Ressource google_container_cluster shouldn't delete and create cluster after image_type change

[etlfg]

Terraform Version

$ terraform -v
Terraform v0.11.7

Affected Resource(s)

google_container_cluster

Terraform Configuration Files

resource "google_container_cluster" "primary" {
  name = "${var.name}"
  zone = "${var.zone}"
  initial_node_count = "${var.node_count}"
  min_master_version = "${var.min_master_version}"
  node_version = "${var.node_version}"
  
  image_type = "COS" # Change this to "Ubuntu"

  node_config {
    disk_size_gb = "${var.disk_size_gb_node}"
    machine_type = "${var.machine_type_node}"
  }
}

Expected Behavior

We should be able to change the image_type of a cluster without deleting it. GCP already manages to do It by updating one node after the other and by provisioning the workload on the others.

Actual Behavior

Terraform plans to destroy and recreate the cluster.

-/+ destroy and then create replacement

Terraform will perform the following actions:

-/+ google_container_cluster.primary (new resource required)

...

node_config.0.image_type:              "UBUNTU" => "COS" (forces new resource)
...

However this feature is designed on GCP side since you're able to do It on the web UI.

It is also possible from the gcloud cli :
gcloud container clusters upgrade --image-type [IMAGE_TYPE] [CLUSTER_NAME] [--node-pool [POOL_NAME]]
See the doc for more information about this.

The call seems to be done with a POST to https://console.cloud.google.com/m/operations?organizationId=<orgId>&operationType=cloud-console.container.updateCluster&pid=<envName> with desiredNodePoolImageType set to "COS" in the params.

Steps to Reproduce

Please list the steps required to reproduce the issue, for example:

1. terraform apply
2. switch image_type from Ubuntu or Cos accordingly to the first applied image
3. terraform plan

References

Seems to be related to :

• google_container_cluster taint triggers new instance #1371

[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. If you feel I made an error 🤖 🙉 , please reach out to my human friends 👉 [email protected]. Thanks!