Cluster Roles

The Auto Deploy cluster role allows for the automatic deployment of portlets that are copied to the Software AG_directory \MWS\server\serverName\deploy directory (of a node that has the Auto Deploy role). A cluster can have multiple Auto Deploy nodes. This role is enabled by default, but is only needed if you want the automatic deployment capability. It is often desirable to have this role disabled in a production environment to reduce the possibility of unauthorized modification of the server. This role does not affect cluster high availability.

The Notification cluster role is responsible for formatting and sending email notifications, and notifications that are sent to a user. A cluster can have multiple Notification nodes. The Notification role is enabled by default, but is needed only if you use notifications in your production environment. For example, if you are running Task Engine, users will not receive task notifications if the Notification role is not enabled. For high availability it is recommended that you have more than one cluster node in this role. If all Notifications nodes go down, notification request messages continue to accumulate in the queue. No notifications are generated or delivered until one or more Notification nodes comes back online and starts processing them.

The Search role is responsible for indexing all content that is exposed to the embedded search engine, maintaining the search index, and performing the searches. The Search role is enabled by default, and should be enabled on all cluster nodes where the search functionality will be needed. The Search role is important to server operation and there are core features that will not operate without it.

Each cluster node maintains a local copy of the search index, improving the performance and reliability of searches. If the search index becomes corrupted on one cluster node, you can remove that node from the load balancer while the search index is rebuilt on that node. In the meantime, the other nodes can continue servicing requests.

The Task engine cluster role is responsible for all Task Engine activities, such as queuing tasks, processing task rules, searching and retrieving task data. A cluster can have multiple Task Engine roles. This role is enabled by default, but is needed only if you actually have Task Engine running on your cluster. For high availability it is recommended that all nodes have this role enabled unless a specific node is not included in the load balancer configuration and never services end-user or Process Engine requests.