When a Broker that is currently assigned an Active role fails, the Failover Manager takes the following actions:

  1. The Failover Manager determines the next candidate to make Active based on the priority and running states of other Brokers.

  2. After identifying the new Active Broker, the Failover Manager updates the Broker on all other Active components (SAM and Domain Managers) and within the Failover Manager as well. This ensures that any component that wants to synchronize with another component will have a Broker that can be used.

    For the SAM server with Notification Cache Publishing, when the Active Broker fails,

    • The Failover Manager uses the OnChangePOST value from the failover.conf file to update the new Broker URL location in the EMC Data Access API (smarts-edaa).

  3. The Failover Manager tries to restart the failed Broker if the host where it resides is functioning.

    • If the InstalledServiceName attribute is defined, the Failover Manager tries to restart the failed Broker service.

      • If the InstalledServiceName attribute is not defined, the Failover Manager tries to start the failed Broker using the brstart command.

  4. If restarting the failed Broker is successful, the Failover Manager marks the restarted Broker as Standby.

    The Failover Manager sends a Broker Restart email that an error occurred during the restart of the Broker <host:port> or the restart of the Broker <host:port> successfully completed.

    If you want to switch the newly-promoted Broker back to Standby and the Standby Broker back to Active, perform a failback action as described in “Optional additional configuration tasks for Broker failover” on page 53.