On Sat, Jan 08, 2005 at 04:39:57PM -0600, Tracy Di Marco White wrote:
> [...]
> I started seeing problems around 1.5.x, all I had was the 7000.  I
> borrowed the 8000 from work, where it just works on the DU alphas.
> IIRC the one machine ran 1.6 for a while, and now runs 2.0.
> 
> I can't get the 8000 to work on NetBSD very long.  It's LVD/SE, but it
> should be running in SE mode, and I've tried two different active
> LVD/SE terminators on it.  In recent tests it's always on the adaptec,
> but it has failed on both cards.

Did you try different cables too ?

> 
> [...]
> 
> >Also, what errors do you get with the tekram ? commands timeout too ?
> 
> Dec  3 05:08:15 afs-1 /netbsd: st0(esiop0:0:5:0):  DEFERRED ERROR, key = 0x4
> Dec  3 05:08:15 afs-1 /netbsd: st0(esiop0:0:5:0): non-media hardware failure, info = 0 (decimal), data = 00 05 a8 d3 00 00 00 00 00 00 85 03 1b 00 00 00 1f 00 13 02 30 00
> Dec  3 05:08:15 afs-1 /netbsd: st0(esiop0:0:5:0): medium error, info = 2 (decimal), data = 00 05 a8 d3 15 02 00 00 00 00 85 03 1b 00 00 00 1f 00 13 02 30 00

You should really build your kernels with "options SCSIVERBOSE" :)
If I got it right, ASC/ASQ is 0x15/0x02 which is "Positioning Error Detected
By Read of Medium"
which really looks like a hardware issue.

> I am not sure the 7000 is not broken, but I am sure the 8000
> is not broken.  The amount of time before errors on the 7000 varies
> with temperature.

Do you get the same errors with the 8000 on siop ?

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