Dear NERSC users, We are preparing for a reboot of Perlmutter, to stabilize the system after a combination of two bugs (one in I/O forwarding and one in the networking software) caused a large number of nodes to fail. We are currently planning the reboot for tomorrow (Wednesday) morning. There is a reservation in the system to avoid starting new jobs that would not complete before the reboot. Some currently-running jobs in the preempt queue are likely to be killed. The timeline for the reboot is still being finalized, with the aim of minimizing further disruptions. We will send an update when the timeline is more firmly established. Thanks for your patience as we work on stabilizing Perlmutter. Sincerely, Steve Leak (he/him/his) NERSC User Engagement _______________________________________________ Users mailing list Users@nersc.gov