We’ve determined the cause of the load spike: the backup HDD in Miller is failing. I’m working with the NOC to schedule a replacement time for tomorrow; I will spend today making secondary backups to my local machine here at the office.
No data has been lost. The load spike was caused by the HDD failures hanging the backup script, which is the only script that ever touches the secondary HDD.
Update: The NOC has to do an analysis of the HDD prior to replacing it; this will occur at 1300 CDT tomorrow. The server will be offline for a few hours after that.