I see no reason why disabling batching should prevent a session from replicating.
However, you should use ATTRIBUTE granularity replication - as opposed to SESSION granularity replication (the default).
The former stores each attribute in a separate Infinispan cache entry, while the latter stores all attributes as a map in a single cache entry.
To use ATTRIBUTE granularity, specify the following in your jboss-web.xml deployment descriptor:
<replication-config> <replication-granularity>ATTRIBUTE</replication-granularity> </replication-config>
I am planning to implement session replication on jboss EAP7 using full-ha profile (I have F5 in front of appServer (NO webServer)), by reading information found online it seems changes are needed both at jboss Configuration and in java code level. What type of configuration changes i am looking at to implement session-replication .?
Thanks for the `replication-granularity` tip, pferraro!
Turns out the problem was in the application. A bad WebFilter was invalidating the session incorrectly.
I wrote a very simple application like a Counter Incremator session and works fine. So a figure out the problem was in my application. Sorry
vamshi1413, the session replication is already pre-configured in standalone-full-ha.xml or domain.xml (ha or full-ha profiles).
You just need to configure the proper `jgroups` stack and socket bindings. And you need the <distributable/> tag in your web.xml.
Thanks, can you please specify what are those changes ? Do I require to update anything in infinispan subsystem ?
If you are using jgroups UDP stack you don't need no other configuration then include the <distributable/> tag in your web.xml