On Sat, Apr 17, 2004 at 07:29:28AM -0700, Wolfgang S. Rupprecht wrote:
> 
> kardel@acm.org (Frank Kardel) writes:
> > a simple pkgsrc compile triggers a reset. Disabling X didn't improve things.
> > It "feels" like things are deteriorating.
> > 
... mumble ...
> 
> Same here.  Asus K8V-SE-DELUXE (slightly different brd from yours)
> w. athlon64 3200+.  The system would reset under high disk-io such as
> during the install phase of a massive build.  The last time it reset I
Yep - installs are very likely to trigger it.
The sys is also very touchy when it is rebuilding raids and doing compiles.

> had just checked /etc into my cvs.  Since both /etc and cvs had large
> numbers of files in buffer caches at the time, it pretty much wiped
> out all my setup files and their online cvs backups.  This was
> Apr. 9th using the previous day's -current.  The fact that it would
> reset without dropping into DDB or at least providing a backtrace was
> a big problem.
It still is - for me at least.
> 
> > Power-, CPU- oder memory-trouble ?
> 
> It didn't feel like a hardware memory problem.  Memory problems, in my
> experience, normally cause builds to die with non-repeatable SEGV's.
Yepp - that's usually it. Another one is uncompilable source because
a bit flip made it to the files during a cvs update... 8-(.

-- 
			Dr.-Ing. Frank Kardel
			   <kardel@acm.org>