Xref: utzoo comp.os.vms:19337 comp.unix.ultrix:2125 comp.sys.dec:2075
Path: utzoo!utgpu!jarvis.csri.toronto.edu!mailrus!wuarchive!brutus.cs.uiuc.edu!rpi!batcomputer!eric
From: eric@batcomputer.tn.cornell.edu (Eric Fielding)
Newsgroups: comp.os.vms,comp.unix.ultrix,comp.sys.dec
Subject: Re: File serving among VMS/Ultrix
Message-ID: <9265@batcomputer.tn.cornell.edu>
Date: 9 Nov 89 22:59:20 GMT
Reply-To: eric@geology.tn.cornell.edu
Organization: Department of Geological Sciences, Cornell University
Lines: 36

In a recent article grr@cbmvax.UUCP (George Robbins) wrote:
>In article <4823@buengc.BU.EDU> bph@buengc.bu.edu (Blair P. Houghton) writes:
>> Is it possible to use a VMS system as a file server for Ultrix workstations?
>It is said that the VMS/Ultrix Connection (aka UCX) can (or will be able)
>do this.  Get the latest dope from your DECperson and make sure you're
>talking about release 1.2, which may or may not be available yet...

We have been running UCX on our VAXcluster to serve a DECstation 3100 for
about a month.

>Rumor has it that there are two modes available - one where you get NFS
>access to ordinary VMS filesystems and the other where VMS/UCS maintains
>a Unix disk image inside a "container file" and serves this up.

This is right.  We have been using the access to the VMS files mode so far.

>It's not clear exactly to what extent UCX supports booting Ultrix diskless
>workstations in a VMS only environment 

I don't think this is presently available.

>...or how, overall, the UCX service
>compares to the usual unix flavor.
>...Works with TCP/IP protocols over ethernet, not sure how/whether DECnet
>is involved or not.

Well, it looks to me like it is pretty much standard NFS service, using TCP/IP.
DECnet is not involved at all.  From the U*IX end, it looks like a U*IX 
NFS disk, with the one exception being the fact that the VMS files ignore
UPPER/lowercase, just like VMS does.  It is amazing that it even seems to
deal with transferring file types most of the time, though not always.

The main disadvantage I can see so far is that the NFS$SERVER process on the
VMS machine grabs a lot of physical memory (6000 pages).

			++Eric