rot@unlisys.UUCP (Robert Rothe) (07/13/90)
hello there. we added a third scsi disk to our system and a strange behavior occured. we used the following components under isc 2.02 : adaptec 1542B WREN V (runner) as drive 00 Quantum P-105 S as drive 01 WREN V (runner) as drive 02 33 MHz motherboard (hauppauge), 64k cache, 8 MB. when adding the second Wrendrive as drive 02 for a moment everything went fine. but then forcing heavy I/O on the drives via a "cpio -p" from drive 01 (quantum) to drive 02 (drive 01 has been the spooldisk ;-) after some copy (~20 MB) the system stops, without any errormessage, It Just Hangs ! it hasn't been the inode bug cause this one has been fixed, and the same happend when just creating one big archive one the disk02. so when i umounted the disk02 and created the archive on disk00 everything was ok, then umounting disk01 and mounting disk02 and unpacking the archive on disk02 was no problem. so now i ask myself if interactive is not able to handle three scsidrives with heavy io ? or if it's the adaptec (i doubt !). robert. -- Unlisys, Hohenzollerndamm 7, 1000 Berlin 31 --- Robert Rothe tmpmbx!unlisys!rot, rot@unlisys, 030 / 853 28 72 like your PAN/IX !
rick@tmiuv0.uucp (07/23/90)
In article <675@unlisys.UUCP>, rot@unlisys.UUCP (Robert Rothe) writes: > hello there. > we added a third scsi disk to our system and a strange behavior occured. > we used the following components under isc 2.02 : > > adaptec 1542B > WREN V (runner) as drive 00 > Quantum P-105 S as drive 01 > WREN V (runner) as drive 02 > 33 MHz motherboard (hauppauge), 64k cache, 8 MB. Whoa! Be careful here! Check your power supply. Three hard disks are NOT a good thing on a single power supply. I had a 230W supply and two ST-506 type drives. When I added the Seagate SCSI drive, I smoked the supply. It's possible that your supplies are crowbarring. If you have one or two of these drives in a separate box and power supply, or a big 300 or 350 watt supply then this is probably not the case. Double check, just to be sure. > robert. > > -- > Unlisys, Hohenzollerndamm 7, 1000 Berlin 31 --- Robert Rothe > tmpmbx!unlisys!rot, rot@unlisys, 030 / 853 28 72 > like your PAN/IX ! -- ---------------------------------------------------------------------------- [- O] Rick Stevens ? EMail: uunet!zardoz!tmiuv0!rick -or- uunet!zardoz!xyclone!sysop V CIS: 75006,1355 (75006.1355@compuserve.com from Internet) "I'm tellin' ya, Valiant! Da whole ting stinks like yesterday's diapers!" - Baby Herman in "Who Framed Roger Rabbit" ----------------------------------------------------------------------------
debra@alice.UUCP (Paul De Bra) (07/25/90)
In article <3491@tmiuv0.uucp> rick@tmiuv0.uucp writes: ?In article <675@unlisys.UUCP>, rot@unlisys.UUCP (Robert Rothe) writes: ?> ... ?> WREN V (runner) as drive 00 ?> Quantum P-105 S as drive 01 ?> WREN V (runner) as drive 02 ? ?Whoa! Be careful here! Check your power supply. Three hard disks are NOT ?a good thing on a single power supply. I had a 230W supply and two ST-506 ?type drives. When I added the Seagate SCSI drive, I smoked the supply. It's ?possible that your supplies are crowbarring. There is a potential problem indeed, but what's most important is how power hungry the drives are. I don't know about the Quantum but the Wren V only needs about 15 Watts (average), while a Seagate 4096 or similar drive needs almost twice as much. The Wren V (mine is a 383H, esdi, not scsi) is an excellent choice if power is a problem. Paul. -- ------------------------------------------------------ |debra@research.att.com | uunet!research!debra | ------------------------------------------------------
src@scuzzy.mbx.sub.org (Heiko Blume) (07/28/90)
rick@tmiuv0.uucp writes: >Whoa! Be careful here! Check your power supply. Three hard disks are NOT >a good thing on a single power supply. I had a 230W supply and two ST-506 >type drives. When I added the Seagate SCSI drive, I smoked the supply. It's >possible that your supplies are crowbarring. well, it's surely better to have more watts, but that seems to depend on the power supply (and the weather on the day of purchase :-). especially those 'generic' (taiwan) power supplies can be real bad. a friend of mine has smoked THREE of them :-( however! i only have a 200 (or 220) watt power supply (*and* the same brand as my friend's!) and had no problems. i also have three scsi drives (ST296N, WREN V, PRO-80S), and i have also used an archive 150MB streamer together with this combo without problems. -- Heiko Blume c/o Diakite blume@scuzzy.mbx.sub.org FAX (+49 30) 882 50 65 Kottbusser Damm 28 blume@netmbx.UUCP VOICE (+49 30) 691 88 93 D-1000 Berlin 61 blume@netmbx.de TELEX 184174 intro d "Have you bugged your source today?"
