- Reorganization of clustering services and conversion to JBoss Microcontainer-based configuration.
- Simplified cluster isolation
- Shared JGroups resources and the JGroups Channel Factory
- On-demand JBoss Cache via the CacheManager service
- Improved JPA/Hibernate Second Level Caching
- Support for passivation to disk of clustered web sessions
- Integration of JBoss Cache 3
- Changes in the Farm Service
- Configuration changes for clustered web applications
- Separate cache used for Clustered SSO, HA-JNDI and DistributedState
- Array interception for FIELD replication granularity clustered web applications
New transaction sticky load balance policies for EJB2 beans.
- Allow use of classes from scoped loaders in HAPartition RPC calls
JBoss AS 5.0.0.GA integrates JBoss Cache 3.0.1.GA as its distributed caching solution. JBC 3 has a significantly different API from the 1.x releases used in JBoss AS 4.x and 3.2.x.
The AS 5.0.0 beta and CR releases integrated JBC 2.x. For those who may have tested those releases and who use JBC directly, there may be some behavioral differences between the JBoss Cache 3.0.1.GA that was included in the AS's GA release and the JBC 2.x versions used earlier, although the JBC team made significant effort to eliminate them. We apologize for any inconvenience this late-in-the-release cycle may have cause. It was clear that with the first GA release from the JBC 3.x branch development on the 2.x branch would halt, so when JBC 3 became available earlier than expected the AS clustering team decided it would be better for both the AS and JBC communities to use a JBC release from an actively developed branch.
In AS 5.0.0 and 5.0.1 the AS 4 Farm Service was removed, with no replacement. Beginning with AS 5.1.0.CR1 an equivalent service based on the AS 5 Profile Service is available. For details, see the "'Farm' Deployments in AS 5.x" page.
The ClusteredSingleSignOn valve no longer shares a JBoss Cache instance with distributable web sessions. Now, like other AS 5 services that use JBoss Cache, it gets its cache from the CacheManager service. By default it will use the ha-partition cache. This can be changed by setting the cacheConfig attribute in the valve's configuration in $JBOSS_HOME/server/all/deploy/jbossweb.sar/server.xml:
<Valve className="org.jboss.web.tomcat.service.sso.ClusteredSingleSignOn" cacheConfig="custom-sso-cache"/>
The ha-partition cache the ClusteredSingleSignOn uses by default is also used by default for HA-JNDI cluster-wide naming tree and for the DistributedState service. The use of JBoss Cache for HA-JNDI and DistributedState is new in JBoss AS 5.