[comp.sys.sun] Sun-Spots Digest, v6n210

Sun-Spots-Request@RICE.EDU (William LeFebvre) (09/01/88)

SUN-SPOTS DIGEST        Wednesday, 31 August 1988     Volume 6 : Issue 210

Today's Topics:
           Re: Ethernet "balancing" when serving many clients?
                       Re: tar: filename too long 
                       Re: pcnfs and sun networks
                         Re: format of a .o file
          getting your login environment in command(shell)tools
           Automation of two invocations of suntools on a 3/60C
                   Stupid YP behavior (services.byname)
                Problem with SunOS 4.0 suntools and tn3270
              Can't use Sony monitors without modification?
                subshell vs. top-level shell in suntools?
             Can someone give me a hand with color canvases?
                         ALM-1 vs ALM-2 question
                     seeking info on NFS for IBM MVS

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:    Tue, 30 Aug 88 11:04:00 EDT
From:    bob@allosaur.cis.ohio-state.edu (Bob Sutterfield)
Subject: Re: Ethernet "balancing" when serving many clients?
Reference: v6n206

biggers@dg-rtp.dg.com (Mark Biggers) writes:
>We have a Sun 3/260 to use as a server and a lot of diskless 3/50s.
>I want to know if it is possible and worthwhile to divide the 3/50s
>over two subnets to minimize Ethernet traffic.
>...
>Is it possible to put 3 Ethernets on a 3/260?

I've heard of as many as four, and you can build a nice router that way.
But you'll have problems doing what you want (distributing diskless
clients onto more Ethernets).

ND only runs on the first interface, the one on the CPU board (ie0).  To
set your server up as an IP router from a subnet full of clients onto a
house backbone, you must put the client subnet on ie0 and the backbone on
ie1.  Any other interfaces become ie2 and up.

This is another one that's presumably "Fixed In 4.0" :-) Sun noted, when
we found the ie0 restriction on ND, that they weren't going to bother
changing it because "ND is going away soon".  That makes sense to me.

>Does anyone have any other ideas about how to cram 18 diskless 3/50s
>onto one server?

We don't like to put more than 15 3/50s on a 3/180 with two XY451s and two
Eagles.  The bottleneck seems to be disk IO before CPU, memory, or
network; so using more modern controllers can probably push you closer to
20 before you bump into something else - perhaps the Ethernet, as you
observe.

	--bob

------------------------------

Date:    30 Aug 88 13:54:42 GMT
From:    mkhaw@teknowledge-vaxc.arpa (Mike Khaw)
Subject: Re: tar: filename too long 
Reference: v6n206

> The tar that is distributed with SunOS 3.4 has an annoying "feature" ...
> it apparently has a 100 character filename buffer.  Why is this?  I now

This is a limitation of the tar header format.  If you look up tar(5), it
says something like:

	...
	#define NAMSIZ 100
	...
	union hblock {
		...
		struct header {
			char name[NAMSIZ];
			...

This limit was in 4.2bsd and continues in both SunOS and Ultrix.  I don't
know about other flavors.

Mike Khaw
-- 
internet: mkhaw@teknowledge.arpa
uucp:	  {uunet|sun|ucbvax|decwrl|uw-beaver}!mkhaw%teknowledge.arpa
hardcopy: Teknowledge Inc, 1850 Embarcadero Rd, POB 10119, Palo Alto, CA 94303

------------------------------

Date:    Tue, 30 Aug 88 12:17:48 +0200
From:    mcvax!olsen!nagler@uunet.uu.net (Robert Nagler)
Subject: Re: pcnfs and sun networks

In reply to monty%tartarus@gargoyle.uchicago.edu:
> we are now so dissatisfied with the product that we'll probably replace it
> (we've had it for about a month).
> ...
> my feeling about this product is that it is truly a second rate and
> inferior product--almost amateur--and that sun should either show some
> professionalism about it or get out of the market.

First, in V6 #187 there was a message about a new version of PC-NFS which
is not made by Sun.   I sent away for information, but haven't received a
reply yet.  (If Carl Beame is listening, I am interested.)

