Synced kserve upstream master branch into odh kserve master branch.

.github/actions/kserve-dep-setup/action.yml

@@ -17,22 +17,14 @@ runs:
       run: |
         shopt -s nocasematch 
 
-        if [[ "${{ inputs.network-layer }}" == "istio" ]]; then
-            echo "Selected network layer ${{ inputs.network-layer }}"
-            ./test/scripts/gh-actions/setup-deps.sh ${{ inputs.deployment-mode }}
-
-            kubectl get pods -n istio-system
-            kubectl describe svc -n istio-system istio-ingressgateway
-        elif [[ "${{ inputs.network-layer }}" == "kourier" ]]; then
-            echo "Selected network layer ${{ inputs.network-layer }}"
-            ./test/scripts/gh-actions/setup-kourier.sh
+        if [[ "${{ inputs.network-layer }}" == "kourier" ]]; then
+          echo "Selected network layer ${{ inputs.network-layer }}"
+          ./test/scripts/gh-actions/setup-kourier.sh
         else
-          echo "Unsupported network layer"
+          echo "Selected network layer ${{ inputs.network-layer }}"
+          ./test/scripts/gh-actions/setup-deps.sh ${{ inputs.deployment-mode }} "${{ inputs.network-layer }}"
         fi
 
-        kubectl get pods -n knative-serving
-        kubectl get pods -n cert-manager
-
     - name: Update test overlays
       shell: bash
       run: |

.github/actions/minikube-setup/action.yml

@@ -1,22 +1,38 @@
 name: 'Minikube setup action'
 description: 'Sets up minikube on the github runner'
 
+inputs:
+  nodes:
+    description: 'Number of nodes to start minikube with'
+    required: false
+    default: '1'
+  driver:
+    description: 'Driver to use for minikube'
+    required: false
+    default: 'none'
+  start-args:
+    description: 'Additional arguments to pass to minikube start'
+    required: false
+    default: ''
+
 runs:
   using: "composite"
   steps:
     - name: Install kubectl
       uses: azure/[email protected]
       with:
-        version: 'v1.29.7'
+        version: 'v1.30.7'
 
     - name: Setup Minikube
       uses: medyagh/setup-minikube@latest
       with:
         minikube-version: '1.33.1'
-        kubernetes-version: 'v1.29.7'
-        driver: 'none'
+        kubernetes-version: 'v1.30.7'
+        driver: ${{ inputs.driver }}
         wait: 'all'
-        start-args: --wait-timeout=6m0s
+        cpus: 'max'
+        memory: 'max'
+        start-args: --wait-timeout=6m0s --nodes=${{ inputs.nodes }} ${{ inputs.start-args }}
 
     - name: Check Kubernetes pods
       shell: bash

.github/workflows/e2e-test.yml

@@ -24,6 +24,7 @@ env:
   # Controller images
   CONTROLLER_IMG: "kserve-controller"
   LOCALMODEL_CONTROLLER_IMG: "kserve-localmodel-controller"
+  LOCALMODEL_AGENT_IMG: "kserve-localmodelnode-agent"
   STORAGE_INIT_IMG: "storage-initializer"
   AGENT_IMG: "agent"
   ROUTER_IMG: "router"
@@ -87,6 +88,14 @@ jobs:
           compression-level: 0
           if-no-files-found: error
 
+      - name: Upload localmodel agent image
+        uses: actions/upload-artifact@v4
+        with:
+          name: ${{ env.BASE_ARTIFACT_PREFIX }}-${{ env.LOCALMODEL_AGENT_IMG }}-${{ github.sha }}
+          path: ${{ env.DOCKER_IMAGES_PATH }}/${{ env.LOCALMODEL_AGENT_IMG }}-${{ github.sha }}
+          compression-level: 0
+          if-no-files-found: error
+
       - name: Upload agent image
         uses: actions/upload-artifact@v4
         with:
@@ -589,7 +598,7 @@ jobs:
 
