Date: Sun, 17 Mar 96 09:07:27 EST From: Dwight McKay (The Moderator) Reply-To: Suns-at-Home@net-kitchen.com Subject: Suns-at-Home Digest V9 #10 To: Suns-at-Home-List Suns-at-Home Digest Sun, 17 Mar 96 Volume 9 : Issue 10 Today's Topics: 4/470 and SCSI-3? Fans fan speeds Free Sun 3-50, needs work Keys for Type 5 Keyboard PPP2-2 too slow ??? serial port patches Sun-4 Fans (2 msgs) Sun 3/80 PPP Sun 4/460 (2 msgs) Sun Hardware Suns-at-Home Digest V9 #9 Sun Sparcstation ELC ttya serial link locks up during kermit/tip downloads +--------------------------------------------------------------------+ | 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: Tue, 12 Mar 1996 15:01:33 -0800 From: gary@ranchogrande.mce.com Subject: 4/470 and SCSI-3? To: Suns-at-Home@tigger.net-kitchen.com Bill, You have probably resolved this by now. If not.. This might be a SCSI revision issue. You need a 501-1236 SCSI board that is Rev 8 or higher. You can identify this version of the board easily. First, it has an engineering change that adds a fuse for TERMPWR. If you have an ALM-2 board in the system, the SCSI needs to be rev 9 or higher for a 4/4xx system. If you add a terminator to an old board, it will not have power supplied by the SCSI host. In most cases, it will fail unless the device is providing TERMPWR. Gary W. Cook Director of Technical Services Minicomputer Exchange, Inc. 610 North Pastoria Avenue Sunnyvale, CA 94086 Email: gary@mce.com Tel: 408-733-4400 Fax: 408-733-8009 - ------------------------------ Date: Wed, 9 Mar 1994 23:35:56 -0500 From: wkearney@attach.net (Bill Kearney) Subject: Fans To: Suns-at-Home@tigger.net-kitchen.com, >I've seen a note in the past, I believe it was on comp.sys.sun.admin or >.hardware, about fans. I wish I had the original to quote; but for what its >worth: the poster stated that if you replace a 12 volt fan with a 24 volt >model (I imagine you should match the amperage rating) the fan will run >quieter and last longer. So you end up running the 24volt units slower? Is this a good thing? I'm going to consider replacing the existing units with brand new fans. I just replaced one in a shoebox and it's a LOT quieter. It seems fans technology has improved as well. I suppose as long as the CFM ratings are the same it doesn't really matter. Unless the wattage is higher; that might affect things. Bill - ------------------------------ Date: Sun, 10 Mar 96 00:41:24 PST From: perryh@pluto.rain.com (Perry Hutchison) Subject: fan speeds To: McClinton_Rick@tmac.com > Date: Mon, 4 Mar 1996 09:53:33 -0500 > From: McClinton_Rick@tmac.com > > I've seen a note in the past, I believe it was on comp.sys.sun.admin or > .hardware, about fans. I wish I had the original to quote; but for what its > worth: the poster stated that if you replace a 12 volt fan with a 24 volt > model (I imagine you should match the amperage rating) the fan will run > quieter and last longer. It will also run slower (that's why it's quieter) and move less air. You may be able to get by with this if you don't have a lot of cards in the cardcage (or a lot of RAM in the SIMM slots, or whatever). - ------------------------------ Date: Sat, 9 Mar 1996 18:27:02 -0500 (EST) From: Joseph C Fineman Subject: Free Sun 3-50, needs work To: suns-at-home@tigger.net-kitchen.com 10-yr-old Sun 3-50, free, if you haul it away yourself from my house in Brookline, MA no later than Friday 22 March. It has an extra disk drive, I forget how big. Everything works but the monitor. Joe Fineman jcf@world.std.com 239 Clinton Road (617) 731-9190 Brookline, MA 02146 - ------------------------------ Date: Sun, 10 Mar 1996 23:55:17 -0800 From: Michael Furman Subject: Keys for Type 5 Keyboard To: suns-at-home@tigger.net-kitchen.com I just replaced my Type 4 keyboard with a Type 5 that I got at a flea market for $10. The only thing is that it is missing two keys, the F5 key and the Close Key. If anyone has a dead Type 5 keyboard they can pop