Imaging SCSI hard disks

From: Paul Koning <pkoning_at_equallogic.com>
Date: Wed Sep 1 08:25:06 2004

>>>>> "Zane" == Zane H Healy <healyzh_at_aracnet.com> writes:

 Zane> My idea is actually to first dd the HD to a disk image, then
 Zane> prep that disk image under SIMH or E11. Copy the data I want
 Zane> on the HD to the image using the emulator, and then dd it back
 Zane> to the HD. Logically it should work just fine. With RT-11,
 Zane> you shouldn't even need to dd the HD to a disk image, just
 Zane> create whatever size disk image and dd it out there. The other
 Zane> OS's would be a bit more picky.

Not necessarily. If you tell the simulator that it has an MSCP disk,
the OS should just accept whatever size it is, because the rule of
MSCP is that the disk tells the OS its size and the OS makes no
assumptions. Certainly RSTS does this correctly.

As for "geometry" -- MSCP doesn't have any that the OS can see, or at
best it has just some vague hints that actually serve no real
purpose. That certainly fits with modern drives -- anything built in
the last 10 years or so has a software geometry that has no connection
to reality, and a real geometry that's (a) not visible, (b) much more
complex than tracks and cylinders of fixed size. Look at IDE disks,
for example; they all claim to have 15 heads and 63 sectors, or
whatever is the max in the old IDE conventions. But those numbers
have nothing to do with reality, they exist only for backward
compatibility.

               paul
Received on Wed Sep 01 2004 - 08:25:06 BST