Replication Invocation

For performance reasons, the Sync Server does not immediately replicate changes received from the synchronization requests to the source, and vice versa. Replication invocation is directed by one of the following:

Batching of synchronization requests by a fixed elapsed period of time.
Configuring a subscription to perform replication at any time you issue a special command.

When batching synchronization requests, the Sync Server begins replicating synchronization requests when the number of seconds defined for the batch window has elapsed since the last replication finished.

For JDBC subscriptions, when you have two devices which both get data from the same subscription, it will take two replication cycles on the server before new data entered in the first device can be synchronized to the second device. The flow of the data goes as follows:

Data is entered on device D1.
Device D1 synchronizes, data goes to the mirror database.
The Sync Server performs the first replication cycle and the data is on the source database.
Device D2 synchronizes, but does not yet receives the data originally entered on device D1.
Sync Server performs the second replication cycle.
Device D2 synchronizes after the second replication cycle, and now D2 receives the data originally entered on device D1.

For information about how to set the Sync Server to perform replication at any given time, see Replicating on demand.

Conceptos relacionados

Data set up and replication

Tareas relacionadas