Event Date Summary
Most CPU nodes are back in service and running jobs; some further examination of power panels for GPU nodes will delay their return to service until as late as noon.

Incident description

System Incident status Start Date End Date
Graham Closed
Created by Mark Hahn on


Planned Outage - Arrêt planifié


Graham compute nodes will be unavailable for about an hour starting at 6:50 AM on December 22. During this time, you WILL be able to login and access storage, but no jobs will be running on compute nodes. The scheduler will refuse to start job that would extend into the outage period. This outage is for placement and removal of phase measurement instrumentation. A second outage for removal will happen starting at 6:50 AM on 2021-01-05 and take approximately an hour. Graham is now operating normally.

Updated by Mark Hahn on