I have two supposedly identical SunFire V240 servers running Solaris 9,
each with four external HP DAT 40 drives off a PCI SCSI card. Both
systems have the drives with SCSI IDs 0,1,2, and 3 with the SCSI chain
terminated after the drive with SCSI ID 3. The primary server loads
its drives as /dev/rmt/0, /dev/rmt/1, /dev/rmt/2, and /dev/rmt/3. For
some reason, the second server is placing the drive with SCSI ID 3 as
/dev/rmt/9.

Crazier yet, it seems that any time I set a drive to a SCSI ID higher
than 2, Solaris sets it up as /dev/rmt/9 or higher (e.g., /dev/rmt/10)
during a reconfigure boot. Once setting the drive(s) to SCSI IDs back
in the 0-2 range, they go back to /dev/rmt/0 - /dev/rmt/2.

Any ideas on why these boxes are behaving differently, or where to look
for the difference? I want both servers to recognize the drives on the
same /dev/rmt aliases to avoid rewriting the backup scripts for each
box.

Michael J. Perkins