large-cluster-test.yaml

name: Large Cluster Test
on:
  workflow_dispatch:
    inputs:
      from_branch:
        description: 'Which branch to source release branch from? (default: master)'
        required: false
        default: 'master'
jobs:
  large-cluster-test:
    runs-on: ubuntu-latest
    steps:
      - uses: actions/checkout@v2
        with:
          ref: ${{github.event.inputs.from_branch}}
      - name: Setup SSH Keys and known_hosts
        env:
          SSH_AUTH_SOCK: /tmp/ssh_agent.sock
        run: |
          ssh-agent -a $SSH_AUTH_SOCK > /dev/null
      - name: install go
        uses: actions/setup-go@v2
        with:
          go-version: '^1.16'
      - name: Install helm
        run: |
          curl https://baltocdn.com/helm/signing.asc | sudo apt-key add -
          sudo apt-get install apt-transport-https --yes
          echo "deb https://baltocdn.com/helm/stable/debian/ all main" | sudo tee /etc/apt/sources.list.d/helm-stable-debian.list
          sudo apt-get update
          sudo apt-get install helm
      - name: Install jq
        run: |
          sudo apt-get install jq
      - name: Install k
        run: |
          sudo curl -o /usr/local/bin/k https://raw.githubusercontent.com/jakepearson/k/master/k
          sudo chmod +x /usr/local/bin/k
      - name: Install sonobuoy
        run: |
          sudo curl -fsSL https://github.com/vmware-tanzu/sonobuoy/releases/download/v0.53.2/sonobuoy_0.53.2_linux_amd64.tar.gz -o sonobuoy_0.53.2_linux_amd64.tar.gz
          sudo tar -xvf ./sonobuoy_0.53.2_linux_amd64.tar.gz
          sudo chmod +x ./sonobuoy
          sudo mv ./sonobuoy /usr/local/bin/sonobuoy
          sudo rm ./sonobuoy_0.53.2_linux_amd64.tar.gz
      - name: Build aks-engine binary
        run: make build-binary
      - name: Validate large cluster scenario
        env:
          SSH_AUTH_SOCK: /tmp/ssh_agent.sock
          ORCHESTRATOR_RELEASE: "1.22"
          CLUSTER_DEFINITION: "examples/largeclusters/kubernetes.json"
          RUN_VMSS_NODE_PROTOTYPE: true
          KAMINO_VMSS_PROTOTYPE_NEW_NODES: 98
          GINKGO_FOCUS: "should be able to install vmss node prototype"
          GINKGO_SKIP: "should create a pv by deploying a pod that consumes a pvc|should have the expected k8s version|should report all nodes in a Ready state|should have node labels specific to masters or agents|should have functional container networking DNS"
          SUBSCRIPTION_ID: ${{ secrets.TEST_AZURE_SUB_ID }}
          CLIENT_ID: ${{ secrets.TEST_AZURE_SP_ID }}
          CLIENT_SECRET: ${{ secrets.TEST_AZURE_SP_PW }}
          LOCATION: "eastus2"
          TENANT_ID: ${{ secrets.TEST_AZURE_TENANT_ID }}
          CLEANUP_ON_EXIT: false
          CLEANUP_IF_FAIL: true
          CONTAINER_RUNTIME: "docker"
          SKIP_LOGS_COLLECTION: true
          AZURE_CORE_ONLY_SHOW_ERRORS: True
        run: make test-kubernetes
      - name: Validate ELB with large cluster
        env:
          SSH_AUTH_SOCK: /tmp/ssh_agent.sock
          ORCHESTRATOR_RELEASE: "1.22"
          CLUSTER_DEFINITION: "examples/largeclusters/kubernetes.json"
          LB_TIMEOUT: "30m"
          GINKGO_FOCUS: "should be able to produce working LoadBalancers"
          GINKGO_SKIP: "should create a pv by deploying a pod that consumes a pvc|should have the expected k8s version|should report all nodes in a Ready state|should have node labels specific to masters or agents|should have functional container networking DNS"
          SUBSCRIPTION_ID: ${{ secrets.TEST_AZURE_SUB_ID }}
          CLIENT_ID: ${{ secrets.TEST_AZURE_SP_ID }}
          CLIENT_SECRET: ${{ secrets.TEST_AZURE_SP_PW }}
          LOCATION: "eastus2"
          TENANT_ID: ${{ secrets.TEST_AZURE_TENANT_ID }}
          CLEANUP_ON_EXIT: true
          CLEANUP_IF_FAIL: true
          CONTAINER_RUNTIME: "docker"
          SKIP_LOGS_COLLECTION: true
          AZURE_CORE_ONLY_SHOW_ERRORS: True
        run: make test-kubernetes
      - name: Run conformance tests
        run: |
          export KUBECONFIG=_output/$(ls -dt1 _output/* | head -n 1 | cut -d/ -f2)/kubeconfig/kubeconfig.eastus2.json
          sonobuoy run --plugin e2e --wait --e2e-skip "Daemon set|SchedulerPreemption|validates resource limits of pods that are allowed to run|validates that NodeSelector is respected|validates that there exists conflict between pods with same hostPort and protocol" --timeout=86400
          k get pod sonobuoy -n sonobuoy -o json | jq -r '.metadata.annotations."sonobuoy.hept.io/status"' | jq -r '.plugins[] | select(.plugin == "e2e") | ."result-counts".failed == 0'