azurerm_kubernetes_cluster - unable to increase min_node value when auto scale is enabled

[leesutcliffe]

Community Note

• Please vote on this issue by adding a 👍 reaction to the original issue to help the community and maintainers prioritize this request
• Please do not leave "+1" or "me too" comments, they generate extra noise for issue followers and do not help prioritize the request
• If you are interested in working on this issue or have submitted a pull request, please leave a comment

Terraform (and AzureRM Provider) Version

Terraform v0.12.29

Affected Resource(s)

• azurerm_kubernetes_cluster

Terraform Configuration Files

resource "azurerm_kubernetes_cluster" "example" {
  name                            = "example-aks"
  location                        = azurerm_resource_group.example.location
  resource_group_name             = azurerm_resource_group.example.name
  kubernetes_version              = "1.16.13"
  node_resource_group             = "example-rg"
  dns_prefix                      = "aks-example"
  enable_pod_security_policy      = false
  private_cluster_enabled         = false
  api_server_authorized_ip_ranges = null
 
  default_node_pool {
    name                = "default"
    enable_auto_scaling = true
    min_count           = 3
    max_count           = 8
    vm_size             = "Standard_D2_v2"
    os_disk_size_gb     = 30
    vnet_subnet_id      = "/subscriptions/*/resourceGroups/example-rg/providers/Microsoft.Network/virtualNetworks/example-vnet/subnets/example-snet-k8s"
    max_pods            = 60
    type                = "VirtualMachineScaleSets"
    availability_zones  = ["1", "2", "3"]
    orchestrator_version = "1.16.13"
  }

Expected Behavior

At the time of testing AKS had 2 nodes in the default node pool, intention was to increase the minimum node_count to 3

After changing the value for min_nodes from 2 to 3, the expected result would be for a Terraform plan to be applied successfully.
AKS would auto-scale from a minimum of 2 nodes, to 3.
It is noted that node_count is optional when enable_auto_scaling is set to true

Actual Behavior

Terraform apply step fails with the following output

Error: expanding `default_node_pool`: `node_count`(2) must be equal to or greater than `min_count`(3) when `enable_auto_scaling` is set to `true`

node_count is not set in the Terraform configuration but is present in the state file.
Removing node_count from the state file yields the same result, moreover, "node_count": 2, is returned to the state file after a terraform apply

Steps to Reproduce

1. AKS Cluster running with two nodes with auto scale configured
2. Change min_nodes = 2 to min_nodes = 3
3. terraform plan
4. terraform apply

Important Factoids

References

https://www.terraform.io/docs/providers/azurerm/r/kubernetes_cluster.html

[neil-yechenwei]

Thanks for opening this issue. Per the document, node_count would be set with min_count when node_count isn't set and enable_auto_scaling is true. So I assume it's expect behavior. I assume you have to explicitly set node_count and it has to be equal or greater than min_count when min_count is updated.

[leesutcliffe]

Thanks for the reply @neil-yechenwei
I cannot see anywhere in the page you reference where it states "node_count would be set with min_count when node_count isn't set and enable_auto_scaling is true"

That said I have attempted to explicitly set node_count to be equal or greater than min_count

The following error is from a cluster that currently has 3 nodes and attempting to set min_count = 4 and node_count is not set

Error: expanding `default_node_pool`: `node_count`(3) must be equal to or greater than `min_count`(4) when `enable_auto_scaling` is set to `true`

The following error is from another attempt to set min_count = 4, this time explicitly setting node_count = 4

Error: expanding `default_node_pool`: cannot change `node_count` when `enable_auto_scaling` is set to `true`

[neil-yechenwei]

I assume the words "node_count - (Optional) The initial number of nodes which should exist in this Node Pool. If specified this must be between 1 and 100 and between min_count and max_count." in document already indicates node_count would be initialized.

I think the error "Error: expanding default_node_pool: cannot change node_count when enable_auto_scaling is set to true" is expected error, which means you cannot update node_count when enable_auto_scaling is set to true.

For more usage problem, please raise question on HashiCorp Community Forums. Thanks.

[leesutcliffe]

I am aware of the expected error: "Error: expanding default_node_pool: cannot change node_count when enable_auto_scaling is set to true" - this was just for demonstration purposes, given your comment previous comment.

But the original issue still stands - being unable to increase min_count, even when node_count is not set (although it does appear in the state file)

[neil-yechenwei]

Yes. I assume you're right. I updating the code for comparing min_count with node_count when enable_auto_scaling is true because node_count at this time is useless while updating. So I submit a PR to fix it. Thanks.

[ghost]

This has been released in version 2.34.0 of the provider. Please see the Terraform documentation on provider versioning or reach out if you need any assistance upgrading. As an example:

provider "azurerm" {
    version = "~> 2.34.0"
}
# ... other configuration ...

[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!