[backport] fix/ci: Linux NPM race condition editing "except" CIDR and NPM pipeline fixes

.github/workflows/cyclonus-netpol-extended-nightly-test.yaml

@@ -73,7 +73,7 @@ jobs:
         uses: actions/upload-artifact@v4
         if: always()
         with:
-          name: logs
+          name: logs-${{ matrix.profile }}
           path: |
             ./npm-logs_${{ matrix.profile }}.txt
             ./cyclonus-test_${{ matrix.profile }}.txt

.github/workflows/cyclonus-netpol-test.yaml

@@ -5,6 +5,7 @@ on:
   push:
     branches:
       - main
+      - release/*
   pull_request:
     paths:
       - 'npm/**'
@@ -49,7 +50,7 @@ jobs:
       - name: Make NPM image
         run: |
           make npm-image NPM_PLATFORM_TAG=cyclonus PLATFORM=linux/amd64 CONTAINER_BUILDER=docker BUILDX_ACTION='--load'
-      
+
       - name: Install Azure NPM
         # set the ConfigMap based on the build matrix
         # currently have to restart the daemonset because changing the ConfigMap doesn't restart NPM
@@ -60,7 +61,7 @@ jobs:
           echo "Applying profile: ${{ matrix.profile }}"
           kubectl apply -f ./npm/profiles/${{ matrix.profile }}
           kubectl rollout restart ds azure-npm -n kube-system
-    
+
       - name: Check Cluster Components
         run: |
           sleep 10
@@ -80,7 +81,7 @@ jobs:
         uses: actions/upload-artifact@v4
         if: always()
         with:
-          name: logs
+          name: logs-${{ matrix.profile }}
           path: |
             ./npm-logs_${{ matrix.profile }}.txt
             ./cyclonus-test_${{ matrix.profile }}.txt

.pipelines/npm/npm-conformance-tests-latest-release.yaml

@@ -1,5 +1,10 @@
 trigger:
-  - master
+  branches:
+    include:
+      - master
+  tags:
+    include:
+      - "*"
 
 variables:
   - name: VNET_NAME
@@ -113,7 +118,7 @@ jobs:
           scriptLocation: "inlineScript"
           failOnStderr: true
           inlineScript: |
-            # get kubectl 
+            # get kubectl
             curl -LO https://dl.k8s.io/release/v1.23.0/bin/linux/amd64/kubectl
             chmod +x kubectl
             echo Cluster $(AZURE_CLUSTER)
@@ -177,7 +182,7 @@ jobs:
               ./kubectl --kubeconfig=./kubeconfig set image daemonset/azure-npm-win -n kube-system azure-npm=$IMAGE_REGISTRY/azure-npm:windows-amd64-ltsc2022-$(TAG)
 
             else
-              echo "Creating Linux Cluster"; 
+              echo "Creating Linux Cluster";
               az aks create --no-ssh-key \
               --resource-group $(RESOURCE_GROUP) \
               --name $(AZURE_CLUSTER) \
@@ -261,7 +266,7 @@ jobs:
               declare -a conformancePIDs
               for round in $(seq 1 $NUM_PARALLEL_JOBS_FOR_STRESS_TEST); do
                   # for each iteration, run the conformance test and echos in the background, and write the output of the conformance test to a file
-                  
+
                    # run the conformance test in the foreground and write the output to stdout and a file
                   if [[ $(AZURE_CLUSTER) == *ws22 ]] # * is used for pattern matching
                   then
@@ -314,133 +319,134 @@ jobs:
         condition: always()
         artifact: NpmLogs_$(AZURE_CLUSTER)
 
