ibm_container_worker_pool_zone_attachment should wait for ALBs to finish in a new zone

[hasueki]

Terraform Version

Terraform v0.12.20

Affected Resource(s)

• ibm_container_worker_pool_zone_attachment

Terraform Configuration Files

data "ibm_container_cluster" "cluster" {
  cluster_name_id = ibm_container_cluster.cluster.id
}

resource "ibm_container_cluster" "cluster" {
  name              = "cluster-1"
  datacenter        = "dal10"
  machine_type      = "b3c.4x16"
  hardware          = "shared"
  private_vlan_id   = "<PRIVATE_VLAN_ID>"
  public_vlan_id    = "<PUBLIC_VLAN_ID>"
  kube_version      = "1.17"
  default_pool_size = "1"
}

resource "ibm_container_worker_pool_zone_attachment" "zone" {
  worker_pool     = "default"
  cluster         = ibm_container_cluster.cluster.name
  zone            = "dal12"
  private_vlan_id = "<PRIVATE_VLAN_ID>"
  public_vlan_id  = "<PUBLIC_VLAN_ID>"
}

resource "ibm_container_alb" "disable_public_alb1" {
  depends_on = [
    ibm_container_cluster.cluster
  ]

  alb_id = [
    for alb in ibm_container_cluster.cluster.albs:
      alb.id if alb.alb_type == "public" && length(regexall("alb1", alb.id)) > 0
  ][0]
  disable_deployment = true
}

resource "ibm_container_alb" "disable_public_alb2" {
  depends_on = [
    ibm_container_cluster.cluster,
    ibm_container_worker_pool_zone_attachment.zone
  ]

  alb_id = [
    for alb in data.ibm_container_cluster.cluster.albs:
      alb.id if alb.alb_type == "public" && length(regexall("alb2", alb.id)) > 0
  ][0]
  disable_deployment = true
}

Expected Behavior

When creating a worker pool attachment in a new zone, it should wait for the default ALBs (private and public) in the new zones to finish associating to the cluster.

Actual Behavior

The worker pool zone attachment only waits for the new workers to become ready, and finishes before the new ALBs to come up in the new zones. This causes issues in my use case above, to go ahead and disable the ALBs and Terraform apply will fail since the ALB creation is still underway. disable_public_alb1 works as expected because ibm_container_cluster waits for ALBs to finish, but is not the case with ibm_container_worker_pool_zone_attachment.

Steps to Reproduce

1. Set appropriate values
2. terraform apply

References

• fix: ibm_container_worker_pool_zone_attachment should wait for ALBs #1373