Skip to content

continuous-delivery #4084

continuous-delivery

continuous-delivery #4084

# This workflow executes the E2E Test Suite for a series of combinations that
# represent different execution environments
name: continuous-delivery
on:
issue_comment:
type: [created]
# Manually or triggered by another workflow
workflow_dispatch:
inputs:
depth:
description: 'Depth (push, pull_request, main (default), schedule)'
required: true
default: 'main'
limit:
description: 'Limit to the specified engines list (local, eks, aks, gke, openshift)'
required: false
test_level:
description: 'Test level: 0(highest) to 4(lowest). Default is 4.'
required: false
default: '4'
feature_type:
description: >
Feature Type (disruptive, performance, upgrade, smoke, basic, service-connectivity, self-healing,
backup-restore, snapshot, operator, observability, replication, plugin, postgres-configuration,
pod-scheduling, cluster-metadata, recovery, importing-databases, storage, security, maintenance,
tablespaces)
required: false
log_level:
description: 'Log level for operator (error, warning, info, debug(default), trace)'
required: false
default: 'debug'
schedule:
- cron: '0 1 * * *'
# set up environment variables to be used across all the jobs
env:
GOLANG_VERSION: "1.23.x"
KUBEBUILDER_VERSION: "2.3.1"
KIND_VERSION: "v0.26.0"
ROOK_VERSION: "v1.16.0"
EXTERNAL_SNAPSHOTTER_VERSION: "v8.2.0"
OPERATOR_IMAGE_NAME: "ghcr.io/${{ github.repository }}-testing"
BUILD_PUSH_PROVENANCE: ""
BUILD_PUSH_CACHE_FROM: ""
BUILD_PUSH_CACHE_TO: ""
REGISTRY: "ghcr.io"
REGISTRY_USER: ${{ github.actor }}
REGISTRY_PASSWORD: ${{ secrets.GITHUB_TOKEN }}
REPOSITORY_OWNER: "cloudnative-pg"
SLACK_USERNAME: "cnpg-bot"
BUILD_MANAGER_RELEASE_ARGS: "build --skip=validate --clean --id manager"
# Keep in mind that adding more platforms (architectures) will increase the building
# time even if we use the ghcache for the building process.
PLATFORMS: "linux/amd64,linux/arm64"
E2E_SUFFIX: "cnpge2e"
defaults:
run:
# default failure handling for shell scripts in 'run' steps
shell: 'bash -Eeuo pipefail -x {0}'
jobs:
# Trigger the workflow on release-* branches for smoke testing whenever it's a scheduled run.
# Note: this is a workaround since we can't directly schedule-run a workflow from a non default branch
smoke_test_release_branches:
runs-on: ubuntu-24.04
name: smoke test release-* branches when it's a scheduled run
if: github.event_name == 'schedule'
strategy:
fail-fast: false
matrix:
branch: [release-1.22, release-1.23, release-1.24]
steps:
- name: Invoke workflow with inputs
uses: benc-uk/workflow-dispatch@v1
with:
workflow: continuous-delivery
ref: ${{ matrix.branch }}
inputs: '{ "depth": "push", "limit": "local", "test_level": "4", "log_level": "debug" }'
check_commenter:
if: |
github.event_name == 'issue_comment' &&
github.event.issue.pull_request &&
startsWith(github.event.comment.body, '/test')
name: Retrieve command
runs-on: ubuntu-24.04
outputs:
github_ref: ${{ steps.refs.outputs.head_sha }}
depth: ${{ env.DEPTH }}
limit: ${{ env.LIMIT }}
test_level: ${{ env.TEST_LEVEL }}
feature_type: ${{ env.FEATURE_TYPE }}
log_level: ${{ env.LOG_LEVEL }}
steps:
- name: Check for Command
id: command
uses: xt0rted/slash-command-action@v2
continue-on-error: false
with:
command: test
reaction: "true"
reaction-type: "eyes"
allow-edits: "false"
permission-level: write
- name: Process arguments
id: args
run: |
ARGS="${{ steps.command.outputs.command-arguments }}"
# Set the defaults
DEPTH="main"
LIMIT="local"
TEST_LEVEL="4"
FEATURE_TYPE=""
LOG_LEVEL="debug"
for ARG in $ARGS; do
IFS='=' read name value <<< $ARG
case "${name}" in
"depth"|"d")
DEPTH="${value}"
;;
"limit"|"l")
LIMIT="${value}"
;;
"test_level"|"level"|"tl")
TEST_LEVEL="${value}"
;;
"feature_type"|"type"|"ft")
FEATURE_TYPE="${value}"
;;
"log_level"|"ll")
LOG_LEVEL="${value}"
;;
*)
;;
esac
done
echo "DEPTH=${DEPTH}" >> $GITHUB_ENV
echo "LIMIT=${LIMIT}" >> $GITHUB_ENV
echo "TEST_LEVEL=${TEST_LEVEL}" >> $GITHUB_ENV
echo "FEATURE_TYPE=${FEATURE_TYPE}" >> $GITHUB_ENV
echo "LOG_LEVEL=${LOG_LEVEL}" >> $GITHUB_ENV
- name: Resolve Git reference
uses: xt0rted/pull-request-comment-branch@v3
id: refs
- name: Create comment
uses: peter-evans/create-or-update-comment@v4
with:
token: ${{ secrets.GITHUB_TOKEN }}
repository: ${{ github.repository }}
issue-number: ${{ github.event.issue.number }}
body: |
@${{ github.actor }}, here's the link to the E2E on CNPG workflow run: https://github.com/${{ github.repository }}/actions/runs/${{ github.run_id }}
test_arguments:
name: Parse arguments
if: |
github.event_name == 'workflow_dispatch' || github.event_name == 'schedule'
runs-on: ubuntu-24.04
outputs:
github_ref: ${{ github.ref }}
depth: ${{ env.DEPTH }}
limit: ${{ env.LIMIT }}
test_level: ${{ env.TEST_LEVEL }}
feature_type: ${{ env.FEATURE_TYPE }}
log_level: ${{ env.LOG_LEVEL }}
steps:
- name: Parse input to env
run: |
# Set the defaults for workflow dispatch
if [[ ${{ github.event_name }} == 'workflow_dispatch' ]]; then
DEPTH=${{ github.event.inputs.depth }}
LIMIT=${{ github.event.inputs.limit }}
TEST_LEVEL=${{ github.event.inputs.test_level }}
FEATURE_TYPE="${{ github.event.inputs.feature_type }}"
LOG_LEVEL="${{ github.event.inputs.log_level }}"
fi
# Set the defaults for schedule dispatch
if [[ ${{ github.event_name }} == 'schedule' ]]; then
DEPTH="schedule"
LIMIT=""
TEST_LEVEL="4"
FEATURE_TYPE=""
LOG_LEVEL="debug"
fi
echo "DEPTH=${DEPTH}" >> $GITHUB_ENV
echo "LIMIT=${LIMIT}" >> $GITHUB_ENV
echo "TEST_LEVEL=${TEST_LEVEL}" >> $GITHUB_ENV
echo "FEATURE_TYPE=${FEATURE_TYPE}" >> $GITHUB_ENV
echo "LOG_LEVEL=${LOG_LEVEL}" >> $GITHUB_ENV
evaluate_options:
name: Evaluate workflow options
needs:
- check_commenter
- test_arguments
runs-on: ubuntu-24.04
if: |
(
needs.check_commenter.result == 'success' ||
needs.test_arguments.result == 'success'
) &&
!cancelled()
outputs:
git_ref: ${{ env.GITHUB_REF }}
depth: ${{ env.DEPTH }}
limit: ${{ env.LIMIT }}
test_level: ${{ env.TEST_LEVEL }}
feature_type: ${{ env.FEATURE_TYPE }}
log_level: ${{ env.LOG_LEVEL }}
steps:
- name: From command
run: |
if [[ ${{ github.event_name }} == 'workflow_dispatch' ]] || [[ ${{ github.event_name }} == 'schedule' ]]; then
echo 'GITHUB_REF=${{ needs.test_arguments.outputs.github_ref }}' >> $GITHUB_ENV
echo 'DEPTH=${{ needs.test_arguments.outputs.depth }}' >> $GITHUB_ENV
echo 'LIMIT=${{ needs.test_arguments.outputs.limit }}' >> $GITHUB_ENV
echo 'TEST_LEVEL=${{ needs.test_arguments.outputs.test_level }}' >> $GITHUB_ENV
echo 'FEATURE_TYPE=${{ needs.test_arguments.outputs.feature_type }}' >> $GITHUB_ENV
echo 'LOG_LEVEL=${{ needs.test_arguments.outputs.log_level }}' >> $GITHUB_ENV
fi
if [[ ${{ github.event_name }} == 'issue_comment' ]]; then
echo 'GITHUB_REF=${{ needs.check_commenter.outputs.github_ref }}' >> $GITHUB_ENV
echo 'DEPTH=${{ needs.check_commenter.outputs.depth }}' >> $GITHUB_ENV
echo 'LIMIT=${{ needs.check_commenter.outputs.limit }}' >> $GITHUB_ENV
echo 'TEST_LEVEL=${{ needs.check_commenter.outputs.test_level }}' >> $GITHUB_ENV
echo 'FEATURE_TYPE=${{ needs.check_commenter.outputs.feature_type }}' >> $GITHUB_ENV
echo 'LOG_LEVEL=${{ needs.check_commenter.outputs.log_level }}' >> $GITHUB_ENV
fi
buildx:
name: Build containers
needs:
- check_commenter
- test_arguments
- evaluate_options
if: |
always() && !cancelled() &&
needs.evaluate_options.result == 'success'
runs-on: ubuntu-24.04
permissions:
contents: read
packages: write
pull-requests: read
outputs:
image: ${{ steps.image-meta.outputs.image }}
# 'branch_name' is used in 'GetMostRecentReleaseTag' in the Go code
branch_name: ${{ steps.build-meta.outputs.branch_name }}
upload_artifacts: ${{ steps.build-meta.outputs.upload_artifacts }}
commit_msg: ${{ steps.build-meta.outputs.commit_msg }}
commit_sha: ${{ steps.build-meta.outputs.commit_sha }}
author_name: ${{ steps.build-meta.outputs.author_name }}
author_email: ${{ steps.build-meta.outputs.author_email }}
controller_img: ${{ env.CONTROLLER_IMG }}
controller_img_ubi8: ${{ env.CONTROLLER_IMG_UBI8 }}
bundle_img: ${{ env.BUNDLE_IMG }}
catalog_img: ${{ env.CATALOG_IMG }}
steps:
-
name: Checkout
uses: actions/checkout@v4
with:
