That's the two places I always change whenever I use dd to make an
exact image onto another drive for backup and need to use the backup.

On Mon, 12 May 2008 15:29:13 -0700, moe wrote:

>opensuse 10.3
>
>Just moved everything to a new hard drive. First I setup swap on the new
>hd, setup partition(s), and copied all files from old to new. Then I
>updated the /etc/fstab, replacing those gnarly disk ids with simple
>device names (like /dev/sda2 for example). Then I used supergrubdisk to
>install grub and make the new hd bootable. Then I had some problems..
>she wouldn't boot. But I figured it out, and also replaced the disk ids
>in /boot/grub/menu.lst with simple device names. Now it boots and
>everything is working good so far.
>
>Just wondering if the old disk ids might be hanging around anywhere else
>and what the ramifications might be? AM I good?

On 2008-05-13 01:11, Greg wrote:
> That's the two places I always change whenever I use dd to make an
> exact image onto another drive for backup and need to use the backup.
>

Since you top post, I cut what you answer.

dd is also saving free blocks, and will create backup images much bigger then
needed.

One good util is partimage, that skip unused blocks.

If one use dd, the output file can be compressed nice if one first zero all
free blocks.

One way to zero free blocks is:

dd if=/dev/zero of=bigfile || rm bigfile

eg. remove bigfile if the command fail, and the command will fail
when the disk if full.

After that, all free blocks has only zero's as data and can be compressed to
near nothing.

But what moe did is a better way, and he got a clean fresh filesystem, the only
thing to check is that filesystems don't mount by the unique ID.

/bb

On Tue, 13 May 2008 13:59:55 +0200, birre
wrote:

>On 2008-05-13 01:11, Greg wrote:
>> That's the two places I always change whenever I use dd to make an
>> exact image onto another drive for backup and need to use the backup.
>>
>
>Since you top post, I cut what you answer.
>
>dd is also saving free blocks, and will create backup images much bigger then
>needed.
>
>One good util is partimage, that skip unused blocks.
>
>If one use dd, the output file can be compressed nice if one first zero all
>free blocks.
>
>One way to zero free blocks is:
>
>dd if=/dev/zero of=bigfile || rm bigfile
>
>eg. remove bigfile if the command fail, and the command will fail
>when the disk if full.
>
>After that, all free blocks has only zero's as data and can be compressed to
>near nothing.
>
>But what moe did is a better way, and he got a clean fresh filesystem, the only
>thing to check is that filesystems don't mount by the unique ID.
>
>/bb

Thanks for the information. Sorry about the top-post - I'm more used
to email. :-)

Greg