Failover modes

With client failover, you can configure two adapters as part of a redundant group, so that the secondary adapter may take the place of the primary in the event of a connection loss to the failover endpoint or data source.

Failover logic

There are three client failover modes that the adapters can be configured to use:

• Hot failover

  In hot failover mode, both the primary and secondary adapter are configured and are actively collecting data. The secondary instance is not transmitting data, but rather continuously buffering it.

  When a failover event occurs, the secondary adapter is promoted to primary and any missing data that was previously collected in the buffer is flushed to the destination. For example, if the primary adapter loses its connection at 12:00 AM, the secondary now becomes the primary. At that time, the secondary adapter flushes any data in its buffer that has occurred after 12:00 AM. In this case, it flushes only what has been missed since the failover event occurred.

  The hot failover mode has the least data loss when compared to the warm and cold failover modes – in many use cases there will be no data loss at all. It is strongly recommended to use hot failover mode in combination with redundant DNP3 outstations.

  

• Warm

  In Warm failover mode, both the primary and secondary adapters are connected to the data source, but only the primary adapter is actively collecting data from the data source. Since data is not being collected with the secondary adapter, data is not buffered nor egressed to the data endpoint(s).

  When compared to cold failover mode, warm failover mode does reduce data loss during a failover event.

  In warm failover, secondary adapters connected to the data source maintain their connection by calling a LinkStatus request to each configured data source. This request is sent at an interval set in the LinkStatusRequestInterval configuration. For additional information on this configuration, see Data source parameters.

  If the Link Status Request encounters an issue making the call to the outstation, the outstation's device status changes to DeviceInError. This lowers the overall failover score. Otherwise, device status in warm failover should be ConnectedNoData.

  It is strongly recommended to use warm failover mode in combination with redundant DNP3 outstations.

  

• Cold

  When in Cold failover mode, one DNP3 adapter is actively collecting data from its data source while the secondary instance is configured but not connected to the data source. The secondary adapter does not collect data from the data source, and data is not egressed to the data endpoint(s).

  For implementations that contain a single DNP3 outstation with no redundancy, cold failover is the recommended failover mode.