On Thursday 22 January 2004 10:08, Marc Haisenko wrote:
> On Thursday 22 January 2004 16:52, Jani Averbach wrote:
> > On 2004-01-22 08:14-0500, Szakmeister, John wrote:
> > > ... If you're truly seeing this behavior, this means one of 2
> > > things: "dump" is broken, or "verify" is broken.
> >
> > You forget the third option: hardware (memory) error which is
> > triggered by intensive disk I/O and memory I/O load.
You're absolutely right Jani. That is a very good thing to check.
> > http://www.memtest86.com/
> > http://people.redhat.com/dledford/memtest.html
> >
> > BR, Jani
>
> Yes, we should check that one, though it's pretty unlikely since the server
> is brand-new and the error always happens in the same revision. But I'll
> run a test this weekend nevertheless, you never know (it's our new
> development server so I can't just take it down and run memtest86).
FWIW, I've bought a brand new server before and found that it had a bad memory
chip (1-bit was stuck to at '1'). After that experience, the first thing I
do is run memtest. :-)
-John
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Received on Fri Jan 23 00:37:10 2004