Error on subsequent terraform plan/apply when no changes

[mmindenhall]

I'm creating the following cluster (nearly identical to the example):

# Create a Cluster
resource "mongodbatlas_cluster" "mongo_env_cluster" {
  name = "vendavo-${lower(var.ENV)}"
  group = "${data.mongodbatlas_project.mongo_env_prj.id}"
  mongodb_major_version = "3.6"
  provider_name = "TENANT"
  backing_provider = "AWS"
  region = "${upper(replace(var.REGION, "-", "_"))}"
  size = "M2"
  backup = false
  disk_gb_enabled = false
}

The cluster is created successfully on the first invocation of terraform apply. However, a subsequent execution of terraform plan/apply results in an error, even when the cluster declaration has not changed:

An execution plan has been generated and is shown below.
Resource actions are indicated with the following symbols:
  + create
  ~ update in-place
Terraform will perform the following actions:

  ~ mongodbatlas_cluster.mongo_env_cluster
      disk_size_gb: "2" => "0"

...

terraform: apply
Acquiring state lock. This may take a few moments...

...

Error: Error applying plan:

1 error(s) occurred:

* mongodbatlas_cluster.mongo_env_cluster: 1 error(s) occurred:

* mongodbatlas_cluster.mongo_env_cluster: Error reading MongoDB Cluster vendavo-dev: MongoDB Atlas: 400 Cannot update a M0/M2/M5 cluster through the public API.

It seems that the optional "disk_size_db" property is set to 0 when not provided. Then the current state is read from Atlas, which correctly reports "2" as the value, so it wants to change the value from "2" to "0".

[mmindenhall]

Have implemented a fix, but will take a while to rebuild docker container and test in CI/CD pipeline. PR to come.

[akshaykarle]

Not sure this is limited to just the shared clusters. I've seen this happen with any sized clusters. Will look at your PR, thank you!

[akshaykarle]

Closed by #39