[lvm-devel] Filtering out LVs?

[Alasdair G Kergon]

On Tue, Mar 29, 2011 at 08:08:23PM +0900, Jun'ichi Nomura wrote:
> So there is a race that one of multiple vgchange instances prevents other;
> "vgchange -an vg0" might fail to deactivate the VG if "vgchange -an vg1"
> opens vg0/lv0 at the very time for scanning.
 
I'm inclined not to do anything about this at the moment, but to wait until
udev+new cache work is complete then revisit it if the problem hasn't
gone away.

Alasdair