[Bug] Terraform tries to reattach a nodegroup to a replaced cluster

[jan-di]

The documentation states, that if the cluster_id of a nodegroup changes, the nodegroup will be recreated.
Actually, terraform tries to reattach the nodegroup by updating the cluster_id in place.

Terraform Version

1.4.2, shouldn't be related to version.

Affected Resource(s)

• openstack_containerinfra_nodegroup_v1

Terraform Configuration Files

resource "openstack_containerinfra_nodegroup_v1" "static" {
  name         = each.value.name
  cluster_id   = local.cluster.id
  node_count   = 1

}

Expected Behavior

Cluster gets replaced, therefore cluster will have a new id
After terraform recreated the cluster, it tries to update the nodegroup with the new ID of the parent cluster. This won't work, as you cannot attach an existing nodegroup to a new cluster.

Actual Behavior

Terraform detects the change of cluster ID and replaces the nodegroup with a new one.

Steps to Reproduce

1. terraform apply -replace "path.to.parent.cluster"

Possible Solution

I'm not too deep into the sourcecode, but I guess the ForceNew property of cluster_id must be set to true.

https://github.com/terraform-provider-openstack/terraform-provider-openstack/blob/main/openstack/resource_openstack_containerinfra_nodegroup_v1.go#LL45C10-L45C10

[mnaser]

Indeed, you're right, if the cluster_id is to be changed, the node group should indeed disappear.