NIH HPC News & Announcements
Some NIH Biowulf cluster nodes will have reduced availability until May 15
Date: 04 May 2023 11:05:08
From: Tim Miller
The HPC staff has been informed that urgent repairs are required to the
data center infrastructure that supports a portion of the Biowulf
cluster, and unfortunately, several groups of Biowulf nodes need to be
shut down during the repair process. These nodes include:
- All CPU nodes with the e7543, x6140, and x6240 Slurm properties
- All A100 and V100X GPU nodes.
The repair has been scheduled for the morning of Monday May 15, 2023.
Currently running jobs on those nodes will not be affected. Queued jobs
requesting these nodes that would extend into the repair schedule will
not be started by the batch system, but will remain queued and will be
started when the nodes are brought back online. Jobs with shorter
walltimes (that would complete before the scheduled repair) will be
started by the batch system depending on resource availability.
Helix, the Biowulf login node, hpcdrive, the nihhpc Globus endpoint, and
the remaining cluster nodes will remain available during this scheduled
repair.
Apologies for the disruption; however, this repair is necessary to avoid
potential damage to the nodes.
########################################################################
Please contact staff@hpc.nih.gov with any questions about the NIH HPC Systems
[Last 12 months of HPC announcements]