AWS GovCloud RDS Postgres Cluster (Non-Aurora) reports InvalidParameterCombination for supported versions

[celestialorb]

Community Note

• Please vote on this issue by adding a 👍 reaction to the original issue to help the community and maintainers prioritize this request
• 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
• If you are interested in working on this issue or have submitted a pull request, please leave a comment

Terraform CLI and Terraform AWS Provider Version

Terraform v1.2.8
on linux_amd64
+ provider registry.terraform.io/hashicorp/aws v4.30.0
...

Affected Resource(s)

• aws_rds_cluster

Terraform Configuration Files

Please include all Terraform configurations required to reproduce the bug. Bug reports without a functional reproduction may be closed without investigation.

data "aws_rds_engine_version" "this" {
  engine             = "postgres"
  preferred_versions = ["13.6"]
}

resource "aws_rds_cluster" "this" {
  ...

  engine         = data.aws_rds_engine_version.this.engine
  engine_version = data.aws_rds_engine_version.this.version

  ...
}

Expected Behavior

I expect to be able to create an RDS cluster with Postgres (non-Aurora) version 13.6 without an error.

Actual Behavior

I receive the following error:

Error: error creating RDS cluster: InvalidParameterCombination: The specified DB engine version isn't valid for the postgres DB cluster. Specify a valid DB engine version.

I'm provisioning this in an AWS GovCloud region, and I have confirmed that the AWS API reports this as being a valid engine version for RDS.

# aws rds describe-db-engine-versions --engine postgres --engine-version 13.6
{
    "DBEngineVersions": [
        {
            "Engine": "postgres",
            "EngineVersion": "13.6",
            "DBParameterGroupFamily": "postgres13",
            "DBEngineDescription": "PostgreSQL",
            "DBEngineVersionDescription": "PostgreSQL 13.6-R1",
            "ValidUpgradeTarget": [
                {
                    "Engine": "postgres",
                    "EngineVersion": "13.7",
                    "Description": "PostgreSQL 13.7-R1",
                    "AutoUpgrade": false,
                    "IsMajorVersionUpgrade": false
                },
                {
                    "Engine": "postgres",
                    "EngineVersion": "14.2",
                    "Description": "PostgreSQL 14.2-R1",
                    "AutoUpgrade": false,
                    "IsMajorVersionUpgrade": true
                },
                {
                    "Engine": "postgres",
                    "EngineVersion": "14.3",
                    "Description": "PostgreSQL 14.3-R1",
                    "AutoUpgrade": false,
                    "IsMajorVersionUpgrade": true
                },
                {
                    "Engine": "postgres",
                    "EngineVersion": "14.4",
                    "Description": "PostgreSQL 14.4-R1",
                    "AutoUpgrade": false,
                    "IsMajorVersionUpgrade": true
                }
            ],
...

Steps to Reproduce

1. Setup a basic AWS RDS cluster configuration in GovCloud with engine = "postgres" and engine_version = "13.6"
2. terraform apply

Important Factoids

• I can provision a single RDS instance via the console with Postgres and version 13.6.
• I can provision an Aurora Postgres RDS cluster with version 13.6 by changing the engine to aurora-postgresql but this is not the desired infrastructure for my use case.
• I have tried provisioning a different version of the Postgres engine (12.10) but received the same error.

This feels as if I'm missing something about AWS RDS, perhaps Postgres 13.6 isn't a valid option for an AWS RDS cluster. I haven't tried this in a non-GovCloud region yet. If it works in a non-GovCloud region then this feels like a bug in the AWS Terraform provider (perhaps there's some translation missing when talking to the GovCloud API versus the commercial AWS API?).

[shengjiangfeng]

same questions

[shengjiangfeng]

no idea

[xandout]

I think I solved it.

It's a bad error message from AWS or the provider, IDK

The solution for me was to use a supported DB instance size. db.t3.medium won't work but db.m5d.large works fine.

Bizarre. I only caught it because I tried to create a multi-az postgres cluster(not aurora) via the console and noticed the instance size options DO NOT include t3.

I assume this is due to the extra horsepower needed to handle replication.

Hope this helps!