Java Reference
In-Depth Information
Deploying clustered applications
If you have tried starting your standalone or domain set of cluster nodes, you will be sur-
prised that there is no information at all about clustering in your server logging. Believe
me, it is not a bug but a feature! One of the key features of WildFly is that only a minimal
set of services is started; therefore, in order to see a cluster's live demonstration, you need
to deploy a cluster-aware application. In order to trigger clustering libraries in your applica-
tion, you can follow two approaches:
• If your application uses Enterprise JavaBeans, you don't have to do anything more.
This area brings some important changes in WildFly. Now, by default, the data of
all stateful session beans is replicated in HA profiles, and all stateless beans are
clustered. If your application is deployed on a container started with the
standalone-ha.xml
configuration, all remote
Stateless Session Bean
(
SLSB
) support failover capabilities by default.
• If your application includes a web application archive, you can use the portable
<distributable />
element in your
web.xml
file.
Let's have a look at both the approaches, starting from clustering EJBs.
