[nycbug-talk] FreeBSD rebooting every 120 seconds =-(

[Okan Demirmen]

On Tue 2005.12.27 at 15:50 -0500, N.J. Thomas wrote:
> * Jim Brown <jpb at sixshooter.v6.thrupoint.net> [2005-12-27 12:30:39 -0500]:
> > > This morning I came back in to find that it is rebooting roughly
> > > every 2-3 minutes or so, but only when I access the disk.
> > 
> >   - excessive heat
> >   - memory chips failing
> >   - an installed card (video, hard disk, usb, serial, etc.)
> >   - static buildup
> 
> I think (but I'm not sure) it is a heating issue.
> 
> I ran memtest86+ on the machine right now and it went through two
> complete passes with no errors.
> 
> However, someone noticed that one of the fans on the side of the case
> (this is one of those fancy see-through cases with two fans built into
> the side covers) was not running...so I cracked open the cover for a bit
> and lo an behold, the machine ran for about 2 hours before rebooting
> again.

ah yes, heating...err, rather, cooling.

> This is an AMD Athlon XP 1800+ (1.5GHz) system, and I understand that
> AMd chips run hot, so I think cooling is more critical than in Intel
> systems.
>
> I went into the BIOS right now and I saw that the CPU temp was hovering
> between 50 and 54 degrees Celsius. It has an alarm more that will warn
> you if your CPU is above some pre-defined temp, but since it allows you
> to set this temperature between 50 and 75C, I'm not exactly sure what
> "normal" is.
 
$ sysctl hw.model
hw.model=AMD Athlon(tm) 64 Processor 3200+

runs at:
hw.sensors.10=lm0, Temp2, temp, 31.50 degC / 88.70 degF

i run all my sensors through rrd, as well as, having sensorsd tell me
when things "look" like they are going too hot/cold/slow/fast...i'd
suggest the same for you.

> In any case, just to rule out unseated cards...I'll open this box up,
> reseat everything and clean out the dust bunnies on the inside of the
> machine (there is quite a bit in there).

dust can also block airflow ;)

cheers,
okan