Another Lustre server crash

On Monday, April 30, 2012 around 9:44 AM, one of the Lustre OSS machines crashed. Just prior to that, at around 9:00 AM, the load on the machine started ramping up. It was not as big of a spike as some others we have seen though. At around 9:20 AM, many jobs failed for at least one user, and at 9:23, compute-9-208 failed while writing the Lustre journal entries.