[LAU] random system lockups, memtest clean, no idea... Magic sysrq?

Loki Davison loki.davison at gmail.com
Tue Nov 18 17:30:02 EST 2008


On Wed, Nov 19, 2008 at 9:02 AM, Kevin Cosgrove <kevinc at cosgroves.us> wrote:
>
> On 18 November 2008 at 15:26, "Christopher Stamper" <christopherstamper at gmail.com> wrote:
>
>> > Look for errors in your /var/log/ files?  Make sure you have a look
>> > at the X log files.  I've had troubles there from time to time.
>> >
>> > I hated it when I had this trouble, which turned out to be the wrong
>> > type of RAM for my mobo.  Good luck!
>> >
>>
>> I have the same problem. I couldn't find any errors in the logs...
>>
>> I'd be interested to know what kind of system you have. I have an intel
>> mobo, 1GB DDR2 RAM at 800mhz, Intel Pentium D @ 3.2ghz, Intel embedded
>> graphics, and a WD SATA HD @ 16GB/3gbps.
>>
>> I'm suspecting a graphics problem, but a fix doesn't seem to be coming
>> easily...
>
> My 1st experience was with a Pentium I system.  I put 4 sticks of
> 64MB EDD RAM in it.  The early generation of my mobo had timing
> trouble with such huge server RAM, and I had to go back to 4
> sticks of 32MB EDO RAM.  After that it never had troubles again.
> I had done all of the memtest86 stuff, and done it for days on
> end without error.  I'd swapped individual 64MB sticks out for
> weeks on end, and never found the problem until I stumbled on an
> old app note about the board indicating its limitations.
>
> My 2nd experience like this was after upgrading my distribution
> and getting new drivers for my ATI Radeon card.  Eventually I
> found out that I needed to run it in "stupid" mode, meaning
> no hardware acceleration, not Xorg, nor ATI native.  Running
> hardware acceleration would cause it to freeze.
>
> Both problems involved a lot of very deliberate, change only one
> thing at a time, troubleshooting.  In the 1st case, the error
> would only occur about every 2 weeks, but it would take down the
> system and corrupt the disks sometimes.  Eventually I found a
> kernel oops message that would occur about 1-2 days in advance,
> and I'd grep'd the log files with cron to detect the trouble and
> reboot preemptively.  I knew I'd nailed that problem when the
> system stayed up for 6 weeks.
>
> I wonder if the original poster's problem would go away if the system
> was booted into console mode only, no graphical interface.  That
> could point at X and/or the video card/driver.
>
> Good luck....
>
>
> --
> Kevin
>
>
> _______________________________________________
> Linux-audio-user mailing list
> Linux-audio-user at lists.linuxaudio.org
> http://lists.linuxaudio.org/mailman/listinfo/linux-audio-user
>


I've tried 2 different graphics card with the same results. Haven't
tried running in console mode though. Is there a good test for systems
stability to run? burncpu seems ancient. Prime tester thingy?  This is
on a quad core if that makes any difference. I haven't noticed
anything in /var/log/messages.

Loki



More information about the Linux-audio-user mailing list