What are best practices for setting a test environment for RDS Aurora PostgreSQL

Question

Customer who periodically simulates an outage in RDS Aurora PostgreSQL by stopping and starting the entire Aurora cluster and observing the effect on different upstream systems. This is not an ideal situation as it takes some time for the stopping/starting and the environment is offline and unavailable for other testing use during the duration. They were thinking of setting up a reader node for testing and isolate testing to this node. However, they indicated that these reader nodes can only be rebooted and not just stopped, this may not be ideal for them.

Is there a way for the customer to have a similar parallel environment for testing purposes for these offline scenarios while maintaining functionality of the database for other testing?

Accepted Answer

Currently Aurora PG does not support multi-primary. Since this testing is done periodically, one option would be to restore a snapshot every time for simulation. Second option would be the promote read replica and delete the read-replica after the testing. With this, you need to create new read-replica every time. https://docs.aws.amazon.com/AmazonRDS/latest/UserGuide/USER_PostgreSQL.Replication.ReadReplicas.html#USER_PostgreSQL.Replication.ReadReplicas.Limitations

Another option would be to keep two active read-write databases. To configure this, you can first restore a snapshot and then keep a DMS instance to replicate changes from source to target instance. https://docs.aws.amazon.com/dms/latest/userguide/CHAP_Target.PostgreSQL.html

Another option would be to keep two active read-write databases. To configure this, you can first restore a snapshot  and then keep a DMS instance  to replicate changes from source to target instance. https://docs.aws.amazon.com/dms/latest/userguide/CHAP_Target.PostgreSQL.html

What are best practices for setting a test environment for RDS Aurora PostgreSQL

関連するコンテンツ