Date: Sun, 2 Nov 97 19:19:54 EST From: Dwight McKay (The Moderator) Reply-To: Suns-at-Home@net-kitchen.com Subject: Suns-at-Home Digest V10 #38 To: Suns-at-Home-List Suns-at-Home Digest Sun, 2 Nov 97 Volume 10 : Issue 38 Today's Topics: Dual Frame Buffers Fujitsu 9 track tape drive Look for an sbus card, Solbourne, 3/60 memory Problem booting Solbourne machine with error code. Removing RAM chips from Sparc 2 board Suns-at-Home Digest V10 #37 +--------------------------------------------------------------------+ | Submissions: suns-at-home \ | | Requests: suns-at-home-request > @net-kitchen.com | | Archives: suns-at-home-archives / | | WWW Archive access: http://www.net-kitchen.com/~sah | +--------------------------------------------------------------------+ ---------------------------------------------------------------------- Date: Mon, 27 Oct 1997 07:48:49 -0500 (EST) From: David Young Subject: Dual Frame Buffers To: adrian@internode.com.au > Date: Thu, 23 Oct 97 17:25:24 CST > From: Adrian Corston > Subject: Two heads are better than one (SS-I with two monitors) > To: Suns-at-Home@tigger.net-kitchen.com > > I have an SS-I with two cgthree cards, and I figure it's about time to > get the second cgthree card up and running. On boot I see: > > cgthree0 at SBus slot 2 0x0 pri 7 > cgthree1 at SBus slot 3 0x0 pri 7 > > So I'm halfway there. I run X11R4 (eek oh well). When I try to run: > > /usr/bin/X11/X -dev /dev/cgthree1 :1 > > I see this error: > > Fatal server bug! > no screens found > Abort (core dumped) > > A trace shows (in part): > > brk (0xbe8c0) = 0 > gettimeofday (0xf7fffa18, 0) = 0 > fcntl (2, 03, 0) = 2 > fcntl (2, 04, 0x6) = 0 > access ("/dev/cgthree1", 06) = 0 > access ("/dev/cgthree1", 06) = 0 > access ("/dev/cgthree1", 06) = 0 > access ("/dev/cgthree1", 06) = 0 > access ("/dev/cgthree1", 06) = 0 > write (2, "\nFatal server bug!\n", 19) = > Fatal server bug! > 19 > write (2, "no screens found", 16) = no screens found16 > write (2, "\n", 1) = > 1 > fcntl (2, 03, 0) = 6 > fcntl (2, 04, 0x2) = 0 > sigblock (0x20) = 0 > sigvec (6, 0xf7fff944, 0xf7fff938) = 0 > sigvec (6, 0xf7fff8cc, 0) = 0 > sigsetmask (0) = 0x20 > > > So where am I going wrong? I've successfully run: It has been my experience that the typical X server is built to support one frame buffer of a particular class. Modifications are required to support two frame buffers of the same class. (We run X11 and Motif, not Openwindows. I don't know if Openwin supports multiple fb's of the same class. We never tried it) Combinations we use: cgthree0 and cgsix0 Hope this helps --- --- --- --- --- --- --- --- David Young - Eaton Corp. - dwy%etnibsd@uunet.uu.net Beverly Ma dyoung@yocms.bev.etn.com - ------------------------------ Date: Sat, 25 Oct 1997 15:06:55 -0600 From: "Brian Lucius" Subject: Fujitsu 9 track tape drive To: Greetings fellow SAH fanatics, I am having some difficulties getting a Fujitsu reel-reel drive to fly on my 3/280. Upon bootup, the system does recognize the xtc0 card installed with an xt0 device conected, however ROM diags give a 'hard error 16' if you try to test the thing. If you try to use the drive with mt, it just says that the device is offline or not connected... If I run the tests from the control panel of the drive, the beast will fault on step 20 with an error code of b3 20. I believe the VME jumpers are set appropriately (according to Birdsall's FAQ), so I am at a loss at this point. Anyone have any experience with these beasts? Or possibly service manuals or related literature? Regards, Brian brianjl@concentric.net - ------------------------------ Date: Sat, 25 Oct 1997 18:28:33 -0700 (PDT) From: Anthony Talltree Subject: Look for an sbus card, Solbourne, 3/60 memory To: Suns-at-Home@tigger.net-kitchen.com >Considering the amount of circuitry on the board >I would say that it would be impossible to make a S-Bus IPI controller Density and integration increase over time, though: http://catalyst.sun.com/all/Hardware_Peripherals/Interface_Devices___Controller s/TURBOstor_S212_IPI-2_Disk_Controller.html > 64Meg RAM each running Sunos4.1. This machine is slowly > dying and we can't get any maintenance support for it anymore (went > out of business). Solbourne had good stuff and some good folks. Grumman bought their hardware inventory and still services it. http://www.nts.gssc.com/solbourne.html > The machine kept on working in spite of this but just lost a good > 64M RAM allocated to this bad processor Eh? The way I remember the Series 4 & 5 machines is that memory just resided on the Kbus. What sort of machine do you have? > My other question is, if the disk has filled up to 100% due to the > sys logs recording boot up information, will this prevent me from > booting my computer? It shouldn't. Boot single-user, though, so you can cleanly clean up. > (I'm not sure what bandwidth 3/60's get out of their >SIMMS, but