Chapter 3 Cache Object Management Tutorial : Configure Cache Cluster Properties

Configure Cache Cluster Properties
BusinessEvents ships with default values set for basic caching functionality, so all you have to do to enable caching in a project is give the cluster a name that is used in all nodes, and enable the multi-engine feature property.
Behind the simple setup is a set of configuration properties with default values. The Learning Points section below provides information on some configuration options that you don’t have to set for the tutorial.
Learning Points
Node discovery  The following default settings are used for multicast node discovery. Properties do not have to be added to the property files unless you use different values.
java.property.tangosol.coherence.clusteraddress=224.3.3.1
java.property.tangosol.coherence.clusterport=31337
To avoid possible collisions with other clusters running on your network, the tutorial uses a non-default cluster address. More cluster discovery settings are available for different situations, such as hosts with multiple NICs. Another method of discovering nodes, using well-known addresses for situations where use of multicast is not an option.
Cache cluster features  Many characteristics of the cache cluster and caching scheme are specified by properties set in the engine properties file, such as the cluster name. More in depth customization is possible but is not often required and is not documented. Contact TIBCO support for assistance.
Multi-engine features  Multiple engines can run concurrently against the same project ontology. You can use a mixture of differently configured agents and identically configured agent groups (see Configure Inference Agents). By default the multi-engine features are disabled, so that only the agents in one EAR file (that is, running in one JVM) are active at one time. To turn on multi-engine features, which offer load balancing (and implicit fault tolerance), you set the be.engine.cluster.multiEngineOn property to true.
Number of cache servers to start  You must also set a property that defines the number of cache servers to start before starting any inference agents. The be.engine.cluster.minCacheServers property ensures that the cache is populated before inference agents begin processing events.
More Information
For details on other cluster-level properties not discussed in this tutorial (because their default values don’t need to be changed) see Chapter 19, Configuring Cache Cluster Discovery and Chapter 20, Configuring Cache Cluster Settings of TIBCO BusinessEvents User’s Guide.
Task G Configure Cluster Properties
1.
   BE_HOME/bin/be-engine.tra
2.

 
#### Cluster Level Properties
 
java.property.tangosol.coherence.cluster=AcmeCluster
java.property.tangosol.coherence.clusteraddress=224.3.3.11
java.property.tangosol.coherence.ttl=0
be.engine.cluster.multiEngineOn=true
be.engine.cluster.minCacheServers=2

 
The properties are briefly described in the table below.
3.
Save the file in the BE_HOME/bin directory as agent1.tra.
You will copy and modify this file for each of the five engines you will start when testing the behavior of the deployed project.
To avoid collisions with any other project deployed on your network, use a non-default address. The tutorial uses 224.3.3.11
Specifies the time-to-live setting for the multicast, that is, the maximum number of "hops" a packet can traverse. A hop is defined as a traversal from one network segment to another via a router.
If you will start all nodes on one machine (which you can do for non-production situations), set the ttl to 0, to keep multicast packets from leaving the originating machine (on most operating systems).
Specifies the minimum number of storage-enabled nodes that must be active in the cluster when the system starts up before the following occur: