[mythtvnz] Pointers on how to track down machine lockup

[Barry Clearwater]

On 12 October 2010 11:20, Wade Maxfield <mythtvnz at hotblack.co.nz> wrote:

> On 12/10/2010, at 10:26 AM, Garrett Stoupe wrote:
>
> > try running memtest leave running over night
> >
> > if it locks up then it is most likely a faulty mother board, if there
> > are memore errors (Usually happens with in the first loop) then you
> > will have faulty memory.
> >
> > if this passes with out failing fidn out the brand of hard drive you
> > have and run there diagnostics
> >
> > if itis a maxtor or seagate run seatools
> >
> > hirens has all the software you need to test it and you can download
> > it from http://www.hiren.info/pages/bootcd
> >
> > I am a hardware engineer if you have any issues contact me off list
> >
> > Regards
> >
> > Garrett
>
> Thanks Garrett,
>
> memtest86+ had been run prior to updating BIOS.  No issues.
>
> Boot drive is a Barracuda 7200.12 500GB, recording space is a pair of
> Barracuda 7200.12 1TB. They were all thoroughly tested using a certify
> process that ran over 12 hours for the 500GB, and 19 hours for the 1TBs. I
> use this process to weed out drives before taking the time to install them
> in a machine.
> SMART shows no errors.
>
> It took a while to figure out how to download Hiren's Boot CD since there
> are no download links on that page. 8)  It's superslow, but will try
> seatools once I've finished downloading.
>
>  - Wade
>

 You say smart shows no errors, did it say how hot the drives had got? The
gui disk managment utility gives good smart tools reports, you can see the
drive temp. but just in passing, I note you have 3 hard drives. We have a
server that did random lock ups and I replaced every single item of hardware
within (January this year), untill the whole thing was a new system apart
from ubuntu 8.04.4 OS.
I ended up modding the case slightly to put a big 120mm fan right in front
of the bank of 4 hard drives to keep them cool. Problem solved. It never
skipped a beat since then. I'd spent hours before then, sweated over
restoring databases that had corrupted in the reboots/lockups or segfaults
(I can't really remember which errors) and a quick replacement of the
MB/CPU/RAM confirmed it wasn't them. (I kept the new dual core setup, only
cost me about $380 all up) - Had already put a big new PS in.

Lesson learned: Extraction of the heat was not enough, CPU core temp did not
tell the whole story. Cool the drives.

Barry Clearwater


Email: barryc at bcsystems.co.nz
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.ourshack.com/pipermail/mythtvnz/attachments/20101013/d42b896f/attachment.htm