': 'router-managed-rolling-upgrade-64e041bd'}, [e2e-llm-inference-service] {'name': 'workload-llmd-simulator-rolling-9660aa23'}, [e2e-llm-inference-service] {'name': 'model-fb-opt-125m-rolling-upgra-30e65219'}]}, [e2e-llm-inference-service] 'status': None}, model_name='facebook/opt-125m'), 900), kwargs={} [e2e-llm-inference-service] INFO e2e.llmisvc.logging:logging.py:34 [get_llm_service_url] [2026-07-07T15:45:08.691055] start - args=(, {'api_version': 'serving.kserve.io/v1alpha1', [e2e-llm-inference-service] 'kind': 'LLMInferenceService', [e2e-llm-inference-service] 'metadata': {'annotations': {'security.opendatahub.io/enable-auth': 'false'}, [e2e-llm-inference-service] 'creation_timestamp': None, [e2e-llm-inference-service] 'deletion_grace_period_seconds': None, [e2e-llm-inference-service] 'deletion_timestamp': None, [e2e-llm-inference-service] 'finalizers': None, [e2e-llm-inference-service] 'generate_name': None, [e2e-llm-inference-service] 'generation': None, [e2e-llm-inference-service] 'labels': None, [e2e-llm-inference-service] 'managed_fields': None, [e2e-llm-inference-service] 'name': 'rolling-upgrade-test', [e2e-llm-inference-service] 'namespace': 'kserve-ci-e2e-test', [e2e-llm-inference-service] 'owner_references': None, [e2e-llm-inference-service] 'resource_version': None, [e2e-llm-inference-service] 'self_link': None, [e2e-llm-inference-service] 'uid': None}, [e2e-llm-inference-service] 'spec': {'baseRefs': [{'name': 'router-managed-rolling-upgrade-64e041bd'}, [e2e-llm-inference-service] {'name': 'workload-llmd-simulator-rolling-9660aa23'}, [e2e-llm-inference-service] {'name': 'model-fb-opt-125m-rolling-upgra-30e65219'}]}, [e2e-llm-inference-service] 'status': None}), kwargs={} [e2e-llm-inference-service] INFO e2e.llmisvc.logging:logging.py:43 [get_llm_service_url] [2026-07-07T15:45:08.732208] end - ✅ in 0.041s [e2e-llm-inference-service] INFO e2e.llmisvc.logging:test_llm_inference_service.py:1092 Calling LLM service at http://ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com/kserve-ci-e2e-test/rolling-upgrade-test/v1/completions with payload {'model': 'facebook/opt-125m', 'prompt': 'KServe is a', 'max_tokens': 20} [e2e-llm-inference-service] WARNING urllib3.connectionpool:connectionpool.py:868 Retrying (Retry(total=7, connect=None, read=None, redirect=None, status=None)) after connection broken by 'ReadTimeoutError("HTTPConnectionPool(host='ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com', port=80): Read timed out. (read timeout=60)")': /kserve-ci-e2e-test/rolling-upgrade-test/v1/completions [e2e-llm-inference-service] WARNING urllib3.connectionpool:connectionpool.py:868 Retrying (Retry(total=6, connect=None, read=None, redirect=None, status=None)) after connection broken by 'ReadTimeoutError("HTTPConnectionPool(host='ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com', port=80): Read timed out. (read timeout=60)")': /kserve-ci-e2e-test/rolling-upgrade-test/v1/completions [e2e-llm-inference-service] WARNING urllib3.connectionpool:connectionpool.py:868 Retrying (Retry(total=5, connect=None, read=None, redirect=None, status=None)) after connection broken by 'ReadTimeoutError("HTTPConnectionPool(host='ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com', port=80): Read timed out. (read timeout=60)")': /kserve-ci-e2e-test/rolling-upgrade-test/v1/completions [e2e-llm-inference-service] WARNING urllib3.connectionpool:connectionpool.py:868 Retrying (Retry(total=4, connect=None, read=None, redirect=None, status=None)) after connection broken by 'ReadTimeoutError("HTTPConnectionPool(host='ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com', port=80): Read timed out. (read timeout=60)")': /kserve-ci-e2e-test/rolling-upgrade-test/v1/completions [e2e-llm-inference-service] WARNING urllib3.connectionpool:connectionpool.py:868 Retrying (Retry(total=3, connect=None, read=None, redirect=None, status=None)) after connection broken by 'ReadTimeoutError("HTTPConnectionPool(host='ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com', port=80): Read timed out. (read timeout=60)")': /kserve-ci-e2e-test/rolling-upgrade-test/v1/completions [e2e-llm-inference-service] WARNING urllib3.connectionpool:connectionpool.py:868 Retrying (Retry(total=2, connect=None, read=None, redirect=None, status=None)) after connection broken by 'ReadTimeoutError("HTTPConnectionPool(host='ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com', port=80): Read timed out. (read timeout=60)")': /kserve-ci-e2e-test/rolling-upgrade-test/v1/completions [e2e-llm-inference-service] WARNING urllib3.connectionpool:connectionpool.py:868 Retrying (Retry(total=1, connect=None, read=None, redirect=None, status=None)) after connection broken by 'ReadTimeoutError("HTTPConnectionPool(host='ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com', port=80): Read timed out. (read timeout=60)")': /kserve-ci-e2e-test/rolling-upgrade-test/v1/completions [e2e-llm-inference-service] WARNING urllib3.connectionpool:connectionpool.py:868 Retrying (Retry(total=0, connect=None, read=None, redirect=None, status=None)) after connection broken by 'ReadTimeoutError("HTTPConnectionPool(host='ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com', port=80): Read timed out. (read timeout=60)")': /kserve-ci-e2e-test/rolling-upgrade-test/v1/completions [e2e-llm-inference-service] ERROR e2e.llmisvc.logging:test_llm_inference_service.py:1108 ❌ Failed to call model: HTTPConnectionPool(host='ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com', port=80): Max retries exceeded with url: /kserve-ci-e2e-test/rolling-upgrade-test/v1/completions (Caused by ReadTimeoutError("HTTPConnectionPool(host='ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com', port=80): Read timed out. (read timeout=60)")) [e2e-llm-inference-service] ERROR e2e.llmisvc.logging:test_llm_inference_service.py:1219 Timed out waiting: ❌ Failed to