dhesi@bsu-cs.UUCP (Rahul Dhesi) (10/05/88)
Given about $20,000 to $30,000 to spend on a disk drive for a VAX-11/785 running 4.3BSD, which already has two rm80s and one rp07 on the massbus, can I do better than buying an RA81 from DEC? How much better? Reliable on-site maintenance is essential. (Location: Muncie, Indiana, about 50 miles from Indianapolis). Please send email if you have suggestions. I will summarize a few weeks from now. -- Rahul Dhesi UUCP: <backbones>!{iuvax,pur-ee}!bsu-cs!dhesi
roy@phri.UUCP (Roy Smith) (10/05/88)
[Note: this has very little to do with either Unix or archicture so I've added comp.periphs to the Newsgroups list and directed followups there] In article <4198@bsu-cs.UUCP> dhesi@bsu-cs.UUCP (Rahul Dhesi) writes: > Given about $20,000 to $30,000 to spend on a disk drive [...] can I do > better than buying an RA81 from DEC? Having had an RA-81 for 4 or 5 years, I can say with some assurance that just about anything else would be better. RA-81s are expensive, slow, and unreliable. What less could you want from a disk drive? For $30k, you should be able to buy a couple of Fuji SuperEagles and a good Emulex controller, or any of several other high-quality drives from Fuji or NEC. If you prefer to buy American, CDC has several good products. -- Roy Smith, System Administrator Public Health Research Institute {allegra,philabs,cmcl2,rutgers}!phri!roy -or- phri!roy@uunet.uu.net "The connector is the network"
nessus@athena.mit.edu (Doug Alan) (10/06/88)
In article <3531@phri.UUCP> roy@phri.UUCP (Roy Smith) writes: > RA-81s are expensive, slow, and unreliable. What less could you > want from a disk drive? For $30k, you should be able to buy a > couple of Fuji SuperEagles and a good Emulex controller, or any of > several other high-quality drives from Fuji or NEC. Why spend big bux on SuperEagles when you can get 800 Meg, high speed 5.25-inch Winchesters these days for $3,300 each? |>oug /\lan (or nessus@athena nessus@mit-eddie.uucp)
seeger@beach.cis.ufl.edu (F. L. Charles Seeger III) (10/06/88)
In article <10199@eddie.MIT.EDU> nessus@athena.MIT.EDU (Doug Alan) writes: |In article <3531@phri.UUCP> roy@phri.UUCP (Roy Smith) writes: | |> RA-81s are expensive, slow, and unreliable. What less could you |> want from a disk drive? For $30k, you should be able to buy a |> couple of Fuji SuperEagles and a good Emulex controller, or any of |> several other high-quality drives from Fuji or NEC. | |Why spend big bux on SuperEagles when you can get 800 Meg, high speed |5.25-inch Winchesters these days for $3,300 each? Performance, of course. The new SMD drives and controllers offer 3.0 MBps interfaces (24 Mbps). I belive that these 5.25" drives are limited to about 10 Mbps. This is important on modern fast machines, but the VAX in question may be too old and slow to qualify. Regards, Chuck
henry@utzoo.uucp (Henry Spencer) (10/07/88)
In article <10199@eddie.MIT.EDU> nessus@athena.MIT.EDU (Doug Alan) writes: >Why spend big bux on SuperEagles when you can get 800 Meg, high speed >5.25-inch Winchesters these days for $3,300 each? Reliability. Actually, I would probably prefer to buy one of Fujitsu's newer (physically smaller) disks, like the Swallow, rather than a Super Eagle. But buying disk drives on capacity and price alone is a serious mistake. -- The meek can have the Earth; | Henry Spencer at U of Toronto Zoology the rest of us have other plans.|uunet!attcan!utzoo!henry henry@zoo.toronto.edu
