Sun-Spots Digest, v6n10
William LeFebvre
Sun-Spots-Request at RICE.EDU
Wed Feb 3 09:30:31 AEST 1988
SUN-SPOTS DIGEST Tuesday, 2 February 1988 Volume 6 : Issue 10
Today's Topics:
The case of the missing Bitnet subscribers
Pixwin operations in SunView panels (2)
Re: Adding a node to a server-based system
Adding Diskless Nodes Without Using Setup
Route additions/deletions hang
client crashes
Problems with Adaptec controller
Spurious level 3 interrupts!
Help with SunView's pw_set_region_rect clipping function
SunDNI question
Suncore view surface windows?
Lisp for Sun-4?
C++ on the Suns: which vendor to choose?
Tektronix 4107 emulator for color Suns?
Adding a super eagle to a Sun?
Experiences with Emulex tc12 or Xylogics 772?
Fujistu 2372 disks on Suns?
Send contributions to: sun-spots at rice.edu
Send subscription add/delete requests to: sun-spots-request at rice.edu
Bitnet readers can subscribe directly with the CMS command:
TELL LISTSERV AT RICE SUBSCRIBE SUNSPOTS My Full Name
Recent backissues are stored on "titan.rice.edu". For volume X, issue Y,
"get sun-spots/vXnY". They are also accessible through the archive
server: mail the word "help" to "archive-server at rice.edu".
----------------------------------------------------------------------
Date: Tue, 2 Feb 88 16:32:49 CST
From: William LeFebvre <phil at Rice.edu>
Subject: The case of the missing Bitnet subscribers
It seems that somewhere around the beginning of the year many of the
Bitnet addresses disappeared from the Bitnet list server at RICE. Further
investigation revealed that the person who maintains the RICE list server
set up a peer list at UBVM to help minimize the Bitnet traffic caused by
sending a Sun-Spots issue. Unfortunately, he forgot to complete all the
work on the Rice end to link the two groups. That has now been finished
and all of our Bitnet subscribers should receive this issue. My apologies
to all who were inconvenienced. I did not discover that there was a
problem until a few days ago.
Now we just have to figure out how to get the first nine issues of the
1988 digest to all who missed them. I would like to put them in the list
server as retrievable files---this would make retrieval for Bitnet people
very easy. As soon as I figure out how to do that I will. If you can't
wait until then, you may retrieve them via the archive server.
William LeFebvre
Department of Computer Science
Rice University
<phil at Rice.edu>
------------------------------
Date: 22 Jan 88 18:48:40 GMT
From: tolerant!vsi1!lmb at ucbvax.berkeley.edu (Larry Blair)
Subject: Pixwin operations in SunView panels (1)
I've seen several inquiries regarding drawing in Sunview panels, so I
thought I'd try to clarify the issue.
There are currently 2 ways to draw on a SunView panel. The most obvious
one is to do pixwin operations to the panel's pixwin. The problem that
everyone who tries this encounters is that a panel's pixwin is not
retained. This means that the application must catch the WIN_REPAINT
event and redraw the pixwin.
The other method is to create memory pixrects and draw your picture in
them with pixrect operations (or load canned pixrects at compile time).
You can then create a PANEL_MESSAGE item for each of the pixrects, using
PANEL_LABEL_IMAGE. The panel package will then redraw the picture when
necessary.
SunOS 4.0 will contain a new feature, panels within canvases. This will
be the best solution to this problem.
Larry Blair - VICOM Systems Inc. (408)432-8660
2520 Junction Ave. San Jose, CA 95134
{sun!pyramid, ucbvax}!tolerant!vsi1!lmb
------------------------------
Date: Thu, 21 Jan 88 11:11:30 PST
From: bickel at nprdc.arpa (Steven Bickel)
Subject: Pixwin operations in SunView panels (2)
If you do not need to move the panels location once it has been displayed
then I would suggest something like the following:
panel = window_create(frame, PANEL,
PANEL_EVENT_PROC, my_event_proc,
PANEL_BACKGROUND_PROC, my_bckgrnd_proc, 0);
my_event_proc(item, event)
Panel_item item;
Event *event;
{
Pixwin *pw;
panel_default_handle_event(item,event);
pw = (Pixwin *) window_get(panel, WIN_PIXWIN);
/* or any vectors etc. */
pw_writebackground(pw,0,0,pw->pw_pixrect->pr_size.x,
pw->pw_pixrect->pr_size.y,PIX_SET);
panel_paint(panel,PANEL_NO_CLEAR);
}
If you "need" to move the panels location once it has been displayed then
what we have added ( no panel refresh function to use ) is something like
the following:
move_panel(panel,rect)
Panel panel;
Rect *rect;
{
show_window(panel,rect)
/* note the time - 0.5 sec must be long enough to occur after display */
set_the_notify_itimer(refresh_our_panel,0.5);
}
show_window(window,rect)
Window window;
Rect *rect;
{
window_set(window,
WIN_X, rect->r_left,
WIN_Y, rect->r_top,
WIN_WIDTH, rect->r_width,
WIN_HEIGHT, rect->r_height,0);
window_set(window, WIN_SHOW, TRUE,0);
notify_client((Notify_client)window);
}
refresh_our_panel()
{
Pixwin *pw;
set_the_notify_itimer(nulltimer,0.0);
pw = (Pixwin *) window_get(our_panel, WIN_PIXWIN);
/* or any vectors etc. */
pw_writebackground(pw,0,0,pw->pw_pixrect->pr_size.x,
pw->pw_pixrect->pr_size.y,PIX_SET);
panel_paint(our_panel,PANEL_NO_CLEAR);
}
set_the_notify_itimer(timer_proc,update_interval)
void (*timer_proc) ();
float update_interval;
{
static char *me = (char *)&me;
struct itimerval timer;
timer.it_interval.tv_usec = (int)(1000000 *
(update_interval - (float)((int)update_interval)));
timer.it_interval.tv_sec = (int) update_interval;
timer.it_interval.tv_sec = 0;
timer.it_value.tv_usec = 1;
timer.it_value.tv_sec = 0;
notify_set_itimer_func(me, timer_proc, ITIMER_REAL, &timer, NULL);
}
nulltimer()
{
}
Note: this is providing you are not already using the notify_itimer for
something else.
If you find anything that works better for panels that are moved I would
appreciate all feedback. Hope this helps.
Steve Bickel bickel at nprdc.arpa
Systems Engineering Assoc. (619) 553-7675
Naval Personel R & D Center.
------------------------------
Date: Thu, 21 Jan 88 11:57:17 EST
From: deke at socrates.ee.rochester.edu (Deke Kassabian)
Subject: Re: Adding a node to a server-based system
Lots of folks do actually keep extra partitions around for the diskless
clients they intend to add. If you didn't do that, and you have a
filesystem you are willing to trash, you won't have to rerun setup.
The trick is to unmount that filesystem, take it out of fstab, make
entries in nd.local, do a mkfs on the appropriate ndl? devices, copy a
clients nd0 partition into that space using tar, fix up things like
"hostname: in rc.boot on the new client, and re-run nd. You're done.
If you are interested in doing this, and need more than that
readers-deigest-condensed explanation I just gave, contact me and I'll be
happy to help.
Deke Kassabian, University of Rochester
Deptartment of Electrical Engineering
deke at ee.rochester.edu
or ...!rochester!ur-valhalla!deke
------------------------------
Date: Thu, 21 Jan 88 20:00:18 EST
From: gotham!escott at sun.com (Scott Menter - NYC SSE)
Subject: Adding Diskless Nodes Without Using Setup
<jipping at cs.hope.edu> (Mike Jipping) forgot to leave extra ND partitions
when he ran setup the first time, and wants to know if he can do anything
besides running setup again to support more diskless nodes. Indeed, there
is another way to add diskless clients, but believe me, it's easier to use
Setup.
Disclaimer: the following is not intended to be a tutorial; please read
the System Administrator's manual for details. This note is only intended
to demonstrate that it is likely best to use Setup as Sun Software Support
advised the original poster, and Sun and I accept no responsibility
whatsoever if you try the procedure below and things don't work out just
as you'd like. Okay?
Start by doing a level 0 dump of *everything*.
You can repartition the disks using the "partition" monitor command (see
your system administration manual for details). Then, bring up mini-root,
and run "newfs" on your newly repartitioned file systems.
Now, unless I've left something out, you can restore your root and /pub
file systems from tape.
Make sure you install boot blocks ("newfs" should take care of this).
Then bring up your system single user and take care of restoring
everything else.
Warning! If any of your partitions shrink in this process, you'll be
bumming when you can't get the stuff off of tape! Also, you'll have to
manually put in the /etc/nd.local entries necessary to boot your new
client. Also, the "partition" monitor command is less than entirely
intuitive.
All in all, it takes about the same amount of time to run Setup and then
restore your files onto the newly generated file system. And, it's a lot
easier to do. And, setup figures out the nasty bits of /etc/nd.local for
you. Still, it's a good question, because one of the things my customers
are always asking me to do is exactly the procedure we are now discussing.
...which is why I've cc'd this back to Sun-Spots.
Which reminds me....
You can always call Sun and pay to have some nice professional sys admin,
like myself, come down and spend a day to do it for you. This will run
you some money, but on the other hand, it will get done right, guaranteed,
the first time. Speak to your sales rep.
Hope this helps.
Scott Menter
Software Support Engineer, Customer Service Division
Sun Microsystems
smenter at sun.com
------------------------------
Date: Wed, 20 Jan 88 16:53:55 PST
From: casey at lll-crg.llnl.gov (Casey Leedom)
Subject: Route additions/deletions hang
I'm constantly running into a problem using route(8) on Sun OS 3.4.
Typically right after a 3.4 system is booted, I can add and delete as many
routes (as are legal) as I want, then it stops letting me and the route
commands just hang. I've even had this happen right after a system was
booted, but that's rare. The only solution I've found so far is to simply
reboot.
We're running a simple class B network with no subnetting. This happens
on 3/50, 3/160, and 3/180. It probably happens on the 3/280 I manage, but
I haven't noticed it there yet.
When I look at the route process, it's waiting on selwait. However, my
4.3BSD sources don't show any select calls at all. Does anyone have any
idea what's going on? And more importantly, does anyone know how to get
around this problem?
Casey
------------------------------
Date: Thu, 21 Jan 88 10:22:35 EST
From: Ned D. Danieley <ndd at cs.duke.edu>
Subject: client crashes
We are having trouble with our diskless clients crashing. This seems to
occur once a day, the first time anyone tries to do something on the
machine. The typical crash results from a user logging in on the console,
but it can also be triggered by a remote login or rsh. We get the
following message:
Jan 21 08:56
login:
trap address 0x8, pid 1669, pc = f01cda8, sr = 2000, stkfmt b, context 2
Bus Error Reg 80<INVALID>
data fault address 30f26b14 faultc 0 faultb 0 dfault 1 rw 1 size 0 fcode 5
KERNEL MODE
page map 23000000
D0-D7 1 140 2 0 0 0 0 1
A0-A7 f1ed3d8 ffff888e 0 effffe0 f070ba4 30f26b14 ffff875a ffff8756
Begin traceback...fp = ffff875a, sp = ffff8756
Called from f01ce58, fp=ffff8766, args=f1ed3d8 ffff8782 f01cc18 f1ed3d8
Called from f01cc18, fp=ffff8782, args=f1ed3d8 f070ba4 ffff87b6 f0709e8
Called from f048f24, fp=ffff87aa, args=ffff886a effffec 0 24288
Called from f0044cc, fp=effff44, args=50 0 50 2
End traceback...
panic: Bus error
syncing disks... 3 3 done
In the case of an rsh, login is replaced by in.rshd. This only happens on
clients that boot off the server: we have three machines that boot off of
local disks, and they don't crash, although they do mount several remote
file systems. The problem began when we added two 3/60s and a 3/110,
giving us a total of 10 diskless clients. Our setup is:
3/280 with xylogics controller and super eagle: serves as nd server and
provides /usr.MC68020 and a local partition to the clients.
3/160 with ciprico rimfire controller and super eagle: provides main user
files and another local file system.
3/52, just boots itself and gets all file systems other than root and swap
elsewhere.
10 diskless machines.
We originally also had a rimfire and super eagle on the 280, and it did
all the serving, but we removed it to see if it was the problem. The 280
hasn't crashed since, but the clients still do. And the 160 hasn't crashed
since we started serving off of it. Ciprico doesn't think it is their
problem, and in order to get Sun to put much time into it, we will
probably have to remove the rimfire completely, which would be a true
pain.
Any ideas about where to look?
Ned Danieley (ndd at sunbar.mc.duke.edu)
Basic Arrhythmia Laboratory
Duke University Medical Center
Durham, NC 27710
(919) 684-6807 or 684-6942
------------------------------
Date: Fri, 22 Jan 88 16:29:19 EST
From: polish at lexington.columbia.edu (Nathaniel Polish)
Subject: Problems with Adaptec controller
I am trying to attach a Seagate ST4096 disk to the Adaptec controller.
The Adaptec attached to the Sun-2 bus via a SCSI board. No matter how
many cyl's I specify in diag, the last two always fail. I can not get any
further as the program will not finish the format. Has anyone else tried
to attach an ST506 type drive to the Adaptec (one that is not in Sun's
table)?
Thanks
Nat Polish at cs.columbia.edu (ARPANET)
------------------------------
Date: Fri, 22 Jan 88 11:24 PST
From: <JON at UCLASTRO.BITNET>
Subject: Spurious level 3 interrupts!
To all Spots,
What does the following message mean when it appears on the console?
spurious level 3 interrupt
These have been happening more frequently on our 3/260 server. They seem
to occur when things are idle. Is the machine going down or what?
Thanks,
Jonathan Eisenhamer
UCLA Astronomyh
jon at uclastro.bitnet
bonnie::jon (span 5828)
------------------------------
Date: Thu, 21 Jan 88 14:29:26 PST
From: Earl <island!earl at sun.com>
Subject: Help with SunView's pw_set_region_rect clipping function
Has anyone tried using SunView's pw_set_region_rect to do clipping on a
retained canvas? The backing pixrect does not get rendered the same way
as the image in the window -- things get offset by the distance of the
origin of the clipping region from the origin of the logical canvas. The
problem manifests itself whenever the backing pixrect is used to repaint
the window, e.g., when hiding the window behind another window and then
re-exposing it, when scrolling, etc.
Earl
email: island!earl at sun.com
phone: (415) 491-1000, from 11:00 a.m. on
------------------------------
Date: Thu, 21 Jan 88 15:55 PST
From: <JON at UCLASTRO.BITNET>
Subject: SunDNI question
To All Spots,
I am running SunOS 3.2 with SunDNI 5.0/vt100tool and I am having some
difficulty. Specifically, when I do a dnalogin to a vax from the sun, the
control characters do not seem to operate properly. For example, the line
editing features ^A, ^H, ^E, left/right arrows do not do their appointed
tasks. ^A, instead of allowing characters to be inserted into the command
line, produces a carriage return. The left/right arrows do not move the
cursor, they seem to type the line (as far as I can tell with the cursor
movement). ^H and ^E do not seem to do anything at all. I know that it
is with the DNI and not the vt100tool emulation because the vt100tool
emulation works fine when I telnet to a terminal server attached to the
vax. Has anyone seen this before and/or suggest what might be the
trouble?
Jonathan Eisenhamer
UCLA Astronomy
jon at uclastro.bitnet
bonnie::jon (span 5828)
------------------------------
Date: Wed, 20 Jan 88 15:11:54 PST
From: merat at uw-ftcl.ee.washington.edu (Merat Bagha)
Subject: Suncore view surface windows?
In Appendix B of the SunCore Ref. Manual it is described how one creates
view surfaces in a new window. As such, one can define the dimensions
and positions of the iconic and normal forms of the window using an
optinal string in the vwsurf.ptr array.
How does one define the icon to be used in the iconic form? As it is, the
"/usr/include/images/core_eye.icon" is used. Is there any way to replace
this?
-- Thanks in advance.
Merat Bagha, Dept. of Elect. Engr., Univ. of Washington
merat at uw-ftcl.ee.washington.edu
------------------------------
Date: Fri, 22 Jan 88 10:33:52 AST
From: David Trueman <dalcs!david at uunet.uu.net>
Subject: Lisp for Sun-4?
What are others doing about Lisp for the Sun-4? Sun is not yet selling
their Common Lisp for the Sun-4. I understand that it is being developed
by Franz Inc. and that pre-release versions may be available, but our Sun
sales rep. has not been forthcoming with information. Does anyone have
any idea what the pricing will be like when it is available?
What alternatives are there? I am a system manager rather than a Lisp
user, so I am not up-to-date on what's available.
UUCP {uunet utai watmath}!dalcs!david
CDN david at cs.dal.cdn
INTERNET david%dalcs at uunet.UU.NET
------------------------------
Date: Fri, 22 Jan 88 13:50:50 EST
From: <lenhart at cs.williams.edu>
Subject: C++ on the Suns: which vendor to choose?
I am interested in acquiring C++ for some Sun workstations (mostly 3/50's
and 3/60's). I have spoken to folks at AT&T and Oasis. AT&T will sell
source for $250 (for any number of machines). Oasis sells a binary
version for $950/station with discounts for multiple stations. I have
heard that AT&T's source is very difficult to port to the Suns (we're
running release 3.4) and I would like to have a useful tool as soon as
possible. So my questions are
1) Has anyone successfully ported the AT&T product to the Sun?
-How difficult is it?
-If I got a licence for the source from AT&T could I obtain
the ported version from someone?
2) Has anyone used the Oasis product? How good is it?
3) Are there any reasonable other choices?
My net address is lenhart at cs.williams.edu. Thanks in advance.
Bill Lenhart
Dept. of Computer Science
Williams College
Williamstown, Ma. 01267
e-mail: lenhart at cs.williams.edu
------------------------------
Date: 20 Jan 88 17:09:57 GMT
From: dave at rosesun.rosemount.com (Dave Marquardt)
Subject: Tektronix 4107 emulator for color Suns?
We're looking for some way to allow our users to get color graphics
output. One suggestion that has come up is using a color Sun to emulate a
Tektronix 4107 terminal, and then print to a Tektronix 4692 or 4696 color
printer. I know that Tektronix will sell us a 4696 and associated drivers
for our Suns, but I'm not sure about finding a 4107 emulator. Does anyone
know where I can find a 4107 emulator for a color Sun?
Dave
------------------------------
Date: Thu, 21 Jan 88 10:27:05 EST
From: Ned D. Danieley <ndd at cs.duke.edu>
Subject: Adding a super eagle to a Sun?
I may have to add a third party super eagle to my existing xylogics
controller. The cables that came with the drive don't match the
connector(s) on the controller back panel. In addition, sun has changed
the back of their drives, making it hard to tell what is going on. I would
appreciate any pointers as to how this could be done.
Ned Danieley (ndd at sunbar.mc.duke.edu)
Basic Arrhythmia Laboratory
Duke University Medical Center
Durham, NC 27710
(919) 684-6807 or 684-6942
------------------------------
Date: Thu, 21 Jan 88 15:33:15 -0500 (EST)
From: "Anthony A. Datri" <aad+ at andrew.cmu.edu>
Subject: Experiences with Emulex tc12 or Xylogics 772?
I plan to use an Emulex tc12 in our VMS 785 and a xylogics 772 in our Sun
3/4 to control Cipher 880 microstreamers. Can anyone give me experiences
either way on such configurations?
anthony a datri
scribe systems
aad%gold.scribe.com at vb.cc.cmu.edu
------------------------------
Date: 21 Jan 88 13:25 -0500
From: naren <naren%sce.carleton.cdn at UBC.CSNET>
Subject: Fujistu 2372 disks on Suns?
Anyone has experience with Fujistu 2372 disks on Suns? This 825 Mb drive
is quoted at $ 9,300 Cdn, compared to $ 12,000 for a 689 Mb 2344. It is
supposed to be a limited time introductory offer, so please mail me your
comments. Thanks.
Narendra Mehta,
Systems & Computer Eng. Department, Carleton University, Ottawa, Canada
UUCP: {allegra,decvax,ihnp4,linus,utzoo}!watmath!clan!naren
ARPA: naren%sce.carleton.cdn%ubc.csnet at csnet-relay.arpa
------------------------------
End of SUN-Spots Digest
***********************
More information about the Comp.sys.sun
mailing list