Sun-Spots-Request@RICE.EDU (William LeFebvre) (08/29/88)
SUN-SPOTS DIGEST Sunday, 28 August 1988 Volume 6 : Issue 200
Today's Topics:
Re: Other network mailing lists and marginal topics
Re: SLIP on SunOS4.0
"open systems"
automounting ... why bother
Proxy ARP daemon source
some curious performance problems
IconEditor problem / At problem
4.0 spline labels for 3" binders
PC-NFS - request for your comments
Questions about 386i communications
Why won't "traffic" work?
Questions about windowing systems: NeWS?
Send contributions to: sun-spots@rice.edu
Send subscription add/delete requests to: sun-spots-request@rice.edu
Bitnet readers can subscribe directly with the CMS command:
TELL LISTSERV AT RICE SUBSCRIBE SUNSPOTS My Full Name
Recent backissues are available via anonymous FTP from "titan.rice.edu".
For volume X, issue Y, "get sun-spots/vXnY". They are also accessible
through the archive server: mail the request "send sun-spots vXnY" to
"archive-server@rice.edu" or mail the word "help" to the same address
for more information.
----------------------------------------------------------------------
Date: Sun, 28 Aug 88 17:00:35 CDT
From: William LeFebvre <phil@Rice.edu>
Subject: Re: Other network mailing lists and marginal topics
I have just been informed that "Info-Postscript", one of the lists I
mentioned in v6n196, no longer exists. There is a Usenet newsgroup
dedicated to the discussion of Postscript ("comp.lang.postscript") but no
corresponding Internet mailing list.
William LeFebvre
------------------------------
Date: Tue, 23 Aug 88 08:57:30 PDT
From: Brent Chapman <capmkt!brent@cogsci.berkeley.edu>
Subject: Re: SLIP on SunOS4.0
Reference: v6n194
In v6n194, Janine Roeth (jar@jessica.stanford.edu) asks:
> Has anyone installed SLIP code on a SunOS 4.0 system? Was it the archived
> slip code on titan.rice.edu under slip.shar?
Sorry, but it won't work, and can't be fixed. In SunOS 4.0, Sun discarded
"line disciplines" as the basis of their TTY drivers, in favor of SysV
"STREAMS" stuff. I'm not saying this was a good or bad thing to do, but
it definitely breaks existing versions of SLIP (including the one in the
Sun-Spots archive), which are written to function as line disciplines.
SLIP will have to be re-implemented using STREAMS before it can be used
with SunOS 4.0.
Now, a few months back, when 4.0 first started appearing, several people
said they would be working on a STREAMS-based SLIP, but I haven't heard
anything since.
An interesting rumor that I've been hearing (I have no idea whether or not
it is true) is that Sun's networking group has SLIP on 4.0 up and running,
and will make it a part of 4.1. I sure hope this is true; lack of SLIP is
one of the two major reasons we haven't upgraded to 4.0 yet (the other is
the incompatibility of .o and .a files from 3.x), and I would like nothing
more than for it to become an official, supported part of SunOS.
Apparently, the Problem is that Sun already sells something similar
(called the Internetwork Router, or IR for short), and the marketroids
might not want to start "giving away" something that is almost as good as
IR, or slightly better, for many applications. From what I hear, IR is a
bit more complex than SLIP, and performs better on faster lines, like 56
kbps and up, than SLIP; on the other hand, IR supposedly has a higher
overhead than SLIP on slower lines, like 9.6 kbps, which is what _I_ am
interested in.
I've never seen IR in action; does anyone use it? How well does it work?
-Brent
Brent Chapman Capital Market Technology, Inc.
Computer Operations Manager 1995 University Ave., Suite 390
brent@capmkt.com Berkeley, CA 94704
{lll-tis,uunet}!capmkt!brent Phone: 415/540-6400
------------------------------
Date: Tue, 23 Aug 88 00:17:32 EDT
From: attcan!utzoo!henry@uunet.uu.net
Subject: "open systems"
>The main place openness is losing is in peripherals...
Amen. My occasional partner in crime, Geoff Collyer, commented aptly on
this trend: "`Open systems' means they want your wallet to open."
