Reference:
This a complete ‘copy’ of the very useful info page at this link:
http://www.pc-professionals.com/blog/cloning-a-linux-heartbeat-3-node/
Some heartbeat problems when starting:
While setting up a Linux Heartbeat failover cluster, you may be tempted to set up one server node and clone the machine to create the other nodes that will make up your cluster.
This is a great approach, however, there is one little caveat you should be aware of…
Heartbeat generates a unique UUID for the node the first time it starts up.
If you had already started Heartbeat on the “template” node then you will need to force Heartbeat to generate a new UUID on the cloned nodes.
To force Heartbeat to generate a new UUID, simply stop Heartbeat, then find and delete this file:
/var/lib/heartbeat/hb_uuid
A new UUID will now be generated the next time you start the Heartbeat service on the node.
If you do not generate a new UUID for the cloned nodes you will experience strange issues, such as spontaneous rebooting of your cloned nodes,
and messages in your Heartbeat logs similar to the following:
Feb 13 11:06:20 lb1 heartbeat: [2611]: WARN: nodename lb1 uuid changed to lb2
Feb 13 11:06:20 lb1 heartbeat: [2611]: debug: displaying uuid table
Feb 13 11:06:20 lb1 heartbeat: [2611]: debug: uuid=xxxxx-xxxx-xxxx-xxxx-xxxxxxxxx, name=lb1
Feb 13 11:06:20 lb1 heartbeat: [2611]: ERROR: should_drop_message: attempted replay attack [lb2]? [gen = 1275513319, curgen = 1275513320]
Happy Heartbeating!