Re: [ext2resize] Ext2resize for big-endian (Ultrasparc)

[Andrew Clausen]

Dave Wapstra wrote:
> Here's the output from 1.4.11-pre1:
> 
> Using /dev/sda
> (parted) print
> Warning: The disk CHS geometry (8637,64,32) does not match the geometry 
> stored on the disk label (4926,27,133).
> Ignore Cancel ? y
> Ignore Cancel ? n
> Ignore Cancel ?
> Ignore Cancel ?
> Ignore Cancel ?
> Ignore Cancel ?
> Ignore Cancel ?
> Ignore Cancel ? q
> Ignore Cancel ?
> Ignore Cancel ?
> Ignore Cancel ?
> Ignore Cancel ?
> [1]+  Stopped                 ./parted/parted

ROFL!  You wanted to type "ignore" or "cancel".  Or "i" would have been ok.
There's even tab completion!  (You can peak at the dangerous 1.5.1-pre1
which contains an even prettier interface, IMHO)

Could you try doing Ignore, and see how Parted copes with it?

> Here's the output from fdisk:
> 
> Command (m for help): p
> 
> Disk /dev/sda (Sun disk label): 27 heads, 133 sectors, 4924 cylinders

So, it's silently taking the geometry from the disk...

> Units = cylinders of 3591 * 512 bytes
> 
>    Device Flag    Start       End    Blocks   Id  System
> /dev/sda1             0      1426   2560316+  83  Linux native
> /dev/sda2          1426      2852   2560383   83  Linux native
> /dev/sda3             0      4924   8841042    5  Whole disk
> /dev/sda4          2852      3423   1025230+  83  Linux native
> /dev/sda5          3423      3716    526081+  82  Linux swap
> /dev/sda6          3716      3774    104139   83  Linux native
> /dev/sda7          3774      3832    104139   83  Linux native
> 
> Command (m for help):
> 
> I'm a little bit puzzled about the "CHS geometry"... This is a SCSI disk
> on a UltraSparc, is there such a thing as CHS on such a plaform?

Apparantly... I think the SCSI standard defines it.  So, there are 2 CHS's
here: the one on the disk label, and the one on the disk.  (Parted and
fdisk get this via an ioctl() directly from Linux).

I don't know where Linux gets the numbers from... it might be directly
from the disk, or from BIOS/firmware.  On x86, it's from the BIOS.
 
> Interestingly enough, there are a number of other disks (exactely the
> same ones) and they look like this:
> 
> Disk /dev/sdb: 64 heads, 32 sectors, 8637 cylinders
> Units = cylinders of 2048 * 512 bytes
> 
> Disk /dev/sdb doesn't contain a valid partition table

Interesting.

> Note that the invalid partition table is because this disk is part of
> a LVM.  This disk however does show the same values as the CHS message
> from parted.
> 
> I think I know what is causing this.  The machine was originally used
> with Solaris, not Linux.  The partition/disk info is left-over from the
> Solaris configuration.

I doubt this is the case.  fdisk obviously isn't detecting the Sun disk
label, so it wouldn't even try to analyse information on it.

> I'm ran into similar wierd problems with LVM, and that is probably why
> I had problems with my VG's instead of ext2resize as I suspected first.

Ah.

> A note somewhere that says "when using Linux with a Sparc platform, first
> do "dd if=/dev/zero of=<your disks>" will probably save a lot of trouble.

If you use Parted, there is no need :-)  Parted removes old signatures.
(Well, we don't support LVM directly, so this doesn't apply yet... but
it will eventually, I hope...)

Andrew Clausen