Troubleshooting

A local test run fails

Check these first:

  • LG_COORDINATOR points to the right shared hardware service.

  • labgrid-client places shows a board slot for the platform.

  • The test’s needs list matches the board slot’s capability tags.

  • The matching envs/<board-slot>.yaml file exists.

  • The hardware host is reachable over SSH.

A driver is missing

The board’s available drivers come from its environment file. If get_driver(...) fails, ask whoever maintains the fixture to add the driver; the mapping is explained in Hardware Model.

Artifacts will not download

If you see Neither 'workflow_run_url' in context or 'GITHUB_REPOSITORY' in environment, the run has no artifact source. Either pass workflow_run_url with a GITHUB_TOKEN, or place local files at the printed _artifacts/<test-name>/0/ path. See Run a test.

The exporter service will not start

On the hardware host:

  • Check status: sudo systemctl --user --machine=labgrid@.host status labgrid-exporter.service.

  • Confirm device permissions for the labgrid-client user:

    ls -l /dev/ttyUSB0
sudo -u labgrid-client test -r /dev/ttyUSB0 && echo client-can-read
sudo -u labgrid-client test -w /dev/ttyUSB0 && echo client-can-write

  • Confirm /var/cache/labgrid exists and is writable by the labgrid group.

A client cannot see resources from an exporter that did not start. Fix the service first. See Set up a hardware host.