lancelot@spock.UUCP (Thor Lancelot Simon) (06/03/90)
<Cut this!> I have an overabundance of DEC RZ23 105-meg drives which are intended to go in a few DecStations that won't arrive for a while. (Please keep in mind that as far as warranty goes, this article is altogether hypothetical.) They look a bit like Quantum 105S, and I figured that they probably were, with the firmware changed so that they report manufacturer DEC, drive type RZ23. However, upon replacing the Q80 in my Amiga HD box with one of these, a 'Drive Scan' from my SCSI format&tools package didn't show up. Didn't even flash the drive light, like it does on everything else I've hooked up to it. I know one can hook up standard SCSI drives (tapes, scanners, and CDROMS as well, though we haven't tried to write drivers yet) to a DECStation, and it will pick them up if you do a 'scsi pb' (drive scan) from the boot prompt, but do the DEC drives have some abnormal query format only understood by other DEC products? If so, could I modify my amiga software to speak it, or does it have to be done at the hardware level (pull a line low or some such oddity)? Just for reference, my Amiga setup is an A500 with a Xetec Fasttrak controller. I have checked all the cables involved, and my controller can talk to just about anything else I hook up to it, except the RZ23s in question. I'm using the utilities that came with the controller, one called 'SCSI Tools', and one called 'Partitions'. The light on the RZ23 does flash a few times when I power up my Amiga, but only then. ******************************************************************************* *Thor Simon * Okay, just a little pin-prick...There'll be no more-* *lancelot@spock.UUCP * Aieeeeaaaugh!-but you may feel a little _sick_. * *uunet!hsi!yale!lancelot* ---Pink Floyd *
jg@zorch.crl.dec.com (Jim Gettys) (06/04/90)
As far as I know, we do everything legit, for the standard command set (or whatever the SCSI terminology is for the command set disk and tape devices are supposed to implement). And while I'm not one of our SCSI experts, I don't think we've done anything strange to the drives we use, other than get them to work properly and put a DEC label in the ROM... Experience has shown, however, that SCSI is far from the tightest specification in the world. When we started hooking up other SCSI devices to DECstations, there was about a 50% chance they would work (to my knowledge, it our case, it was always microcode problems with the SCSI peripherals; in your case, it could be either the drive or your controller); this percentage has been increasing with time, as people get their implementations conformant to the specification, and the specification gets better understood by the implementors (moral is: KISS... SCSI violates this rule...). But your experience is not all that unusual, from what I've seen, with hooking up new devices on a SCSI... Sigh... Of course, make sure you have the SCSI unit number set to something reasonable... I once had a RZ55 and an RZ23 set to the same address on my DS3100, and shall we just say that the machine acted wierd..... Almost booted up multi-user, but not quite... - Jim
wherry@alazif.enet.dec.com (bradley g wherry) (06/04/90)
In article <1990Jun2.213812.13909@spock.UUCP>, lancelot@spock.UUCP (Thor Lancelot Simon) writes: <stuff deleted> > bit like Quantum 105S, and I figured that they probably were, with the firmware > changed so that they report manufacturer DEC, drive type RZ23. However, upon > replacing the Q80 in my Amiga HD box with one of these, a 'Drive Scan' from my > SCSI format&tools package didn't show up. Didn't even flash the drive light, > like it does on everything else I've hooked up to it. I know one can hook up The RZ23 is really SCSI, it is honest. The drive is not a Quantum 105S but rather a Conner 105 MB drive. yes the drive does report the drive type as RZ23. The drive does not spin up when power is applied. I believe you need to send an SCSI INIT command to spin up the drive. They don't spin-up by default or rather the DEC version does not do this by default. <stuff deleted> > called 'Partitions'. The light on the RZ23 does flash a few times when I power > up my Amiga, but only then. not being really familiar with the drive I could only speculate that its the power up diags which are blinking the light. -- brad wherry | Ex ignorantia ad sapientiam; wherry@alazif.enet.dec.com | e luce ad tenebras.