-  - job: Create_Windows_Cluster_and_Run_Test
-    timeoutInMinutes: 360
-    displayName: "Run Windows Cyclonus"
-    pool:
-      name: $(BUILD_POOL_NAME_DEFAULT)
-      demands:
-        - agent.os -equals Linux
-        - Role -equals Build
-    dependsOn: [setup]
-    variables:
-      RESOURCE_GROUP: $[ dependencies.setup.outputs['EnvironmentalVariables.RESOURCE_GROUP'] ]
-      TAG: $[ dependencies.setup.outputs['EnvironmentalVariables.TAG'] ]
-      FQDN: empty
-    strategy:
-      matrix:
-        v2-windows:
-          PROFILE: "cyc-ws22"
-    steps:
-      - checkout: self
-      - download: none
-
-      - task: AzureCLI@2
-        displayName: "Create AKS Cluster"
-        inputs:
-          azureSubscription: $(BUILD_VALIDATIONS_SERVICE_CONNECTION)
-          scriptType: "bash"
-          scriptLocation: "inlineScript"
-          failOnStderr: true
-          inlineScript: |
-            az extension add --name aks-preview
-            az extension update --name aks-preview
-
-            export CLUSTER_NAME=$(RESOURCE_GROUP)-$(PROFILE)
-
-            echo "Creating resource group named $CLUSTER_NAME"
-            az group create --name $CLUSTER_NAME -l $(LOCATION) -o table
-
-            echo "Creating resource group named $CLUSTER_NAME"
-            az aks create \
-                --resource-group $CLUSTER_NAME \
-                --name $CLUSTER_NAME \
-                --generate-ssh-keys \
-                --windows-admin-username e2eadmin \
-                --windows-admin-password alpha@numeric!password2 \
-                --network-plugin azure \
-                --vm-set-type VirtualMachineScaleSets \
-                --node-vm-size Standard_D4s_v3 \
-                --node-count 1
-
-            # don't schedule anything on the linux system pool
-            echo "Updating $CLUSTER_NAME to not schedule anything on linux pool..."
-            az aks nodepool update \
-              --cluster-name $CLUSTER_NAME \
-              -g $CLUSTER_NAME \
-              -n nodepool1 \
-              --node-taints CriticalAddonsOnly=true:NoSchedule
-
-            echo "Adding Windows nodepool to $CLUSTER_NAME"
-            az aks nodepool add \
-                --resource-group $CLUSTER_NAME \
-                --cluster-name $CLUSTER_NAME \
-                --name awin22 \
-                --os-type Windows \
-                --os-sku Windows2022 \
-                --node-vm-size Standard_D4s_v3 \
-                --node-count 3
-
-            echo "Getting credentials to $CLUSTER_NAME"
-            az aks get-credentials -g $CLUSTER_NAME -n $CLUSTER_NAME --overwrite-existing --file ./kubeconfig
-            mkdir -p ~/.kube/
-            cp ./kubeconfig ~/.kube/config
-
-      - task: AzureCLI@2
-        displayName: "Deploy NPM to Test Cluster"
-        inputs:
-          azureSubscription: $(BUILD_VALIDATIONS_SERVICE_CONNECTION)
-          scriptType: "bash"
-          scriptLocation: "inlineScript"
-          failOnStderr: true
-          inlineScript: |
-            export CLUSTER_NAME=$(RESOURCE_GROUP)-$(PROFILE)
-
-            curl -LO https://dl.k8s.io/release/v1.23.0/bin/linux/amd64/kubectl
-            chmod +x kubectl
-
-            # deploy azure-npm
-            ./kubectl --kubeconfig=./kubeconfig apply -f https://raw.githubusercontent.com/Azure/azure-container-networking/master/npm/examples/windows/azure-npm.yaml
-
-            # swap azure-npm image with one built during run
-            ./kubectl --kubeconfig=./kubeconfig set image daemonset/azure-npm-win -n kube-system azure-npm=$IMAGE_REGISTRY/azure-npm:windows-amd64-ltsc2022-$(TAG)
-
-            echo "sleeping 3 minutes to allow NPM pods to restart"
-            sleep 180
-
