[Bug]: Applying aws_dms_s3_endpoint attempts to change state even if state change has already been applied.

Terraform Core Version

Terraform v1.6.2

AWS Provider Version

aws v5.64.0

Affected Resource(s)

aws_dms_s3_endpoint

Expected Behavior

After removing the properties date_partition_delimiter and date_partition_sequence on a prior apply and making modifications to other resources, a subsequent Terraform plan and apply should not see these properties in the state for the aws_dms_s3_endpoint

Actual Behavior

A first apply set the properties date_partition_delimiter and date_partition_sequence to null. On a subsequent plan and apply, a difference was noticed where the state showed a previous version of these 2 properties where they were set to values.

This causes subsequent applies to fail because they want to change these values while the DMS Replication Task is running.

Relevant Error/Panic Output Snippet

# module.dms.aws_dms_s3_endpoint.target will be updated in-place
  ~ resource "aws_dms_s3_endpoint" "target" {
      - date_partition_delimiter                    = "SLASH" -> null
      - date_partition_sequence                     = "yyyymmddhh" -> null
        id                                          = "dms-endpoint-id"
        tags                                        = {}
        # (35 unchanged attributes hidden)
    }

Terraform Configuration Files

TerraformConfig.txt.gpg.txt

Steps to Reproduce

Apply changes to a DMS S3 Endpoint to remove date partition settings
Run a subsequent plan/apply with no further changes.

Debug Output

No response

Panic Output

No response

Important Factoids

No response

References

No response

Would you like to implement a fix?

None

hashicorp / terraform-provider-aws