I found the following message on
http://mail-index.netbsd.org/port-pmax/1997/03/10/0001.html
Since I'm having exactly the same problem (OpenBSD 2.6, DECstation 2100,
24 MB RAM, two disks, freezes under load, no error messages). I'm curios
if this problem has already been identified or maybe even fixed with
either NetBSD or OpenBSD.

Subject: Re: DS3100 kernel clashes -- Re: updated snapshot on
ftp.netbsd.org
To: Michael L. Hitch <mhitch@lightning.oscs.montana.edu>
From: Thomas Graichen <graichen@rzpd.de>
List: port-pmax
Date: 03/10/1997 09:18:30=20
hasn't Michael L. Hitch said ? ...
> On Mar 10, 10:01am, Toru Nishimura wrote:
> > Just to report my case;
> >=20
> > Although I don't have any PMAX other than several MAXINE (DS5000/25),
> > I have never experienced kernel troubles with stock NetBSD/pmax
> > (formal releases and snapshots) since NetBSD/1.1.
> >=20
> > > Now I'm having another problem: Though it works generally, for some
> > > reason or another, my DECstation 3100 (KN01, Version 7.0, with two RZ=
55
> > > disks, CFB, 24M), crashes when it's doing something - anything,
> > > randomly.  It just sits there, frozen.  No panic message, no savecore
> > > dump.  I haven't been able to compile a kernel since it would crash i=
n
> > > the middle somewhere and I'd have to restart compiling from the
> > > beginning (is this normal???  Do I have hardware trouble?)
>=20
>   The 3100 has a complete different SCSI interface than the 5000 machines=
.
> It's quite possible that the SCSI driver for the 3100 may still have a
> bug or two.  It's been a while since I've worked on the 3100 SCSI driver,
> and I haven't used the 3100 much since then, so I'm not certain how well
> it works.  I've been using a 5000/25 for most of my development work
> (and hope to be able to move it to a 5000/50 before too long).
>=20
>   If I get a chance, I can move my two RZ56 drives over to the 3100 and
> run with it a while to see if the driver still works.
>=20
i had a very similar problem running OpenBSD/pmax on a 2100 (i first
thought that it was a OpenBSD specific problem - but now i have seen
several reports of such behaviour in this list so i think it's a
general *BSD/pmax problem - and OpenBSD/pmax has no real difference
from the hardware part to NetBSD/pmax) - the system runs for weeks if
idle - but if i put some disk activity onto it it hangs after a while
(from half an hour to more than a day) - without any panic messages or
something like - some time ago i switched from using both disks

  rz0 at sii0 drive 0 slave 0 FUJITSU M2624S-512 rev 0202,
    1015812 512 byte blocks
  rz3 at sii0 drive 3 slave 0 SEAGATE ST1480 rev 5830,
    832527 512 byte blocks

to only using the second one (the dec disk) and since then it runs
just fine - may be it's broken hardware - or the driver has problems
with multiple disks or with non dec disks - i just say this here
because maybe it can be useful for finding out the source of the
problem

hope that helps

t

--=20
Georg Schwarz     schwarz@physik.tu-berlin.de, kuroi@cs.tu-berlin.de
Institut f=FCr Theoretische Physik       +49 30 314-24254, FAX -21130
Technische Universit=E4t Berlin        http://home.pages.de/~schwarz/