a few keys off of, please let me know! Thanks, --Mike - ------------------------------ Date: Tue, 12 Mar 1996 12:09:00 -0500 From: "fadi (f.) mehanna" Subject: PPP2-2 too slow ??? To: Suns-at-Home@tigger.net-kitchen.com Hi, this is in regard to using ppp-2.2. Is it my immagination, or ppp is too slow ??? Starting an xterm from work takes almost two minutes. I have a 28.8k Microcom modem ??? I have X11R5 installed locally, and most other utilities. But still when using ftp, it takes a long time to download a file .. I havn't benchmarked the speed, but I can tell it is very slow, most ftp speeds are 1 - 2kbps ???? Is there some king of magic configuration to use ?? I am allready using bsdcomp and mtu 296 to try to speed up things ... Thanks for any pointers, - ------------------------------ Date: Mon, 11 Mar 96 01:34:29 -0500 From: Kevin Kim Subject: serial port patches To: Suns-at-Home-List@tigger.net-kitchen.com Hi, I'm trying to run a 28.8k modem off my sun 3/60 serial port. Now I was told by someone else that they are running their serial ports at 38.4k. I've never been able to set my ports to anything above 19.2k. It's been suggested that I need to install a patch to my OS (SunOS 4.1.1U1). Anyone know what patch(es) I need to install? Thanks, -kevin - ------------------------------ Date: Sat, 9 Mar 1996 14:55:23 -0500 (EST) From: gregc@pm-tech.com (Greg Cronau) Subject: Sun-4 Fans To: wkearney@access.digex.net You wrote to the suns-at-home-list: > Date: Sun, 03 Mar 1996 17:23:14 -0500 > From: Bill Kearney > Subject: Temperature controlled fans? > To: Suns-at-Home@tigger.net-kitchen.com > > Well, now being the owner of both a 4/330 and a 4/470 I've suddenly > become concerned with noise levels. Criminy! Are these things LOUD or > what? If I hadn't lugged it's 12 slot butt up a flight of stairs I'd > think this thing was preparing to FLY when I turned it on. > > Anyway, I'm not running an abundance of cards in these things. Mostly > just CPU, 1 RAM board and a SCSI controller. I will eventually add more > RAM and thus the 12 slot cages to start with. > > [...] > > I have no where else that's convenient to stash these monsters so > anything I can do to quiet them is necessary. The units are tolerable > when only 1 top and 1 bottom fan are running. I wouldn't think of > running with no fans. I actually tried this and it gets funny when the > RAM overheats. Probably not a *good* thing. *Definately* not a good thing! I just finished doing a test of a similiar machine with just a 4/200 CPU, 1 32meg memory board and a SCSI-3 controller to determine it's power consumption. It uses 320Watts. DO NOT RUN THIS SETUP WITH JUST 2 FANS!!! In fact, one of Sun's early modifications to the 12-slot X/X60 chassis was a fan tray upgrade that replaced the older 4 fan tray with a 6 fan tray. They were running into overheating problems when people upgraded older 3/160 systems to 3/260's. A 4/400 CPU runs ever hotter. You do have a sparsely filled chassis, so you could probably get away with turning off 1 of the 2 top fans, and *maybe* 3 out of the 6 bottom fans, but I wouldn't go further. A 4/400 CPU is still fairly expensive to replace. Greg Cronau gregc@pm-tech.com - ------------------------------ Date: Sat, 9 Mar 1996 15:33:07 -0500 (EST) From: Bill Kearney Subject: Sun-4 Fans To: Greg Cronau On Sat, 9 Mar 1996, Greg Cronau wrote: > > Well, now being the owner of both a 4/330 and a 4/470 I've suddenly > > become concerned with noise levels. Criminy! Are these things LOUD or > > what? If I hadn't lugged it's 12 slot butt up a flight of stairs I'd > > think this thing was preparing to FLY when I turned it on. I tried a few less fans and it doesn't affect the noise appreciably. Granted, I'm now probably used to the racket so it bugs me a lot less. - ------------------------------ Date: Sat, 9 Mar 1996 20:00:49 -0500 (EST) From: Mike Pepe Subject: Sun 3/80 PPP To: Dwight McKay Hi there. I'm wondering if