chip and RRStor Downtime Scheduled for August 25th

To all chip users,

On Tuesday August 25th, the chip cluster will be inaccessible from 0800ET - 1800ET. This maintenance window will focus on installing critical security patches for the chip cluster, upgrading the version of ceph running on RRStor to improve performance, and implementing Slurm scheduler tweaks to reduce queue bottlenecks caused by large job volumes.

We understand that this downtime will be disruptive, but the Research Computing & Data Team will endeavor to complete this and other routine maintenance quickly and alert the user base to resumed operations.

Please note that the current downtime plan does not include cancelling running jobs, but any jobs requiring access via the login nodes (e.g., interactive jobs) will be disrupted. Though it is not the plan, do note that circumstances may arise that could interrupt access to storage volumes or require the compute nodes to be rebooted, which would cause running jobs to be killed; please plan your jobs accordingly.

Remember to submit any issues you have here:https://rtforms.umbc.edu/rt_authenticated/doit/DoIT-support.php?auto=Research%20Computing

and check out our documentation here:https://umbc.atlassian.net/wiki/spaces/faq/pages/1082589207/UMBC+HPCF+-+chip

--

Give me all the bits!

Gregory Ballantine

HPC Specialist