使用 terraform 和 count 部署 gcp 实例（使用 attach_disk）答案

【问题标题】：deploy gcp instance using terraform and count (with attached_disk)使用 terraform 和 count 部署 gcp 实例（使用 attach_disk）
【发布时间】：2018-06-01 15:49:45
【问题描述】：

环境：

terraform v0.10.7
谷歌云平台
用于创建后端、变量等的各种 .tf 文件

问题：

我能够创建多个 vm 实例以及多个附加磁盘（boot_disk 在每个实例上都可以正常工作）但我希望能够将这些附加磁盘相应地附加到每个 vm 而不必为每个 vm 单独添加（如果这是有道理的！）。

到目前为止，我的代码是（它适用于构建多个计算实例和多个额外的磁盘）：注意（我已经注释掉了attached_disk which errors atm）

# vm1.tf

variable "node_count" {
  default = "3"
 }

resource "google_compute_disk" "test-node-1-index-disk-" {
    count   = "${var.node_count}"
    name    = "test-node-1-index-disk-${count.index}-data"
    type    = "pd-standard"
    zone    = "${var.zone}"
    size    = "5"
}
resource "google_compute_instance" "test-node-" {
    count = "${var.node_count}"
    name = "test-node-${count.index}"
    machine_type = "${var.machine_type}"
    zone = "${var.zone}"

    boot_disk {
    initialize_params {
    image = "${var.image}"
    }
   }
#    attached_disk {
#        source      = "${google_compute_disk.test-node-1-index-disk-0}"
#        device_name = "${google_compute_disk.test-node-1-index-disk-0}"
#   }


    network_interface {
      network = "default"
      access_config {
        // Ephemeral IP
      }

    }
}

如果我做单独的 .tf，attached_disk 工作没问题。

我想要的最终状态，is to be able to build multiple vm's, multiple additional disks using count and attach/assign each added disk to each vm instance with a relationship of 1:1 but preferable within a single .tf and block...

我想，我可以考虑应用一个 post gcloud compute 命令来附加（知道预期的命名约定），但我希望它更加动态并在创建时完成。

我是不是搞错了？非常感谢任何帮助/指针！

谢谢布里

【问题讨论】：

标签： terraform gcp

【解决方案1】：

# vm1.tf

variable "node_count" {
  default = "3"
 }

resource "google_compute_disk" "test-node-1-index-disk-" {
    count   = "${var.node_count}"
    name    = "test-node-1-index-disk-${count.index}-data"
    type    = "pd-standard"
    zone    = "${var.zone}"
    size    = "5"
}
resource "google_compute_instance" "test-node-" {
    count = "${var.node_count}"
    name = "test-node-${count.index}"
    machine_type = "${var.machine_type}"
    zone = "${var.zone}"

    boot_disk {
    initialize_params {
    image = "${var.image}"
    }
   }
    attached_disk {
        source      = "${element(google_compute_disk.test-node-1-index-disk-.*.self_link, count.index)}"
        device_name = "${element(google_compute_disk.test-node-1-index-disk-.*.name, count.index)}"
   }


    network_interface {
      network = "default"
      access_config {
        // Ephemeral IP
      }

    }
}

【讨论】：

【解决方案2】：

如果你想要静态 ip

variable "node_count" {
 default = "3"
}

resource "google_compute_address" "static-ip-address" {
  count = "${var.node_count}"
  name = "${var.tag}-static-ip-${count.index + 1}"
}

resource "google_compute_disk" "test-node-1-index-disk-" {
  count   = "${var.node_count}"
  name    = "test-node-1-index-disk-${count.index}-data"
  type    = "pd-standard"
  zone    = "${var.zone}"
  size    = "5"
}
resource "google_compute_instance" "test-node-" {
  count = "${var.node_count}"
  name = "test-node-${count.index}"
  machine_type = "${var.machine_type}"
  zone = "${var.zone}"

boot_disk {
  initialize_params {
  image = "${var.image}"
  }
}
attached_disk {
    source      = "${element(google_compute_disk.test-node-1-index-disk-.*.self_link, count.index)}"
    device_name = "${element(google_compute_disk.test-node-1-index-disk-.*.name, count.index)}"
}


network_interface {
  network = "default"
  access_config {
    nat_ip = "${element(google_compute_address.static-ip-address.*.address, count.index)}"
  }

}

【讨论】：