Azure AKS Terraform - how to specify VM Size - terraform

Question

In my Azure subscription, I'm trying to create an AKS cluster using Terraform.
My main.tf looks like this:
## Azure resource provider ##
provider "azurerm" {
version = "=1.36.1"
}
## Azure resource group for the kubernetes cluster ##
resource "azurerm_resource_group" "aks_demo" {
name = var.resource_group_name
location = var.location
}
## AKS kubernetes cluster ##
resource "azurerm_kubernetes_cluster" "aks_demo" {
name = var.cluster_name
resource_group_name = azurerm_resource_group.aks_demo.name
location = azurerm_resource_group.aks_demo.location
dns_prefix = var.dns_prefix
linux_profile {
admin_username = var.admin_username
## SSH key is generated using "tls_private_key" resource
ssh_key {
key_data = "${trimspace(tls_private_key.key.public_key_openssh)} ${var.admin_username}#azure.com"
}
}
agent_pool_profile {
name = "default"
count = var.agent_count
vm_size = "Standard_D2"
os_type = "Linux"
os_disk_size_gb = 30
}
service_principal {
client_id = var.client_id
client_secret = var.client_secret
}
tags = {
Environment = "Production"
}
}
## Private key for the kubernetes cluster ##
resource "tls_private_key" "key" {
algorithm = "RSA"
}
## Save the private key in the local workspace ##
resource "null_resource" "save-key" {
triggers = {
key = tls_private_key.key.private_key_pem
}
provisioner "local-exec" {
command = <<EOF
mkdir -p ${path.module}/.ssh
echo "${tls_private_key.key.private_key_pem}" > ${path.module}/.ssh/id_rsa
chmod 0600 ${path.module}/.ssh/id_rsa
EOF
}
}
## Outputs ##
# Example attributes available for output
output "id" {
value = "${azurerm_kubernetes_cluster.aks_demo.id}"
}
output "client_key" {
value = "${azurerm_kubernetes_cluster.aks_demo.kube_config.0.client_key}"
}
output "client_certificate" {
value = "${azurerm_kubernetes_cluster.aks_demo.kube_config.0.client_certificate}"
}
output "cluster_ca_certificate" {
value = "${azurerm_kubernetes_cluster.aks_demo.kube_config.0.cluster_ca_certificate}"
}
output "kube_config" {
value = azurerm_kubernetes_cluster.aks_demo.kube_config_raw
}
output "host" {
value = azurerm_kubernetes_cluster.aks_demo.kube_config.0.host
}
output "configure" {
value = <<CONFIGURE
Run the following commands to configure kubernetes client:
$ terraform output kube_config > ~/.kube/aksconfig
$ export KUBECONFIG=~/.kube/aksconfig
Test configuration using kubectl
$ kubectl get nodes
CONFIGURE
}
My variables.tf looks like this:
## Azure config variables ##
variable "client_id" {}
variable "client_secret" {}
variable location {
default = "Central US"
}
## Resource group variables ##
variable resource_group_name {
default = "aksdemo-rg"
}
## AKS kubernetes cluster variables ##
variable cluster_name {
default = "aksdemo1"
}
variable "vm_size" {
default = "Standard_A0"
}
variable "agent_count" {
default = 3
}
variable "dns_prefix" {
default = "aksdemo"
}
variable "admin_username" {
default = "demo"
}
When I run terraform apply, I get this error:
Error: Error creating Managed Kubernetes Cluster "aksdemo1" (Resource Group "aksdemo-rg"):
containerservice.ManagedClustersClient#CreateOrUpdate: Failure sending request: StatusCode=400 --
Original Error: Code="BadRequest"
Message="The VM size of AgentPoolProfile:default is not allowed in your subscription in location 'centralus'. The available VM sizes are Standard_A2,Standard_A2_v2,Standard_A2m_v2,Standard_A3,Standard_A4,Standard_A4_v2,Standard_A4m_v2,
Standard_A5,Standard_A6,Standard_A7,Standard_A8_v2,Standard_A8m_v2,Standard_B12ms,Standard_B16ms,Standard_B20ms,Standard_B2ms,Standard_B2s,Standard_B4ms,Standard_B8ms,Standard_D11_v2,Standard_D12_v2,
Standard_D13_v2,Standard_D14_v2,Standard_D15_v2,Standard_D16_v3,Standard_D16s_v3,Standard_D1_v2,Standard_D2_v2,Standard_D2_v3,Standard_D2s_v3,Standard_D32_v3,Standard_D32s_v3,Standard_D3_v2,Standard_D48_v3,
Standard_D48s_v3,Standard_D4_v2,Standard_D4_v3,Standard_D4s_v3,Standard_D5_v2,Standard_D64_v3,Standard_D64s_v3,Standard_D8_v3,Standard_D8s_v3,Standard_DS1,Standard_DS11,Standard_DS11_v2,Standard_DS12,Standard_DS12_v2,Standard_DS13,Standard_DS13-2_v2,Standard_DS13-4_v2,Standard_DS13_v2,Standard_DS14,Standard_DS14-4_v2,Standard_DS14-8_v2,Standard_DS14_v2,Standard_DS15_v2,Standard_DS1_v2,Standard_DS2,Standard_DS2_v2,Standard_DS3,Standard_DS3_v2,Standard_DS4,Standard_DS4_v2,Standard_DS5_v2,Standard_E16_v3,Standard_E16s_v3,Standard_E2_v3,Standard_E2s_v3,Standard_E32-16s_v3,Standard_E32-8s_v3,Standard_E32_v3,Standard_E32s_v3,Standard_E48_v3,Standard_E48s_v3,Standard_E4_v3,Standard_E4s_v3,Standard_E64-16s_v3,Standard_E64-32s_v3,Standard_E64_v3,Standard_E64i_v3,Standard_E64is_v3,Standard_E64s_v3,Standard_E8_v3,Standard_E8s_v3,Standard_F16,Standard_F16s,Standard_F16s_v2,Standard_F2,Standard_F2s,Standard_F2s_v2,Standard_F32s_v2,Standard_F4,Standard_F48s_v2,Standard_F4s,Standard_F4s_v2,Standard_F64s_v2,Standard_F72s_v2,Standard_F8,
Standard_F8s,Standard_F8s_v2
For more details, please visit https://aka.ms/cpu-quota"
This is confusing to me, as there is clearly a variable named vm_size
What can I change in order for this to work?

Answer 1

As I see from the code you provided and the error you got, you made the mistake in the code.
What the code you made:
agent_pool_profile {
name = "default"
count = var.agent_count
vm_size = "Standard_D2"
os_type = "Linux"
os_disk_size_gb = 30
}
It should be like this when you use the variable for the VM size:
agent_pool_profile {
name = "default"
count = var.agent_count
vm_size = var.vm_size
os_type = "Linux"
os_disk_size_gb = 30
}
And the VM size should be an appropriate one on yourself and for the requirements. For example, just like it shows in the Terraform example.

Answer 2

The error message is telling you that you're trying to use a VM size, or VM type if you will that's not available for your subscription in that location, it's also giving you all the VM sizes you can choose from.
Note that you have probably copy pasted this:
agent_pool_profile {
name = "default"
count = var.agent_count
vm_size = "Standard_D2"
os_type = "Linux"
os_disk_size_gb = 30
}
The VM size is hardcoded there, so you're default Standard_A0value is not being picked up.
You have more than one way to debug here, first I'd start by making sure the right value is being used, and second change the VM type to see if that works.