[linux-lvm] pvscan connect() failed on local socket: Connection refused

[linux-lvm] pvscan connect() failed on local socket: Connection refused

[Madison Kelly]

Hi all,

   I've got a two-node Xen cluster I am trying to setup with a DRBD 
partition under an LVM pv. My layout is:

* Both node's hard drive is a PV used by one VG.
* Each node has the host OS  (dom0) install on an LV using a small 
portion of the available space.
* The majority of the space is assigned to an LV on either node that 
DRBD uses.
* The DRBD partition is itself setup as a new PV.

   The problem is; Once I setup '/dev/drbd0' as a VG, I had to enable 
cluster-aware mode by changing lvm.conf to:

filter = [ "a|drbd.*|", "r|.*|" ]
locking_type = 3

   And then running:

lvmconf --enable-cluster

   After that, when I run 'pvscan', I get this output on both nodes:

[root@an_san02 ~]# pvscan
   connect() failed on local socket: Connection refused
   WARNING: Falling back to local file-based locking.
   Volume Groups with the clustered attribute will be inaccessible.
   PV /dev/drbd0                      lvm2 [399.99 GB]
   Total: 1 [399.99 GB] / in use: 0 [0   ] / in no VG: 1 [399.99 GB]

   How can I tell what or where the connection is failing? It *should* 
be going over 'eth1' on both nodes which is outside of Xen's control. 
I've also checked iptables and I don't see any relevant rules (I've 
flushed the tables and there is only now Xen-related bridging rules). 
Also, I can ping the other node over the dedicated DRBD interfaces.

   Thanks for any insight!

Madi

Re: [linux-lvm] pvscan connect() failed on local socket: Connection refused

[Luca Berra]

On Fri, Feb 05, 2010 at 08:42:05PM -0500, Madison Kelly wrote:
> Hi all,
>
>   I've got a two-node Xen cluster I am trying to setup with a DRBD 
> partition under an LVM pv. My layout is:
>
> * Both node's hard drive is a PV used by one VG.
> * Each node has the host OS  (dom0) install on an LV using a small portion 
> of the available space.
> * The majority of the space is assigned to an LV on either node that DRBD 
> uses.
> * The DRBD partition is itself setup as a new PV.
>
>   The problem is; Once I setup '/dev/drbd0' as a VG, I had to enable 
> cluster-aware mode by changing lvm.conf to:
>
I'm not sure what you are trying to achieve there.
Active-Active drbd as shared storage???

> [root@an_san02 ~]# pvscan
>   connect() failed on local socket: Connection refused
>   WARNING: Falling back to local file-based locking.
>   Volume Groups with the clustered attribute will be inaccessible.
>   PV /dev/drbd0                      lvm2 [399.99 GB]
>   Total: 1 [399.99 GB] / in use: 0 [0   ] / in no VG: 1 [399.99 GB]
sounds like the cluster daemon is not running

L.
-- 
Luca Berra -- bluca@comedia.it
         Communication Media & Services S.r.l.
  /"\
  \ /     ASCII RIBBON CAMPAIGN
   X        AGAINST HTML MAIL
  / \