[sf-lug] Pointers for how to track down source of system hang

[Rick Moen]

Quoting Mark K. Zanfardino (mark at zanfardinoconsulting.com):

[Trying to track down mysterious system hangs, you've run your system
with and without Compiz, and with and without some sound driver or other:]

> As an aside, it has at times remained very stable over as long as a week 
> or more with both Compiz running and the sound driver installed, which 
> makes me think it's going to be something else.  The key for me is to be 
> sure I'm checking all relevant logs and that I have all possible logging 
> enabled.

Well, whether you'd find anything useful in logfiles depends on the root
cause of your symptom.  _Some_ root causes, including most hardware
glitches, wouldn't leave much, or might leave nothing, at the software
level.   

You've definitely looked at all the places (system logfiles) I'd check
if I suspected some system-wide problem such as a major hardware glitch.
Since that didn't help (yielded no information), a change of tactics may
be in order.

The really annoying bit is that diagnosis in such cases, although it
always reaches a definitive conclusion if you persist and are careful
about your assumptions and judgements, takes time away from, y'know,
using the computer for doing something useful.  For example, in your
shoes, I might be tempted to use one or two live-CD Linux distributions
for a few days, deliberately pushing the hardware hard, e.g., chewing up
lots of RAM, banging on the CPU, etc., and see if the symptom persists.

If it doesn't, then it's something about the installed software.  If it
does, then it's probably a hardware problem.

It's helpful to maintain a list, either on a notepad or in your mind (if
you can) of all the suspects that _could_ be causing your problem.  As
you try various things, watch carefully to see if you've ruled any of
them out.

There are also a couple of ways to torture-test ("burn-in") your
hardware in cases where you suspect a problem there, which I can post
about separately.

> I'm still relatively new to Linux....

Ah, that explains the relative sanity.  We can help with _that_ problem,
at least.  ;->