Sun-Spots-Request@RICE.EDU (William LeFebvre) (08/05/88)
SUN-SPOTS DIGEST Thursday, 4 August 1988 Volume 6 : Issue 166
Today's Topics:
Re: Suntools startup bug/feature (3)
Re: Kic and MAGIC for Sun
Re: (Answer) on Unused Client
Re: NeWS Question
Exporting under 4.0
vt100 session messes up Sun screen
4.0 Sunview question: "alert mouse" cursor color
Problem with Fujitsu M2372K switch settings
Question about Sun386i
on-line documentation for SunView?
Scrollable Manual pages with bold and italics?
Sun manual binders: how many?
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, 24 Jul 88 12:21:30 EDT
From: asmith@tcgould.tn.cornell.edu (Arthur Smith)
Subject: Re: Suntools startup bug/feature (1)
Reference: v6n149
Mark G. Brown writes:
>I thought I'd seen everything, but when one of our staff came up with this
>one this morning, I was completely baffled.
... .suntools file
>when he started suntools from his home directory, the cmdtool, clock, and
>1st shelltool appeared on the screen. A ps ax showed the 2nd shelltool
>process, but without any associated tty....
>Has anyone ever experienced this phenomenon?
I have experienced exactly the same problem on our diskless sun 3/50's,
and others here have also.
Aside from re-booting, another way of "fixing" the problem is to start TWO
shelltools in background from another shelltool, with the result that the
second one starts up immediately, but there is no sign of the first until
the original shelltool is killed. On the first occasion I failed to notice
that one of my shelltools was missing, and allowed it to use up nearly 60
minutes of cpu time running on in the background.
It is certainly not due to running out of pty's, but it does sound like
both shelltools are trying to use the same pty. The problem is also
definitely repeatable, even though it started rather suddenly: it has been
happening every time I log in for the past few days. I attempted to start
a shelltool under dbx, and the same problem seemed to be happening - it
was stuck on line 245 in shelltool.c, where the second window_create
occurs. But I couldn't figure out what was going on on any finer scale
than that.
I don't know if this helps, but it would be nice to have problem fixed.
-- Arthur
------------------------------
Date: Mon, 25 Jul 88 09:45:41 EDT
From: Dave Goldberg <dsg@mitre-bedford.arpa>
Subject: Re: Suntools startup bug/feature (2)
We ran into this problem recently, too. What I found was that some
applications (I still haven't figured out which) seem to leave the pty's
with mode 622 when they exit as opposed to the proper 666. Thus, when the
tool tries to access that pty, it can't read it and hangs (I suppose that
this much is really a bug in the shell tool). One thing that I know
causes it is to toggle a getty on the pty. Occasionally, the wtmp file
will not be updated properly after a shell exits and w(1) will show a huge
amount of idle time for a process that doesn't even exist. (It doesn't
seem to generate any accounting data though fortunately). Anyway, the fix
to that was posted to info-unix or unix-wizards a while back, and it is to
toggle a getty on the offending pty. To do that, edit /etc/ttys and
toggle the first char in the appropriate line to 1. Then kill -1 1,
re-edit /etc/ttys to change the 1 back to a 0 and then kill -1 1 again.
This will effectively get rid of the w problem, but will leave the pty in
mode 622 instead of 666. The mode problem is new to us in this case (ie
toggling getty's), which has worked with no problem since version 3.2
through 3.4. It just started happening under 3.5. I believe that other
programs left the modes incorrect under earlier versions, though but as I
said, I have been unable to track down which ones as yet.
hope this helps,
-dave
Dave Goldberg ARPA: dsg@mitre-bedford.arpa
The Mitre Corporation UUCP: linus!mbunix!dsg
MS B015
Bedford, MA 01730
617-271-2460
------------------------------
Date: Tue, 26 Jul 88 15:39:29 EST
From: munnari!yabbie.rmit.oz.au!rcodi@uunet.uu.net (Ian Donaldson)
Subject: Re: Suntools startup bug/feature (3)
I frequently see this, but its due to a particular pty (eg: /dev/ttyp3)
being not mode 666 like it should be, when nobody else is using it.
suntools goes into an infinite loop chewing up cpu when it sees this. If
you go and su, chmod the pty to 666 you will find the appropriate tool pop
up instantly and suntools stops chewing cpu.
