Giter VIP home page Giter VIP logo

github-runner-docker's Introduction

Ephemeral self-hosted Github Runners in Docker

Objectives:

  • Avoid limitations of Github's hosted runner hardware
    • More CPU cores
    • More disk space
    • GPUs
  • Run for more than Github's free tier (2000 minutes/month as of this writing)
  • Fresh environment for each run
  • Isolated environment from the system hosting the runner

This was based off of a testdriven.io blog post.

Design

The foundation is a docker image that contains a specific version of the Github Runner. Docker is nice for this case because it simplifies accomplishing isolation and a fresh environment for each run.

The dockerfile is the foundation of the runner. RUNNER_VERSION describes which version of the runner to embed in the image. Find the latest version at the Github runner releases page curl and ca-certificates are needed in the image to download the runner. jq is needed to associate a runner with an organization. Then the runner binary is retireved, its dependencies are installed.

The start.sh script is the entry point for the docker image. It needs to be modified for the particular repository you want to connect the runner to. In your repository:

  • Go to settings > actions > runners > add runner
    • modify start.sh lines that have config.sh with the provided snippet

The start.sh script connects the runner to the Github repository. Then, it registers a cleanup function, which unregisters the runner with Github. This cleanup is called at the end of the script, and also if sigint or sigterm is sent to the container (e.g. from docker stop). This prevents a bunch of old runners from cluttering up your list of runners.

There is active work on the ephemeral runner tracked here: actions/runner#660

Getting Started

  • Modify dockerfile to install any packages you want to include for your action. (You can also just specify the installation command in your action).
  • Create a personal access token that the runner will use to request a registration token (user settings > developer settings > personal access tokens)
    • Optionally, you can use a registration token alone (repo settings > actions > runners > add runner), but those tokens are short-lived and after a bit the runner won't be able to connect without an updated token

The build the runner

docker build -t test-runner-2.280.3 .

Start the runner and check the logs

Your repository URL is something like https://github.com/cwpearson/github-runner-docker (no trailing slash!)

docker run \
--name runner-name
-e GITHUB_ACCESS_TOKEN=<your personal access token> \
-e RUNNER_REPOSITORY_URL=<your repository URL> \
--detach --restart unless-stopped \
test-runner-2.280.3

docker logs -f runner-name

You should see something like this:

--------------------------------------------------------------------------------
|        ____ _ _   _   _       _          _        _   _                      |
|       / ___(_) |_| | | |_   _| |__      / \   ___| |_(_) ___  _ __  ___      |
|      | |  _| | __| |_| | | | | '_ \    / _ \ / __| __| |/ _ \| '_ \/ __|     |
|      | |_| | | |_|  _  | |_| | |_) |  / ___ \ (__| |_| | (_) | | | \__ \     |
|       \____|_|\__|_| |_|\__,_|_.__/  /_/   \_\___|\__|_|\___/|_| |_|___/     |
|                                                                              |
|                       Self-hosted runner registration                        |
|                                                                              |
--------------------------------------------------------------------------------

# Authentication


√ Connected to GitHub

# Runner Registration

Enter the name of the runner group to add this runner to: [press Enter for Default] 
Enter the name of runner: [press Enter for 899dec636070] 
This runner will have the following labels: 'self-hosted', 'Linux', 'X64' 
Enter any additional labels (ex. label-1,label-2): [press Enter to skip] 
√ Runner successfully added
√ Runner connection is good

# Runner settings

Enter name of work folder: [press Enter for _work] 
√ Settings Saved.


√ Connected to GitHub

2021-08-23 12:44:56Z: Listening for Jobs

You can go into your repository "actions" tab and manually start an action. The runner should pick it up and execute it. Then you will notice that the container has stopped. Check setttings > actions > runners to make sure the runner unregistered itself.

Now, when your runner completes an action, it will unregister itself and exit. Then Docker will spin up a new container with a fresh runner to grab the next action.

Limitations

Runnner token expires

It may be possible to get an access token and then ask Github for a runner token dynamically

one runner per repository

The current configuration only supports one runner per repository. It is possible to associate a runner with an enterprise or organization instead. To do an organization, start.sh will need to take the organization access token and ask github to convert it to a registration token. That code is currently commented out.

runner runs as root

To allow the docker image to install packages, the runner runs as root.

I think this is bad if you're having your runner run PRs. Basically, it allows anyone to run arbitrary code in your container as root. This might be a problem especially if Docker is configured in certain ways or has a vulnerability.

--once flag is not official

The --once flag is not documented, and seems to mostly work. However, sometimes it seems to run more than one job. One day, Github may provide an official way for a runner to run a single job, then terminate.

References:

Ideas:

github-runner-docker's People

Contributors

cwpearson avatar

Watchers

James Cloos avatar

Recommend Projects

  • React photo React

    A declarative, efficient, and flexible JavaScript library for building user interfaces.

  • Vue.js photo Vue.js

    🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.

  • Typescript photo Typescript

    TypeScript is a superset of JavaScript that compiles to clean JavaScript output.

  • TensorFlow photo TensorFlow

    An Open Source Machine Learning Framework for Everyone

  • Django photo Django

    The Web framework for perfectionists with deadlines.

  • D3 photo D3

    Bring data to life with SVG, Canvas and HTML. 📊📈🎉

Recommend Topics

  • javascript

    JavaScript (JS) is a lightweight interpreted programming language with first-class functions.

  • web

    Some thing interesting about web. New door for the world.

  • server

    A server is a program made to process requests and deliver data to clients.

  • Machine learning

    Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.

  • Game

    Some thing interesting about game, make everyone happy.

Recommend Org

  • Facebook photo Facebook

    We are working to build community through open source technology. NB: members must have two-factor auth.

  • Microsoft photo Microsoft

    Open source projects and samples from Microsoft.

  • Google photo Google

    Google ❤️ Open Source for everyone.

  • D3 photo D3

    Data-Driven Documents codes.