ref: ${{ needs.evaluate_options.outputs.git_ref }}
# To identify the commit we need the history and all the tags.
fetch-depth: 0
-
name: Install Go
uses: actions/setup-go@v5
with:
go-version: ${{ env.GOLANG_VERSION }}
check-latest: true
-
name: Build meta
id: build-meta
env:
GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
run: |
images='${{ env.OPERATOR_IMAGE_NAME }}'
tags=''
labels=''
commit_sha=${{ needs.evaluate_options.outputs.git_ref }}
commit_date=$(git log -1 --pretty=format:'%ad' --date short "${commit_sha}" || : )
# use git describe to get the nearest tag and use that to build the version (e.g. 1.4.0-dev24 or 1.4.0)
commit_version=$(git describe --tags --match 'v*' "${commit_sha}"| sed -e 's/^v//; s/-g[0-9a-f]\+$//; s/-\([0-9]\+\)$/-dev\1/')
# shortened commit sha
commit_short=$(git rev-parse --short "${commit_sha}")
# multiline strings are weird
commit_message=$(git show -s --format=%B "${commit_sha}")
commit_message=${commit_message//$'%'/'%25'}
commit_message=${commit_message//$'\n'/'%0A'}
commit_message=${commit_message//$'\r'/'%0D'}
# get git user and email
author_name=$(git show -s --format='%an' "${commit_sha}")
author_email=$(git show -s --format='%ae' "${commit_sha}")
# extract branch name
if [[ ${{ github.event_name }} == 'workflow_dispatch' ]] || [[ ${{ github.event_name }} == 'schedule' ]]
then
branch_name=${GITHUB_REF#refs/heads/}
fi
if [[ ${{ github.event_name }} == 'issue_comment' ]]
then
branch_name=$(gh pr view "${{ github.event.issue.number }}" --json headRefName -q '.headRefName' 2>/dev/null)
fi
# extract tag from branch name
tag_name=$(echo "$branch_name" | sed 's/[^a-zA-Z0-9]/-/g')
upload_artifacts=false
if [[ ${branch_name} == main || ${branch_name} =~ ^release- ]]; then
upload_artifacts=true
fi
echo "IMAGES=${images}" >> $GITHUB_ENV
echo "TAGS=${tags}" >> $GITHUB_ENV
echo "LABELS=${labels}" >> $GITHUB_ENV
echo "DATE=${commit_date}" >> $GITHUB_ENV
echo "VERSION=${commit_version}" >> $GITHUB_ENV
echo "COMMIT=${commit_short}" >> $GITHUB_ENV
echo "commit_sha=${commit_sha}" >> $GITHUB_OUTPUT
echo "commit_msg=${commit_message}" >> $GITHUB_OUTPUT
echo "author_name=${author_name}" >> $GITHUB_OUTPUT
echo "author_email=${author_email}" >> $GITHUB_OUTPUT
echo "branch_name=${branch_name}" >> $GITHUB_OUTPUT
echo "tag_name=${tag_name,,}" >> $GITHUB_OUTPUT
echo "upload_artifacts=${upload_artifacts}" >> $GITHUB_OUTPUT
-
name: Set GoReleaser environment
run: |
echo GOPATH=$(go env GOPATH) >> $GITHUB_ENV
echo PWD=$(pwd) >> $GITHUB_ENV
-
name: Run GoReleaser
uses: goreleaser/goreleaser-action@v6
with:
distribution: goreleaser
version: v2
args: ${{ env.BUILD_MANAGER_RELEASE_ARGS }}
env:
DATE: ${{ env.DATE }}
COMMIT: ${{ env.COMMIT }}
VERSION: ${{ env.VERSION }}
-
name: Docker meta
id: docker-meta
uses: docker/metadata-action@v5
with:
images: ${{ env.IMAGES }}
tags: |
type=raw,value=${{ steps.build-meta.outputs.tag_name }}
-
name: Docker meta UBI8
id: docker-meta-ubi8
uses: docker/metadata-action@v5
with:
images: ${{ env.IMAGES }}
flavor: |
suffix=-ubi8
tags: |
type=raw,value=${{ steps.build-meta.outputs.tag_name }}
-
name: Set up QEMU
uses: docker/setup-qemu-action@v3
with:
platforms: ${{ env.PLATFORMS }}
-
name: Set up Docker Buildx
uses: docker/setup-buildx-action@v3
-
name: Login into docker registry
uses: docker/login-action@v3
with:
registry: ${{ env.REGISTRY }}
username: ${{ env.REGISTRY_USER }}
password: ${{ env.REGISTRY_PASSWORD }}
-
name: Build and push
uses: docker/build-push-action@v6
with:
platforms: ${{ env.PLATFORMS }}
context: .
file: Dockerfile
push: true
build-args: |
VERSION=${{ env.VERSION }}
tags: ${{ steps.docker-meta.outputs.tags }}
labels: ${{ env.LABELS }}
provenance: ${{ env.BUILD_PUSH_PROVENANCE }}
cache-from: ${{ env.BUILD_PUSH_CACHE_FROM }}
cache-to: ${{ env.BUILD_PUSH_CACHE_TO }}
-
name: Build and push UBI8
uses: docker/build-push-action@v6
with:
platforms: ${{ env.PLATFORMS }}
context: .
file: Dockerfile-ubi8
push: true
build-args: |
VERSION=${{ env.VERSION }}
tags: ${{ steps.docker-meta-ubi8.outputs.tags }}
labels: ${{ env.LABELS }}
provenance: ${{ env.BUILD_PUSH_PROVENANCE }}
cache-from: ${{ env.BUILD_PUSH_CACHE_FROM }}
cache-to: ${{ env.BUILD_PUSH_CACHE_TO }}
-
name: Image Meta
id: image-meta
env:
TAGS: ${{ steps.docker-meta.outputs.tags }}
run: |
# If there is more than one tag, take the first one
# TAGS could be separated by newlines or commas
image=$(sed -n '1{s/,.*//; p}' <<< "$TAGS")
echo "image=${image}" >> $GITHUB_OUTPUT
-
name: Output images
env:
TAGS: ${{ steps.docker-meta.outputs.tags }}
TAGS_UBI8: ${{ steps.docker-meta-ubi8.outputs.tags }}
run: |
LOWERCASE_OPERATOR_IMAGE_NAME=${OPERATOR_IMAGE_NAME,,}
TAG=${TAGS#*:}
TAG_UBI=${TAGS_UBI8#*:}
echo "CONTROLLER_IMG=${LOWERCASE_OPERATOR_IMAGE_NAME}:${TAG}" >> $GITHUB_ENV
echo "CONTROLLER_IMG_UBI8=${LOWERCASE_OPERATOR_IMAGE_NAME}:${TAG_UBI}" >> $GITHUB_ENV
echo "BUNDLE_IMG=${LOWERCASE_OPERATOR_IMAGE_NAME}:bundle-${TAG}" >> $GITHUB_ENV
echo "CATALOG_IMG=${LOWERCASE_OPERATOR_IMAGE_NAME}:catalog-${TAG}" >> $GITHUB_ENV
-
name: Generate manifest for operator deployment
id: generate-manifest
env:
CONTROLLER_IMG: ${{ steps.image-meta.outputs.image }}
run: |
make generate-manifest
-
name: Upload the operator manifest as artifact in workflow
uses: actions/upload-artifact@v4
with:
name: operator-manifest.yaml
path: dist/operator-manifest.yaml
retention-days: 7
-
# In order to test the case of upgrading from the current operator
# to a future one, we build and push an image with a different VERSION
# to force a different hash for the manager binary.
# (Otherwise the ONLINE upgrade won't trigger)
#
# NOTE: we only fire this in TEST DEPTH = 4, as that is the level of the
# upgrade test
name: Build binary for upgrade test
uses: goreleaser/goreleaser-action@v6
if: |
always() && !cancelled() &&
needs.evaluate_options.outputs.test_level == '4'
with:
distribution: goreleaser
version: v2
args: ${{ env.BUILD_MANAGER_RELEASE_ARGS }}
env:
DATE: ${{ env.DATE }}
COMMIT: ${{ env.COMMIT }}
VERSION: ${{ env.VERSION }}-prime
-
# In order to test the case of upgrading from the current operator
# to a future one, we build and push an image with a different VERSION
# to force a different hash for the manager binary.
# (Otherwise the ONLINE upgrade won't trigger)
#
# We push the "prime" binary using a tag with the suffix "-prime"
# NOTE: we only fire this in TEST DEPTH = 4, as that is the level of the
# upgrade test
name: Build and push image for upgrade test
uses: docker/build-push-action@v6
if: |
always() && !cancelled() &&
needs.evaluate_options.outputs.test_level == '4'
with:
platforms: ${{ env.PLATFORMS }}
context: .
file: Dockerfile
push: true
build-args: |
VERSION=${{ env.VERSION }}-prime
tags: ${{ steps.docker-meta.outputs.tags }}-prime
labels: ${{ env.LABELS }}
provenance: ${{ env.BUILD_PUSH_PROVENANCE }}
cache-from: ${{ env.BUILD_PUSH_CACHE_FROM }}
cache-to: ${{ env.BUILD_PUSH_CACHE_TO }}
# This will only execute in cloudnative-pg org
publish-artifacts:
name: Publish artifacts
needs:
- buildx
if: |
(always() && !cancelled()) &&
needs.buildx.result == 'success' &&
needs.buildx.outputs.upload_artifacts == 'true' &&
github.repository_owner == 'cloudnative-pg'
runs-on: ubuntu-24.04
steps:
-
name: Checkout artifact
uses: actions/checkout@v4
with:
repository: cloudnative-pg/artifacts
token: ${{ secrets.REPO_GHA_PAT }}
ref: main
fetch-depth: 0
-
name: Configure git user
run: |
git config user.email "${{ needs.buildx.outputs.author_email }}"
git config user.name "${{ needs.buildx.outputs.author_name }}"
-
name: Switch to or create the right branch
env:
BRANCH: ${{ needs.buildx.outputs.branch_name }}
run: |
git checkout "${BRANCH}" 2>/dev/null || git checkout -b "${BRANCH}"
# Remove the previous operator manifest if present because the next
# step doesn't overwrite existing files
rm -fr manifests/operator-manifest.yaml
-
name: Prepare the operator manifest
uses: actions/download-artifact@v4
with:
name: operator-manifest.yaml
path: manifests
-
name: Prepare the commit
env:
COMMIT_MESSAGE: |
${{ needs.buildx.outputs.commit_msg }}
https://github.com/cloudnative-pg/cloudnative-pg/commit/${{ needs.buildx.outputs.commit_sha }}
run: |
# Skip creating the commit if there are no changes
[ -n "$(git status -s)" ] || exit 0
git add .
