Subject: Re: 2.0_RC4 and -current instability, data corruption and system hang ups
To: None <port-cobalt@NetBSD.org>
From: Matt Dainty <matt@bodgit-n-scarper.com>
List: port-cobalt
Date: 11/05/2004 09:03:53
* Lists - Kush-T <lists@kush-t.co.uk> [2004-11-05 01:23:22]:
> On Thu, 2004-11-04 at 20:21, Markus W Kilbinger wrote:
> > >>>>> "Matt" == Matt Dainty <matt@bodgit-n-scarper.com> writes:
> > 
> >     >> - Sudden hangups under heavier load, mostly under continuous
> >     >>   i/o traffic (ata and/or network, beyond 100 mb data volume).
> >     >>   There is no panic message, no way to get into ddb (-> manual
> >     >>   reset necessary).
> > 
> >     Matt> I've seen this under 1.6.x too, my Qube2 runs as an NFS
> >     Matt> server and too much continuous traffic tickles this every so
> >     Matt> often.
> > 
> > Ah, then it seems to be a quite long standing problem...
> > .. and hopefully I don't have to blame my hardware!
> > 
> > Thanks for your comment!
> > 
> > Markus.
> > 
> 
> Could be Galileo driver?
> 
> http://forums.gentoo.org/viewtopic.php?p=824486#824486

Possibly. I've actually got two Qube2's, the other one I've put Gentoo
on. I know it's not really a direct comparison, but with Linux on it, it
stays up for much longer, and I'm fairly sure I get measurably faster
NFS performance out of it too.

I think it crashed once initially in the same sort of high network/disk
I/O situation, but it's been updated a few times since, kernel included,
and I haven't had it crash since. Certainly the traffic I throw at it I
know would crash the NetBSD-running Qube.

I've also tried using a 3Com network card in the PCI slot, and as
another poster has pointed out (and myself in an earlier thread) the
performance is abysmal.

Matt
-- 
"I've seen his style before, terminal psychotic!"