On Mon, 2009-03-30 at 10:11 -0500, Eric Sandeen wrote:
Bob Gustafson wrote:
> On Mon, 2009-03-30 at 09:12 -0500, Eric Sandeen wrote:
>> Bob Gustafson wrote:
>>> On Mon, 2009-03-30 at 08:55 -0500, Eric Sandeen wrote:
>>>> FWIW there is supposed to be an ext3->ext4 migration via anaconda if
>>>> boot with "ext4migrate" on the cmdline.
>>> Hmm, by definition an 'update' - and failure is not an option -
>>> back to FC10.
>>> Testing of this option on Anaconda would include having a FC10 whole
>>> disk image - to re-image disks for the next test.
>> Yep. I guess I missed the FC10 part.
> Also, for testing RAID and ext4migrate, I think the testers will need a
> disk image for both RAID disks (assuming RAID 1). The data on both disks
> should be the same, but I think there are differences in the metadata
> between each disk in a pair. Is this true?
The fs metadata on each leg of the mirror should be identical.
But depending on the RAID used the RAID metadata may not be. If you're
preserving images of members for re-tests this will matter (though I'd
have thought it would be easier to keep an image of the fs and restore
that to the array each time you need to re-test).