Sun-Spots-Request@RICE.EDU (William LeFebvre) (09/23/87)
SUN-SPOTS DIGEST Tuesday, 22 September 1987 Volume 5 : Issue 44
Today's Topics:
Moderator's note: bug reports again
bug reports
/etc/rc fix (and fsck -p)
Re: mathematical WYSIWYG text processor for Sun 3
Sun 60
Re: Sun as a Personal Computer at Home
Getting info about Ciprico products
Quick VAXMail handling
Sun-3 console problem
Re: tar tape troubles
vmstat parameter "de"?
Killed processes in 3.4?
tn3270 help
Yacc grammar for FORTRAN 77?
Non-Sun serial interface?
VME / multibus adaptor driver experience?
advice sought on interesting(?) Sun project
----------------------------------------------------------------------
Date: Tue, 22 Sep 87 14:23:18 CDT
From: William LeFebvre <phil@Rice.edu>
Subject: Moderator's note: bug reports again
This message is a followup to the one in the last issue (obviously).
Someone suggested that posting a plea for help to Sun-Spots concurrently
with contacting Sun support is a very reasonable thing to do. I agree.
But in all fairness to Sun, I would like to encourage everyone who does
this to say, at the very least, "I have contacted Sun about this, but I
also want to see if anyone on the list has seen this before...." If you
feel that Sun has mistreated you, then by all means say so. A long, drawn
out, explanation is not necessary. Just a brief one-line summary will
usually suffice. But the point is, if you are on Sun support then you
should most definitely contact Sun about your problem (whether it is in
addition to or instead of posting to Sun-Spots). Additionally, please
hold the griping down to a minimum. If you haven't talked to Sun about
the problem, then you probably don't have much to gripe about!
Those sites that are not on a software service contract are in a very
different position in Sun's eyes. I am trying to get the "official" word
from Sun about what such customers are to do when they find a genuine bug
in Sun's software. Naturally, Sun-Spots can help these people out a great
deal and that is one reason this list exists. But when a customer who is
not on a service contract finds a bug, how can they report it to Sun? Can
they expect Sun to fix it for them? These are questions I hope to have
answered later this week.
I am not suggesting that anyone curtail the use of Sun-Spots to get bugs
fixed. This forum has been most effective in that capacity and I feel
that it can and should continue that way. I guess all I'm saying is that
everyone should make sure that Sun is in the "loop" of bug reporting.
In the last issue, I stated: "I was recently contacted by a member of the
Sun technical staff regarding, specifically, the report about problems
with the GKS support. No one had ever contacted Sun about this problem."
This statement was in error and I apologize for this error. I only heard
one side of the story. It turns out that Greg Earle (the one who posted
the report about GKS) had contacted Sun about the problem and had not
gotten any suitable answers. Ironically, the people at Sun that were able
to help him contacted him after seeing his message in Sun-Spots!
William LeFebvre
Department of Computer Science
Rice University
<phil@Rice.edu>
------------------------------
Date: Sat, 19 Sep 87 21:55:50 EDT
From: hedrick@topaz.rutgers.edu (Charles Hedrick)
Subject: bug reports
The traditional software support contract used by universities for Sun
software has recently been replaced by a contract for the same price, but
which offers no service other than updates. To get actual support is
sufficiently more expensive that we, and I suspect many other
universities, do not take that contract. It is certainly possible that
people who choose to do support themselves would want to co-operate with
others who are in the same position by sharing bug fixes. Such people
obviously can't wait for Sun to fix the problem for them before reporting
it in this group. Of course we should still report the bug to Sun, so
that they can fix it in the next release. Also, we generally like bug
reports to vendors to be sufficiently precise to allow them to fix the
problem. If a problem is happening intermittently or under circumstances
that are hard to characterize, it is often useful to see whether other
people know more about it, before attempting to write an SPR.
------------------------------
Date: Thu 17 Sep EDT 1987 18:40
From: geoff@utstat.toronto.edu
Subject: /etc/rc fix (and fsck -p)
Oops, our "fsck -p" just encountered a root file system it couldn't (or
more accurately, wouldn't) fix, and this brought to light a bug in my
/etc/rc* files. I had forgotten that "exit 1" is equivalent to "false" in
both of these shell commands:
(cmds1; exit 1); cmds2
if cmds1; then exit 1; fi >/dev/console; cmds2
A corrected version of the /etc/rc* files (using "exec" for i/o
redirection) is now in the archive server, replacing the incorrect
original version.
And now on to fsck -p: Has anyone yet modified fsck such that fsck can be
told `fix things up as per -y, except print messages in the style of -p'?
It's not important that fsck in this mode should run parallel fsck's;
Chris Torek's preen command does a better job of that anyway. Sun's fsck
-p seems to be incredibly wimpy and won't even fix incorrect link counts
(at least on our ND roots).