jbs@fenchurch.MIT.EDU (Jeff Siegal) (10/07/88)
In article <18529@uflorida.cis.ufl.EDU> seeger@beach.cis.ufl.edu (F. L. Charles Seeger III) writes: >|Why spend big bux on SuperEagles when you can get 800 Meg, high speed >|5.25-inch Winchesters these days for $3,300 each? >Performance, of course. The new SMD drives and controllers offer >3.0 MBps interfaces (24 Mbps). I belive that these 5.25" drives are >limited to about 10 Mbps. Not true. Maxtor offers a version of its drive which supports synchronous SCSI and claims 4MB/s transfers. I do not believe it costs much (any?) more than the slower (15 Mb/s, I believe) ESDI version. Jeff Siegal
nessus@athena.mit.edu (Doug Alan) (10/07/88)
>> [Doug Alan:] Why spend big bux on SuperEagles when you can get 800 >> Meg, high speed |5.25-inch Winchesters these days for $3,300 each? > [Chuck Seeger:] Performance, of course. The new SMD drives and > controllers offer 3.0 MBps interfaces (24 Mbps). I belive that > these 5.25" drives are limited to about 10 Mbps. This is important > on modern fast machines, but the VAX in question may be too old and > slow to qualify. Well, I'm not sure of this "of course". Our ESDI controllers (Sigma SCD-RQD11-EC's) certainly provides 24 Mbps also. The 5.25-inch disk drives we use (Maxtor XT8760E's) only do 15 Mbps, but if we multiply the number of sectors per track (52) by the number of rotations per second (60) we find that the disk drive can only get data off the disk at 12.5 Mbps, so having a higher transfer rate would probably have no effect. Future ESDI drives that have more sectors per track, will have higher transfer rates. In order for an SMD disk to provide much better performance, it will have to either spin much faster, have many more sectors per track, or have a much smaller seek time (or a less emphatic combination of the above). In no way is SMD inately superior to ESDI. (Except that SMD allows much longer cable distances.) On the other hand, the 5.25-inch drives are also available in SCSI versions. SCSI offers potentially higher performance and versatility than either SMD or SCSI. The only reason we didn't go with SCSI is because we already have a base of ESDI here. |>oug /\lan (or nessus@athena.mit.edu nessus@mit-eddie.uucp)
m5@lynx.UUCP (Mike McNally) (10/07/88)
In article <10209@eddie.MIT.EDU> jbs@fenchurch.MIT.EDU (Jeff Siegal) writes: >Not true. Maxtor offers a version of its drive which supports >synchronous SCSI and claims 4MB/s transfers. One must be careful when looking at performance specifications published by drive manufacturers. Often, the performance analyses are done by reading a megabyte or two in one command. This is all well and good, but doesn't say much about performance in a UNIX environment. It's important to know how fast the drives execute SCSI commands. We tried some drives here (I can't remember which manufacturer) that resonded to SCSI read requests so slowly that it lost a revloution on each read! We tried an 8-inch Fujitsu thing that was really slow, even though it had very high performance claims. In principle, I agree that SCSI can have very good performance. It's danged convenient, too. One interesting thing about Maxtors: we have a 5.25 inch full-height SCSI drive (170MB). It only responds to the "test unit ready" command when it's ready; otherwise, the command isn't acknowledged. Think about it. -- Mike McNally Lynx Real-Time Systems uucp: {voder,athsys}!lynx!m5 phone: 408 370 2233 Where equal mind and contest equal, go.
