Hi.  I guess I'm suffering from one of those infamous SCSI bugs (details
below) and I'm wondering whether there's any chance the problem is fixed
or will be fixed soon.  Unfortunately I don't really have any other spare
SCSI disk so I'm kind of stuck.  Plus, installing the damn thing took
eight hours.  EIGHT HOURS!  And that's just the time from "yes, install
these packages I have selected" until it was done. (I was installing from
an oldskool Appletalk network (that's 19200 bps isn't it?) which might
explain it)

  The disk in question is an IBM DSAS-3540 which is jumpered according to
an FAQ on the IBM website regarding using the disk in a Mac.

 The computer is a Mac IIsi (68030 @ 20 MHz, 5 MB RAM) (with an Apple
Portrait Display which I now know has a resolution of 640x870 thanks to
the NetBSD kernel boot messages).

  Speaking of boot messages, here is the relevant part:


scsibus0: waiting 2 seconds for devices to settle...
sd0 at scsibus0 target 0 lun 0: <CO\000NER, CP\000040\000-40\000b-3\0005, 2.\0001> SCSI0 0/direct fixed
sd0: 41040 KB, 1024 cyl, 2 head, 40 sec, 512 bytes/sect x 82080 sectors
sd1 at scsibus0 target 1 lun 0: <IB, DS\000S-3\00040 \000   \000W, S4\000Y> SCSI0 0/direct fixed
sd1: 516 MB, 3840 cyl, 3 head, 91 sec, 512 bytes/sect x 1056768 sectors
boot device: sd1
root on sd1a dumps on sd1b
sd1(ncrscsi0:1:0) Sense Error Code 0x6d
sd1(ncrscsi0:1:0) Sense Error Code 0x6d
sd1(ncrscsi0:1:0) Sense Error Code 0x6d
sd1(ncrscsi0:1:0) Sense Error Code 0x6d
sd1(ncrscsi0:1:0) Sense Error Code 0x6d
sd1(ncrscsi0:1:0) Sense Error Code 0x6d
sd1(ncrscsi0:1:0) Sense Error Code 0x6d
sd1(ncrscsi0:1:0) Sense Error Code 0x6d
sd1(ncrscsi0:1:0) Sense Error Code 0x6d
sd1(ncrscsi0:1:0) Sense Error Code 0x6d
sd1(ncrscsi0:1:0) Sense Error Code 0x6d
sd1(ncrscsi0:1:0) Sense Error Code 0x6d
no file system for sd1 (dev 0x408)
cannot mount root, error = 79
root device (default sd1a):

The "CO\000NER, CP\000040\000-40\000b-3\0005, 2.\0001" is really a CONNER
CP3040A.

The "IB, DS\000S-3\00040 \000 \000W, S4\000Y" is the IBM drive in
question.

  The bug is quite interesting too.  Compare the strings given in the boot
process with what the drive probably really says (for clarity, I've
replaced the \000s with percent signs) :

CO%NER, CP%040%-40%b-3%5, 2.%1

CONNER  CP3040A


IB%, DS%S-3%40 %   %W, S4%Y

IBM  DSAS-3540

("IB" = "IB%" makes sense if we suppose that trailing nulls don't count)

  I know that "IBM, DSAS-3540 -W, S4-Y" doesn't seem right either, but
that is actually what the drive reports to HDToolBox on my Amiga.

  Anyway, it seems that every fourth character is replaced with a null;
except for the first null, which seems to be off by one.

  And a database of un/successful installs on specific hardware
configurations might be a good idea, I think.  Please use my data if you
want.


^M