Posted On: Oct 11, 2012
We are pleased to announce that Amazon RDS for MySQL now supports “Promote Read Replica” functionality. You can now convert a MySQL Read Replica into a “standalone” DB Instance using the “Promote Read Replica” option. This option stops replication and converts the Read Replica in its existing state into a “standalone” DB Instance.
You can use this option for a number of use cases including:
- Perform DDL operations: DDL operations such as creating/re-building indexes etc. could take a long time and impose significant performance penalty on your DB Instance. You can perform these operations on a Read Replica, and once the operations are complete and the updates are caught up with the Source DB Instance, you can promote the Read Replica, and point your applications to it.
- Sharding embodies the "share-nothing" architecture and essentially involves breaking a larger database up into smaller databases. Common ways to split a database are: Splitting tables that are not joined in the same query onto different hosts or duplicating a table across multiple hosts and then deciding on a hashing algorithm to figure out into which host a row goes. You can create Read Replicas corresponding to each of your “shards” and promote them when you decide to convert them into “standalone” shards. You can then delete the rows or tables that belong to the other shards.
- Recovery against failures: Amazon RDS provides multiple options for data recovery during failures including Multi-AZ deployments and Point in Time Recovery. With the ability to “Promote“, Read Replica can potentially be considered as an additional recovery alternative against failures. However, it is to be noted that with asynchronous replication, database writes occur on a Read Replica after they have already occurred on the Source DB Instance, and this replication “lag” can vary significantly depending on the workload. If your use case requires synchronous replication, automatic failure detection and failover, we recommend you run your DB Instance as a Multi-AZ deployment.
Please refer to the Read Replicas section of the User Guide to learn more.