Do the following to simulate a failover event and then restore the cluster to the original configuration:
- View the current cluster status: ./dmcluster status
- On the designated-active system, shut down the Data Mover repository: /etc/init.d/postgresql stop
- Check the cluster status to verify the initiation of failover: ./dmcluster statusThe failover monitor recognizes that the Data Mover repository on the designated-active system is down. However, the failover monitor makes retry attempts to check the repository status before initiating failover. So the process can take time.Stopping the Data Mover services, such as the daemon or agent, does not result in a failover event. The failover monitor instead detects that those services are down and attempts to restart the services before initiating a failover.
- On the designated-active system, restart the Data Mover repository: /etc/init.d/postgresql stop
- On the designated-active system, run the command: ./dmcluster switchback
- Verify the restoration of the original cluster configuration: ./dmcluster status
Simulating a Failover Event
Restoring the Cluster to the Original Configuration: