Skip to content
Permalink
Branch: master
Find file Copy path
Find file Copy path
8 contributors

Users who have contributed to this file

@fweikert @buchgr @philwo @meteorcloudy @joeleba @alexeagle @dslomov @brandjon
450 lines (340 sloc) 17 KB

Bazel Continuous Integration

tl;dr:

Bazel uses Buildkite for continuous integration. The user interface and the orchestration of CI builds is fully managed by Buildkite, but Bazel brings its own CI machines. The buildkite folder contains all the scripts and configuration files necessary to setup Bazel's CI on Buildkite.

Bazel on Buildkite 101

When you first log into Buildkite you are presented with a list of pipelines. A pipeline is a template of steps that are executed either in sequence or in parallel and that all need to succeed in order for the pipeline to succeed. The Bazel organisation has dozens of pipelines. Here are a selected few:

pipelines

  • The bazel postsubmit pipeline builds and tests each commit to Bazel's repository on all supported platforms.
  • The bazel presubmit pipeline is triggered on every pull request to Bazel.
  • The rules_go postsubmit pipeline is triggered on every commit to the rules_go repository.
  • The TensorFlow pipeline builds and tests TensorFlow at HEAD every four hours.

Builds

When you click on a pipeline you can see the last few builds of this pipeline. Clicking on a build then gives you access to the details of the build. For example, the below image shows a failed build step on Ubuntu 16.04.

failed build step

One can see which tests failed by clicking on the Test section. In the below example, the //src/test/shell/bazel:external_path_test was flaky as it failed in 1 out of 5 runs.

flaky test

You can view the failed test attempt's test.log file in the Artifacts tab.

flaky test log

Useful Links

buildkite useful buttons

Pull Requests

Bazel accepts contributions via pull requests. Contributions by members of the bazelbuild organisation as well as members of individual repositories (i.e. rule maintainers) are whitelisted automatically and will immediately be built and tested on Buildkite.

An external contribution, however, first needs to be verified by a project member and therefore will display a pending status named Verify Pull Request.

status verify pull request

A member can verify a pull request by clicking on Details, followed by Verify Pull Request.

buildkite verify pull request

Please vet external contributions carefully as they can execute arbitrary code on our CI machines

Build and Test Results

After a pull request has been built and tested, the results will be displayed as a status message on the pull request. A detailed view is available when clicking on the corresponding Details link. Click here for an example.

pull request details

Presubmit for downstream projects

You can preview the effect of an unmerged commit on downstream projects. See Testing Local Changes With All Downstream Projects.

Checking incompatible changes status for downstream projects

Bazelisk + Incompatible flags pipeline runs bazelisk --migrate on all downstream projects and reports a summary of all incompatible flags and migrations statuses of downstream projects.

The pipeline runs every night against the latest Bazel release. You can also schedule manual builds and set USE_BAZEL_VERSION to run against a specific release or a release candidate (e.g. USE_BAZEL_VERSION=0.29rc3 to test against RC3 of release 0.29)

Culprit Finder

Bazel downstream projects is red? Use culprit finder to find out which bazel commit broke it!

First you should check if the project is green with the latest Bazel release. If not, probably it's their commits that broke the CI.

If a project is green with release Bazel but red with Bazel nightly, it means some Bazel commit broke it, then culprit finder can help!

Create "New Build" in the Culprit Finder project with the following environment variable:

  • PROJECT_NAME (The project name must exist in DOWNSTREAM_PROJECTS in bazelci.py)
  • (Optional) TASK_NAME (The task name must exist in the project's config file, eg. macos_latest). For old config syntax where platform name is essentially the task name, you can also set PLATFORM_NAME instead of TASK_NAME. If not set, culprit finder will bisect for all tasks of the specified project.
  • (Optional) GOOD_BAZEL_COMMIT (A full Bazel commit, Bazel built at this commit still works for this project). If not set, culprit finder will use the last green bazel commit in downstream pipeline as the good bazel commit.
  • (Optional) BAD_BAZEL_COMMIT (A full Bazel commit, Bazel built at this commit fails with this project). If not set, culprit finder will use the lastest Bazel commit as the bad bazel commit.
  • (Optional) NEEDS_CLEAN (Set NEEDS_CLEAN to true to run bazel clean --expunge before each build, this will help reduce flakiness)
  • (Optional) REPEAT_TIMES (Set REPEAT_TIMES to run the build multiple times to detect flaky build failure, if at least one build fails we consider the commit as bad)

eg.

