Re: [linux-lvm] Installation fails if VG is exported

[Misc Things <formisc@gmail.com>]

On Tue, Jun 7, 2011 at 11:38 AM, Stuart D. Gathman <stuart@bmsi.com> wrote:
> On Tue, 7 Jun 2011, Stuart D. Gathman wrote:
> ...
> I get it - you wanted to prevent your OS experimentation from activating and
> possibly writing over your VG. �I can understand why vgexport seemed like
> a solution, but it only prevents activation and removes references from
> local LVM configuration (and prevents booting when that depends on the VG as
> you discovered).
>
> It doesn't prevent a rogue/buggy OS from scribbling on the disks. �So,
> physically disconnecting (or powering off) the disks is the best way
> to prevent that.

Right you are. i think i should reconsider my point of view - indeed
if i'm that paranoid i should eliminate any chances of losing data -
disconnect the dirve(s) is the best solution for this.

And then, i think, that the question is more relevant to anaconda
mailing group.support then to LVM group. Since it's anaconda that
tries to activate the VG.
Thank you Stuart!
Andrew