The following illustrates a simplified synchronization topology with a single failover server on the source, destination, and PingDataSync, respectively. The gray lines represent possible failover connections in the event the server is down. The external servers are prioritized so that src1 has higher priority than src2; dest1 has higher priority than dest2.
The main PingDataSync and its redundant failover
instance communicate with each other over LDAP and bind using
User,cn=Root DNs,cn=config. The servers run periodic health checks on each
other and share information on all changes that have been processed. Whenever the failover
server loses connection to the main server, it assumes that the main server is down and
begins processing changes from the last known change. Control reverts back to the main
server once it is back online.
Unlike the PingDataSyncs, the external servers and their corresponding failover server(s) do not run periodic health checks. If an external server goes offline, the failover server will receive transactions and remain connected to PingDataSync until the Sync Pipe is restarted, regardless of if the main external server comes back online.