Random Restarts?

[Pieter de Goeje]

On Sunday 20 May 2007, Edward Ruggeri wrote:
> On Sun, 20 May 2007 11:41:36 -0500, Roland Smith <rsmith at xs4all.nl> wrote:
> > On Sun, May 20, 2007 at 11:03:25AM -0500, Edward Ruggeri wrote:
> >>  Hi,
> >>
> >>  My system randomly reboots, usually in the evening.  It is definitely
> >> not a
> >>  soft reboot, since the filesystems don't get properly dismounted.  My
> >>  suspicion is that it is a heat related issue -- I do leave the computer
> >>  running just about all day long, and it has started to get warm.  Then
> >>  again, it's coolest in the evening...
> >
> > Try and install the mbmon port, and see if it works on your machine. If
> > so, start a cron job that appends mbmon output to a file say every 15
> > minutes. If it's a heat buildup issue in a monitored component, it would
> > show.
> >
> > I wonder though. My machine usually doesn't need a day to heat up after
> > a cold start. An hour or so usually suffices.

In my experience this kind of problem can also be caused by bad memory.

> >
> > Other causes could be a spike in the line voltage due to a large device
> > switching on or off nearby. Or an underrated power supply overloaded
> > through a cron job.
> >
> >
> > Roland
>
> Thanks for the ideas, Roland (and Tamouh)!  I forgot about the possibility
> of power issues.  I do have an Antec power supply, 500W, which certainly
> doesn't mean it isn't the problem, but it _ought_ to be able handle this
> system...
>
> But here's what I get if I run mbmon.
>
> %mbmon -A -t -r 1
> TEMP0 : 39.0
> TEMP1 : 34.0
> TEMP2 : 25.0
> FAN0  :    0
> FAN1  : 5818
> FAN2  :    0
> VC0   :  +1.28
> VC1   :  +1.50
> V33   :  +3.30
> V50P  :  +4.97
> V12P  : +11.13
> V12N  : -11.52
> V50N  :  -3.76
> Sun May 20 12:32:10 CDT 2007
>
> That CPU temp is at about 97% idle -- high, no?  The voltage on the 12V
> lines seems pretty bad, and especially so on the -5V line.  Should I trust
> the sensors and think about a new PSU?  In the meantime, I'll write to a
> file, and see if it records a dip in power before the next reboot.
I wouldn't trust it to report correct (absolute) values. I've seen 
motherboards report values _way_ off.

HTH,
Pieter de Goeje