> > Well, no. If I partitionate a new drive on the secondary interface and
> > move it to the primary, it will change it's geometry. Not good.
>
> Not true. Look further on, at the partition check.
> Thats where the translated geometry is reported.
Obviously you have not tested.
I'm doing this right here right now, please follow my instructions:
Take a HD with physically more than 1024 cylinders, make sure you don't
have anything important on it.
Put it on the secondary interface.
In BIOS make sure AUTO, LBA is enabled on all drives.
Boot from some Linux boot/root diskette. The Debian one is perfect.
As root do "dd if=/dev/zero of=/dev/hdc bs=512 count=1".
Reboot on the same Linux diskette (this reboot is important).
Note the CHS value.
As root do "fdisk /dev/hdc", and make some partitions.
Shutdown.
Move HD to primary interface
Boot from the same Linux diskette.
Note the CHS value (ops it is not the same).
As root run "fdisk /dev/hda", print the partitions. By this I hope it
should be obvious for everyone that the geometry has changed.
Convinced?
> Not that it matters. Geometries are used only by fdisk,
> not normal system operation.
And lilo, the bootloader. Rather important.
/Ryde