[ISSUE] AWS S3 Mount does not use cluster_id during plan

[theesen]

Terraform Version

Terraform v0.12.30

• provider.aws v3.32.0
• provider.databricks v0.3.2

Affected Resource(s)

Please list the resources as a list, for example:

• databricks_aws_s3_mount

Terraform Configuration Files

resource "databricks_aws_s3_mount" "generic" {
  for_each = toset(var.buckets)

  cluster_id       = "999"
  instance_profile = "some_instance_profile"
  s3_bucket_name   = each.key
  mount_name       = "data/${replace(each.key, "/[^0-9A-Za-z_]/", "_")}"
}

Debug Output

  # module.mountpoints.databricks_aws_s3_mount.generic["com.my.bucket"] must be replaced
-/+ resource "databricks_aws_s3_mount" "generic" {
      ~ cluster_id       = "123" -> "999" # forces replacement
      ~ id               = "com.my.bucket" -> (known after apply)
        instance_profile = "arn:aws:iam::XXXXX:instance-profile/databricks-shared-instance-profile"
        mount_name       = "data/com.my.bucket"
        s3_bucket_name   = "com.my.bucket"
      ~ source           = "s3a://com.my.bucket" -> (known after apply)
    }

Cluster ID: 999 is the ID of the actual cluster I would like the mount to use, while 123 is the ID of a temporary cluster that is auto created by the terraform plan.

Expected Behavior

Mounting should happen on the provided cluster_id during terraform plan & apply.

Actual Behavior

A new cluster is being launched every time terraform plan is executed.

Steps to Reproduce

1. Create a mountpoint using a cluster_id with terraform.
2. Rerun the same terraform plan without any changes.
3. Terraform will plan and apply the changes on a different cluster then specified (In my example on 123 instead of 999)

[nfx]

will be fixed in #497