Henry Spencer @ U of Toronto Zoology
uunet!attcan!utzoo!henry henry@zoo.toronto.edu
------------------------------
Date: Tue, 23 Aug 88 08:39:22 CDT
From: strong%stetsun@mcc.com (Michael Strong)
Subject: automounting ... why bother
Reference: v6n193
If you are at a big site, the number cross mounts that are possible grows
exponentially as you add systems. If you have a relatively small site
with everyone's home directories on a few servers, you won't appreciate
the big win of having automounting. We haven't used it yet, but have
wanted it for a long time. Now that it is more common for Suns to come
packaged with their own 327Meg (or more) on local pedestals, we are moving
more home directories to the local machines for performance. When that
individual remotely logs in, it would be nice if his home directory would
come and go on remote machines as he or she does.
The machine hanging problem is ministered to more by not allowing nfs
mounts immediately under /, but having them go away completely when not
being used, of course, protects you even further.
Mike Strong
@ MCC VLSI CAD Program [512] 338-3642
P.O. Box 200195, Austin, TX 78720
3500 West Balcones Center Drive, Austin, Tx 78759
ARPA: strong@mcc.com
UUCP: {ihnp4,seismo,harvard,gatech,pyramid}!ut-sally!strong%mcc.com
------------------------------
Date: Fri, 12 Aug 88 12:11:04 +0200
From: Haavard Eidnes IDT <he@idt.unit.no>
Subject: Proxy ARP daemon source
As "promised" on the TCP-IP list, I am submitting my proxy ARP daemon for
Suns running SunOS 3.[345] for inclusion in the Sun-Source archive. The
program is useful if you have a subnetted IP network (who doesn't), where
some of the hosts on the network do not understand subnetting. For other
documentation, see the man page, or look up the description in RFC1027.
If you ever wondered how to get at the ethernet address of the local
machine (asked about in this forum some Sun-Spots ago), the program
includes code to do just that.
By the way, I am aware that a program with the same functionality (or
better? don't know) is available from Sun on special request (I was
unaware of this when I wrote the program), and a program by Barry Shein
(only partly source code) inspired this program (I could not make his
program work for me).
Hope this proves useful to some.
Haavard Eidnes, Division of Computer Systems and Telematics
Norwegian Institute of Technology, N-7034 Trondheim, Norway
E-mail:
he@idt.unit.no (will cease to work for ~2 months from now)
he%idt.unit.no@norunix.bitnet
[[ The source has been placed in the archives under "sun-source" as
"proxyarpd.shar" (it is 25913 bytes long). 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: Tue, 23 Aug 88 01:12:05 EDT
From: dan@wind.bellcore.com (Daniel Strick)
Subject: some curious performance problems
During the last several months I have noticed some rather serious
performance problems when writing files on my Sun workstations. These are
not really new problems. They have been around since at least release 3.4
and are still present in release 4.0. I just didn't notice them before.
I was wondering if anyone else has noticed. (These are BIG performance
problems. Is it just me? Am I crazy?)
The program I use to measure disk i/o rates is too long to include here,
but the basic procedure is equivalent to:
#include <stdio.h>
char b[8192];
main ()
{
int i;
long t1, t2;
t1 = time(0);
for (i = 0; i < 3000; ++i) write (1, b, sizeof b);
t2 = time(0);
fprintf (stderr, "i/o rate = %d kb/sec\n", 24000/(t2-t1));
}
The test program is run with a command like "disktest >junkfile". I often
view system activity with the iostat program when I run the test program.
The i/o rates reported by iostat should only be slightly higher than the
i/o rates reported by the test program. This is what I observe when I
test local disk systems.
When I use the test program to write a file on a remote file system, I
find that the i/o rate on the server is many times the i/o rate reported
by the test program. Typical numbers are 250 kb/sec on the server and 60
kb/sec on the client running the test program. It has been suggested that
this happens because NFS is "stateless" and must do the equivalent of an
open, a close, and a sync for each file block written. (I don't know. I
just observe the results. I also note that since under release 4.0
diskless clients now swap to NFS files, ...)
