SUMMARY: Exabyte problems with 4.1.1

From: Paul Davison (pd@dcs.qmw.ac.uk)
Date: Thu Oct 10 1991 - 04:56:02 CDT


I wrote asking for further information concerning difficulties I was
having running a SCSI exabyte on an SLC running 4.1.1, with a Wren VII
on the same bus. I received some very helpful replies. Right now I am
running the machine with the kernel patch sent to me by Forrest Cook
<cook@stout.atd.ucar.edu>. I have also recently grabbed patch 100280-02,
which purports to improve various st-driver related things, but haven't
tried it. I'm still getting the occasional error, which makes me think
that Debra Tivetsky <debra@fennel.acc.com> is more than likely right as
well.

It's worth noting that I went through an entire serious of exabyte
problem under 4.1 until we made the exabyte the first device on the
chain so Konradin Stoehr <topaz@uni-augsburg.de> also has useful advice.
I haven't looked at the exabyte summaries he mentions.

Thanks to everyone that replied. Salient details follow.

----
From: "Dick St.Peters" <stpeters@dawn.crd.ge.com>

> I remember being told a while ago that there were problems under 4.1.1 > when a synchronous SCSI device was on the same bus as an exabyte, and > was wondering if this was the problem.

The Exabyte 8200 (2.3 GB model) runs just fine on a sync SCSI bus; I have one doing backups every night on a 4.1.1 (no patches) SS2 with three 660MB SCSI disks running sync.

I've heard reports of problems with the Exabyte 8500 (5 GB model) on sync SCSI. ---- From: John Sin <jsin@janet.ucla.edu>

I had a same problem with my SLC with 4.1.1, but I knew it wasn't the OS problem since it had been running fine for few months. I started getting those messages once in a while, and then more frequently until it just wouldn't write to tape anymore.

We paid money to have it serviced - it turned out to be a hardware problem. If you find out otherwise, please summarize.

---- From: Debra Tivetsky <debra@fennel.acc.com>

About a year ago we put an Exabyte tape drive on an SLC running SunOS4.1.1 and attempted to remotely dump several systems to it at night. We got many SCSI bus resets during lengthy dumps. Sun acknowledged at the time that the SLC's have a SCSI hardware problem and were working on the fix. The workaround, believe it or not, was to remove the terminator from the end of the bus!

I recently called Sun to check on the status of the hardware fix and it is out. However, I don't have any identifying revision number to pass on to you.

Your problem sounds as if it may be related to this. Anyway, hope this is of help to you.

Debra Tivetsky Advanced Computer Communications email: debra@acc.com

---- From: Konradin Stoehr <topaz@uni-augsburg.de>

Hi -

there is a collection of exabyte-summaries at ftphost.Uni-Augsburg.de in the directory pub/cronos.

All I can add is that you should make sure, that your ExaByte is the first in the chain.

After we had done this, it's running now *local* without any problems.

---- From: northrup@cps.msu.edu

Sorry Dick, but there is a problem with a subset of the 8200's and 4.1.1. Paul's message is the 3rd about this topic on sun-managers. One of our three 8200's has this same exact problem. Another one works perfectly fine. The third has a different problem.

I have a call into Sun on this problem. Neither Sun, Exabyte, nor the suggestions I've seen on sun-managers have solved the problem.

--Frank Northrup Computer Science Department Michigan State University northrup@cps.msu.edu

---- From: Forrest Cook <cook@stout.atd.ucar.edu>

Greetings Paul,

I had a similar problem running an Exabyte 8500 on an SLC with no disks. The quick solution was to disable sync-scsi. If your disk drive uses sync scsi, you will probably have some loss of disk performance. Here is a quick fix that can be applied to your kernel to disable sync-scsi:

> # adb -k /vmunix /dev/mem > scsi_options/X > _scsi_options: 78 >> $W > scsi_options?W 58 > scsi_options/W 58 > $q

A better long-term solution would be for exabyte to fix the problem in the drive. I was told that the Exabyte responds positive when asked whether it supports sync-scsi, then fails to do sync-scsi. Also, Sun is supposed to be coming out with a new exabyte driver that supports the faster file skip operations on the 8500 drives. I heard the driver should be out sometime early in 1992.

-- Paul Davison (postmaster for dcs.qmw.ac.uk)

UUCP: pd@qmw-dcs.uucp | Computer Science Dept | Mail 'stormcock-request Internet: pd@dcs.qmw.ac.uk | QMW University of London | @dcs.qmw.ac.uk' for JANET: pd@uk.ac.qmw.dcs | Mile End Road | info on the Roy Voice: +44 71 975 5250 | London E1 4NS | Harper mailing list!



This archive was generated by hypermail 2.1.2 : Fri Sep 28 2001 - 23:06:18 CDT