What’s the issue?
When a database is added to an Always On Availability Group, the secondary replicas need an initial copy of the database before synchronization can begin. SQL Server supports two seeding modes for this initial copy: MANUAL, where the administrator restores a backup of the primary on each secondary, and AUTOMATIC, where SQL Server streams the database directly from the primary to the secondary over the AG endpoint.
The seeding mode is set per replica using the SEEDING_MODE option in CREATE AVAILABILITY GROUP or ALTER AVAILABILITY GROUP. Once a replica has been fully seeded and is synchronizing normally, the seeding mode setting continues to apply for any new databases added to the AG in the future.
This finding identifies replicas that are currently configured with a seeding mode of AUTOMATIC. The setting is often left in place after initial seeding without being reviewed for whether it should remain.
Why is this a problem?
Automatic seeding is a good option for the initial copy of a database to a secondary, particularly when paired with trace flag 9567 (often used for compression) to reduce the network and time impact of seeding large databases. The mechanism handles the file copy and synchronization handoff automatically, removing several manual steps from the deployment process.
After the replica is fully seeded and synchronizing, however, the automatic seeding mode continues to apply to any new database added to the AG. This means the next database added is automatically streamed to every secondary configured for automatic seeding, which may not be the desired behavior. In environments where new databases are added regularly, this can produce unexpected resource usage on the primary and secondaries during the seeding process.
Automatic seeding also adds a small amount of ongoing overhead to maintain the readiness state on the primary, since SQL Server must be prepared to seed at any time. The cost is small but real, and it provides no benefit once existing replicas are fully synchronized and stable.
What should you do about this?
Confirm with the team whether automatic seeding is still desired or whether it was left in place from the initial AG deployment.
For replicas that are fully seeded and stable, change the seeding mode to manual using ALTER AVAILABILITY GROUP [AGName] MODIFY REPLICA ON N’ReplicaName’ WITH (SEEDING_MODE = MANUAL);. The change takes effect immediately and does not affect existing synchronization, only how new databases are added in the future.
For environments where automatic seeding is genuinely the preferred method for adding new databases (often the case in heavily automated deployments), document the decision and confirm that the operational impact of automatic seeding is acceptable for the size and frequency of new database additions. Pair the configuration with appropriate use of trace flag 9567 if compression during seeding is beneficial for your network and database sizes.