Technical Staff Blog

Last update on .

ibm-gpfs.jpg The department experienced a GPFS blip shortly after 10am this morning. The outage become pregressively worse and lasted about 10 minutes in total.

For those interested in more details, one of the quorum and manager nodes for the cluster experienced a system hard drive failure a few days ago. After replacing the hard drive and reinstalling the OS, there was a mismatch between the version of GPFS it should have been running and the version that was installed. This caused an inconsistency in the GPFS cluster. The package archive was out of sync with the running version of GPFS, causing the version of GPFS to be downgraded on the reinstalled machine.