e2e-tests.yaml

name: E2E Tests

on:
  workflow_call:
    inputs:
      environment:
        required: true
        type: string
      ref:
        required: true
        type: string

concurrency:
  group: e2e-cluster-${{ inputs.environment }}
  queue: max

env:
  IMAGE_NAME: "s3-csi-driver"
  BENCHMARK_ARTIFACTS_FOLDER: ".github/artifacts"
  BRANCH_NAME: ${{ github.head_ref || github.ref_name }}
  BENCHMARK_BUCKET: "s3://${{ vars.BENCHMARK_BUCKET }}"
  TAG_UNTESTED: "untested_${{ inputs.ref }}"
  TAG_PASSED: "test_passed_${{ inputs.ref }}"
jobs:
  build_matrix:
    name: Build Matrix
    uses: ./.github/workflows/build_matrix.yaml
    permissions:
      contents: read
  build:
    runs-on: ubuntu-22.04 # FIXME - https://github.com/actions/runner-images/issues/11471
    environment: ${{ inputs.environment }}
    permissions:
      id-token: write
      contents: read
    steps:
      - name: Checkout
        uses: actions/checkout@v6
        with:
          ref: ${{ inputs.ref }}
          persist-credentials: false
      - name: Set up Go
        uses: actions/setup-go@v6
        with:
          go-version-file: "go.mod"
      - name: Set up Docker Buildx
        uses: docker/setup-buildx-action@v3
      - name: Configure AWS Credentials
        uses: aws-actions/configure-aws-credentials@master
        with:
          role-to-assume: ${{ vars.IAM_ROLE }}
          aws-region: ${{ vars.AWS_REGION }}
      - name: Login to Amazon ECR
        id: login-ecr
        uses: aws-actions/amazon-ecr-login@v2
      - name: Build, tag, and push docker image to Amazon ECR Private Repository
        env:
          REGISTRY: ${{ steps.login-ecr.outputs.registry }}
          PLATFORM: "linux/amd64,linux/arm64"
          TAG: "${{ env.TAG_UNTESTED }}"
        run: |
          make -j `nproc` all-push-skip-if-present
  test:
    needs: ["build", "build_matrix"]
    strategy:
      # Failing fast causes some resources created during the test to leak,
      # so we disable it to ensure all resources created during test are properly cleaned up.
      fail-fast: false
      matrix: ${{ fromJson(needs.build_matrix.outputs.matrix) }}
    runs-on: ubuntu-22.04 # FIXME - https://github.com/actions/runner-images/issues/11471
    environment: ${{ inputs.environment }}
    permissions:
      id-token: write
      contents: read
    env:
      AWS_REGION: "${{ vars.AWS_REGION }}"
      CLUSTER_TYPE: "${{ matrix.cluster-type }}"
      ARCH: "${{ matrix.arch }}"
      AMI_FAMILY: "${{ matrix.family }}"
      TAG: "untested_${{ inputs.ref }}"
      K8S_VERSION: "${{ matrix.kubernetes-version }}"
      SELINUX_MODE: "${{ matrix.selinux-mode }}"
    steps:
      - name: Checkout
        uses: actions/checkout@v6
        with:
          ref: ${{ inputs.ref }}
          persist-credentials: false
      - name: Set up Go
        uses: actions/setup-go@v6
        with:
          go-version-file: "go.mod"
      - name: Configure AWS Credentials
        uses: aws-actions/configure-aws-credentials@v6
        with:
          role-to-assume: ${{ vars.IAM_ROLE }}
          aws-region: ${{ vars.AWS_REGION }}
          role-duration-seconds: 7200
      - name: Install tools
        env:
          ACTION: "install_tools"
        run: |
          tests/e2e-kubernetes/scripts/run.sh
      - name: Create cluster
        id: create_cluster
        env:
          ACTION: "create_cluster"
        run: |
          tests/e2e-kubernetes/scripts/run.sh
      - name: Update kubeconfig
        env:
          ACTION: "update_kubeconfig"
        run: |
          tests/e2e-kubernetes/scripts/run.sh