-            ./kubectl --kubeconfig=./kubeconfig get po -n kube-system -owide -A
-
-            echo "Showing cluster status for $CLUSTER_NAME"
-            FQDN=`az aks show -n $CLUSTER_NAME -g $CLUSTER_NAME --query fqdn -o tsv`
-            echo "##vso[task.setvariable variable=FQDN]$FQDN"
-
-      - script: |
-          cat ~/.kube/config
-          curl -fsSL github.com/mattfenwick/cyclonus/releases/latest/download/cyclonus_linux_amd64.tar.gz | tar -zxv
-        name: download_cyclonus
-        displayName: "Download Cyclonus"
-        failOnStderr: false
-        condition: always()
-
-      - script: |
-          ./test/cyclonus/test-cyclonus-windows.sh
-        name: cyclonus
-        displayName: "Run Cyclonus Test"
-        failOnStderr: false
-        condition: always()
-
-      - bash: |
-          export CLUSTER_NAME=$(RESOURCE_GROUP)-$(PROFILE)
-          cp cyclonus-$CLUSTER_NAME $(System.DefaultWorkingDirectory)/$CLUSTER_NAME/cyclonus-$CLUSTER_NAME
-          echo "Getting cluster state for $CLUSTER_NAME"
-          mkdir -p $(System.DefaultWorkingDirectory)/$CLUSTER_NAME
-          kubectl logs -n kube-system -l k8s-app=azure-npm --tail -1 --prefix > $(System.DefaultWorkingDirectory)/$CLUSTER_NAME/npm-logs_$(PROFILE).txt
-          cp ./kubeconfig $(System.DefaultWorkingDirectory)/$CLUSTER_NAME/.kubeconfig
-        condition: always()
-
-      - publish: $(System.DefaultWorkingDirectory)/$(RESOURCE_GROUP)-$(PROFILE)
-        condition: always()
-        artifact: NpmLogs-$(RESOURCE_GROUP)-$(PROFILE)
+  # 2024/07/23: Windows Cyclonus is consistently timing after 6 hours
+  # - job: Create_Windows_Cluster_and_Run_Test
+  #   timeoutInMinutes: 360
+  #   displayName: "Run Windows Cyclonus"
+  #   pool:
+  #     name: $(BUILD_POOL_NAME_DEFAULT)
+  #     demands:
+  #       - agent.os -equals Linux
+  #       - Role -equals Build
+  #   dependsOn: [setup]
+  #   variables:
+  #     RESOURCE_GROUP: $[ dependencies.setup.outputs['EnvironmentalVariables.RESOURCE_GROUP'] ]
+  #     TAG: $[ dependencies.setup.outputs['EnvironmentalVariables.TAG'] ]
+  #     FQDN: empty
+  #   strategy:
+  #     matrix:
+  #       v2-windows:
+  #         PROFILE: "cyc-ws22"
+  #   steps:
+  #     - checkout: self
+  #     - download: none
+
+  #     - task: AzureCLI@2
+  #       displayName: "Create AKS Cluster"
+  #       inputs:
+  #         azureSubscription: $(BUILD_VALIDATIONS_SERVICE_CONNECTION)
+  #         scriptType: "bash"
+  #         scriptLocation: "inlineScript"
+  #         failOnStderr: true
+  #         inlineScript: |
+  #           az extension add --name aks-preview
+  #           az extension update --name aks-preview
+
+  #           export CLUSTER_NAME=$(RESOURCE_GROUP)-$(PROFILE)
+
+  #           echo "Creating resource group named $CLUSTER_NAME"
+  #           az group create --name $CLUSTER_NAME -l $(LOCATION) -o table
+
+  #           echo "Creating resource group named $CLUSTER_NAME"
+  #           az aks create \
+  #               --resource-group $CLUSTER_NAME \
+  #               --name $CLUSTER_NAME \
+  #               --generate-ssh-keys \
+  #               --windows-admin-username e2eadmin \
+  #               --windows-admin-password alpha@numeric!password2 \
+  #               --network-plugin azure \
+  #               --vm-set-type VirtualMachineScaleSets \
+  #               --node-vm-size Standard_D4s_v3 \
+  #               --node-count 1
+
+  #           # don't schedule anything on the linux system pool
+  #           echo "Updating $CLUSTER_NAME to not schedule anything on linux pool..."
+  #           az aks nodepool update \
+  #             --cluster-name $CLUSTER_NAME \
+  #             -g $CLUSTER_NAME \
+  #             -n nodepool1 \
