Terraform模块中AKS与节点池关联报错求助

我有一个包含azurerm_kubernetes_cluster和azurerm_kubernetes_cluster_node_pool资源的Terraform模块main.tf，以及存储变量值的dev1.tfvars文件。尝试引用第一个Kubernetes集群的kubernetes_cluster_id时，配置kubernetes_cluster_id = azurerm_kubernetes_cluster.aks.DEV-AKS-11.name出现错误：

A managed resource "azurerm_kubernetes_cluster" "aks" has not been declared in module.AKS.

相关代码如下：

main.tf #

resource "azurerm_kubernetes_cluster" "AKS" {
  for_each            = var.AKS
  name                = each.value.name
  location            = each.value.location
  resource_group_name = each.value.resourceGroupName
  dns_prefix          = each.value.dnsprefix
  node_resource_group = each.value.noderesourcegroup

  default_node_pool {
    name       = each.value.nodename
    node_count = each.value.nodecount
    vm_size    = each.value.vmsize
  }

  identity {
    type = "SystemAssigned"
  }
}

resource "azurerm_kubernetes_cluster_node_pool" "nodepool" {
  for_each              = var.nodepool
  name                  = each.value.nodepoolname
  kubernetes_cluster_id = azurerm_kubernetes_cluster.aks.DEV-AKS-11.id
  vm_size               = each.value.nodepoolvmsize
  node_count            = each.value.nodepoolcount
  enable_auto_scaling   = each.value.nodepoolenableautoscaling
}

dev1.tfvars #

AKS = { 
    "DEV-AKS-11" = {
        name     = "DEV-AKS-11"
        location = "UK South"
        resourceGroupName = "DEV-RG-11"
        dnsprefix = "devaks11"
        nodename = "testnode"
        nodecount = "1"
        vmsize = "Standard_D2_v2"
        noderesourcegroup = "DEV-RG-AKS-11"
  }
}

nodepool = { 
    "agentpool" = {
        name     = "agentpool"
        nodepoolvmsize = "Standard_D2_v2"
        nodepoolcount = "1"
        nodepoolenableautoscaling = "false"
        nodepoolmode = "system"
  }

  "appspool" = {
        name     = "appspool"
        nodepoolvmsize = "Standard_D2_v2"
        nodepoolcount = "1"
        nodepoolenableautoscaling = "false"
        nodepoolmode = "user"
  }

  "minappspool" = {
        name     = "minappspool"
        nodepoolvmsize = "Standard_D2_v2"
        nodepoolcount = "1"
        nodepoolenableautoscaling = "false"
        nodepoolmode = "user"
  }
}

错误原因 #

1. 资源名称大小写不匹配：定义的Kubernetes集群资源是azurerm_kubernetes_cluster.AKS（大写AKS），但引用时用了azurerm_kubernetes_cluster.aks（小写aks），Terraform对资源名称大小写严格敏感，因此找不到对应资源。
2. for_each实例引用方式错误：当资源通过for_each创建时，不能直接用.DEV-AKS-11的方式引用实例，必须通过键值索引["DEV-AKS-11"]访问。

修正代码 #

单集群场景直接修正 #

将azurerm_kubernetes_cluster_node_pool资源中的kubernetes_cluster_id修改为：

kubernetes_cluster_id = azurerm_kubernetes_cluster.AKS["DEV-AKS-11"].id

多集群场景灵活关联 #

如果未来需要扩展多集群，建议在nodepool变量中添加集群关联键，让每个节点池明确绑定对应集群：

1. 修改dev1.tfvars的nodepool变量，增加cluster_key字段：

nodepool = { 
    "agentpool" = {
        name     = "agentpool"
        cluster_key = "DEV-AKS-11"
        nodepoolvmsize = "Standard_D2_v2"
        nodepoolcount = "1"
        nodepoolenableautoscaling = "false"
        nodepoolmode = "system"
  }

  "appspool" = {
        name     = "appspool"
        cluster_key = "DEV-AKS-11"
        nodepoolvmsize = "Standard_D2_v2"
        nodepoolcount = "1"
        nodepoolenableautoscaling = "false"
        nodepoolmode = "user"
  }

  "minappspool" = {
        name     = "minappspool"
        cluster_key = "DEV-AKS-11"
        nodepoolvmsize = "Standard_D2_v2"
        nodepoolcount = "1"
        nodepoolenableautoscaling = "false"
        nodepoolmode = "user"
  }
}

2. 修改main.tf中的节点池资源，通过each.value.cluster_key关联集群：

resource "azurerm_kubernetes_cluster_node_pool" "nodepool" {
  for_each              = var.nodepool
  name                  = each.value.name
  kubernetes_cluster_id = azurerm_kubernetes_cluster.AKS[each.value.cluster_key].id
  vm_size               = each.value.nodepoolvmsize
  node_count            = each.value.nodepoolcount
  enable_auto_scaling   = each.value.nodepoolenableautoscaling
}

此修改既解决当前引用错误，又提升配置扩展性，后续添加新集群或节点池时，只需更新变量文件即可，无需修改资源定义。

内容的提问来源于stack exchange，提问作者Ashish Mishra