Sun-Spots-Request@Rice.edu (William LeFebvre) (11/07/88)
SUN-SPOTS DIGEST Saturday, 5 November 1988 Volume 7 : Issue 6 Today's Topics: Re: Problems with pc/nfs + IBM worm disk Re: /var/spool/lpd/err files incorrect logfile handling Re: UUCP, bug or feature Comments (flame) about lack of 3/50 memory upgrade vt100tool from mod.sources vol 6 Problem with TTY windows yp: problems with the outside world babbling while booting Postscript previewer for Sun? Molecular modeling packages (graphics) on Sun-3 or Sun 386i? Cheap LaserWriter Alternative for Sun 386i? Need Sun/ethernet performance info 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: 28 Oct 1988 1257-EDT From: hsw@tycho.arpa (Howard Weiss) Subject: Re: Problems with pc/nfs + IBM worm disk We have had problems along similar lines installing the 3Com 3C501 ethernet cards in some of our PCs and then getting PC/NFS to work with the system. as it had before installing the Ethernet interface. Turns out, we had two interfaces at the same vector (both the 3Com card and a Frontier serial board). When we moved the vector jumper on the 3Com card to another unused vector and then started PC/NFS with that new vector in the startup file, the problems went away. The symptoms we were seeing were that PC/NFS started up OK and we could access the network, but when accessing the Frontier board the PC would lock up solid and would have to be re-booted. Howard Weiss ------------------------------ Date: 28 Oct 88 19:58:40 GMT From: bparent%sdcc7@ucsd.edu (Brian Parent) Subject: Re: /var/spool/lpd/err files Reference: v6n259 In issue v6n259 of comp.sys.sun grunwald@m.cs.uiuc.edu (Dirk Grunwald) writes: >We're using TranScript 1.1 on SunOS 4.0 to drive a DEC ScriptPrinter. 2.1-^ (He wrote to me later to correct this.) >The TransScript documentation (and the SunOS documentation) claim that >error messages get put into the file specified by ``lf'' in the printcap >entry. This is, in part, true: the output from psbanner gets put there. >However, there's another file that gets created in the spool directory, >labeled errABCDE where ABCDE is probably a process number. This file >contains the useful error messages from the PostScript printer (i.e. all >the messages that tell you why your print didn't come out of the printer). >I can find no incantation that makes these messages go to the log file. >Has anyone else run into this problem? Any clues? I first saw this problem when we updated to BSD 4.3 on our vaxes. It turns out to be a combination of newer 4.3 lpd, and Adobe's Transcript pscomm filter. Judging by Dirk's problems, Sun is using the 4.3 version of lpd in their SunOS 4.0. (verifications anyone?) I've sent a bug report to Adobe. Here are the relevant parts: Index: transcript/src/pscomm.c Subject: incorrect logfile handling Description: 4.3BSD lpr sets up stderr so that it writes to a logfile as defined in the printcap entry per printer. In older versions (4.2 and earlier) it left stderr that way for the filters to use. Now, however, lpd (actually in printjob.c) sets up a temporary file for the stderr descriptor in the filter. I'm guessing, after looking at the sources, that they (Berkeley folks) want the logfile as defined in the printcap to be used only for lpr/lpd stuff, and not the filters that get exec'ed by lpr/lpd. The temp file that receives all output aimed at stderr from the filter, gets unlinked as soon as the filter goes away. Therefore, if the user wants a record of errors, the filter (pscomm) will need to close stderr, and reopen it, (possibly after aiming it at a different log file). Fix: I don't think Adobe would appreciate me distributing their source (or even a part of it), but there is a fairly simple work around. Use a .options file in the printer's spool directory that looks something like this: # .options JOBOUTPUT=/tmp/printer1_diags VERBOSELOG=1 export JOBOUTPUT VERBOSELOG log=/usr/spool/lpd/lpd-errs test -r $JOBOUTPUT && cat $JOBOUTPUT >> $log test -w $JOBOUTPUT && cp /dev/null $JOBOUTPUT This works in sort of a two pass way. The .options file is executed prior to starting pscomm. The first print job to go through will set up the JOBOUTPUT variable for pscomm, and the second job to go through will cat that output to the lpd-errs file. ------------------------------ Date: 28 Oct 88 03:12:04 GMT From: heather@cs.ucla.edu Subject: Re: UUCP, bug or feature Reference: v6n266 I would like to publically thank all the people who took the time to respond to my question about a uucico bug (I'll call it a bug now) that results in the auto-answer bit of a modem being turned on. Several of the responses were quite detailed and I believe could be of help to others, so I have included them here. Other respondents touched on the same issues as those included, so I have omitted them. The original problem: >The problem is that when uucico runs, it sets the auto-answer bit of the >modem to on. This condition is not reset when uucico ends.... Two fine responses: >From: Preston Mullen <mullen@nrl-css.arpa> To get around uucp's setting the modem's autoanswer bit, patch the dial string in /usr/lib/uucp/uucico: ATV0Q0E0S0=1S2=255S12=255 ^ change 1 to 0 [[ This was posted in v6n273. --wnl ]] >From: vsi1!lmb@ames.arc.nasa.gov (Larry Blair) ... If you begin the phone number in the L.sys file with a alphabetic string, the value for that string in the L-dialcodes file is prepended to the number that follows. More importantly, if the replacement string begins with an "S", the Hayes dialer prepends only "AT" rather than the usual "ATDT".... [[ This was posted in v6n274. --wnl ]] >From: Joseph Kwan <rabbit@hera> I would double check the modem you were using for the proper switch settings. Most modems have a switch that will always put the modem in auto answer mode. "ATS0=0" can be used to disable that but that would be temporary until the modem is reset (which is usually when DTR is toggled). But, if there aren't any switches or whatever you can just add the "ATS0=0" command to the uucico chat script. Something like: system Any DIR baud device "" ATS0=0 OK ATDT... (that assuming your modem is return status codes in English). I've noticed that Hayes command compatible 2400 baud modems have the ability to save settings in software (I think it's the AT&W command), so typing "ATS0=0" then "AT&W" might turn off auto answer and save that as a default setting for when the modem is reset. On 1200 baud modems there's normally switches. Joseph Kwan Thanks also to the following respondents: Herb Barad <barad@bourbon.ee.tulane.edu> auspex!guy@uunet.UU.NET (Guy Harris) esosun!sol!keith@ucsd.edu (Keith Pilotti) david@Sun.COM (David DiGiacomo) ames!harvard!yale!hsi!stepstone.com!aad (Anthony A. Datri) ho@tis-w.arpa (Hilarie K. Orman) Sergei A. Gourevitch <asg@space.mit.edu> ------------------------------ Date: Fri, 28 Oct 88 08:19:40 PDT From: brent%sprite.Berkeley.EDU@ginger.berkeley.edu (Brent Welch) Subject: Comments (flame) about lack of 3/50 memory upgrade I got a little postcard from Helios about a 3/50 upgrade.... I have no connection to Helios, etc. etc. and haven't called them yet to find out about pricing. It's about time, eh? *flame on* I've heard that it was marketing/managment that inhibited a Sun upgrade path for the 3/50, not the engineering of the box, and that it's possible (technically) to upgrade memory or to upgrate to a 3/60. *flame off*. Our experiences are that more memory is the best way to make your workstation go faster for the simple reason that you don't page. 4 Megs is about enough room for the operating system and the window system, and anyting interesting on top of that causes paging... Spend your money on more memory before adding a local disk. honest. Brent Welch University of California brent%sprite@ginger.Berkeley.EDU (Sprite is alive and well, in house anyway) ------------------------------ Date: Tue, 25 Oct 88 19:24:38 BST From: Bruce Arden <mcvax!tcom.stc.co.uk!arden@uunet.uu.net> Subject: vt100tool from mod.sources vol 6 I recently extracted vt100tool from a comp.sources.unix archive but found that it doesn't compile (presumably it was for an old version of SunOS). Has anyone got the patches (or even a complete new copy) to get it to work on SunOS 3.5 or 4.0. I've got vttool and that works, but I was hoping for something that could do the vt100 line drawing and double height/width characters. Thanks, Bruce Arden arden@tcom.stc.co.uk ...!ukc!stc!arden STC Telecommunications, London N11 1HB. +44 1 368 1234 Ext 2795 ------------------------------ Date: Fri, 28 Oct 88 14:55:07 EDT From: Chuck Musciano <chuck@trantor.harris-atd.com> Subject: Problem with TTY windows I have a tty window and I start an application in it. The first thing the application does is call termcap routines to get the window size. The problem is that I don't think that the values are correctly set until the window is actually displayed via a call to window_main_loop(). Until then, some bogus values are returned. The application often executes far enough to get the values before I can call window_main_loop() and get the stuff on the screen. Has anyone experienced this problem? Does anyone have a solution to this problem? Note that I set the window values when I do the window_create(), but the termcap stuff is not updated until I clal window_main_loop(). Chuck Musciano Advanced Technology Department Harris Corporation (407) 727-6131 ARPA: chuck@trantor.harris-atd.com ------------------------------ Date: Thu, 27 Oct 88 11:47:28 +0100 From: mcvax!ifi!ikaros!gutorm@uunet.uu.net Subject: yp: problems with the outside world babbling while booting Dear Spots, We have a net of Suns connected with the campus backbone through a lan-bridge (iso-lan primary bridge). We run SunOS 3.5 on all 10 machines (3/50s and 3/60s), and are using yp with only one yp-server. If I turn off the lan-bridge, I am able to boot all Suns: no problem. I can then turn on the bridge, and all is fine. However, if the world can send us goodies during boot-time, yp will flip: Even the yp-server will not be able to talk to the yp-server. Because of this, we cannot NFS mount filesystems, and the machines will hang, waiting for the yp-server to respond. (yp: server not responding for domain "solarium"; still trying . At least I've learned to love _that_ one ... ) This problem looks similar to another one discussed here earlier, but we run 3.5 and not 4.0 . The machines here will boot, but yp will not work. A ypwhich will give 'Domain solarium not bound', if the outside world was babbling during boot-time. It seems that the problem started to occur after the ip-addresses and the netmask was changed for everyone on our backbone net (they are now typically 129.240.