UPDATE: 5:12 PM: The Lustre management tools have caught up and are reporting normally. Moving status to operational.
UPDATE: 5:02 PM: Reboot is being reconsidered as performance seems OK. Suspicion that the diagnostics are not reporting up-to-date info.
UPDATE 4:50 PM: A full reboot of Regal is necessary at this point. Jobs will be suspended, but due to timeouts which have likely already happened, some jobs reading from/writing to Regal may fail.
Regal scratch is experiencing performance issues again. We are actively working on the problem. Updates to follow.