Note that we will start pausing running jobs at 5AM. This will permit your work to have as much access to compute resources as we can allow while still performing scheduled maintenance tasks. We will re-enable all jobs after the slurm upgrade is complete.
Research Computing's July 6 maintenance activities will be:
1) Upgrading the slurm cluster scheduling software for bug fixes and performance upgrades.
2) Testing the spunnel and x11 plug-ins. Their proper operation is critical to some of Odyssey's users.
3) Rolling reboot of all login (rcloginXX) and nx (holynx01 and rcnx01) nodes.
During this time, the Odyssey cluster will be unavailable. If at all possible, we hope to have all services up and available prior to the official 11AM event close.
Please remember that Research Computing is continuing to enforce data retention policies on all of RC's scratch filesystems (regal and holyscratch). All files that have not been modified for the last 90 days will be deleted.
Thank you for your anticipated cooperation and patience.