debra@alice.UUCP (Paul De Bra) (08/01/90)
In article <1990Jul28.011957.301@scuzzy.mbx.sub.org> src@scuzzy.mbx.sub.org (Heiko Blume) writes: >rick@tmiuv0.uucp writes: >>Whoa! Be careful here! Check your power supply. Three hard disks are NOT >>a good thing on a single power supply... >well, it's surely better to have more watts, but that seems to depend on the >power supply (and the weather on the day of purchase :-). especially those >'generic' (taiwan) power supplies can be real bad. a friend of mine has >smoked THREE of them :-( ... Yep, power supply smoking can be very unrelated to power consumption. The power supply in my Everex Step 386/25 smoked after less than 6 months, with only 1 drive and 4meg in the system. I put in another one, a 230 Watt thing from a local store, and it smoked a couple of months later, shortly after adding a second drive and 4 more meg. I pulled a 5 year old (!) 200 Watt power supply from an old AT-clone, and the Everex has been running happily every since. So beware of these new taiwan 'heavy duty' 230 or more watt power supplies. Paul. -- ------------------------------------------------------ |debra@research.att.com | uunet!research!debra | ------------------------------------------------------
src@scuzzy.mbx.sub.org (Heiko Blume) (08/03/90)
debra@alice.UUCP (Paul De Bra) writes: >?possible that your supplies are crowbarring. >There is a potential problem indeed, but what's most important is how >power hungry the drives are. I don't know about the Quantum but the >Wren V only needs about 15 Watts (average), while a Seagate 4096 or similar >drive needs almost twice as much. The Wren V (mine is a 383H, esdi, not scsi) >is an excellent choice if power is a problem. the danger is how much the drives need when you power them up. i RTFM'ed a bit and found out that my 600MB WREN V uses up to 4.5 ampere on startup (on 12V). well, most 220 watt power supplies can give you 8 ampere on 12V so if you connect two of these beast you might get problems. i also phoned around, and bigger power supplies like 300+ watts are very rare and *very* expensive. i'll get more info next week, anybody interested? -- Heiko Blume c/o Diakite blume@scuzzy.mbx.sub.org FAX (+49 30) 882 50 65 Kottbusser Damm 28 blume@netmbx.UUCP VOICE (+49 30) 691 88 93 D-1000 Berlin 61 blume@netmbx.de TELEX 184174 intro d scuzzy Any ACU,e 19200 6919520 ogin:--ogin: nuucp ssword: nuucp
jeffl@comix.UUCP (Jeff Liebermann) (08/05/90)
In article <11119@alice.UUCP>, debra@alice.UUCP (Paul De Bra) writes: > In article <1990Jul28.011957.301@scuzzy.mbx.sub.org> src@scuzzy.mbx.sub.org (Heiko Blume) writes: > >rick@tmiuv0.uucp writes: > Yep, power supply smoking can be very unrelated to power consumption. > The power supply in my Everex Step 386/25 smoked after less than 6 months, > with only 1 drive and 4meg in the system. I put in another one, a 230 Watt > thing from a local store, and it smoked a couple of months later, shortly I do component level repair on some hardware including power supplies. About half of the power supplies I've seen have a problem with poor solder connections on the *SINGLE* *SIDED* circuit boards. What happens is the vibrations (at 40Khz or so) from the transformer(s) mechanically couple thru the leads and fracture the solder connections. These physically large components also absorb considerable heat during wave soldering that tends to cause cold (crummy) solder connections. I can usually spot an annular ring around the pins. Re-soldering with a big iron solves the problem. Another common disgrace is that the switching transistor(s) and diodes are spring clipped instead of screwed to the heat sink(s). Clips do not guaranttee sufficient heat transfer and will result in elevated device temperatures, and early failures. Semiconductor failure due to inadequate heat sinking accounts for another 30% of the repairs. The rest are mechanical damager, filth, paper clips, staples, screws, fan failure, and such. -- # Jeff Liebermann Box 272 1540 Jackson Ave Ben Lomond CA 95005 # (408)336-2558 voice (408)429-0483 digital pager CIS:73557,2074 # PC REPAIR & RF DESIGN. uunet!comix!jeffl pyramid!vdx!comix!jeffl # universe!milky_way!solar_system!earth!na!us!uunet!comix!jeffl
rcd@ico.isc.com (Dick Dunn) (08/07/90)
src@scuzzy.mbx.sub.org (Heiko Blume) writes: ...debra@alice.UUCP (Paul De Bra) writes: > >There is a potential problem indeed, but what's most important is how > >power hungry the drives are... ... > the danger is how much the drives need when you power them up. i RTFM'ed > a bit and found out that my 600MB WREN V uses up to 4.5 ampere on startup > (on 12V)... Blume's point is a good one: the limiting factor for powering the drives is almost certainly the startup power. It's time to re-invent something else the mainframe world has known for a long time: power sequencing. One of the folks here has a box'o'disks (SCSI) which are designed to power- sequence themselves. The idea is very simple: Each drive delays its spin- up from the time it gets logic power by an interval directly related to the drive's SCSI ID. Seems simple and eminently sensible. (How common is this, anyway?) -- Dick Dunn rcd@ico.isc.com -or- ico!rcd Boulder, CO (303)449-2870 ...Are you making this up as you go along?
nvk@ddsw1.MCS.COM (Norman Kohn) (08/09/90)
In article <1990Aug03.132554.13326@scuzzy.mbx.sub.org> src@scuzzy.mbx.sub.org (Heiko Blume) writes: >debra@alice.UUCP (Paul De Bra) writes: >>?possible that your supplies are crowbarring. > >>There is a potential problem indeed, but what's most important is how >>power hungry the drives are. I don't know about the Quantum but the >>Wren V only needs about 15 Watts (average), while a Seagate 4096 or similar >>drive needs almost twice as much.> >the danger is how much the drives need when you power them up. i RTFM'ed >a bit and found out that my 600MB WREN V uses up to 4.5 ampere on startup >(on 12V). well, most 220 watt power supplies can give you 8 ampere on 12V Some drives (my CDC-Imprimis, for example) have a jumper option of not starting the motor until receiving a "motor start" SCSI command. The drives could thus be started one at a time. This could even be done within the unix startup processes: the drive with the root filesystem could be left to start at power on, and the rest powered up via commands in /etc/rc2.d. I haven't tried this, though, and wouldn't know how to issue the commands to the SCSI interface. ISC clearly doesn't want us talking to the interface, and I imagine that you'd have to dig quite a bit in the driver code. It would probably be easier to write a separate driver that does nothing but send the startup signal to the SCSI port: if it functions only during system startup, perhaps it can exist in ignorance of the state of the "real" SCSI driver. -- Norman Kohn | ...ddsw1!nvk Chicago, Il. | days/ans svc: (312) 650-6840 | eves: (312) 373-0564