Mystery resets = less coding time..

[Steven M. Klass]

Hey all,

	I've run up on a stumper and I can't to figure out how to debug this one.  
I've got a home built PIII 500, 256 Mb ECC, 30 Gb HD, basic workstation.  
It's running a very heavily modified slack system - complete with a full 
reiserfs - X 4.03 - and many other changes.  
Anyway, about a week ago we had a power outage, and for the first time in 
months it was shutdown (actual power removed vs reboot)  After power was 
restored the system came back up, and I logged into X.  The KDE Splash screen 
stayed there forever (like 5 minutes) and then all was fine. I logged out and 
logged back in at which time everything acted normal again.  I had one 
problem which seemed wierd and unrelated - I could no longer nfs (encrypted 
cipe) to my work.  Two nights ago I (like many nights) stopped my work and 
went to bed - No save, just turned off my monitor and went to bed.  I woke up 
the next morning to an xlogin screen.  Huh??  What happened?  So I looked at 
my logs - No hint of the reboot 

Jun 26 21:53:33 ethel -- MARK --
Jun 26 22:13:33 ethel -- MARK --
Jun 26 22:33:33 ethel -- MARK --
Jun 26 22:53:33 ethel -- MARK --
Jun 26 23:13:32 ethel -- MARK --
Jun 26 23:33:32 ethel -- MARK --
Jun 26 23:53:32 ethel -- MARK --
Jun 27 00:13:00 ethel syslogd 1.3-3: restart.
Jun 27 00:13:01 ethel kernel: klogd 1.3-3, log source = /proc/kmsg started.
Jun 27 00:13:01 ethel kernel: Inspecting /boot/System.map
Jun 27 00:13:02 ethel kernel: Loaded 14801 symbols from /boot/System.map.
Jun 27 00:13:02 ethel kernel: Symbols match kernel version 2.4.4.

So I looked at my .xsession-errors - Nothing

So I logged in and began working.  All of the sudden <no warning> reset.  
Like someone pushed the reset button.  Fortunately (or so I think) I have 
reiserfs, which seems to be coping (at least better than me) with this 
irritation.  This happend 4 times last night.  

So now how do I begin to troubleshoot this?  I haven't touched my OS in a 
while (>week) and I use it daily.  If this is a hardware problem how do I 
find it?  It's not an obvious fault (like a dead harddrive).  Then look at 
the time for the reboot?  The time certainly seems suspicious - like a cron 
job.  But I don't know of a "hard reset" ability with a computer..    Any 
ideas are welcome.  

Thanks
-- 

Steven M. Klass
Physical Design Engineering Manager

Andigilog Inc.
7404 W. Detroit Street, Suite 100
Chandler, AZ 85226
Ph: 480-940-6200 ext. 18
Fax: 480-940-4255

sklass@andigilog.com
http://www.andigilog.com