gconcat and gpt

[Rong-en Fan]

On 1/4/07, Eric Anderson <anderson at centtech.com> wrote:
> On 01/03/07 00:00, Rong-en Fan wrote:
> > Hi,
> >
> > I'm running 6.2-RC1 on an i386 box. I'm testing gjournal patch.
> > I found that something is annoying... I use
> >
> > gconcat label big da1 da2 da3
> >
> > and then running gpt on the concat/big device. So, I got
> > concat/bigp1, concat/bigp2, ... etc *AND* I also have
> > da1p1, da1p2... Well, this is not a problem for normal use. However,
> > when I enable gjournal, at boot time, gconcat first activates the
> > big one then, gjournal finds there is journal data on da1p1, so
> > it uses! Then, da1 is deactivated from gconcat...
> >
> > My question is, why I got da1p1 when da1 is used by gconcat?
>
> Did you use gjournal against the da* devices, or the gconcat device?  If
> you ever used it on the da* devices, the meta data might still be there
> confusing things.

I use gjournal against gpt'ed gconcat device.