*.*). And yes, there is all kind of software running on the net ( but no ultrix bridge I know about ). I have tried different broadcast-addresses and netmasks without success. I've also tried to put the clients IP addresses direct in its ifconfig line in rc.boot, as suggested by Steve Miller. All help will be more appreciated than you might believe. gutorm thomas hogasen | Internet: gutorm%ikaros@ifi.uio.no University of Oslo | Bitnet: gutorm%ikaros@oslo-vax.arpa Inst. of Mathematics | Box 1053 Blindern | N-0316 OSLO | NORWAY | Tlf: +47 2 455879 ------------------------------ Date: Fri, 28 Oct 88 15:33 EST From: MARKIEWI@CSHLAB.BITNET Subject: Postscript previewer for Sun? Does anyone know of a Postscript previewer for the Sun that runs on the workstations and let you preview your document in a separate window? [[ Won't "NeWS" do that for you? It is, after all, Postscript-based. --wnl ]] Please E-mail responses directly to me if possible. Thanks in advance. Peter Markiewicz Cold Spring Harbor Labs Bungtown Road Cold Spring Harbor, NY 11724 MARKIEWI@CSHLAB.Bitnet ------------------------------ Date: Fri, 28 Oct 88 16:34:25 EDT From: shenkin@cubsun.bio.columbia.edu (Peter Shenkin) Subject: Molecular modeling packages (graphics) on Sun-3 or Sun 386i? If anyone knows of molecular modeling packages that will run on a Sun-3 or a Sun 386i, I'd like to hear about them. (I don't think I'm interested in the ms-dos packages I could run in an ms-dos window.) Peter S. Shenkin, Department of Chemistry, Barnard College, Columbia University New York, NY 10027 Tel: (212) 280-5517 (work); (212) 829-5363 (home) shenkin@cubsun.bio.columbia.edu shenkin%cubsun.bio.columbia.edu@cuvmb.BITNET ------------------------------ Date: Fri, 28 Oct 88 17:02:44 EDT From: shenkin@cubmol.bio.columbia.edu (Peter Shenkin) Subject: Cheap LaserWriter Alternative for Sun 386i? I've just received a Sun 386i (the 25MHz version), and love it! My Fortran number-crunching benchmarks come in at just twice Sun 3-180 speeds, where the 3-180 was also running SunOS 4.0 with the new Sun Fortran. (This exhibited no improvement over 3.x Fortran speeds, which were Ultrix-VAX11-780-like). Now for the question: soon Sun will have LaserWriter support for SunOS 4.0, and I expect to purchase it. I also want to get an Apple LaserWriter, which at academic discounts will come in at something more than $4k. A friend pointed out an intriguing alternative. You can, I believe, buy an HP Laserjet for about $1200, and an AT card with Postscript on it that talks to the Laserjet for about $800, giving the moral equivalent of a LaserWriter for about $2k. In fact, it should be even better, because this card has enough memory on it to buffer an entire bitmapped page. Since the 386i is an AT-bus machine, I can't think of any reason this wouldn't work, unless I would have to write a device driver or something to get UNIX to talk to the proper slot with the proper protocol. I'd rather not have to do this. Has anyone out there tried this sort of print arrangement, or have a better idea than I how easy this arrangment would be to make work? Also on the same subject, someone has told me that NEC has a 300dpi PostScript LED printer for under 4k. Does anyone have any experience with these? Finally, one area I don't wish to neglect is the niceness of the fonts. This is where the LaserWriter really shines over some of the competition, such as the Talaris. Does anyone know whether this $800 card (I don't know the name; sorry!) has the Adobe fonts on it? And likewise for the NEC? Peter S. Shenkin, Department of Chemistry, Barnard College, Columbia University New York, NY 10027 Tel: (212) 280-5517 (work); (212) 829-5363 (home) shenkin@cubsun.bio.columbia.edu shenkin%cubsun.bio.columbia.edu@cuvmb.BITNET ------------------------------ Date: 28 Oct 88 18:40:12 GMT From: mrr@amanpt1.zone1.com (Mark Rinfret) Subject: Need Sun/ethernet performance info Hello. I don't usually follow the discussions on networking, so please forgive me if the following request includes items that have been recently or frequently discussed on Usenet. I am interested in obtaining any information available related to the performance of ethernet-based Sun workstation networks. If any of you have done any performance analysis or simulations for various configurations (disk/diskless workstations, gateways, clustering, interconnection with other machine types, etc.), I would be most appreciative if you'd email your results to me. Please include any specific do's and do-not's that your experience has revealed. I realize that I'm asking for a lot, but any info I can gather will be very helpful for an upcoming project. Thanks very much, Mark Mark R. Rinfret, mrr@amanpt1.ZONE1.COM | ...rayssd!galaxia!amanpt1!mrr HyperView Systems Corp. Home: 401-846-7639 28 Jacome Way Work: 401-849-9390 x301 Middletown, RI 02840 Hypermedia R Us! ------------------------------ End of SUN-Spots Digest ***********************