Daniel Carosone <dan@geek.com.au> wrote:

> > I need to think about this a bit more... the idea with "raidctl -A 
> > root" was that there would only ever be *one* RAID set configured 
> > that way, per machine. 
> The typical situation where you wind up here is when plugging the raid
> set of a failed machine into a working machine, to get access to the
> data.  So as much as its an odd situation and warrants at least a
> warning and probably user interaction, we need to handle it well.

That's exactly why I did that change: when you move a set of disk from
one machine to another, it can be an emergency, and you really don't
want to have to refusing to boot because it does not find the root
anymore. Especially if you have hardwired it in the kernel config file.

-- 
Emmanuel Dreyfus
http://hcpnet.free.fr/pubz
manu@netbsd.org