it could be more than the VME bus can deliver which is >probably quite a bit less than the theoretical maximum of 40MB/s). I recall that the 3/180 in practice was significantly faster than the 3/50. I've read the claim that this was because of the memory architecture, though the framebuffer could be involved, too. >It didn't matter so much for my server but I also noted that X11R6 under >NetBSD on a CG2 was a lot more sluggish than X11R5 under SunOS on a >BW2. I think the cg2's weren't relatively fast boards, plus remember that any color FB is naturally going to be moving around more data than a mono FB. - ------------------------------ Date: Mon, 27 Oct 1997 17:35:44 +1100 (EST) From: Craig Dewick Subject: Problem booting Solbourne machine with error code. To: Suns-at-Home@tigger.net-kitchen.com > Date: Mon, 20 Oct 1997 14:44:17 -0400 > From: swegerm@baileys-emh5.army.mil (Michael Sweger) > Subject: Problem booting Solbourne machine with error code. > To: suns-at-home@tigger.net-kitchen.com > > Hello, > > I'm having trouble booting a Solbourne dual processor machine with > 64Meg RAM each running Sunos4.1. This machine is slowly > dying and we can't get any maintenance support for it anymore (went > out of business). Here are the problems with it: Solbourne sold their hardware maintenance and support to Grumman Systems Support Corporation sometime in 1993. Take a look at 'http://www.nts.gssc.com/solbourne.html' for more info. Stephen Dowdy of the University of Colorado has taken James Birdsall's ideas from the Sun Hardware Reference and created something like it for Solbourne machines. The URL is 'http://www.cs.colorado.edu/~dowdy/Solbourne'. > - one of the processors died out of the two. > The machine kept on working in spite of this but just lost a good > 64M RAM allocated to this bad processor. It's been running ok for > 2 years now. Is your machine either a 5/502 or a 5e/502? I've got a 5e/502 here and one of my processor boards failed a while ago, so the machine's performance is noticably down. The other one works fine though. > - the video card is slowly dying. At first the characters would > have missing pixels, then garbage then the whole screen would have > a black square hole on it where all the characters would rapidly > disappear into before I could read them. The outside of this box > was still white. Since this was usless I shutoff all output being > sent to console when the Unix system boots. However, the computer > would still boot. I've never used the colour framebuffer card that came with mine, but I did plug it into the system to test it out. I don't have a hi-res ECL mono framebuffer so I can't use the native console (I presume it's hi-res ECL since my lo-res ECL Sun monitor didn't do a thing when I connected it). > - Recently we had a power outage, and I had to do a reboot after > 96 days running. During this time I was trying to shrink the > system log files since they were pretty close to filling up the > disk where /tmp is located. After varying the temperature in the > room from 60-75F ( it previously liked 65F) to see if the > temperature point had changed - made no difference - I removed > and reseated the video card. Now it won't boot to the point where > you are supposed to enter a boot tape (a flashing tape drive light). > Note: it only got to this point if it couldn't boot correctly from > disk. This either worked or didn't work depending on the room temp. I think this is why one of my 5e CPU boards died - it got too hot. Here in Sydney it doesn't get swelteringly hot all that much in summer, but the humidity is a killer since we're right on the Pacific Ocean. That combined with high ambient temps is enough to cause significant problems. > Now, after reseating the video card I can't even boot the computer. > The disks spin up and click (releasing the disk heads I assume) and > trying to read the disk. However, at this point the disk drives > lights go dark and I read an error code on the CPU LED with a > self-test code of 18-45. My solbourne books don't mention what test > this code represents so that I could determine if it is due to the > reseat of the video card. Can anybody shed some light? Try unplugging the keyboard, restarting the machine with a terminal as the console and/or see if you can telnet in from another machine. I've got no documentation at all on my machine, otherwise I might be able to give you some more advice. > My other question is, if the disk has filled up to 100% due to the > sys logs recording boot up information, will this prevent me from > booting my computer? I've never encountered this situation before > so I don't know what would happen. I realize the /tmp directory > is supposed to clear, but if the syslog files even ate up this > available space then the bootup of the kernel may not have any > disk space to work with in /tmp to finish booting