dick@cimcor.mn.org (Dick Schlotfeldt) (09/22/90)
I used MS-DOS 3.3 fdisk and format to partition and format a Wren IV with an ST-02 SCSI host adapter in a 12 MHz 286 machine. The partitions were C (boot), D, E, and F. It worked fine. After moving the disk drive to a 25 MHz 386 machine I tried booting with an Adaptec 1542B adapter. The SCSI bios found the drive (id 0) but could not find the system files. A boot from floppy in A: was ok and DIR C: found everything just fine. However DIR D: (also E: and F:) reported partitions not found (or non-existent or whatever the word is). 1. Is a SCSI drive sensitive to the host adapter (or the computer)? If so, why? But the controller (where really low level stuff happens) is on the drive! 2. Do I really have to re-partition and re-format the drive? 3. Or is something else causing my problem? Thanks for any help. Dick Schlotfeldt dick@cimcor.mn.org (612) 785-4621 (work) (612) 490-1252 (home)
als@bohra.cpg.oz (Anthony Shipman) (09/24/90)
In article <1990Sep21.191238.23217@cimcor.mn.org>, dick@cimcor.mn.org (Dick Schlotfeldt) writes: > I used MS-DOS 3.3 fdisk and format to partition and format a Wren IV > with an ST-02 SCSI host adapter in a 12 MHz 286 machine. The partitions > were C (boot), D, E, and F. It worked fine. > > After moving the disk drive to a 25 MHz 386 machine I tried booting with > an Adaptec 1542B adapter. The SCSI bios found the drive (id 0) but > could not find the system files. A boot from floppy in A: was ok and > DIR C: found everything just fine. However DIR D: (also E: and F:) > reported partitions not found (or non-existent or whatever the word is). > > 1. Is a SCSI drive sensitive to the host adapter (or the computer)? If so, why? > But the controller (where really low level stuff happens) is on the drive! > 2. Do I really have to re-partition and re-format the drive? > 3. Or is something else causing my problem? Yes you must reformat the drive. I had the same problem when changing from a Future Domain controller to the Adaptec. The release notes for the AHA1542A mention that if you change controllers the behaviour is "erratic". What an understatement! The Adaptec controller maps my drive to 64 heads/cyl, 32 sectors per track. The FT controller had 38 heads per cyl and 17 sectors per track. The result is that the disk data reads as totally garbled. For DOS you could get away with rerunning fdisk and "format c:". For UNIX a low level format is needed as the controller stores format data on the drive, which Interactive Unix reads. -- Anthony Shipman ACSnet: als@bohra.cpg.oz.au Computer Power Group 9th Flr, 616 St. Kilda Rd., St. Kilda, Melbourne, Australia D
lawson@iphase.UUCP (David Lawson PER ) (09/24/90)
In article <1990Sep21.191238.23217@cimcor.mn.org> dick@cimcor.mn.org (Dick Schlotfeldt) writes: >I used MS-DOS 3.3 fdisk and format to partition and format a Wren IV >with an ST-02 SCSI host adapter in a 12 MHz 286 machine. The partitions >were C (boot), D, E, and F. It worked fine. > >After moving the disk drive to a 25 MHz 386 machine I tried booting with >an Adaptec 1542B adapter. The SCSI bios found the drive (id 0) but What is probably happening is that the different boards' BIOS extensions handle drive geometry in different manners. What will happen, since DOS talks to the BIOS in cylinders, heads, and sectors, the BIOS has to translate this to a logical address for SCSI. Some boards will assume a specific geometry of the drive, say 16 heads, 32 sectors per track, and vary the number of cylinders according to the size of the drive. The next board may assume its own geometry which would cause problems if they do not match. The only solution is to completely back up the drive with the original controller and then set it up again with fdisk on the new controller and restore the files.