[ejabberd] Clustering for redundancy

Pico Geyer pico at trispen.com
Wed Dec 19 11:06:38 MSK 2007

Hi all.

I'm currently trying to test an ejabberd cluster consisting of two nodes.
What I'm hoping to achieve is, when either node goes down the other will 
be able to continue serving clients.
I'm simulating one of the nodes going down by temporarily firewalling it 
I then receive the following messages on the first node's console:   

 ** Node ejabberd at jabber2 not responding **
 ** Removing (timedout) connection **

and a corresponding message on the second nodes console.
If I now unfirewall the firewalled node, it seems that the nodes never 
reestablish the connection between each other.
Is there a way to get erjabberd/erlang to retry the connection between 
nodes every X minutes.


More information about the ejabberd mailing list