About Polling and Pruning Roles

One requirement of the Advanced Clustering plugin is that all participating servers in the cluster be connected to a single, shared database. Because of this requirement, the plugin needs to enforce that only one server in the cluster is assigned to execute the data pruner on its defined schedule.

Likewise, a similar problem exists for channels with polling source connectors (File Reader, Database Reader, JavaScript Reader). Typically, a polling source connector polls a shared file system or an external database, in which case, not all servers in the cluster should perform this operation. If they did, you may end up with the same HL7 message file being processed once for each server in the cluster. To avoid this problem, the Advanced Clustering plugin designates one server in the cluster to perform these polling operations.

To designate a particular server as the preferred "polling" or "pruning" server, select the appropriate server by selecting its radio button in the Node Configuration list on the Clustering page. If the preferred polling or pruning server goes offline, another server will automatically assume the role until the preferred server is back online.