azurerm_kubernetes_cluster: state not recorded in case of failure

[jambt]

Hi,

I ran into a bug with the azurerm_kubernetes_cluster. The deployment state is not recorded properly in case of a failure.

Terraform Version

Terraform v0.11.3
azurerm (1.1.2)

Affected Resource(s)

Please list the resources as a list, for example:

• azurerm_kubernetes_cluster

Terraform Configuration Files

resource "azurerm_kubernetes_cluster" "aks_container" {
  name = "${var.name}"
  location = "${var.location}"
  resource_group_name = "${var.resource_group}"
  kubernetes_version = "1.8.7"
  dns_prefix = "default"

  linux_profile {
    admin_username = "admin"
    ssh_key {
      key_data = "${file("~/.ssh/id_rsa.pub")}"
    }
  }

  agent_pool_profile {
    name = "agentpool"
    count = "${var.node_count}"
    vm_size = "${var.node_vm_size}"
    os_type = "Linux"
  }

  service_principal {
    client_id = "${var.client_id}"
    client_secret = "${var.client_secret}"
  }
}

Please note the usage of admin as admin_username. Thus, resource creation will fail as this default value is not valid.

Expected Behavior

The resource should be marked as created, but tainted.

Actual Behavior

The resource is not recorded as created.

Steps to Reproduce

1. terraform apply
2. Change admin_username to something different
3. terraform apply will not work, since the SDK tries to update the username which is not possible.

[katbyte]

Hello @jambt,

Thank you for finding and reporting this bug.

You are correct, it appears that when admin_username is changed a new resource should be created. To fix this we will just have the property force a new resource when it changes.

To address the original issue, when the creation fails the resource enters into a failed state. As far as terraform is concerned the creation call failed and the resource was not created. As the resource is in a unknown state it is left up to the user to decide how to proceed.

[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!