[173] in Info-AFS_Redistribution
Re: AFS 3.1 under Risc Ultrix
daemon@ATHENA.MIT.EDU (Steve Mattson)
Mon Jun 24 22:52:35 1991
To: Liz_Hines@transarc.com
Cc: khua@husc.harvard.edu, Richard Basch <basch@MIT.EDU>,
In-Reply-To: Yo message of Wed, 12 Jun 91 14:15:26 -0400
Date: Mon, 24 Jun 91 21:29:29 -0400
From: Steve Mattson <hobbes@caen.engin.umich.edu>
Date: Wed, 12 Jun 1991 14:15:26 -0400 (EDT)
To: khua@husc.harvard.edu, Richard Basch <basch@mit.edu>
cc: heft@husc.harvard.edu
From: Liz_Hines@transarc.com
Subject: Re: AFS 3.1 under Risc Ultrix
We did some initial testing with the beta release of Ultrix 4.2. It
appeared that AFS 3.1 for system type pmax_ul4 worked on it, but we are
not willing to say that it does definitely work until we have tested it on
the Ultrix 4.2 release. We received this release last week and we are
in the process of installing it and testing AFS 3.1. We will let you
know the results of this testing.
As Richard Basch mentioned, AFS does not work on multiprocessor systems.
We do not currently have plans to make AFS work on multiprocessor systems.
If any of you do try AFS 3.1 on Ultrix 4.2 and find problems, please let
us know.
Thanks,
Liz Hines
Product Support Manager, File Systems
Transarc Corporation
One of our staff has been setting up Ultrix 4.2 on a DS5000/120 here
this week. He's come to some conclusions about AFS and the CD ROM
filesystems that I want to make sure everyone is aware of.
It seems that using kernels built both with and without the CD ROM
filesystem type, running AFS and DEC's proprietary CD format work fine.
This is the format that Bookreader uses for the Ultrix Online Help CD.
However, it also seems that reading ISO 9660 format CD's and using AFS
are incompatible. At least without recompiling AFS to use a different
filesystem type. If you are running AFS and put a 9660 CD in the drive
and mount it, you'll see the AFS root instead of the contents of the CD
at the mount point. Now since the Digital Software Store Sampler CD is in
9660 format, we're a bit loathe to just comment out the format in our
kernels.
For this reason we'd also like to see the aforementioned change made to the
pmax_ul4 AFS code to change AFS's filesystem type. Personally, I wouldn't
mind if this was the only change made and you reissued the whole thing as
pmax_ul42. But that's just because I need to build a new system tree for the
new DEC release and at the same time keep my pmax_ul4 tree around for my
existing machines. I could be dreaming here. :)
Any plans along these lines?
-Steve Mattson
University of Michigan, CAEN
