[mythtv-users] Crashes, then fsck..

[James L. Paul]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On Tuesday 24 February 2004 8:09 am, Mark Hodges wrote:
> manu wrote:
> > Le 24.02.2004 02:57:59, Mark Hodges a écrit :
> >> I have a problem with MythTV Crashing. After a random time period of
> >> recording or watching live TV (paused or unpaused) mythtv hard-locks
> >> up. Ctrl-Alt-Del does nothing, I can't switch out to a console or
> >> kill X. After I do reboot I have to run a full fsck and rebuild the
> >> file system tree. I am using reiserfs and I have to run reiserfsfsck
> >> --rebuild-tree. I always lose the recent data after this, that
> >> includes logs and whatever was being recorded.  I have tried
> >> changing  settings, and change the hard drive that is being used for
> >> recording  (I have three).  My motherboard is an dfi nf2ultra
> >> (nforce2)  motherboard.  I have two hard drives running on the
> >> primary channel  on the second ide controller (an hpt370 ide
> >> controller). The master (hde) drive is where myth is installed, and
> >> the secondary drive(hdf)  is where I planned to store the videos.
> >> After the first lockup I  tried saving the videos to the primary
> >> drive(hde), and still got a  lock-up. The only other drive that I
> >> have is hda, and I can't save to  it because it is formatted ntfs.  I
> >> am running gentoo with a 2.6.3  kernel, and an ATI  tvwonder VE
> >> capture card.  My vid card is an  nvidia card, running the newest
> >> nvidia drivers. Any help would be  greatly appreciated, and tomorrow
> >> after the fsck finishes I will try  running the cvs version to see if
> >> that helps. But if there is any  sort of a ide issue or something
> >> simple that I am overlooking, help  would be greatly appreciated
> >
> > For NForce2 mobos you have to disable local apic (you have to pass
> > nolapic or/and noapic as kernel boot parameter). At least this was
> > necessary with 2.4.x kernels, I don't know if you still have to do
> > that  with 2.6.x.
> > Bye
> > Ma
> > nu
> >
> >------------------------------------------------------------------------
>
> Yeah, I forgot to include that I have tried passing the noacpi option to
> the kernel, in fact that is how I run things all of the time due to some
> other issues, I also use hdparm to enable dma, and have the idebus set
> at 66mhz. I have tried fiddling with all the hd settings that I know of
> and nothing gets any better.

You don't mention if you see anything in the logs to indicate an error at 
crash time, so I assume the logs show nothing interesting.

The fsck after the crash is nothing unusual. That's simply to be expected when 
a volume is not unmounted cleanly due to any systemlock during write 
operations.

Have you considered the possibility of the CPU overheating? When I have had 
symptoms of things working fine for some period of time, yet consistent 
system lockups after some period of non-idle activity, I've often found the 
problem to be the CPU overheating. Especially with my Athlons.

If you are able to ssh into the box and monitor logs and processes, do so. If 
you are able to monitor CPU temp, do so. Do you best to watch what is 
happening at the time of the crash across multiple incidents and look for a 
common factor.

I certainly don't have enough information to suggest all the possibilities, 
but your symptoms seem consistent with a SocketA CPU cooler that isn't able 
to transfer heat quickly enough away from the CPU. I suggest investigating 
that and eliminating it as a possibility.

> _______________________________________________
> mythtv-users mailing list
> mythtv-users at mythtv.org
> http://mythtv.org/cgi-bin/mailman/listinfo/mythtv-users
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.2 (GNU/Linux)

iD8DBQFAOxGvT8BYaKRUpkQRAsbRAJ9ckNV1nHSQrZuBe4lJmPW57z9VEwCgnypn
pXmAKJHnMyfVB6tOq0MOjmE=
=h46S
-----END PGP SIGNATURE-----