I walked into the control room this morning to a symphony of names of optics being announced on the verbal saturation alarm.  The ISC_LOCK guardian had connection errors, and so was not able to recognize the fact that the lock was lost, so it had not run the DOWN state.  I stopped and restarted the ISC_LOCK guardian, which cleared the error long enough for it to start the DOWN state. 

While the DOWN state was resetting things, I checked the alog and see that Jamie predicted this behavior in alog 44383, since a similar thing happened at LLO recently. 

I think the take-away here is that if a front end computer has crashed, we should ensure that the IFO is in its DOWN state.  (To do this, in ISC_LOCK guardian, request 'manual' control, then select DOWN, then request 'auto' control.  You may also have to ask the node to STOP (pulldown menu that is usually on EXEC), then request EXEC again.)  I'm happy to receive a call to do this remotely if someone is busy with other recovery. 

Carlos and DaveB are already working on recovery of the susetmy computer.