QRHHomelab Quick Reference Handbook
AbnormalA5

YouTrack Hub calls fail with "Item of type AuthorityHolder … is not found"

When
Every Hub API call on issues.webathome.org fails with that error while YouTrack's own /api/* keeps working.
Ansible Edit on GitHub

A stale reference in Hub's permission cache, not in the database. On 2026-09-16 a <project> admins group was deleted through DELETE /hub/api/rest/usergroups/{id}; the project's Team group kept pointing at it and every Hub call from the token's account failed, even after recreating the group under its old id. API repairs do not help. A restart of the YouTrack pod fixed it completely.

YouTrack is one replica with strategy Recreate, so the restart is about a minute of downtime (43 s to YouTrack init complete on 2026-09-17). It needs ~/.kube/config-prd-write.

Steps #

  1. Confirm it is Hub, not YouTrack. With an admin token (YOUTRACK_TOKEN in a YouTrackMCPServer environment):

    curl -s -o /dev/null -w '%{http_code}\n' -H "Authorization: Bearer $YOUTRACK_TOKEN" \
  'https://issues.webathome.org/hub/api/rest/users?fields=login'
kubectl -n youtrack-prd logs deploy/youtrack --since=10m | grep -E 'ERROR|AuthorityHolder'

  2. Restart, and wait for the new pod:

    kubectl -n youtrack-prd rollout restart deploy/youtrack
kubectl -n youtrack-prd rollout status deploy/youtrack

  3. YouTrack init completein the pod log

  4. The curl of step 1200, and no new AuthorityHolder line

Don't

Never delete Hub users or groups on this instance directly. The import creates <project> admins groups and reuses them: leave them. Hub Integration's quick fixes are untested production writes; the runbook below has them, behind a database backup.

Runbooks #

In YouTrackMCPServerSpecs (private):

End of procedure