Terraform:迭代和跳过特定的索引值

Terraform: Iterating and skipping over specific index values

我正在尝试更新我现有的 azurerm_subnet terraform 代码 被许多其他人消耗掉了。所以我们真的不想破坏任何现有的 tfvars

使用 network_security_group_id 的折旧字段从 azurerm_subnet 到名为 azurerm_network_interface_security_group_association 的新资源,从版本 2.0.

开始

现有代码

resource "azurerm_subnet" "generic_subnet" {
  count = "${length(var.subnet_names)}"

  name = "${element("${var.subnet_names}", count.index)}"
  resource_group_name = "${var.resource_group_name}"
  virtual_network_name = "${azurerm_virtual_network.generic_vnet.name}"
  address_prefix = "${element("${var.subnet_address_ranges}", count.index)}"
  service_endpoints = "${var.service_endpoints[count.index]}"


  network_security_group_id = "${element(var.subnet_nsg_ids, count.index)}"
  route_table_id = "${element(var.subnet_route_tables, count.index)}"
}

tfvars:


subnet_address_ranges = [
 "10.102.40.0/22",
 "10.102.44.0/24", 
 "10.102.45.0/25"
]
subnet_names = [
 "private-subnet-01",
 "public-subnet-01", 
 "protected-subnet-01"
 ]
service_endpoints = [    
    ["Microsoft.AzureCosmosDB","Microsoft.KeyVault", "Microsoft.Storage","Microsoft.Sql","Microsoft.AzureActiveDirectory","Microsoft.ContainerRegistry","Microsoft.EventHub","Microsoft.ServiceBus","Microsoft.Web"],
    [],
    ["Microsoft.KeyVault", "Microsoft.Storage"]
]
subnet_nsg_ids = [
    "/subscriptions/yyyy/resourceGroups/yyy/providers/Microsoft.Network/networkSecurityGroups/nsg-01",
    "",
"/subscriptions/yyyy/resourceGroups/yyy/providers/Microsoft.Network/networkSecurityGroups/nsg-02" 
]

更新代码(通过引入 azurerm_network_interface_security_group_association 资源)

resource "azurerm_subnet" "generic_subnet" {
  count = "${length(var.subnet_names)}"

  name = "${element("${var.subnet_names}", count.index)}"
  resource_group_name = "${var.resource_group_name}"
  virtual_network_name = "${azurerm_virtual_network.generic_vnet.name}"
  address_prefix = "${element("${var.subnet_address_ranges}", count.index)}"
  service_endpoints = "${var.service_endpoints[count.index]}"


  network_security_group_id = "${element(var.subnet_nsg_ids, count.index)}"
  route_table_id = "${element(var.subnet_route_tables, count.index)}"
}

resource "azurerm_subnet_network_security_group_association" "generic_nsg_association" {
  count= "${length(var.subnet_nsg_ids)}" 
  subnet_id = "${element(azurerm_subnet.generic_subnet.*.id, count.index)}"
  network_security_group_id = "${element(var.subnet_nsg_ids, count.index)}"
}

显然我的新资源会通过抛出像

这样的错误来破坏所有 tf​​vars

Error: Can not parse "network_security_group_id" as a resource id: Cannot parse Azure ID: parse "": empty url

on ../../../main.tf line 41, in resource "azurerm_subnet_network_security_group_association" "generic_nsg_association": 41: network_security_group_id = "${element(var.subnet_nsg_ids, count.index)}"

原因是我们需要给出 network_security_group_id 参数。但是,我所有现有的 tfvar 都有一个结构支持 subnet_nsg_ids 列表,其中包含一个空字符串。

所以我的问题是有没有办法只循环访问我的 azurerm_subnet_network_security_group_association 资源列表中的特定索引。例如,仅循环遍历 [0] 和 [2],但跳过 [1](因为 [1] 是一个空字符串

它不像我只需要使用 count。如果可以进行这种跳过,我也很乐意使用 for_each

我不确定我是否完全理解您的问题,但是如果您想在有空字符串时跳过 network_security_group_id,您可以使用:

network_security_group_id = element(var.subnet_nsg_ids, count.index) != "" ? element(var.subnet_nsg_ids, count.index) : null

如果您想过滤掉包含空字符串的项目,您可以这样做:

for_each = [for idx, val in var.subnet_names): val if var.subnet_nsg_ids[idx] != ""]