Purdue University

Link to update at March 8, 2023 4:08pm EST UPDATE: March 8, 2023 4:08pm EST

As of 4:08pm EST, engineers have completed all maintenance tasks and returned the Negishi cluster back to service. Jobs scheduling has been resumed.

Important changes

• Dedicated queues are enabled and configured for all PIs and users;
• The standby queue is back to its usual walltime limit of 4 hours, and will reject new submissions requesting longer walltimes (use your dedicated queue instead);

Please see slist output for details on available queues.

How this may affect you?

If you had jobs waiting in the standby queue before the maintenance, and the requested walltime of these jobs exceeds the new standby limit (4 hours), these jobs will not start after the maintenance, and will stay in the queue indefinitely. If you notice such jobs, you can choose any of the following remediation strategies:

1. Stay in standby, but modify the job to use shorter walltime: scontrol update job JOBID timelimit=4:00:00
2. Keep the time limit, but modify the job to use your lab's dedicated queue: scontrol update job JOBID account=MYLAB partition=a
3. Delete the job altogether and resubmit it into a different queue or with a different time limit: scancel JOBID

Please report any issues to rcac-help@purdue.edu and your EUP liason.

Link to original posting ORIGINAL: March 8, 2023 8:00am - 5:00pm EST

The Negishi cluster will be unavailable Wednesday, March 8, 2023 from 8:00am - 5:00pm EST for a scheduled weekly EUP maintenance. The cluster will return to its early user operations mode by Wednesday, March 8th, 2023 at 5:00pm EST.

Any Slurm jobs which request a walltime which would take them past Wednesday, March 8th, 2023 at 8:00am EST will not start and will remain in the queue until after the maintenance is completed.