Delete kubernetes node pull with Terraform without cluster recreating

Question

I have created DigitalOcean Kubernetes cluster with additional node pull. Everything works great.

resource "digitalocean_kubernetes_cluster" "k8s_cluster" {
  depends_on = [digitalocean_container_registry.docker_registry]
  name    = "dev"
  region  = var.region
  version = "1.30.2-do.0"
  registry_integration = true

  node_pool {
    name       = "dev-pool"
    size       = "s-2vcpu-2gb"
    auto_scale = true
    min_nodes  = 2
    max_nodes  = 4

  }
}

resource "digitalocean_kubernetes_node_pool" "new_k8s_node_pool" {
  cluster_id = digitalocean_kubernetes_cluster.k8s_cluster.id

  name       = "dev-pool-new"
  size       = "s-2vcpu-4gb"
  auto_scale = true
  min_nodes  = 2
  max_nodes  = 4
}

But now I want to delete old node pull (dev-pool) and keep the cluster working with the new node pool (dev-pool-new). But when I do terraform plan I can see that terraform is going to recreate cluster with the new node pull. I don't want to recreate cluster, I just want to delete default node pool and keep only new node pull. How can I do that? Is it available to delete default node pull without cluster recreating?

resource "digitalocean_kubernetes_cluster" "k8s_cluster" {
  depends_on = [digitalocean_container_registry.docker_registry]
  name    = "dev"
  region  = var.region
  version = "1.30.2-do.0"
  registry_integration = true

  node_pool {
    name       = "dev-pool-new"
    size       = "s-2vcpu-4gb"
    auto_scale = true
    min_nodes  = 2
    max_nodes  = 4

  }
}

terraform plan output samples:

digitalocean_kubernetes_cluster.k8s_cluster must be replaced

digitalocean_kubernetes_node_pool.new_k8s_node_pool will be destroyed

Answer 1

The cluster must be re-created if the default node pool (the block node_pool within digitalocean_kubernetes_cluster) is modified for certain parameters. This is a requirement enforced by the Digital Ocean API.

If you want to modify a node pool without cluster re-creation, then you must use the additional node pool resource instead.

resource "digitalocean_kubernetes_node_pool" "new" {
  cluster_id = digitalocean_kubernetes_cluster.k8s_cluster.id

  name       = "dev-pool-new"
  size       = "s-2vcpu-4gb"
  auto_scale = true
  min_nodes  = 2
  max_nodes  = 4
}

Note that for your current cluster you cannot remove the default node pool without causing a cluster re-creation, but can add the new node pool.