kaufman@polya.Stanford.EDU (Marc T. Kaufman) (10/07/88)
In article <4559@lynx.UUCP> m5@lynx.UUCP (Mike McNally) writes: >One interesting thing about Maxtors: we have a 5.25 inch full-height >SCSI drive (170MB). It only responds to the "test unit ready" command >when it's ready; otherwise, the command isn't acknowledged. Think >about it. From ANS X3.121-1986 (p.62): "The TEST UNIT READY command (table 7-2) provides a means to check if the logical unit is ready. This is not a request for a self test. If the logical unit would accept an appropriate medium-access command without returning CHECK CONDITION status, this command shall return a GOOD status." See -- it says nothing about what to do if the drive is not ready :-) I expect that the correct response would be BUSY. Marc Kaufman (kaufman@polya.stanford.edu)
jbs@fenchurch.MIT.EDU (Jeff Siegal) (10/07/88)
In article <1988Oct6.181050.788@utzoo.uucp> henry@utzoo.uucp (Henry Spencer) writes: >In article <10199@eddie.MIT.EDU> nessus@athena.MIT.EDU (Doug Alan) writes: >>Why spend big bux on SuperEagles when you can get 800 Meg, high speed >>5.25-inch Winchesters these days for $3,300 each? >Reliability. Our experience with the Maxtors has been very good. While we've only been using them for about 6 months, none have failed at all (we have quite a few; I'm not sure of the exact number now). Similarly with 5 1/4" ESDI Fujitsu drives (although not quite as cost effective as Maxtors), which we have been using for over a year. We do test them when they first arrive, and we have had some minor problems (e.g. 1/1,000 reads would result in soft errors) which caused us to return the drive for a replacement. Of course, we do regular backups (onto Exabyte 8mm tape--also very cost effective, and occasionally onto 1/2" GCR) and we have spare drives to swap in the case of failure. No maintenance cost, and since the drives cost less than half as much as the "big clunkers," keeping a spare or two is quite cost effective. Jeff Siegal
phil@amdcad.AMD.COM (Phil Ngai) (10/08/88)
In article <10209@eddie.MIT.EDU> jbs@fenchurch.MIT.EDU (Jeff Siegal) writes: >Not true. Maxtor offers a version of its drive which supports >synchronous SCSI and claims 4MB/s transfers. I do not believe it >costs much (any?) more than the slower (15 Mb/s, I believe) ESDI version. Ah, but is that 4 MB/s throughput, or just the rate it runs on the synchronous SCSI? It's like the difference between a computer with a 10 Mbit/s Ethernet and 85 Kbit/s throughput. The reason I ask is because 4 MB/s sounds like a generic synch SCSI figure. -- "In the West, to waste water is not to consume it, to let it flow unimpeded and undiverted down rivers. Use of water is, by definition, beneficial use." (from _Cadillac Desert_) Phil Ngai, {ucbvax,decwrl,allegra}!amdcad!phil or phil@amd.com
jlohmeye@entec.Wichita.NCR.COM (John Lohmeyer) (10/09/88)
In article <4329@polya.Stanford.EDU> kaufman@polya.Stanford.EDU (Marc T. Kaufman) writes: >In article <4559@lynx.UUCP> m5@lynx.UUCP (Mike McNally) writes: > >>One interesting thing about Maxtors: we have a 5.25 inch full-height >>SCSI drive (170MB). It only responds to the "test unit ready" command >>when it's ready; otherwise, the command isn't acknowledged. Think >>about it. > >From ANS X3.121-1986 (p.62): ^^^ This should be "X3.131-1986". > "The TEST UNIT READY command (table 7-2) provides a means to check if the >logical unit is ready. This is not a request for a self test. If the logical >unit would accept an appropriate medium-access command without returning >CHECK CONDITION status, this command shall return a GOOD status." > >See -- it says nothing about what to do if the drive is not ready :-) >I expect that the correct response would be BUSY. Actually, the correct response is CHECK CONDITION status with a sense key of NOT READY. There used to be a statement in the draft SCSI standard right after the one about not being a self test that said, "A fast response is expected." This sentence was dropped because no one really wanted to define "fast" and the sentence didn't mean much without such a definition. I definitely agree with the sentiment that Maxtor blew it on this one -- it certainly was not the SCSI committee's idea that a device not respond merely because it wasn't ready. (I strongly suspect that Maxtor is retrieving their controller microcode from the disk--thus they can't respond to SCSI commands until the disk is ready. They should have put in a bit more in ROM.) John Lohmeyer j.lohmeyer@wichita.ncr.COM .
