John F. Morse wrote:
> Robert Harris wrote:
>> John F. Morse wrote:
>>
>>> I have an old AMD-K6 266 MHz PC with 128 MB of RAM that I upgraded from
>>> Debian 3.1 to Debian 4.0 by doing the "apt-get dist-upgrade" while on
>>> the Internet.
>>>
>>> It was crashing for a few days, but it isn't the Debian OS because it
>>> also crashed when using a DSL LiveCD and a Knoppix LiveCD. I thought
>>> that those tests would eliminate the Debian OS as well as the HDD.
>>>
>>> It ran for a week without crashing when running from tomsrtbt on a
>>> floppy.
>>>
>>> I fired it up again on the Debian 4.0 installed on the HDD and it ran OK
>>> for a week until I accidentally killed the AC power feed which ended the
>>> week's uptime. Then I added a UPS and rebooted it.
>>>
>>> It has ran for 5 days without failure until a short while ago.
>>>
>>> I had a tail -f running on the syslog and just now when I checked it I
>>> noticed the box was in trouble.
>>>
>>> [snip]
>>>
>>
>> Try doing a memory test, like memtest68+
>>
>> Robert
>>
>
>
> Took your advice, even though I didn't suspect a memory issue, since
> this server has ran for many days in the past with various OSes (see
> above).
>
> The memtest86 has ran for seven straight days now without any failure.
>
> The kerneloops.org site says to report oopses from distros to the
> various distros instead of their site. So it looks like I'll have to
> report it to Debian, but would like to figure it out myself since it
> must be my hardware (other Etch PCs aren't oopsing).
>
> Do you have any knowledge of what the oops message indicates, such as a
> particular piece of hardware?
>
> Here it is again:
>
> john@optima12:~$ date
> Tue Apr 29 22:57:46 CDT 2008
> john@optima12:~$ uptime
> 08:18:02 up 4 days, 10:23, 3 users, load average: 0.04, 0.02, 0.00
> john@optima12:~$
> Message from syslogd@optima12 at Wed Apr 30 11:56:39 2008 ...
> optima12 kernel: Oops: 0000 [#1]
>
> Message from syslogd@optima12 at Wed Apr 30 11:56:39 2008 ...
> optima12 kernel: CPU: 0
>
> Message from syslogd@optima12 at Wed Apr 30 11:56:39 2008 ...
> optima12 kernel: EIP is at drain_array+0x10/0x7f

Well, that is in the slab allocator, so it could well be a driver
problem. Does your kernel include any non-free drivers (e.g. nvidia or a
Windows wireless driver)?

Robert
>
> [snip]