[[ You really don't want to do this. Although I agree that the set of
failures that "-p" should fix could be larger, I don't think that it
should be all-inclusive (which is essentially what you are suggesting).
There are times when I have run fsck interactively that I have said "n".
I can't remember specifics, unfortunately, but I know that I have done
it. --wnl ]]
Geoff Collyer utzoo!utstat!geoff, utstat.toronto.{edu,cdn}!geoff
------------------------------
Date: Thu, 17 Sep 87 17:47:07 +0100
From: whbr%cgcha.UUCP%cernvax.bitnet@jade.berkeley.edu (Hellmuth Broda)
Subject: Re: mathematical WYSIWYG text processor for Sun 3
> I am interested in a WYSIWYG text processor, being able to handle
> mathematical formulas, and to create special characters.
If you really want to do Math formulas---consider TeX/LaTeX with a
preview. You can edit your text in one window and `preview' it in a
different window on the Sun workstation.
You can buy the Sun-Implementation from Arbortext (asa Textset, Ann Arbor,
MI) or get the Tapes from Stanford and do it yourself (;-).
For more information look into the desktop publishing posters in this net.
Also, comp.text will give you info on this.
There is also a TeXhax newsletter which you can subscribe to.
[[ Do you mean TUGBoat? That's the "TeX User's Group" Newsletter and is
only related to TeXhax in that many people subscribe to both. --wnl ]]
Hellmuth Broda, CIBA-GEIGY, Scientific Computing Center, R-1032.444, P.O.Box
CH-4002 Basel, Switzerland (+4161) 37-7109
whbr@cgcha.UUCP
mcvax!cernvax!cgcha!whbr
------------------------------
Date: Thu, 17 Sep 87 13:29:05 PDT
From: acad!robert@sun.com (Robert Wenig ext 609)
Subject: Sun 60
We received a few Sun 60's with the High Resolution Monochrome monitor
(1600 by 1280) and the EEPROM was set wrong. Here are the correct settings
for those who are interested. With the EEPROM set wrong, you will get
a 20 character screen when not in suntools.
Get the monitor prompt first. (L1A does it fast, but it's brutal)
> q50 <cr>
EEPROM 050: 00? 78 <cr>
EEPROM 051: 00? . <cr>
> q51 <cr>
EEPROM 051: 00? 30 <cr>
EEPROM 052: 00? . <cr>
>k2 <cr>
Basically you need to change EEPROM locations 50 and 51 to 78 and 30.
Make sure you reset the EEPROM with a k2.
Robert
------------------------------
Date: Fri, 18 Sep 87 09:05:59 PDT
From: weiser.pa@xerox.com
Subject: Re: Sun as a Personal Computer at Home
I have done this for the past couple of years. Some opinions:
The ideal Sun to take home varies with new models. For a time it was the
75, then the 140 (or 110), and now the 60. The very smallest Suns have
always not been sufficiently expandable in memory, which is why I do not
recommend the 50's. 4M just does not cut it for an actively used Sun
these days--8M minimum. 70MB disk is too small. Depending on what you
do, you will want at least 141MB disk, and possibly two 141's. You will
want at least a 32MB swap partition, for instance, unless you want to
constantly be bumping up against the VM limits, closing down temporarily
idle windows, etc. Add to that a 7M root, and 30M for system files...
I have never had Sun maintenace. It isn't necessary for small machines:
they don't break much, and when they do the board swap is relatively easy.
However, I recommend you keep the machine at work for the 90 warranty
period. Sun can easily get to the machine then during working hours, and
if you are going to have a failure it will be probably be during that 90
days.
I lived without a tape drive at home, and then got one. You can live
without it---I missed it when I didn't have it, but then when I did the
main use was for archiving. What I would find it hard to live without is
SLIP, the serial line IP that makes my home sun appear to be on the
ethernet over ordinary phone lines and modems (a 9600 baud modem helps a
lot). I would rcp new vmunix's down this line, when necessary.
Hope this helps.
-mark
------------------------------
Date: 18 Sep 87 17:02:55 GMT
From: mee@ciprico.mn.org (Mark E. Erickson)
Subject: Getting info about Ciprico products
Ciprico
2955 Xenium Lane
Plymouth, MN 55441
(612)559-2034
For information on any controllers please contact Customer Engineering.
------------------------------
Date: 18 Sep 87 14:11 EST
From: blackje%sungod.tcpip@ge-crd.arpa
Subject: Quick VAXMail handling
I began to get annoyed at the local mail system for taking too long to
deliver mail forwarded from my VAX/VMS account to my Sun. So I hacked
together a quicky [DCL] command file to process the mail on the VAX and
ship it to me quickly at my Sun. Think of it as stopping at the post
office, grabbing all the mail, stuffing it in a box and shipping it via
UPS to where it really belongs. It's not pretty, but it works most of the
time. And it's certainly saved me a LOT of time.
--Emmett
BlackJE@GE-CRD.ARPA
...!steinmetz!crd!blackje
[[ The VMS com file is in the archives under the name "sun-source/ups.com".
It can be retrieved via anonymous FTP from the host "titan.rice.edu" or
via the archive server. For more information about the archive server,
send a mail message containing the word "help" to the address
"archive-server@rice.edu". --wnl ]]
------------------------------
Date: Fri, 18 Sep 87 11:20:13 mdt
From: ed arnold <era@scdsw1.ucar.edu>
Subject: Sun-3 console problem
We have found that our Sun-3/280s crash when the [ascii terminal] console
is either powered off, or its rs232 plug disconnected. Apparently, the
console tty interface is biased such that received data goes from MARK to
SPACE (at least for awhile) when one of these events occurs, making the
system think that BREAK has been hit, ergo the crash.
In our case, this is more than a nuisance, since we want to connect
several 3/280s to a single console through an ABCD switch. However, we
can't do this until we find a way to prevent the system from crashing when
the switch is used.
Is there anyone out there who has found a method to get around this,
either hardware (hopefully simple) or software? (Sun hasn't been any help
with this matter.) If so, please mail me directly, since I don't
regularly read sun-spots digest. Thanks in advance for your advice ...
Ed Arnold * NCAR (Nat'l Center for Atmospheric Research)
PO Box 3000 * Boulder, CO 80307-3000 * 303-497-1253
era@scdsw1.{arpa,ucar.edu} * era@scdpyr.uucp
era%scdsw1.arpa@wiscvm.bitnet * era@ncar.csnet
------------------------------
Date: Thu, 17 Sep 87 10:00:42 EDT
From: Bernard Silver <bs30%gte-labs.csnet@relay.cs.net>
Subject: Re: tar tape troubles
Thanks for the suggestion but it doesn't work. Nothing seems to be able
to induce the tape drive to skip forward. It reports it does, but always
returns to the begining of the tape (I did use /dev/nrst0, do that wasnt
the problem.) It's more like I'm going to have to find a tape drive
expert...
Bernard
------------------------------
Date: 18 Sep 1987 00:30-EST
From: bassen@ifi.uio.no (Tor Sverre Lande)
Subject: vmstat parameter "de"?
Could anybody explain to be what the "de" parameter in vmstat means? The
manual quotes something like "anticipated short term memory shortfall".
We have 7 Sun-3/280 installed with 8Mbyte memory and aprox. 10 Sun-3/50
clients on each. What makes me curious is the difference between my seven
servers. Five of them have de=0 constantly independent of load. One server
has de=40 constantly and the last one de=72 constantly. Does this
parameter really mean something? If so how should the values above be
interpreted?
Tor Sverre Lande
Institute of Informatics
University of Oslo
Norway
------------------------------
Date: 16 Sep 87 15:02:33 GMT
From: ihnp4!uokmax!rob@ucbvax.berkeley.edu
Subject: Killed processes in 3.4?
We have 8 diskless Sun 2/120's and a Sun 2/170 fileserver. After upgrading
to 3.4 a short time ago, we began having problems on 4 of our
workstations. On boot, some processes die with the following message:
Sorry, pid ### was killed due to swap problems in getxfile: i/o error mapping
pages.
This occasionally happens to other processes (after boot time) and always
to the ifconfig command when I try to check the Ethernet status. The
kernel used on all of our workstations is the same, but the problem
appears only on some of them. All have 3 meg of memory. Any suggestions
on solving this would be greatly appreciated.
Thanks,
Robert Shull
University of Oklahoma
rob@uokmax
Robert K. Shull
University of Oklahoma, Engineering Computer Network
ihnp4!occrsh!uokmax!rob
CIS 73765,1254 Delphi RKSHULL
[[ I seem to recall that this is caused by stripping an executable with
the wrong version of "strip". Specifically, running "strip" from a
version 2 machine (a Sun running version 2 of the OS) on a version 3
executable. Unless you have the source or the object files, I'm not sure
what you can do about it. --wnl ]]
------------------------------
Date: Thu, 17 Sep 87 21:42 EST
From: John H. Yates <YATES@a.chem.upenn.edu>
Subject: tn3270 help
I am posting this to aid another user, please respond to him directly at
shelley%sun0.chem.upenn.edu@cis.upenn.edu
He has run the makefile for the public domain tn3270, but when it is run
it complains that the keboard map file map3270.dat is missing. Is there a
template file under another name on the distribution or is a system
terminal definition file (as /etc/termcap) to be used? They want to be
able to have their sun 3s talk intelligently to an IBM 3090, which is what
tn3270 is supposed to do.
