MDM Trouble Report for 2007 Sep 18

Telescope: 1.3m
Observer(s): Eastman (OSU)
Instrument: 4K

Problem(s) Encountered:
mcgraw became completely unresponsive at 12:30 am (local time) 
while changing filters/transferring an image during the 
execution of a prospero script (katie15p.pro). 

The mouse froze, I couldn't ssh to it, it didn't respond to 
alt-ctrl-delete/alt-ctrl-backspace, and I couldn't even ping 
it from hiltner.

I did a hard reboot. On start up, I saw this warning: 
 Raid array is not clean starting background reconstruction . 
I didn't force a system check.

Everything started back up without any problems.

The file that was writing as it crashed 
(/lhome/data/4kks070918.0067.fits) was truncated to 
218,752 bytes (the normal size is 33,560,640 bytes -- 
so it got 0.6% done). The fact that it crashed just as this 
was being written is probably not a coincidence.

Only the standard windows were open (ds9, iraf, 
telescope/instrument control, terminals, and maybe firefox), 
and I wasn't doing anything else out of the ordinary.

After I restarted my script, I remembered about Rick's new 
 RECOVER  command. When the script was complete, I tried it 
( RECOVER 5 DISK1 ), which is supposed to grab the last 5 
images written to DISK1 (my script took 4 images).

It said:

> Re-initialize all data taking system tables to recover/updates the system <y|n>?

I said no, and it did nothing, so I tried again and said yes. It responded

> OK, this will take a little time...
> 
>  Unknown keyword
>  DISK1
> 
> in the command RECOVER
> Usage: recover [verbose]
> Execution error in Prospero command.
> Command: RECOVER 5 DISK1 
   
I also tried it with  RECOVER 5 DISK2  and got the same results.
As this was more of a curiosity and it only took 3 minutes to 
replace the lost exposure, I didn't bother with it anymore.

A few days ago, mcgraw crashed during the day and I forgot to 
report it. Unfortunately I don't remember what I was doing at 
the time. I do remember that I couldn't ssh to it, it didn't 
respond to alt-ctrl-delete/backspace, but I could ping it. 
I did a hard reboot and when it restarted, the X windows didn't
start. I rebooted it through the command line and then it 
started normally. When I started up prospero, the sequencer 
didn't start. It took a seqinit from the IC to get it to go.


------------------------------

Submitted on 2007 Sep 19 [6:11:55]