Affected services:

  • Crane

Crane /work partial outage, jobs held

Opened on Thursday 18th June 2020, last updated

Resolved

A reboot of the Lustre metadata server for the /work filesystem has resolved the issue. We've made changes to Lustre locking parameters to avoid the issue in the future.

Jobs on crane are no longer held, so will be starting normally.

Posted by John Thiltges

Investigating

We are investigating slow performance on Crane's /work filesystem. Pending jobs will be held in queue (not started) while we work on the issue.

Further updates will be added to the status website entry as they're available.

Posted by John Thiltges