[linux-lvm] devices.filter changed behaviour in 80ac8f37d6

[Chris Webb]

Peter Rajnoha <prajnoha at redhat.com> wrote:

> Yes, without lvmetad daemon, each LVM command needs to run full device
> scan at its start to look for PVs and then when a PV is found, it needs
> to read all LVM metadata it finds (and check for consistency etc etc).
> Of course, not including those devices which are filtered out.
>
> However, with lvmetad, you can save some disk access here since all the
> metadata are taken from lvmetad cache...
>
> Lvmetad doesn't have any extra deps except what lvm binary already has.

Sounds like a clear performance win for us. I'll have a play.

> If you're not using udev to handle kernel uevents (hmm, what are you using
> then?)

Nothing particularly clever or interesting I'm afraid: just a very trivial  
daemon which translates events from a NETLINK_KOBJECT_UEVENT socket into a  
continuous blocking stream.

   https://github.com/arachsys/init/blob/master/uevent.c

This is then consumed by a simple read loop to implement the handful of  
rules we actually want.

It'd be easy to add pvscan --cache when the relevant devices appear or  
disappear. Thanks for the pointer.

(I note from the pvscan(8) man page that global_filter applies to pvscan  
--cache but not standard filter, which sounds like another good reason to  
switch to global_filter!)

Cheers,

Chris.