git commit -m "${COMMIT_MESSAGE}"
-
name: Push changes
uses: ad-m/[email protected]
with:
github_token: ${{ secrets.REPO_GHA_PAT }}
repository: cloudnative-pg/artifacts
branch: ${{ needs.buildx.outputs.branch_name }}
generate-jobs:
name: Generate jobs for E2E tests
needs:
- buildx
- evaluate_options
# We try to avoid running the E2E Test Suite in general, to reduce load on
# GitHub resources.
# Currently, it's executed in the following cases:
# - When dispatched via chatops commands
# - On a push in main and release branches
# - On scheduled executions
if: |
(always() && !cancelled()) &&
needs.buildx.result == 'success'
runs-on: ubuntu-24.04
outputs:
image: ${{ needs.buildx.outputs.image }}
localMatrix: ${{ steps.generate-jobs.outputs.localMatrix }}
localEnabled: ${{ steps.generate-jobs.outputs.localEnabled }}
localTimeout: ${{ steps.generate-jobs.outputs.localE2ETimeout }}
eksMatrix: ${{ steps.generate-jobs.outputs.eksMatrix }}
eksEnabled: ${{ steps.generate-jobs.outputs.eksEnabled }}
eksTimeout: ${{ steps.generate-jobs.outputs.eksE2ETimeout }}
aksMatrix: ${{ steps.generate-jobs.outputs.aksMatrix }}
aksEnabled: ${{ steps.generate-jobs.outputs.aksEnabled }}
aksTimeout: ${{ steps.generate-jobs.outputs.aksE2ETimeout }}
gkeMatrix: ${{ steps.generate-jobs.outputs.gkeMatrix }}
gkeEnabled: ${{ steps.generate-jobs.outputs.gkeEnabled }}
gkeTimeout: ${{ steps.generate-jobs.outputs.gkeE2ETimeout }}
openshiftMatrix: ${{ steps.generate-jobs.outputs.openshiftMatrix }}
openshiftEnabled: ${{ steps.generate-jobs.outputs.openshiftEnabled }}
openshiftTimeout: ${{ steps.generate-jobs.outputs.openshiftE2ETimeout }}
steps:
-
name: Checkout code
uses: actions/checkout@v4
with:
ref: ${{ needs.evaluate_options.outputs.git_ref }}
-
id: generate-jobs
# Generates the jobs that will become different matrix branches,
# according to the event, or to the "depth" parameter if set manually
name: Generate Jobs
shell: bash
run: |
python .github/e2e-matrix-generator.py \
-m '${{ needs.evaluate_options.outputs.depth }}' \
-l '${{ needs.evaluate_options.outputs.limit }}'
e2e-local:
name: Run E2E on local executors
if: |
(always() && !cancelled()) &&
needs.generate-jobs.outputs.localEnabled == 'true' &&
needs.generate-jobs.result == 'success'
needs:
- buildx
- generate-jobs
- evaluate_options
strategy:
fail-fast: false
matrix: ${{ fromJSON(needs.generate-jobs.outputs.localMatrix) }}
runs-on: ubuntu-24.04
env:
# TEST_DEPTH determines the maximum test level the suite should be running
TEST_DEPTH: ${{ needs.evaluate_options.outputs.test_level }}
# FEATURE_TYPE, when defined, determines the subset of E2E tests that will be executed, divided by feature type
FEATURE_TYPE: ${{ needs.evaluate_options.outputs.feature_type }}
K8S_VERSION: "${{ matrix.k8s_version }}"
POSTGRES_VERSION: ${{ matrix.postgres_version }}
POSTGRES_KIND: ${{ matrix.postgres_kind }}
MATRIX: ${{ matrix.id }}
POSTGRES_IMG: "${{ matrix.postgres_img }}"
# The version of operator to upgrade FROM, in the rolling upgrade E2E test
E2E_PRE_ROLLING_UPDATE_IMG: "${{ matrix.postgres_pre_img }}"
TEST_TIMEOUTS: ${{ needs.generate-jobs.outputs.localTimeout }}
BRANCH_NAME: ${{ needs.buildx.outputs.branch_name }}
DEBUG: "true"
BUILD_IMAGE: "false"
CONTROLLER_IMG: ${{ needs.generate-jobs.outputs.image }}
E2E_DEFAULT_STORAGE_CLASS: standard
E2E_CSI_STORAGE_CLASS: csi-hostpath-sc
E2E_DEFAULT_VOLUMESNAPSHOT_CLASS: csi-hostpath-snapclass
LOG_DIR: ${{ github.workspace }}/kind-logs/
DOCKER_REGISTRY_MIRROR: https://mirror.gcr.io
TEST_CLOUD_VENDOR: "local"
steps:
-
name: Cleanup Disk
uses: jlumbroso/free-disk-space@main
with:
android: true
dotnet: true
haskell: true
tool-cache: true
large-packages: false
swap-storage: false
-
name: Cleanup docker cache
run: |
echo "-------------Disk info before cleanup----------------"
df -h
echo "-----------------------------------------------------"
docker system prune -a -f
echo "-------------Disk info after cleanup----------------"
df -h
echo "-----------------------------------------------------"
-
name: Checkout code
uses: actions/checkout@v4
with:
ref: ${{ needs.evaluate_options.outputs.git_ref }}
-
name: Install Go
uses: actions/setup-go@v5
with:
go-version: ${{ env.GOLANG_VERSION }}
check-latest: true
-
## In case hack/setup-cluster.sh need pull operand image from registry
name: Login into docker registry
uses: docker/login-action@v3
with:
registry: ${{ env.REGISTRY }}
username: ${{ env.REGISTRY_USER }}
password: ${{ env.REGISTRY_PASSWORD }}
-
# 'Retry' preparing the E2E test ENV
name: Prepare the environment
uses: nick-fields/retry@v3
with:
timeout_seconds: 300
max_attempts: 3
on_retry_command: |
# Clear-ups before retries
sudo rm -rf /usr/local/bin/kind /usr/local/bin/kubectl
command: |
sudo apt-get update
sudo apt-get install -y gettext-base
sudo hack/setup-cluster.sh prepare /usr/local/bin
-
name: Prepare patch for customization
env:
## the following variable all need be set if we use env_override_customized.yaml.template
## this is customization for local kind
LEADER_ELECTION: "true"
LEADER_LEASE_DURATION: 15
LEADER_RENEW_DEADLINE: 10
LIVENESS_PROBE_THRESHOLD: 3
LOG_LEVEL: ${{ needs.evaluate_options.outputs.log_level }}
run: |
LOG_LEVEL=${LOG_LEVEL:-info}
envsubst < hack/e2e/env_override_customized.yaml.template > config/manager/env_override.yaml
cat config/manager/env_override.yaml
-
name: Run Kind End-to-End tests
env:
ENABLE_APISERVER_AUDIT: true
run:
make e2e-test-kind
-
# Summarize the failed E2E test cases if there are any
name: Report failed E2E tests
if: failure()
run: |
set +x
chmod +x .github/report-failed-test.sh
./.github/report-failed-test.sh
-
# Create an individual artifact for each E2E test, which will be used to
# generate E2E test summary in the follow-up job 'summarize-e2e-tests'
name: Create individual artifact for each E2E test
if: (always() && !cancelled())
env:
RUNNER: "local"
RUN_ID: ${{ github.run_id }}
REPOSITORY: ${{ github.repository }}
GIT_REF: ${{ needs.evaluate_options.outputs.git_ref }}
run: |
set +x
python .github/generate-test-artifacts.py \
-o testartifacts-${{ env.MATRIX }} \
-f tests/e2e/out/report.json \
--environment=true
if [ -f tests/e2e/out/upgrade_report.json ]; then
python .github/generate-test-artifacts.py \
-o testartifacts-${{ env.MATRIX }} \
-f tests/e2e/out/upgrade_report.json \
--environment=true
fi
-
name: Archive test artifacts
if: (always() && !cancelled())
uses: actions/upload-artifact@v4
with:
name: testartifacts-${{ env.MATRIX }}
path: testartifacts-${{ env.MATRIX }}/
retention-days: 7
-
name: Cleanup test artifacts
if: always()
run:
rm -rf testartifacts-${{ env.MATRIX }}/
-
name: Cleanup ginkgo JSON report
# Delete report.json after the analysis. File should always exist.
# Delete upgrade_report.json. It may not exist depending on test level.
if: always()
run: |
if [ -f tests/e2e/out/upgrade_report.json ]; then
rm tests/e2e/out/upgrade_report.json
fi
if [ -f tests/e2e/out/report.json ]; then
rm tests/e2e/out/report.json
fi
-
# Archive logs for failed test cases if there are any
name: Archive Kind logs
if: failure()
uses: actions/upload-artifact@v4
with:
name: kind-logs-${{ matrix.id }}
path: kind-logs/
retention-days: 7
-
name: Archive e2e failure contexts
if: failure()
uses: actions/upload-artifact@v4
with:
name: test-failure-contexts-${{ matrix.id }}
path: |
tests/*/out/
retention-days: 7
if-no-files-found: ignore
-
name: Archive e2e logs
if: failure()
uses: actions/upload-artifact@v4
with:
name: cluster-logs-${{ matrix.id }}
path: |
tests/e2e/cluster_logs/**
retention-days: 7
if-no-files-found: ignore
# AKS Secrets required
# secrets.AZURE_CREDENTIALS
# secrets.AZURE_SUBSCRIPTION
# secrets.AZURE_RESOURCEGROUP
# secrets.AZURE_RESOURCENAME
# secrets.AZURE_WORKSPACE_RESOURCE_ID
e2e-aks-setup:
name: Setup shared resources for Microsoft AKS E2Es
if: |
(always() && !cancelled()) &&
vars.AKS_ENABLED == 'true' &&
needs.generate-jobs.outputs.aksEnabled == 'true' &&
needs.generate-jobs.result == 'success'
needs:
- buildx
- generate-jobs
- evaluate_options
runs-on: ubuntu-24.04
outputs:
azure_storage_account: ${{ steps.setup.outputs.azure_storage_account }}
steps:
-
name: Azure Login
uses: azure/[email protected]
with:
creds: ${{ secrets.AZURE_CREDENTIALS }}
-
name: Create AKS shared resources
uses: nick-fields/retry@v3
id: setup
with:
timeout_minutes: 10
max_attempts: 3
command: |
az extension add --allow-preview true --name aks-preview
az account set --subscription ${{ secrets.AZURE_SUBSCRIPTION }}
AZURE_STORAGE_ACCOUNT="${{ github.run_number }}${{ env.E2E_SUFFIX }}"
az storage account create \
--resource-group ${{ secrets.AZURE_RESOURCEGROUP }} \
--name ${AZURE_STORAGE_ACCOUNT} \
--sku Standard_LRS -o none
# Output storage account name
echo "azure_storage_account=${AZURE_STORAGE_ACCOUNT}" >> $GITHUB_OUTPUT
e2e-aks:
name: Run E2E on Microsoft AKS
if: |
(always() && !cancelled()) &&
vars.AKS_ENABLED == 'true' &&
