teoh@uicbert.eecs.uic.edu (08/20/90)
I have a seagate 80MB w/ MFM interface running at 28ms hardrive. But many people have pointed out that there's hardly any room for a complete SVR4 package. So, I decided to add another faster drive (~15ms) but of different interface type. Is it possible to allow 2 different types of hard disks in a pc UNIX environment (using 2 different controllers)? If so, what are comments on AT IDE vs ESDI drives? Is it true to say that the AT IDE drive is dependent on AT architecture machine? How machine independent is the ESDI drive outside the AT arena? I was planning to use the slower 80MB as secondary drive for back up purposes or storing files of lower priority, and leaving all the executables in the faster drive. Will there be any substantial loss of performance w/ this configuration? Or, is it better just to have a faster hardrive. All comments are welcome, your help is greatly appreciated. -------------------------------------------------------------------------- Hai-Chuan Teoh * Email: teoh@uicbert.eecs.uic.edu Univ of Illinois * Tel: (312) 930-1942 Electrical Engr & Computer Sci * --------------------------------------------------------------------------
teoh@uicbert.eecs.uic.edu (08/20/90)
Message-ID: <1990Aug19.182733.8095@uicbert.eecs.uic.edu> Organization: University of Illinois at Chicago Distribution: comp Date: Sun, 19 Aug 90 18:27:33 GMT Lines: 25 I have a seagate 80MB w/ MFM interface running at 28ms hardrive. But many people have pointed out that there's hardly any room for a complete SVR4 package. So, I decided to add another faster drive (~15ms) but of different interface type. Is it possible to allow 2 different types of hard disks in a pc UNIX environment (using 2 different controllers)? If so, what are comments on AT IDE vs ESDI drives? Is it true to say that the AT IDE drive is dependent on AT architecture machine? How machine independent is the ESDI drive outside the AT arena? I was planning to use the slower 80MB as secondary drive for back up purposes or storing files of lower priority, and leaving all the executables in the faster drive. Will there be any substantial loss of performance w/ this configuration? Or, is it better just to have a faster hardrive. All comments are welcome, your help is greatly appreciated. -------------------------------------------------------------------------- Hai-Chuan Teoh * Email: teoh@uicbert.eecs.uic.edu Univ of Illinois * Tel: (312) 930-1942 Electrical Engr & Computer Sci * --------------------------------------------------------------------------
rcd@ico.isc.com (Dick Dunn) (08/22/90)
teoh@uicbert.eecs.uic.edu writes: > Is it possible to allow 2 different types of hard disks in a pc UNIX > environment (using 2 different controllers)? If so, what are comments > on AT IDE vs ESDI drives? It is possible to mix controller types, as long as you apply the care you'd expect, to avoid interrupt and I/O port conflicts. IDE and ESDI are both fairly inexpensive ways to get disks a bunch faster than old AT ST-506 (MFM) drives. IDE is usually cheaper; IDE drives tend to be smaller. >...Is it true to say that the AT IDE drive is dependent on AT architecture > machine?... Yes, in the sense that the "IDE" interface is mostly a way to extend the relevant parts of the "ISA" (AT) bus. That's not to say you can't make an IDE controller for some other bus; it's just that IDE is most closely matched to ISA. The win with either IDE or ESDI--or, to look at the other side, the loss in the old ST506 interface--is that the old style has a separate drive and controller with a rather slow interface between the two. It is this interface (which is standard) that limits the data rate. ESDI keeps the drive/controller split but ups the data rate. IDE puts the controller on the drive; the maximum IDE data rate is determined by the ISA bus. One thing to keep in mind about IDE: The interface matches that of a primary disk controller on the AT bus, and you get a max of two IDE drives. After that, to add more disks you'd have to add them on a "secondary" AT-style controller (standard or ESDI) or SCSI. > I was planning to use the slower 80MB as secondary drive for back up > purposes or storing files of lower priority, and leaving all the executables > in the faster drive. Will there be any substantial loss of performance w/ > this configuration?... Performance shouldn't be that much of a problem--just move things around if the slow disk is getting too much use. Even staying with a slow disk, you can improve performance if you've got an old 2:1 or 3:1 controller by going to a 1:1 controller. These are in the neighborhood of $100. -- Dick Dunn rcd@ico.isc.com -or- ico!rcd Boulder, CO (303)449-2870 ...Are you making this up as you go along?
palowoda@fiver (Bob Palowoda) (08/22/90)
From article <1990Aug21.235008.13039@ico.isc.com>, by rcd@ico.isc.com (Dick Dunn): > One thing to keep in mind about IDE: The interface matches that of a > primary disk controller on the AT bus, and you get a max of two IDE > drives. After that, to add more disks you'd have to add them on a > "secondary" AT-style controller (standard or ESDI) or SCSI. Ahh, somebody mentioned the magic word "secondary" controller. Has anyone ever tested this setup. Say with the second disk controller on interupt 15? ISC? ESIX? or SCO UNIX? I have run some disk bench tests on a Conner and Maxtor IDE drive (under ESIX) and I would perfer to go the IDE over SCSI. It seems to be a better cost performance ratio. I wouldn't mind moveing my ESDI drives over to the secondary controller but I'm not sure how to set this up. What's more amazeing is in the ATT manuals they reference a secondary controller (or at least reference to the device names) but I cannot find any setup info? ---Bob -- Bob Palowoda palowoda@fiver | *Home of Fiver BBS* Home {sun}!ys2!fiver!palowoda | 415-623-8809 1200/2400 {pacbell}!indetech!fiver!palowoda | An XBBS System Work {sun,pyramid,decwrl}!megatest!palowoda| 415-623-8806 1200/2400/19.2k TB+
stevewa@upvax.UUCP (Steve Ward) (08/23/90)
In article <1990Aug22.061223.15564@fiver> palowoda@fiver (Bob Palowoda) writes: > Ahh, somebody mentioned the magic word "secondary" controller. Has anyone >ever tested this setup. Say with the second disk controller on interupt 15? >ISC? ESIX? or SCO UNIX? I have run some disk bench tests on a Conner and >Maxtor IDE drive (under ESIX) and I would perfer to go the IDE over SCSI. >It seems to be a better cost performance ratio. I wouldn't mind moveing my >ESDI drives over to the secondary controller but I'm not sure how to set >this up. What's more amazeing is in the ATT manuals they reference a >secondary controller (or at least reference to the device names) but >I cannot find any setup info? I called Western Digital on this very subject...I bought an ESDI controller, and wanted to use my old ST506 as a secondary controller. They *assured* me that even if I got the addressing and interrupts changed, it wouldn't work. If that's true, then why do all the controller cards have secondary address jumpers???? Is there anyone out there who has two controllers (SCSI doesn't count!) running in the same box happily??? Steve -- | Steve Ward Jr. appears courtesy of | stevewa@upvax.UUCP | | Univ. of Portland, Portland, OR | !tektronix!upvax!stevewa | | (insert disclaimer here) | upvax!stevewa@tektronix.TEK.COM | | --If all else fails, try: tektronix.TEK.COM!upvax!stevewa@uunet.uu.net |
dougp@ico.isc.com (Doug Pintar) (08/23/90)
In article <1990Aug22.061223.15564@fiver> palowoda@fiver (Bob Palowoda) writes: > > Ahh, somebody mentioned the magic word "secondary" controller. Has anyone >ever tested this setup. Say with the second disk controller on interupt 15? >ISC? ESIX? or SCO UNIX? Yes, I have. The ISC HPDD will support a secondary AT controller, given 1) secondary I/O addresses, 2) if it has a floppy controller, it must be disabled, and 3) [the biggie] a secondary interrupt (usually 15). The reason 2 is of concern is that very few controllers offer it. The only ones I've seen to date are the Adaptec 232x ESDI and 237x RLL. I have used a WD-1007 (Compaq flavor) with a 300 MB ESDI drive as primary and an Adaptec 2320 with some other flavor ESDI drive on it. Works fine, BUT... when you get both drives going at the same time (the AT controller is PIO, remember) the CPU starts to get swamped handling interrupts and moving data to/from the controllers. I found my max (raw, large-buffers) transfer rate on either drive alone was around 850 KB/sec and that both together topped out at an aggregate of about 1.3 MB/sec. This is on a 20 MHz Compaq, faster CPU would help with the interrupt overhead, but probably not with the 16-bit string moves to/from the controllers. The HPDD will also support a SCSI adapter as a secondary controller (the machine I'm writing this on has 1.4 GB of Maxtor 8380s hanging off an Adaptec 1542B in addition to the ESDI primary). For pure performance, having the first-party DMA of the 1542 off-loads BUNCHES of overhead from the CPU. I just wish I could run mine faster than 5.7 MB/sec burst rate, sniff sniff... DLP
palowoda@fiver (Bob Palowoda) (08/23/90)