there's anyone out there who has gotten PPP running on a 3/80. I've been fooling with DP 4.0 but haven't had much success, it seems far more complicated than it need be. I'm aware the 3/80 serial ports are brain-damaged in some way, but I've gotten terminal emulation software to work and I seem to be able to send and receive even large files without any problem. The system is a 3/80, 16 megs RAM, 900 megs disk space, 150 meg tape drive. Running SunOS 4.1.1U1, with X11R6 and gcc 2.6.2 - ------------------------------ Date: Sat, 9 Mar 1996 14:38:35 -0500 (EST) From: gregc@pm-tech.com (Greg Cronau) Subject: Sun 4/460 To: wkearney@access.digex.net You wrote in the sun-at-home-digest: > Date: Sun, 03 Mar 1996 11:44:24 -0500 > From: Bill Kearney > Subject: 4/470 and SCSI-3? > To: Suns-at-Home@tigger.net-kitchen.com > > I've got a 4/400 motherboard in a 12 slot 3/260 cage. This sort of > qualifies it as a 4/470. I'm attempting to use a SCSI-3 controller with > it. I can't get it to boot anything. It's got 4.1.1 boot ROM and no > OpenBoot stuff; ie no probe-scsi. > > I've taken a known good CD ROM and the Solaris 2.4 CD and tried > b sd(0,30,1) > with the settings on the SCSI for the 1st controller I get nothing. If > I set the controller to the 2nd device I get a watchdog reset after it > flashes the drive a couple times. This is not a scsi addressing issue. > I've got the devices terminated properly and set an HD as ID0 and the CD > as ID6. I've also tried sr(0,30,1) ,2) and damn near everything else. > > What must I do to get this to function? > > Bill I run the same configuration. I call it a 4/460, but Sun never did. :-) I've been running such a setup for at least a year now with few problems. However, I run SunOS 4.1.3_U1, not Solaris. I'm not sure if you will be able to use Solaris as they removed support for a number of VME devices. Also, are you aware of the DMA and Interupt chain jumpers that have to be installed on the back side of the motherboard for each empty slot? If you don't have those jumpers installed, you won't be able to access the SCSI controller. If you don't know what I'm talking about, then they are probably your problem. Write me and I'll explain further. Another thing to keep in mind, if you decide to go to SunOS 4.1.3_U1, be aware that Sun screwed up several of it's patches for the Sun4(not the sun4c or sun4m) archetecture beginning about January 1995. We found this out the hard way and had to patch a number of sun's patches. I have fixed versions available if you decide to go that route. Greg Cronau gregc@pm-tech.com - ------------------------------ Date: Thu, 14 Mar 1996 07:28:10 -0500 (EST) From: gregc@pm-tech.com (Greg Cronau) Subject: Sun 4/460 To: wkearney@access.digex.net (Bill Kearney) Bill Kearney > >Thanks for the reply. The nearest Sun designation for a 12 slot unit >with a 4/400 CPU was the 4/470. At least according to the Field >Engineers Handbook. Just to let you know: The Sun 3100, 3200, and 4200 CPU's were typically put into an older style of 12 slot deskside chassis and called 3/160, 3/260, and 4/260. The x/x60 designation applied to the older cabinet. When those boards are placed into the older style Rack-mount chassis they were called 3/180, 3/280, and 4/280, x/x80 implies the older rack-mount. When the 4/470 and 4/490 series came out. Sun had designed a much newer set of chassis. These chassis used a new type of backplane, power supply, and internal shelving for the mounting of SCSI or IPI drives. A x/x70 implies the *new* style 12 slot deskside chassis, and the x/x90 implies the new style rack mount chassis. The 3400 CPU was also installed into these newer cabinets I believe, but 3/470 and 3/490 systems are rare. Technically, installing a 4400 CPU in an old-style 3/260 chassis creates a 4/460 system(although Sun never had that as an official designation), but if you ever have to deal with Sun, or any other company, for any kind of tech support or repair, if you tell them you have a 4/470 system it will result in alot of confusion, because that's not what you have. >I must have the VME slot settings correct since it functions. Unless >there's something about them that affects the use of the slot 7 pass >through. If you unscrew the 2 screws holding the top of the power supply in place and hinge down the front panel, you will see sets of jumpers for each slot near the top of the backplane. For each slot there are 5 jumpers. They are grouped as 1 group of 3 and 1 group of 2. All jumpers in the group-of-3 should be installed for all 12 slots. For empty slots, both jumpers in the group-of-2 should be installed, for filled slots, the jumpers should be removed. However, the memory boards don't use these signals, so it's simpler to just install the jumpers on the 5 slots to the right of the CPU and forget about them. Greg Cronau gregc@pm-tech.com - ------------------------------ Date: Thu, 14 Mar 1996 07:10:21 -0500 (EST) From: gregc@pm-tech.com (Greg Cronau) Subject: Sun Hardware To: rvacante@verinet.com Hi, You listed the following items in the Suns-at-home mailing list: > I also have an 501-1102 RAM board that won't work on my machine. (But will > work on a 3/2xx, 3/4xx, 4/2xx.) I'd like to sell this (unless someone can > tell me how to get this working with my 3/160.) You can't. The 3/100 CPU board uses parity memory and the 3/200 CPU uses ECC memory. Can't put 3/200 memory in a 3/100 and vice versa. Those 8meg boards are currently worth about $40.00. I've got a big stack of them, don't need any more. > Finally, I have some "mystery boards". I can only guess what they do. Can > anyone provide any info about: > Solflower SF-375MEM (Memory board with simm sockets. How to configure?) No idea on that one. > Xylogics 7053 (???) That's a higher speed SMD interface. It's capable(depending on the rev) of doing SMD-H at 16mbit/sec. Not mbytes/sec. What would you want for this? I might be interested if the price is right. > Dawn 9u400/6u-5 holding a Ciprico RF - 3500 (???) Also an SMD Disk controller. Should be capable of doing SMD-E. I'm also interested in this. Quote me a price. > 501-1054 holding a mystery board..two 25 pin "data" connectors, and > one 25 pin "command" conn., and one 37 pin "command". > labelled 301-160-600 That sounds like a good old Xylogics 451 SMD controller, in a Multibus to VME adapter frame. I have plenty of these. I don't need another one. They're worth about $25.00. As for trade. I have alot of Misc Sun VME hardware laying around. Are you looking for anything in particular? Greg Cronau gregc@pm-tech.com PS: For more Sun info that you can probably handle, do a finger of "faq@case.com" and it will send you the current comp.sun.sun.hardware FAQ. Warning: it's currently about 650K, so you want to redirect the finger to a file. - ------------------------------ Date: Sat, 9 Mar 1996 21:20:28 +0100 (MET) From: Wilko Bulte Subject: Suns-at-Home Digest V9 #9 To: Suns-at-Home@tigger.net-kitchen.com > > Finally, I have some "mystery boards". I can only guess what they do. Can > anyone provide any info about: > Solflower SF-375MEM (Memory board with simm sockets. How to configure?) > Xylogics 7053 (???) That is a SMD disk controller. > Dawn 9u400/6u-5 holding a Ciprico RF - 3500 (???) Guess: Ciprico Rimfire. Some kind of a disk controller (faint recollection here..) > 501-1054 holding a mystery board..two 25 pin "data" connectors, and > one 25 pin "command" conn., and one 37 pin "command". > labelled 301-160-600 Sounds like another SMD controller, now probably a Xylogics 450 or 451 If you have any SMD disks use the XD7053, that is the better of the SMD controller.s Wilko - ------------------------------ Date: Sun, 10 Mar 96 16:50 PST From: bkis@nanaimo.island.net (Jonathan Thornburg) Subject: Sun Sparcstation ELC ttya serial link locks up during kermit/tip downloads To: suns-at-home@tigger.net-kitchen.com I'm having (what appears to be) trouble with serial-port flow control on a Sun 4/25 (Sparcstation ELC). I'm using SunOS 4.1.3, with Sun's Jumbo TTY patch (#100513-04) applied. I'm using a Telebit WorldBlazer modem (14.4K V.32bis modulation, V.42 error control, V.42bis compression) attached to the Sun's ttya serial port, which is set for 38400 baud, 8 data bits, 1 stop bit, no parity, and RTS/CTS flow control. I've followed the setup advice of both the Celeste FAQ (pointer in the comp.sys.sun.admin FAQ) and the Telebit setup guides from their ftp archives. (These recommend basically the same setups.) ttysoftcar reports /dev/ttya is n and the eeprom settings are ttya-rts-dtr-off=false ttya-ignore-cd=true ttya-mode=38400,8,n,1,h I've experimented with both tip, using the /etc/remote entry hardwire|modem|cua0:\ :dv=/dev/cua0:br#38400:p8:hf:nt:el=^C^S^Q^U^D:ie=%$:oe=^D: and kermit, using the .kermrc entries ; ; parameters for the kermit file transfer protocol ; set file type binary set file names literal ; I always call unix systems ==> no conversions set block 3 ; rigorous error-checking set send packet-length 2048 set receive packet-length 2048 set window 10 ; sliding-window size for kermit protocol ; ; parameters for the modem ; set modem unknown ; prevents kermit from tryint to be "smart" ; and mucking with all my S-register settings set dial dial-command ATD%s\{13} set dial display on ; watch dialing progress set dial modem-hangup off ; hang up by dropping dtr, rather than ath0 set dial timeout 60 ; ; parameters for the computer <--> modem serial link ; set line /dev/cua0 set speed 38400 ; this speed should be considerably faster than ; the modem <--> modem modulation-protocol speed, ; in order for the modem <--> modem automatic ; data compression to work effectively set flow rts/cts set terminal bytesize 8 set parity none Using either tip or kermit, everything works fine when interactively logging on to a remote system and transferring small files in either direction. However, using either tip or kermit, trying to download larger files from the remote system to the Sun, causes the connection to lock up after anywhere from 20K to 500K bytes have been transferred. I can see no pattern in how much data I get before a lock-up. When in the locked-up state, no data will flow in either direction, and the modem will *not* respond to its (pause) +++ (pause) escape signal. The only way I've found to un-lock-up the connection is to exit tip/kermit, causing the Sun to drop RTS. If I then start up (a new) tip, the modem responds normally to its escape signal, and reports nothing wrong with the connection, which is still up (modem off-hook, carrier detected, etc). Alas, I can't get any responce from the remote tip/kermit at this point, so I'm forced to break the connection (and let the remote system kill the remote processes). When I manage to download a file without the connection locking up, there's no corruption of the data, even when I use tip (which has no error correction of its own). (When I use kermit, it reports no packet retransmits.) When the connection does lock up, there are *no* console or 'dmesg' error messages, and in particular I haven't seen any reports of zs silo overflows. Using either tip or kermit, everything works fine when *uploading* large files from the Sun to the remote system. I've successfully sent a 1.2 MB tar.gz.uu uuencoded gzipped tar archive using both tip and kermit. Kermit reported *no* packet retransmissions. Watching the modem status lights during these transfers, the Sun --> modem flow control seems to work fine: DTR = off RTS = on continuously CTS = alternates approx 0.5sec off, 0.5 sec on SD = same as CTS RD = off (tip) or brief blink approx each 0.5 sec (kermit) [[The V.42bis compression can probably squeeze out the uuencoding redundancy in the data stream (a factor of 4/3), but presumably it can't further compress the underlying gzip archive, so the net data rate