To set up the project for cache object management, you create a new CDD file using the template provided for cache OM. The CDD template defines an inference agent class and a cache agent class, and default values for cluster level properties, so it doesn’t take much to configure an example project to work. Other configuration options are available for different situations.
How do I cluster nodes discover each other? For the tutorial you can use the default multicast node discovery settings. You can also change the defaults as needed to avoid collisions with another cluster in the same network. Cluster discovery settings are available for different situations, such as hosts with multiple NICs. Another method of cluster discovery uses well-known addresses for situations where use of multicast is not an option.
How do I configure agents and processing units? A processing unit deploys as a TIBCO BusinessEvents engine. At a minimum, you list the agents you want to run in one processing unit. For this example, you don’t have to do any processing unit configuration because defaults will work out of the box. However, for real-world scenarios, where concurrency features are used, and perhaps add-on products, processing unit configuration requires some thought and planning. The processing unit definitions are also used in the TIBCO BusinessEvents Monitoring and Management component, for deploy-time configuration.
Number of cache agents to start For a production environment you can define how many cache agents must be started before the system startup can continue. The default is one. When a backing store is used, cache agents can preload objects from a backing store at system startup. More cache agents make the preloading phase quicker. See the Cache Agent Quorum setting in the Object Management Configuration panel.
Limited Cache The Entity Cache Size and Object Table Cache Size settings in the Object Management tab take effect only if you use a limited cache. By default, the cache is not limited—In Domain Objects > Default, the setting Is Cache Limited is not checked. However when you add a backing store, it is common to use a limited cache, because you can store all objects in the backing store. You can override the limited cache setting at the object level, as desired.
You have configured the CDD for cache cluster discovery, and you configured the processing units and agents as you did before. Now you can start the project and see how things work at runtime.