Use the dmcluster service script, found in /opt/teradata/client/nn.nn/datamover/failover/, to run commands on the failover cluster. Whendmcluster config completes the configuration of the cluster, the following occurs:
- Services are stopped and restarted in active mode on the designated-active servers
- Standby components are stopped
- Active component servers are monitored at regular intervals by the monitoring service through an SSH connection making sure services are running
If the active Data Mover daemon, REST service, repository, ActiveMQ, or all agents used by the active daemon are unavailable, the following occurs:
- A failover sequence is initiated
- The failover sequence stops the Data Mover daemon, REST, agents, ActiveMQ, and bundled DSA (if applicable) on the designated-active servers and starts these components on the designated-standby server
When the failover sequence completes, the monitoring service shifts to the standby monitoring server and begins monitoring the now-active standby components. If the monitoring service finds a running Data Mover component or process on the standby server, it automatically stops to prevent two services from being active at the same time.
When the Data Mover cluster is actively running on the designated-standby servers, the high availability service is limited with the following changes:
- The Data Mover repositories are no longer kept in synchronization, the synchronization service is disabled and the sync monitoring service is not in use.
- No additional failover occurs when the main components are unavailable. The cluster must be manually switched back to bring the cluster back to the original active state.
The dmFailover.log file on the monitor servers provide more information on the failover sequence.