Sun-Spots-Request@RICE.EDU (William LeFebvre) (01/23/88)
SUN-SPOTS DIGEST Thursday, 21 January 1988 Volume 6 : Issue 6
Today's Topics:
Archive server source and other administrivia
Re: Sun 3 keyboards drop characters when you type too fast
Re: rows,cols (window sizing)
Re: ping bug
Strange failure of bind(2) on diskless clients (2)
UREP runs on a Sun
UREP for SUN/4 (2)
Andataco
Attempt to use dumpregion.c
Wanted: info on 3rd party SCSI disks for Sun 3/50
Problems with chained shoeboxes with SunOs 3.4
Need help with shared memory routines in 3.4
Filesystem integrity - help
Problems starting up Suntools
Software to connect UNIX and VMS email?
Adding accessable directories to /usr/lib/uucp/L.cmds?
How big a Safety Net?
Omnipage on a Sun?
Two ND servers on one cable?
SEE Editor for UNIX?
----------------------------------------------------------------------
Date: Thu, 21 Jan 88 09:45:00 CST
From: William LeFebvre <phil@Rice.edu>
Subject: Archive server source and other administrivia
Many people have asked me recently about the source for the archive server
we use for the Sun-Spots archives. The server was written by Brian Reid
for his own purposes. He is perfectly willing to give away copies of the
source to anyone who wants them, but he does not guarantee that the system
will be usable for anyone else's application. In other words, it is "as
is". I had to make several changes to adapt it for our use here. With
that caveat, I will tell you that it is available from the archive server
at "decwrl.dec.com". Simply send the request
"send systems archive-system.shar" to the address
"archive-server@decwrl.dec.com" and it will try to mail you a copy. If
anyone knows of any other archive server systems, please send me mail. I
am especially interested in one that will do keyword lookup, so that
people can use it to find out what digest issues contain messages about
given topics.
There is still a large backlog of messages left over from the Christmas
hiatus. If you haven't seen your message sppear yet, please have
patience. I have gotten as far as January 14 in the backlog. Hopefully I
will be all caught up in a week. Requests are not being processed very
fast, either. Hang in there, things can only get better!
William LeFebvre
Department of Computer Science
Rice University
<phil@Rice.edu>
------------------------------
Date: Fri, 8 Jan 88 15:27:51 EST
From: bob@allosaur.cis.ohio-state.edu (Bob Sutterfield)
Subject: Re: Sun 3 keyboards drop characters when you type too fast
mike@thumper.bellcore.com (Michael Caplinger) writes:
>
>Occasionally when I switch into furious typing mode on my Sun 3, the
>keyboard starts beeping and my keystrokes are dropped.
I didn't notice this on older 3/50s, but on our latest delivery batches
(since July or so) that have the monitor separated from the CPU on a
pedestal rather than nestled into the top of the CPU, the keyboard
beeping/keystroke dropping problem has become pretty bad. It doesn't even
take particularly furious typing to cause it.
Also, running GNU Emacs 18.49 under X10r4, the arrow keys (cursor
movement) work on old keyboards but not on new ones. This causes some of
our users great consternation.
Was there maybe a microcode change inside the keyboard that coincided with
the system packaging? Would it have caused these symptoms?
Bob Sutterfield, Department of Computer and Information Science
The Ohio State University; 2036 Neil Ave. Columbus OH USA 43210-1277
bob@ohio-state.{arpa,csnet} or ...!cbosgd!osu-cis!bob
soon: bob@cis.ohio-state.edu
------------------------------
Date: Fri, 08 Jan 88 14:43:33 PST
From: Craig Leres <leres%lbl-helios@lbl-rtsg.arpa>
Subject: Re: rows,cols (window sizing)
I've been using the following aliases to size my X10 and suntool windows
for some quite some time:
if ("$TERM" == xterm) then
a sunsize 'echo -n "[8;\!:1;\!:2t"; stty rows \!:1 cols \!:2'
endif
if ("$TERM" == sun) then
a sunsize 'echo -n "[8;\!:1;\!:2t"; stty rows \!:1 cols \!:2; echo ""'
endif
You must supply both the row and column arguments, but at least you
don't have to lug a program around.
Craig
------------------------------
Date: Mon, 11 Jan 88 07:47:17 EST
From: straw@nbs-enh.arpa (Mike Strawbridge)
Subject: Re: ping bug
We have noticed the same behavior. It seems that if more that one PING is
outstanding, that is, if more than one ICMP ECHO packet has been
transmitted without receiving the corresponding ECHO REPLY packet, then
when one ECHO REPLY does arrive, it satisfies all outstanding ECHOs and
ping erroneously reports all systems as 'alive'.
mike strawbridge
straw@nbs-enh.arpa
------------------------------
Date: Fri, 08 Jan 88 17:27:07 EST
From: jenkins@purdue.edu
Subject: Strange failure of bind(2) on diskless clients (1)
Many thanks to the numerous people who responded. Lee Camel of Sun
pinpointed the problem correctly. We use yellow pages to advertise
/etc/hosts, however at boot time, the local /etc/hosts is apparently
consulted first. I am not clear as to why the network interface showed
the correct IP address though. The solution to the problem was to add the
diskless client to its own local copy of /etc/hosts. That minor detail
was overlooked when the nd partition was created. Thanks again for
everybody's help.
Colin Jenkins
Systems Programmer
Purdue CS Dept.
jenkins@cs.purdue.edu
------------------------------
Date: Sat, 9 Jan 88 04:41:20 EST
From: dupuy@westend.columbia.edu (Alexander Dupuy)
Subject: Strange failure of bind(2) on diskless clients (2)
The fact that the symptoms only appear on diskless clients leads me to
believe that your network interfaces are not being ifconfig'd properly
(despite what netstat may say). Diskless machines must do a certain
amount of network interface configuration in order to boot (this is done
entirely within the kernel) and so it is possible to use the machine
without even having run ifconfig. However, if ifconfig hasn't been run,
the association of internet address to interface is not completely set up,
and so programs are unable to bind sockets with the host's internet
address.
I would suggest replacing these standard lines in /etc/rc.boot
/etc/ifconfig ec0 $hostname -trailers up
/etc/ifconfig ie0 $hostname -trailers up
/etc/ifconfig le0 $hostname -trailers up
with the following lines:
/etc/ifconfig ec0 $hostname -trailers up ||
/etc/ifconfig ie0 $hostname -trailers up ||
/etc/ifconfig le0 $hostname -trailers up ||
/etc/ifconfig lo0 $hostname -trailers up >/dev/console 2>&1
/etc/ifconfig lo0 localhost up >/dev/console 2>&1
which will cause ifconfig errors to be reported at boot time. In your
case, I suspect that the local host name is not in /etc/hosts, or that you
have altered the rc.boot script, and $hostname isn't being set
(correctly).
@alex
[[ It was the first problem you suggested. See previous message. --wnl ]]
------------------------------
Date: Wed, 13 Jan 88 20:27:29 +0200
From: leonid@TAURUS.BITNET
Subject: UREP runs on a Sun
Good news for those who plan to swap a VAX for a SUN: The Penn-State Unix
RSCS Emulation Program runs on SUN without additional hardware. Using a
pseudo device driver (Public Domain) one of the zs ports on CPU board is
made to talk BSC to an IBM running RSCS-V1. The user-level code originally
from PSU compiles and runs smoothly on SUN - no machine depenent bugs,
(just ordinary UREP bugs ;-)). So, if you want to have a SUN connected to
BITNET - there you go! Get the original UREP package from PSU (send e-mail
to Dae@PSUVAX1.BITNET) which is $800 for an educational site, and drop me
a line and I'll send you the device driver.
Leonid Rosenboim, System Manager, Tel-Aviv University, Math Dept, Israel.
E-Mail: leonid@Math.Tau.Ac.IL, leonid@taurus.BITNET,
eonid%taurus.BITNET@Cunyvm.cuny.edu
P.S.
TAURUS.BITNET was a VAX/780 and is now a SUN-3/180 !
