[Bug]: RDS SQLServer — InvalidParameterCombination: You can't specify storage throughput for storage type io1.

[marceloboeira]

Terraform Core Version

1.5.2

AWS Provider Version

4.67.0

Affected Resource(s)

• aws_db_instance

Expected Behavior

Error during plan stating this "target state" is not possible, something along the lines of:

terraform plan ...
...
Validation error: You can't specify storage throughput for storage type io1.

Actual Behavior

• Error after applying, changes applied partially (gp3 -> io1 works, storage throughput stays pending)
• A new terraform plan shows a diff

~ resource "aws_db_instance" "this" {
       ~ storage_throughput                    = 0 -> 1000
...

And if you try to apply that, it fails with the mentioned error

Relevant Error/Panic Output Snippet

│ Error: updating RDS DB Instance (<redacted>): operation error RDS: ModifyDBInstance, 
https response error StatusCode: 400, RequestID: <redacted>, 
api error InvalidParameterCombination: 
You can't specify storage throughput for storage type io1.

Terraform Configuration Files

In the steps to reproduce...

Steps to Reproduce

1. Provision a DB with this config:

resource "aws_db_instance" "this" {
  ...
  storage_type         = "gp3"
  storage_throughput   = 1000
}

2. Then after plan/apply, switch the config to be:

resource "aws_db_instance" "this" {
  ...
-  storage_type         = "gp3"
+  storage_type         = "io1"
  storage_throughput    = 1000
}

3. Then plan/apply that change (which "works")

4. The gp3 to io1 change works, apply "finishes"

5. The next terraform plan shows this diff

~ resource "aws_db_instance" "this" {
       ~ storage_throughput                    = 0 -> 1000
...

6. And if you try to apply that, it fails with the mentioned error

Important Factoids

• Only observed with SQLServer, it is also probably engine specific to some extend
• Interestingly enough, AWS Accepts the first API call to switch the storage type but not the next ones ...

References

• [Bug]: "aws_db_instance", "storage_type" (io1 > gp3) can't be changed if "allocated_storage" is under 400 GB  #28589 — it seems we have other related issues, should we gather together and build some plan-time validation for those?

Would you like to implement a fix?

Yes

[github-actions]

Community Note

Voting for Prioritization

• Please vote on this issue by adding a 👍 reaction to the original post to help the community and maintainers prioritize this request.
• Please see our prioritization guide for information on how we prioritize.
• Please do not leave "+1" or other comments that do not add relevant new information or questions, they generate extra noise for issue followers and do not help prioritize the request.

Volunteering to Work on This Issue

• If you are interested in working on this issue, please leave a comment.
• If this would be your first contribution, please review the contribution guide.