SunOS 4.1 Query Responses
Len Jacobs
lj at ursa-major.SPDCC.COM
Fri Jul 13 22:46:00 AEST 1990
[[Ed's Note: Let me get up on my editorial shobeox for a moment and talk
about 'summaries'. When you do a summary, please remember a few things:
(a) always summarize your original question so we all know what you were
talking about (b) *summarize* the responses in a brief format (c) include
a few key responses in their entirety if they really make a definitive
point. This is the second time I've had to send this out (evidentally it
didn't make it the first time) and it's kind of got to me. I think this
could have been _summarized_ as 'Of x respondents, y said they were
running 3.5, y said 4.0.03, y said 4.1: those who were running 4.1
reported the following problems: {list of problems}. This commpedium of
responses could have been mailed to those who really wanted it. Also, one
additional point: when you summarize, you should always leave the 'From:'
field in your quoted messages, that way your original authors get their
credit. I've tried to clean this up a little, but I don't personally have
the time to summarize it correctly - future submitters should be warned;
however that "summaries" in this compedium of responses format will be
rejected for rewriting. -bdg]]
Well, the responses have quieted down so I am summarizing the responses I
received and mailing the whole file to you. Please respond to
comp.sys.sun along with whatever you received. And please send me a copy
of what you discovered.
Unfortunately I have less valuable information than I had counted on.
Several weeks ago I requested experiences managers were having with the
latest upgrade question. I received responses from ~125 people, with the
majority included below. Hopefully there is some value in this exercise.
------
We have an automatic software update agreement with sun that means that we
automatically get new releases at no additional cost (that is, no cost
above the agreement...which does cost something).
Consequently, all of our suns are running 4.0.3/4.1 and all will almost
certainly move to 4.1 during the current break between semesters.
Since we have not yet upgraded anything to 4.1 (our only 4.1 systems are
preinstalled ss1 units) I can't respond to your second query.
Frank W. Peters Systems Programmer Computing Center & Services
fwp1 at CC.MsState.Edu Peters at MsState.Bitnet (601)325-2942
-----
Our group has two Sun workstations. The Sun 4/110 has been upgraded as
far as SunOS 4.0.3. The Sun 3/110 is still sitting at SunOS 3.4, awaiting
the day when we can get 4 Mb more memory for it. (Watching a Sun 3/110
try to run under SunOS 4.0.3 in 4 Mb memory is not the prettiest of
sights. :-) )
It is uncertain whether we will upgrade to 4.1. The upgrade from 4.0 to
4.0.3 changed the interface Sun uses to talk over the SCSI port to our
(unsupported) Ciprico M990 tape drive. While we were able to get new ROMs
from Ciprico to deal with this change, we are no longer at a point where
we can afford to find other such surprises awaiting us in future upgrades.
Tony Nardo, INET: trn at warper.jhuapl.edu, trn at aplcen.apl.jhu.edu
Johns Hopkins Univ./APL UUCP: {backbone!}mimsy!aplcen!trn
-----
We still have our gateway systems running 3.x, as they need a working
hyperchannel driver. All other systems are running 4.0.1, with the
exception of some sparcstations that came with 4.0.3c only...
Bob Van Cleef - vancleef at nas.nasa.gov
-----
Ok, here's my idiotic anecdote:
1) we get SunOS 4.1, tapes and docs - no time to install it, we have a
conference next week, etc.
2) the hard disk goes, so... do I spend the time to re-install 4.0.3
and the install 4.1 next week? Naaah! So we install 4.1
3) now, we need FORTRAN (don't look surprised, we are physicists, Ok? :-)
I load the tape, run extract_unbundled and... oops! Somebody at Sun
hard-coded a line with SOS_COMPAT="4.0" and PROD_BASE_SOS_RELEASE="4.0"
in the installation script (/usr/tmp/unbundled/1.2_fortran).
So the beast refuses to install Fortran 1.2 on the 4/260... I went ahead
and modified the script, and everything worked. Compilation and execution
of my usual fortran stuff doesn't show any problems.
I put a call to Sun, just in case. I wanted to know if could possibly be
that something in 4.1 would cause Fortran 1.2 *not* to work. The bottom
line is they didn't quite know. I was told the software compatibility
information I wanted must be someplace in the Sysadmin manuals... I
haven't got the time to wade through a foot of paper, so I still don't
know the official answer.
Beware if you keep any system running 4.0.3, your fortran programs
compiled on a machine running 4.1 may not run on them, even if compiled
with -Bstatic. Just found that one today, I can't quite put the finger on
the problem, though it looks like there are unresolved modules in some
system library. I don't know what happens going the other way. I'm bound
to find out, since we haven't got the upgrade tapes for the Sparcstations
yet.
Roberto Gomez | Arpa: roberto at bondi.phyast.pitt.edu
-----
We just upgraded to 4.0.3. I dislike using anything from Sun of the form:
X.X.0 for the obvious reasons. Thus we're often one sub-minor revision
or 1 minor revision behind.
-dave fetrow- fetrow at bones.biostat.washington.edu
dfetrow at uwalocke (bitnet) {uunet}!uw-beaver!uw-entropy!fetrow
-----
>>>>> On 11 May 90 23:30:13 GMT, lj at ursa-major.spdcc.com (Len Jacobs) said:
Len> I am wondering how many Sun users are still working within 3.? as opposed
Len> to 4.? I know that any survey will be biased since only a small minority
Len> of users will respond, but respond nonetheless. I will evaluate the
Len> response and summarize.
We're still working with 3.5 (all two of us). Needless to say, cost is
the contributing factor.
Evan Harris
evan at goanna.cs.rmit.OZ.AU Dept. of Computer Science
...!uunet!goanna.cs.rmit.OZ.AU!evan RMIT, Australia
-----
In a small lab with 2 sun3/60 and 1 sun3/80, I have one of them running
sunos 3.5 in order to run some old products. The others are running sunos
4.0.3. The 3.5 will soon be upgraded to 4.03 too.
-----
We are running 4.0.3 on sun3 & sun4 and 4.0.2 on sun386i. We have just
received 4.1 but are holding off installing it because:
a) We have a customer base on 4.0.3 to support and 4.1 binaries apparently
will not run on 4.0.3 (according to Sun)
b) We are waiting a while to see if anyone reports any major problems (kinda
selfish I know)
-----
We are still running SunOS 3.5 at our site. I'm using mine as an X Window
[ DECwindow ] development platform; MIT's X11r4 Server, Saber C,
FrameMaker. I'm currently considering an upgrade to 4.1 sometime this
summer, if the reviews are good.
WHO: Richard Hess CORP: Consilium
UUCP: ...!uunet!cimshop!rhess Mountain View, CA 94043
-----
Brown University Computer Science is running almost all 4.0.3 with a
couple of 4.1 for evaluation. We have found a couple of bugs in 4.1 but
they are less severe than those in 4.0.3, so we will probably upgrade to
it shortly.
Peter Baer Galvin (401) 863-7623
Systems Manager, Brown Univ. Comp. Sci. pbg at cs.brown.edu
-----
All 20 of our Sun3's (2 servers,6 standalones, 12 workstations) are
running 4.0.3c. I have the 4.1 tapes sitting here on my desk and an
installation manual, but I'm waiting for some documentation sets.
Probabaly be another month before I attempt an installation of 4.1. Also
keeping an eye and ear on comp.sys.sun and sunmanagers to see how 4.1
'flys' out there.
Bill Vaughn
bill at cvs.rochester.edu
-----
We are running 4.0.3 on all our sun4s and 3.5 on our sun3s. This is
because we still have software from vendors that hasn't been ported to 4.x
yet... 4.0.3 is stable enough that I do want to upgrade.
Christopher Welty --- Act. Director, RPI CS Labs | "Porsche: Fahren in
weltyc at cs.rpi.edu ...!njin!nyser!weltyc | seiner schoensten Form"
-----
We run SunOS 4.0.3 since about august 1989 (Sun3's only). I haven't
received SunOS 4.1 yet. I'll probably upgrade to 4.1 in two month (as soon
as our new Sun 4/490 arrives).
PTT Research Neher Laboratories [ E-mail : BJ_Lippolt at pttrnl.nl ]
P.O. Box 421, 2260 AK Leidschendam, [ BITnet : LIPPOLT at HLSDNL5 ]
The Netherlands. Tel: +31 70 3325439 [ UUCP : hp4nl!dnlunx!lippolt ]
-----
We're still on 3.4 and 3.5. We have 4.0.3 and 4.1, but only recently got
all of our add-in software for 4.0. NSE 1.2 took 18 months after 4.0 to
be 4.0-compatible. After we got NSE, we started ordering all of our
add-in software that wasn't under maintenance. It took about four or five
months to accomplish that (one month of delay ours, the rest Sun's). Now,
we need a good breaking point to start upgrading things. -- Craig Jackson
dricejb at drilex.dri.mgh.com
-----
We are still running 3.5. I had no hesitation in going from 3.2 to 3.4 to
3.5, buyt the stories I heard about 4.0 made me wait, and the sloth of
3/50's under 4.0 have kept me back at 3.5 all this time. I now have funds
for more memory for our 3/50's, and 4.1 is a full install, so we're going
to make the move soon. Thus, your summary will be very helpful if it
arrives in time... :-)
I don't know what you want to know, but we have 1 3/260 and 8 3/50's and
(just arrived) 4 4/65's supporting 40-50 occasional users, 10-20 frequent
users.
Robert Bruner rrb at math.wayne.edu
-----
We are using SunOS 4.0.3 and 4.0.1 for 13 of our suns, and plan to upgrade
to 4.1 sometime this summer. The other five are SS1s that came with 4.1
installed.
Charles H. Buchholtz chip at ee.upenn.edu
-----
We have:
Sun 3/180 -- 3.5
12 * 3/50 (diskless) -- 3.5
4 * 3/50 standalones -- 3.2
Bob Noonan Internet: noonan at cs.wm.edu
-----
Hello, I saw your posting in SunSpots today. We at McGill's School of
Computer Science are still largely SunOS 3.5-based, since most of our CPUs
are 68020s, and we were not happy with the decreased performance of SunOS
4.0 on a Sun-4/280, let alone anything like a Sun-3/50! The number of
serious reported bugs was also quite alarming--it was quite obvious that
Sun rushed the release of 4.0, and I'm sure that they suffered for that
mistake.
For the record, we have 2 Sun-3/280s and 19 Sun-3/50s, all running SunOS
3.5; and 2 Sun-4/280s running SunOS 4.0.3. The number of CPUs in the
School is likely to double over the summer: I expect that a good number of
them will be SPARCs, no doubt running SunOS 4.1. I would be surprised if
any of the Sun-3s will be upgraded to SunOS 4.1 given the time
constraints, but stranger things have happened... :-)
Christopher Rabson McGill University
E-mail: albion at milo.cs.mcgill.ca Montreal, Quebec
-----
We still run 3.5.2 on all production machines, except the one Sun-4/260,
which runs 4.0.3. We do run 4.0.3 on machines on which we've had to test
drivers. From everything I've seen, SunOS 4.0 was an amateur release.
(For that matter, everything after 3.2 was pretty crufty. Maybe 4.1 will
be OK.)
-----
I use SunOS3.5 right now. I've been thinking about upgrading to 4.X, but
have hesetated to do so because of my hardware.
I've got two Sun2/120 workstations, one with 4meg of ram and the other
with 7meg. I'm told my 7meg machine might run *faster* under 4.X, but
everyone says that running the newer OS on the 4meg beast would be a big,
big loose.
Any idea if Sun is going to (or has already) release the source code for
some hideously outdated version of their OS? How about just their device
drivers? I'd like to be able to fix bugs...
-----
At our site, we have three Suns -- a Sparc at 4.0.3c, a 386 at 4.0.2, and
a Sun3 at 3.5 (!). The reason we have the 3 at 3.5 is because we are a
software development company, and we believe most of our customers have
3.5 on their Sun3s, and further that they would not be able to execute
binaries that were made with 4.0 or 4.1. I don't know if this is true or
not.
Eric Hanchrow yamada-sun!eric at nosun.west.sun.com
-----
We are still running SunOS 3.5 on our 4MB 3/50s, 3/60s and 3/180 server
because:-
1) At the time of purchase/SunOS 4.0/SunOS 4.0.1 some software, eg Oracle
Ingres and UNIRAS was not available in a full form.
2) We did/do not want to upgrade in mid term/course
3) There was no money available to get round the "4MB performance problem"
which many people reported. We are hoping that SunOS 4.1 gives us
equally good(!) performance as SunOS 3.5 otherwise we are in a mess.
-----
Our project has 8 Sun-3/50s all still using Sun OS 3.4 or so. We are
planning an upgrade to 4.1 soon, though.
Dan Franklin
BBN Systems and Technologies
-----
Our site still currently runs S.OS 3.5 and has been for about 1.5-2 yrs
now. (Sun 3/50m) However, we are about to upgrade the system. We are
planning on adding our 3rd SCSI drive (660M) and moving to S.OS 4.1. I
would be *VERY* interested in the summary of problems should there be a
good response. It may even turn out our site will send you a list of our
problems, kinda depends on how soon all of this comes together. If you
want any other info email me and let me know. Actually email me and let
me know why you're interested also.
-----
Is not real clear what metric for "Sun users" you intend to use: sites,
cpus, wc -l /etc/passwd, average actual users. here's our rough breakdown
by cpu:
sun cpus sunos3.5 sunos4.0.3
3/160 1
3/60 9
3/50 2 6
4/60 5
Bob Gregory Data Processing Department UUCP ...!uunet!bwc!bob
-----
The Suns at here are in mixed OS environment. All Sun-3s are still running
SunOS 3.5. But all Sun-4s are running 4.0.3. One reason we are forced to
stay on 3.5 for Sun-3s is that some of our Sun-3s have only 4MB memory and
running 4.0 on them would cause performance problem (we don't have enough
money to buy more memory). Another reason is that some expensive software
packages, such as Interleaf's TPS, running on them and the object codes
between 3.x and 4.0 are not compatible. Of course there're a lot of bugs
in 3.5 but most of them have been fixed.
For the Sun-4s, we have only one choice, i.e. running 4.0.3. But the
performance is the biggest problem we have. All of them are configured
with 8Mb to 16MB memory. It seems to me that 8MB is not enough for heavy
jobs, as we are mainly doing graphics and cad/cae. I don't know what will
happen when 4.1 is installed.
If you have a summary of responses, I would like to have a copy.
-----
We still have sites running 3.2, although we plan to upgrade them
(probably to 4.1) in the coming months. We found 3.2 the most stable for
our purposes (which include SCSI device driver development), so never
bothered with >3.2. Porting to SPARC took a little while so may well end
up going out first on 4.1. Quite a jump!
Martin Reed, Racal Imaging Systems Ltd
uucp: mr at ritd.co.uk, uunet!ukc!ritd!mr
-----
We are using 4.0.3. We will not upgrade to 4.1 as it is our invariable
experience that an x.x release is full of nasty bugs. From SunSpots, it
looks as though 4.1 is no exception. We'll wait for 4.1.1 (or more likely
4.1.2,...)
David Carter
SRI International Cambridge Research Centre
Cambridge, U.K.
-----
We run 3.5 on all our Sun2 systems, and a few inertia bound Sun3s. On
most of our 3's and all of our Sparcs, we run 4.0.3. Given the history of
Sun releases, I'm inclined to wait for 4.1.1 or 4.1.2 before switching to
4.1, (1.4 was the first decent 1.X release, 2.2 the first decent 2.X, 3.2
the first decent 3.X, 4.0.3 the first decent 4.0.X ...), but will be
watching the net to see other people's experiences.
Ted Nolan
ted at usasoc.soc.mil
-----
At Polaroid we are running SunOS 3.5 (68020 machines, CV CADDS stations),
SunOS4.0.2 (386i), and SunOS 4.0.3 (SPARC's). In absolute count the 60+
CADDS systems dominate. The decision here is driven primarily by
applications. The reliable CADDS still runs on 3.5 and we will not switch
a major application in mid project. We wait until project lulls.
R Horn horn%hydra at polaroid.com
-----
We still run 3.5 on all of our non-sparc machines (we have only two
sparcs). The only trouble is with the dynamic-linking stuff added by Sun
in 4.0. It should be possible to write a program to take a dynamically
linked executable & produce a statically linked version of the same, but
(a) Sun didn't do so, and
(b) they didn't document the dynamic link structures so that we could do
it instead.
without such a program we must be able to boot SunOS 4 to run a few
imported programs for which people insist on sending (only) dynamically
linked binaries. (they could use -Bstatic and produce something that we
could run easily, but they don't.)
I should add that we use our own bitmapped-graphics support libraries
which are tiny, so that the usual arguments for dynamic linking & shared
libraries don't apply here. Our 3.5 machines rarely page things out once a
working set has been established.
Some machines here run an experimental version of the system in which
Berkeley's silly pageout daemon has been replaced by a proper per-process
working set system (with copy-on-write fork instead of vfork). We also
have our own streams implementation, and many other things. it mounts all
its file systems using NFS (ND is, optionally, used for swapping, since
its performance seems better than 3.5's NFS impl.) I'm using that kernel
now. The experimental kernel is really rather smaller than the 3.5 one
that Sun supplies (even when that is sensibly configured). Currently it
is about 270k bytes without ND, but 70k bytes of that is taken by
SunWindow support code. NFS takes a large chunk of the remainder, but I
can't do much about that...
-----
We have a number of machines running 4.0.* here. We still have a couple
running 3.5. The reason for that is that the NIT driver in 4.0.* is
broken, and we need to use Ethernet monitoring tools such as "tcpdump" and
"statspy" that cannot be made to work on 4.0.*. I haven't had a real
chance to play with 4.1 yet to find out if this is fixed, but if it is,
3.5's days here will be numbered.
--Greg
-----
Here at Lynxys we are still doing just fine with SunOS 3.5. We are
running a standalone 3/160 which started life in 1985(!!) as a 68010 based
system (because the 68020 wasn't available yet). We have a mere 4M of
memory (that's why we still run 3.5). It is unclear that we will ever get
to 4.x because the power supply probably won't hack any more memory and is
NOT upgradeable!! For disks we have a Fuji Eagle hanging off a XY451 and
the 60M internal SCSI drive in the pedastal. Laserwriter and various
terminals hang off of a 16 port Integrated Solutions serial board (Sun's
16 port serial board was vaporware at the time this started.) We are
connected via ethernet to a network of Compaq 286 Deskpros running Xenix
and an IBM AT fileserver running Novelle accounting software. As you can
see, we talk to the world via a UUCP connection at Purdue University.
Interesting the mixed hardware you can have around!!
Jim Wildman, Production Manager
Lynxys, Inc UUCP = {backbone}!pur-ee!lynxys!jaw
Lafayette, IN or
317/497-0309 jaw%lynxys.uucp at ee.ecn.purdue.edu
-----
I still run at 3.5. Why?
1) All the (real) horror stories.
2) Known 4.0 and 4.1 nfs bugs
3) Problems with PD Franz Lisp at 4.0
Mike Clarkson mike at ists.ists.ca
-----
We upgraded to 4.1, and the only complaint I have is that every once in a
while we get "not enough memory" errors. We never got them under 4.0.3.
Kevin Brady
brady at osi.ncsl.nist.gov
-----
The Computing Centre here operates a rather small Sun network:
File server: 3/180
Clients: 3/110, 4 x 3/50
Stand-alone (but configured as servers for simplicity): 3/110, 3/50
We moved earlier this year to 4.0.3 on the file server & its clients
and subsequently on the separate 3/110 when it was recalled to the main
building for some testing. The separate 3/50 still runs 3.5 as it
seems quite happy with it, nobody has complained, and it is annoyingly
far from our centre of operations and so needs an excursion to do anything
with it. Anyway, I am hoping for 4.1 before too long - Maths here
have it already although they are not installing it for a few days yet.
Frankly, I feel that if the file server still operated 3.5 nobody would
be in the least concerned except as a matter of principle.
-----
I was involved in upgrading the last two companies I was working for
(Alldata Corp and Martine Marietta, Sacramento) to 4.1. We were already
running 4.1 by the time I got to Boeing, though, so I don't know if any
upgrade problems were encountered.
At Martin Marietta, the upgrade went whole completely smoothly with no
problems whatsoever.
We had a couple of problems at Alldata, partly because we had a number of
3/50s on the network and partly because we had some hardware and software
that hadn't yet been tested under 4.0 by the manufacturer.
We had a half dozen or so 3/50 with only 4 MB of RAM in them. As you
probably know by now, SunOS 4.x does not run well in only 4 MB of RAM.
These machines were dog slow, and they ended up upgrading them all to 8 MB
sometime after I left.
One problem we ran in to was with a device driver for a Maxtor CD-ROM
drive. I believe the driver was by Ciprico, but that was two years ago
and my memory may be failing me. We figured out the differences between
3.x and 4.0 device drivers (4.0 wants an extra parameter in some
structure) and made the changes.
We also had an AnaTech optical scanner whose software was written on VMS
and ported to Sun. I don't remember the details of this, but I seem to
remember having some problems recompiling under 4.0.
David L. Kensiski sunkist.west.sun.com!oahu!bcshaw!dkensis
Boeing Computer Services
-----
I spent a few months trying to get serial ports to work under Sunos 4.0.1
pl 5 without success. I went back to 3.5 and remain there today. The
4.0.3 and 4.1 boxes just sit in the corner. If I get another disk, I'll
probably load up 4.1 to see if it works, but my estimate of the
probability doesn't outweigh the known effort to swap operating systems.
I'm hoping someone will post their experiences with 4.1 soon.
-----
We are running SunOS 3.2 on 3 3/50s; will switch to 4.1 this summer.
----
We upgraded to 4.0 two summers ago, currently 4.0.3 and will go to 4.1
after the school years ends.
My old company devnet?.hac.com (Hughes Aircraft Co.) is still running 3.x
due to requirements of various third party software packages. Contact
wdelv at devnet3.hac.com if want more comments.
-----
I manage about 60 Suns. 12 of them have disk and the rest are diskless.
Currently, 9 diskless clients and 4 of the diskfull machines are running
4.0.3. The rest of the machines are split between 3.5 and 3.4. I've been
holding off on upgrading everybody to 4.x because the rest of the machines
are going to all have to make the jump at once and it's going to be an
enormous job. I'd do it one server at a time, but many of our users run
SunLink 3270 and the gateway machine has to be running the same version of
the software as all the clients. :-(
We had plans to take all the machines to 4.0.3 the first weekend in June,
but then 4.1 arrived. It's looking like we'll slide our schedule about
six weeks and shoot for a 4.1 install in mid-July. The extra time is for
testing user-written software under 4.1 and verifying that all our third-
party software will work. It'll be nice to be on the latest release, but
I'm not looking forward to the upgrade.
Paul Allen
pallen at atc.boeing.com
-----
The system adminstrators at our site would like to upgrade the 100+ Sun 4s
to SunOS 4.1. We are interested in the Advantages of 4.1. What will
users gain by upgrading to 4.1? What application software will not work
under 4.1? Is there a way to install 10+ systems at a time? If there are
disadvantages to upgrading we would like that that information as well?
We are running YP on 1/3 of our machines and the automounter on 90% of our
machines. Are there problems with machines running 4.0.3 on a network of
machines running 4.1?
As usual, thanks in advance and I will post a summary.
| Name: Darren Curtis | Disclaimer: My opinion in my own! |
| email: d3c572 at pnl.gov | This space left blank intentionally! |
-----
We are still using SunOS 3.2 (I know, they just unsupported it at Sun),
but we intend on upgrading to 4.1 in the next few months. The reasons we
are still where we are are (1) the Suns before I got here were set up and
left to themselves, sort of the forgotten child in our DEC environment;
(2) more importantly, we have 13 diskless 3/50's with 4MB of memory and we
were told that we needed to have at least 8MB of mem. on those clients.
We are close to having the funds necessary to do the upgrade of those
clients. As soon as that's done, we'll upgrade the OS.
Jon Geld
geld at mail.physics.lsa.umich.edu
-----
We have a Sun 3/160 with a 16 port ALM-1 (serial board). We have recently
upgraded from Sun Os 4.0.3 to Os 4.1 and have had several problems which
appear to be related to the serial board.
1. A qms laser printer will no longer work on the serial board.
Everything queues fine but no job ever becomes active. This occurred
while we had the speed of the printer at 19200 baud. We changed the
speed to 9600 and still no luck. So we moved the printer to one of the
CPU card's serial ports and the printer worked. However at 19200 baud
parts of the files were being lost so we moved back to 9600 baud and
all is well.
2. We have vt100 clones (Ampex 220 and Wyse) attached directly to the
serial board. To give us multiple sessions we use a program called
screen. We were running our clones at 19200 baud as well but the
Ampex terminals started to hang when the screen program was being
used (note the Wyse terminals are having no problems). When the
speed of the Ampex terminals was cut back to 9600 baud the hanging
problem decreased in frequency but didn't go away.
3. We also have an X.25 (PAD) connected to datapac with four lines going
to the serial board. Typically we use tip and kermit to communicate
over these lines. We have found that file transfers using tip hang
after a very small amount of data has been transfered. The hanging
only occurs when files are tranferred from the 3/160 to the remote
machine. Transfers from the remote machine to the 3/160 work fine.
Kermit works without errors but is too slow for our use.
If anyone has any information about any of these problems I'd love to hear
from you. Sun is working on them but as of now they only say that the
serial drivers have been re-written.
..uunet!lsuc!array!gene
gene at array.uucp
-----
We have two suns, a 3/60 and a SS1. We just upgraded the SS1 to 4.1 with
not too many problems. There may be a few little gotchas, but by and
large I think it would be fair to say that we're happy with the outcome.
We will probably upgrade the 3/60 in the near future.
Jeff Hagen (jhagen at nrao.edu) is the individual who actually did the upgrade
for us and he may have more complaints about it (any gripes, Jeff?)
- Pat Murphy (pmurphy at nrao.edu)
-----
We are running 3.5.2 on a 3/280 serving two diskless 3/50's and a 3/160
serving another diskless 3/50 in another office (Toronto, Ontario). After
all of the initial hassles other people had with 4.0.X, I decided to hold
off until a 'reliable' release came along.
R. M. (Bob) Harvey, Corporate Computer Systems Administrator
E-mail: bobh at zaphod.uucp -=- Phone: (306) 931-1390 -=- Fax: (306) 931-1386
-----
I just upgraded my Suns to SunOS 4.1. Most things are fine, but I'm
having two problems.
I want to use a modem attached to serial port "b" both for incoming and
outgoing calls. I had no problems with this before (SunOS 3.5, 3.5.2 and
4.0.3), but now things are not working so well. When soft carrier is
turned off (either in the old way, by changing flags from 3 to 1 in the
kernel config file, or in the new way by using ttytab(5) and
ttysoftcar(8)) the line can still be used for incoming calls. But when I
try it for outging I run into problems. I use tip(1) to connect to the
modem, and then type commands to the modem to dial the remote site. Now,
after the upgrade tip's connection is immediatley closed when the
connection is made:
ANC > Dxxxxxxx
ANC > NUMBER = xxxxxxx
AUTO-DIALING ...
DIAL TONE
DIALING xxxxxxx....
RINGING
ANSWER TONE
INITIATING
Connection Closed
[EOT]
%
(xxxxxxx represents the phone number typed. The modem is compatible with
the Concorde CDS 224, but has a different prompt). If I type E (exit) to
the modem tip disconnects directly, without me typing the tip disconnect
sequence:
ANC > E
ANC > BY
Connection Closed
[EOT]
%
This was not the case with previous versions of SunOS. Uucp also suffers
from these problems. uucico(8) is also disconnected immediately when the
call is completed.
My other problem is with uucp. I used to use uucp from 4.3BSD, but now
that SunOS has HDB uucp I figured I should switch. I configured it today,
and it seems to work fine. But there is one thing from the BSD uucp I
miss. My L.sys file looked something like this:
xxxx Any/A,NonPeak ACU 2400 1234567 ogin:--ogin: Ufront word: xxxxxxxx
The interesting part is the second field (Any/A,NonPeak) meaning that uucp
jobs of grade A and higher can be transfered at any time and all other
jobs only during non peak hours.
Can I accomplish the same thing with HDB? I've tried a bunch of different
things in the Systems file, but all I am able to do is either do any kind
of transfer at any time or only do transfers at off-peak hours. Neither
is acceptable. I can kludge around this by having a script run from cron
that edits the Systems file (yuck!).
Does anybody have any useful input about either of these problems?
Reply by mail as usual, I'll summarize to the list.
Thanks in advance,
Svante
--
Len Jacobs
lj at ursa-major.SPDCC.COM
More information about the Comp.sys.sun
mailing list