should be about 4/3 * 14.4K = about 19.2K bits/sec, which matches the observed ~50% duty cycle for the 38.4K bits/sec Sun --> modem serial link. The modem --> Sun data for kermit is just what would be expected for kermit's packet acknowledgments.]] When either uploading or downloading files, with the Sun running multiuser but otherwise idle except for an xclock, an xmeter, and the usual system daemons, vmstat reports tip kermit 3000-3500 1700-1800 interrupts/second 1800-2200 100-200 system calls/second 9-13% 1-3% user cpu time 50-65% 2-5% system cpu time This suggests that the Sun *is* able to handle the data rates, with a little margin for tip, and ample margin for kermit. I've tried lowering the Sun <--> modem baud rate to 19.2K by changing the /etc/remote entry to modem2:\ :dv=/dev/cua0:br#19200:p8:hf:nt:el=^C^S^Q^U^D:ie=%$:oe=^D: or using set speed 19200 for kermit, but downloads still lock up just as badly as at 38.4K. I haven't (yet) tried lowering the baud rate to 9600... I know about S-bus and SCSI serial servers, but (a) only the latter would be usable here, the ELC doesn't have an off-the-cpu-board S-bus, (b) I can't afford one, and (c) I shouldn't need one, other people seem to be able to drive the build-in ports at 38.4K without difficulty. Questions: - HELP!! Can anyone suggest just what's going on here and/or how to get this hardware configuration to work? - How is it that on one occasion I successfully downloaded a 500K tar.gz gzipped tar archive (using kermit), yet I have had several other downloads lock up with as little as 20K bytes transferred? - Are there other SunOS patches which might help? - Would upgrading to SunOS 4.1.4 help? I noticed that Sun's Jumbo TTY patch (#100513-04) includes fixes for 4.1.[123], but makes no mention of 4.1.4. - Could some kind soul with a WorldBlazer hooked to a sun4c built-in serial port please E-mail me her/his modem S-register settings and/or /etc/remote entry and/or .kermrc commands? - Could I find out anything useful by putting a serial-port breakout box between the Sun and the modem and watching the various handshaking signals? - With the Jumbo TTY patch (#100513-04) applied, does SunOS 4.1.3 really support RTS/CTS handshaking in both directions? The patch README lists | BugID 1012954: | SunOS does not do RTS/CTS flow control for incoming datastreams, previously | Xon/Xoff flow control had to be used, (this has been fixed for CPU (zs_async) | serial ports only). as a BUGId fixed with this patch, suggesting that RTS/CTS *is* supported, but I'd like an independent confirmation of this. - An anonymous person posted a replacement zs_async.o back in 1993 (archived in ftp.lysator.liu.se:/pub/sun/kernel-hacks/) which | ... was built from a | 4.1.2 base (probably not much changed in 4.1.3) that incorporates the old known | serial patches all into one binary, AND it adds a special hack to suck in | all of the available characters from the FIFO if there are some left (i.e., | basically adds a "do { original code } while (characters available)" to the | "zsa_rxint()" routine, that's all). This is useful for things like PPP where | you want to suck things in fast as you can at 38,400 baud, and it helps | reduce the number of "zs0: silo overflow" messages that are all too familar | to those of you doing high-speed asynch serial I/O ... What experience do people have with this? I'm a little surprised that Sun didn't include the "do { original code } while (characters available)" hack in its own Jumbo TTY patch, and I'm nervous about trying this in the kernel until I know more about it. Thanks 2^20 for any assistance anyone can offer, - Jonathan Thornburg (personal E-mail) U of British Columbia / Physics Dept / "Washing one's hands of the conflict between the powerful and the powerless means to side with the powerful, not to be neutral." - Freire / OXFAM - ------------------------------ End of Suns-at-Home Digest ******************************