call model: HTTPConnectionPool(host='ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com', port=80): Max retries exceeded with url: /kserve-ci-e2e-test/rolling-upgrade-test/v1/completions (Caused by ReadTimeoutError("HTTPConnectionPool(host='ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com', port=80): Read timed out. (read timeout=60)")) [e2e-llm-inference-service] ERROR e2e.llmisvc.logging:logging.py:48 [wait_for_model_response] [2026-07-07T16:00:13.347454] end - ❌ 904.656s: ❌ Failed to call model: HTTPConnectionPool(host='ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com', port=80): Max retries exceeded with url: /kserve-ci-e2e-test/rolling-upgrade-test/v1/completions (Caused by ReadTimeoutError("HTTPConnectionPool(host='ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com', port=80): Read timed out. (read timeout=60)")) [e2e-llm-inference-service] ERROR e2e.llmisvc.logging:logging.py:48 [test_rolling_upgrade_coordination] [2026-07-07T16:00:13.347558] end - ❌ 948.996s: ❌ Failed to call model: HTTPConnectionPool(host='ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com', port=80): Max retries exceeded with url: /kserve-ci-e2e-test/rolling-upgrade-test/v1/completions (Caused by ReadTimeoutError("HTTPConnectionPool(host='ae063886a26074181a63b39cd9058f0e-258571161.us-east-1.elb.amazonaws.com', port=80): Read timed out. (read timeout=60)")) [e2e-llm-inference-service] =============================== warnings summary =============================== [e2e-llm-inference-service] llmisvc/test_llm_inference_service.py::test_llm_inference_service[cluster_cpu-cluster_single_node-router-managed-scheduler-with-precise-prefix-cache-inline-config-workload-llmd-simulator-kvcache] [e2e-llm-inference-service] llmisvc/test_llm_inference_service.py::test_llm_inference_service[cluster_cpu-cluster_single_node-router-with-gateway-ref-router-with-managed-route-model-fb-opt-125m-workload-llmd-simulator] [e2e-llm-inference-service] /workspace/source/python/kserve/.venv/lib64/python3.11/site-packages/pytest_asyncio/plugin.py:761: DeprecationWarning: The event_loop fixture provided by pytest-asyncio has been redefined in [e2e-llm-inference-service] /workspace/source/test/e2e/conftest.py:43 [e2e-llm-inference-service] Replacing the event_loop fixture with a custom implementation is deprecated [e2e-llm-inference-service] and will lead to errors in the future. [e2e-llm-inference-service] If you want to request an asyncio event loop with a scope other than function [e2e-llm-inference-service] scope, use the "scope" argument to the asyncio mark when marking the tests. [e2e-llm-inference-service] If you want to return different types of event loops, use the event_loop_policy [e2e-llm-inference-service] fixture. [e2e-llm-inference-service] [e2e-llm-inference-service] warnings.warn( [e2e-llm-inference-service] [e2e-llm-inference-service] llmisvc/test_llm_inference_service.py::test_llm_inference_service[cluster_cpu-cluster_single_node-router-managed-scheduler-with-precise-prefix-cache-inline-config-workload-llmd-simulator-kvcache] [e2e-llm-inference-service] llmisvc/test_llm_inference_service.py:245: PytestWarning: The test is marked with '@pytest.mark.asyncio' but it is not an async function. Please remove the asyncio mark. If the test is not marked explicitly, check for global marks applied via 'pytestmark'. [e2e-llm-inference-service] @pytest.mark.llminferenceservice [e2e-llm-inference-service] [e2e-llm-inference-service] llmisvc/test_llm_inference_service.py::test_llm_inference_service[cluster_cpu-cluster_single_node-router-managed-workload-llmd-simulator0] [e2e-llm-inference-service] llmisvc/test_llm_inference_service.py:245: PytestWarning: The test is marked with '@pytest.mark.asyncio' but it is not an async function. Please remove the asyncio mark. If the test is not marked explicitly, check for global marks applied via 'pytestmark'. [e2e-llm-inference-service] @pytest.mark.llminferenceservice [e2e-llm-inference-service] [e2e-llm-inference-service] llmisvc/test_llm_inference_service.py::test_llm_inference_service[cluster_cpu-cluster_single_node-router-managed-workload-llmd-simulator1] [e2e-llm-inference-service] llmisvc/test_llm_inference_service.py:245: PytestWarning: The test is marked with '@pytest.mark.asyncio' but it is not an async function. Please remove the asyncio mark. If the test is not marked explicitly, check for global marks applied via 'pytestmark'. [e2e-llm-inference-service] @pytest.mark.llminferenceservice [e2e-llm-inference-service] [e2e-llm-inference-service] llmisvc/test_llm_inference_service.py::test_llm_inference_service[cluster_cpu-cluster_single_node-router-managed-workload-llmd-simulator2] [e2e-llm-inference-service] llmisvc/test_llm_inference_service.py:245: PytestWarning: The test is marked with '@pytest.mark.asyncio' but it is not an async function. Please remove the asyncio mark. If the test is not marked explicitly, check for global marks applied via 'pytestmark'. [e2e-llm-inference-service] @pytest.mark.llminferenceservice [e2e-llm-inference-service] [e2e-llm-inference-service] llmisvc/test_llm_inference_service.py::test_llm_inference_service[cluster_cpu-cluster_single_node-router-with-gateway-ref-router-with-managed-route-model-fb-opt-125m-workload-llmd-simulator] [e2e-llm-inference-service] llmisvc/test_llm_inference_service.py:245: PytestWarning: The test is marked with '@pytest.mark.asyncio' but it is not an async function. Please remove the asyncio mark. If the test is not marked explicitly, check for global marks applied via 'pytestmark'. [e2e-llm-inference-service] @pytest.mark.llminferenceservice [e2e-llm-inference-service] [e2e-llm-inference-service] llmisvc/test_llm_inference_service.py::test_llm_inference_service[cluster_cpu-cluster_single_node-router-managed-workload-single-cpu-model-fb-opt-125m-with-lora-hf0] [e2e-llm-inference-service] llmisvc/test_llm_inference_service.py:245: PytestWarning: The test is marked with '@pytest.mark.asyncio' but it is not an async function. Please remove the asyncio mark. If the test is not marked explicitly, check for global marks applied via 'pytestmark'. [e2e-llm-inference-service] @pytest.mark.llminferenceservice [e2e-llm-inference-service] [e2e-llm-inference-service] llmisvc/test_llm_inference_service.py::test_llm_inference_service[cluster_cpu-cluster_single_node-router-managed-workload-single-cpu-model-fb-opt-125m] [e2e-llm-inference-service] llmisvc/test_llm_inference_service.py:245: PytestWarning: The test is marked with '@pytest.mark.asyncio' but it is not an async function. Please remove the asyncio mark. If the test is not marked explicitly, check for global marks applied via 'pytestmark'. [e2e-llm-inference-service] @pytest.mark.llminferenceservice [e2e-llm-inference-service] [e2e-llm-inference-service] llmisvc/test_llm_inference_service.py::test_llm_inference_service[cluster_cpu-cluster_single_node-router-managed-workload-single-cpu-model-fb-opt-125m-with-lora-hf1] [e2e-llm-inference-service] llmisvc/test_llm_inference_service.py:245: PytestWarning: The test is marked with '@pytest.mark.asyncio' but it is not an async function. Please remove the asyncio mark. If the test is not marked explicitly, check for global marks applied via 'pytestmark'. [e2e-llm-inference-service] @pytest.mark.llminferenceservice [e2e-llm-inference-service] [e2e-llm-inference-service] llmisvc/test_llm_inference_service.py::test_llm_inference_service[cluster_cpu-cluster_single_node-router-custom-route-timeout-scheduler-managed-workload-single-cpu-model-fb-opt-125m] [e2e-llm-inference-service] llmisvc/test_llm_inference_service.py:245: PytestWarning: The test is marked with '@pytest.mark.asyncio' but it is not an async function. Please remove the asyncio mark. If the test is not marked explicitly, check for global marks applied via 'pytestmark'. [e2e-llm-inference-service] @pytest.mark.llminferenceservice [e2e-llm-inference-service] [e2e-llm-inference-service] llmisvc/test_llm_inference_service.py::test_llm_inference_service[cluster_cpu-cluster_single_node-router-with-refs-scheduler-managed-workload-single-cpu-model-fb-opt-125m] [e2e-llm-inference-service] llmisvc/test_llm_inference_service.py:245: PytestWarning: The test is marked with '@pytest.mark.asyncio' but it is not an async function. Please remove the asyncio mark. If the test is not marked explicitly, check for global marks applied via 'pytestmark'. [e2e-llm-inference-service] @pytest.mark.llminferenceservice [e2e-llm-inference-service] [e2e-llm-inference-service] llmisvc/test_llm_inference_service.py::test_llm_inference_service[cluster_cpu-cluster_single_node-router-managed-workload-single-cpu-model-pvc] [e2e-llm-inference-service] llmisvc/test_llm_inference_service.py:245: PytestWarning: The test is marked with '@pytest.mark.asyncio' but it is not an async function. Please remove the asyncio mark. If the test is not marked explicitly, check for global marks applied via 'pytestmark'. [e2e-llm-inference-service] @pytest.mark.llminferenceservice [e2e-llm-inference-service] [e2e-llm-inference-service] llmisvc/test_llm_inference_service.py::test_llm_inference_service[cluster_cpu-cluster_single_node-router-managed-workload-pd-cpu-model-pvc] [e2e-llm-inference-service] llmisvc/test_llm_inference_service.py:245: PytestWarning: The test is marked with '@pytest.mark.asyncio' but it is not an async function. Please remove the asyncio mark. If the test is not marked explicitly, check for global marks applied via 'pytestmark'. [e2e-llm-inference-service] @pytest.mark.llminferenceservice [e2e-llm-inference-service] [e2e-llm-inference-service] llmisvc/test_llm_inference_service.py::test_llm_inference_service[cluster_cpu-cluster_single_node-router-managed-workload-pd-cpu-model-fb-opt-125m] [e2e-llm-inference-service] llmisvc/test_llm_inference_service.py:245: PytestWarning: The test is marked with '@pytest.mark.asyncio' but it is not an async function. Please remove the asyncio mark. If the test is not marked explicitly, check for global marks applied via 'pytestmark'. [e2e-llm-inference-service] @pytest.mark.llminferenceservice [e2e-llm-inference-service] [e2e-llm-inference-service] llmisvc/test_llm_inference_service.py::test_llm_inference_service[cluster_cpu-cluster_single_node-router-custom-route-timeout-pd-scheduler-managed-workload-pd-cpu-model-fb-opt-125m] [e2e-llm-inference-service] llmisvc/test_llm_inference_service.py:245: PytestWarning: The test is marked with '@pytest.mark.asyncio' but it is not an async function. Please remove the asyncio mark. If the test is not marked explicitly, check for global marks applied via 'pytestmark'. [e2e-llm-inference-service] @pytest.mark.llminferenceservice [e2e-llm-inference-service] [e2e-llm-inference-service] llmisvc/test_llm_inference_service.py::test_llm_inference_service[cluster_cpu-cluster_multi_node-router-managed-workload-simulated-dp-ep-cpu-model-pvc] [e2e-llm-inference-service] llmisvc/test_llm_inference_service.py:245: PytestWarning: The test is marked with '@pytest.mark.asyncio' but it is not an async function. Please remove the asyncio mark. If the test is not marked explicitly, check for global marks applied via 'pytestmark'. [e2e-llm-inference-service] @pytest.mark.llminferenceservice [e2e-llm-inference-service] [e2e-llm-inference-service] llmisvc/test_llm_inference_service.py::test_llm_inference_service[cluster_cpu-cluster_single_node-router-with-refs-pd-scheduler-managed-workload-pd-cpu-model-fb-opt-125m] [e2e-llm-inference-service] llmisvc/test_llm_inference_service.py:245: PytestWarning: The test is marked with '@pytest.mark.asyncio' but it is not an async function. Please remove the asyncio mark. If the test is not marked explicitly, check for global marks applied via 