[Pacemaker] Multicast corosync packets and default route

[Daniel Dehennin]

Hello,

I'm trying to setup a pacemaker/corosync on Ubuntu Trusty to access a
SAN to use with OpenNebula[1]:

- pacemaker .1.10+git20130802-1ubuntu2.1
- corosync 2.3.3-1ubuntu1

I have a dedicated VLAN for cluster communications.

Each bare metal node have a dedicated interface eth0 on that VLAN, 3
other interfaces are used as a bond0 integrated to an Open vSwtich as
VLAN trunk.

One VM have two interfaces on this Open vSwitch:

- one for cluster communication
- one to provide services, with default route on it

My 3 bare metal nodes are OK, with pacemaker up running dlm/cLVM/GFS2,
but my VM is always isolated.

I setup a dedicated quorum (standby=on) VM with a single interface
plugged to the cluster communication VLAN and it works
(corosync/pacemaker).

I run ssmping to debug multicast communication and found that the VM can
only make unicast ping to the bare metal nodes.

I finish by adding a route for multicast:

    ip route add 224.0.0.0/4 dev eth1 src 192.168.1.111

But it does not work.

I only manage to have my VM as corosync member like others when default
the route is on the same interface as my multicast traffic.

I'm sure there is something I do not understand in corosync and
multicast communication, do you have any hints?

Regards.

Footnotes: 
[1]  http://opennebula.org/

-- 
Daniel Dehennin
Récupérer ma clef GPG: gpg --recv-keys 0xCC1E9E5B7A6FE2DF
Fingerprint: 3E69 014E 5C23 50E8 9ED6  2AAD CC1E 9E5B 7A6F E2DF
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 342 bytes
Desc: not available
URL: <http://lists.clusterlabs.org/pipermail/pacemaker/attachments/20141106/36b3a375/attachment-0002.sig>