rehi,

"Gwilym Evans" <meatgroup@iprimus.com.au> writes:

> The first crash produced no error, it just stopped, dead, after about 12
> days up. The 2nd two showed the above on the local terminal.
> 
> I doubt I can get a definate answer about what it is straight away, so I'm
> asking how can I debug this further? What info should I try and get next
> time so someone may be able to help me further?

I had a similar problem, which I posted here some time ago. Though I got
no useful answer, I now know a little more.

The way I tracked things: My machine hangs occasionally with no log
footprint at all. Nothing. And nothing on console.

The machine is almost dead then, but still pingable by IP number, but
does not resolve host names. So services affected would be DNS, syslog,
NFS and some more which I thought all rely on syslog.

But it seems simple here, they all rely on disk writing and syslog is
the first one to "die" if some process tries to report a problem.

So please try to isolate how dead your machine becomes. If the kernel
complains heavily, it's pretty dead, but otherwise it might be not.

 - do logged in sessions still work, try something without disk access
   (swapping included) like "echo 1" or typing at all first

 - can you ping some ip number (from another machine to the suspect)

 - can you read and write to disk

 - do you have lots of disk space on a small machine (not enough RAM to
   handle caching data etc.). There is a limit I don't know how to deal
   with

 - if the latter, setting up a virtual file system (vnconfig) and
   filling it up freezes my machine. Right now I don't have time to
   track this, so I just avoid it.

cu Norbert
-- 
Today is the first day of the rest of your life.