Re: fdisk and cylinders > 1024
> I have a 1.0G IDE disk. (Micropolis 2210A) I am told that there should be
> no problem for linux fdisk to partition the clinders above 1024. So long as
> the boot partition remains below 1024. I am running Debian (its really great
> folks - good work).
> Is there something special about Debian that causes this problem?
No, Slackware fdisk reports similar problems.
> When I try to use cylinders above 1024, fdisk indicates that partition logical
> beginnings and endings do not match physical beginnings and endings.
> partition4: cyl 1025 to cyl 1891 Phys: (63,15,1023) to (63,15,866)
> Any ideas?
> (I'll even accept suggestions to go to SCSI but then I would appreciate a
> controller card suggestion (I have a VLB IDE card now, and one other IDE
> harddisk for DOS ) please).
I'm using an Adaptec 1542-C and a Seagate N12400 (2.1 GB formatted
capacity - 2048 cylinders!) so it isn't a problem with the IDE or SCSI
standards. I'm willing to bet that something in fdisk is complaining.
Mostly because there is a 10-bit limit on a field in the PC partition table
IMO, the Linux fdisk should not enforce this limit unless the partition
type is a DOS partition type.
I just ignored the complaint about the logical/physical discrepancy, and
left the paartition table as I configured it. Booting DOS and examining the
table with Norton shows that the extra bits of precision *are* stored on
the disk, and Linux picks up all the bits when actually mounting the
partitions. It's only some software that complains about the extra bits.