DevPortal

What does a Cluster Definition Involve?


A cluster definition involves:

The initial definition of a cluster is made in the XML of the first node of the cluster. Once the basic cluster attributes are defined in the cluster-config.xml, the cluster type is defined by the cluster-node definition in the cluster-config.xml as shown in the following code snippet:

<node>
  <name>node</name>
  <computer-name>127.0.0.1</computer-name>
  <naming-port>9191</naming-port>
  <http-port>8181</http-port>
  <node-type>BOTH</node-type>
  <node-id>001</node-id>
</node>

Here, when node-type is J2EE, enabling the cluster service would make it a J2EE Cluster. A J2EE Cluster can be of type BOTH (both EJB and Web), or just EJB, or Web.

J2EE Cluster cannot start with Message Server in an embedded mode, and a Message Server cluster cannot have EJB or Web Containers enabled. If node-type is JMS, then it becomes an HA Message Server cluster. J2EE clustering can involve both EJB and Web clustering. However, Web and Message Server instances cannot exist in a single cluster. The cluster solution is based on a common cluster backbone called Cluster service, which also runs on the Pramati Services framework.

Once a cluster is enabled in server-config.xml, define the cluster topology and properties in the cluster-config.xml. Specify cluster nodes, topology, persistence (EJB/Web), and Load Balancer here.

Pramati Cluster being a self-organizing, non master-slave homogeneous cluster, does not have a separate entity for cluster housekeeping or processing. Based on the shared persistence store, the cluster runtime in each cluster node jointly performs all the housekeeping and operational tasks.


Related Topics:

© Pramati Technologies 2007 Runs on Pramati Server | Feedback | Legal