Synchronization involves one replica sending data changes and the relative replica receiving changes. Data changes include inserts, updates, and deletes made in the replica version. To perform a synchronization, you must be connected as the same database user that was used to create the replica, or the geodatabase administrator.
For two-way and one-way replication, the filters and relationship class rules that were applied at creation are also applied to determine what changes to synchronize. Changes outside the filters and relationship class rules are not synchronized. See Replication and related data to learn how to find the filters and relationship class rules for a replica. Logic is also used to prevent unnecessarily resending changes that have already been sent. For checkout replicas, all edits made to the checkout replica are synchronized.
Data transfer is based on exchanging replication messages. Message exchange between replicas is designed to be sequential. This means that the replicas are set up to exchange messages in a manner similar to the way voice messages are exchanged between parties engaged in a telephone conversation: each party sends a message to the other that is received and processed before the other party responds. The nature of the system ensures that only one replica can send changes at a time.
During synchronization, changes are reconciled and posted into the replica version. During this reconcile, conflicts can occur. You can choose a reconcile policy to define how to handle these conflicts. You can also choose between column-level and row-level conflict detection for this reconcile.
Note:
Metadata for the data you choose to replicate is copied during the replica creation process. However, changes to the metadata are not applied during replica synchronization.