[Linux-cluster] LVM Cluster errors on bootup -- ok?

[Ray Van Dolson]

On Wed, Nov 25, 2009 at 09:30:17AM -0800, Ray Van Dolson wrote:
> Noticed that on bootup of a cluster node usings cLVM and GFS2 I see the
> following:
> 
>   Scanning logical volumes
>     connect() failed on local socket: Connection refused
>     WARNING: Falling back to local file-based locking.
>     Volume Groups with the clustered attribute will be inaccessible.
>     Reading all physical volumes.  This may take a while...
>     Found volume group "VolGroup00" using metadata type lvm2
>   Activating logical volumes
>     connect() failed on local socket: Connection refused
>     WARNING: Falling back to local file-based locking.
>     Volume Groups with the clustered attribute will be inaccessible.
>     2 logical volume(s) in volume group "VolGroup00" now active
>   
> VolGroup00 is a vg on the local disk and is not shared.
> 
> Later, cluster services start (including clvmd) and the clustered
> volumes and associated filesystems come up fine.
> 
> I'm assuming the errors above are because clvmd isn't running.  Can
> they be safely ignored?  Any way to configure so my clustered volumes
> aren't scanned until after clvmd starts?  I know I can edit filter
> settings in lvm.conf, but don't see any way to specify that some block
> devices should be skipped until clvmd is running.
> 

Hmm, thinking out loud here.  I am using multipath, and I'm seriously
doubting multipath is running when the errors above are spit out.  My
filter is as follows for lvm.conf:

  filter = [ "a|/dev/mapper/.*|", "a|/dev/hd[a-z].*|", "r|/dev/sd[a-z].*|" ]

Which should be filtering out my FC attached devices in lieu of stuff
under /dev/mapper.

So I guess maybe the fact that I have locking_type 3 set, my local hard
drive (non-clustered) is getting scanned and clustered locking is
attempted with it which fails.

So in the end it's probably OK that it falls back to local file locking
and I should just ignore these "errors".

Ray