PROJECT_NAME=rules_go
PLATFORM_NAME=ubuntu1604
GOOD_BAZEL_COMMIT=b6ea3b6caa7f379778e74da33d1bd0ff6477f963
BAD_BAZEL_COMMIT=91eb3d207714af0ab1e5812252a0f10f40d6e4a8

Note: Bazel commit can only be set to commits after 63453bdbc6b05bd201375ee9e25b35010ae88aab, Culprit Finder needs to download Bazel at specific commit, but we didn't prebuild Bazel binaries before this commit.

Configuring a Pipeline

Each pipeline is configured via a Yaml file. This file either lives in $PROJECT_DIR/.bazelci/presubmit.yml (for presubmits) or in an arbitrary location whose path or URL is passed to the CI script (as configured in the Buildkite settings of the respective pipeline). Projects should store the postsubmit configuration in their own repository, but we keep some configurations for downstream projects in https://github.com/bazelbuild/continuous-integration/tree/master/buildkite/pipelines.

Basic Syntax

The most important piece of the configuration file is the tasks dictionary. Each task has a unique key, a platform and usually some build and/or test targets:

---
tasks:
  ubuntu_build_only:
    platform: ubuntu1604
    build_targets:
    - "..."
  windows:
    platform: windows
    build_targets:
    - "..."
    test_targets:
    - "..."

If there is exactly one task per platform, you can omit the platform field and use its value as task ID instead. The following code snippet is equivalent to the previous one:

---
tasks:
  ubuntu1604:
    build_targets:
    - "..."
  windows:
    build_targets:
    - "..."
    test_targets:
    - "..."

Setting Environment Variables

You can set environment variables for each individual task via the environment field:

---
tasks:
  ubuntu1804:
    environment:
      CC: clang
    build_targets:
    - "..."

Running Commands, Shell Scripts or Binary Targets

The presubmit configuration allows you to specify a list of shell commands that are executed at the beginning of every job. Simply add the batch_commands (Windows) or shell_commands field (all other platforms).

You can even run executable targets via the run_targets field. The following example demonstrates all of these features:

---
tasks:
  ubuntu1804:
    shell_commands:
    - rm -f obsolete_file
    run_targets:
    - "//whatever"
    build_targets:
    - "..."
  windows:
    batch_commands:
    - powershell -Command "..."
    build_targets:
    - "..."

Using Specific Build & Test Flags

The build_flags and test_flags fields contain lists of flags that should be used when building or testing (respectively):

---
tasks:
  ubuntu1804:
    build_flags:
    - "--define=ij_product=clion-latest"
    build_targets:
    - "..."
    test_flags:
    - "--define=ij_product=clion-latest"
    test_targets:
    - ":clwb_tests"

Specifying a Display Name

Each task may have an optional display name that can include Emojis. This feature is especially useful if you have several tasks that run on the same platform, but use different Bazel binaries. Simply set the name field:

---
tasks:
  windows:
    name: "some :emoji:"
    build_targets:
    - "..."

Legacy Format

Most existing configuration use the legacy format with a "platforms" dictionary:

---
platforms:
  ubuntu1804:
    build_targets:
    - "..."
    test_targets:
    - "..."

The new format expects a "tasks" dictionary instead:

---
tasks:
  arbitrary_id:
    platform: ubuntu1804
    build_targets:
    - "..."
    test_targets:
    - "..."

In this case we can omit the platform field since there is a 1:1 mapping between tasks and platforms. Consequently, the format looks almost identical to the old one:

---
tasks:
  ubuntu1804:
    build_targets:
    - "..."
    test_targets:
    - "..."

The CI script still supports the legacy format, too.

Using a specific version of Bazel

The CI uses Bazelisk to support older versions of Bazel, too. You can specify a Bazel version for each pipeline (or even for individual platforms) in the pipeline Yaml configuration:

---
bazel: 0.20.0
tasks:
  windows:
    build_targets:
    - "..."
  macos:
    build_targets:
    - "..."
  ubuntu1804:
    bazel: 0.18.0
    build_targets:
    - "..."
[...]

