Advisory Messages and Impact on Fault Tolerance Processing
The advisory heartbeat message is sent when a resource failure is identified. This message is sent to the database and the JMS at a regular interval of heartbeat beat interval configured for the cluster. The messages are sent to JMS and the database, and the resource that is available will be able to process the respective message.
The message contains information such as NodeId, Update Timestamp and a Message.
Database advisory messages are stored in DOMAINMEMBERS_ADVISORY table. A record is inserted in this table till there is resource failure. The DOMAINMEMBERS_ADVISORY table is a new table.
JMS advisory messages are sent to the topic tibco.aff.orchestrator.cluster.advisory.heartbeat, and this message is received by the corresponding listener if the JMS resource is available along with the updates the in memory data and the message content.
This information helps to identify the nodes that are able to process advisory message and subsequently perform fault tolerance activity within the nodes. When the nodes resume the normal state, then the cluster manager will perform certain fault tolerance related activities for the failed nodes. Before performing any fault tolerance activities, the cluster manager will check if there is not any advisory message available for the failed node that is older than the Fault Tolerance threshold interval. If advisory message is available for the particular failed node then the fault tolerance activities will not be performed for the failed node.
The messages are deleted if they are older than twice the Fault Tolerance threshold for the sources.