needs.generate-jobs.outputs.aksEnabled == 'true' &&
needs.generate-jobs.result == 'success' &&
needs.e2e-aks-setup.result == 'success'
needs:
- buildx
- generate-jobs
- evaluate_options
- e2e-aks-setup
strategy:
fail-fast: false
max-parallel: 8
matrix: ${{ fromJSON(needs.generate-jobs.outputs.aksMatrix) }}
runs-on: ubuntu-24.04
env:
# TEST_DEPTH determines the maximum test level the suite should be running
TEST_DEPTH: ${{ needs.evaluate_options.outputs.test_level }}
# FEATURE_TYPE, when defined, determines the subset of E2E tests that will be executed, divided by feature type
FEATURE_TYPE: ${{ needs.evaluate_options.outputs.feature_type }}
K8S_VERSION: "${{ matrix.k8s_version }}"
POSTGRES_VERSION: ${{ matrix.postgres_version }}
POSTGRES_KIND: ${{ matrix.postgres_kind }}
MATRIX: ${{ matrix.id }}
POSTGRES_IMG: "${{ matrix.postgres_img }}"
# The version of operator to upgrade FROM, in the rolling upgrade E2E test
E2E_PRE_ROLLING_UPDATE_IMG: "${{ matrix.postgres_pre_img }}"
TEST_TIMEOUTS: ${{ needs.generate-jobs.outputs.aksTimeout }}
BRANCH_NAME: ${{ needs.buildx.outputs.branch_name }}
AZURE_STORAGE_ACCOUNT: ${{ needs.e2e-aks-setup.outputs.azure_storage_account }}
# AZURE_STORAGE_KEY: this one is gathered during a subsequent step
DEBUG: "true"
BUILD_IMAGE: "false"
CONTROLLER_IMG: ${{ needs.generate-jobs.outputs.image }}
E2E_DEFAULT_STORAGE_CLASS: rook-ceph-block
E2E_CSI_STORAGE_CLASS: rook-ceph-block
E2E_DEFAULT_VOLUMESNAPSHOT_CLASS: csi-rbdplugin-snapclass
TEST_CLOUD_VENDOR: "aks"
steps:
-
name: Checkout code
uses: actions/checkout@v4
with:
ref: ${{ needs.evaluate_options.outputs.git_ref }}
-
name: Install Go
uses: actions/setup-go@v5
with:
go-version: ${{ env.GOLANG_VERSION }}
check-latest: true
-
name: Prepare the environment
uses: nick-fields/retry@v3
with:
timeout_seconds: 300
max_attempts: 3
command: |
sudo apt-get update
sudo apt-get install -y gettext-base
-
name: Install ginkgo
uses: nick-fields/retry@v3
with:
timeout_minutes: 1
max_attempts: 3
command: |
go install github.com/onsi/ginkgo/v2/ginkgo
-
## In case hack/setup-cluster.sh need pull operand image from registry
name: Login into docker registry
uses: docker/login-action@v3
with:
registry: ${{ env.REGISTRY }}
username: ${{ env.REGISTRY_USER }}
password: ${{ env.REGISTRY_PASSWORD }}
-
name: Azure Login
uses: azure/[email protected]
with:
creds: ${{ secrets.AZURE_CREDENTIALS }}
-
name: Install kubectl
uses: azure/setup-kubectl@v4
with:
version: v${{ env.K8S_VERSION }}
-
name: Create AKS cluster
uses: nick-fields/retry@v3
with:
timeout_minutes: 10
max_attempts: 3
command: |
az extension add --allow-preview true --name aks-preview
az account set --subscription ${{ secrets.AZURE_SUBSCRIPTION }}
# name of the AKS cluster
AZURE_AKS="${{ secrets.AZURE_RESOURCENAME }}-${{ github.run_number }}-$( echo ${{ matrix.id }} | tr -d '_.-' )"
echo "AZURE_AKS=${AZURE_AKS}" >> $GITHUB_ENV
# gather the storage account Key
AZURE_STORAGE_KEY=$(az storage account keys list -g "${{ secrets.AZURE_RESOURCEGROUP }}" -n "${{ env.AZURE_STORAGE_ACCOUNT }}" --query "[0].value" -o tsv)
echo "::add-mask::$AZURE_STORAGE_KEY"
echo "AZURE_STORAGE_KEY=${AZURE_STORAGE_KEY}" >> $GITHUB_ENV
# name of the cluster's blob container in the storage account
AZURE_BLOB_CONTAINER="$( echo ${{ matrix.id }} | tr -d '_.-' | tr '[:upper:]' '[:lower:]' )"
echo "AZURE_BLOB_CONTAINER=${AZURE_BLOB_CONTAINER}" >> $GITHUB_ENV
# create and login to the AKS cluster
az aks create --resource-group ${{ secrets.AZURE_RESOURCEGROUP }} \
--name ${AZURE_AKS} \
--tier standard \
--node-count 3 -k v${K8S_VERSION} --generate-ssh-keys --enable-addons monitoring \
--workspace-resource-id ${{ secrets.AZURE_WORKSPACE_RESOURCE_ID }} \
--aks-custom-headers EnableAzureDiskFileCSIDriver=true
az aks get-credentials --resource-group ${{ secrets.AZURE_RESOURCEGROUP }} \
--name ${AZURE_AKS}
# create diagnostic settings for monitoring kube-apiserver logs
AKS_CLUSTER_RESOURCE_ID=$(az aks show --resource-group ${{ secrets.AZURE_RESOURCEGROUP }} --name ${AZURE_AKS} --query id -o tsv --only-show-errors)
az monitor diagnostic-settings create \
--resource-group ${{ secrets.AZURE_RESOURCEGROUP }} \
--resource ${AKS_CLUSTER_RESOURCE_ID} \
--name diagnostic-kube-apiserver-logs \
--workspace ${{ secrets.AZURE_WORKSPACE_RESOURCE_ID }} \
--logs '[ { "category": "kube-apiserver", "enabled": true } ]'
-
# Azure is slow in provisioning disks, and we can't wait two minutes
# every time we create a pod, otherwise all the tests will time out.
# We set up a few large disks now, we run Rook on top of them and we
# use rook to get the small PV we use in the tests.
# It can still take a while to deploy rook.
name: Set up Rook
uses: nick-fields/retry@v3
with:
timeout_minutes: 27
max_attempts: 1
command: |
STORAGECLASSNAME=default
go install github.com/mikefarah/yq/v4@v4
ROOK_BASE_URL=https://raw.githubusercontent.com/rook/rook/${{ env.ROOK_VERSION }}/deploy/examples
kubectl apply -f ${ROOK_BASE_URL}/crds.yaml
kubectl apply -f ${ROOK_BASE_URL}/common.yaml
kubectl apply -f ${ROOK_BASE_URL}/operator.yaml
curl ${ROOK_BASE_URL}/cluster-on-pvc.yaml | \
sed '/^ *#/d;/^ *$/d' | \
yq e ".spec.storage.storageClassDeviceSets[].volumeClaimTemplates[].spec.resources.requests.storage = \"50Gi\" |
.spec.storage.storageClassDeviceSets[].volumeClaimTemplates[].spec.storageClassName = \"${STORAGECLASSNAME}\" |
.spec.mon.volumeClaimTemplate.spec.storageClassName = \"${STORAGECLASSNAME}\" " - | \
kubectl apply -f -
while true; do
output=$( kubectl get deploy -n rook-ceph -l app=rook-ceph-osd --no-headers -o name )
if [[ $(wc -w <<< $output) == 3 ]]; then
break
fi
done
echo "Waiting for Rook OSDs to be available"
kubectl wait deploy -n rook-ceph --for condition=available --timeout 480s -l app=rook-ceph-osd
kubectl apply -f ${ROOK_BASE_URL}/csi/rbd/storageclass.yaml
kubectl apply -f ${ROOK_BASE_URL}/csi/rbd/snapshotclass.yaml
kubectl annotate storageclass ${{env.E2E_DEFAULT_STORAGE_CLASS}} storage.kubernetes.io/default-snapshot-class=${{env.E2E_DEFAULT_VOLUMESNAPSHOT_CLASS}} --overwrite
-
name: Prepare patch for customization
env:
## the following variable all need be set if we use env_override_customized.yaml.template
## this is customization for aks
LEADER_ELECTION: "true"
LEADER_LEASE_DURATION: 15
LEADER_RENEW_DEADLINE: 10
LIVENESS_PROBE_THRESHOLD: 3
LOG_LEVEL: ${{ needs.evaluate_options.outputs.log_level }}
run: |
LOG_LEVEL=${LOG_LEVEL:-info}
envsubst < hack/e2e/env_override_customized.yaml.template > config/manager/env_override.yaml
cat config/manager/env_override.yaml
-
name: Run E2E tests
run: hack/e2e/run-e2e.sh
-
# Summarize the failed E2E test cases if there are any
name: Report failed E2E tests
if: failure()
run: |
set +x
chmod +x .github/report-failed-test.sh
./.github/report-failed-test.sh
-
# Create an individual artifact for each E2E test, which will be used to
# generate E2E test summary in the follow-up job 'summarize-e2e-tests'
name: Create individual artifact for each E2E test
if: (always() && !cancelled())
env:
RUNNER: "aks"
RUN_ID: ${{ github.run_id }}
REPOSITORY: ${{ github.repository }}
GIT_REF: ${{ needs.evaluate_options.outputs.git_ref }}
run: |
set +x
python .github/generate-test-artifacts.py \
-o testartifacts-${{ env.MATRIX }} \
-f tests/e2e/out/report.json \
--environment=true
if [ -f tests/e2e/out/upgrade_report.json ]; then
python .github/generate-test-artifacts.py \
-o testartifacts-${{ env.MATRIX }} \
-f tests/e2e/out/upgrade_report.json \
--environment=true
fi
-
name: Archive test artifacts
if: (always() && !cancelled())
uses: actions/upload-artifact@v4
with:
name: testartifacts-${{ env.MATRIX }}
path: testartifacts-${{ env.MATRIX }}/
retention-days: 7
-
name: Cleanup test artifacts
if: always()
run:
rm -rf testartifacts-${{ env.MATRIX }}/
-
name: Cleanup ginkgo JSON report
# Delete report.json after the analysis. File should always exist.
# Delete upgrade_report.json. It may not exist depending on test level.
if: always()
run: |
if [ -f tests/e2e/out/upgrade_report.json ]; then
rm tests/e2e/out/upgrade_report.json
fi
if [ -f tests/e2e/out/report.json ]; then
rm tests/e2e/out/report.json
fi
-
name: Archive e2e failure contexts
if: failure()
uses: actions/upload-artifact@v4
with:
name: test-failure-contexts-${{ matrix.id }}
path: |
tests/*/out/
retention-days: 7
if-no-files-found: ignore
-
name: Archive e2e logs
if: failure()
uses: actions/upload-artifact@v4
with:
name: cluster-logs-${{ matrix.id }}
path: |
tests/e2e/cluster_logs/**
retention-days: 7
if-no-files-found: ignore
-
name: Clean up
if: always()
run: |
set +e
az extension add --allow-preview true --name monitor-control-service
az account set --subscription ${{ secrets.AZURE_SUBSCRIPTION }}
attempt=1
max_attempts=3
while [ "${attempt}" -le "${max_attempts}" ]; do
echo "Deleting cluster. Attempt ${attempt} of ${max_attempts}"
az aks delete --resource-group ${{ secrets.AZURE_RESOURCEGROUP }} -y --name ${{ env.AZURE_AKS }}
status=$?
