Why is my Amazon RDS DB instance is in a failed state?
Last updated: 2020-01-02
My Amazon Relational Database Service (Amazon RDS) instance is in a failed state. What does this mean, and how do I fix it?
The failed state for RDS DB instances usually indicates that the hardware associated with your RDS DB instance failed, and that AWS wasn’t able to recover it.
If you previously configured a backup solution for your RDS DB instance, restore it from one of the following backups:
- An automatic or manual backup.
- Point-in-time recovery (PITR).
Note: PITR is available only for RDS DB instances with automatic backups. To retain the same endpoint name for your RDS DB instance, perform a PITR of your instance using a different name before deleting the instance that is in the failed state.
After recovering your instance by using your preferred backup solution, delete the failed instance.
After the failed DB instance is deleted, you can optionally rename the newly restored DB instance using the old DB instance's endpoint name, which means you don't need to change the configuration of your application to connect to the new DB instance.
When you restore a DB instance from a snapshot, the I/O latency might be high to start. As you access the data on the instance, the latency decreases.
To help mitigate failures, consider running your RDS DB instances in a Multi-AZ deployment, and keep the best practices for Amazon RDS in mind when designing your architecture.