------------------------------
Date: Thu, 14 Jan 1988 11:12
From: Peter W. Day <OSPWD@EMUVM1.BITNET>
Subject: UREP for SUN/4 (1)
LEONID@TAURUS reports that he has UREP running on a SUN/3 without
additional hardware using a public domain pseudo device driver and one of
the zs ports on the cpu board.
Any thoughts on what would be required to get UREP to run on a SUN/4 ?
Peter Day
Emory University
------------------------------
Date: Thu, 14 Jan 88 19:39:47 +0200
From: leonid@TAURUS.BITNET
Subject: UREP for SUN/4 (2)
Peter, Sun claims that SUN-4 is source-compatible with SUN-3, therefore I
expect that this device driver will compile and run smoothly on SUN-4 as
well. Please note that this device driver is in fact a "pseudo-device"
which uses Sun's zs_common.c code to access the Zilog-SIO chip.
Leonid
------------------------------
Date: 8 Jan 88 23:11:27 GMT
From: horizon!rob@seismo.css.gov (Rob Lewis)
Subject: Andataco
A while back, someone asked how and where to get SUN peripherals at a
reasonable cost. We have had a lot of success with Andataco, they are
technically competent and helpfull. They are also very helpfull in
getting the new stuff in first. Anyway, the address of the company is:
10054 Mesa Ridge Ct. Ste 102
San Diego, CA 92121
The contact is Dave Sykes. He is a real mad man.
His phone number is 619-453-9191
He sells Xylogics, Clearpoint, Fujitsu, CDC, and a bunch of others
that I have not had to use.
------------------------------
Date: Wed, 13 Jan 88 17:42:42 EST
From: ras@sabre.bellcore.com (Raymond A. Schnitzler)
Subject: Attempt to use dumpregion.c
I got dumpregion to compile okay, but I've had trouble printing the result
to an imagen (8/300 ?). I have a program here called idump, which can be
successfully used with "screendump | idump | ipr", but
"dumpregion | idump | ipr" doesn't work. My best guess is that idump is
assuming a fixed raster width, or something; I get reproduceable garbage
(well, almost garbage, it does have some sort of pattern).
Anyone have any ideas?
Ray Schnitzler
Bell Communication Research
uucp: ...!bellcore!schnitz
internet: schnitz@bellcore.com
[[ From personal experience I can tell you that it is much easier to build
a bitmap suitable for imPRESS if you can assume that the bitmap's sizes
(in both X and Y) are multiples of 32. Likely, someone needs to teach
"idump" how to pad out the bitmap in both directions before attempting the
conversion. --wnl ]]
------------------------------
Date: Mon, 11 Jan 88 14:42:29 est
From: tropix!den@cs.rochester.edu (David E. Nichols)
Subject: Wanted: info on 3rd party SCSI disks for Sun 3/50
I am interested in experiences (good or bad) in hooking up 3rd party SCSI
disks to Sun 3/50's. I realize that this has been a well discussed topic
but now I need to know and I don't have any of the previous articles
around.
I am particularly interested in knowing whether you bought a pre-packaged
shoebox look-alike or if you assembled pieces yourself, etc. Did you need
a network to boot from for the first time??
Any stories will be greatly apreciated. Please e-mail directly to me at
tropix!den@cs.rochester.edu.
Dave Nichols
tropix!den@cs.rochester.edu
------------------------------
Date: 8 Jan 88 08:49:11 GMT
From: Hans van Staveren <mcvax!cs.vu.nl!sater@uunet.uu.net>
Subject: Problems with chained shoeboxes with SunOs 3.4
We just upgraded to 3.4 and while most things seem to work we have one
3/50 with two shoeboxes(Adaptec ctl, 1 Fujitsu, 1 Micropolis) that does
not appear to work. In our own tailored 3.4 kernel the labels can not
even be read, and with the generic 3.4 the system crashes during fsck.
We did the chaining of the shoeboxes ourself, so we might have done
something wrong(we *do* know about SCSI targets, termination etc..), but
our 3.2 kernels had no problem whatsoever.
A clue anyone?
Hans van Staveren
Vrije Universiteit
Amsterdam, Holland
------------------------------
Date: 8 Jan 88 21:10:32 GMT
From: rfox@amelia.nas.nasa.gov (Richard Fox)
Subject: Need help with shared memory routines in 3.4
I have been trying to use the shared memory routines provided with sun os 3.4.
The routines are called: shmget shmat shmctl.
The documentation on line doesn't really explain how to use them at all,
and when trying to use the routines the errors I get back don't always
make sense. I have been told that the package is buggy and will of course
be outdated in os 4.0. However, for the interm I do need to try and get
the stuff to work.
My question:
does anybody out there have a real document on how to use the stuff, with
possibly a bug list? And, if not, does anybody have a working example
that they could share?
Thanks,
Richard Fox
rfox@ames-amelia.ARPA
(415)694-4358
------------------------------
Date: Fri Jan 8 20:29:35 1988
From: Paulo L de Geus <paulo%unix.computer-science.manchester.ac.uk@nss.cs.ucl.ac.uk>
Subject: Filesystem integrity - help
Environment: 5 3/160 diskless hooked to a 3/160 server, some of them with
imaging hardware attached. These ones have to boot a hacked
kernel in order to make use of the imaging hardware.
Problem: The porting of this hardware to the suns is new and problems do
occur. Sometimes I notice that the workstation I am running my
application gets under heavy load, with annoying delays for
anything you try to do. A few minutes more and a window frame
menu may take several minutes to show up. When I realize I've
gone that far it's very hard to try and kill any process, and in a
couple of moments the system hangs inexorably. Forget about
remote logins. What is left then is L1-a to kill unix and call
the monitor to re-boot. As expected the filesystem is corrupted
and a message is shown during the booting process saying that "the
root was damaged" and also "has been fixed".
Questions: (I don't know much about unix and suns, nor where I should look
for an answer in the sun manuals. Though I had several
workstation crashes like that in the past weeks, only one
filesystem crash was reported a few days ago, and what a bad
one! Am I the culprit?).
1- How faithfull is that message "has been fixed"?
2- Does a sync done on the server's console insure filesystem
integrity?
3- Is there a way to sync the disk "on behalf" of
the crashed system from the server's console?
Any suggestions as to how to do a safe reboot of the workstation in that
case? I hope I'm missing something easy.
Any help will be greatly appreciated.
Paulo L de Geus JANET: paulo@uk.ac.man.cs.ux
Dept of Computer Science Internet: paulo%ux.cs.man.ac.uk@nss.cs.ucl.ac.uk
Univ of Manchester USENET:...!mcvax!ukc!man.cs.ux!paulo
Manchester M13 9PL U.K.
------------------------------
Date: Wed, 13 Jan 88 15:54:35 EST
From: Bob 3-8661 <bob@steinr.pa.msu.edu>
Subject: Problems starting up Suntools
For the past few days I have developed a problem when I open suntools.
Not all the windows in my .suntools file are opened. When I rebooted the
system (3.4) the problem dissapeared for a day, but is now back. I am
running a 3/160. Can anyone help me locate the source of the problem?
Bob Stein
bob@steinr.pa.msu.edu [35.8.48.10]
stein@msupa.bitnet
steinrt@clvax1.cl.msu.edu
------------------------------
Date: 7 Jan 88 15:06:33 GMT
From: heather@oac.ucla.edu (Heather Burris)
Subject: Software to connect UNIX and VMS email?
We would like to attach a Sun workstation to a Vaxstation running VMS and
exchange electronic mail. I am aware of software packages that support
this, for example the Wollongong group produces one. I was wondering if
any of you have used any software to do this. If so, could you please
send me the name of the software package, the company that makes it,
whether it has performed to your satisfaction and whether it requires
hardware other than an ethernet connection. Other information (such as
prices, etc.) is appreciated too.
Send responses to me and I'll summarize to the net. Thanks,
Heather Burris
Advanced Workstation Consultant, UCLA
------------------------------
Date: Sun, 10 Jan 88 16:03:54 PST
From: John Bossert <bossert@thebes.thalatta.com>
Subject: Adding accessable directories to /usr/lib/uucp/L.cmds?