up. It never happened to me with SunOS 4.1.3 on real Sun boxes, so it shouldn't happen with OS/MP, but if /tmp does fill up it could well slow booting down a lot. Regards, Craig. PS. While we're talking about Solbournes, I am in need of some Solbourne items: - a 128 meg RAM board to replace the 32 meg board currrently in my series-5 chassis, - another series-5e processor board (or maybe 2 of them) with boot roms as recent as possible, - *very desirable* an OS/MP 1.1C CD-ROM. I'm trying to set up a new hard drive and I need the CD-ROM to copy the prototype boot block code from since it's *not* on the drive that came in the chassis when I received it, so the new drive is all ready to go, but I've got no way to install a boot block on it! Grrrr! 8-) - if I can get an OS/MP 1.1C CD-ROM, I'd like to try and get some series-6 processor boards sometime down the track so I can build a series-6 machine in a spare series-5 chassis. BTW, I've been quoted what seem to me to be very unrealistic prices for the hardware items from a couple of surplus dealers, such as $400 each for series-5e processor boards, and over $1000 (these are US$ prices - not A$ prices!) for series-6 processor boards and the 128 meg RAM boards. If anyone reading this list knows of places that have loads of Solbourne gear to sell (at reasonable prices), please let me know. -- Craig Dewick. Send email to 'cdewick@lios.apana.org.au' Point a web browser at 'http://lios.apana.org.au/~cdewick/sun_ark.html' to access my archive of Sun information and links to other places. For info about Sun Ripened Kernels, go to 'http://www.zeta.org.au/~craig/srk' - ------------------------------ Date: Fri, 31 Oct 1997 22:07:46 -0800 From: editors@kspace.com (Kaleidospace) Subject: Removing RAM chips from Sparc 2 board To: suns-at-home@tigger.net-kitchen.com Hi: We've got a Sparc 2 running SunOS 4.1.4 with 64 megs of RAM. One of the RAM chips appears to have failed judging from a "bad trap" error on boot. However, when we opened the case, we couldn't figure out how to pull out the RAM! Unlike PC motherboards, the Sun SIMM slots don't allow the chips to swing to the side. It looks like you have to pull straight up, but there's no way to get leverage to do so. Is there a strange tool out there that is necessary to remove the chips? Knowing this would save a lot of $$ since we can pop in our own RAM instead of hiring someone to do it. Thanks! Pete Markiewicz [Yes, Pete, there's a silly little tool that has a pin which fits into the] [hole on the end of each SIMM allowing you to pry them loose. My memory ] [vendor gave me one along with my last memory order. Maybe you can get one] [along with a new SIMM? --ddm ] ********************************* Kaleidospace (editors@kspace.com) P.O. Box 341556 Los Angeles, CA 90034 310/399-4349 (voice) 310/396-5489 (fax) Kaleidospace Independent Internet Artists http://kspace.com Kspace's new book on Creating Internet Entertainment http://kspace.com/intertainment Multispace Internet Services http://multispace.com **Join our general newsletter** send email to majordomo@kspace.com with the following message: subscribe kspace-list (followed by your email address) - ------------------------------ Date: Mon, 27 Oct 1997 12:20:53 -0500 From: John DiMarco Subject: Suns-at-Home Digest V10 #37 To: Suns-at-Home@tigger.net-kitchen.com In message <199710250149.UAA06721@tigger.net-kitchen.com> raub@kushana.aero.ufl.edu (Mauricio Tavares) writes: >> AAA is the starting cylinder of the partition. BBB is the size of the >> partition in blocks; CCC/D/E is the size in cylinders/tracks/sectors. >> [tracks and sectors should always be 0 and 0, for historical reasons] >> > How much is a cylinder in blocks? cylinder size in sectors = sectors-per-track x number-of-heads >> if you don't have it already, scsiinfo can interrogate any drive and >> spit out a format.dat-style entry for it. get it. use it. love it. This is true only if the drive is labelled using the settings that scsiinfo generates. It's possible to label a SCSI drive with any number of different values for sector-per-track, number-of-heads, etc. Use format to find out what settings your disk label currently contains. > I must have an old copy of scsiinfo because mine won't give me >that... unless I do not know how to ask it for that. =( The -f option (old versions of scsiinfo) or -r option (new versions of scsiinfo) will do the trick. -F generates an actual format.dat entry, which you can also use to compute the number of sectors per cylinder. FYI, the latest version of scsiinfo (currently 4.4) is available from ftp://ftp.cs.toronto.edu/pub/jdd/scsiinfo/ Regards, John -- John DiMarco Office: SF2101 CSLab Systems Manager Phone: 416-978-5300 University of Toronto Fax: 416-978-1931 http://www.cs.toronto.edu/~jdd - ------------------------------ End of Suns-at-Home Digest ******************************