We've been using PC-NFS for quite a while now for doing development with
the PC.  Although the product isn't perfect (I have many complaints), we
have found it acceptable for our purposes.  One must keep in mind what
type of hardware and software with which PC-NFS must contend.  (Put your
mind back to CPM/80, then speed up the picture a tad ;-)  In other words,
PC-NFS is trying to make the system do what it was never intended to do.

Considering the other network products available for the IBM PC, Sun's
products (IPC and NFS) are a dream.  The diversity of PC-NFS users is
quite amazing.  We use for remote compilation (and believe it or not, this
works with a bit of glue).  You want it to print perfectly.  Other people
use it instead of hard disks.   If you really think PC-NFS is a horrible
product, you might try some of the other network products available for
the PC.  Just go to any PC dealer and I'm sure you will be overwhelmed
with alternatives.

Rob Nagler		olsen!nagler@uunet.uu.net
Olsen & Associates/Seefeldstr 233 CH-8008 Zuerich/+411 55'22'24

PS. I would be wary of harsh statements such as "amateur", i.e.
    ``people who work in software houses shouldn't throw bugs''.

------------------------------

Date:    Tue, 30 Aug 88 11:22:30 edt
From:    mp%allegra@research.att.com
Subject: Re: format of a .o file

In v6n185, utzoo!henry writes:
>Danger!  Your Sun software license (if it's like ours) specifically forbids
>decompiling or disassembling!

If this is the case, what do people with source code do when "dot-dot"
releases come out?  Sun doesn't supply source code for dot-dot releases -
I asked - and it's always desirable to have the source match the binary.

	Mark Plotnick
	allegra!mp

------------------------------

Date:    Tue, 30 Aug 88 11:38:25 EDT
From:    Bennett Todd <bent!bet@mcnc.org>
Subject: getting your login environment in command(shell)tools

The problem: you want to have your personalized environment with aliases,
internal shell variables, and an artificially intelligent prompt:-). You
don't want to waste the time, and cause the compatibility problems, that
you get when this is done in .cshrc for *every* child shell which is
created (such as those launched for globbing). The canonical solution:
start .cshrc with the following incantation.

	if ($?prompt == 0) then
		exit
	endif

Then follow it with all the amazing wizardry you like. Any non-interactive
shells will not have the prompt variable set and so will exit. The only
shells which will be customized are those which will be prompting you for
interaction.

-Bennett
bet@orion.mc.duke.edu

------------------------------

Date:    Tue, 30 Aug 88 13:48:21 BST
From:    Nick Holloway <alfie%OPAL.WARWICK.AC.UK@cunyvm.cuny.edu>
Subject: Automation of two invocations of suntools on a 3/60C

After reading how various people have got two invocations of suntools
running, I eventually got it to work (well almost -- see later).  What I
found, however, was that there are a lot of magic incantations to get
right, and in the right order, so I thought it could be automated.  Here
is my attempt.  (Note that although I have used the word color, this is
only because this is what suntools uses, as it comes from US background,
but I feel that it should be spelt colour, because I have a UK background
:-)

This was for a Sun 3/60C, with /dev/fb and a /dev/cgfour0.

To get the two invocations run/exec the program the program twoscreens as
below (This could always be done as a csh alias etc)

    #!/bin/sh
    # File: twoscreens

    suntools -d /dev/fb -s ".suntools.b&w" -toggle_enable -overlay_only

This starts up the first invocation of suntools on the overlay plane.  The
startup file starts up the other invocation of suntools and any other
tools wanted.

    # File: .suntools.b&w
    cmdtool  -Wl "CONSOLE -- Black & White" -WL console -C
    # Start other invocation of suntools ...
    suntools -d /dev/cgfour0 -s .suntools.col -toggle_enable -8bit_color_only
    # ... and a way to reach it.
    switcher -WL "To Color" -d /dev/cgfour0 -s f
    # Insert your tools to startup here

The second suntools start up file contains

    # File: .suntools.col
    # Provide means to get back to other invocation
    switcher -WL "To B&W" -d /dev/fb -s f
    # Insert your tools to startup here


When it comes to quitting suntools, quit the colour (second) invocation
first, and this will dump you back in the black&white invocation which can
then be exitted.  I found I got read errors and a locked console
otherwise.

