As long as the database is in a remote server you can create a second Harvest broker as the fail over system.
Install the same Harvest release binaries and configure and test connectivity with the Harvest database.
Use RT_FORCE_NODE_NAME system environment variable in the fail over server to set the host name
to be the same as that of the main Harvest broker and that the rtserver will recognize.
At fail over time when the Harvest main broker is down, then start the fail over broker. At that point when the
broker is up and confirmed running, then all clients should be able to connect to the fail over broker using
that same down broker name (this is transparency).