if [[ $status == 0 ]]; then
echo "AKS cluster deleted"
break
fi
echo "Failed deleting cluster ${{ env.AZURE_AKS }}, retrying"
sleep 5
attempt=$((attempt+1))
done
attempt=1
AZURE_RESOURCEGROUP_LOCATION="$( az group show --resource-group ${{ secrets.AZURE_RESOURCEGROUP }} --query location -o tsv --only-show-errors )"
DATA_COLL_RULE_NAME="MSCI-${AZURE_RESOURCEGROUP_LOCATION}-${{ env.AZURE_AKS }}"
while [ "${attempt}" -le "${max_attempts}" ]; do
echo "Deleting data-collection rule ${DATA_COLL_RULE_NAME}. Attempt ${attempt} of ${max_attempts}"
az monitor data-collection rule show --name ${DATA_COLL_RULE_NAME} --resource-group ${{ secrets.AZURE_RESOURCEGROUP }} --query name
# if not found, let it go
status=$?
if [[ $status != 0 ]]; then
echo "AKS data-collection rule not found"
break
fi
az monitor data-collection rule delete -y --name ${DATA_COLL_RULE_NAME} --resource-group ${{ secrets.AZURE_RESOURCEGROUP }}
status=$?
if [[ $status == 0 ]]; then
echo "AKS data-collection rule deleted"
break
fi
echo "Failed deleting data-collection rule ${DATA_COLL_RULE_NAME}, retrying"
sleep 5
attempt=$((attempt+1))
done
e2e-aks-teardown:
name: Teardown Microsoft AKS shared resources
if: |
always() &&
vars.AKS_ENABLED == 'true' &&
needs.generate-jobs.outputs.aksEnabled == 'true' &&
needs.generate-jobs.result == 'success' &&
needs.e2e-aks-setup.result == 'success'
needs:
- buildx
- generate-jobs
- e2e-aks-setup
- e2e-aks
runs-on: ubuntu-24.04
env:
AZURE_STORAGE_ACCOUNT: ${{ needs.e2e-aks-setup.outputs.azure_storage_account }}
steps:
-
name: Azure Login
if: always()
uses: azure/[email protected]
with:
creds: ${{ secrets.AZURE_CREDENTIALS }}
-
name: Teardown AKS shared resources
if: always()
uses: nick-fields/retry@v3
with:
timeout_minutes: 5
max_attempts: 3
command: |
az account set --subscription ${{ secrets.AZURE_SUBSCRIPTION }}
az storage account delete -y --resource-group ${{ secrets.AZURE_RESOURCEGROUP }} --name ${{ env.AZURE_STORAGE_ACCOUNT }}
# EKS Secrets required
# secrets.AWS_EKS_ADMIN_IAM_ROLES
# secrets.AWS_ACCESS_KEY_ID
# secrets.AWS_SECRET_ACCESS_KEY
e2e-eks:
name: Run E2E on Amazon EKS
if: |
(always() && !cancelled()) &&
vars.EKS_ENABLED == 'true' &&
needs.generate-jobs.outputs.eksEnabled == 'true' &&
needs.generate-jobs.result == 'success'
needs:
- buildx
- generate-jobs
- evaluate_options
strategy:
fail-fast: false
max-parallel: 6
matrix: ${{ fromJSON(needs.generate-jobs.outputs.eksMatrix) }}
runs-on: ubuntu-24.04
env:
# TEST_DEPTH determines the maximum test level the suite should be running
TEST_DEPTH: ${{ needs.evaluate_options.outputs.test_level }}
# FEATURE_TYPE, when defined, determines the subset of E2E tests that will be executed, divided by feature type
FEATURE_TYPE: ${{ needs.evaluate_options.outputs.feature_type }}
K8S_VERSION: "${{ matrix.k8s_version }}"
POSTGRES_VERSION: ${{ matrix.postgres_version }}
POSTGRES_KIND: ${{ matrix.postgres_kind }}
MATRIX: ${{ matrix.id }}
POSTGRES_IMG: "${{ matrix.postgres_img }}"
# The version of operator to upgrade FROM, in the rolling upgrade E2E test
E2E_PRE_ROLLING_UPDATE_IMG: "${{ matrix.postgres_pre_img }}"
TEST_TIMEOUTS: ${{ needs.generate-jobs.outputs.eksTimeout }}
BRANCH_NAME: ${{ needs.buildx.outputs.branch_name }}
DEBUG: "true"
BUILD_IMAGE: "false"
CONTROLLER_IMG: ${{ needs.generate-jobs.outputs.image }}
E2E_DEFAULT_STORAGE_CLASS: gp3
E2E_CSI_STORAGE_CLASS: gp3
E2E_DEFAULT_VOLUMESNAPSHOT_CLASS: ebs-csi-snapclass
AWS_REGION: eu-central-1
AWS_EKS_ADMIN_IAM_ROLES: ${{ secrets.AWS_EKS_ADMIN_IAM_ROLES }}
TEST_CLOUD_VENDOR: "eks"
steps:
-
name: Set cluster name
run: |
echo "CLUSTER_NAME=${{ env.E2E_SUFFIX }}-test-${{ github.run_number }}-$( echo ${{ matrix.id }} | tr -d '_.-' )" >> $GITHUB_ENV
-
name: Checkout code
uses: actions/checkout@v4
with:
ref: ${{ needs.evaluate_options.outputs.git_ref }}
-
name: Install Go
uses: actions/setup-go@v5
with:
go-version: ${{ env.GOLANG_VERSION }}
check-latest: true
-
## In case hack/setup-cluster.sh need pull operand image from registry
name: Login into docker registry
uses: docker/login-action@v3
with:
registry: ${{ env.REGISTRY }}
username: ${{ env.REGISTRY_USER }}
password: ${{ env.REGISTRY_PASSWORD }}
-
name: Prepare the environment
uses: nick-fields/retry@v3
with:
timeout_seconds: 300
max_attempts: 3
command: |
sudo apt-get update
sudo apt-get install -y gettext-base
-
name: Install ginkgo
uses: nick-fields/retry@v3
with:
timeout_minutes: 1
max_attempts: 3
command: |
go install github.com/onsi/ginkgo/v2/ginkgo
-
name: Configure AWS credentials
uses: aws-actions/configure-aws-credentials@v4
with:
aws-access-key-id: ${{ secrets.AWS_ACCESS_KEY_ID }}
aws-secret-access-key: ${{ secrets.AWS_SECRET_ACCESS_KEY }}
aws-region: ${{ env.AWS_REGION }}
-
name: Install eksctl
uses: nick-fields/retry@v3
with:
timeout_minutes: 1
max_attempts: 3
command: |
mkdir -p "$HOME/.local/bin"
curl -sL "https://github.com/weaveworks/eksctl/releases/latest/download/eksctl_$(uname -s)_amd64.tar.gz" \
| tar xz -C $HOME/.local/bin
echo "$HOME/.local/bin" >> $GITHUB_PATH
-
name: Configure EKS setup
run: |
envsubst < hack/e2e/eks-cluster.yaml.template > hack/e2e/eks-cluster.yaml
-
name: Setup EKS
run: |
# Setting up EKS cluster
echo "create cluster"
eksctl create cluster -f hack/e2e/eks-cluster.yaml
# Create iamidentitymapping
echo "$AWS_EKS_ADMIN_IAM_ROLES" | while read role
do
# Masking variables to hide values
echo "::add-mask::$role"
eksctl create iamidentitymapping --cluster "${CLUSTER_NAME}" --region="${AWS_REGION}" --arn "${role}" --group system:masters --username admin
done
# Updating .kubeconfig to use the correct version of client.authentication.k8s.io API
aws eks update-kubeconfig --name ${CLUSTER_NAME} --region ${AWS_REGION}
# Installing CRD for support volumeSnapshot
SNAPSHOTTER_BASE_URL=https://raw.githubusercontent.com/kubernetes-csi/external-snapshotter/${{env.EXTERNAL_SNAPSHOTTER_VERSION}}
kubectl apply -f ${SNAPSHOTTER_BASE_URL}/client/config/crd/snapshot.storage.k8s.io_volumesnapshotclasses.yaml
kubectl apply -f ${SNAPSHOTTER_BASE_URL}/client/config/crd/snapshot.storage.k8s.io_volumesnapshotcontents.yaml
kubectl apply -f ${SNAPSHOTTER_BASE_URL}/client/config/crd/snapshot.storage.k8s.io_volumesnapshots.yaml
## Controller
kubectl apply -f ${SNAPSHOTTER_BASE_URL}/deploy/kubernetes/snapshot-controller/rbac-snapshot-controller.yaml
kubectl apply -f ${SNAPSHOTTER_BASE_URL}/deploy/kubernetes/snapshot-controller/setup-snapshot-controller.yaml
# Install volume snapshot class
kubectl apply -f hack/e2e/volumesnapshotclass-ebs-csi.yaml
kubectl get volumesnapshotclass
# Change to use gp3 as default storage account
kubectl annotate storageclass gp2 storageclass.kubernetes.io/is-default-class=false --overwrite
kubectl apply -f hack/e2e/storage-class-gp3.yaml
kubectl annotate storageclass ${{env.E2E_DEFAULT_STORAGE_CLASS}} storage.kubernetes.io/default-snapshot-class=${{env.E2E_DEFAULT_VOLUMESNAPSHOT_CLASS}} --overwrite
kubectl get storageclass
-
name: Setup Velero
uses: nick-fields/retry@v3
env:
VELERO_VERSION: "v1.15.0"
VELERO_AWS_PLUGIN_VERSION: "v1.11.0"
with:
timeout_minutes: 10
max_attempts: 3
on_retry_command: |
# Clean up buckets
output=$( aws s3api delete-bucket --bucket "${VELERO_BUCKET_NAME}" --region "${AWS_REGION}" 2>&1 )
status=$?