      # Test SystemdMounter
      # SystemdMounter tests are skipped if SELinux is enabled as SystemdMounter won't work with SELinux.
      - name: Install the driver (SystemdMounter)
        if: matrix.selinux-mode != 'enforcing'
        env:
          ACTION: "install_driver"
        run: |
          tests/e2e-kubernetes/scripts/run.sh
      - name: Run E2E Tests (SystemdMounter)
        if: matrix.selinux-mode != 'enforcing'
        env:
          ACTION: "run_tests"
        run: |
          tests/e2e-kubernetes/scripts/run.sh
      - name: Post e2e cleanup (SystemdMounter)
        if: matrix.selinux-mode != 'enforcing' && always()
        env:
          ACTION: "e2e_cleanup"
        run: |
          tests/e2e-kubernetes/scripts/run.sh
      - name: Uninstall the driver (SystemdMounter)
        if: matrix.selinux-mode != 'enforcing' && always()
        env:
          ACTION: "uninstall_driver"
        run: |
          tests/e2e-kubernetes/scripts/run.sh

      # Test PodMounter
      - name: Install the driver (PodMounter)
        env:
          ACTION: "install_driver"
          MOUNTER_KIND: "pod"
        run: |
          tests/e2e-kubernetes/scripts/run.sh
      - name: Run E2E Tests (PodMounter)
        env:
          ACTION: "run_tests"
          MOUNTER_KIND: "pod"
        run: |
          tests/e2e-kubernetes/scripts/run.sh
      - name: Post e2e cleanup (PodMounter)
        if: always()
        env:
          ACTION: "e2e_cleanup"
        run: |
          tests/e2e-kubernetes/scripts/run.sh
      - name: Uninstall the driver (PodMounter)
        # Skip uninstall if cluster creation did not succeed - no cluster to communicate with
        if: ${{ always() && steps.create_cluster.conclusion == 'success' }}
        env:
          ACTION: "uninstall_driver"
        run: |
          tests/e2e-kubernetes/scripts/run.sh
      - name: Delete cluster
        if: always()
        env:
          ACTION: "delete_cluster"
        run: |
          tests/e2e-kubernetes/scripts/run.sh
  post_test:
    if: always()
    needs:
      - test
    runs-on: ubuntu-latest
    environment: ${{ inputs.environment }}
    permissions:
      id-token: write
      contents: read
    steps:
      - name: Verify jobs succeeded
        run: echo '${{ toJSON(needs) }}' | jq -e 'to_entries | all(.value.result == "success")'
      - name: Set up Docker Buildx
        uses: docker/setup-buildx-action@v3
      - name: Configure AWS Credentials
        uses: aws-actions/configure-aws-credentials@v6
        with:
          role-to-assume: ${{ vars.IAM_ROLE }}
          aws-region: ${{ vars.AWS_REGION }}
      - name: Login to Amazon ECR
        id: login-ecr
        uses: aws-actions/amazon-ecr-login@v2
      - name: Mark tests as passed
        env:
          AWS_REGION: "${{ vars.AWS_REGION }}"
          REPOSITORY: "${{ steps.login-ecr.outputs.registry }}/${{ env.IMAGE_NAME }}"
          REPOSITORY_NAME: "${{ env.IMAGE_NAME }}"
        run: |
          docker buildx imagetools create --tag ${REPOSITORY}:${TAG_PASSED} ${REPOSITORY}:${TAG_UNTESTED}
          for ARCH in amd64 arm64; do
            MANIFEST=$(aws ecr batch-get-image --repository-name ${REPOSITORY_NAME} \
              --image-ids imageTag=${TAG_UNTESTED}-linux-${ARCH}-amazon \
              --region ${AWS_REGION} --query 'images[].imageManifest' --output text)

            aws ecr put-image \
              --repository-name ${REPOSITORY_NAME} \
              --image-tag ${TAG_PASSED}-linux-${ARCH}-amazon \
              --image-manifest "$MANIFEST" \
              --region ${AWS_REGION} 2>&1 | grep -q "ImageAlreadyExistsException" || \
              { err=$?; [ $err -eq 1 ] || exit $err; }
          done