As a result of installing news, and wishing to keep everything in
/usr/local/bin, I need to add /usr/local/bin to the list of directories
that uuxqt can access.
Under 4.3 [BSD], I can accomplish this by placing the following in L.cmds:
PATH=/bin:/usr/bin:/usr/local/bin
rmail
rnews
lpr
uusend
Under [SunOS] 3.3, however, uuxqt spits at me, (effectively) saying it
can't find rnews (/usr/local/bin/rnews). How can I add to the search
path? I really don't want to "pollute" /usr/bin ...
In-Real-Life: John Bossert, Thalatta Corporation, (+1 206 643 7187)
Domain: bossert@Thalatta.COM Path: uw-beaver!uw-entropy!thebes!bossert
------------------------------
Date: Mon, 11 Jan 88 12:37:19 EST
From: Kevin Brown <KCB1@APLVM.BITNET>
Subject: How big a Safety Net?
For each partition on a disk drive, 10% of the space is reserved as a
"whoops buffer" (got a better name?) to keep the disk from being
completely filled. Why 10 percent? Why should this be a function of the
total size of the disk? Setting aside 14 Meg for a 140 Meg drive as
opposed to setting aside 7 Meg for a 70 Meg drive only seems to make sense
if you assume that larger disks will deal with commensurately larger files
instead of the same mix of file sizes but more of them. On my machine the
latter is true, we just have more files whose average size didn't increase
because we are using a larger disk.
Is there a better way to choose an amount of disk space to set aside? Am
I missing the point? Does anyone know the history on the choice of 10% ?
How big a disk drive was being used by the person who choose 10% ?
Kevin Brown
KCB1@APLVM.BITNET
na.kbrown@score.stanford.edu
Applied Physics Lab
Laurel, MD
[[ My understanding is that this decision was made for 4.2 BSD because the
file system algorithms (remember, they rewrote a large part of the file
system code between 4.1 and 4.2) started flogging when the disk reached
about 90% capacity. Maybe it was just a nasty rumor. --wnl ]]
------------------------------
Date: 14 Jan 88 14:43:17 GMT
From: lj@spdcc.com (Len Jacobs)
Subject: Omnipage on a Sun?
If anyone is using Omnipage on a Sun, please send me your comments on how
well it serves your purposes. I will summarize.
Thanks.
Len Jacobs
{ihnp4,harvard}!spdcc!lj
------------------------------
Date: 13 Jan 88 22:23:05 GMT
From: km@emory.UUCP (Ken Mandelberg)
Subject: Two ND servers on one cable?
We see lots of collisions on our ethernet. Netstat reports collisions as
more than 30% of output packets over extended periods on the diskless
clients. This happens even when there is very little traffic (maybe 100
packets/min per client). The clients are divided amoung two ND servers
(all one one wire) which themselves show lower collisions rates (5-10%).
This behavior shows up even with only 6 clients (3 per ND server).
Some experiments show that the collisions rate goes way down when we only
run 1 ND server. I suspect the two ND servers on one wire is the problem.
Does anyone recognize this?
Ken Mandelberg | {decvax,sun!sunatl,gatech}!emory!km UUCP
Emory University | km@emory BITNET
Dept of Math and CS | km@emory.ARPA ARPA,CSNET
Atlanta, GA 30322 | Phone: (404) 727-7963
------------------------------
Date: 12 Jan 88 16:28:42 GMT
From: boulder!seri!sst!arch@hao.ucar.edu (Arch Mott [0274200])
Subject: SEE Editor for UNIX?
Greetings-
I am trying to find a version of the SEE Editor to run on UNIX systems. I
called C-ware to ask about it and they said that they had sold copies of
the source to some folks who had said they were going to port it to UNIX.
Anyone know anything about it?
Thanks in advance.
Arch Mott
Solar Energy Research Institute
1617 Cole Blvd
Golden, Co. 80401
------------------------------
End of SUN-Spots Digest
***********************