Subject: Re: resets under high load
To: Stephen Degler <stephen@degler.net>
From: Andreas Kahari <ak+netbsd@freeshell.org>
List: port-amd64
Date: 03/10/2004 19:36:01
On Wed, Mar 10, 2004 at 05:01:44PM +0000, Andreas Kahari wrote:
> On Wed, Mar 10, 2004 at 11:47:02AM -0500, Stephen Degler wrote:
> > 
> > On Wednesday 10 March 2004 11:30 am, Andreas Kahari wrote:
> > 
> > > Hmmm... I might have run into this one as well actually.  I
> > > first thought it was a power cut, but it was too short to be
> > > that and nothing else in the apartment registred it.  I was
> > > compiling three packages from pkgsrc and listening to a CD at
> > > the time.
> > >
> > > I don't really feel encouraged to try to provoke this to
> > > happen again... fsck reported a lot of (non-disasterous) disk
> > > inconsistencies.
> > >
> > 
> > I haven't had serious file system damage due to this, even though I have 
> > softdeps enabled.  I'm guessing that these are mostly the recently created .o 
> > files and such from the compilation.
> > 
> > If you could try to reproduce it again I'd be most greatful.  Also I'll 
> > send-pr the issue once I'm sure that its not just me who is experiencing it.
> 
> I'll hook up a serial console later tonight and see if I can
> reproduce it with a ddd-and-distribution-building loop as you
> suggested earlier.  It might be interesting to note that I have
> a K8T800-based motherboard (GA-K8VNXP).

I'm sorry, but I'm going to chicken out.  I'm currently
replacing all checked out sources on my /usr partition since
many of the files on there had nul-filled holes which prevented
me from rebuilding my kernel.

I don't think it's safe to provoke a reset like that.

If a reset does happen again, I'll certainly tell you about it.


Andreas

-- 
Andreas Kähäri                      |[][]|
                                    |][][|
East Anglia                         |[][]|
England                             |][][|