Re: storage rewrite bugs

[Bruno Wolff III <bruno wolff to>]

On Mon, Nov 24, 2008 at 15:27:55 -0500,
  Jeremy Katz <katzj redhat com> wrote:
> On Mon, 2008-11-24 at 14:19 -0600, Bruno Wolff III wrote:
> > The advantage of doing this is that an element of a raid array has slightly
> > less space available than it would otherwise. So that changing to a raid
> > array later is not easy to do. The alternative is to manually set up the
> > device yourself, but it is easy to make mistakes do that as you don't get
> > the defaults from anaconda any more.
> 
> ... and while I'm not really a fan of making it so that you can easily
> set up non-complete arrays as might have been noticed over the years,
> I'm *not* against actually leaving the 64k or so at the end of the
> device so that you could turn it into a RAID array later.

That might be an ever better solution. For raid 1 you could just make it
a raid array and then add the other disks. For other raid levels you should
be able to set up a degraded array and dd the block device over. Afterwards
the space could be used to complete the array. I think as long as fstab
mounted things by label switching wouldn't be too hard. Probably just
booting in rescue mode and running mkinitrd?