Continuous Integration is important component of making Apache Beam robust and stable.
Our execution environment for CI is mainly the Jenkins which is available at https://ci-beam.apache.org/. See .test-infra/jenkins/README for trigger phrase, status and link of all Jenkins jobs. See Apache Beam Developer Guide for Jenkins Tips.
An additional execution environment for CI is GitHub Actions. GitHub Actions (GA) are very well integrated with GitHub code and Workflow and it has evolved fast in 2019/2020 to become a fully-fledged CI environment, easy to use and develop for, so we decided to use it for building python source distribution and wheels.
The following GA CI Job runs are currently run for Apache Beam, and each of the runs have different purpose and context.
Those runs are results of PR from the forks made by contributors. Most builds for Apache Beam fall into this category. They are executed in the context of the "Fork", not main Beam Code Repository which means that they have only "read" permission to all the GitHub resources (container registry, code repository). This is necessary as the code in those PRs (including CI job definition) might be modified by people who are not committers for the Apache Beam Code Repository.
The main purpose of those jobs is to check if PR builds cleanly, if the test run properly and if the PR is ready to review and merge.
Those runs are results of direct pushes done by the committers or as result of merge of a Pull Request by the committers. Those runs execute in the context of the Apache Beam Code Repository and have also write permission for GitHub resources (container registry, code repository). The main purpose for the run is to check if the code after merge still holds all the assertions - like whether it still builds, all tests are green.
This is needed because some of the conflicting changes from multiple PRs might cause build and test failures after merge even if they do not fail in isolation.
Those runs are results of (nightly) triggered job - only for master branch. The
main purpose of the job is to check if there was no impact of external dependency changes on the Apache
Beam code (for example transitive dependencies released that fail the build). Another reason for the nightly
build is that the builds tags most recent master with nightly-master.
All runs consist of the same jobs, but the jobs behave slightly differently or they are skipped in different run categories. Here is a summary of the run categories with regards of the jobs they are running. Those jobs often have matrix run strategy which runs several different variations of the jobs (with different platform type / Python version to run for example)
| Job | Description | Pull Request Run | Direct Push/Merge Run | Scheduled Run | Requires GCP Credentials |
|---|---|---|---|---|---|
| Build python source distribution | Builds python source distribution and uploads it to artifacts. Artifacts from release branch are used in release process (build_release_candidate.sh) |
Yes | Yes | Yes | - |
| Prepare GCS | Clears target path on GCS if already exists. | - | Yes | Yes | Yes |
| Upload python source distribution to GCS bucket | Uploads python source distribution to GCS bucket for path unique for specific workflow run. | - | Yes | Yes | Yes |
| Build python wheels on linux/macos/windows | Builds python wheels on linux/macos/windows platform with usage of cibuildwheel and uploads it to artifacts. Artifacts from release branch are used in release process ( build_release_candidate.sh ) |
Yes | Yes | Yes | - |
| Upload python wheels to GCS bucket | Uploads python wheels to GCS bucket for path unique for specific workflow run. Additionally uploads workflow run data. | - | Yes | Yes | Yes |
| List files on Google Cloud Storage Bucket | Lists files on GCS for verification purpose. | - | Yes | Yes | Yes |
| Tag repo nightly | Tag repo with nightly-master tag if build python source distribution and python wheels finished successfully. |
- | - | Yes | - |
Some of the jobs require variables stored as a GitHub Secrets to perform operations on Google Cloud Platform. Currently these jobs are limited to Apache repository only. These variables are:
GCP_SA_EMAIL- Service account email address. This is usually of the format<name>@<project-id>.iam.gserviceaccount.com.GCP_SA_KEY- Service account key. This key should be created and encoded as a Base64 string (eg.cat my-key.json | base64on macOS).
Service Account shall have following permissions:
- Storage Object Admin (roles/storage.objectAdmin)
- If you introduce changes to the workflow it is possible that your changes will not be present in the check run triggered in Pull Request. In this case please attach link to the modified workflow run executed on your fork.
- Possible timeouts with macOS runner - existing issue: (X) This check failed - sometimes happens on macOS runner #841
- GitHub Actions Documentation