Mysterious Server Shutdown

Virtually no UNIX admin experience. Any admin duties are shared by several folks with no special training. Today we had our Sun v880 server, running Solaris 5.8, shutdown for no apparent reason.

When we checked on server we found it completely powered down, yet still connected to a fully charged UPS. When we booted the system back up a few of the file systems required FSCHK be run on them. Other then that...no damage done.

Can anyone point us to any system logs that might give us some insight as to why our server shutdown?

Any help would be greatly appreciated!

Try these for starters and see what you find..

last -20
who -b
view /var/adm/messages

Oh.. since you had to run an fsck on some of your filesystems I guess you don't have logging enabled.

To enable logging, edit your /etc/vfstab file and add logging to your ufs..
ie:
/dev/dsk/c0t0d0s0 /dev/rdsk/c0t0d0s0 / ufs 1 no logging

You might want to run a prtdiag aswell to make sure nothing else has failed..
/usr/platform/`uname -i`/sbin/prtdiag -v

Another quick way to check your messages files is to type dmesg

Some time ago I expreienced the same problem.

The problem was that I did't have any real console connected to the machine. Isted I used the hyperterm in windows and connected via the serial port.
After I left this connection to my unixserver I just klicked the (X) box up in the right corner of the hyperterm window. When I did the the machine halted but first it seemed like it was totaly hanged. This happend 3-4 times until I figured out what was wrong.

The fix was to just type 'go' and the machine came up in runlevel 3 again.
Nothing was in the messages...
If this is your problem sun have a patch for this. Or you can comment out
KEYBOARD_ABORT=alternate in /etc/default/kbd

if people have easy physical access to the server, i wouldn't be surprised if somebody hit the power button or kicked/loosened the power cords by "accident" ... if the server is within a secure facility, check who has access and who would have been in the room at the time of the powerdown ...

also, check for ventilation --- make sure the server did not get too hot and there is nothing obstructing the ventilation fans ...

i'm more inclined to say that a finger or a foot touched the server based on what you have written but there would be no way to confirm that if you can't prove somebody was in the same room as the server when it went down ...

good luck!

check /var/adm messages file.
to check other physical devices run prtdiag -v, format, and iostat -En.
Run an extended post or create an explorer file and analyze it if the problem continues

Just came accross these pdf files that would be usefull for you...
It will help you with the sysadmin of your sun servers...Well get you started anyway

http://www.fisica.uson.mx/carlos/Software/RefCards/Solaris_quickref.pdf