August Maintenance Complete

Posted on

We have completed our August maintenance downtime, and we appreciate everybody's patience while our systems were unavailable. During the maintenance, we accomplished critical upgrades to our HPC, storage systems, power, and networking infrastructures.

Specifically, we accomplished the following items:

  • Upgraded our storage system, GPFS (Spectrum Scale) from version 4 to v5.0.4
  • Upgraded the firmware on the GPFS and Research Archival storage hardware
  • Upgraded Slurm Scheduler from v18.08 to v20.02 (release notes)
  • Completed our InfiniBand rollout.
  • Completed some critical power upgrades and tests to ensure reliability for our colocation customers

Our next downtime is planned for December, when we will update the firmware on our switches, and it will be much shorter than the August maintenance. Look for an announcement on the HPC Notice List as our plans firm up.