Cannot contact other cluster node

  • 7006864
  • 17-Sep-2010
  • 27-Apr-2012


SUSE Linux Enterprise Server 11
SUSE Linux Enterprise Server 11 Service Pack 1


After setting up openais/corosync software the cluster nodes do not see each other despite the network working and firewall rules where checked for interference.

Changing in the interface setting of the cluster communication ring the setting from

         interface {
                mcastport:      6467
                ringnumber:     0      


(the values for bindnetaddr mcastaddr mcastport are only examples here)


        interface {
                broadcast: yes
                mcastport:      6467
                ringnumber:     0      


the cluster nodes see each other after a restart of the cluster software.


The reason for this behaviour is in the switch connecting the nodes. This switch is doing igmp snooping. It expects any multicast traffic that does not run on 224/24 to provide LEAVE JOIN and other information.
And as can be seen in the above example the multicast address used in the interface section is not one of 224/24. The address ranges can be found documented at

and the Local Network Control Block ( -  (224.0.0/24)) is the range that is treated like broadcast and not filtered by the switch.

To address this issue otherwise there is a few options:

One would be to disable igmp snooping on all hardware relevant to the cluster network, considering that switches should not do package filtering, which might not be applicable in the specific scenario

The last option is to use the broadcast: yes setting in the interface section

Special thanks at this point to Mr Markus Hoeschen for his invaluable help to identify this issue

Feedback service temporarily unavailable. For content questions or problems, please contact Support.