From article <1990Aug22.194548.25743@ico.isc.com>, by dougp@ico.isc.com (Doug Pintar): > In article <1990Aug22.061223.15564@fiver> palowoda@fiver (Bob Palowoda) writes: >> >> Ahh, somebody mentioned the magic word "secondary" controller. Has anyone >>ever tested this setup. Say with the second disk controller on interupt 15? >>ISC? ESIX? or SCO UNIX? > > Yes, I have. The ISC HPDD will support a secondary AT controller, given > 1) secondary I/O addresses, > 2) if it has a floppy controller, it must be disabled, and > 3) [the biggie] a secondary interrupt (usually 15). > The reason 2 is of concern is that very few controllers offer it. The only > ones I've seen to date are the Adaptec 232x ESDI and 237x RLL. I have used Some good news here. Actually I had the setup almost done on ESIX (rev C) with an Adaptec 2370 RLL controller as the secondary controller. I did a low level format on the first controller and moved it over to the second with the proper int and base address set. Made the correct node's in the dev directories. I got cold feet when it came to makeing the partitions with 'mkpart' because I was wasn't sure if it would mess up my root partition if I did it wrong. I wasn't even sure if there was kernel support for the secondary controller. What is the 'mkpart' command line to do this? Anyone tried this on ESIX? Any help would be appreciated. The additional storeage goes over to more drive space for PD source on the bbs. ---Bob -- Bob Palowoda palowoda@fiver | *Home of Fiver BBS* Home {sun}!ys2!fiver!palowoda | 415-623-8809 1200/2400 {pacbell}!indetech!fiver!palowoda | An XBBS System Work {sun,pyramid,decwrl}!megatest!palowoda| 415-623-8806 1200/2400/19.2k TB+
campbell@redsox.bsw.com (Larry Campbell) (08/29/90)
In article <1106@upvax.UUCP> stevewa@upvax.UUCP (Steve Ward) writes:
-
-Is there anyone out there who has two controllers (SCSI doesn't count!)
-running in the same box happily???
I have a WD1007A and a WD1007V (both ESDI) running in the same box -- the
box I'm typing on right now. It works fine. I'm running ISC 386/ix
V2.0.2. I did not have to disable the floppy part of the second controller,
either -- I just set its I/O ports to the secondary addresses. (I have not,
however, hooked a floppy driver up to it to see if it would actually work.)
The only annoyance about this is that the WD1007A *must* be the boot
controller, since you can't change its IRQ line, and ISC insists that
controller 0 be on IRQ14 and controller 1 on IRQ 15.
--
Larry Campbell The Boston Software Works, Inc.
campbell@redsox.bsw.com 120 Fulton Street
wjh12!redsox!campbell Boston, MA 02109jdeitch@jadpc.cts.com (Jim Deitch) (08/29/90)
In article <1586@redsox.bsw.com> campbell@redsox.bsw.com (Larry Campbell) writes: >In article <1106@upvax.UUCP> stevewa@upvax.UUCP (Steve Ward) writes: >- >-Is there anyone out there who has two controllers (SCSI doesn't count!) >-running in the same box happily??? > >I have a WD1007A and a WD1007V (both ESDI) running in the same box -- the >box I'm typing on right now. It works fine. I'm running ISC 386/ix >V2.0.2. I did not have to disable the floppy part of the second controller, >either -- I just set its I/O ports to the secondary addresses. (I have not, >however, hooked a floppy driver up to it to see if it would actually work.) > >The only annoyance about this is that the WD1007A *must* be the boot >controller, since you can't change its IRQ line, and ISC insists that >controller 0 be on IRQ14 and controller 1 on IRQ 15. >-- I have 2 WD1006MV-2's in the same box with no problems. I had to cut a trace and solder a wire on one for IRQ 15 and set one of the floppy controllers to the secondary address. Running ISC 2.0.2 and SCO Xenix 2.3.X and under SCO Unix. Jim -- UUCP: {nosc ucsd hplabs!hp-sdd}!crash!jadpc!jdeitch ARPA: crash!jadpc!jdeitch@nosc.mil INET: jdeitch@jadpc.cts.com
sl@van-bc.wimsey.bc.ca (Stuart Lynne) (08/29/90)
In article <1586@redsox.bsw.com> campbell@redsox.bsw.com (Larry Campbell) writes: >The only annoyance about this is that the WD1007A *must* be the boot >controller, since you can't change its IRQ line, and ISC insists that >controller 0 be on IRQ14 and controller 1 on IRQ 15. Ever heard of a soldering iron? Tack solder a piece of wire wrap wire onto the appropriate spot on the board and either cut the trace going to the edge connector or cover it with a piece of non conductive tape. Run the wire to another board that has the interrupt line and attach. (Serial cards are good candidates for this, they usually have a large range interrupt options and often selection is made with a jumper. The jumper post is convienent for wiring wrapping.) -- Stuart.Lynne@wimsey.bc.ca ubc-cs!van-bc!sl 604-937-7532(voice)