Diagnostic tools for tracking down machine freeze

[Nigel Henry]

On Sunday 25 October 2009 22:25, D. R. Evans wrote:
> Nigel Henry said the following at 10/25/2009 03:02 PM :
> > I've always had the odd freeze up on this new machine of mine, but mainly
> > with some distros, like Archlinux for example. But now I'm getting
> > frequent freezeups on Kubuntu Dapper, and Debian Etch was freezing up
> > after about 5 mins.
> >
> > I've been running Dappers Memtest86+ v1.70 for over 10hrs now. 14 passes
> > done with no errors, so the memory appears to be ok, and the machine
> > hasn't locked up.
> >
> > This machine is on a KVM switch with another machine which is
> > infrequently booted up. Are KVM switches known for creating such
> > freezeups?
> >
> > Any suggestions as to diagnostic tools I could use to try and track down
> > the problem?
>
> Are you running lmsensors?

It's only running on one of the distros on the machine (9 distros on it), so 
that I could see if Gkrellm would display the cpu temp. Just checked on 
Hardy, which hasn't frozen yet, and lmsensors is not installed.
>
> If so, remove it and any modules it loads.
>
> It took me a year of intermittent freezes to figure out that that was the
> problem on one of my machines.

I'll keep that in mind.
>
> I have used a KVM switch for years; I've never known it cause a freeze. (I
> have, though, occasionally fingered it as responsible for getting a system
> into such a confused state that I've ended up rebooting; I've learned now
> not to switch amongst my machines [I have four on the switch] when I'm
> booting or closing down one of them.)

Thanks for the feedback on the KVM switch question.

Could the problem be clock related. I ask because I checked the clock in the 
BIOS, and it was 2hrs ahead of UTC. I reset it to UTC, rebooted Hardy, and it 
showed the correct time of UTC +1 (CET) clocks changed last night in France 
to add to the confusion from CEST to CET. Just rebooted, and checked in the 
BIOS, and clock shows time being saved correctly as UTC. 

I think I'll have to check all the distros clocks on this machine, as I know 
one of them messes the time up, and when I boot up a different distro, it 
shows it 2 hrs ahead, until ntp sets it correctly.

Darn. I've just noticed that Hardy has frozen up 3mins ago, and it had only 
been up and running for about 10mins. Weird eh.
>
>   Doc

Thanks for the suggestions.

Nigel.