작업 전후에 스크립트 실행

About pre- and post-job scripts

You can automatically execute scripts on a self-hosted runner, either before a job runs, or after a job finishes running. You could use these scripts to support the job's requirements, such as building or tearing down a runner environment, or cleaning out directories. You could also use these scripts to track telemetry of how your runners are used.

The custom scripts are automatically triggered when a specific environment variable is set on the runner; the environment variable must contain the absolute path to the script. For more information, see "Triggering the scripts" below.

The following scripting languages are supported:

• Bash: Uses bash and can fallback to sh. Executes by running -e {pathtofile}.
• PowerShell: Uses pwsh and can fallback to powershell. Executes by running -command \". '{pathtofile}'\".

Writing the scripts

Your custom scripts can use the following features:

• Variables: Scripts have access to the default variables. The full webhook event payload can be found in GITHUB_EVENT_PATH. For more information, see "Variables."
• Workflow commands: Scripts can use workflow commands. For more information, see "Workflow commands for GitHub Actions", with the exception of save-state and set-output, which are not supported by these scripts. Scripts can also use environment files. For more information, see Environment files.

Your script files must use a file extension for the relevant language, such as .sh or .ps1, in order to run successfully.

Handling exit codes

For pre-job scripts, exit code 0 indicates that the script completed successfully, and the job will then proceed to run. If there is any other exit code, the job will not run and will be marked as failed. To see the results of your pre-job scripts, check the logs for Set up runner entries. For more information on checking the logs, see "Using workflow run logs."

The continue-on-error setting is not supported for use by these scripts.

Triggering the scripts

The custom scripts must be located on the runner, but should not be stored in the actions-runner application directory. The scripts are executed in the security context of the service account that's running the runner service.

The scripts are automatically executed when the runner has the following environment variables containing an absolute path to the script:

• ACTIONS_RUNNER_HOOK_JOB_STARTED: The script defined in this environment variable is triggered when a job has been assigned to a runner, but before the job starts running.
• ACTIONS_RUNNER_HOOK_JOB_COMPLETED: The script defined in this environment variable is triggered after the job has finished processing.

To set these environment variables, you can either add them to the operating system, or add them to a file named .env within the self-hosted runner application directory. For example, the following .env entry will have the runner automatically run a script named cleanup_script.sh before each job runs:

ACTIONS_RUNNER_HOOK_JOB_STARTED=/cleanup_script.sh

Troubleshooting

No timeout setting

There is currently no timeout setting available for scripts executed by ACTIONS_RUNNER_HOOK_JOB_STARTED or ACTIONS_RUNNER_HOOK_JOB_COMPLETED. As a result, you could consider adding timeout handling to your script.

Reviewing the workflow run log

To confirm whether your scripts are executing, you can review the logs for that job. The scripts will be listed within separate steps for either Set up runner or Complete runner, depending on which environment variable is triggering the script. For more information on checking the logs, see "Using workflow run logs."