Currently the system allows for the DPR client to send files for ingest to a single instance of the DPR Server and then rely on processes on the server to perform replication of this data (and data in the database that the client has also updated) (which we have termed "single ingest"). This is currently done by replicating the files on the file-system with rsync and by replicating the DR database between the master and the replica. The main issue with this approach is that the database contains information about multiple repository locations that exist on the server. By replicating the database exactly the repository locations have to be the same between the master and the replica which places limitations on how logical disk space is configured for the replica (and note that the replica is intentionally using a different file-system from the master). There are also some issues with the use of bandwidth to pass data from the master to the replica.
This enhancement request is to alter the DPR Server to have it handle the replication so as to allow for the database on the replica to differ in terms of repository location setup and so as to allow more control of the use of bandwidth to the replica. It may also be desirable to move a lot of the onus on updating the database from the DPR Client to the DPR Server as part of this change.