In this example the jobs on Windows and MacOS would use 0.20.0, whereas the job on Ubuntu would run 0.18.0.

CI supports several magic version values such as latest, last_green and last_downstream_green. Please see the Bazelisk documentation for more details.

Testing with incompatible flags

Similar to the aforementioned downstream pipeline you can configure individual pipelines to run with bazelisk --migrate. As a result, the pipeline runs your targets with all incompatible flags that will be flipped in the next major Bazel release and prints detailed information about which flags need to be migrated.

You can enable this feature by adding the following code to the top of the pipeline steps in Buildkite at https://buildkite.com/bazel/YOUR_PIPELINE_SLUG/settings, not in the pipeline configuration yaml file:

---
env:
  USE_BAZELISK_MIGRATE: true

If you want your pipeline to fail if at least one flag needs migration, you need to add this code instead:

---
env:
  USE_BAZELISK_MIGRATE: FAIL

If you want to enable this feature for a single build, but not for the entire pipeline, you can follow these steps instead:

  1. Navigate to your pipeline in Buildkite.
  2. Click on the "New Build" button in the top right corner.
  3. Expand the pipeline options via a click on "Options".
  4. Enter USE_BAZELISK_MIGRATE=FAIL into the "Environment Variables" text field.
  5. Click on "Create Build".

macOS: Using a specific version of Xcode

We upgrade the CI machines to the latest version of Xcode shortly after it is released and this version will then be used as the default Xcode version. If required, you can specify a fixed Xcode version to test against in your pipeline config.

Warning: We might have to run jobs that specify an explicit Xcode version on separate, slower machines, so we really advise you to not use this feature unless necessary.

The general policy is to not specify a fixed Xcode version number, so that we can update the default version more easily and don't have to update every single CI configuration file out there.

However, if you know that you need to test against multiple versions of Xcode or that newer versions frequently break you, you can use this feature.

tasks:
  # Test against the latest released Xcode version.
  macos:
    build_targets:
    - "..."
  # Ensure that we're still supporting Xcode 10.1.
  macos_xcode_10_1:
    platform: macos
    xcode_version: "10.1"
    build_targets:
    - "..."

Take care to quote the version number, otherwise YAML will interpret it as a floating point number.

Running Buildifier on CI

For each pipeline you can enable Buildifier to check all WORKSPACE, BUILD, BUILD.bazel and .bzl files for lint warnings and formatting violations. Simply add the following code to the top of the particular pipeline configuration:

---
buildifier: latest
[...]

As a consequence, every future build for this pipeline will contain an additional "Buildifier" step that runs the latest version of Buildifier both in "lint" and "check" mode. Alternatively you can specify a particular Buildifier version such as "0.20.0".

There is also a more advanced syntax that allows you to specify which warnings should be checked in lint mode:

---
buildifier:
  version: latest
  warnings: "positional-args,duplicated-name"
[...]

Using multiple Workspaces in a single Pipeline

Some projects may contain one or more WORKSPACE files in subdirectories, in addition to their top-level WORKSPACE file. All of these workspaces can be tested in a single pipeline by using the working_directory task property. Consider the configuration for a project that contains a second WORKSPACE file in the examples_dir/ directory:

---
tasks:
  production_code:
    name: "My Project"
    platform: ubuntu1804
    test_targets:
    - //...
  examples:
    name: Examples
    platform: ubuntu1804
    working_directory: examples_dir
    test_targets:
    - //...

Validating changes to pipeline configuration files

You can set the top-level validate_config option to ensure that changes to pipeline configuration files in the .bazelci directory will be validated. With this option, every build for a commit that touches a configuration file will contain an additional validation step for each modified configuration file.

Example usage:

---
validate_config: 1
tasks:
  macos:
    build_targets:
    - "..."

Exporting JSON profiles of builds and tests

Bazel's JSON Profile is a useful tool to investigate the performance of Bazel. You can configure your pipeline to export these JSON profiles on builds and tests using the include_json_profile option.

Example usage:

---
tasks:
  ubuntu1604:
    include_json_profile:
    - build
    - test
    build_targets:
    - "..."
    test_targets:
    - "..."

When include_json_profile is specified with build, the builds will be carried out with the extra JSON profile flags. Similarly for test. Other values will be ignored.

The exported JSON profiles are available as artifacts after each run.

You can’t perform that action at this time.