> I would never use dd to copy complete installations unless the discs
> were identical and the machines' BIOSes had the same idea of the disc
> geometry. YMMV.
When I did my "brain transplant" from my old to my new PC I was
very concerned about this, especially since many people, like you,
cite disk geometry as the reason why one can't simply use dd to
copy partitions.
Having given the matter more thought, I wonder if the disk geometry
argument isn't specious. I would expect disk geometry to be a
spoiler only if C/H/S addressing were embedded in the partition
data, e.g, if individual clusters were identified by their C/H/S
coordinates.
This is not the case with Linux, nor as far as I know Windows 98,
which both use logical block addressing rather than C/H/S. As far
as I know this is not the case with DOS either: C/H/S values are
present in only one place, namely the partition table while the dos
partition itself uses logical addressing to identify clusters.
This would seem to imply that so long as you make byte for byte
copies of the individual partitions rather than the whole disk (e.g.,
hda1, hda2, hda3 instead of hda), disk geometry shouldn't be an
issue. This is born out by my experience copying individual dos
partitions to linux files e.g, by using:
dd if=/dev/hda1 of=dos_partition.raw
I am able to mount these files via the loopback device and access
the contents of the dos partition without a problem. Since files
don't have any "disk geometry", let alone the "right" one, this
shouldn't work, it seems to me, if what people say about disk
geometry were really true.
I don't mean to be rude, but I really don't see why disk geometry is
an issue. Perhaps you can help me understand why it is. Have I
missed something? Have you personally had bad experiences
using dd to copy partitions?