if [[ $status == 0 ]]; then
echo "S3 Bucket deleted"
break
fi
if ( grep "NoSuchBucket" <<< "$output" ); then
echo "S3 Bucket doesn't exist, nothing to remove"
break
fi
# Uninstall Velero
kubectl delete namespace/velero clusterrolebinding/velero
kubectl delete crds -l component=velero
command: |
VELERO_BUCKET_NAME="${CLUSTER_NAME,,}-velero"
echo "VELERO_BUCKET_NAME=${VELERO_BUCKET_NAME}" >> $GITHUB_ENV
# Create S3 bucket
aws s3api create-bucket \
--bucket "${VELERO_BUCKET_NAME}" \
--region "${AWS_REGION}" \
--create-bucket-configuration LocationConstraint="${AWS_REGION}"
# Download Velero, extract and place it in $PATH
curl -sL "https://github.com/vmware-tanzu/velero/releases/download/${VELERO_VERSION}/velero-${VELERO_VERSION}-linux-amd64.tar.gz" | tar xz
mv velero-${VELERO_VERSION}-linux-amd64/velero $HOME/.local/bin
# Set Velero-specific credentials
echo -e "[default]\naws_access_key_id=${{ secrets.AWS_ACCESS_KEY_ID }}\naws_secret_access_key=${{ secrets.AWS_SECRET_ACCESS_KEY }}" >> credentials-velero
# Install Velero
velero install \
--provider aws \
--plugins velero/velero-plugin-for-aws:${VELERO_AWS_PLUGIN_VERSION} \
--bucket "${VELERO_BUCKET_NAME}" \
--backup-location-config region="${AWS_REGION}" \
--snapshot-location-config region="${AWS_REGION}" \
--secret-file ./credentials-velero \
--wait
-
name: Prepare patch for customization
env:
## the following variable all need be set if we use env_override_customized.yaml.template
## this is customization for eks
LEADER_ELECTION: "true"
LEADER_LEASE_DURATION: 15
LEADER_RENEW_DEADLINE: 10
LIVENESS_PROBE_THRESHOLD: 3
LOG_LEVEL: ${{ needs.evaluate_options.outputs.log_level }}
run: |
LOG_LEVEL=${LOG_LEVEL:-info}
envsubst < hack/e2e/env_override_customized.yaml.template > config/manager/env_override.yaml
cat config/manager/env_override.yaml
-
name: Run E2E tests
run: hack/e2e/run-e2e.sh
-
# Summarize the failed E2E test cases if there are any
name: Report failed E2E tests
if: failure()
run: |
set +x
chmod +x .github/report-failed-test.sh
./.github/report-failed-test.sh
-
# Create an individual artifact for each E2E test, which will be used to
# generate E2E test summary in the follow-up job 'summarize-e2e-tests'
name: Create individual artifact for each E2E test
if: (always() && !cancelled())
env:
RUNNER: "eks"
RUN_ID: ${{ github.run_id }}
REPOSITORY: ${{ github.repository }}
GIT_REF: ${{ needs.evaluate_options.outputs.git_ref }}
run: |
set +x
python .github/generate-test-artifacts.py \
-o testartifacts-${{ env.MATRIX }} \
-f tests/e2e/out/report.json \
--environment=true
if [ -f tests/e2e/out/upgrade_report.json ]; then
python .github/generate-test-artifacts.py \
-o testartifacts-${{ env.MATRIX }} \
-f tests/e2e/out/upgrade_report.json \
--environment=true
fi
-
name: Archive test artifacts
if: (always() && !cancelled())
uses: actions/upload-artifact@v4
with:
name: testartifacts-${{ env.MATRIX }}
path: testartifacts-${{ env.MATRIX }}/
retention-days: 7
-
name: Cleanup test artifacts
if: always()
run:
rm -rf testartifacts-${{ env.MATRIX }}/
-
name: Cleanup ginkgo JSON report
# Delete report.json after the analysis. File should always exist.
# Delete upgrade_report.json. It may not exist depending on test level.
if: always()
run: |
if [ -f tests/e2e/out/upgrade_report.json ]; then
rm tests/e2e/out/upgrade_report.json
fi
if [ -f tests/e2e/out/report.json ]; then
rm tests/e2e/out/report.json
fi
-
name: Archive e2e failure contexts
if: failure()
uses: actions/upload-artifact@v4
with:
name: test-failure-contexts-${{ matrix.id }}
path: |
tests/*/out/
retention-days: 7
if-no-files-found: ignore
-
name: Archive e2e logs
if: failure()
uses: actions/upload-artifact@v4
with:
name: cluster-logs-${{ matrix.id }}
path: |
tests/e2e/cluster_logs/**
retention-days: 7
if-no-files-found: ignore
-
name: Clean up
if: always()
run: |
set +e
CLUSTER_NAME="${{ env.CLUSTER_NAME }}"
REGION_NAME="${{ env.AWS_REGION }}"
STACK_NAME="eksctl-${CLUSTER_NAME}-cluster"
CLOUDFORMATION_STATUS_CURRENT=$(aws cloudformation describe-stacks --stack-name "${STACK_NAME}" | jq -r '.Stacks[].StackStatus')
if [[ -z "${CLOUDFORMATION_STATUS_CURRENT}" ]]; then
echo "CloudFormation stack not found. Nothing to cleanup."
exit 0
fi
# Attempt to remove any leftover PDB (and Cluster that would recreate it)
# that could prevent the EKS cluster deletion
kubectl delete cluster --all --all-namespaces --now --timeout=30s || true
kubectl delete pdb --all --all-namespaces --now --timeout=30s || true
kubectl delete pvc --all --all-namespaces --now --timeout=30s || true
# Remove any LoadBalancer service
kubectl get service --all-namespaces -o json | jq -r '.items[] | select(.spec.type=="LoadBalancer") | .metadata | "kubectl delete service --now --timeout=30s -n " + .namespace + " " + .name' | xargs -rI X bash -c X || true
NODEGROUP_STACK_NAMES=$(eksctl get nodegroup --cluster "${CLUSTER_NAME}" -o json | jq -r '.[].StackName' || true)
attempt=1
bucket_attempt=1
max_attempts=3
# Attempting three times to remove the Velero S3 bucket
VELERO_BUCKET_NAME=${VELERO_BUCKET_NAME:-"${CLUSTER_NAME,,}-velero"}
while [ "${bucket_attempt}" -le "${max_attempts}" ]; do
echo "Deleting S3 Bucket. Attempt ${bucket_attempt} of ${max_attempts}"
output=$( aws s3api delete-bucket --bucket "${VELERO_BUCKET_NAME}" --region "${AWS_REGION}" 2>&1 )
status=$?
if [[ $status == 0 ]]; then
echo "S3 Bucket deleted"
break
fi
if ( grep "NoSuchBucket" <<< "$output" ); then
echo "S3 Bucket doesn't exist, nothing to remove"
break
fi
echo "Failed deleting S3 Bucket ${VELERO_BUCKET_NAME}, retrying"
sleep 5
bucket_attempt=$((bucket_attempt+1))
done
# Attempting three times to cleanly remove the cluster via eksctl
while [ "${attempt}" -le "${max_attempts}" ]; do
echo "Deleting cluster. Attempt ${attempt} of ${max_attempts}"
output=$( eksctl delete cluster -n "${CLUSTER_NAME}" -r "${REGION_NAME}" --wait --force 2>&1 )
status=$?
if [[ $status == 0 ]]; then
echo "EKS cluster deleted"
break
fi
if ( grep "ResourceNotFoundException: No cluster found for name: ${CLUSTER_NAME}" <<< "$output" ); then
echo "EKS cluster doesn't exist, nothing to remove"
break
fi
echo "Failed deleting cluster ${CLUSTER_NAME}, retrying"
sleep 5
attempt=$((attempt+1))
done
# Recheck if something got stuck, and use harder methods to clean up
CLOUDFORMATION_STATUS_CURRENT=$(aws cloudformation describe-stacks --stack-name "${STACK_NAME}" | jq -r '.Stacks[].StackStatus')
if [ -n "${CLOUDFORMATION_STATUS_CURRENT}" ] ; then
echo "::warning file=continuous-delivery.yml::eksctl failed deleting a cluster cleanly"
# When the status of CloudFormation stack managed by eksctl reports an error, try to delete resources directly with AWS CLI
pip install boto3
for vpc_id in $(aws ec2 describe-vpcs | jq -r '.Vpcs[] | select(.Tags?[]? | .Key == "Name" and (.Value | contains("'"${STACK_NAME}"'"))).VpcId'); do
python .github/vpc_destroy.py --vpc_id "${vpc_id}" --region "${REGION_NAME}" --services ec2
done
# Then we try to delete the cluster cleanly and the cloudformation
if aws eks describe-cluster --name "${CLUSTER_NAME}" --region "${REGION_NAME}" ; then
eksctl delete cluster -n "${CLUSTER_NAME}" -r "${REGION_NAME}" --wait --force
fi
if [ -n "${NODEGROUP_STACK_NAMES}" ] ; then
for NODEGROUP_STACK_NAME in ${NODEGROUP_STACK_NAMES}; do
if aws cloudformation describe-stacks --stack-name "${NODEGROUP_STACK_NAME}" --region "${REGION_NAME}" ; then
aws cloudformation delete-stack --stack-name "${NODEGROUP_STACK_NAME}" --region "${REGION_NAME}"
fi
done
fi
if aws cloudformation describe-stacks --stack-name "${STACK_NAME}" --region "${REGION_NAME}" ; then
aws cloudformation delete-stack --stack-name "${STACK_NAME}" --region "${REGION_NAME}"
fi
fi
# Clear up leftover volumes
while read -r volume; do
echo "Deleting $volume of cluster $CLUSTER_NAME ..."
if ! aws ec2 delete-volume --region "${REGION_NAME}" --volume-id "$volume" ; then
echo "::warning file=continuous-delivery.yml::Failed deleting $volume of cluster $CLUSTER_NAME"
fi
done < <(aws ec2 describe-volumes --region "${REGION_NAME}" --query 'Volumes[?not_null(Tags[?Key == `kubernetes.io/cluster/'"$CLUSTER_NAME"'` && Value == `owned`].Value)].VolumeId' | jq -r '.[]' || true)
# GKE Secrets required
# secrets.GCP_SERVICE_ACCOUNT
# secrets.GCP_PROJECT_ID
e2e-gke:
name: Run E2E on Google GKE
if: |
(always() && !cancelled()) &&
vars.GKE_ENABLED == 'true' &&
needs.generate-jobs.outputs.gkeEnabled == 'true' &&
needs.generate-jobs.result == 'success'
needs:
- buildx
- generate-jobs
- evaluate_options
strategy:
fail-fast: false
max-parallel: 6
matrix: ${{ fromJSON(needs.generate-jobs.outputs.gkeMatrix) }}
runs-on: ubuntu-24.04
env:
# TEST_DEPTH determines the maximum test level the suite should be running
TEST_DEPTH: ${{ needs.evaluate_options.outputs.test_level }}
# FEATURE_TYPE, when defined, determines the subset of E2E tests that will be executed, divided by feature type
FEATURE_TYPE: ${{ needs.evaluate_options.outputs.feature_type }}
K8S_VERSION: "${{ matrix.k8s_version }}"
POSTGRES_VERSION: ${{ matrix.postgres_version }}
POSTGRES_KIND: ${{ matrix.postgres_kind }}
MATRIX: ${{ matrix.id }}
POSTGRES_IMG: "${{ matrix.postgres_img }}"
# The version of operator to upgrade FROM, in the rolling upgrade E2E test
E2E_PRE_ROLLING_UPDATE_IMG: "${{ matrix.postgres_pre_img }}"
TEST_TIMEOUTS: ${{ needs.generate-jobs.outputs.gkeTimeout }}
BRANCH_NAME: ${{ needs.buildx.outputs.branch_name }}
DEBUG: "true"
BUILD_IMAGE: "false"
CONTROLLER_IMG: ${{ needs.generate-jobs.outputs.image }}
E2E_DEFAULT_STORAGE_CLASS: standard-rwo
E2E_CSI_STORAGE_CLASS: standard-rwo
E2E_DEFAULT_VOLUMESNAPSHOT_CLASS: pd-csi-snapclass
REGION: europe-west3
TEST_CLOUD_VENDOR: "gke"
steps:
-
name: Checkout code
uses: actions/checkout@v4
with:
ref: ${{ needs.evaluate_options.outputs.git_ref }}
-
name: Install Go
uses: actions/setup-go@v5
with:
go-version: ${{ env.GOLANG_VERSION }}
check-latest: true
-
## In case hack/setup-cluster.sh need pull operand image from registry
name: Login into docker registry
uses: docker/login-action@v3
with:
registry: ${{ env.REGISTRY }}
username: ${{ env.REGISTRY_USER }}
password: ${{ env.REGISTRY_PASSWORD }}
-
name: Prepare the environment
uses: nick-fields/retry@v3
with:
timeout_seconds: 300
max_attempts: 3
command: |
sudo apt-get update
sudo apt-get install -y gettext-base
-
name: Install ginkgo
uses: nick-fields/retry@v3
with:
timeout_seconds: 120
max_attempts: 3
command: |
go install github.com/onsi/ginkgo/v2/ginkgo
-
name: Set cluster name
run: |
# GKE cluster names rules:
# only lowercase alphanumerics and '-' allowed, must start with a letter and end with an alphanumeric,
# and must be no longer than 40 characters
# We need to shorten the name and lower the case
SHORT_ID=$( echo ${{ matrix.id }} | tr -d '_.-' | tr '[:upper:]' '[:lower:]')
echo "CLUSTER_NAME=${{ env.E2E_SUFFIX }}-test-${{ github.run_number }}-${SHORT_ID}" >> $GITHUB_ENV
-
name: Authenticate to Google Cloud
id: 'auth'
uses: google-github-actions/auth@v2
with:
credentials_json: '${{ secrets.GCP_SERVICE_ACCOUNT }}'
-
name: Set up Cloud SDK and kubectl
uses: google-github-actions/setup-gcloud@v2
with:
project_id: ${{ secrets.GCP_PROJECT_ID }}
install_components: 'kubectl,gke-gcloud-auth-plugin'
-
name: Create GKE cluster
run: |
set +e
# We may go over the amount of API requests allowed
# by Google when creating all the clusters at the same time.
# We give a few attempts at creating the cluster before giving up.
# The following command will create a 3 nodes cluster, with each
# node deployed in its own availability zone.
for i in `seq 1 5`; do
if gcloud container clusters create ${{ env.CLUSTER_NAME }} \
--num-nodes=1 \
--cluster-version=${{ env.K8S_VERSION }} \
--region=${{ env.REGION }} \
--disk-size=20 \
--machine-type=e2-standard-2 \
--labels=cluster=${{ env.CLUSTER_NAME }}
then
exit 0
fi
echo "Couldn't create the cluster. Retrying in 100s."
sleep 100
done
echo "Couldn't create the cluster. Failing."
exit 1
-
name: Get GKE kubeconfig credentials
env:
USE_GKE_GCLOUD_AUTH_PLUGIN: "True"
run: |
gcloud container clusters get-credentials ${{ env.CLUSTER_NAME }} --region ${{ env.REGION }} --project ${{ secrets.GCP_PROJECT_ID }}
-
name: Configure Storage
run: |
# Install volume snapshot class
kubectl apply -f hack/e2e/volumesnapshotclass-pd-csi.yaml
# Change to use standard-rwo as default storage account
kubectl annotate storageclass ${{env.E2E_DEFAULT_STORAGE_CLASS}} storage.kubernetes.io/default-snapshot-class=${{env.E2E_DEFAULT_VOLUMESNAPSHOT_CLASS}} --overwrite
kubectl get storageclass
-
name: Prepare patch for customization
env:
## the following variable all need be set if we use env_override_customized.yaml.template
## this is customization for gke
LEADER_ELECTION: "false"
LEADER_LEASE_DURATION: 240
LEADER_RENEW_DEADLINE: 230
LIVENESS_PROBE_THRESHOLD: 9
LOG_LEVEL: ${{ needs.evaluate_options.outputs.log_level }}
run: |
LOG_LEVEL=${LOG_LEVEL:-info}
envsubst < hack/e2e/env_override_customized.yaml.template > config/manager/env_override.yaml
cat config/manager/env_override.yaml
-
name: Run E2E tests
run: hack/e2e/run-e2e.sh
-
name: Report failed E2E tests
if: failure()
run: |
set +x
chmod +x .github/report-failed-test.sh
./.github/report-failed-test.sh
-
# Create an individual artifact for each E2E test, which will be used to
# generate E2E test summary in the follow-up job 'summarize-e2e-tests'
name: Create individual artifact for each E2E test
if: (always() && !cancelled())
env:
RUNNER: "gke"
RUN_ID: ${{ github.run_id }}
REPOSITORY: ${{ github.repository }}
GIT_REF: ${{ needs.evaluate_options.outputs.git_ref }}
run: |
set +x
python .github/generate-test-artifacts.py \
-o testartifacts-${{ env.MATRIX }} \
-f tests/e2e/out/report.json \
--environment=true
if [ -f tests/e2e/out/upgrade_report.json ]; then
python .github/generate-test-artifacts.py \
-o testartifacts-${{ env.MATRIX }} \
-f tests/e2e/out/upgrade_report.json \
--environment=true
fi
-
name: Archive test artifacts
if: (always() && !cancelled())
uses: actions/upload-artifact@v4
with:
name: testartifacts-${{ env.MATRIX }}
path: testartifacts-${{ env.MATRIX }}/
retention-days: 7
-
name: Cleanup test artifacts
if: always()
run:
rm -rf testartifacts-${{ env.MATRIX }}/
-
name: Cleanup ginkgo JSON report
# Delete report.json after the analysis. File should always exist.
# Delete upgrade_report.json. It may not exist depending on test level.
if: always()
run: |
if [ -f tests/e2e/out/upgrade_report.json ]; then
rm tests/e2e/out/upgrade_report.json
fi
if [ -f tests/e2e/out/report.json ]; then
rm tests/e2e/out/report.json
fi
-
name: Archive e2e failure contexts
if: failure()
uses: actions/upload-artifact@v4
with:
name: test-failure-contexts-${{ matrix.id }}
path: |
tests/*/out/
retention-days: 7
if-no-files-found: ignore
-
name: Archive e2e logs
if: failure()
uses: actions/upload-artifact@v4
with:
name: cluster-logs-${{ matrix.id }}
path: |
tests/e2e/cluster_logs/**
retention-days: 7
if-no-files-found: ignore
-
name: Clean up
if: always()
run: |
set +e
# Attempt to remove any leftover resource
kubectl delete cluster --all --all-namespaces --now --timeout=30s || true
kubectl delete pdb --all --all-namespaces --now --timeout=30s || true
kubectl delete pvc --all --all-namespaces --now --timeout=30s || true
# Wait until all the PVs provisioned are actually reclaimed
kubectl wait --for delete --all pv --timeout=60s || true
attempt=1
max_attempts=3
while [ "${attempt}" -le "${max_attempts}" ]; do
gcloud container clusters delete ${{ env.CLUSTER_NAME }} --region=${{ env.REGION }} --quiet
status=$?
if [[ $status == 0 ]]; then
echo "GKS cluster ${{ env.CLUSTER_NAME }} deleted from region ${{ env.REGION }}"
break
fi
echo "Failed deleting cluster ${{ env.CLUSTER_NAME }} from region ${{ env.REGION }}, retrying"
sleep 5
attempt=$((attempt+1))
done
# The node's disks are not automatically deleted when the cluster is removed.
# We delete all the disks tagged with the name of the cluster that are not
# owned by anyone.
attempt=1
max_attempts=3
while [ "${attempt}" -le "${max_attempts}" ]; do
IDS=$(gcloud compute disks list --filter="labels.cluster=${{ env.CLUSTER_NAME }} AND region:${{ env.REGION }} AND -users:*" --format="value(id)")
amount="$(echo $IDS | awk '{print NF}')"
if [[ "$amount" == 3 ]]; then
echo -e "Found the 3 disks to be removed:\n$IDS"
break
fi
echo "Expected 3 disks to delete but found $amount, waiting and retrying"
sleep 20
attempt=$((attempt+1))
done
for ID in ${IDS}
do
attempt=1
max_attempts=3
while [ "${attempt}" -le "${max_attempts}" ]; do
gcloud compute disks delete --region "${{ env.REGION }}" --quiet "${ID}"
status=$?
if [[ $status == 0 ]]; then
echo "computer disk ${ID} deleted"
break
fi
echo "Failed deleting disk ${ID} from region ${{ env.REGION }}, retrying"
sleep 5
attempt=$((attempt+1))
done
done
# OpenShift Secrets required
# secrets.AWS_EKS_ADMIN_IAM_ROLES
# secrets.AWS_ACCESS_KEY_ID
# secrets.AWS_SECRET_ACCESS_KEY
e2e-openshift:
name: Run E2E on OpenShift
if: |
always() && !cancelled() &&
vars.OPENSHIFT_ENABLED == 'true' &&
needs.generate-jobs.outputs.openshiftEnabled == 'true' &&
needs.generate-jobs.result == 'success'
needs:
- buildx
- generate-jobs
- evaluate_options
strategy:
fail-fast: false
max-parallel: 6
matrix: ${{ fromJSON(needs.generate-jobs.outputs.openshiftMatrix) }}
runs-on: ubuntu-24.04
env:
# TEST_DEPTH determines the maximum test level the suite should be running
TEST_DEPTH: ${{ needs.evaluate_options.outputs.test_level }}
# FEATURE_TYPE, when defined, determines the subset of E2E tests that will be executed, divided by feature type
FEATURE_TYPE: ${{ needs.evaluate_options.outputs.feature_type }}
K8S_VERSION: "${{ matrix.k8s_version }}"
POSTGRES_VERSION: ${{ matrix.postgres_version }}
POSTGRES_KIND: ${{ matrix.postgres_kind }}
MATRIX: ${{ matrix.id }}
POSTGRES_IMG: "${{ matrix.postgres_img }}"
# The version of operator to upgrade FROM, in the rolling upgrade E2E test
E2E_PRE_ROLLING_UPDATE_IMG: "${{ matrix.postgres_pre_img }}"
TEST_TIMEOUTS: ${{ needs.generate-jobs.outputs.openshiftTimeout }}
BRANCH_NAME: ${{ needs.buildx.outputs.branch_name }}
DEBUG: "true"
BUILD_IMAGE: "false"
CONTROLLER_IMG: ${{ needs.generate-jobs.outputs.image }}
E2E_DEFAULT_STORAGE_CLASS: gp3-csi
E2E_CSI_STORAGE_CLASS: gp3-csi
E2E_DEFAULT_VOLUMESNAPSHOT_CLASS: csi-aws-vsc
TEST_CLOUD_VENDOR: "ocp"
# AWS configuration
AWS_BASE_DOMAIN: ${{ secrets.AWS_BASE_DOMAIN }}
AWS_ACCESS_KEY_ID: ${{ secrets.AWS_ACCESS_KEY_ID }}
AWS_SECRET_ACCESS_KEY: ${{ secrets.AWS_SECRET_ACCESS_KEY }}
AWS_REGION: eu-central-1
AWS_EKS_ADMIN_IAM_ROLES: ${{ secrets.AWS_EKS_ADMIN_IAM_ROLES }}
REDHAT_PULL: ${{ secrets.REDHAT_PULL }}
SSH_PUBLIC_KEY: ${{ secrets.SSH_PUBLIC_KEY }}
steps:
-
name: Set cluster name
run: |
echo "CLUSTER_NAME=${{ env.E2E_SUFFIX }}-ocp-${{ github.run_number}}-$( echo ${{ matrix.k8s_version }} | tr -d '.' )" >> $GITHUB_ENV
-
name: Checkout code
uses: actions/checkout@v4
with:
ref: ${{ needs.evaluate_options.outputs.git_ref }}
fetch-depth: 0
-
name: Install Go
uses: actions/setup-go@v5
with:
go-version: ${{ env.GOLANG_VERSION }}
check-latest: true
-
name: Set up QEMU
uses: docker/setup-qemu-action@v3
with:
platforms: ${{ env.PLATFORMS }}
-
name: Set up Docker Buildx
uses: docker/setup-buildx-action@v3
-
## In case hack/setup-cluster.sh need pull operand image from registry
name: Login into docker registry
uses: docker/login-action@v3
with:
registry: ${{ env.REGISTRY }}
username: ${{ env.REGISTRY_USER }}
password: ${{ env.REGISTRY_PASSWORD }}
-
name: Build and push the operator and catalog
env:
CONTROLLER_IMG: ${{ needs.buildx.outputs.controller_img_ubi8 }}
BUNDLE_IMG: ${{ needs.buildx.outputs.bundle_img }}
CATALOG_IMG: ${{ needs.buildx.outputs.catalog_img }}
run: |
make olm-catalog
-
name: Install OC Installer and client
uses: redhat-actions/openshift-tools-installer@v1
with:
source: "mirror"
openshift-install: ${{ matrix.k8s_version }}
oc: ${{ matrix.k8s_version }}
-
name: Install OpenShift Cluster ${{ matrix.k8s_version }}
run: |
envsubst < hack/install-config.yaml.template > hack/install-config.yaml
openshift-install create cluster --dir hack/ --log-level warn
-
name: Run E2E tests
if: (always() && !cancelled())
run: |
# Before running on OpenShift we make sure that the catalog is created
# in the openshift-marketplace namespace
sed -i -e 's/namespace: operators/namespace: openshift-marketplace/' cloudnative-pg-catalog.yaml
find -type f -name "cloudnative-pg-catalog.yaml"
cat cloudnative-pg-catalog.yaml
KUBECONFIG=$(pwd)/hack/auth/kubeconfig bash -x hack/e2e/run-e2e-ocp.sh
-
# Summarize the failed E2E tests cases if there are any
name: Report failed E2E tests
if: failure()
run: |
set +x
chmod +x .github/report-failed-test.sh
./.github/report-failed-test.sh
-
# Create an individual artifact for each E2E test, which will be used to
# generate E2E test summary in the follow-up job 'summarize-e2e-tests'
name: Create individual artifact for each E2E test
if: (always() && !cancelled())
env:
RUNNER: "openshift"
RUN_ID: ${{ github.run_id }}
REPOSITORY: ${{ github.repository }}
GIT_REF: ${{ needs.evaluate_options.outputs.git_ref }}
run: |
set +x
python .github/generate-test-artifacts.py \
-o testartifacts-${{ env.MATRIX }} \
-f tests/e2e/out/report.json \
--environment=true
if [ -f tests/e2e/out/upgrade_report.json ]; then
python .github/generate-test-artifacts.py \
-o testartifacts-${{ env.MATRIX }} \
-f tests/e2e/out/upgrade_report.json \
--environment=true
fi
-
name: Archive test artifacts
if: (always() && !cancelled())
uses: actions/upload-artifact@v4
with:
name: testartifacts-${{ env.MATRIX }}
path: testartifacts-${{ env.MATRIX }}/
retention-days: 7
-
name: Cleanup test artifacts
if: always()
run:
rm -rf testartifacts-${{ env.MATRIX }}/
-
name: Cleanup ginkgo JSON report
# Delete report.json after the analysis. File should always exist.
# Delete upgrade_report.json. It may not exist depending on test level.
if: always()
run: |
if [ -f tests/e2e/out/upgrade_report.json ]; then
rm tests/e2e/out/upgrade_report.json
fi
if [ -f tests/e2e/out/report.json ]; then
rm tests/e2e/out/report.json
fi
-
name: Archive e2e failure contexts
if: failure()
uses: actions/upload-artifact@v4
with:
name: test-failure-contexts-${{ matrix.id }}
path: |
tests/*/out/
retention-days: 7
if-no-files-found: ignore
-
name: Archive e2e logs
if: failure()
uses: actions/upload-artifact@v4
with:
name: cluster-logs-${{ matrix.id }}
path: |
tests/e2e/cluster_logs/**
retention-days: 7
if-no-files-found: ignore
-
name: Destroy OpenShift Cluster ${{ matrix.k8s_version }}
if: always()
run: |
openshift-install destroy cluster --dir hack/
# Summarize E2E test results, display in the GitHub 'summary' view
summarize-e2e-tests:
name: E2E test suite
needs:
- evaluate_options
- e2e-local
- e2e-eks
- e2e-aks
- e2e-gke
- e2e-openshift
if: |
(always() && !cancelled()) &&
((
needs.e2e-local.result == 'success' ||
needs.e2e-local.result == 'failure'
) ||
(
needs.e2e-eks.result == 'success' ||
needs.e2e-eks.result == 'failure'
) ||
(
needs.e2e-aks.result == 'success' ||
needs.e2e-aks.result == 'failure'
) ||
(
needs.e2e-gke.result == 'success' ||
needs.e2e-gke.result == 'failure'
) ||
(
needs.e2e-openshift.result == 'success' ||
needs.e2e-openshift.result == 'failure'
))
runs-on: ubuntu-24.04
steps:
- name: Create a directory for the artifacts
run: mkdir test-artifacts
- name: Download all artifacts to the directory
uses: actions/download-artifact@v4
with:
path: test-artifacts
pattern: testartifacts-*
- name: Flatten all artifacts onto directory
# The download-artifact action, since we did not give it a name,
# downloads all artifacts and creates a new folder for each.
# In this step we bring all the JSONs to a single folder
run: |
mkdir test-artifacts/data
mv test-artifacts/*/*.json test-artifacts/data || true
- name: Display the structure of the artifact folder
run: ls -R test-artifacts/data
- name: Compute the E2E test summary
id: generate-summary
uses: cloudnative-pg/[email protected]
with:
artifact_directory: test-artifacts/data
- name: If there is an overflow summary, archive it
if: steps.generate-summary.outputs.Overflow
uses: actions/upload-artifact@v4
with:
name: ${{ steps.generate-summary.outputs.Overflow }}
path: ${{ steps.generate-summary.outputs.Overflow }}
retention-days: 7
- name: Send the Ciclops view over Slack
# Send the Ciclops thermometer on every scheduled run on `main`.
# or when there are systematic failures in release branches
uses: rtCamp/action-slack-notify@v2
if: |
github.repository_owner == env.REPOSITORY_OWNER &&
(
github.event_name == 'schedule' ||
(
steps.generate-summary.outputs.alerts &&
startsWith(needs.evaluate_options.outputs.git_ref, 'refs/heads/release-')
)
)
env:
# SLACK_COLOR is where we distinguish a run with/without alerts. It's where the
# action has hooks for conditionality in the message body (yeah, weird)
SLACK_COLOR: ${{ steps.generate-summary.outputs.alerts && 'failure' || 'success' }}
SLACK_ICON: https://avatars.githubusercontent.com/u/85171364?size=48
SLACK_USERNAME: ${{ env.SLACK_USERNAME }}
SLACK_WEBHOOK: ${{ secrets.SLACK_WEBHOOK }}
SLACK_TITLE: CICLOPS view for ${{ github.repository }}
SLACK_MESSAGE_ON_SUCCESS: |
${{ steps.generate-summary.outputs.thermometer }}
SLACK_MESSAGE_ON_FAILURE: |
${{ steps.generate-summary.outputs.thermometer }}
:warning: *Systematic failures!*
${{ steps.generate-summary.outputs.alerts }}
SLACK_FOOTER: |
<${{ github.server_url }}/${{ github.repository }}/actions/runs/${{ github.run_id }}|*See full CI run*>
- name: Delete the downloaded files
run: rm -rf test-artifacts
# Adds the 'ok-to-merge' label to workflows that have run successfully and
# have adequate test and matrix coverage.
# This label is a prerequisite to be able to merge a PR.
# Also see to 'require-labels.yml'
ok-to-merge:
name: Label the PR as "ok to merge :ok_hand:"
needs:
- evaluate_options
- e2e-local
if: |
always() &&
needs.e2e-local.result == 'success' &&
github.event_name == 'issue_comment' &&
needs.evaluate_options.outputs.test_level == '4'
runs-on: ubuntu-24.04
steps:
- name: Check preconditions
id: get_pr_number_and_labels
env:
GITHUB_TOKEN: ${{ secrets.REPO_GHA_PAT }}
run: |
ok_label=$(gh pr view "${{ github.event.issue.number }}" --json labels -q ".labels.[].name" 2>/dev/null | grep "ok to merge :ok_hand:" || :)
echo "OK_LABEL=${ok_label}" >> $GITHUB_ENV
- name: Label the PR as "ok to merge :ok_hand:"
if: |
env.OK_LABEL == ''
uses: actions-ecosystem/[email protected]
with:
github_token: ${{ secrets.REPO_GHA_PAT }}
number: ${{ github.event.issue.number }}
labels: "ok to merge :ok_hand:"
# Remove the "ok to merge :ok_hand:" label if the E2E tests or previous steps failed
unlabel-ok-to-merge:
name: Remove the "ok to merge :ok_hand:" label from the PR
needs:
- evaluate_options
- e2e-local
if: |
always() &&
needs.e2e-local.result == 'failure' &&
github.event_name == 'issue_comment'
runs-on: ubuntu-24.04
steps:
- name: Check preconditions
id: get_pr_number_and_labels
env:
GITHUB_TOKEN: ${{ secrets.REPO_GHA_PAT }}
run: |
ok_label=$(gh pr view "${{ github.event.issue.number }}" --json labels -q ".labels.[].name" 2>/dev/null | grep "ok to merge :ok_hand:" || :)
echo "OK_LABEL=${ok_label}" >> $GITHUB_ENV
- name: Remove "ok to merge :ok_hand:" label from PR
if: |
env.OK_LABEL != ''
uses: actions-ecosystem/[email protected]
with:
github_token: ${{ secrets.REPO_GHA_PAT }}
number: ${{ github.event.issue.number }}
labels: "ok to merge :ok_hand:"