Terraform didn't wait for RDS to finish modifying

[hashibot]

This issue was originally opened by @davidvuong as hashicorp/terraform#16170. It was migrated here as a result of the provider split. The original body of the issue is below.

---

Terraform Version

Terraform v0.10.5

Debug Output

First terraform apply

...
module.tf_rds.aws_db_instance.xxxx: Modifying... (ID: xxx)
  allocated_storage: "30" => "60"
module.tf_rds.aws_db_instance.xxx: Still modifying... (ID: xxx, 10s elapsed)
module.tf_rds.aws_db_instance.xxx: Still modifying... (ID: xxx, 20s elapsed)
module.tf_rds.aws_db_instance.xxx: Still modifying... (ID: xxx, 30s elapsed)
module.tf_rds.aws_db_instance.xxx: Modifications complete after 31s (ID: xxx)

Apply complete! Resources: 0 added, 1 changed, 0 destroyed.

Subsequent terraform apply

Error applying plan:
1 error(s) occurred:
* module.tf_rds.aws_db_instance.xxx: 1 error(s) occurred:
* aws_db_instance.xxx: Error modifying DB Instance xxx: InvalidDBInstanceState: Database instance is not in available state.
	status code: 400, request id: xxx

Expected Behavior

terraform apply should wait for RDS to finish modifying.

Actual Behavior

1. First apply completes within 31 seconds of modifying then mentions it successfully completes
2. The plan immediately after the apply causes a -/+ on the already modified RDS
3. Subsequent apply tries to make changes
4. AWS complains and mentions it is not in an available state.

To me, on the surface it seems like there are multiple issues:

1. Terraform shouldn't say the modifications are complete when they aren't
2. Terraform should wait for the modifications to complete (perhaps poll AWS?)
3. The state possibly wasn't updated after the 1st "successful" modification (or it was but was updated again during the state refresh?) when I went to plan immediately after.

[dvishniakov]

This is still happening on TF 0.11. Haven't tried it on 0.12. In our case the same issue happens after modifying DB cluster (i.e. changing SG groups), which leads to modification of DB instances. So when cluster modification is complete, different changes to cluster_db_instance (i.e instance type) fail, because DB instances are still being modified due to SG change.

[bflad]

Hi folks 👋 The original issue here regarding the aws_db_instance resource not waiting for success on updates should have been resolved in version 1.34.0 of the Terraform AWS Provider. If you are still having lingering issues on more recent versions of the Terraform AWS Provider (2.69.0 is the current latest), please file a new bug report issue following the template and we can take a fresh look. Thanks.

[ghost]

I'm going to lock this issue because it has been closed for 30 days ⏳. This helps our maintainers find and focus on the active issues.

If you feel this issue should be reopened, we encourage creating a new issue linking back to this one for added context. Thanks!