Some OAK targets are unavailable
This issue has been identified: it turned out to be a monitoring problem, and not a real file system issue on Oak.
Compute nodes are currently re-opening to new jobs, and everything should be operating normally again very shortly.
Some storage targets are currently unavailable on $OAK, and to prevent potential job failures, all of the compute nodes are currently draining and won’t accept new jobs at this time.
We’re investigating the cause of the issue.
3 Affected Services:
- Job scheduler / Public partitions
- Job scheduler / Owner partitions
- Filesystems / $OAK