Presumably suntools is being smart and using the error code returned from
the open (which should be EPERM) to loop and try again. Why it doesn't
try the next available pty I have no idea. Maybe it thinks that any error
code other than EBUSY means there are no free pty's??
I have no idea why the mode of a pty gets left wrong either, but I presume
that telnetd or rlogind is dying unexpectedly, forgetting to reset the
mode before it exits.
A reboot fixes it because there is a "chmod 666 /dev/tty[pq]*" in /etc/rc.
Maybe there should be some daemon that regularly wakes up and chmods
unused pty's back again... ugh! but its better than a reboot of a heavily
used multi-user machine.
Ian D
[[ But if this is the problem, then why doesn't the program simply choose
a different pty when it discovers that it can't open that one? If it did,
there would be no problem. If it doesn't, then explain why a functioning
tool exiting would cause a non-functioning one to start working. I'm not
trying to insult anyone here, I'm just trying to fit all the symptoms to
the explanation. --wnl ]]
------------------------------
Date: Sun, 24 Jul 88 15:42:16 EST
From: munnari!cad.oz.au!shand@uunet.uu.net (Mark Shand)
Subject: Re: Kic and MAGIC for Sun
Firstly with respect to magic, I think if you read the magic manual a
little more closely you'll find you can make color work on a 3/60C. I
can't remember the deatils, but I know we initially had it only in B/W,
but later found that setting the right options made color work.
Secondly with respect to kic, I have a driver for SUNVIEW that works on
our 3/60Cs running SUNOS3.4. I'd be happy to send it to you if you can't
get one from someone in the US. (Seems silly to generate bulk
transpacific traffic gratuitously).
Mark Shand - shand@cad.unsw.oz.AU
------------------------------
Date: Sun, 24 Jul 88 16:46:55 EDT
From: mcgrew@topaz.rutgers.edu (Charles)
Subject: Re: (Answer) on Unused Client
The answer is to be found in your nd.local file if you're running the 3.X
version of the operating system, or with a simple 'du' if you're running
4.0. For 3.X, /etc/nd.local looks something like this:
clear
version 1
user 0 0 /dev/xy0e 0 21440 -1
user frik 0 /dev/xy0c 49580 5360 0
user frik 1 /dev/xy0c 54940 25460 -1
user frak 0 /dev/xy0c 80400 5360 1
user frak 1 /dev/xy0c 85760 25460 -1
son
the line "user 0 0 /dev/xy0e 0 21440 -1" declares the size of the /pub
partition, in sectors; so this /pub is ~10.5 Meg. The next 4 lines decide
the size of the root and swap areas for the two clients.
"user frik 0 /dev/xy0c 49580 5360 0" says that the root for 'frik' is 2.6
Meg (we do a lot of amusing things to cut down on the size of the client
root partition, which surpisingly enough showed up in 4.0), and "user frik
1 /dev/xy0c 54940 25460 -1" sets aside 12.5 Meg for swap space. So the
client uses 15 Meg. (This is low on swap space, even for a sun2 - which
this is - but it's usable.) Note the usage of xy0c. This goes back to
the decision by sun to allow pub.MC68010 and pub.MC68020 and so on, which
runs you out of paritions on the disk. The clients wind up in a 'hole'
between xy0b and xy0d. There's nothing to stop you from using some other
unused parition instead, and in that case the offsets start at 0. (See
'man nd' for more info.)
For 4.0, rather than nd, Sun uses nfs. For the each client, sun_install
creates a directory structure in /export/root, e.g. /export/root/frik,
and a big file to swap onto, e.g. /export/swap/frik. You can just look
at how big these are with normal unix commands (du, ls -l, etc).
Whether you want to use the extra client space depends on how accessible
the extra space is (and how badly you want it). If its in /export (4.0),
you'd have to do a repartition /export to grab back the extra space
(requiring a refresh of the old client's file system and swap space, since
its in a real filesystem on the server) - unless you want to just use the
space as-is, which could be slightly messy; if its 3.X, you can probably
do some creative reparitioning and access that space without affecting
anything else - remember that the partitions don't necessarily have to
follow one another when selecting an unused parition for that 'new'
space).
Hope this helps,
Charles
------------------------------
Date: Mon, 25 Jul 88 10:19:18 -0400
From: Henry B.J. Krempel <krempel@pacrat.npac.syr.edu>
Subject: Re: NeWS Question
> From: ssc-vax!dmg@beaver.cs.washington.edu (David Geary)
>
> If I write a NeWS-based application, can I run the application on another
> Sun that does NOT have NeWS loaded on it? If not, how much of the ~16
> Megs of NeWS software must be loaded on the workstation?
If you simply need to use the other sun to perform calculations (like a
remote Sun-4 or other Unix-based thing) all you need is the client code.
Source for this is usually under /usr/NeWS/clientsrc (depending on where
you loaded NeWS.
If you want to run the NeWS server, I think you can do without the
directories: demo and smi, both of which are pretty big.
None of the files in the NeWS directory are written to by processes
running NeWS, so the best solution, is to use NFS to share a single copy,
assuming you have a network in place.
------------------------------
Date: Sat, 23 Jul 88 18:02:10 PDT
From: eggert@sm.unisys.com (Paul Eggert)
Subject: Exporting under 4.0
In Sun-Spots v6n141 david@elroy.jpl.nasa.gov (David Robinson) writes that
a fileserver need not have an /export partition under SunOS 4.0, because
you can put the client areas under /usr, and "When you export /usr, this
implicitly exports ALL subdirectories and files below it." His statement
is true, but unfortunately his method doesn't handle client superuser
access correctly. E.g. our fileserver's /etc/exports is:
/usr -ro,access=allclients
/export/root/client1 -root=client1,access=client1
/export/swap/client1 -root=client1,access=client1
etc.
Were we to put client root filesystems under the server's /usr, we'd have
to give all the clients root access to /usr -- not always a good idea. I
believe this is partly why Sun recommends /export.
------------------------------
Date: Sat, 23 Jul 88 16:11:01 PDT
From: cnrdean2@violet.berkeley.edu
Subject: vt100 session messes up Sun screen
We sometimes use our Sun 3/50s (Sun OS 3.2) to emulate terminals to other
campus computers via rlogin. When we do this, we run bare Unix, no
windowing. If an application requires a terminal type of vt100, we find
the workstation will not scroll properly for subsequent non-vt100
applications: it clears the screen after 24 lines of text and redraws from
the top. This is very disconcerting for `vi' users!
`reset' or `clear' will not clear this up: the only solution we find that
works is to reboot.
Is there a better (software reset) that will solve our problem?
Jim Bradley, CNR Computer Facility, UC Berkeley
jim@insect.berkeley.edu
------------------------------
Date: Sat, 23 Jul 88 03:42:45 CDT
From: "Rich Winkel" <MATHPG1@UMCVMB.BITNET>
Subject: 4.0 Sunview question: "alert mouse" cursor color
I've just booted up OS4 on our first Sun workstation here. (A 4/110)
Question 1 of 7M: Certain dialog boxes in sunview (such as the box
resizing prompt and the 'Are you sure you want to quit?' box) insist on
setting the mouse arrow to the background color, making it impossible to
see. Changing the default background makes no difference. Can anyone
help?
Thanks,
Rich Winkel
(MATHPG1@UMCVMB.MISSOURI.EDU)
------------------------------
Date: Sun, 24 Jul 88 18:54:14 MDT
From: davis@groucho.ucar.edu (Glenn P. Davis)
Subject: Problem with Fujitsu M2372K switch settings
Hi. We have hooked up Fuji M2372K 824 Meg drive as the second drive on our
system. The other drive is a 2361 'Super Eagle'. The controller is a
Xylogics 451.
When using Sun 4.0 format, I get
"Warning: disk not set for slip-sectoring."
Sun's docs refer me to the controller and drive manuals.
(The 'format.dat' entry is
disk_type = "Fujitsu M2372K" \
: ctlr = XY450 : fmt_time = 4 \
: ncyl = 743 : acyl = 2 : pcyl = 745 : nhead = 27 : nsect = 67 \
: rpm = 3600 : bpt = 40960 : bps = 604 : drive_type = 0
)
I'm lacking the controller manual, and don't see anything in the drive
manual that would do this. Anybody know what one should do to enable
slip-sectoring on this drive?
Glenn P. Davis davis@groucho.ucar.edu
UCAR / Unidata
PO Box 3000 1685 38th St.
Boulder, CO 80307-3000 Boulder, CO 80303
(303) 497 8643
------------------------------
Date: Sat, 23 Jul 88 15:15:23 EDT
From: "Paul H. Calamai" <phcalamai@water.waterloo.edu>
Subject: Question about Sun386i
I'm considering the purchase of a Sun386i but I have some unanswered
questions (the people at Sun couldn't help either). I own an Apple
LaserWriter Plus and would like to be able to access it from the Sun.
Currently it hangs off my IBM-PC and my Macintosh Plus over a TOPS network
(note: TOPS is a product distributed by SUN). The PC has a TOPS board in
it that essentially provides an appletalk interface (hardware and
software). I don't want to spend mega-bucks (ie. by transcript binaries)
in order to access the LaserWriter from the SUN386i. Is there an easy
way? Can I simply drop the TOPS card into one of the SUN AT slots and use
it that way? Any advice would be greatly appreciated!
Paul H. Calamai
Department of Systems Design Engineering
University of Waterloo
Waterloo, Ontario N2L 3G1
(519) 885-1211 x3182
------------------------------
Date: 23 Jul 88 22:15:27 GMT
From: foo@portia.stanford.edu (castor fu)
Subject: on-line documentation for SunView?
Does anyone have anything like the Sunview reference guide available on
line? Having something like manual pages would be really nice for when
working at home, when physical manuals are not readily available.
-Castor Fu
foo@fizzle.stanford.edu
------------------------------
Date: Sun, 24 Jul 88 18:02:14 CDT
From: drl@vuse.vanderbilt.edu (David R. Linn)
Subject: Scrollable Manual pages with bold and italics?
Has anyone devised a method for getting bold-face and italics with man(1)
in a cmdtool? The tty page for defaultsedit(1) seems to imply that the
various types of bold face, underlined, and inverse text are availble only
when the cmdtool when is in "ttymode" instead of "textmode". I am aware of
less(1n) and the PAGER variable used by man(1) but I want a
bold/underlined manual page I can reference after exiting from man(1).
David
David Linn - drl@vuse.vanderbilt.edu
------------------------------
Date: Sat, 23 Jul 88 9:59:18 CDT
From: Herb Barad <barad@bourbon.ee.tulane.edu>
Subject: Sun manual binders: how many?
I saw the posting for the Sun manual binders. They're selling six-packs
of different size binders. Given that I have version 4.0 of the software,
how many binders will I need? I suppose the answer will be different
depending upon the binder size, so what combinations to complete a set are
popular?
Herb Barad [Electrical Engineering Dept. - Tulane Univ.]
INTERNET: barad@ee.tulane.edu
USENET: barad@bourbon.uucp
------------------------------
End of SUN-Spots Digest
***********************