Table 29 Inference Agent Engine Properties Specifies the maximum number of active instances of the agent. This value is used to limit the number of active instances in an agent group.Must be the same for all nodes where instances of this agent (that is, agent group members) are deployed. The priority indicates the order in which inactive agents become active, and conversely, the order in which active agents become inactive, when new agents join the cluster.The lower the number, the higher the agent is in the activation priority list. For example, an agent with priority 2 has a higher priority than an agent with a priority of 6.Set differently in each node’s TRA file to determine the order of each agent instance for startup, as well as failover and failback in fault tolerance situations. Specifies a value that uniquely identifies an instance of an agent within its agent group.Set differently in each node’s TRA file to identify each agent uniquely. (BusinessEvents also maintains an internal ID in addition to this key to uniquely identify agents.)Required for recovery of scorecards. Recommended in all cases, for situations that require an agent instance to be uniquely identified. Specifies the maximum number of objects in each agent’s L1Cache. The L1 cache is a local cache used by the inference agent for local access to recently used objects. It is used to optimize access to objects.The value can be 10,000 or above. Values less than 10,000 are interpreted as 10,000.Note that this setting does not relate to local storage. It is required for all inference agent nodes. java.property.tangosol.coherence.distributed.localstorage When you use a distributed cache scheme, you decide whether to use some JVM memory for local cache storage based on your needs.Set to true to enable the object management layer to use the BusinessEvents server’s JVM memory for storing cached objects.TIBCO recommends that you disable local storage on inference agents and query agents, and enable it only on cache server nodes. Specifies the number of threads used by the inference agent L1 cache to write to the cache cluster.The default value is the same number as there are processors available to the JVM.See also Configuring Caching Scheme, Multi-Engine, and Cluster Properties for a related thread property. Specifies the number of entries per page to be used while recovering objects from the cache.For example, if you set the value to 10,000, then the engine loads handles in blocks of 10,000, instead of trying to load them in a single batch. Smaller batch sizes result in slower recovery. Experiment with batch size to establish the best batch size to use for your environment.
Copyright © TIBCO Software Inc. All Rights Reserved.