I recently replaced my sun-3/160 with a sun-3/60. The sun-3/160 had the
old sun scsi disk system (70 MB ST-506 disk, Adaptec ACB-4000 scsi
converter, "sc" scsi host adapter). Typical i/o rates were 200 kb/sec
write, 240 kb/sec read. I kept the old disk. After the cpu swap (which
also changed the scsi host adapter to "si"), paging performance was often
awful. I eventually ran my disk test program and discovered: 250 kb/sec
read, 50 kb/sec write. Iostat gives the same numbers. Can someone
explain this? Is this strange behavior a result of mixing the old drive
with the new host adapter? Does it happen with new drives?
------------------------------
Date: Mon Aug 22 19:27:23 1988
From: portal!cup.portal.com!DMasterson@sun.com
Subject: IconEditor problem / At problem
Two separate problems that I've discovered in using a couple of Sun 3s
with SunOS 3.5. Any help would be appreciated.
1) I keep a directory of icons for general playing with. In using the
IconEditor, I tried typing in the directory name and then hit the BROWSE
gadget to look at the directory. Boom! Exit with core dump in THAT
DIRECTORY! If I give it the name of the directory and the name of a file
in that directory, then hit BROWSE, I get a popup requestor that shows the
icon for me to select ==> no problems. Am I using browse wrong?
[[ It works for me. How many icons are in that directory? --wnl ]]
2) Has anyone tried the AT command with SunOS 3.5? 'At' accepts my
request fine and builds a spool file to run at the later time, but, when
that time rolls around (or CRON's approximation thereof), the command
doesn't run. Checking all the logs and testing things out shows that CRON
does run /usr/lib/atrun as expected, but /usr/lib/atrun bombs with a 'bad
header block' (or something close to that) error. The spool file looks
ok, but I'm not sure what its format is supposed to be (not documented
anywhere that I can find). The spool file is set up as a /bin/sh command
file with a header block at the beginning that has lines of '# info: data'
type. Have I forgot to set something up to use it properly?
Thanx for your help
David Masterson
DMasterson@cup.portal.com
------------------------------
Date: Mon, 22 Aug 88 16:34:58 CDT
From: selig@xanth.msfc.nasa.gov
Subject: 4.0 spline labels for 3" binders
>In v6n133 Jim Morton writes...
>
>I have taken the previous posting of the Sun logo binder spine labels for
>the SunOS documentation and modified them for 4.0....
I needed a slightly different organization of the manuals here. I had 3"
binders, I wanted to bind the beginners guides also, and I had SPARC
specific stuff.
Additionally, I intended to have multiple sets of the manuals. One would
be my SysAdmin copy that no one could have. A copy in the machine room
that was not supposed to be removed, and a copy (or two) available for
checkout.
Thus, I took Jim Mortons version and added the 'ownership' lines in the
label definition and redid the binder organization.
Enclosed is the SysAdmin copy. For the others I change SYS MGR COPY in the
'sunlabel' definition to either MACHINE COPY or USER COPY and I change DO
NOT REMOVE to SIGNOUT REQD as necessary. This organizes everything into 13
3" binders.
Hope this helps someone. Consider it partial payment for all the help I've
received from this list.
--bill
US MAIL: Bill Selig, NASA/MSFC, ES53
Huntsville, Al 35812 USA
TPC: (205)-544-7608
FTS: 824-7608
SPAN: sam::selig
INTERNET: selig@xanth.msfc.nasa.gov -OR-
selig%sam.span@fedex.msfc.nasa.gov
BITNET: selig%sam.span@star.stanford.edu
UUCP: uunet!selig@sam.SPAN -OR- uunet!sam.span!selig
[[ The Postscript file has been stored in the archives under "sun-source"
as "seliglabels.ps" (I'm running out of names). It is 6753 bytes long and
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: Tue, 23 Aug 88 00:38:00 EDT
From: jdh@bu-it.bu.edu (Jason Heirtzler)
Subject: PC-NFS - request for your comments
There has been a lot of talk about PC-NFS and I'd like to hear (good &
bad) from those who have been using it. Please mail to me directly and
I'll post an abridged summary (providing there's sufficient interest.)
