Role of the Cluster Coordinator

The Cluster Coordinator provides configuration and state management and distributed coordination services to members of the Endeca Server cluster. It ensures high availability of the query processing provided by the nodes in the Endeca Server cluster and in the data domain clusters.

Any node in the Endeca Server cluster can optionally be configured to run the Cluster Coordinator. Since the availability of a clustered Endeca Server deployment depends on the running of the Cluster Coordinator services, it is highly recommended that an odd number (at least three) Endeca Server nodes are configured as Cluster Coordinators to avoid the Cluster Coordinator being a single point of failure.

The Cluster Coordinator service has the following characteristics:

It is a shared information repository that provides a set of distributed coordination services. It ensures that all systems in the cluster coordinate their actions relative to all other systems running in your environment. If one of the nodes communicates any cluster information, all other nodes recognize it and react to it in a manner that ensures synchronization, event notification, and coordination between the nodes.
The communication and coordination mechanisms continue to work in the case when connections or data domain cluster nodes fail.
It provides communication between the Endeca Server nodes, ensuring that if one of these nodes fails, requests are sent to other active nodes, until the node rejoins the Endeca Server cluster.
It provides communication between Dgraphs in a data domain cluster. It prevents node isolation, ensures that updates from a cluster client are applied in the order that they were sent, guarantees that updates are either applied in full or not at all, and ensures that all data domain nodes have access to the most recent version of the index after updates. It also controls the election of the Dgraph leader node in the data domain cluster, in case the current leader node fails. The newly-elected leader node identifies the most recent version of the index, and using the Cluster Coordinator, informs other nodes of the current version.

To summarize, in order to run, the Cluster Coordinator service requires a majority of its nodes to be active. Therefore, it is recommended that the Cluster Coordinator instance runs on an odd number (at least three) of the Endeca Server nodes in the deployed Endeca Server cluster. For information on deploying Cluster Coordinator services on a subset of Endeca Server nodes, see the Oracle Endeca Server Installation Guide.