[Techtalk] kernel panic

[Maria McKinley]

Rudy Zijlstra wrote:
> Op donderdag 23-10-2008 om 09:24 uur [tijdzone -0700], schreef Maria
> McKinley:
>> Hello there,
>>
>> I had a mail server have a kernel panic this morning, and I am trying to 
>> find out what happened, but it seems totally mysterious. The last entry 
>> in the log is a standard postfix disconnect. All systems seemed fine 
>> right before the panic:
>>
>> http://www.shadlen.org/munin/Servers/ella.shadlen.org.html
>>
>> None of the partitions or drives are remotely close to being full. There 
>> may have been a network outage around the time of the panic, but this 
>> should not have caused a kernel panic. This machines runs a mail server, 
>> web server, and bind9. It booted back up just fine, and seems to be 
>> working happily now.
>>
>> Is there something I should be logging that I am not, or something else 
>> that may give me a clue as to why this happened?
> 
> Hi Maria,
> 
> I am impressed by all that you log.
> 
> I fear though, that this is a case where the information on the console
> is of crucial import. 
> 
> If you start getting this more often, and you are running with X
> enabled, then i would suggest to enable either network console logging
> or serial console. In both cases you can capture the kernel panic dump.
> It is the dump on the console that contains the information to trace
> which aspect caused the panic. Without this, all you can do is guess.
> 
> Cheers,
> 
> 
> Rudy
> 
> (who has had some kernel panics.. )

Thanks Rudy.

This is a server, so I don't have X enabled, as I don't really have user 
accounts on it. Anybody know of a way to log the kernel panic dump that 
doesn't involve X? I've always wondered if there was a way to log this 
stuff.

thanks,
maria