Re: [cobalt-users] Logging CPU temp

["Richard Donahue" <sales@xxxxxxxxxxxxx>]

>
> > How can I log my CPU temp? My RaQ4 sometimes crashes at 4:02 AM. I've
> > removed the scripts that I installed to run in cron.daily and
> > everything is
> > fine. I added a couple (logcheck.sh and chkrootkit.sh) back
> > to cron.daily
> > and it crashed again this morning. Someone said I should
> > check my fans and
> > CPU temp because it may be causing the server to restart. I
> > never get any
> > warning messages, and the logs don't show anything useful
> > that I can see.
> > The CPU temp is always around 35C when I check it.
> >
> > Thanks,
> >
> > Rich
> > EBS
> >
> You are probably experiencing heavy cpu loads around 4 am, due to logs
> rotating. Your system wouldn't crash at 4 am all the time, due to fans.
They
> are more likely to lock up in the day time. I would suspect a job running
> around that time.
> We get a memory dangerously low on 2 of our Raqs every day. The email
comes
> in between 4am and 5 am. Our log rotation is causing that.
> Tom
>

I need to solve this, so I just ran logrotate and watched the CPU temp and
free memory. The CPU temp got up to 58.5C and stayed around that for about 3
minutes then the server restarted. Right now it's at 45C. I guess 58.5C is
too high for my server. I'll have to look at a few things.

The free memory stayed around 2 to 9 MB's. I have 512 MB's and it use any of
the swap that I could see. The server memory use hang around 100 to 150 MB's
free.  I think there's something wrong with the memory. I'll swap the chips
out.

Rich
EBS