wallace@oldtmr.enet.dec.com (Ray Wallace) (06/04/90)
In article <1990Jun2.213812.13909@spock.UUCP>, lancelot@spock.UUCP (Thor Lancelot Simon) writes... > However, upon >replacing the Q80 in my Amiga HD box with one of these, a 'Drive Scan' from my >SCSI format&tools package didn't show up. Didn't even flash the drive light, >like it does on everything else I've hooked up to it. Can you tell if the drive actualy spun up or not? I've read that the RZ's need a "spin up SCSI command" sent to them when ever you turn them on. --- Ray Wallace (INTERNET,UUCP) wallace@oldtmr.enet.dec.com (UUCP) ...!decwrl!oldtmr.enet!wallace (INTERNET) wallace%oldtmr.enet@decwrl.dec.com ---
minow@mountn.dec.com (Martin Minow) (06/04/90)
First, I would want to make it clear that I do not recommend that you connect a Dec RZ23 to a non-Dec computer. The RZ23 is a normal SCSI device. Drive formatters, however (I'm only familiar with the Macintosh) do a "device inquiry" to see what they're connected to. The RZ23 responds with something like "DEC RZ23". If your drive formatter is looking for "QUANTUM P105S", it will fail with "don't understand this drive." Also, vendors (not specifically Dec) often add proprietary microcode to the raw hardware. For example, they may need to control certain kinds of error handling. The Ansi standard reserves certain "mode select" and "mode sense" pages (records) to the manufacturer. Some vendors (again, not specifically Dec) send a Mode Sense command and expect to see "Copyright 1990 Mumble Corporation". In the PC market, when you buy a drive from an end-user vendor, it generally comes with a formatting package that understands its specific quirks. Martin Minow minow@thundr.enet.dec.com The above does not represent the position of Digital Equipment Corporation.
abstine@image.soe.clarkson.edu (Arthur Stine) (06/04/90)
From article <1990Jun2.213812.13909@spock.UUCP>, by lancelot@spock.UUCP (Thor Lancelot Simon): > <Cut this!> > > I have an overabundance of DEC RZ23 105-meg drives which are intended to go in > a few DecStations that won't arrive for a while. (Please keep in mind that as > far as warranty goes, this article is altogether hypothetical.) They look a > bit like Quantum 105S, and I figured that they probably were, with the firmware > changed so that they report manufacturer DEC, drive type RZ23. However, upon > replacing the Q80 in my Amiga HD box with one of these, a 'Drive Scan' from my The RZ23 and 22 are Conner drives... the RZ55 is a Micropolis to the best of my knowledge. -- Art Stine Sr Network Engineer Clarkson U ABStine@CLVMS.Clarkson.Edu
mjacob@wonky.Eng.Sun.COM (Matt Jacob) (06/04/90)
In article <5454@crltrx.crl.dec.com> jg@zorch.crl.dec.com (Jim Gettys) writes: >Experience has shown, however, that SCSI is far from the >tightest specification in the world. When we started hooking up >other SCSI devices to DECstations, there was about a 50% chance >they would work (to my knowledge, it our case, it was always microcode >problems with the SCSI peripherals; in your case, it could be either >the drive or your controller); this percentage has >been increasing with time, as people get their implementations >conformant to the specification, and the specification gets better >understood by the implementors (moral is: KISS... SCSI violates >this rule...). Weak, buddy. Real weak. This would have been a truer statement 3-4 years ago, but not for the last couple of years. -matt jacob Sun Microsystems
leech@ornat.cs.unc.edu (Jonathan Leech) (06/04/90)
In article <136580@sun.Eng.Sun.COM> mjacob@wonky.Eng.Sun.COM (Matt Jacob) writes: >In article <5454@crltrx.crl.dec.com> jg@zorch.crl.dec.com (Jim Gettys) writes: >>Experience has shown, however, that SCSI is far from the >>tightest specification in the world. (...) >Weak, buddy. Real weak. This would have been a truer statement 3-4 >years ago, but not for the last couple of years. Can we assume, then, that any SCSI drive manufactured in the last couple of years will work correctly when connected to a Sun SLC? (This is not a snide question; I'm thinking about buying either an SLC or a DS2100, and can't afford Sun/DEC mass storage prices). -- Jon Leech (leech@cs.unc.edu) __@/ "We were driving along, minding our own business, when there was a sudden flash of blue light which blotted out the stars. I thought it was a nuclear bomb going off and despaired for my career." - Keith Hughes
jesup@cbmvax.commodore.com (Randell Jesup) (06/04/90)