'pytestmark'. [e2e-llm-inference-service] @pytest.mark.llminferenceservice [e2e-llm-inference-service] [e2e-llm-inference-service] llmisvc/test_llm_inference_service.py::test_llm_inference_service[cluster_cpu-cluster_single_node-router-no-scheduler-workload-single-cpu-model-fb-opt-125m] [e2e-llm-inference-service] llmisvc/test_llm_inference_service.py:245: PytestWarning: The test is marked with '@pytest.mark.asyncio' but it is not an async function. Please remove the asyncio mark. If the test is not marked explicitly, check for global marks applied via 'pytestmark'. [e2e-llm-inference-service] @pytest.mark.llminferenceservice [e2e-llm-inference-service] [e2e-llm-inference-service] llmisvc/test_llm_inference_service_stop.py::test_llm_stop_feature[cluster_cpu-cluster_single_node-router-managed-workload-single-cpu-model-fb-opt-125m] [e2e-llm-inference-service] llmisvc/test_llm_inference_service_stop.py:40: PytestWarning: The test is marked with '@pytest.mark.asyncio' but it is not an async function. Please remove the asyncio mark. If the test is not marked explicitly, check for global marks applied via 'pytestmark'. [e2e-llm-inference-service] @pytest.mark.llminferenceservice [e2e-llm-inference-service] [e2e-llm-inference-service] llmisvc/test_llm_inference_service.py::test_llm_inference_service[cluster_cpu-cluster_multi_node-router-managed-workload-simulated-dp-ep-cpu-model-fb-opt-125m] [e2e-llm-inference-service] llmisvc/test_llm_inference_service.py:245: PytestWarning: The test is marked with '@pytest.mark.asyncio' but it is not an async function. Please remove the asyncio mark. If the test is not marked explicitly, check for global marks applied via 'pytestmark'. [e2e-llm-inference-service] @pytest.mark.llminferenceservice [e2e-llm-inference-service] [e2e-llm-inference-service] llmisvc/test_llm_inference_service.py::test_llm_inference_service[cluster_cpu-cluster_single_node-router-managed-scheduler-with-inline-config-workload-llmd-simulator] [e2e-llm-inference-service] llmisvc/test_llm_inference_service.py:245: PytestWarning: The test is marked with '@pytest.mark.asyncio' but it is not an async function. Please remove the asyncio mark. If the test is not marked explicitly, check for global marks applied via 'pytestmark'. [e2e-llm-inference-service] @pytest.mark.llminferenceservice [e2e-llm-inference-service] [e2e-llm-inference-service] llmisvc/test_llm_inference_service.py::test_llm_inference_service[cluster_cpu-cluster_single_node-router-managed-workload-llmd-simulator-model-qwen2.5-0.5b] [e2e-llm-inference-service] llmisvc/test_llm_inference_service.py:245: PytestWarning: The test is marked with '@pytest.mark.asyncio' but it is not an async function. Please remove the asyncio mark. If the test is not marked explicitly, check for global marks applied via 'pytestmark'. [e2e-llm-inference-service] @pytest.mark.llminferenceservice [e2e-llm-inference-service] [e2e-llm-inference-service] llmisvc/test_llm_inference_service.py::test_llm_inference_service[cluster_cpu-cluster_single_node-router-managed-scheduler-with-configmap-ref-workload-llmd-simulator] [e2e-llm-inference-service] llmisvc/test_llm_inference_service.py:245: PytestWarning: The test is marked with '@pytest.mark.asyncio' but it is not an async function. Please remove the asyncio mark. If the test is not marked explicitly, check for global marks applied via 'pytestmark'. [e2e-llm-inference-service] @pytest.mark.llminferenceservice [e2e-llm-inference-service] [e2e-llm-inference-service] llmisvc/test_llm_tls.py::test_llm_tls_resources[cluster_cpu-cluster_single_node-router-managed-workload-single-cpu-model-fb-opt-125m] [e2e-llm-inference-service] llmisvc/test_llm_tls.py:93: PytestWarning: The test is marked with '@pytest.mark.asyncio' but it is not an async function. Please remove the asyncio mark. If the test is not marked explicitly, check for global marks applied via 'pytestmark'. [e2e-llm-inference-service] @pytest.mark.llminferenceservice [e2e-llm-inference-service] [e2e-llm-inference-service] llmisvc/test_llm_inference_service.py::test_llm_inference_service[cluster_cpu-cluster_single_node-router-managed-scheduler-with-replicas-workload-llmd-simulator] [e2e-llm-inference-service] llmisvc/test_llm_inference_service.py:245: PytestWarning: The test is marked with '@pytest.mark.asyncio' but it is not an async function. Please remove the asyncio mark. If the test is not marked explicitly, check for global marks applied via 'pytestmark'. [e2e-llm-inference-service] @pytest.mark.llminferenceservice [e2e-llm-inference-service] [e2e-llm-inference-service] llmisvc/test_llm_inference_service.py::test_llm_inference_service[cluster_cpu-cluster_single_node-router-managed-scheduler-with-custom-template-workload-llmd-simulator] [e2e-llm-inference-service] llmisvc/test_llm_inference_service.py:245: PytestWarning: The test is marked with '@pytest.mark.asyncio' but it is not an async function. Please remove the asyncio mark. If the test is not marked explicitly, check for global marks applied via 'pytestmark'. [e2e-llm-inference-service] @pytest.mark.llminferenceservice [e2e-llm-inference-service] [e2e-llm-inference-service] llmisvc/test_llm_inference_service.py::test_llm_inference_service[cluster_cpu-cluster_single_node-router-managed-scheduler-v06-pd-config-migration-workload-llmd-simulator-pd] [e2e-llm-inference-service] llmisvc/test_llm_inference_service.py:245: PytestWarning: The test is marked with '@pytest.mark.asyncio' but it is not an async function. Please remove the asyncio mark. If the test is not marked explicitly, check for global marks applied via 'pytestmark'. [e2e-llm-inference-service] @pytest.mark.llminferenceservice [e2e-llm-inference-service] [e2e-llm-inference-service] llmisvc/test_llm_inference_service.py::test_llm_inference_service[cluster_cpu-cluster_single_node-router-managed-scheduler-v06-nonzero-threshold-migration-workload-llmd-simulator-pd] [e2e-llm-inference-service] llmisvc/test_llm_inference_service.py:245: PytestWarning: The test is marked with '@pytest.mark.asyncio' but it is not an async function. Please remove the asyncio mark. If the test is not marked explicitly, check for global marks applied via 'pytestmark'. [e2e-llm-inference-service] @pytest.mark.llminferenceservice [e2e-llm-inference-service] [e2e-llm-inference-service] -- Docs: https://docs.pytest.org/en/stable/how-to/capture-warnings.html [e2e-llm-inference-service] ---------- generated xml file: /workspace/artifacts-dir/junit_e2e.xml ---------- [e2e-llm-inference-service] --------------------------------- JSON report ---------------------------------- [e2e-llm-inference-service] report saved to: /workspace/artifacts-dir/e2e_results.json [e2e-llm-inference-service] =========================== short test summary info ============================ [e2e-llm-inference-service] FAILED llmisvc/test_llm_inference_service.py::test_llm_inference_service[cluster_cpu-cluster_single_node-router-managed-workload-single-cpu-model-fb-opt-125m-with-lora-hf0] [e2e-llm-inference-service] FAILED llmisvc/test_llm_inference_service.py::test_llm_inference_service[cluster_cpu-cluster_single_node-router-managed-workload-single-cpu-model-fb-opt-125m] [e2e-llm-inference-service] FAILED llmisvc/test_llm_inference_service.py::test_llm_inference_service[cluster_cpu-cluster_single_node-router-managed-workload-single-cpu-model-fb-opt-125m-with-lora-hf1] [e2e-llm-inference-service] FAILED llmisvc/test_llm_inference_service.py::test_llm_inference_service[cluster_cpu-cluster_single_node-router-custom-route-timeout-scheduler-managed-workload-single-cpu-model-fb-opt-125m] [e2e-llm-inference-service] FAILED llmisvc/test_rolling_upgrade.py::test_rolling_upgrade_coordination[cluster_cpu-cluster_single_node-router-managed-workload-llmd-simulator-model-fb-opt-125m] [e2e-llm-inference-service] !!!!!!!!!!!!!!!!!!!!!!!!!! stopping after 5 failures !!!!!!!!!!!!!!!!!!!!!!!!!!! [e2e-llm-inference-service] !!!!!!!!!!!! xdist.dsession.Interrupted: stopping after 5 failures !!!!!!!!!!!!! [e2e-llm-inference-service] ====== 5 failed, 37 passed, 3 skipped, 29 warnings in 5920.97s (1:38:40) ======= [must-gather] [must-gather ] OUT 2026-07-07T16:00:14.840019738Z Using must-gather plug-in image: quay.io/modh/must-gather:rhoai-2.24 [must-gather] When opening a support case, bugzilla, or issue please include the following summary data along with any other requested information: [must-gather] ClusterID: b54a7a34-f018-4011-8b3a-987919d37ffa [must-gather] ClientVersion: 4.21.10 [must-gather] ClusterVersion: Stable at "4.21.23" [must-gather] ClusterOperators: [must-gather] clusteroperator/authentication is missing [must-gather] clusteroperator/cloud-credential is missing [must-gather] clusteroperator/cluster-autoscaler is missing [must-gather] clusteroperator/config-operator is missing [must-gather] clusteroperator/etcd is missing [must-gather] clusteroperator/machine-api is missing [must-gather] clusteroperator/machine-approver is missing [must-gather] clusteroperator/machine-config is missing [must-gather] clusteroperator/marketplace is missing [must-gather] [must-gather] [must-gather] [must-gather ] OUT 2026-07-07T16:00:14.892367134Z namespace/openshift-must-gather-h6gn8 created [must-gather] [must-gather ] OUT 2026-07-07T16:00:14.897715142Z clusterrolebinding.rbac.authorization.k8s.io/must-gather-mfcfv created [must-gather] [must-gather ] OUT 2026-07-07T16:00:14.929920166Z pod for plug-in image quay.io/modh/must-gather:rhoai-2.24 created [must-gather] [must-gather-dnqjn] POD 2026-07-07T16:00:22.044287559Z [disk usage checker] Started [must-gather] [must-gather-dnqjn] POD 2026-07-07T16:00:22.047506332Z [disk usage checker] Volume usage percentage: current = 16 ; allowed = 70 [must-gather] [must-gather-dnqjn] POD 2026-07-07T16:00:22.312347697Z Error from server (NotFound): namespaces "redhat-ods-operator" not found [must-gather] [must-gather-dnqjn] POD 2026-07-07T16:00:22.483039706Z Wrote inspect data to must-gather. [must-gather] [must-gather-dnqjn] POD 2026-07-07T16:00:22.483071611Z error: inspection completed with the errors occurred while gathering data: [must-gather] [must-gather-dnqjn] POD 2026-07-07T16:00:22.483071611Z namespaces "redhat-ods-operator" not found [must-gather] [must-gather-dnqjn] POD 2026-07-07T16:00:22.485531801Z Error getting logs from redhat-ods-operator [must-gather] [must-gather-dnqjn] POD 2026-07-07T16:00:22.649852129Z Wrote inspect data to must-gather. [must-gather] [must-gather-dnqjn] POD 2026-07-07T16:00:22.649883345Z error: inspection completed with the errors occurred while gathering data: [must-gather] [must-gather-dnqjn] POD 2026-07-07T16:00:22.649883345Z namespaces "redhat-ods-monitoring" not found [must-gather] [must-gather-dnqjn] POD 2026-07-07T16:00:22.651929658Z Error getting logs from redhat-ods-monitoring [must-gather] [must-gather-dnqjn] POD 2026-07-07T16:00:22.823924224Z Wrote inspect data to must-gather. [must-gather] [must-gather-dnqjn] POD 2026-07-07T16:00:22.823977265Z error: inspection completed with the errors occurred while gathering data: [must-gather] [must-gather-dnqjn] POD 2026-07-07T16:00:22.823977265Z namespaces "redhat-ods-applications" not found [must-gather] [must-gather-dnqjn] POD 2026-07-07T16:00:22.826267645Z Error getting logs from redhat-ods-applications [must-gather] [must-gather-dnqjn] POD 2026-07-07T16:00:22.991350259Z Wrote inspect data to must-gather. [must-gather] [must-gather-dnqjn] POD 2026-07-07T16:00:22.991445589Z error: inspection completed with the errors occurred while gathering data: [must-gather] [must-gather-dnqjn] POD 2026-07-07T16:00:22.991445589Z namespaces "rhods-notebooks" not found [must-gather] [must-gather-dnqjn] POD 2026-07-07T16:00:22.994695461Z Error getting logs from rhods-notebooks [must-gather] [must-gather-dnqjn] POD 2026-07-07T16:00:23.157351491Z Wrote inspect data to must-gather. [must-gather] [must-gather-dnqjn] POD 2026-07-07T16:00:23.157394197Z error: inspection completed with the errors occurred while gathering data: [must-gather] [must-gather-dnqjn] POD 2026-07-07T16:00:23.157394197Z namespaces "rhoai-model-registries" not found [must-gather] [must-gather-dnqjn] POD 2026-07-07T16:00:23.159555100Z Error getting logs from rhoai-model-registries [must-gather] [must-gather-dnqjn] POD 2026-07-07T16:00:23.322352191Z Wrote inspect data to must-gather. [must-gather] [must-gather-dnqjn] POD 2026-07-07T16:00:23.322423614Z error: inspection completed with the errors occurred while gathering data: [must-gather] [must-gather-dnqjn] POD 2026-07-07T16:00:23.322423614Z namespaces "istio-system" not found [must-gather] [must-gather-dnqjn] POD 2026-07-07T16:00:23.324668018Z Error getting logs from istio-system [must-gather] [must-gather-dnqjn] POD 2026-07-07T16:00:23.489798437Z Wrote inspect data to must-gather. [must-gather] [must-gather-dnqjn] POD 2026-07-07T16:00:23.489828935Z error: inspection completed with the errors occurred while gathering data: [must-gather] [must-gather-dnqjn] POD 2026-07-07T16:00:23.489828935Z namespaces "knative-serving" not found [must-gather] [must-gather-dnqjn] POD 2026-07-07T16:00:23.492421957Z Error getting logs from knative-serving [must-gather] [must-gather-dnqjn] POD 2026-07-07T16:00:23.661519559Z Wrote inspect data to must-gather. [must-gather] [must-gather-dnqjn] POD 2026-07-07T16:00:23.661552297Z error: inspection completed with the errors occurred while gathering data: [must-gather] [must-gather-dnqjn] POD 2026-07-07T16:00:23.661552297Z namespaces "redhat-ods-applications-auth-provider" not found [must-gather] [must-gather-dnqjn] POD 2026-07-07T16:00:23.664078264Z Error getting logs from redhat-ods-applications-auth-provider [must-gather] [must-gather-dnqjn] POD 2026-07-07T16:00:23.964976520Z Wrote inspect data to must-gather. [must-gather] [must-gather-dnqjn] POD 2026-07-07T16:00:24.250800401Z Wrote inspect data to must-gather. [must-gather] [must-gather-dnqjn] POD 2026-07-07T16:00:24.393784189Z error: the server doesn't have a resource type "auths" [must-gather] [must-gather-dnqjn] POD 2026-07-07T16:00:24.476124540Z error: arguments in resource/name form must have a single resource and name [must-gather] [must-gather-dnqjn] POD 2026-07-07T16:00:24.478303414Z Error collecting info from [must-gather] [must-gather-dnqjn] POD 2026-07-07T16:00:24.615043091Z error: the server doesn't have a resource type "monitorings" [must-gather] [must-gather-dnqjn] POD 2026-07-07T16:00:24.697211289Z error: arguments in resource/name form must have a single resource and name [must-gather] [must-gather-dnqjn] POD 2026-07-07T16:00:24.699238293Z Error collecting info from [must-gather] [must-gather-dnqjn] POD 2026-07-07T16:00:24.833096017Z error: the server doesn't have a resource type "featuretrackers" [must-gather] [must-gather-dnqjn] POD 2026-07-07T16:00:24.914434314Z error: arguments in resource/name form must have a single resource and name [must-gather] [must-gather-dnqjn] POD 2026-07-07T16:00:24.916414851Z Error collecting info from [must-gather] [must-gather-dnqjn] POD 2026-07-07T16:00:25.055196406Z error: the server doesn't have a resource type "codeflares" [must-gather] [must-gather-dnqjn] POD 2026-07-07T16:00:25.139403759Z error: arguments in resource/name form must have a single resource and name [must-gather] [must-gather-dnqjn] POD 2026-07-07T16:00:25.141858676Z Error collecting info from [must-gather] [must-gather-dnqjn] POD 2026-07-07T16:00:25.276190826Z error: the server doesn't have a resource type "dashboards" [must-gather] [must-gather-dnqjn] POD 2026-07-07T16:00:25.359832148Z error: arguments in resource/name form must have a single resource and name [must-gather] [must-gather-dnqjn] POD 2026-07-07T16:00:25.362034543Z Error collecting info from [must-gather] [must-gather-dnqjn] POD 2026-07-07T16:00:25.498042940Z error: the server doesn't have a resource type "datasciencepipelines" [must-gather] [must-gather-dnqjn] POD 2026-07-07T16:00:25.578568022Z error: arguments in resource/name form must have a single resource and name [must-gather] [must-gather-dnqjn] POD 2026-07-07T16:00:25.580826113Z Error collecting info from [must-gather] [must-gather-dnqjn] POD 2026-07-07T16:00:25.718023957Z error: the server doesn't have a resource type "feastoperators" [must-gather] [must-gather-dnqjn] POD 2026-07-07T16:00:25.799752832Z error: arguments in resource/name form must have a single resource and name [must-gather] [must-gather-dnqjn] POD 2026-07-07T16:00:25.801907102Z Error collecting info from [must-gather] [must-gather-dnqjn] POD 2026-07-07T16:00:25.943304958Z error: the server doesn't have a resource type "kserves" [must-gather] [must-gather-dnqjn] POD 2026-07-07T16:00:26.027388084Z error: arguments in resource/name form must have a single resource and name [must-gather] [must-gather-dnqjn] POD 2026-07-07T16:00:26.029555483Z Error collecting info from [must-gather] [must-gather-dnqjn] POD 2026-07-07T16:00:26.164540023Z error: the server doesn't have a resource type "kueues" [must-gather] [must-gather-dnqjn] POD 2026-07-07T16:00:26.245811179Z error: arguments in resource/name form must have a single resource and name [must-gather] [must-gather-dnqjn] POD 2026-07-07T16:00:26.247900387Z Error collecting info from [must-gather] [must-gather-dnqjn] POD 2026-07-07T16:00:26.378984435Z error: the server doesn't have a resource type "modelcontrollers" [must-gather] [must-gather-dnqjn] POD 2026-07-07T16:00:26.460520279Z error: arguments in resource/name form must have a single resource and name [must-gather] [must-gather-dnqjn] POD 2026-07-07T16:00:26.462752646Z Error collecting info from [must-gather] [must-gather-dnqjn] POD 2026-07-07T16:00:26.593064499Z error: the server doesn't have a resource type "modelmeshservings" [must-gather] [must-gather-dnqjn] POD 2026-07-07T16:00:26.681913587Z error: arguments in resource/name form must have a single resource and name [must-gather] [must-gather-dnqjn] POD 2026-07-07T16:00:26.684523650Z Error collecting info from [must-gather] [must-gather-dnqjn] POD 2026-07-07T16:00:26.814102781Z error: the server doesn't have a resource type "modelregistries" [must-gather] [must-gather-dnqjn] POD 2026-07-07T16:00:26.900062745Z error: arguments in resource/name form must have a single resource and name [must-gather] [must-gather-dnqjn] POD 2026-07-07T16:00:26.902483300Z Error collecting info from [must-gather] [must-gather-dnqjn] POD 2026-07-07T16:00:27.036045873Z error: the server doesn't have a resource type "rays" [must-gather] [must-gather-dnqjn] POD 2026-07-07T16:00:27.051860557Z [disk usage checker] Volume usage percentage: current = 16 ; allowed = 70 [must-gather] [must-gather-dnqjn] POD 2026-07-07T16:00:27.121002183Z error: arguments in resource/name form must have a single resource and name [must-gather] [must-gather-dnqjn] POD 2026-07-07T16:00:27.123333746Z Error collecting info from [must-gather] [must-gather-dnqjn] POD 2026-07-07T16:00:27.252426136Z error: the server doesn't have a resource type "trainingoperators" [must-gather] [must-gather-dnqjn] POD 2026-07-07T16:00:27.336157254Z error: arguments in resource/name form must have a single resource and name [must-gather] [must-gather-dnqjn] POD 2026-07-07T16:00:27.338293945Z Error collecting info from [must-gather] [must-gather-dnqjn] POD 2026-07-07T16:00:27.473865518Z error: the server doesn't have a resource type "trustyais" [must-gather] [must-gather-dnqjn] POD 2026-07-07T16:00:27.555858577Z error: arguments in resource/name form must have a single resource and name [must-gather] [must-gather-dnqjn] POD 2026-07-07T16:00:27.558353870Z Error collecting info from [must-gather] [must-gather-dnqjn] POD 2026-07-07T16:00:27.686634109Z error: the server doesn't have a resource type "workbenches" [must-gather] [must-gather-dnqjn] POD 2026-07-07T16:00:27.774555377Z error: arguments in resource/name form must have a single resource and name [must-gather] [must-gather-dnqjn] POD 2026-07-07T16:00:27.776579967Z Error collecting info from [must-gather] [must-gather-dnqjn] POD 2026-07-07T16:00:27.910436599Z error: the server doesn't have a resource type "hardwareprofiles" [must-gather] [must-gather-dnqjn] POD 2026-07-07T16:00:27.992933172Z error: arguments in resource/name form must have a single resource and name [must-gather] [must-gather-dnqjn] POD 2026-07-07T16:00:27.995120311Z Error collecting info from [must-gather] [must-gather-dnqjn] POD 2026-07-07T16:00:28.128219425Z error: the server doesn't have a resource type "llamastackoperators" [must-gather] [must-gather-dnqjn] POD 2026-07-07T16:00:28.208878387Z error: arguments in resource/name form must have a single resource and name [must-gather] [must-gather-dnqjn] POD 2026-07-07T16:00:28.211244686Z Error collecting info from [must-gather] [must-gather-dnqjn] POD 2026-07-07T16:00:28.924050407Z error: the server doesn't have a resource type "predictors" [must-gather] [must-gather-dnqjn] POD 2026-07-07T16:00:29.056726569Z error: the server doesn't have a resource type "localmodelnodegroups" [must-gather] [must-gather-dnqjn] POD 2026-07-07T16:00:29.184579704Z error: the server doesn't have a resource type "smcp" [must-gather] [must-gather-dnqjn] POD 2026-07-07T16:00:29.314341529Z error: the server doesn't have a resource type "smm" [must-gather] [must-gather-dnqjn] POD 2026-07-07T16:00:29.447971048Z error: the server doesn't have a resource type "smmr" [must-gather] [must-gather-dnqjn] POD 2026-07-07T16:00:30.070035501Z error: the server doesn't have a resource type "knativeservings" [must-gather] [must-gather-dnqjn] POD 2026-07-07T16:00:30.202047237Z error: the server doesn't have a resource type "configurations" [must-gather] [must-gather-dnqjn] POD 2026-07-07T16:00:30.333042330Z error: the server doesn't have a resource type "routes" [must-gather] [must-gather-dnqjn] POD 2026-07-07T16:00:30.468762647Z error: the server doesn't have a resource type "services" [must-gather] [must-gather-dnqjn] POD 2026-07-07T16:00:30.596021702Z error: the server doesn't have a resource type "revisions" [must-gather] [must-gather-dnqjn] POD 2026-07-07T16:00:31.179265915Z Gathering data for ns/kserve-ci-e2e-test... [must-gather] [must-gather-dnqjn] POD 2026-07-07T16:00:32.056517960Z [disk usage checker] Volume usage percentage: current = 16 ; allowed = 70 [must-gather] [must-gather-dnqjn] POD 2026-07-07T16:00:32.427674990Z Warning: apps.openshift.io/v1 DeploymentConfig is deprecated in v4.14+, unavailable in v4.10000+ [must-gather] [must-gather-dnqjn] POD 2026-07-07T16:00:33.216473529Z Warning: v1 Endpoints is deprecated in v1.33+; use discovery.k8s.io/v1 EndpointSlice [must-gather] [must-gather-dnqjn] POD 2026-07-07T16:00:33.574199335Z Wrote inspect data to must-gather. [must-gather] [must-gather-dnqjn] POD 2026-07-07T16:00:33.741921686Z Gathering data for ns/kserve-ci-e2e-test... [must-gather] [must-gather-dnqjn] POD 2026-07-07T16:00:34.932860636Z Warning: apps.openshift.io/v1 DeploymentConfig is deprecated in v4.14+, unavailable in v4.10000+ [must-gather] [must-gather-dnqjn] POD 2026-07-07T16:00:35.764352317Z Warning: v1 Endpoints is deprecated in v1.33+; use discovery.k8s.io/v1 EndpointSlice [must-gather] [must-gather-dnqjn] POD 2026-07-07T16:00:36.157048556Z Wrote inspect data to must-gather. [must-gather] [must-gather-dnqjn] POD 2026-07-07T16:00:36.338183219Z Gathering data for ns/kserve-ci-e2e-test... [must-gather] [must-gather-dnqjn] OUT 2026-07-07T16:00:36.616202407Z gather logs unavailable: Get "https://10.0.135.220:10250/containerLogs/openshift-must-gather-h6gn8/must-gather-dnqjn/gather?follow=true&sinceSeconds=2×tamps=true": unexpected EOF [must-gather] [must-gather-dnqjn] OUT 