Openstack - Terraform assign servergroup to instances

Question

I currently use Terraform to create Openstack instances with different modules. Here is the module instances and a small bit of its definition :

// Module
module "instances" {
  source = "./modules/instance"

  for_each = local.instances

  env              = var.env
  type             = each.key
  icount           = each.value.count
  image            = each.value.image
  flavor           = each.value.flavor
  volume           = each.value.volume
  volumes          = each.value.volumes
  network_id       = local._network.id
  servergroup_name = var.servergroups ? each.value.servergroup_name : null
  dns              = each.value.dns
  networks         = each.value.networks
//[...]
}

// Instance definition
resource "openstack_compute_instance_v2" "instance" {
  count = var.icount

  name        = "${var.env}-${var.type}-${count.index + 1}"
  flavor_name = var.flavor

  block_device {
    uuid                  = openstack_blockstorage_volume_v3.boot[count.index].id
    source_type           = "volume"
    boot_index            = 0
    destination_type      = "volume"
    delete_on_termination = var.volume.delete
  }

  //[...]

  scheduler_hints {
    group = openstack_compute_servergroup_v2.servergroup.id // This bit does not work
  }
}

// The .tfvars file used to declare variables that are use to create
// the whole infrastructure
instances = {
  nginx = {
    image = "rhel-9.1"
    servergroup_name = "nginx-aa"
    // [...]
    flavor = "m4.medium"
    volume = { size = 50, delete = false }
  }
}

+ servergroups = {
+   "nginx-aa" = {
+     policy = "soft-anti-affinity"
+   }
+ }

Just in case, here is the directory architecture of my Terraform

├── 00-locals.tf
├── 01-cloud_init.tf
├── 02-network.tf
├── 03-security_groups.tf
├── 04-infra.tf
├── modules
│   ├── instance
│   │   ├── main.tf
│   │   ├── output.tf
│   │   ├── secgroup
│   │   │   ├── main.tf
│   │   │   ├── variable.tf
│   │   │   └── versions.tf
│   │   ├── variable.tf
│   │   └── versions.tf
│   ├── secgroup
│   │   ├── main.tf
│   │   ├── main.tf.ori
│   │   ├── output.tf
│   │   ├── variable.tf
│   │   └── versions.tf
│   └── servergroup
│       ├── main.tf
│       ├── variables.tf
│       └── versions.tf
├── variable.tf
└── versions.tf

I would like now to add a feature to first create a servergroup and add some instances to this servergroup. How to find the id of the servergroup called "nginx-aa" and use it in the group directive of the scheduler_hints block in the resource openstack_compute_instance_v2 ?

Answer 1

At that time I already wrote a servergroups module, but what I did wrong was the way I summoned the servergroup id. That's what actually worked in the instance module :

module "instances" {
//[...]
servergroup    = each.value.servergroup
servergroup_id = module.servergroups[each.value.servergroup].srvgrp_id 
}

srvgrp_id is an output containing this :

output "srvgrp_id" {
  value = openstack_compute_servergroup_v2.servergroup.id
}