karp-brad@CS.YALE.EDU (Brad Karp) (10/05/90)
I recently purchased an Adaptec AHA-1542B SCSI controller, a Maxtor 8760S SCSI drive, and an Archive 2150s SCSI tape drive. My system's configuration before purchasing these items was with a single MFM standard AT disk controller and two MFM hard disks attached to this MFM controller. I intended to back up the contents of my two MFM drives using my new SCSI tape drive, and then to restore the data from tape to the new Maxtor SCSI drive. To do this, though, I needed to install the AHA-1542B and 2150s along with my two old MFM drives. So, I removed the jumper on the AHA-1542B which enables the floppy drive controller functions to prevent a conflict with my existing AT MFM combination hard/floppy controller, set the SCSI ID on the tape drive to 2, the SCSI ID on the AHA-1542B to 7, and installed them in my machine. I then built a new kernel with the HPDD configured properly; a "standard AT controller" as my "primary" controller, and a "secondary SCSI controller with a tape drive attached to it." I installed this kernel. Unfortunately, the kernel I produced hangs my machine. My machine makes it to Booting the UNIX system ..., reads /unix in, and then hangs when it starts executing the kernel (right when the screen normally clears and presents the memory available, etc.). I am almost positive that my machine has no interrupt conflicts or I/O address range conflicts; my AHA-1542B is configured to use interrupt 11, while my MFM controller is configured to use interrupt 14. I have the MFM controller configured to use the standard I/O port range, and my AHA-1542B to use the factory-set 330h I/O address. Other cards in my system: Two serial/one parallel I/O card, Paradise VGA 1024, Logitech Bus Mouse card. Can anyone suggest anything that I might try to fix this problem? Thanks. -- Brad Karp, (203) 436-3060 (voice) | The views expressed in the text my 386: karp%softshop.uucp@cs.yale.edu | you have just perused are not my via Yale CS Dept: karp-brad@cs.yale.edu | own; rather, they are those of the Box 2443 Yale Station, New Haven, CT 06520 | heavenly muse who sings in me. -JM
itwaf@wafbox.UUCP (Bill Fulton) (10/06/90)
In article <26531@cs.yale.edu> karp-brad@CS.YALE.EDU (Brad Karp) writes: >I recently purchased an Adaptec AHA-1542B SCSI controller, a Maxtor 8760S >SCSI drive, and an Archive 2150s SCSI tape drive. >I intended to back up the contents of my two MFM drives using my new SCSI >tape drive, and then to restore the data from tape to the new Maxtor SCSI >drive. To do this, though, I needed to install the AHA-1542B and 2150s along >with my two old MFM drives. >Can anyone suggest anything that I might try to fix this problem? Thanks. >Brad Karp, (203) 436-3060 (voice) | The views expressed in the text I have just installed a 1542B controller, SCSI disk, and SCSI tape on my system at work, where the SCSI controller is the secondary and an ESDI controller is the primary. Flushed with victory, and hoping that I'm not posting something already posted ... ** IMPORTANT Thingies to know when installing a AHA-154x as secondary cu ** Following modifications should be made to factory defaults of 154x: 1) Change the DMA channel from default 5 to * 6 * (*NOTE* that this requires changing *three* jumpers!) If you don't do this - on boot the system will display the Unix boot message and then just hang. Sound familier? 2) Be sure that the BIOS addresses of the two controllers do not conflict. (Mine did, changed default DC000 to D8000) Following are other factoids: 3) LU 0 - 3 are (arbitrarily) reserved for disks LU 4 - 7 are (arbitrarily) reserved for tape devices I made the disk LU0 and the tape LU4. (Pretty radical, eh?) 4) Use kconfig and reconfig HPDD. 5) Run "sysadm addharddisk". This utility was not QA'd worth a damn. For example, note that you must define an "active" partition on the 2nd disk. Also, you'll loose at least 2 of the cylinders. Just let it have its way. 6) You will probably (*very* probably) need to manually create the /dev entries for the tape. Use "mknod <name> c 41 <#>". See the "Maintenance" section for the actual names and the minor numbers. Example: mknod /dev/tape c 41 0 (Well, the last number is either 0 or 3. One means "no rewind". Sorry, forgot which is which) 7) I had to manually add the new partition(s) to /etc/fstab. 8) Note that the SCSI devices must be up when you boot the machine - they won't be recognized if you power them up after boot. Mine are working damn well. Thank goodness we've got a tech support contract; my manager wasn't at all happy with "gee - it SHOULD work". Minor flames - ISC should have gotten this info out! - The addharddisk utility should not treat all disks like boot disks. Also, it should not crash. When I ran "addharddisk", the disk params were pretty zany. (64 Heads!) I just let it think that, because; 1) It would crash if I tried to change it. 2) I figured that it didn't matter much for SCSI. (Logical devices - what a concept!) BTW - Does anyone know a good, safe way to speed up the mount's/umount's ? For example, a way override the mount's fsck? I know it's important; but when you're just doing some maint work it's not good to wait for a 600 Mb fsck. Also, why does it take so long to unmount? I need a "fastboot" and "fasthalt". I saw the call to "fsck" in /etc/mountall, but just renaming or spoofing /etc/fsck doesn't seem to help. Hope it helps, waf Real men don't use tech support. I'm a wimp.