[Pacemaker] Active/Active issue

David david at adurotec.com
Thu Jun 10 15:47:07 UTC 2010


Got my first install of pacemaker/corosync running for a 2 node apache 
cluster in active/active mode yesterday on CentOS 5.5 x86_64.

Everything was working just fine until I tested disaster recovery by 
failing both servers and restarting them.

When the servers came back I noticed that even though both nodes are 
online, the ClusterIP:0 and ClusterIP:1 are running on the same node and 
only one instance of apache is running on one of the nodes.

I am assuming that apache is only running on one node because there is 
no valid IP running on the other node since both are running on one 
node.  Why would both clone IPs be running on the same node?  Is there a 
way to manually move one of these instances to the node that has no IP?

Thanks,
David




More information about the Pacemaker mailing list