On Mon, Jul 24, 2006 at 01:16:30PM +0300, Jukka Marin wrote:
> I'm setting up this SMP amd64 system, now running the i386 port, -current
> kernel and binaries built yesterday.
> 
> I have a large raidframe "disk" (raid1, two SATA disks).  I used tar
> to archive some 30 GB of files into a partition (no softdep) on the
> raid disk.  The source files were read over NFS.  Nothing else (but
> systat) was running on the system.
> 
> A few times during the copy, the system froze completely and then
> woke up again.  When this happens, the system does not respond to
> ping, it doesn't update the systat display or do anything visible.
> During the copy, over 40 ping packets were lost.
> 
> Once, systat compalined something about a dead timer and using another
> one (I couldn't save the message).
> 
> I have seen pauses like this on my single-CPU i386 system running 2.0
> without raidframe, so this is not a new problem.  I'll try to do more
> testing when I get back to the new machine (it's kind of slow over
> GPRS).

Is it AMD CPU or Intel CPU ?

There seems to be issues with the cycle counter on some amd CPUs,
and maybe this could cause this behavior ...

-- 
Manuel Bouyer <bouyer@antioche.eu.org>
     NetBSD: 26 ans d'experience feront toujours la difference
--