Oracle Cloud Infrastructure Documentation

Overview of Inbound Replication

Inbound replication uses replication channels configured in MySQL HeatWave Service to copy transactions from different locations to a DB system. Each channel connects a source (a MySQL instance or another DB system) to the replica (a DB system), and copies data from the source to the replica.

Asynchronous replication means that the replica does not need to be running and connected to the source all the time. It can pick up new updates whenever it is online and reconnects.

A replica DB system can connect to up to 18 MySQL sources by using one replication channel for each source. The source servers do not control the DB system replica and do not need to have permission to write to it. The replica uses each configured replication channel to connect to a source, providing a set of replication user credentials. Over the connection formed by the channel, the replica retrieves committed transactions from the source. Then the replica writes those transactions to its own copy of the databases.

Each replication channel uses relay logs to store temporary transactions before they are applied. Disk space usage on the replica DB system can increase as the number of channels increases.

Note

Inbound replication is not a managed functionality. You are responsible for configuring and maintaining the channels, and for ensuring that the traffic between each source and the replica is properly configured.

Impact of the DB System Operations on the Replication Channels

The DB system operations affect the replication channels on the DB system. These operations are applied across all configured channels on the DB system. As the number of channels increases, the operations might take longer to complete.

  • If you stop the DB system, it stops any enabled channels and changes their state to Needs Attention.
  • If you delete the DB system, it deletes any channels attached to the DB system.
  • If you restart, upgrade, or update a DB system, enable or disable high availability, or perform a switchover, it temporarily suspends replication on all channels while the operation is ongoing. When the operation is complete, replication is resumed. Each channel state changes to Updating while the channels are being resumed.