sql_mode in an RDS parameter group always causes an update

[bitglue]

Given this:

resource "aws_db_parameter_group" "test" {
    name = "test"
    family = "mysql5.6"
    description = "just a test"

    parameter {
        name = "sql_mode"
        value = "STRICT_ALL_TABLES"
    }
}

You can apply that, and a parameter group will be made with sql_mode=STRICT_ALL_TABLES. The state file will reflect that fact:

                "aws_db_parameter_group.test": {
                    "type": "aws_db_parameter_group",
                    "primary": {
                        "id": "test",
                        "attributes": {
                            "description": "just a test",
                            "family": "mysql5.6",
                            "id": "test",
                            "name": "test",
                            "parameter.#": "1",
                            "parameter.1506707202.name": "sql_mode",
                            "parameter.1506707202.value": "strict_all_tables"
                        }
                    }
                },

But now do a plan:

~ aws_db_parameter_group.test
    parameter.2228267888.name:  "" => "sql_mode"
    parameter.2228267888.value: "" => "STRICT_ALL_TABLES"

If I'm reading this correctly, this means Terraform read the resource, and thinks that the parameter is empty and it needs to create it. I'm not expecting any update because the parameter group already exists with sql_mode=STRICT_ALL_TABLES in RDS.

Confusingly, my attempts to instrument the conversation with RDS indicates that Terraform is getting sql_mode=STRICT_ALL_TABLES from RDS, at least as far as flattenParameters. And oddly, I can reproduce this only with sql_mode -- every other parameter I've tried works fine. Not sure what's special about that one, and I'm out of ideas.

[catsby]

Fixed in #1225

[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 have found a problem that seems similar to this, please open a new issue and complete the issue template so we can capture all the details necessary to investigate further.