Packaging and Testing with Crossbow

The content of arrow/dev/tasks directory aims for automating the process of Arrow packaging and integration testing.

Packages:

• C++ and Python conda-forge packages for Linux, macOS and Windows

• Python Wheels for Linux, macOS and Windows

• C++ and GLib Linux packages for multiple distributions

• Java for Gandiva

Integration tests:

• Various docker tests

• Pandas

• Dask

• Turbodbc

• HDFS

• Spark

Architecture

Executors

Individual jobs are executed on public CI services, currently:

• Linux: GitHub Actions, Travis CI, Azure Pipelines

• macOS: GitHub Actions, Azure Pipelines

• Windows: GitHub Actions, Azure Pipelines

Queue

Because of the nature of how the CI services work, the scheduling of jobs happens through an additional git repository, which acts like a job queue for the tasks. Anyone can host a queue repository (usually named <ghuser>/crossbow).

A job is a git commit on a particular git branch, containing the required configuration files to run the requested builds (like .travis.yml, azure-pipelines.yml, or crossbow.yml for GitHub Actions ).

Scheduler

Crossbow handles version generation, task rendering and submission. The tasks are defined in tasks.yml.

Install

The following guide depends on GitHub, but theoretically any git server can be used.

If you are not using the ursacomputing/crossbow repository, you will need to complete the first two steps, otherwise proceed to step 3:

1. Create the queue repository

2. Enable Travis CI and Azure Pipelines integrations for the newly created queue repository.

3. Clone either ursacomputing/crossbow if you are using that, or the newly created repository next to the arrow repository:

   By default the scripts looks for a crossbow clone next to the arrow directory, but this can configured through command line arguments.

   git clone https://github.com/<user>/crossbow crossbow
   

   Important note: Crossbow only supports GitHub token based authentication. Although it overwrites the repository urls provided with ssh protocol, it’s advisable to use the HTTPS repository URLs.

4. Create a Personal Access Token with repo and workflow permissions (other permissions are not needed)

5. Locally export the token as an environment variable:

   export CROSSBOW_GITHUB_TOKEN=<token>
   

   or pass as an argument to the CLI script --github-token

6. Add the previously created GitHub token to Travis CI:

   Use CROSSBOW_GITHUB_TOKEN encrypted environment variable. You can set it at the following URL, where ghuser is the GitHub username and ghrepo is the GitHub repository name (typically crossbow):

   https://travis-ci.com/<ghuser>/<ghrepo>/settings

   • Confirm the auto cancellation feature is turned off for branch builds. This should be the default setting.

7. Install Python (minimum supported version is 3.9):

   Miniconda is preferred, see installation instructions:

   https://conda.io/docs/user-guide/install/index.html

8. Install the archery toolset containing crossbow itself:

   $ pip install -e "arrow/dev/archery[crossbow]"
   

9. Try running it:

   $ archery crossbow --help
   

Usage

The script does the following:

1. Detects the current repository, thus supports forks. The following snippet will build kszucs’s fork instead of the upstream apache/arrow repository.

   $ git clone https://github.com/kszucs/arrow
   $ git clone https://github.com/kszucs/crossbow
   
   $ cd arrow/dev/tasks
   $ archery crossbow submit --help  # show the available options
   $ archery crossbow submit conda-win conda-linux conda-osx
   

2. Gets the HEAD commit of the currently checked out branch and generates the version number based on setuptools_scm. So to build a particular branch check out before running the script:

   $ git checkout ARROW-<ticket number>
   $ archery crossbow submit --dry-run conda-linux conda-osx
   

   Note that the arrow branch must be pushed beforehand, because the script will clone the selected branch.

3. Reads and renders the required build configurations with the parameters substituted.

4. Create a branch per task, prefixed with the job id. For example, to build conda recipes on linux, it will create a new branch: crossbow@build-<id>-conda-linux.

5. Pushes the modified branches to GitHub which triggers the builds. For authentication it uses GitHub OAuth tokens described in the install section.

Query the build status

Build id (which has a corresponding branch in the queue repository) is returned by the submit command.

$ archery crossbow status <build id / branch name>

Download the build artifacts

$ archery crossbow artifacts <build id / branch name>

Examples

Submit command accepts a list of task names and/or a list of task-group names to select which tasks to build.

Run multiple builds:

$ archery crossbow submit debian-stretch conda-linux-gcc-py37-r40
Repository: https://github.com/kszucs/arrow@tasks
Commit SHA: 810a718836bb3a8cefc053055600bdcc440e6702
Version: 0.9.1.dev48+g810a7188.d20180414
Pushed branches:
 - debian-stretch
 - conda-linux-gcc-py37-r40

Just render without applying or committing the changes:

$ archery crossbow submit --dry-run task_name

Run only conda package builds and a Linux one:

$ archery crossbow submit --group conda centos-7

Run wheel builds:

$ archery crossbow submit --group wheel

There are multiple task groups in the tasks.yml like docker, integration and cpp-python for running docker based tests.

archery crossbow submit supports multiple options and arguments, for more see its help page:

$ archery crossbow submit --help