System Downtime December 15 2020

Date: 
Tuesday, December 15, 2020 - 7:00am to 9:00pm
Location: 

Ohio Supercomputer Center

A downtime for all OSC HPC systems is scheduled from 7 a.m. to 9 p.m., Tuesday, December 15, 2020. The downtime will affect the Pitzer and Owens Clusters, web portals, state-wide licenses, and HPC file servers. Login services will not be available during this time, including MyOSC.

In preparation for the downtime, the batch scheduler will begin holding jobs that cannot be completed before 7 a.m., December 15, 2020:

  • Jobs that are not started on Pitzer will be held until after the downtime and then started once the system is returned to production status. 
  • Jobs that are submitted to the current Owens cluster but do not get started before 7 a.m. on December 15, 2020 will be removed from the system; we will inform the impacted users to resubmit the jobs. To help reduce the number of jobs to be removed, we will temporarily disable the 'longserial' queue and start to reduce the maximum walltime by 24 hours per day until the max walltime reaches 24 hours on all queues on Owens beginning on Tuesday, December 8. The maximum walltime will be changed back to the original value and the 'longserial' queue will be reenabled once the downtime completes.

Highlights of the downtime activities include:

  • Switch to Slurm for job scheduling and resource management on Owens. See slurm migration page for details.
  • Updates to the software environment on Owens for Slurm compatibility. See owens slurm environment changes page for details.
  • Regular systems maintenance
  • Changes to network routing and firewall configuration
  • Switching to clustered cron for cluster login nodes
  • Upgrade to Globus v5.4

To stay up to date on system notices, follow @HPCNotices  on Twitter. As always, you can contact us at OSC Help