ALG106@psuvm.psu.edu (02/12/91)
The company I work for was thinking of getting a new fileserver -
we currently have a 33MHz 386. We were thinking of a 33MHz 486 EISA
bus machine. However, someone told us that the i486 has a bug in it
that shows up only when running Netware.
First of all, is this true? If so, are there plans to fix it?
Any suggestions you may have on a new fileserver would be helpful -
we were thinking of getting 2 1Gig drives (fault tolerance).
Andy
------------------------------------------------------------------------
Andrew Gross BITNET: ALG106@PSUVM
Penn State University INTERNET: alg106@psuvm.psu.edu6600sirt@ucsbuxa.ucsb.edu (Mike O'Brien) (02/13/91)
From article <91042.192216ALG106@psuvm.psu.edu>, by ALG106@psuvm.psu.edu: > The company I work for was thinking of getting a new fileserver - > we currently have a 33MHz 386. We were thinking of a 33MHz 486 EISA > bus machine. However, someone told us that the i486 has a bug in it > that shows up only when running Netware. > > First of all, is this true? If so, are there plans to fix it? > > Any suggestions you may have on a new fileserver would be helpful - > we were thinking of getting 2 1Gig drives (fault tolerance). > There were bugs in the early versions of both the i386 chip and the i486 chip that affected Netware. Netware checks for bugs in your CPU whenever you boot it. However, these bugs were fixed a long time ago, so if you buy a new 486 now, you will have no problem. Note that Netware 386 actually contains optimizations for a 486, so there is quite a large difference in speed between Netware 386 on a 386 and on a 486. Our Logix 486/25 file server gets a speed rating (from the Netware 386 SPEED command) of 686! Finally, if you plan to use more than 16 MB of memory on your server, make sure you get an EISA machine as opposed to an ISA one. Otherwise Netware will not recognize the extra memory. Mike O'Brien 6600sirt@ucsbuxa.ucsb.edu
mallsop@suna.mqcc.mq.oz.au (Mark Allsop) (02/13/91)
In article <9057@hub.ucsb.edu> 6600sirt@ucsbuxa.ucsb.edu writes: >From article <91042.192216ALG106@psuvm.psu.edu>, by ALG106@psuvm.psu.edu: <stuff deleted> >Note that Netware 386 actually contains optimizations for a 486, so >there is quite a large difference in speed between Netware 386 on a >386 and on a 486. Our Logix 486/25 file server gets a speed rating >(from the Netware 386 SPEED command) of 686! Could someone let me know the speed rating for a 386 server? We also have a 486 server and I'd be interested to know the difference. Cheers, +Mark. -- Mark Allsop Computer Scientist email: mallsop@suna.mqcc.mq.oz.au The Statistical Laboratory Phone: At MacUni: (61 2) 805-8792 / \ Macquarie University, Australia Fax : : (61 2) 805-7433 | This one goes up to 11.....
ckinsman@eecs.wsu.edu (Chris Kinsman - EE major) (02/14/91)
In article <1202@macuni.mqcc.mq.oz> mallsop@suna.mqcc.mq.oz.au (Mark Allsop) writes: >In article <9057@hub.ucsb.edu> 6600sirt@ucsbuxa.ucsb.edu writes: >>From article <91042.192216ALG106@psuvm.psu.edu>, by ALG106@psuvm.psu.edu: > ><stuff deleted> > >>Note that Netware 386 actually contains optimizations for a 486, so >>there is quite a large difference in speed between Netware 386 on a >>386 and on a 486. Our Logix 486/25 file server gets a speed rating >>(from the Netware 386 SPEED command) of 686! > >Could someone let me know the speed rating for a 386 server? We also have >a 486 server and I'd be interested to know the difference. > we have a 16Mhz Model 70 and it rates a 120. Chris
dwd@usenet.umr.edu (Dan DeNise) (02/14/91)
In article <1202@macuni.mqcc.mq.oz> mallsop@suna.mqcc.mq.oz.au (Mark Allsop) writes: > >Could someone let me know the speed rating for a 386 server? We also have >a 486 server and I'd be interested to know the difference. > >Cheers, >+Mark. All 6 of our 33Mhz 386s get a rating of 319. Dan DeNise Computing Services University of Missouri-Rolla Missouri's Technological University c0016@umrvmb.umr.edu
heflin@cod.NOSC.MIL (Greg R. Heflin) (02/14/91)
I have a noname, unknown 386 w/coprocessor SPEED ratting of 316. (The computer is in a tower configuration, but was put together from pieces by someother people here.) -gregory heflin NAVELEX,SD -- -gregory Just my two cents which shouldn't make a flame, unless you rub them real hard.
samba@sm.sony.co.jp (Masaaki Hoshino) (02/14/91)
In article <1202@macuni.mqcc.mq.oz>, mallsop@suna.mqcc.mq.oz.au (Mark Allsop) writes: > In article <9057@hub.ucsb.edu> 6600sirt@ucsbuxa.ucsb.edu writes: > >From article <91042.192216ALG106@psuvm.psu.edu>, by ALG106@psuvm.psu.edu: > > <stuff deleted> > > >Note that Netware 386 actually contains optimizations for a 486, so > >there is quite a large difference in speed between Netware 386 on a > >386 and on a 486. Our Logix 486/25 file server gets a speed rating > >(from the Netware 386 SPEED command) of 686! > > Could someone let me know the speed rating for a 386 server? We also have > a 486 server and I'd be interested to know the difference. > > Cheers, > +Mark. > Our server machine's speed rating for a 386 server are as follows. Model Name Specification Speed rate --------------------------------------------------------------------------- Sony PCX-700DR7 486 33MHz EISA Bus SCSI 670M HD 915 Sony PCX-300E62 386sx 16Mhz ISA Bus SCSI 250M HD 102 Thank you. Masaaki Hoshino samba@sm.sony.co.jp Engineer System Engineering Div. Supermicro System Group Sony Corporation Tokyo, Japan
kenh@techbook.com (Ken Haynes) (02/15/91)
In article <91042.192216ALG106@psuvm.psu.edu> ALG106@psuvm.psu.edu writes: >The company I work for was thinking of getting a new fileserver - >we currently have a 33MHz 386. We were thinking of a 33MHz 486 EISA >bus machine. However, someone told us that the i486 has a bug in it >that shows up only when running Netware. > >First of all, is this true? If so, are there plans to fix it? > >Any suggestions you may have on a new fileserver would be helpful - >we were thinking of getting 2 1Gig drives (fault tolerance). > There are several i486 systems that are approved for use as a NW fileserver. Two are 33 Mhz, last I checked. Most are 25Mhz. There are reported problems using Bustech SCSI controllers on 486 systems running NW 3.1 with two differing types of ethernet NICs. So far, we haven't been able to isolate the problem to any one device in the machine, but we suspect the Bustech controller, since we have had other reports as well. Be sure to check the IMSP bulletins for approved hardware configurations first. Ken -- ****************************************************************************** Ken Haynes, Certified Netware Engineer Technical Support Product Mgr. 900 Support
karinc@isc.intel.com (Karin Coffee) (02/16/91)
In article <91042.192216ALG106@psuvm.psu.edu> ALG106@psuvm.psu.edu writes: >The company I work for was thinking of getting a new fileserver - >we currently have a 33MHz 386. We were thinking of a 33MHz 486 EISA >bus machine. However, someone told us that the i486 has a bug in it >that shows up only when running Netware. > >First of all, is this true? If so, are there plans to fix it? > >Any suggestions you may have on a new fileserver would be helpful - >we were thinking of getting 2 1Gig drives (fault tolerance). > I have been running NetWare 386 version 3.0 and now 3.1 on a 486 machine since last August. I am currently beta testing a 486 33 MHz EISA bus machine. The only problems we have had are with the EISA/SCSI adapter card (they die easily), but have had no problems with the 486. What type of bug have you heard about? I benchmarked 3 486 boxes. I looked at the HP Vectra, the Compaq SystemPro, and the generic Intel box. All three functioned with basically the same speed and functionality. We (of course) ended up with the Intel box since we didn't have to pay retail cost for it. My biggest gripe with the SystemPro is the proprietary hardware for the disk array. It's pretty spiffy how that works, but we like to fix our own hardware, and proprietary stuff doesn't go over too well. I really liked the Vectra, it was just too expensive. Good Luck! -=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= Karin Coffee Intel Supercomputer Systems Division Network/System Administration 15201 NW Greenbrier Parkway The LAN Lords Beaverton, OR 97201 karinc@isc.intel.com (503) 629-7693 =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-
todd@Quotron.COM (Todd_Booth) (02/19/91)
karinc@isc.intel.com (Karin Coffee) writes: >In article <91042.192216ALG106@psuvm.psu.edu> ALG106@psuvm.psu.edu writes: >>The company I work for was thinking of getting a new fileserver - >>we currently have a 33MHz 386. We were thinking of a 33MHz 486 EISA >>bus machine. However, someone told us that the i486 has a bug in it >>that shows up only when running Netware. If there are no bugs with the EISA 486..., why doesn't Novell certify them, as they do with at least 20 386 machines? --todd -- todd (booth) todd@quotron.com 213 302-4368
karinc@isc.intel.com (Karin Coffee) (02/23/91)