+  #             --node-taints CriticalAddonsOnly=true:NoSchedule
+
+  #           echo "Adding Windows nodepool to $CLUSTER_NAME"
+  #           az aks nodepool add \
+  #               --resource-group $CLUSTER_NAME \
+  #               --cluster-name $CLUSTER_NAME \
+  #               --name awin22 \
+  #               --os-type Windows \
+  #               --os-sku Windows2022 \
+  #               --node-vm-size Standard_D4s_v3 \
+  #               --node-count 3
+
+  #           echo "Getting credentials to $CLUSTER_NAME"
+  #           az aks get-credentials -g $CLUSTER_NAME -n $CLUSTER_NAME --overwrite-existing --file ./kubeconfig
+  #           mkdir -p ~/.kube/
+  #           cp ./kubeconfig ~/.kube/config
+
+  #     - task: AzureCLI@2
+  #       displayName: "Deploy NPM to Test Cluster"
+  #       inputs:
+  #         azureSubscription: $(BUILD_VALIDATIONS_SERVICE_CONNECTION)
+  #         scriptType: "bash"
+  #         scriptLocation: "inlineScript"
+  #         failOnStderr: true
+  #         inlineScript: |
+  #           export CLUSTER_NAME=$(RESOURCE_GROUP)-$(PROFILE)
+
+  #           curl -LO https://dl.k8s.io/release/v1.23.0/bin/linux/amd64/kubectl
+  #           chmod +x kubectl
+
+  #           # deploy azure-npm
+  #           ./kubectl --kubeconfig=./kubeconfig apply -f https://raw.githubusercontent.com/Azure/azure-container-networking/master/npm/examples/windows/azure-npm.yaml
+
+  #           # swap azure-npm image with one built during run
+  #           ./kubectl --kubeconfig=./kubeconfig set image daemonset/azure-npm-win -n kube-system azure-npm=$IMAGE_REGISTRY/azure-npm:windows-amd64-ltsc2022-$(TAG)
+
+  #           echo "sleeping 3 minutes to allow NPM pods to restart"
+  #           sleep 180
+
+  #           ./kubectl --kubeconfig=./kubeconfig get po -n kube-system -owide -A
+
+  #           echo "Showing cluster status for $CLUSTER_NAME"
+  #           FQDN=`az aks show -n $CLUSTER_NAME -g $CLUSTER_NAME --query fqdn -o tsv`
+  #           echo "##vso[task.setvariable variable=FQDN]$FQDN"
+
+  #     - script: |
+  #         cat ~/.kube/config
+  #         curl -fsSL github.com/mattfenwick/cyclonus/releases/latest/download/cyclonus_linux_amd64.tar.gz | tar -zxv
+  #       name: download_cyclonus
+  #       displayName: "Download Cyclonus"
+  #       failOnStderr: false
+  #       condition: always()
+
+  #     - script: |
+  #         ./test/cyclonus/test-cyclonus-windows.sh
+  #       name: cyclonus
+  #       displayName: "Run Cyclonus Test"
+  #       failOnStderr: false
+  #       condition: always()
+
+  #     - bash: |
+  #         export CLUSTER_NAME=$(RESOURCE_GROUP)-$(PROFILE)
+  #         cp cyclonus-$CLUSTER_NAME $(System.DefaultWorkingDirectory)/$CLUSTER_NAME/cyclonus-$CLUSTER_NAME
+  #         echo "Getting cluster state for $CLUSTER_NAME"
+  #         mkdir -p $(System.DefaultWorkingDirectory)/$CLUSTER_NAME
+  #         kubectl logs -n kube-system -l k8s-app=azure-npm --tail -1 --prefix > $(System.DefaultWorkingDirectory)/$CLUSTER_NAME/npm-logs_$(PROFILE).txt
+  #         cp ./kubeconfig $(System.DefaultWorkingDirectory)/$CLUSTER_NAME/.kubeconfig
+  #       condition: always()
+
+  #     - publish: $(System.DefaultWorkingDirectory)/$(RESOURCE_GROUP)-$(PROFILE)
+  #       condition: always()
+  #       artifact: NpmLogs-$(RESOURCE_GROUP)-$(PROFILE)
 
   - job: clean_up
     displayName: "Cleanup"
@@ -450,7 +456,7 @@ jobs:
         - agent.os -equals Linux
         - Role -equals Build
     dependsOn:
-      [Create_Cluster_and_Run_Test, Create_Windows_Cluster_and_Run_Test, setup]
+      [Create_Cluster_and_Run_Test, setup]
     variables:
       RESOURCE_GROUP: $[ dependencies.setup.outputs['EnvironmentalVariables.RESOURCE_GROUP'] ]
     steps: