Sun-Spots-Request@RICE.EDU (William LeFebvre) (04/08/88)
SUN-SPOTS DIGEST Thursday, 7 April 1988 Volume 6 : Issue 49 Today's Topics: Re: Clues about "tty grp reset" message Re: YP Problem Re: Can backups be done with system in multi-user mode SUN-4 problems (2) Error messages: MC68881 not available CDC Sabre 850 Mb SMD disk working on Sun with Xylogics controller Memory modules for Sun 3/60's? driver for Maxtor worm wanted Any problems with calentool (SunOS 2.2..)? VAX vs SUN? vt100 emulator that supports CPR? VT100 emulator? Difference between thin and "thick" wire networks? 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 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@rice.edu". ---------------------------------------------------------------------- Date: Tue, 29 Mar 88 14:48:37 CST From: maeder@symcom.math.uiuc.edu Subject: Re: Clues about "tty grp reset" message The problem described by darkstar!brian@uc.msc.umn.edu (Brian Utterback) was that a new cmdtool died with a message "tty grp reset...". We have observed the same problem on our 3/50s that are used by many different people (they are all in a public location). I found that when this happenend there always was a csh process from a previous user that was still connected to a pty. When the new cmdtool was opened it got the same pty and the two processes competed for input. I don't know under which circumstances these csh processes stay around. All I can do is kill them. If it happens to you open a new cmdtool (it will get the same pty) and before typing anything (which would trigger the error) close it and never touch it. You can then open a new one that will behave normally. Roman E. Maeder Dept. of Mathematics UUCP: ...!uiucuxc!symcom!maeder Univ. of Illinois at Urbana-Champaign Internet: maeder@symcom.math.uiuc.edu ------------------------------ Date: Tue, 29 Mar 88 00:07:35 -0500 From: Henry B. J. Krempel <krempel@pacrat.npac.syr.edu> Subject: Re: YP Problem Reference: v6n34 A few issues back Brian Clapper reported a problem using yppasswd after moving the password file from /etc into another location. I had the same problem when I did this, moving my file from /etc/passwd to /etc/yp/src/passwd. Luckily, I was still under 90-day support from Sun, so my friendly Sun support person suggested I move the password file back into /etc. I now keep my yp password file in /etc/passwd.yp and link /etc/yp/src/passwd to it (for completeness). Everything works fine. One more thing: you have to modify the Makefile in /etc/yp so that the right password file is checked when doing a make. Henry B. J. Krempel <krempel@pacrat.npac.syr.edu> Northeast Parallel Architectures Center (NPAC) 250 Machinery Hall Syracuse University Syracuse, N.Y. 13244 ------------------------------ Date: Wed, 30 Mar 88 11:11:06 EST From: montnaro@sprite.steinmetz.ge.com (Skip Montanaro) Subject: Re: Can backups be done with system in multi-user mode Reference: v6n38 We do our backups in multi-user mode with no problem. As I understand it, the problem with multi-user backups is that a directory inode might get deleted and then reallocated between dump passes. One thing I've toyed with, but never tried, is to send long-running processes a SIGQUIT signal, causing them to core dump. After running the (single user) backups and rebooting, the core image should be restartable using dbx or adb. Is this feasible? [[ There is no way to recover information about open files without the co-operation of the program itself. I suppose it would be possible to write a signal handler that could write unrecoverable information (such as open files, signals, etc.) out to a file before dumping core.... If only I had the time... For those who are interested: the Unix-TeX distribution includes a program called "undump" that reads a core file and its corresponding executable and builds a new executable from them. But you lose the open files and the new executable starts again at the beginning (instead of the point at which it was interrupted). But it still might be interesting to look at. --wnl ]] ------------------------------ Date: Tue, 29 Mar 88 21:58:39 CST From: steve@ncsa.uiuc.edu (Steve Christensen) Subject: SUN-4 problems (1) Reference: v6n37 I think that your experience (like others I have heard) is due primarily to problems with your local SUN people rather than a problems with SUN itself. Here at the University of Illinois we have been blessed I think with sales and service people to go way out of their way to be accurate and helpful. I have seen them push up to the highest corporate levels to get us what we need. In almost all cases, the SUN people have been successful in getting us what we need. This has happened because many on campus here have spent many hours developing close relationships with the Chicago SUN office and the local SUN service rep. Further, we have had visits by two SUN Vice Presidents and many other SUN Headquarters experts to help with our special needs. This kind of behavior by SUN has caused the SUN workstation to be the top selling workstation on this campus with numbers well above any other vendor. In fact, we hold up SUN to other vendors we deal with as an example of how to do things. It is common for our SUN reps to admit that they do not know about a certain set of SUN technical information. In all cases, they find someone who can give us a reasonable answer and have spent many hours visiting the campus holding Help Sessions to answer questions at our local SUN user group. Not all users are always happy or satisfied, but this is not for the lack of trying by the SUN people. Incidentally, the Apollo 10000 you refer to at the end of your message is still a "vaporware" product that is not on the market and is still in a research stage according to all the information I have. If you read the trade announcements carefully, you will see this. Further, that machine only obtains its "speed" by a multi-processing mechanism which is not a general purpose system like the SUN-4. Steve Christensen Research Scientist National Center for Supercomputing Applications Director, UIUC/NCSA SUN User Group University of Illinois at Urbana-Champaign steve@spock.ncsa.uiuc.edu ------------------------------ Date: Wed, 30 Mar 88 09:56:20 EST From: weltyc@cs.rpi.edu (Christopher A. Welty) Subject: SUN-4 problems (2) Reference: v6n37 [[ Sun-Spots was cc-ed on this message to "steve@ncsa.uiuc.edu". --wnl ]] We have always been a sun customer. We started with Suns in 1982, I believe, those old tanks called SUN-1s (we still have a few). We have always chosen Sun over other vendors, and at this point Suns here outnumber all other vendors combined. I have no complaints about SUN technical support or with the local sales folk in general. In fact, with the exception of the SUN4 fiasco, we are quite happy with the machines, and are still ordering more. My posting was merely telling the truth about the SUN4, since SUN won't. The machine is NOT ready yet. It shouldn't have been announced if there was only going to be partial support. We feel especially angry at being lied to when we see ourselves as a good and long time customer. We still have not received the second of the ALM2s we ordered, nor the pascal compiler we were promised. Keep in mind that we were assured we could have a complete system available by the start of classes in January. This was part of the reason we chose the SUN4. When that failed we decided to shoot for Spring break (March 28-April 1st). At the time of delivery (February) we were assured that would not be a problem. Since we are in the midst of that week and still do not have all our hardware and software, we will not be able to convert to the SUN4 until the summer. Since that is the case, and since we haven't paid for our SUN4 yet (despite the constant yelling from the SUN accounting office), we may very well send it back if there is a chance of getting another, better machine before the start of classes next Fall. [[ And in a separate note, Mr. Welty says: ]] The saga continues.... And I forgot to stress the most important point here. One of the main reasons we needed to switch to the SUN4 is that we are paying maintenance on the VAX 11/780 we want to replace. This amounts to a good $2K a month. That means we have lost $8,000 so far on the deal. Christopher Welty --- Asst. Director, RPI CS Labs weltyc@cs.rpi.edu ...!rutgers!nysernic!weltyc ------------------------------ Date: Tue, 29 Mar 88 10:19:24 EST From: Chuck Musciano <chuck@trantor.harris-atd.com> Subject: Error messages: MC68881 not available We seem to get erroneous error messages stating: MC68881 floating point not available -- program requires it even though the '881 is installed, and /usr/etc/mc68881version can find and use the device. Some programs get the error, others do not, even though all were compiled with the environment variable FLOAT_OPTION=f68881 set. What could be causing this? Sometimes, exiting and reentering suntools solves the problem. Chuck Musciano Advanced Technology Department Harris Corporation (305) 727-6131 ARPA: chuck@trantor.harris-atd.com [[ This suggestion may make it sound like I don't believe you know what you are doing, but I've seen very experienced people make the same mistake: are you absolutely sure that you are trying to run all the programs on the same (local) machine? I've seen experienced people get very puzzled when the tool they tried failed because they were running it in an rlogin window. If you are, then just ignore me. --wnl ]] ------------------------------ Date: 29 Mar 88 11:35:39 GMT From: Hans van Staveren <mcvax!cs.vu.nl!sater@uunet.uu.net> Subject: CDC Sabre 850 Mb SMD disk working on Sun with Xylogics controller We recently acquired a CDC Sabre 850Mb disk and got it to work using our old friend diag. To spare other people the trouble here is what we did to get it working on our Sun 4/280 with Xylogics 451 controller. First the jumpering of the drive: The sectoring should be set to 68 data sectors, round down method. Table 3-11 in the manual then gives the sector switches: O C C C O C C O O C C C C C for a sector length of 603 bytes. The RUNT jumper should be connected to disable RUNT sector pulses. The B/8 Sector Clock switch should be 8(off). This makes it possible to format the disk. Start diag and answer the questions: Use 1379 data tracks, 2 spare tracks, sector length 603 bytes, disk type 1, interleave 1, 15 heads, 67 data sectors per track. The disk type 1 is not very important, the Xylogics controller can handle four different disk types at a time. The Eagle is type 0, and Sun also has ideas about the other types. As long as you don't mix them with other drives you could pick any type number between 0 and 3. Before formatting type in the bad block info that accompanies the disk. I typed in all defects of length 11 and larger, we had 25 of those. This can be done by the diag-command "sformat". Format the drive, this will take some 4-5 hours including verification. Go to it and good luck. Hans van Staveren Vrije Universiteit Amsterdam, Holland ------------------------------ Date: Tue, 29 Mar 88 14:27:47 EST From: mstan!s2!dpk@uunet.uu.net (Douglas P. Kingston) Subject: Memory modules for Sun 3/60's? Sun appears unable to supply memory modules in a timely fashion. Does anyone know of alternate sources of supply for the memory modules used in Sun 3/60's? -Doug- [[ I doubt that anyone in the US can supply memory modules for 60's in a timely fashion. They are constructed from 1Meg DRAM chips, which are currently only supplied by Japan, who is mad at our government for accusing them of dumping chips on our market (not to mention the weakened dollar driving up the price of Japanese chips). --wnl ]] ------------------------------ Date: 30 Mar 88 04:24:27 GMT From: ut-emx!boyter@sally.utexas.edu (Maj Brian Boyter) Subject: driver for Maxtor worm wanted We are looking at buying a worm optical disk from Maxtor... It comes with software and controller for an IBM-PC... We would also like to use the drive with our SUN workstations... The Maxtor interface is SCSI, so the hardware interface is no problem... Has anyone written a driver, or know of a vendor that sells the driver for a Maxtor drive on a SCSI controller for the SUN workstation??? Thanks in advance Brian Boyter boyter@emx.utexas.edu emx!boyter (804)973-9440 ------------------------------ Date: Mon, 28 Mar 88 21:27:13 EST From: ted@braggvax.arpa Subject: Any problems with calentool (SunOS 2.2..)? Has anyone had any problems with calentool? I compiled it on our Sun2s running 2.2 without much trouble (had to change one #include I think), but after giving it to our secretary, she noticed that it got weird after a while. The sympton is basically that it seems to stack up button events in such a way that you have to click in the bottom subwindow to get it to do the next thing. If you have buttoned ahead, you may have to click in the bottom several times to run through all the buttons you have pushed. We're upgrading to 3.4 soon, so I don't want to spend a lot of time on this if 3.X will fix it. Has anybody seen this behavoir under 3.X (on 3s or 2s)? Or anywhere else for that matter? Thanks, Ted Nolan ted@braggvax.arpa ------------------------------ Date: Tue, 29 Mar 88 09:00 N From: <PETIT@HGRRUG5.BITNET> Subject: VAX vs SUN? We are planning to change a VAX 750 in a VAX-cluster by a VAX 8350, our VAX's are on ethernet. I am curious of a SUM alternative ( We have very little experience with SUN, we have only a SUN 3/50), we plan to use it for SIMULA courses and data base applications running INGRES. I have some questions : 1. How many users can I support on a SUN 3/SUN 4 , how much memory ? 2. Is there any benchmark figures for this two systems ? Thank you Marc Petit University of Groningen Computing Center The Netherlands My address on EARN/BITNET : PETIT@HGRRUG5 ------------------------------ Date: Tue, 29 Mar 88 08:39:24 PST From: kay%sdcc7@ucsd.edu (Don Kay) Subject: vt100 emulator that supports CPR? I am looking for a vt100 emulator for SunView that supports "Cursor Position Report" CPR, in terms of Row and Column. The vt100 escape sequence for CPR is \E[6n where \E is ESC character (0x1b). I know that xterm under X11 supports the sequence, but when I tried Bruce Barnett's vttool it did not support the CPR esacpe sequence on the SUN console. Does anyone know where I might get a vt100 emulator for SunView that supports CPR? [[ Especially handy when you have a heart attack...... Seriously, check the "vt100tool" that was distributed on the Sun User Group tape a few months ago. --wnl ]] ------------------------------ Date: Mon, 28 Mar 88 16:29:35 EST From: chuck@wooglin.scc.com (Charles Williams) Subject: VT100 emulator? HELP!!! I am in need of a VT100 terminal emulator that runs under suntools. I know that SUN sells one which has been ordered, but been delayed by the lack of manuals, and I need one in a pinch. I heard that there is one floating around in the public domain. Can anyone direct me to it?? Thanks in advance, Chuck Williams Contel Federal Systems chuck@wooglin.scc.com [[ The sun-spots source archives has a copy of "vttool". It is available in the files "sun-source/vttool.shar.1" and " "sun-source/vttool.shar.2". 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". There is also "vt100tool" which is (unfortunately) not in the archives. --wnl ]] ------------------------------ Date: Tue, 29 Mar 88 13:02:09 EST From: edsews!cowan@uunet.uu.net (Andy Cowan) Subject: Difference between thin and "thick" wire networks? I need some real facts and figures comparing thinwire and thickwire Sun networks. If this has already been done to death, then I apologize. But, as a novice to networking I'm looking to put together a cheat sheet that quickly explains the differences in simple language that even management types could understand. (Yes, really. THAT simple.) Some things to consider ( but not limited to these ): cost max/min length max/min number of attached devices signal rate data rate other equipment needed to connect sheilding best/worst environment in which to use each ease of installation/cabling anything else that might help Stress the differences using real numbers. Not just 'gut feel' which my lab guys are giving me. Thanks...... cowan@edsews.eds.com Andy Cowan EDS-TSD ------------------------------ End of SUN-Spots Digest ***********************