Why is an Amazon RDS DB instance stuck in the modifying state when I try to increase the allocated storage?
Last updated: 2021-06-15
I want to increase the allocated storage for an Amazon Relational Database Service (Amazon RDS) DB instance, but the operation is stuck in the modifying state.
By design, storage scaling operations for an Amazon RDS DB instance have minimal impact on ongoing database operations. In most cases, the storage scaling operations are completely offloaded to the Amazon Elastic Block Store (Amazon EBS) layer and transparent from the database. This process is typically completed in a few minutes. However, for some legacy Amazon RDS storage volumes, you might require a different process for modifying the size, IOPS, or volume type of your Amazon RDS storage. You might need to make a full copy of the data using a potentially long-running I/O operation.
Most RDS volume geometries include either one Amazon EBS volume or four striped EBS volumes in a RAID0 configuration depending on the size of the allocated storage. You must use the legacy method under either of the following conditions:
- Your RDS instance doesn't have either one or four volumes.
- The target size for your modification increases the allocated storage beyond 400 GB.
You can view the number of volumes in use on your RDS instances using the Enhanced Monitoring metrics. Also, any source volume that uses previous generation EBS volumes require the legacy method for modifying the size of the allocated storage.
The following factors can affect the time required to increase the allocated storage of an RDS DB instance:
- The legacy method uses I/O resources, and this could increase your database workload. It's a best practice to use the minimal impact method whenever possible. The minimal impact method doesn't use any resources on the database. If you must use the legacy method, then it's a best practice to schedule the storage increase operations outside of the peak hours. This might reduce the time required to complete the storage increase operations.
- If you have high load conditions and must use the legacy method, then you can create a read replica for the RDS DB instance. You can perform the storage scaling operations on the read replica and then promote the read replica DB instance to the primary DB instance.
- If you have high-load conditions, then do the following:
- Create a read replica for the RDS DB instance.
- Perform storage scaling operations on the read replica.
- Promote the read replica DB instance to the primary DB instance.
After a storage modification has started, the operation can't be canceled. The DB instance status is in the modifying state until the Amazon EBS operations are complete. You can restore a DB instance to a specified time or restore from a DB snapshot to create a new DB instance with the original storage configuration. A DB instance that's restored is not in the modifying status.