The normal disaster recovery methods that you use on standard databases also apply to replicated databases. Checkpoints, journaling, and rollforwards must be used on the Ingres Replicator system to bring a failed database to a consistent state. For a description of these recovery tools, see the Database Administrator Guide.

If these methods recover the failed database to the exact point of failure, the remote databases queue all the changes that are due to happen to the failed database, ready for the moment when the failed database becomes available. However, if journals, log files, or dump areas are lost, a gap of missing data exists on the failed database. You can recover part or this entire gap of missing data if the records still exist and reflect the data that was lost for the duration of the gap. For more information, see Recovering Transactions from a Peer Database Using the Reconciler.