In article <todd.666899540@premises1> todd@Quotron.COM (Todd_Booth) writes: >karinc@isc.intel.com (Karin Coffee) writes: > Todd, if you are going to leave my header on the article, at least leave in what I wrote (or part of it). Thanks! What I wrote was that I hadn't heard of a bug, and I've been running NetWare 386 on a 486 25MHz machine since August. As of today, I have a 486 33MHz machine with an EISA bus ready to set up and beta test. Hopefully this time there won't be any hardware crashes. :-) >>In article <91042.192216ALG106@psuvm.psu.edu> ALG106@psuvm.psu.edu writes: >>>The company I work for was thinking of getting a new fileserver - >>>we currently have a 33MHz 386. We were thinking of a 33MHz 486 EISA >>>bus machine. However, someone told us that the i486 has a bug in it >>>that shows up only when running Netware. > >If there are no bugs with the EISA 486..., why doesn't Novell certify >them, as they do with at least 20 386 machines? > I don't know why Novell hasn't certified any, yet. I am currently using a beta version of a 486 33 EISA machine. To my knowledge, Novell has one that they are certifying. The certification process takes awhile. Who knows how many 486 33MHz EISA machines are attempting to get certified? -=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= Karin Coffee Intel Supercomputer Systems Division Network/System Administration 15201 NW Greenbrier Parkway The LAN Lords Beaverton, OR 97201 karinc@isc.intel.com (503) 629-7693 =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-
james@bigtex.cactus.org (James Van Artsdalen) (02/26/91)
In <1258@pcgw.sm.sony.co.jp>, samba@sm.sony.co.jp (Masaaki Hoshino) wrote: > Now, Our 486 33MHz machine has the following trouble with Netware 386 v3.1. > 1. Pressing Caps Lock, Numeric Lock, or Scroll Lock keys will cause > a lost interrupt detected message as follows. > "Interrupt controller detected a lost hardware interrupt" > 2. When using NE3200 EISA Bus Master LAN card with our SCSI disk driver > (NLM), the same message is often indicated on the server console. > I asked Novell IMSP division about it. They said this problem is caused by > 486 CPU with Programmable Interrupt Controller(PIC), and the above messages > will occur with Netware 3.1 on all 486 systems regardless of bus type( ISA, > EISA, MCA) with varying frequency. These should not happen with a C step 486 at all, I believe. Even with a B step 486, I think Intel published an errata with suggested hardware fixes to put on the motherboard. The bug is independent of bus type and CPU frequency. It isn't Netware specific either - the system BIOS usually doesn't say anything when an interrupt is lost. Even with a B step 486, some interrupt controllers will not confused by the behavior of the CPU, so just because you have B step doesn't mean you will get errors. Have your engineers contact Intel about the tech note on this 486 bug. -- James R. Van Artsdalen james@bigtex.cactus.org "Live Free or Die" Dell Computer Co 9505 Arboretum Blvd Austin TX 78759 512-338-8789
samba@sm.sony.co.jp (Masaaki Hoshino) (02/28/91)
In article <55206@bigtex.cactus.org>, james@bigtex.cactus.org (James Van Artsdalen) writes: > > I asked Novell IMSP division about it. They said this problem is caused by > > 486 CPU with Programmable Interrupt Controller(PIC), and the above messages > > will occur with Netware 3.1 on all 486 systems regardless of bus type( ISA, > > EISA, MCA) with varying frequency. > > These should not happen with a C step 486 at all, I believe. Even > with a B step 486, I think Intel published an errata with suggested > hardware fixes to put on the motherboard. The bug is independent of > bus type and CPU frequency. It isn't Netware specific either - the > system BIOS usually doesn't say anything when an interrupt is lost. > > Even with a B step 486, some interrupt controllers will not confused > by the behavior of the CPU, so just because you have B step doesn't > mean you will get errors. > > Have your engineers contact Intel about the tech note on this 486 bug. Yes. Our hardware engineers did. He sais that we use C step 486 and next release of 486 ( D step ?) will fix this problem. The 1st and 2nd problem I said before are not so serious problem for us, because the server does not ABEND. :-> Thank you. P.S. Sony PCX-700(486 33MHz EISA PC, SCSI HD) is domestic model in Japan. not available in U.S.A or another foreign country. *************************************************************************** Masaaki Hoshino Sony corp. Tokyo Japan samba@sm.sony.co.jp ***************************************************************************
root@xstor.com (Superuser) (03/22/91)
In article <1991Feb15.142012.1978@techbook.com> kenh@techbook.com (Ken Haynes) writes: >There are several i486 systems that are approved for use as a NW fileserver. >Two are 33 Mhz, last I checked. Most are 25Mhz. There are reported >problems using Bustech SCSI controllers on 486 systems running NW 3.1 with >two differing types of ethernet NICs. So far, we haven't been able to >isolate the problem to any one device in the machine, but we suspect the >Bustech controller, since we have had other reports as well. I'm about to invest in a Compaq Systempro, and I intended to use the Bustech Controller with some Seagate Elite drives. Anything else I should know before I try to go live? David Potter, Network Administrator Storage Dimensions dpotter@xstor.com ..!xstor!stdim!dpotter 408/879-3368 FAX: 408/879-3382