CHPC - Status Page

Another set of critical vulnerability patches requires updating OS kernel on CHPC Linux clusters. We will be starting rolling reboots of compute nodes momentarily that will update to the latest kernel. Unfortunately, just today another vulnerability has been found which does not have a patch, which means that another rolling reboot will be needed as soon as the patch is released. That is, there is a potential for two rolling reboots in the near future.

This process will proceed as follows:

• Compute nodes will be drained so they stop accepting new jobs. Running jobs will be allowed to finish normally.

• Once a node is idle it will be rebooted onto the new kernel and go back into service automatically when it comes up.

• Nodes still to be updated will show the drain reason "auto_kernel_update" in sinfo -R.

• With fewer nodes accepting work, some jobs may wait in the queue longer than usual, especially large or long-walltime jobs.

• To accelerate the rollong reboot process, guest, OSG and freecycle jobs may be cancelled.

We thank you for your understanding in this difficult period of many newly discovered vulnerabilities. Contact helpdesk@chpc.utah.edu if you notice any irregularities.