With virtualization so prevalant these days it's so easy get one server all setup the way you like and then clone it on off for the next server. The linux heartbeat daemon is still very prevalent in use to keep various services up and running during maintenance, hardware failure, etc. However when you clone one box that has heartbeat up and running to create that next node that will be joined into the cluster, there is a small issue that becomes problematic. Of course after adding the new node you have to do your necessary configs to
/etc/ha.d/ha.cf to get the new box on, but if you look at the logs after adding it, you may see some of these flying by and filling up your ha logs.
Jun 13 10:35:24 server1 heartbeat: : WARN: nodename server1 uuid changed to server2
Jun 13 10:35:24 server1 heartbeat: : debug: displaying uuid table
Jun 13 10:35:24 server1 heartbeat: : debug: uuid=6cefa610-ef8c-4f1f-b9fe-55de41b1168f, name=server1
Jun 13 10:35:24 server1 heartbeat: : ERROR: should_drop_message: attempted replay attack [server2]? [gen = 1450123238, curgen = 1450123239]
This has to do with the fact that you cloned server1 to make server2, changed and updated some configs and then fired everything back up. At the time of heartbeat starting for the first time it generates a uuid on the server to identify itself in the cluster. Since you just spun up server2 from server1, this uuid has already been generated. No fear though. Easy fix
#> service heartbeat stop
#> rm /var/lib/heartbeat/hb_uuid
#> service heartbeat start
You can do this on one or both of the servers. It doesn't hurt to re-do it.
Clustering, Linux, Networking