home *** CD-ROM | disk | FTP | other *** search
- Path: sparky!uunet!spool.mu.edu!olivea!sgigate!odin!sgihub!zola!zuni!anchor!olson
- From: olson@anchor.esd.sgi.com (Dave Olson)
- Newsgroups: comp.sys.sgi
- Subject: Re: Hard drives...
- Message-ID: <u83fjps@zuni.esd.sgi.com>
- Date: 30 Dec 92 07:53:11 GMT
- References: <34820@adm.brl.mil> <1hr2viINNq25@calvin.NYU.EDU> <72606@cup.portal.com>
- Sender: news@zuni.esd.sgi.com (Net News)
- Organization: Silicon Graphics, Inc. Mountain View, CA
- Lines: 77
-
- In <72606@cup.portal.com> thad@cup.portal.com (Thad P Floryan) writes:
-
- | In article <1hr2viINNq25@calvin.NYU.EDU>
- | roy@mchip00.med.nyu.edu (Roy Smith) writes:
- |
- | | [...]
- | | As far as I'm concerned, SCSI means Some Crappy Stupid Interface. I
- | | guess it beats SMD, but it sure doesn't live up the the wonderful promises
- | | it made (not to mention the random connector confusion; I know of 3 commonly
- | | used connectors, and I havn't even mentioned the wierdo PowerBook SCSI
- | | connector, but I guess that's understandable).
- |
- | As far as I'm concerned, "SGI = Stupid Goddamn Idiots" re: all the braindamage
- | on the SCSI bus per comments earlier this year. Wrong ID for the host adapter,
- | use of WD chip, inability to support a 6 meter SCSI bus with 4 devices, extreme
- | sensitivity to cable impedance, etc.
-
- Umm, where did you get the idea that there was something magical about
- using some ID (I assume you mean 7) for the host? Just because a lot
- of people happen to use it doesn't make it 'correct'. I have posted
- here several times why we use 0. The short summary is 'it was easy'.
- The longer answer is that it can actually (very) slightly improve
- *overall* system performance in a multiple target system (in a single
- target system it doesn't matter what you do, of course), although that
- testing came long after the decision to use 0. Certainly none of the
- SCSI standards ever came close to specifying/recommending an adapter
- ID, since they supported multiple initiators from the first SCSI
- standard (not SASI, of course).
-
- I don't know what you have against the WD chip, or why it
- matters to you, but see a long series of discussions about
- 2 months ago on comp.periphs.scsi on various SCSI chips for
- my comments on the advantages and disadvantges of the WD 33C93
- chip series.
-
- I *personally* have used a 6 meter overall cable length on a number of
- different SGI systems (with 1-7 targets). I do hope you allowed for
- the internal cable length (obscenely long in the predator chassis!).
-
- Any high speed SCSI bus (single ended, at least) will be sensitive to
- cable impedance.
-
- If you believe you have legal configurations that don't work,
- call support and beat them over the head about it until they
- make it work, or show you that something isn't in spec!
-
- | NO such problems with HP, Sun, Amiga, DEC, and Mac equipment on any of multiple
- | systems (and the bad experiences with SGI SCSI is on more than one platform).
-
- Then by all means, get it fixed. I hear about most of the
- SCSI problems on the ESD platforms, and a fair number on the
- ASD platforms, and I know that *many* people have 7 devices
- per bus working just fine, at least some with full length
- cables (sometimes even exceeding the specs, although that
- is pretty iffy).
-
- I'm sorry to hear you have had more problems with our equipment than
- with our competitors (after all, I own stock in SGI), but without more
- details, I don't know what is different.
-
- By the way, I've heard you complain more than once on comp.sys.hp about
- the lack of partitioning, and other HP problems (and similarly on the
- Sun newsgroups) , so I assume you aren't just bashing SGI for the hell
- of it, and are really having problems. I haven't heard from you on
- comp.sys.sgi about SGI SCSI problems, as far as I can remember though;
- have you at least contacted the support folks? What hardware
- platforms, and what kind of problems have you had?
-
- I've been running 5-7 SCSI devices on every SGI system I've had (4D/20,
- 25, 35, Indigo, r4k Indigo) for long periods of time, with no problems;
- similarly I've set up many in the labs, and in or test setups. These
- were all ESD products, and I can't speak from quite so much experience
- on the power series, but I know it works there also.
- --
- Let no one tell me that silence gives consent, | Dave Olson
- because whoever is silent dissents. | Silicon Graphics, Inc.
- Maria Isabel Barreno | olson@sgi.com
-