LBA and CHS issue redux for disk imaging with parted

[Bill Catlan]

Hello All,

I am trying to gather a good conceptual and practical understanding of the 
implications of the access mode used by a disk on a given system.  I have 
already read the parted docs and looked through the archives (and discovered the 
CHS discussions - there is a good one at 
http://lists.gnu.org/archive/html/bug-parted/2004-06/msg00056.html).

To start this discussion, I want to ask the list to comment on what are the 
implications of imaging the first partition on the first drive (ie, /dev/hda1) - 
call it D1P1 - on Box A, where Box A's BIOS access method set to CHS, in the 
following scenarios:

1) What will happen when I restore the image of D1P1 to a new drive on Box A, 
assuming everything is the same on Box A, except at restore time the restore 
drive - call it D2P1 - has been swapped in for D1P1 and is now the first drive 
and first partition on Box A?  Assume that the image of D1P1 is stored on some 
third drive - say D3P1.  I assume here that I have to verify the CHS geometry of 
D2P1, since it is starting out blank. Is that correct?  Does D2P1's geometry 
need to match D1P1, or will parted be able to restore the image of D1P1 to D2P1 
despite a differing geometry on D2P1?

2) What will happen when I restore the image of D1P1 from Box A to a new drive 
on Box B (a new computer with a different BIOS)?  If I set Box B's BIOS to 
access the first drive using CHS, will I be able to easily restore the image of 
D1P1 to the first drive in Box B?  Is geometry stored in the image at all - or 
is it merely detected at restore time?

3) Same as scenario 1, but the source drive on Box A (D1P1) and the destination 
drive (D2P1) are accessed using LBA mode as set in the BIOS.

4) Same as scenario 2, but the destination drive on Box B (D2P1) is accessed 
using LBA mode as set in the BIOS.

Thanks in advance for any comments!

Bill