--Jason Heirtzler
Boston University
(617) 353-2780
jdh@bu-it.bu.edu
[[ Just a reminder: there is an Internet mailing list devoted to NFS
implementations, especially for personal computers. Submissions should be
mailed to "nfs@tmc.edu" and requests to "nfs-request@tmc.edu". --wnl ]]
------------------------------
Date: Mon, 22 Aug 88 16:38:55 MDT
From: alberta!dutchyn@uunet.uu.net
Subject: Questions about 386i communications
I have relatively simple questions, yet ones which I think my be of some
general appeal. A company which I do some part-time consulting has a
client who just purchased a Sun 386i. The 386i is to be used for CADD,
but the client is interested in doing other things on it also. In
particular, they have a Compaq acting as a 3270/3780 emulator (with file
transfer) over a bi-synchronous modem on an (obviously) dedicated line to
a 4300 series IBM mainframe. Currently, the Compq is used for remote job
entry, and some file transfer. They would like to have the Sun 386i take
over this function. As a result, I need to get information on two aspects
of this communication:
1. Are there any bi-sync modems which operate on a Sun 386i?
(I need to know about baud rates (if applicable),
device drivers, and the like...)
2. I have head some things about SunLink (primarily from SunSpots),
is it available for the 386i, and will it perform file transfer
and allow the Sun to act as a remote job entry system?
Any information or pointers to information would be greatly appreciated.
Christopher Dutchyn
email: dutchyn@alberta.edu
snail: Christopher Dutchyn
#64 Lancaster Terrace
Edmonton, Alberta
CANADA
------------------------------
Date: 23 Aug 88 02:52:08 GMT
From: pyramid!leadsv!laic!cygnus!darin@decwrl.dec.com (Darin Johnson)
Subject: Why won't "traffic" work?
I am having trouble running traffic on some machines. On machine A, with
a local root partition, but remote /usr partition), it works just fine.
On machine B (actually, all of the diskless clients), it doesn't work. I
have run "rpc.etherd ie0" and it gives no errors. In fact, from machine
A, I can type "traffic -h B" and it works fine also. (as a side note, the
version of traffic is the one from Sunlink/DNI that has decnet info)
Here is some output:
B# traffic
Can't contact rpc.etherd running on B.
RPC_PMAP_FAILURE
cygnus# rpcinfo -p
program vers proto port
100007 2 tcp 1024 ypbind
100007 2 udp 1028 ypbind
100007 1 tcp 1024 ypbind
...
100001 1 udp 1099 rstatd
100001 2 udp 1099 rstatd
100001 3 udp 1099 rstatd
100010 1 udp 1117 etherstatd
100010 2 udp 1117 etherstatd
100015 6 udp 1126 selection_svc
1073741824 6 udp 1154
1073741824 6 tcp 1052
B#
The last two lines weren't there before traffic was run, and two more
lines get added each time traffic is run.
Darin Johnson (...pyramid.arpa!leadsv!laic!darin)
(...ucbvax!sun!sunncal!leadsv!laic!darin)
------------------------------
Date: Tue, 23 Aug 88 08:42:43 EDT
From: Steven.Hoover@edrc.cmu.edu
Subject: Questions about windowing systems: NeWS?
There seems to be a proliferation of graphics tools and windowing systems
for Sun machines and I'm confused. So please excuse my ignorance and help
me by answering some very basic questions. I realize that these may
require some long uninteresting answers to the experienced SUN hackers out
there so please mail your responses to me. I will summarize if there is
interest.
1.) Under SunOS 4.0 what are the functions of Suntools, SunView,
SunCGI, Xwindows and NeWs?
2.) To what extent are any of the above compatible or mutually exclusive?
and most importantly
3.) If I have a Sun-3/160 running 4.0 and NeWs networked with another
unix workstation (Specfically an IRIS 4-D system) which runs NeWs
what functionality does this give me?
Thanks,
Steve Hoover
sph@edrc.cmu.edu
Department of Mechanical Engineering
Carnegie-Mellon University
Pittsburgh, PA 15217
------------------------------
End of SUN-Spots Digest
***********************