adding ha = true to existing do_kubernetes_cluster (1.23.14) fails

[gladiatr72]

Bug Report

---

Describe the bug

according to DO, adding HA backplane functionality is supported post kube 1.21. It works from the web console; however, the plan after adding ha = true to an existing cluster indicates a destroy/create cycle would be necessary. It appears the effort to support ha as a valid on-create parameter might pre-date DO's support for enabling HA post-creation.

Affected Resource(s)

• digitalocean_kubernetes_cluster

Expected Behavior

applying a plan with a newly truthful ha parameter would provide a non-destructive update

Terraform Configuration Files

  resource digitalocean_kubernetes_cluster cluster {
    name    = "do-2"
    region  = var.do_region
    version = "1.23.14-do.0"

    ha = true /* post creation setting */

    node_pool {
      name       = "default-node-pool"
      size       = "s-2vcpu-4gb"
      node_count = 4
    }
  }

Terraform version

Terraform v0.15.5 on linux_amd64

• provider registry.terraform.io/digitalocean/digitalocean v2.26.0

[maiku1008]

Any news on this? It would actually be very helpful.

The workaround is to:

• Add HA from the web console.
• in the terraform code, change ha to true.
• run terraform refresh to update the state file.