Possible improvements are to give the colour console a coloured border, so
it is clear which invocation you are currently using.

The problem that I hinted at earlier is that in the colour suntools (the
one started second) the shelltools are not set up as they should be, and a
reset is necessary.  If anyone has any suggestions as to why, and a cure I
would be interested.

--
JANET : alfie@uk.ac.warwick.cs
UUCP  : ..!mcvax!ukc!warwick!alfie 
BITNET: alfie%uk.ac.warwick.cs@ukacrl
ARPA  : alfie%cs.warwick.ac.uk@cunyvm.cuny.edu

[[ Nick sent changes in at a later time.  These changes were incorporated
into this message.  --wnl ]]

------------------------------

Date:    Tue, 30 Aug 88 10:29:50 EDT
From:    dpk@morgan.com (Douglas P. Kingston)
Subject: Stupid YP behavior (services.byname)

>   From:    Ronald G Minnich <rminnich@super.org>
>   Subject: questions on sun3/280+sunOS4.0
...
>   In a slightly related question, i notice that the key in the ypmap for
>   services.byname is portnumber/protocol, e.g. the key for the map is the
>   thing you are looking for....Any good reason for this?

This is one of the bigger performance hits in YP, and there is virtually
no good reason for it.  I found this on our Gould implementation back at
BRL and informed Unix-Wizards and others about it (We didn't have Suns
then), but I guess its still not fixed.  This note is going to
hotline@sun.com.

The problem is clear.  The table they have, services.byname, should really
be called services.bynumber, and a new table should be created called
services.byname with key-pairs that look like:
	ntp/udp
	telnet/tcp
	smtp/tcp
	...
Then getservbyname needs to be modified to use it.  I did this on the
Gould and it was a big speed up for things like netstat, where you could
really see the change.  Other programs just got quicker and took less CPU.

Cheers,
	-Doug-

------------------------------

Date:    Tue, 30 Aug 88 10:00:17 LCL
From:    "John M. Wobus" <JMWOBUS@suvm.acs.syr.edu>
Subject: Problem with SunOS 4.0 suntools and tn3270

I am trying to use tn3270 with suntools, specifically "(beta test) version
3" of tn3270 (with a patch to prevent core dumps on Sun 4s) with suntools
running on a Sun 3/50 under SunOS 4.0.

My problem is that one line of the display is often garbled--usually the
first boldface line.  The garbling consists of missing characters, often
space characters.  Asking suntools to redisplay the screen or the window
doesn't fix the line.  However, asking tn3270 to redisplay (in our case,
by pressing control-g) does fix the line.

The problem arose before we applied the patch, but it doesn't arise when I
run tn3270 directly rather than under suntools.  It also wasn't evident
under suntools/SunOS 3.5.

Anyone else run into this?  If so, did you solve it?

John Wobus
Syracuse University

------------------------------

Date:    Tue, 30 Aug 88 10:06:36 EDT
From:    segall@caip.rutgers.edu (Ed Segall)
Subject: Can't use Sony monitors without modification?

I did a bit of checking into the useability of Sony GDM-1953, GDM-1602
monitors for Sun color displays, and found that they are probably not set
up to work properly.  The problem (I'm told) is that Sun uses a 66Hz
refresh rate, while Sony standard is 60Hz (actually 59.something).  I am
no video expert, but this difference is probably significant.  It impacts
internal timings as well as video bandwidth.  A video guy who sells Sonys
told me he didn't think the monitor would lock onto a 66Hz signal. 

However, I could be dead wrong.  Maybe all you need to do is increase the
main oscillator frequency by 10%. Or, increase it by 5% and have the
monitor make up the difference (there is some leeway in these numbers).
However, I wouldn't bet $3000 on it.  Anyone out there have better
information?  Sun doesn't like to let this sort of information out, and
Sony has an agreement with them whereby they won't tell you how to modify
their montors to work with any different specs.  However, complete Sun
video interface specs are available, so perhaps some video guru out there
can tell us if it really is possible to make the Sonys work with the Suns.
I just wouldn't count on it.

--Ed Segall

uucp:   {...}!rutgers!caip.rutgers.edu!segall
arpa:   segall@caip.rutgers.edu

------------------------------

Date:    30 Aug 88 14:14:56 GMT
From:    mkhaw@teknowledge-vaxc.arpa (Mike Khaw)
Subject: subshell vs. top-level shell in suntools?

How can a csh running in suntools tell if it's the top level shell in the
window or not?  Currently I "set ignoreeof" in .cshrc if $prompt and $term
are both set, so that I don't accidentally kill a shell/cmdtool window
with ^D; but then if I run "script" or some other interactive csh, I can't
exit with ^D.  I'd like to avoid using "ps" to find out, as it can be
slow.

Thanks,
Mike Khaw
-- 
internet: mkhaw@teknowledge.arpa
uucp:	  {uunet|sun|ucbvax|decwrl|uw-beaver}!mkhaw%teknowledge.arpa
hardcopy: Teknowledge Inc, 1850 Embarcadero Rd, POB 10119, Palo Alto, CA 94303

------------------------------

Date:    Tue, 30 Aug 88 11:47:14 EDT
From:    Anne Becker <anne@cvl.umd.edu>
Subject: Can someone give me a hand with color canvases?

I am trying to write a program in SunView that opens up a retained canvas
that I will draw to in either color or monochrome (specified eventually by
the user.)  I would like the canvas to be resizeable upon the whim of the
user.  I haven't had any problems with this when I work strictly in
monochrome, but when I set up the program for color, the following
symptoms occur:

My little colored squares appear to be fine until I choose to change the
size of the screen greatly (i.e., zoom or change to full screen). When I
do this, the image disappears, and I can't figure out how to retrieve it.
If I create the canvas with scrollbars, after I scroll across the canvas,
the image reappears, so I know it hasn't completely disappeared.

The second problem is that after a call to refresh from the window's menu,
all the blocks turn purple (the highest entry in the colormap--I'm using
cms_rainbow).  Hereafter, anything I draw only comes out in this purple.

I assume I need to write a repaint procedure of some sort.  I don't want
to keep a running record of all the commands sent to the canvas to run
through each time a redraw is called, since much of what will be displayed
is rows of image data, and this could take up way too much space.  Has
anybody had experience in this area, and how did you resolve this sort of
problem?  Am I overlooking something blatantly obvious?  I really don't
want to reinvent the wheel on this one.

Many thanks in advance for any help or pointers.  (By the way, I am using
Sun OS v3.4.on 3/50s, 3/160s and a 3/280.)

	-Anne Becker
	ARPA: anne@cvl.umd.edu
	UUCP: uunet!mimsy!cvl!anne

------------------------------

Date:    Tue, 30 Aug 88 09:44:48 PDT
From:    limes@sun.com (Greg Limes)
Subject: ALM-1 vs ALM-2 question

The Systech MTI is a multibus card which may be installed in the Sun3 VME
bus via an adapter card. This guy is known affectionately as the ALM-1, or
sometimes just as an ALM. The internal drivers are all known by the
generic name "mti", and the device is generally accessed via devices named
"/dev/tty00" through "/dev/tty3f".

The ALM-2 is a VME card designed and built by Sun. It uses a different set
of drivers (all known internally as "mcp"), and is accessed via devices
named "/dev/ttyh0" through "/dev/ttykf".

You mention that you have an MTI in your 3/180, and that occasionally
lines will wedge. First off, have you called Sun about it? Second, there
is not much anybody is going to be able to do to help unless you mention
what version of the operating system you are running ...

Greg Limes [limes@sun.com]

------------------------------

Date:    Tue, 30 Aug 88 09:34:50 EDT
From:    dunigan@msr.epm.ornl.gov (Tom Dunigan 576-2522)
Subject: seeking info on NFS for IBM MVS

Our IBM group is interested on any information, products, or plans for
Sun's NFS on IBM MVS systems.

thanks
  tom
  dunigan@msr.epm.ornl.gov

------------------------------

End of SUN-Spots Digest
***********************