In article <1644@mountn.dec.com> minow@thundr.enet.dec.com (Martin Minow) writes: >First, I would want to make it clear that I do not recommend that you >connect a Dec RZ23 to a non-Dec computer. > >The RZ23 is a normal SCSI device. Drive formatters, however (I'm only >familiar with the Macintosh) do a "device inquiry" to see what they're >connected to. The RZ23 responds with something like "DEC RZ23". If your >drive formatter is looking for "QUANTUM P105S", it will fail with "don't >understand this drive." Note that almost all Amiga HD controllers (including Commodore's A590/A2091/A3000 controller and software) can deal with most any SCSI drive that implements a reasonable portion of the CCS. In particular, ReadCapacity, Inquiry, and the standard R/W/request-sense/FormatUnit/etc commands. In the HDToolBox for the above-mentioned controllers, there's a "Read Configuration from Drive" button that uses Inquiry, ReadCapacity, etc to attempt to figure out a reasonable configuration for the drive. This works pretty well, even on zone-recorded things like quantums. It may not be optimum, but it will never waste much space on a disk. I think having to get custom formatting software for each drive is a massive kludge/ripoff, IMHO. >Also, vendors (not specifically Dec) often add proprietary microcode to >the raw hardware. For example, they may need to control certain kinds of >error handling. The Ansi standard reserves certain "mode select" and "mode >sense" pages (records) to the manufacturer. Some vendors (again, not >specifically Dec) send a Mode Sense command and expect to see "Copyright 1990 >Mumble Corporation". > >In the PC market, when you buy a drive from an end-user vendor, it generally >comes with a formatting package that understands its specific quirks. Not on the Amiga. Just about any SCSI drive can be used, even stupid drives that (for example) respond on all LUNs of their target address (we have a dip-switch for that) or take 20 seconds to spin up (another dip-switch, otherwise known as "seagate" mode). The symptoms imply either a) it's set up for address 7 (all amiga scsi controllers I've seen default to 7), or b) it's totally stone dead, and not even responding to SCSI messaging. -- Randell Jesup, Keeper of AmigaDos, Commodore Engineering. {uunet|rutgers}!cbmvax!jesup, jesup@cbmvax.cbm.commodore.com BIX: rjesup Common phrase heard at Amiga Devcon '89: "It's in there!"
jg@zorch.crl.dec.com (Jim Gettys) (06/04/90)
>>Experience has shown, however, that SCSI is far from the >>tightest specification in the world. When we started hooking up >>other SCSI devices to DECstations, there was about a 50% chance >>they would work (to my knowledge, it our case, it was always microcode >>problems with the SCSI peripherals; in your case, it could be either >>the drive or your controller); this percentage has >>been increasing with time, as people get their implementations >>conformant to the specification, and the specification gets better >>understood by the implementors (moral is: KISS... SCSI violates >>this rule...). >Weak, buddy. Real weak. This would have been a truer statement 3-4 >years ago, but not for the last couple of years. > >-matt jacob > Sun Microsystems Boy, it looks like you don't like to read.... If you read my statement above, you'll see that I was talking about several years ago, and noted that the percentage has been increasing with time. And it was a very small sample. (statistics of small numbers....) Most of the problems we saw were of the form: Disk firmware has timing window that only a fast machine can hit... Drive might generate error or hang after pounding on it for some hours. Getting solid implementations of complex specifications is always difficult. Are you claiming that Sun makes slow machines and slow SCSI implementations that don't see that kind of problem? :-). Sun certainly didn't build such fast machines 18 months ago... :-). - Jim
mjacob@wonky.Eng.Sun.COM (Matt Jacob) (06/04/90)
In article <14442@thorin.cs.unc.edu> leech@ornat.cs.unc.edu (Jonathan Leech) writes: >>>Experience has shown, however, that SCSI is far from the >>>tightest specification in the world. (...) >>Weak, buddy. Real weak. This would have been a truer statement 3-4 >>years ago, but not for the last couple of years. > > Can we assume, then, that any SCSI drive manufactured in the last >couple of years will work correctly when connected to a Sun SLC? >(This is not a snide question; I'm thinking about buying either an SLC >or a DS2100, and can't afford Sun/DEC mass storage prices). >-- > Jon Leech (leech@cs.unc.edu) __@/ Well- I hope so. I realize that my statement was a bit too polemical. I've put in a *lot* of hours trying to tie down at least the DIRECT ACCESS devices. Ignoring formatting issues (which with respect to supported MODE SELECT pages are still a little shakey), basic reading and writing, disconnect/reconnect, sync negotiation, work pretty well now. Now- as a member of my company- and a very cautious one at that- I have to put up a disclaimer: If I haven't tried it personally, I don't warrant it. As a user and observer of SCSI over the last five years, I have to, on the other hand, state that the confusion and uncertainty over what a proper implementation of the SCSI specs are has died down tremendously- to the point that a new and unvalidated disk drive has, maybe, a 80-90% like- lihood of working well. This is not the ideal, but it sure as hell is a lot better than things were a couple of years ago. -matt
steven@pacific.csl.uiuc.edu (Steven Parkes) (06/04/90)
|> Can we assume, then, that any SCSI drive manufactured in the last |> couple of years will work correctly when connected to a Sun SLC? |> (This is not a snide question; I'm thinking about buying either an SLC |> or a DS2100, and can't afford Sun/DEC mass storage prices). The latest I heard on this group was that although non-DEC disks could be used for data on 21/31/5k's, they couldn't be used for booting. Is this still the concesus? Also, regarding prices, University people should ask their DEC rep's about consortium prices -- maybe not the best, but at near half off, they're very good for what you get.
lancelot@spock.UUCP (Thor Lancelot Simon) (06/04/90)
I think my controller automagically spins the drive up whenever I reboot the machine - I can feel vibration in the drive somewhere begin and it makes tracking noises. Re the drive light, I feel kinda dumb now. A RZ23 doesn't HAVE a drive light... Thus I saw no telltale flash from within my expansion chassis. If it had a light, I guess I'd probably see it. I still don't get the DEC label and device type like I do with any other arbitrary SCSI device I've tried. ******************************************************************************* *Thor Simon * Okay, just a little pin-prick...There'll be no more-* *lancelot@spock.UUCP * Aieeeeaaaugh!-but you may feel a little _sick_. * *uunet!hsi!yale!lancelot* ---Pink Floyd *
mjacob@wonky.Eng.Sun.COM (Matt Jacob) (06/04/90)
In article <5461@crltrx.crl.dec.com> jg@zorch.crl.dec.com (Jim Gettys) writes: >>>Experience has shown, however, that SCSI is far from the >>>tightest specification in the world. When we started hooking up >>>other SCSI devices to DECstations, there was about a 50% chance >>>they would work (to my knowledge, it our case, it was always microcode >>>problems with the SCSI peripherals; in your case, it could be either >>>the drive or your controller); this percentage has >>>been increasing with time, as people get their implementations >>>conformant to the specification, and the specification gets better >>>understood by the implementors (moral is: KISS... SCSI violates >>>this rule...). > >>Weak, buddy. Real weak. This would have been a truer statement 3-4 >>years ago, but not for the last couple of years. > >> >>-matt jacob >> Sun Microsystems > >Boy, it looks like you don't like to read.... First of all- apologies for being so acerbic. And yes- I didn't read it as carefully as I might- I was dialed in from home (23 line tty pages don't latch up in my mind like they should...*sigh*) > >If you read my statement above, you'll see that I was talking about several >years ago, and noted that the percentage has been increasing with time. >And it was a very small sample. (statistics of small numbers....) > >Most of the problems we saw were of the form: > >Disk firmware has timing window that only a fast machine can hit... >Drive might generate error or hang after pounding on it for some hours. > >Getting solid implementations of complex specifications is always >difficult. True. > >Are you claiming that Sun makes slow machines and slow SCSI implementations >that don't see that kind of problem? :-). Sun certainly didn't >build such fast machines 18 months ago... :-). > - Jim Speed of the CPU or the Host Adapter is largely irrelevant. A good and careful s/w implementation of the transport layer (Host adapter driver) can undertake various strategies to make up for deficient hardware. -matt p.s.: Let me present my credentials, which you can then use specifically defame me with: I am 99% responsible for the SCSI S/W on the SparcStation-1 (and its clones, e.g., SLC, etc.).
leech@homer.cs.unc.edu (Jonathan Leech) (06/04/90)
In article <1990Jun4.160328.22033@ux1.cso.uiuc.edu> steven@pacific.csl.uiuc.edu writes: >Also, regarding prices, University people should ask their DEC rep's about >consortium prices -- maybe not the best, but at near half off, they're very >good for what you get. DEC Consortium Random PC Week Vendor -- -- RZ23 $1000 Conner CP3100 $565 RZ24 $1500 Conner CP3200 $875 But for a major manufacturer, I suppose that is pretty good. I'm certainly not flaming DEC for setting the prices they do, just exploring far cheaper alternatives. -- Jon Leech (leech@cs.unc.edu) __@/ "I met a wonderful new man. He's fictional, but you can't have everything." - Cecelia, _The Purple Rose of Cairo_
jg@zorch.crl.dec.com (Jim Gettys) (06/05/90)
Matt, I think we are converging on general agreement... Short of making signals run more slowly, or slowing down interrupt response (something I doubt you'd be willing to do), the kind of problems we were seeing weren't fixable in host software. Sounds like you've had somewhat better luck than we did (or your machines are slower :-)). In any case, general experience shows that disk incompatibilty problems have been diminishing with time, and are much less a problem than they once were. I've heard much more often recently of things plugging in and working fine than was true 20 months ago. And to answer Steven Parkes question about booting... Ultrix boots just fine off of other people's disks. The only problem I am aware of is doing an initial installation of Ultrix onto someone else's disk is more involved (since Ultrix won't know anything about the disk, at the time it needs to). Even this can be worked around if you are sufficiently motivated/informed; it is the same procedure needed when installing on an RZ56 or RZ57 under Ultrix 3.0 or 3.1, which won't be able to recognise the disks. Generally, the easiest solution is to hook up the other disk to an existing machine, and set the disk up there, and recable, rather than do a fresh install. - Jim
chris@com50.c2s.mn.org (Chris Johnson) (06/08/90)
In article <14442@thorin.cs.unc.edu> leech@ornat.cs.unc.edu (Jonathan Leech) writes: >In article <136580@sun.Eng.Sun.COM> mjacob@wonky.Eng.Sun.COM (Matt Jacob) writes: >>In article <5454@crltrx.crl.dec.com> jg@zorch.crl.dec.com (Jim Gettys) writes: >>>Experience has shown, however, that SCSI is far from the >>>tightest specification in the world. (...) >>Weak, buddy. Real weak. This would have been a truer statement 3-4 >>years ago, but not for the last couple of years. > > Can we assume, then, that any SCSI drive manufactured in the last >couple of years will work correctly when connected to a Sun SLC? >(This is not a snide question; I'm thinking about buying either an SLC >or a DS2100, and can't afford Sun/DEC mass storage prices). >-- > Jon Leech (leech@cs.unc.edu) __@/ Oh heck, I'll jump into the fray. SCSI is pretty much plug-n-play for disk drives, if the software driver has been written with such in mind. However, Sun's SCSI driver is just barely crawling out of the stone age, in that aspect. Generic, it is not, although with SunOS 4.X it has improved a lot over previous versions in the arena of compatibility. It still thinks SCSI devices are out there on an Emulex MD21 bridge controller, for pete's sake. So the real answer is: Most SCSI drives manufactured in the last couple years can be made to work with a Sun SLC with a small amount of fiddling. -- ...Chris Johnson chris@c2s.mn.org ..uunet!bungia!com50!chris Com Squared Systems, Inc. St. Paul, MN USA +1 612 452 9522