Missing update-variant keepers, breaking gpu_driver_version update

[S4ckBr]

TL;DR

There is no node-pool recreation after editing the gpu_driver_version field

Expected behavior

• The terraform apply succeeds
• The node-pool suffix name should change (update-variant)

Observed behavior

• The terraform apply fails:

module.gke_cluster.google_container_node_pool.pools["my-node-pool"]: Creating...
╷
│ Error: resource - projects/myproject/locations/us-central1/clusters/my-cluster/nodePools/my-node-pool-XXXX - already exists
│ 
│   with module.gke_cluster.google_container_node_pool.pools["my-node-pool"],
│   on .terraform/modules/gke_cluster/modules/beta-public-cluster-update-variant/cluster.tf line 565, in resource "google_container_node_pool" "pools":
│  565: resource "google_container_node_pool" "pools" {
│ 
╵

• The node-pool name suffix is not updated. GKE does not support updating in place the gpu_driver_version field. This explains the failure

Terraform Configuration

module "gke_cluster" {
  source  = "terraform-google-modules/kubernetes-engine/google//modules/beta-public-cluster-update-variant"
  version = "30.2.0"

  # Cluster basics
  project_id      = google_project.this.project_id
  name            = "my-cluster"
  region          = "us-central-1"

  ## Node Pools
  remove_default_node_pool = true
  initial_node_count       = 1

  node_pools = [{
    name = "my-node-pool"

    accelerator_count  = 1
    accelerator_type   = "nvidia-tesla-t4"
    gpu_driver_version = "DEFAULT"
  }]
}

Terraform Version

Terraform v1.5.7
on linux_amd64

Additional information

Seems to be fixed by #1913

[DrFaust92]

this can be closed.

cc: @apeabody

[apeabody]

Closed by #1874