Since a few days ago, I'm testing a RDS Aurora Writer-Reader (like classic mysql Master-Slave). Specifically in its recovery capacity agains a failover in the writer, among other task, RDS should swap itself the Writer and Reader Role (all good in this step) and swap the endpoint CNAME so that the application can continue normally.

In my case, this final step not happend, and my application still pointing to the same machine (this machine is in failover and then it's degraded to reader).

To test this, I use "Test Failover" button in RDS panel.

The feature of change CNAME endpoint is mentioned in the official docs: https://aws.amazon.com/rds/faqs/#45

In the following paragraph:

Failover is automatically handled by Amazon RDS so that you can resume database operations as quickly as possible without administrative intervention. When failing over, Amazon RDS simply flips the canonical name record (CNAME) for your DB instance to point at the standby, which is in turn promoted to become the new primary. We encourage you to follow best practices and implement database connection retry at the application layer.

  • 1
    Are you using the cluster hostnames? They include the words cluster and cluster-ro in the hostname. – Michael - sqlbot Mar 14 '18 at 16:47
  • Thank you @Michael - sqlbot, l, that was the problema. To use the aurora HA capacity, I must use the cluster endpoint (Inside cluster menu). When I read the faq I misunderstood the change of name in the instance endpoint. – JDaniel Mar 14 '18 at 17:23

Based in @Michael - sqlbot, I fix the problem. When you use aurora in HA configuration, you need to use, in your application config, the cluster endpoint, not the instance endpoint, because the name change inside the cluster.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.