Path: utzoo!attcan!uunet!samsung!olivea!apple!julius.cs.uiuc.edu!psuvax1!psuvm!cunyvm!ndsuvm1!mtus5!tomiii
From: TOMIII@MTUS5.BITNET (Thomas Dwyer III)
Newsgroups: comp.protocols.nfs
Subject: Re: PC-NFS vs. Beame
Message-ID: <90309.175619TOMIII@MTUS5.BITNET>
Date: 5 Nov 90 22:56:19 GMT
References: <10083@uhccux.uhcc.Hawaii.Edu>
 <1990Oct31.213006.12428@naitc.naitc.com> <1430@travis.csd.harris.com>
 <3160@jaytee.East.S <1990Nov04.001340.20629@naitc.naitc.com>
Organization: Computing Technology Services, Michigan Technological Univ.
Lines: 62

>>>Most of our large customers buy right-to-copy (RTC) kits, in which
>>>we sell them a set of diskettes and a serialization kit. They are
>>>free to use this to manufacture multiple full kits, or (more
>>>commonly) to create a serialized PCNFS.SYS for each system. The
>>>downside: documentation must be purchased separately.
>>>some of what we're thinking about long-term, with no dates,
>
>That doesn't help most of us.
>
>You see, here's the classic problem:
>1) You load PC/NFS on a machine, and give it a serial number.
>2) The system crashes it's hard disk, losing the serialized copy.
>
>Now, you have to have (1) the original diskettes, or (2) knowledge of which
>copy was on that machine, or (3) you burn another copy (with the RTC kit).
>
>If you don't do this, you end up with two that have the same serial number,
>and they stop one another when they detect the "license violation".


In our situation, we have some 300+ copies of PC-NFS, which we made with the
"serialization kit".  The problem here is that our machines are in student
labs, where they boot off floppys.  The boot disks seem to mysteriously
"wander off" every once in a while.  How in the world are we supposed to
determine what serial number to use for the new disk?  If Sun insists on
making PC-NFS serialized, why can't the server handle this, based on the
PC IP address or machine name or anything?  The software on the PC should
NEVER have anything to do with serialization.  It should be handled by the
server.

Every PC on our campus boots with the same serial number.  We then obtain
the correct serial number from a yp map.  Why can't Sun integrate this idea
with the distributed PC-NFS?  Since we cannot use the toolkit until *after*
the NET START RDR command, all our machines complain for the first few
minutes of the boot process.  This is a major influence on our abandonment
of Sun.

Our evaluation of Beame's BW-NFS so far has been quite satisfying.  First of
all, Beame has NO ANNOYING serial numbers.  It boots faster, has a very nice
telnet (with multiple sessions AND user-defined keyboard maps), and supports
packet drivers out of the box.  As far as customer support goes, I must admit
that Beame leaves Sun in the dust.  I have spent many an hour on the phone
with Sun, only to hear the phrase, "...in the next release, due out sometime
in the 90's..." or "Ok, I'll file a bug report."  On the other hand, I had
a few questions about BW-NFS and left a message at Beame.  Within an hour,
Carl Beame himself called back, was able to answer all my questions, and
was quite receptive to new ideas and modifications.

By the way, my previous posting regarding Beame's failure to allow the
user to determine when a file is spooled was incorrect.  After re-reading
the chapter on printing, it appears that Beame DOES have a hotkey which
causes the file to be printed.


The question is this:  So many people are hacking PC-NFS to make it useable.
Why isn't Sun?


Thomas Dwyer III                        Email: tomiii@mtu.edu
Network Programmer                             tomiii@mtus5.BITNET
Computing Technology Services           Voice: (906) 487-2110
Michigan Technological University       Fax:   (906) 487-2787