[ejabberd] Preferred MUC clustering method

Eric Rivas kc2hmv at jla.rutgers.edu
Thu Mar 19 19:48:28 MSK 2009


I've setup a two node ejabberd cluster with clustered mod_muc.  I would 
like to set it up so if any node fails, the other can resume with limited 
degradation (complete with mucs).

Right now, I have it so that most mnesia table is duplicated on each node 
(I basically did an add_table_copy with the same table type on the second 
node).

I had an incident where both node were brought down. I brought up the 
second node and it appears that the persistent mucs were not showing up. 
Does anyone know of a way to remedy this?

The cluster is using ejabberd-2.0.4.

-- 
Eric Rivas
kc2hmv at jla.rutgers.edu



More information about the ejabberd mailing list