2026-07-07T16:00:36.616268891Z waiting for gather to complete [must-gather] [must-gather-dnqjn] OUT 2026-07-07T16:02:36.621008369Z gather never finished: pods "must-gather-dnqjn" not found [must-gather] [must-gather ] OUT 2026-07-07T16:02:36.627352548Z namespace/openshift-must-gather-h6gn8 deleted [must-gather] [must-gather] [must-gather] Reprinting Cluster State: [must-gather] When opening a support case, bugzilla, or issue please include the following summary data along with any other requested information: [must-gather] ClusterID: b54a7a34-f018-4011-8b3a-987919d37ffa [must-gather] ClientVersion: 4.21.10 [must-gather] ClusterVersion: Stable at "4.21.23" [must-gather] ClusterOperators: [must-gather] clusteroperator/dns is not available (DNS "default" is unavailable.) because DNS default is degraded [must-gather] clusteroperator/image-registry is not available (Available: The deployment does not have available replicas [must-gather] NodeCADaemonAvailable: The daemon set node-ca does not have available replicas [must-gather] ImagePrunerAvailable: Pruner CronJob has been created) because Degraded: The deployment does not have available replicas [must-gather] clusteroperator/ingress is not available (The "default" ingress controller reports Available=False: IngressControllerUnavailable: One or more status conditions indicate unavailable: DeploymentAvailable=False (DeploymentUnavailable: The deployment has Available status condition set to False (reason: MinimumReplicasUnavailable) with message: Deployment does not have minimum availability.)) because The "default" ingress controller reports Degraded=True: DegradedConditions: One or more other status conditions indicate a degraded state: DeploymentAvailable=False (DeploymentUnavailable: The deployment has Available status condition set to False (reason: MinimumReplicasUnavailable) with message: Deployment does not have minimum availability.). [must-gather] clusteroperator/network is progressing: DaemonSet "/openshift-multus/multus" is not available (awaiting 2 nodes) [must-gather] DaemonSet "/openshift-multus/multus-additional-cni-plugins" is not available (awaiting 2 nodes) [must-gather] DaemonSet "/openshift-multus/network-metrics-daemon" is not available (awaiting 2 nodes) [must-gather] DaemonSet "/openshift-ovn-kubernetes/ovnkube-node" is not available (awaiting 2 nodes) [must-gather] DaemonSet "/openshift-network-operator/iptables-alerter" is not available (awaiting 2 nodes) [must-gather] Deployment "/openshift-network-console/networking-console-plugin" is not available (awaiting 1 nodes) [must-gather] clusteroperator/node-tuning is not available (DaemonSet "tuned" has no available Pod(s)) because DaemonSet "tuned" available [must-gather] clusteroperator/storage is not available (AWSEBSCSIDriverOperatorCRAvailable: AWSEBSDriverNodeServiceControllerAvailable: Waiting for the DaemonSet to deploy the CSI Node Service) because AWSEBSCSIDriverOperatorCRDegraded: All is well [must-gather] clusteroperator/authentication is missing [must-gather] clusteroperator/cloud-credential is missing [must-gather] clusteroperator/cluster-autoscaler is missing [must-gather] clusteroperator/config-operator is missing [must-gather] clusteroperator/etcd is missing [must-gather] clusteroperator/machine-api is missing [must-gather] clusteroperator/machine-approver is missing [must-gather] clusteroperator/machine-config is missing [must-gather] clusteroperator/marketplace is missing [must-gather] [must-gather] [must-gather] error: gather never finished for pod must-gather-dnqjn: pods "must-gather-dnqjn" not found [git-push-artifacts] WORK_DIR: /workspace/odh-ci-artifacts [git-push-artifacts] REPO_PATH: opendatahub-io/odh-build-metadata [git-push-artifacts] REPO_BRANCH: ci-artifacts [git-push-artifacts] SPARSE_FILE_PATH: test-artifacts/docs [git-push-artifacts] SOURCE_PATH: /workspace/artifacts-dir [git-push-artifacts] DEST_PATH: test-artifacts/kserve-group-test-mc25r [git-push-artifacts] ALWAYS_PASS: false [git-push-artifacts] configuring gh token [git-push-artifacts] taking github token from Konflux bot [git-push-artifacts] Initialized empty Git repository in /workspace/odh-ci-artifacts/.git/ [git-push-artifacts] Using partial fetch with sparse checkout for: test-artifacts/docs [git-push-artifacts] From https://github.com/opendatahub-io/odh-build-metadata [git-push-artifacts] * branch ci-artifacts -> FETCH_HEAD [git-push-artifacts] * [new branch] ci-artifacts -> origin/ci-artifacts [git-push-artifacts] Already on 'ci-artifacts' [git-push-artifacts] branch 'ci-artifacts' set up to track 'origin/ci-artifacts'. [git-push-artifacts] TASK_NAME=kserve-group-test-mc25r-e2e-llm-inference-service [git-push-artifacts] PIPELINERUN_NAME=kserve-group-test-mc25r [git-push-artifacts] From https://github.com/opendatahub-io/odh-build-metadata [git-push-artifacts] * branch ci-artifacts -> FETCH_HEAD [git-push-artifacts] Already up to date. [git-push-artifacts] -rw-r--r--. 1 root 1001540000 193543 Jul 7 16:02 /workspace/odh-ci-artifacts/test-artifacts/kserve-group-test-mc25r/e2e-llm-inference-service.tar.gz [git-push-artifacts] [ci-artifacts c583d33] Updating CI Artifacts in e2e-llm-inference-service [git-push-artifacts] 1 file changed, 0 insertions(+), 0 deletions(-) [git-push-artifacts] create mode 100644 test-artifacts/kserve-group-test-mc25r/e2e-llm-inference-service.tar.gz [git-push-artifacts] From https://github.com/opendatahub-io/odh-build-metadata [git-push-artifacts] * branch ci-artifacts -> FETCH_HEAD [git-push-artifacts] Already up to date. [git-push-artifacts] To https://github.com/opendatahub-io/odh-build-metadata.git [git-push-artifacts] 20d0bb3..c583d33 ci-artifacts -> ci-artifacts [fail-if-needed] Failing pipeline because deploy-and-e2e step failed container step-fail-if-needed has failed : [{"key":"StartedAt","value":"2026-07-07T16:02:42.839Z","type":3}]