RDS instance stays in pending-reboot after apply

[pioneerit]

Provider: AWS, resource: aws_db_instance
In case if RDS instance restored from the snapshot after terraform apply, the instance is in pending-reboot status, because of parameter_group values. It doesn't happen with new instances.

resource "aws_db_parameter_group" "db" {
  name        = "${var.environment}}-db"
  description = "${var.environment}-db parameter group by Terraform"
  family      = "mysql5.6"

  parameter {
# ...
# The are different from the parameters in the snapshot
  }
}

resource "aws_db_instance" "db" {
  identifier = "test"
  snapshot_identifier  = "${var.latest_snapshot}"
  skip_final_snapshot  = true
  parameter_group_name = "${aws_db_parameter_group.db.name}"
  instance_class       = "db.m4.large"
  apply_immediately    = true
}

I manage to do a workaround with:

resource "null_resource" "pending-reboot" {
  triggers {
    db_host = "${aws_db_instance.db.parameter_group_name}"
  }

  provisioner "local-exec" {
    command = "aws rds reboot-db-instance --db-instance-identifier ${aws_db_instance.db.identifier} && aws rds wait db-instance-available --db-instance-identifier ${aws_db_instance.db.identifier}"
  }
}

I believe that it should be a way to adjust the code, to make terraform apply work even in this case. Unfortunately, I'm not using GoLang to do a proper PR.

[bflad]

In version 1.33.0 of the AWS provider, likely releasing later today, the aws_db_instance will now correctly apply parameter_group_name on resource creation when using snapshot_identifier. Please note though that it will still be in pending-reboot state, which will also be fixed shortly.

[bflad]

The handling to prevent pending-reboot on creation (by issuing RebootDBInstance) has been merged into master and will release with version 1.34.0 of the AWS provider, likely later today.

[bflad]

This has been released in version 1.34.0 of the AWS provider. Please see the Terraform documentation on provider versioning or reach out if you need any assistance upgrading.

[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!