Thanks, John H. Yates
------------------------------
Date: 18 Sep 87 13:31 EST
From: JARMOLOWSKI%VAXSCS.decnet@ge-crd.arpa
Subject: Yacc grammar for FORTRAN 77?
Does anyone know of a YACCable grammer for FORTRAN 77. Please send
responses directly to me.
INTERNET address: Jarmolowski%vaxts.decnet@ge-crd.arpa
Postal Address: Tom Jarmolowski
RCA-ESD
Borton Landing Road
Moorestown NJ 08057
Mail-stop 108-204
Phone: (609) 722-4298
------------------------------
Date: Mon, 21 Sep 87 18:19:55 -0100
From: Richard Tobin <richard%aiva.edinburgh.ac.uk@nss.cs.ucl.ac.uk>
Subject: Non-Sun serial interface?
Mike Irving from Oxford asked me to post this:
Does anyone have experience with third-party multi-terminal serial
interfaces for Sun 3s?
Information we would like is:
1) Price
2) Availability
3) Number of lines
4) Does it need/include a driver other than the standard sun one?
5) Anything else we should know.
Richard Tobin, JANET: R.Tobin@uk.ac.ed
AI Applications Institute, ARPA: R.Tobin%uk.ac.ed@nss.cs.ucl.ac.uk
Edinburgh University. UUCP: ...!ukc!ed.ac.uk!R.Tobin
------------------------------
Date: Mon, 21 Sep 87 21:16:45 CDT
From: Simon John Gibbs <simon@mcc.com>
Subject: VME / multibus adaptor driver experience?
Has anyone had any experience porting drivers for Sun-2 Multibus boards to
run on a Sun-3 using a VME / Multibus adaptor board?
I'm working with a Votan VSP-1010 - it does audio AD/DA - and have the
Sun-2 driver. I've been hoping it's easy to modify the driver but haven't
attempted anything like this before. If anyone can cast any light on this
I'd sure appreciate hearing from them.
Simon Gibbs
uucp: ut-sally!im4u!milano!simon
Internet: simon@mcc.COM
------------------------------
Date: 20 Sep 87 04:33:51 GMT
From: bunce@unccvax.UUCP (Timothy Bunce)
Subject: advice sought on interesting(?) Sun project
Hi. This is my first posting so please bear with me.
I am a final year student from Kingston Polytechnic, England. I'm
currently on a short exchange visit to the University of North Carolina at
Charlotte. It is this visit that has given me my first access to a network
such as this. I am leaving Charlotte to return to England (via Washington
DC and New York) on September the 30th so I have little time to establish
a useful dialogue, but I am forever hopeful.
In the final year of my degree course (BSc (honours) Computer Science) I
have to complete a 'significant' software project. The Poly has recently
aquired a network of 10 Sun 3/50 workstations and a large fileserver. When
I caught sight of this set-up my eyes popped and my head started spinning
with all sorts of ideas for projects, but there is one thats been on my
mind for quite a while. Many of you may be familier with the Atari vector-
graphics arcade game 'Star-Wars'. Well, to put it simply, I want to create
a system capable of 'real-time' multi-user graphics of comparable quality.
At this early stage in my thinking I envisage the system as comprising of
a number of layers. ie:
Application (games/simulation/CAD/etc) Layer
|
Communication (Sync and Motion control) Layer
|
Object Attribute Database Layer
|
3D Transform Engine (Sun Core?) Layer
|
Display Device Interface (PixRect?)
Developing the system as a set of layers would have several advantages
(apart from simplicity). Firstly, if I were unable to complete the
'Application' layer, for example, I would still have a demonstrable
multi-user graphics system on which to be graded. Secondly, as the name
implies, I hope in future to develop several different 'Applications' on
top of the lower levels.
I have much to do and much to learn to realise this project (which I'm
planning to write in Objective or just plain C). I am posting this message
in the hope of tapping into a much larger source of wisdom than I have
access to in England. If you any experience or words of wisdom on any of
the areas covered by this project (especially getting good bandwidth out
of the Ethernet and Sun Core) please get in touch.
Thank you in anticipation -- Tim Bunce.
An Englishman abroad. New to nets (any nets!). Temporarily at UNC-Charlotte
USENET: bunce@unccvax.UUCP should work till September 30th (returning to UK)
TPhone: {USofA} (704)549-1463 or (704)549-1517. {UK(London)} 01-541-1306
Snail: Tim Bunce, 155 Villiers Road, Kingston-upon-Thames, Surrey, England.
------------------------------
End of SUN-Spots Digest
***********************