after many desperate tries, I have installed Openfire on two Linux machines and (with the new clusering plugin 1.2.1_beta) it seems to work. It is just a test installation. We want to install the productive system on a solaris cluster.
We use a solaris cluster with two nodes. Only one node is active at the time. At the moment we use two openfire servers, one on each node. Both servers use its own embedded database. That means that we have to make changes to the configuration on both servers. Now we want to start using the clustering plugin. We use many processes that are connected to the openfire server only on the current node(via cluster alias). Via XMPP we can send them commands (Stop, Restart, etc). After a fail-over we have to restart them all, because they have to reconnect to the new current openfire server otherwise they are not visible anymore.
We want to avoid this in the future. We want that all processes keep the connection to the server after a fail-over. Does the openfire clustering plugin offer this functionality or is it just a kind of load balancer? We need a permanent connection to the openfire server come what may.
Does anybody have some experinces with it?
Thanking you in anticipation