The purpose of cache servers is to store and serve cache data for the cluster. A cache server is a non-reasoning agent used as a storage node only. Storage nodes are responsible for object management. They participate in distribution, partitioning and storage of the objects in the cluster.
Local storage The local storage engine property is true by default, so you only have to set it when you want to disable cache data storage for a node. It is a best practice in production systems to use dedicated cache server nodes for cache data storage and to disable cache data storage in all other nodes.
Deploying a node as a cache server Cache server nodes are deployed using any EAR file that contains a BAR resource configured for an inference agent (or a query agent). All other agent-level properties in the TRA file are ignored when the following property is set: be.engine.cacheServer=true.
Number of cache servers to start At the cluster level you used the
be.engine.cluster.minCacheServers property to define the number of cache servers that must be started before inference agents start. This means that you can start all the nodes in any order. The inference agents won’t become active until the specified number of cache servers have started, ensuring that the agents can write to and read from the cache.