Stopping the replication job

Stopping the job by using a savepoint

You must always stop the replication job by using a savepoint. A savepoint is a consistent snapshot of the entire job state, including the enumerator position, in-flight splits, and the lastEndTimestamp property, which the system writes to durable storage.

Stopping the job with a savepoint allows you to restart the job later from exactly where it left off. This process prevents the need to re-run the initial full snapshot scan, saving time and resources.