       - name: Patch inferenceservice config
         run: |
-          kubectl patch configmaps -n kserve inferenceservice-config --patch-file config/overlays/test/configmap/inferenceservice-ingress.yaml
+          kubectl patch configmaps -n kserve inferenceservice-config --patch-file config/overlays/test/configmap/inferenceservice-path-template.yaml
           kubectl describe configmaps -n kserve inferenceservice-config
 
       - name: Run E2E tests with path-based routing
@@ -720,8 +729,13 @@ jobs:
 
   test-raw:
     runs-on: ubuntu-22.04
+    strategy:
+      fail-fast: false
+      matrix:
+        # Test with Ingress and Gateway API
+        network-layer: ["istio-ingress", "envoy-gatewayapi", "istio-gatewayapi"]
     needs:
-      [kserve-image-build, predictor-runtime-build]
+      [kserve-image-build, predictor-runtime-build, explainer-runtime-build]
     steps:
       - name: Checkout source
         uses: actions/checkout@v4
@@ -733,6 +747,7 @@ jobs:
         uses: actions/setup-go@v5
         with:
           go-version: "1.22"
+
       - name: Setup Python
         uses: actions/setup-python@v5
         with:
@@ -745,6 +760,7 @@ jobs:
         uses: ./.github/actions/kserve-dep-setup
         with:
           deployment-mode: "raw"
+          network-layer: ${{ matrix.network-layer }}
 
       - name: Download base images
         uses: ./.github/actions/base-download
@@ -767,11 +783,12 @@ jobs:
           name: ${{ env.PREDICTOR_ARTIFACT_PREFIX }}-${{ env.CUSTOM_MODEL_GRPC_IMG }}-${{ github.sha }}
           path: ./tmp
 
-      - name: Download transformer image
+      - name: Download transformer and explainer artifacts
         uses: actions/download-artifact@v4
         with:
-          name: ${{ env.TRANSFORMER_ARTIFACT_PREFIX }}-${{ env.IMAGE_TRANSFORMER_IMG }}-${{ github.sha }}
           path: ./tmp
+          pattern: +(${{ env.TRANSFORMER_ARTIFACT_PREFIX }}|${{ env.EXPLAINER_ARTIFACT_PREFIX }})-*
+          merge-multiple: true
 
       - name: Load docker images
         uses: ./.github/actions/load-docker-images
@@ -783,7 +800,7 @@ jobs:
 
       - name: Install KServe
         run: |
-          ./test/scripts/gh-actions/setup-kserve.sh "raw"
+          ./test/scripts/gh-actions/setup-kserve.sh "raw" ${{ matrix.network-layer }}
 
           kubectl get pods -n kserve
           kubectl describe pods -n kserve
@@ -792,10 +809,27 @@ jobs:
         run: |
           kubectl describe configmaps -n kserve inferenceservice-config
 
+      - name: Enable Gateway API
+        if: matrix.network-layer == 'envoy-gatewayapi' || matrix.network-layer == 'istio-gatewayapi'
+        run: |
+          kubectl patch configmaps -n kserve inferenceservice-config --patch-file config/overlays/test/configmap/inferenceservice-enable-gateway-api.yaml
+
       - name: Run E2E tests
         timeout-minutes: 30
         run: |
-          ./test/scripts/gh-actions/run-e2e-tests.sh "raw" "6"
+          ./test/scripts/gh-actions/run-e2e-tests.sh "raw" "6" ${{ matrix.network-layer }}
+
+      - name: Patch inferenceservice config for path based routing
+        if: matrix.network-layer == 'envoy-gatewayapi' || matrix.network-layer == 'istio-gatewayapi'
+        run: |
+          kubectl patch configmaps -n kserve inferenceservice-config --patch-file config/overlays/test/configmap/inferenceservice-path-template.yaml
+          kubectl describe configmaps -n kserve inferenceservice-config
+
+      - name: Run E2E tests with path based routing
+        if: matrix.network-layer == 'envoy-gatewayapi' || matrix.network-layer == 'istio-gatewayapi'
+        timeout-minutes: 30
+        run: |
+          ./test/scripts/gh-actions/run-e2e-tests.sh "raw" "6" ${{ matrix.network-layer }}
 
