Mysterious reboot

[Tamouh H.]

 > 
> Alex de Kruijff <mailto:freebsd at akruijff.dds.nl> wrote:
> > On Thu, Feb 16, 2006 at 01:46:19AM -0600, Mike Loiterman wrote:
> >> Afte being up for more than two months, I just noticed that my ver 
> >> reliable and stable server mysteriously rebooted.  I've
> had the same
> >> hardware and system setup for more than a year and haven't ever 
> >> experienced anything like this before.  I haven't added any new 
> >> hardware in a very long time.  I did upgrade mysql this
> morning via
> >> portupgrade, however.
> >> 
> >> The system appears to have crashed, as dmesg shows that
> the various
> >> filesystems weren't unmounted correctly upon reboot, but
> aside from
> >> that single error, nothing else points to what the problem was.
> >> 
> >> /var/log/messages just shows a user connectig via pop, and
> then the
> >> next line is the machine booting.
> >> 
> >> We are having a pretty good rainstorm, but nothing else
> seems to have
> >> lost power and I have several other machines in this space
> that did
> >> not go down (servers, TVs, XBOX, laptops, etc).
> >> 
> >> The machine seems to have gone down around 11:23 PM, so I checked 
> >> /etc/cron to see if I had scheduled a job to be run at
> that time, but
> >> there are no jobs scheduled to be run on or near this time.
> >> 
> >> Where can I begin to look for more clues as to why this
> machine seems
> >> to have crashed?  Where would a kernel core dump be or a
> mysql core
> >> dump?
> > 
> > You can look at /var/log/messages and find core files with 'find / 
> > -name \*.core'
> > 
> 
> The only real explanation was some sort of power fluctuation. 
>  I'll continue to monitor it carefully, but there doesn't seem to be 
> anyway to definitivly identify the cause, but the machine crashing 
> under load is looking less and less likely.
> 
> If there is a problem, I suspect it will show up again within the next 
> week or so, otherwise I'll chalk it up to a random event in the 
> unverse.

I have had the exact issue with a dual xeon and FreeBSD 5.4 with 4GB RAM. It was working perfectly fine and after a month or two all the sudden the machine rebooted by itself. Well, I shouldn't say all the sudden, one of the circuit breakers blow up taking the server down, but since that time , the server would do sudden reboots.

No errors, no dumps, no panics. Just a simple reboot. Then it started becoming more often (it reboots every 15 days, then 10 days, then 5 days). So I began to suspect it is a power issue.

We use a remote reboot power switch, so I moved it to another socket, and so far 20 days, no reboots. Not sure if it will go crazy again, but I hope it doesn't.

The power supply is 500W so should be more than sufficient for 4 SCSI drives and dual xeons.