How to recover config

Recover with a saved active revision

When disk config is invalid, the daemon tries to load the last active revision.

If recovery succeeds:

• Controllers start with the last active config.
• The Console reports Disk config invalid.
• The invalid disk file does not replace the active config.

To restore a revision:

1. Open Config.
2. Find a known good item under Recent revisions.
3. Select Restore to disk.
4. Validate the restored disk config.
5. Open Jobs and wait for running jobs to finish.
6. Restart the service manually.
7. Confirm the state is In sync.

Restore writes the selected revision to disk. It does not change running controllers and does not restart the service.

The daemon keeps the newest ten config revisions.

Recover when no active revision exists

If disk config is invalid and no active revision exists, the daemon starts in config mode.

In config mode:

• Runner controllers do not start.
• The Console uses the default local address.
• Admin setup and sign in remain available.
• The Config page remains available.

To repair the config:

1. Open http://127.0.0.1:8080.
2. Set up or sign in to the Console.
3. Open Config.
4. Replace the invalid YAML with a complete config.
5. Select Validate.
6. Fix every error.
7. Save the config.
8. Restart the service manually.

After restart, confirm that Runner Sets contains the expected sets and Overview shows GitHub connection state.

Recover without Console access

Edit the configured YAML file with a local text editor.

Run the daemon in the foreground to see validation output:

elastic-fruit-runner --config /path/to/config.yaml

The error includes the config path and reason. Fix the file, then start the normal service again.

Use Configuration Reference to check fields and limits.