       - name: Patch inferenceservice config for cluster ip none
         run: |
@@ -805,7 +839,7 @@ jobs:
       - name: Run E2E tests - cluster ip none
         timeout-minutes: 30
         run: |
-          ./test/scripts/gh-actions/run-e2e-tests.sh "rawcipn" "1"
+          ./test/scripts/gh-actions/run-e2e-tests.sh "rawcipn" "1" ${{ matrix.network-layer }}
 
       - name: Check system status
         if: always()
@@ -900,7 +934,7 @@ jobs:
   test-llm:
     runs-on: ubuntu-22.04
     needs:
-      [ kserve-image-build, predictor-runtime-build]
+      [kserve-image-build, predictor-runtime-build]
     steps:
       - name: Checkout source
         uses: actions/checkout@v4
@@ -960,7 +994,7 @@ jobs:
   test-huggingface-server-vllm:
     runs-on: ubuntu-22.04
     needs:
-      [ kserve-image-build, predictor-runtime-build]
+      [kserve-image-build, predictor-runtime-build]
     steps:
       - name: Checkout source
         uses: actions/checkout@v4
@@ -1015,4 +1049,129 @@ jobs:
       - name: Check system status
         if: always()
         run: |
+          ./test/scripts/gh-actions/status-check.sh
+
+  test-modelcache:
+    runs-on: ubuntu-22.04
+    needs:
+      [kserve-image-build, predictor-runtime-build]
+    steps:
+      - name: Checkout source
+        uses: actions/checkout@v4
+
+      - name: Free-up disk space
+        uses: ./.github/actions/free-up-disk-space
+
+      - name: Setup Go
+        uses: actions/setup-go@v5
+        with:
+          go-version: "1.22"
+
+      - name: Setup Python
+        uses: actions/setup-python@v5
+        with:
+          python-version: "3.9"
+
+      - name: Create tmp-images directory
+        run: |
+          sudo mkdir -p /tmp-images
+          sudo chown -R $USER /tmp-images
+
+      - name: Setup Minikube
+        uses: ./.github/actions/minikube-setup
+        with:
+          nodes: 3
+          driver: "docker"
+          start-args: "--mount --mount-string=/tmp-images:/tmp-images"
+
+      - name: Create minikube tunnel
+        run: |
+          nohup minikube tunnel > minikube-tunnel.log 2>&1 &
+
+      - name: KServe dependency setup
+        uses: ./.github/actions/kserve-dep-setup
+
+      - name: Download base images
+        uses: actions/download-artifact@v4
+        with:
+          path: /tmp-images
+          pattern: ${{ env.BASE_ARTIFACT_PREFIX }}-*
+          merge-multiple: true
+
+      - name: Load base docker images
+        run: |
+          ls -l /tmp-images
+          minikube ssh -n minikube -- ls -l /tmp-images
+          files=$(find /tmp-images -maxdepth 1 -type f)
+          for file in ${files[@]};do
+            echo "Loading image $(basename ${file})"
+            minikube ssh -n minikube -- docker image load -i ${file}
+            minikube ssh -n minikube-m02 -- docker image load -i ${file}
+            minikube ssh -n minikube-m03 -- docker image load -i ${file}
+          done
+          sudo rm -rf /tmp-images/*
+          minikube ssh -n minikube -- docker image ls
+          minikube ssh -n minikube-m02 -- docker image ls
+          minikube ssh -n minikube-m03 -- docker image ls
+
+      - name: Download huggingface server image
+        uses: actions/download-artifact@v4
+        with:
+          name: ${{ env.HUGGINGFACE_IMG }}-${{ github.sha }}
+          path: /tmp-images
+
+      - name: Load runtime docker images
+        run: |
+          files=$(find /tmp-images -maxdepth 1 -type f)
+          for file in ${files[@]};do
+            echo "Loading image $(basename ${file})"
+            minikube ssh -n minikube-m02 -- docker image load -i ${file}
+          done
+          sudo rm -rf tmp-images/*
+          minikube ssh -n minikube-m02 -- docker image ls
+
+      - name: Create model root directory
+        run: |
+          minikube ssh -n minikube-m02 -- sudo mkdir -p -m=777 /models
+          minikube ssh -n minikube-m03 -- sudo mkdir -p -m=777 /models
+
+      - name: Install Poetry and version plugin
+        run: ./test/scripts/gh-actions/setup-poetry.sh
+
+      - name: Install KServe
+        run: |
+          ./test/scripts/gh-actions/setup-kserve.sh
+
+          kubectl get pods -n kserve
+          kubectl describe pods -n kserve
+
+      - name: Enable modelcache
+        run: |
+          sed -i -e "s/latest/${GITHUB_SHA}/g" config/overlays/test/configmap/inferenceservice-enable-modelcache.yaml
+          kubectl patch configmaps -n kserve inferenceservice-config --patch-file config/overlays/test/configmap/inferenceservice-enable-modelcache.yaml
+          kubectl describe configmaps -n kserve inferenceservice-config
+
+      - name: Create localmodel job namespace
+        run: |
+          kubectl create ns kserve-localmodel-jobs
+
+      - name: Label worker nodes for modelcache
+        run: |
+          kubectl label nodes -l '!node-role.kubernetes.io/control-plane' kserve/localmodel=worker
+
+      - name: Enable nodeselector in knative
+        run: |
+          kubectl patch configmaps -n knative-serving config-features --patch '{"data": {"kubernetes.podspec-nodeselector": "enabled"}}'
+
+      - name: Run E2E tests
+        timeout-minutes: 15
+        run: |
+          ./test/scripts/gh-actions/run-e2e-tests.sh "modelcache" "1"
+
+      - name: Check system status
+        if: always()
+        run: |
+          echo "::group::Minikube tunnel logs"
+          cat minikube-tunnel.log
+          echo "::endgroup::"
           ./test/scripts/gh-actions/status-check.sh

.github/workflows/golangci-lint.yml

@@ -18,11 +18,47 @@ jobs:
       - uses: actions/checkout@v4
       - uses: actions/setup-go@v5
         with:
-          go-version: '1.22'
+          go-version: "1.22"
           cache: false
       - name: golangci-lint
         uses: golangci/golangci-lint-action@v4
         with:
           version: v1.56
           args: --out-format=line-number
 
+  verify-go-mod:
+    runs-on: ubuntu-latest
+    steps:
+      - name: Checkout source code
+        uses: actions/checkout@v4
+
+      - name: Setup Go
+        uses: actions/setup-go@v5
+        with:
+          go-version: 1.22
+          cache: false
+
+      - name: Run go mod tidy
+        run: |
+          go mod tidy
+
+      - name: Verify
+        shell: bash
+        run: |
+          # From: https://backreference.org/2009/12/23/how-to-match-newlines-in-sed/
+          # This is to leverage this workaround:
+          # https://github.com/actions/toolkit/issues/193#issuecomment-605394935
+          urlencode() {
+            sed ':begin;$!N;s/\n/%0A/;tbegin'
+          }
+
+          if [ -z "$(git status --porcelain)" ]; then
+              echo "${{ github.repository }} up to date."
+          else
+              echo "Found diffs in: $(git diff-index --name-only HEAD)"
+              for x in $(git diff-index --name-only HEAD); do
+                  echo "::error file=$x::Please run 'go mod tidy'.%0A$(git diff $x | urlencode)"
+              done
+              echo "${{ github.repository }} is out of date. Please run 'go mod tidy'"
+              exit 1
+          fi