1610610 : Jupyterhub login issues

Created: 2026-06-19T10:57:52Z - current status: new

Anonymized Summary:

A user reports frequent automatic disconnections from JupyterHub. Upon reloading, an error message appears (referenced in an attachment).

Issue Core:

Recurring disconnections and potential deadlocks or misconfigurations in the notebook/spawner system.

Possible Causes & Solutions:

  1. Deadlocked Notebook Entry
  2. If the notebook state is stuck in the Hub’s database, it may prevent normal shutdown/restart.

  3. Action: Try manually stopping/starting the notebook via the Hub GUI. If unresolved, contact admins to clear the stale entry (patch since Feb 2026 allows self-resolution). (Ref: Known issue with race conditions or route deletion failures.)

  4. Unresponsive CVMFS Mounts

  5. If the underlying worker node has unresponsive mounts, notebooks may fail to report their status, causing timeouts.

  6. Action: Retry spawning; healthchecks should now skip faulty nodes. Persistent issues require admin intervention.

  7. Quota Exceeded

  8. Check $HOME for disk usage (ls -lh ~/.local/share/jupyter/runtime/). Clean up files or request quota increase.

  9. Error clue: "OSError(122, 'Disk quota exceeded')".

  10. Misconfigured Local Settings

  11. Verify no open_browser = False exists in ~/.jupyter/jupyter_server_config.json. Comment it out if present.

Next Steps:

  • Inspect .jupyterhub.condor.err in $HOME for specific errors.
  • Test with a fresh browser session or incognito mode.
  • Report recurring node-specific issues to support.

(Sources: NAF documentation on notebook spawning, deadlocks, CVMFS mounts.)