markb@denali (10/11/88)
In article <10209@eddie.MIT.EDU>, jbs@fenchurch.MIT.EDU (Jeff Siegal) writes: > In article <18529@uflorida.cis.ufl.EDU> seeger@beach.cis.ufl.edu (F. L. Charles Seeger III) writes: > >|Why spend big bux on SuperEagles when you can get 800 Meg, high speed > >|5.25-inch Winchesters these days for $3,300 each? > >Performance, of course. The new SMD drives and controllers offer > >3.0 MBps interfaces (24 Mbps). I belive that these 5.25" drives are > >limited to about 10 Mbps. > > Not true. Maxtor offers a version of its drive which supports > synchronous SCSI and claims 4MB/s transfers. I do not believe it > costs much (any?) more than the slower (15 Mb/s, I believe) ESDI > version. > > Jeff Siegal Some clarification: 3 MB/s is data xfer rate from the media itself, and SMD can xfer at that rate. (actually SMDE) 4 MB/s claim by Maxtor and others is actually burst rate on SCSI, whether sync or async. xfer rate from the media itself is 10-15 MHz. ESDI will blow the doors off SCSI if done correctly. ESDI can go to 24 MHz, which compares nicely with SMDE. Currently available drives are at 15 Mhz. ESDI bus times are measured in microseconds, while SCSI can be looked at in milliseconds. For performance today, use SMD of some flavor. ESDI comes in 2nd, and SCSI is a far 3rd. ESDI will approach SMD within a year, and the 5.25" drives are coming with SMD interfaces now, too. markb
henry@utzoo.uucp (Henry Spencer) (10/12/88)
In article <20391@sgi.SGI.COM> markb@denali writes: >ESDI will blow the doors off SCSI if done correctly. ESDI can go to >24 MHz, which compares nicely with SMDE. Currently available drives >are at 15 Mhz. Uh, say what? SCSI transfer rate, at full bore, is 4 MB/s (note, not Mb/s), which is 32 MHz. Sounds comparable to me. Of course, there are a lot of cruddy SCSI controllers which can't hack that kind of speed, but then, there are cruddy ESDI and SMDE controllers too. >ESDI bus times are measured in microseconds, while >SCSI can be looked at in milliseconds. References, please. I've seen both SCSI and ESDI specs, and somehow I failed to notice any such disparity. In the specs, not the current (often lousy, for both) implementations. Do remember that this is, to some extent, an apples-and-oranges comparison, since SMDE and ESDI are drive-to-controller interfaces and SCSI is a controller-to-host interface. Since there *has* to be a controller between a disk drive and a SCSI bus, the quality of the controller makes a big difference. -- The meek can have the Earth; | Henry Spencer at U of Toronto Zoology the rest of us have other plans.|uunet!attcan!utzoo!henry henry@zoo.toronto.edu
friedl@vsi.COM (Stephen J. Friedl) (10/14/88)
[SCSI vs ESDI speed claims & flames deleted] In article <1988Oct12.164433.17763@utzoo.uucp>, henry@utzoo.uucp (Henry Spencer) writes: > > Do remember that this is, to some extent, an apples-and-oranges comparison, > since SMDE and ESDI are drive-to-controller interfaces and SCSI is a > controller-to-host interface. Since there *has* to be a controller > between a disk drive and a SCSI bus, the quality of the controller makes > a big difference. Indeed; the AT&T 3B2/600 uses a SCSI host adaptor with an ESDI disk controller module. Isn't this a hardware equivalent of the C-vs-assembler debate? Steve -- Steve Friedl V-Systems, Inc. +1 714 545 6442 3B2-kind-of-guy friedl@vsi.com {backbones}!vsi.com!friedl attmail!vsi!friedl ---------Nancy Reagan on the Three Stooges: "Just say Moe"---------