Page 1 of 1

Clustering for load balancing and fail over

Posted: Mon Oct 17, 2016 3:16 pm
by nbe
There are two reasons clustering would be desired: fail over and load balancing.

fail over:
JBoss provides different configurations of the cluster depending on what is desired.
In case of node failure, is flawless switching to the next node with transfer of not saved data expected?

load balancing:
This setup only makes sense if there are many messages in the JMS and the machine is busy with processing (other node will help with processing).

In either case, some items are needed:
-LSPS3.1.1014 or later
-jBoss app serverEAP 6.4 or 7.0
-Terracotta server to share ehcache between nodes. (Open source servers can be used if desired)
-replicate sessions, timers, jms messages
-Activate sticky-session="true"

Re: Clustering for load balancing and fail over

Posted: Mon Oct 31, 2016 3:13 am
by Byron Glueck
:idea: