Error 'no such host' when adding a volume to Terraform AKS cluster on Azure

Question

I'm using the below Terraform file to create an AKS cluster:

resource "random_pet" "prefix" {}

resource "kubernetes_persistent_volume" "example" {
  metadata {
    name = "example"
  }
  spec {
    capacity = {
      storage = "1Gi"
    }
    access_modes = ["ReadWriteOnce"]
    persistent_volume_source {
      azure_disk {
        caching_mode  = "None"
        data_disk_uri = azurerm_managed_disk.example.id
        disk_name     = "example"
        kind          = "Managed"
      }
    }
  }
}

resource "azurerm_kubernetes_cluster" "example" {
  name                = "${random_pet.prefix.id}-aks"
  location            = azurerm_resource_group.example.location
  resource_group_name = azurerm_resource_group.example.name
  dns_prefix          = "${random_pet.prefix.id}-k8s"

  default_node_pool {
    name            = "example"
    node_count      = 2
    vm_size         = "Standard_D2_v2"
    os_disk_size_gb = 30
  }

  identity {
    type = "SystemAssigned"
  }

  role_based_access_control {
    enabled = true
  }

  addon_profile {
    kube_dashboard {
      enabled = true
    }
  }

  tags = {
    environment = "Demo"
  }
}

provider "azurerm" {
  version = ">=2.20.0"
  features {}
}

resource "azurerm_resource_group" "example" {
  name     = "${random_pet.prefix.id}-rg"
  location = "westus2"
}


resource "azurerm_managed_disk" "example" {
  name                 = "example"
  location             = azurerm_resource_group.example.location
  resource_group_name  = azurerm_resource_group.example.name
  storage_account_type = "Standard_LRS"
  create_option        = "Empty"
  disk_size_gb         = "1"
  tags = {
    environment = azurerm_resource_group.example.name
  }
}

I've derived the above file from Terraform's tutorial on setting up an AKS cluster: https://learn.hashicorp.com/tutorials/terraform/aks

And I've used Terraform's example of setting up an Azure managed disk and k8s volume here: https://registry.terraform.io/providers/hashicorp/kubernetes/latest/docs/resources/persistent_volume

When I try to run the above config with Terraform I get the following error:

Error: Post "https://pumped-llama-k8s-419df981.hcp.westus2.azmk8s.io:443/api/v1/persistentvolumes": dial tcp: lookup pumped-llama-k8s-419df981.hcp.westus2.azmk8s.io on 192.168.1.1:53: no such host

  on main.tf line 3, in resource "kubernetes_persistent_volume" "example":
   3: resource "kubernetes_persistent_volume" "example" {

I get this same error whenever I attempt to use any non-azurerm Terraform resource. Eg. when trying to configure roles and role binding: resource "kubernetes_role"

I'm assuming by the error message's url - trying to connect to HashiCorp Cloud Platform, which I assume is the default - that I need to explicitly tell these non-azurerm resources that I am connecting to an Azure hosted Kubernetes; however I can't figure out how to do that.

It seems it cannot connect to the AKS from your local machine, how about the output of the command kubectl config view or kubectl get nodes? — Charles Xu
Sounds more like a DNS name resolution issue.. is nslookup working? — harshavmb
I don't think it's an issue with connectivity to AKS, as the remainder of the Terraform resources are created; I can go to the AKS cluster on Azure, and it's all there and working. It's just any Terraform resources that are kubernetes specific like 'kubernetes_persistent_volume" or "kubernetes_role" that seem to fail. But it looks like it's failing because Terraform is attempting to connect to a non-extant HashiCorp Cloud server rather than my AKS instance. — Sam Carswell

Sam Carswell Sam Carswell · Accepted Answer · 2020-11-30T03:43:55

Turns out I needed to define the kubernetes provider in the Terraform file - surprised I wouldn't get some kind of a warning for not including it, considering I'm interacting with it's resources.

Here's what I did to fix it:

outputs.tf:

output "host" {
  value = azurerm_kubernetes_cluster.default.kube_config.0.host
}

output "client_key" {
  value = azurerm_kubernetes_cluster.default.kube_config.0.client_key
}

output "client_certificate" {
  value = azurerm_kubernetes_cluster.default.kube_config.0.client_certificate
}

output "kube_config" {
  value = azurerm_kubernetes_cluster.default.kube_config_raw
}

output "cluster_ca_certificate" {
  value = azurerm_kubernetes_cluster.default.kube_config.0.cluster_ca_certificate
}

main.tf:

...
provider "kubernetes" {
  version = "=1.13.2"
  load_config_file = "false"

  host = azurerm_kubernetes_cluster.default.kube_config.0.host
  
  client_certificate     = "${base64decode(azurerm_kubernetes_cluster.default.kube_config.0.client_certificate)}"
  client_key             = "${base64decode(azurerm_kubernetes_cluster.default.kube_config.0.client_key)}"
  cluster_ca_certificate = "${base64decode(azurerm_kubernetes_cluster.default.kube_config.0.cluster_ca_certificate)}"
}
...

Error 'no such host' when adding a volume to Terraform AKS cluster on Azure

1 Answers