antony@notes:~/ops-automation$ cat "Deploying-GKE-Clusters-with-Terraform-from-scratch.md"

Deploying GKE Clusters with Terraform from scratch

2025-06-06· ops-automation

Deploying GKE Clusters with Terraform from scratch

1. 先決條件

建立 Google Mail 帳號
申請免費試用 Google Cloud Platform 資格 (注意，需要一張 VISA 或信用卡) https://console.cloud.google.com/getting-started
在 Google Cloud Platform 建立 project
- 點選左上角 “My First Project” 專案的按鈕 :::spoiler 點我展開圖片 :::
- 點選右上角 “新增專案” 按鈕 :::spoiler 點我展開圖片 :::
- 輸入 testgke，點選 “建立” 按鈕 :::spoiler 點我展開圖片 :::
確認在剛建立的 gketest 專案後，點選左上角按鈕 :::spoiler 點我展開圖片 :::
點選 Kubernetes Engine 按鈕 :::spoiler 點我展開圖片 :::
點選 “啟用” 按鈕，啟用 Kubernetes Engine API :::spoiler 點我展開圖片 :::
點選右上角 “Cloud Shell” 按鈕，並授權 Cloud Shell :::spoiler 點我展開圖片 :::

2. 開始部署

$ mkdir gcp-terraform; cd gcp-terraform

$ nano gke.tf
resource "google_container_cluster" "gke_cluster" {
  name               = "k1"
  location           = "asia-east1-b"
  initial_node_count = 1

  network    = google_compute_network.vpc.name
  subnetwork = google_compute_subnetwork.subnet.name  
  networking_mode = "VPC_NATIVE"

  logging_service = "none"
  monitoring_service = "none"

  private_cluster_config {
    enable_private_endpoint = false
    enable_private_nodes    = true
    master_ipv4_cidr_block  = "10.13.0.0/28"
  }
  ip_allocation_policy {
    cluster_ipv4_cidr_block  = "10.98.0.0/16"
    services_ipv4_cidr_block = "10.244.0.0/16"
  }
  node_config {
    machine_type = "e2-custom-4-12288" 
  }
}

output "cluster_endpoint" {
  value = google_container_cluster.gke_cluster.endpoint
}

output "cluster_ca_certificate" {
  value     = google_container_cluster.gke_cluster.master_auth.0.cluster_ca_certificate
  sensitive = true
}

$ nano variable.tf
#PROJECT INFO
variable "PROJECT_ID" {
  default = "gketest-462015"
}

variable "REGION" {
  default = "asia-east1"
}

#VPC
variable "VPC_NAME" {
  default = "k8s-vpc"
}

variable "VPC_NAME2" {
  default = "k8s-vpc2"
}

variable "VPC_SUBNET_NAME" {
  default = "asia-east1"
}

variable "VPC_SUBNET_NAME2" {
  default = "asia-east1-2"
}

variable "IP_CIDR_RANGE" {
  default = "10.10.0.0/24"
}

variable "IP_CIDR_RANGE2" {
  default = "10.10.1.0/24"
}

variable "ROUTER_NAME" {
  default = "demo-route"
}

variable "ROUTER_NAME2" {
  default = "demo-route2"
}

variable "NAT_NAME" {
  default = "demo-nat"
}

variable "NAT_NAME2" {
  default = "demo-nat2"
}

#GKE

variable "GKE_LOCATION" {
  default = "asia-east1-b"
}

variable "GKE_MACHINE_TYPE" {
  default = "e2-custom-2-8192"
}

$ nano vpc.tf


resource "google_compute_network" "vpc" {
  name                    = var.VPC_NAME
  auto_create_subnetworks = "false"
}


resource "google_compute_subnetwork" "subnet" {
  name          = "${var.VPC_SUBNET_NAME}-subnet"
  region        = var.REGION
  network       = google_compute_network.vpc.name
  ip_cidr_range = var.IP_CIDR_RANGE
  private_ip_google_access   = true
}

#resource "google_compute_subnetwork" "subnet2" {
#  name          = "${var.VPC_SUBNET_NAME}2-subnet"
#  region        = var.REGION
#  network       = google_compute_network.vpc.name
#  ip_cidr_range = var.IP_CIDR_RANGE2
#  private_ip_google_access   = true
#}


resource "google_compute_router" "router" {
  name    = var.ROUTER_NAME
  region  = google_compute_subnetwork.subnet.region
  network = google_compute_network.vpc.id
}

#resource "google_compute_address" "address" {
#  count  = 2
#  name   = "nat-manual-ip-${count.index}"
#  region = google_compute_subnetwork.subnet.region
#}


resource "google_compute_address" "r1-address" {
  name   = "nat-address"
  region = google_compute_subnetwork.subnet.region
}

#resource "google_compute_address" "r1-address-2" {
#  name   = "nat-address2"
#  region = google_compute_subnetwork.subnet.region
#}


resource "google_compute_router_nat" "nat_manual" {
  name   = var.NAT_NAME
  router = google_compute_router.router.name
  region = google_compute_router.router.region

  nat_ip_allocate_option = "MANUAL_ONLY"
  nat_ips                = [google_compute_address.r1-address.self_link]
  enable_dynamic_port_allocation      = true
  enable_endpoint_independent_mapping = false
  min_ports_per_vm                    = 4096
  max_ports_per_vm                    = 65536     
  source_subnetwork_ip_ranges_to_nat = "LIST_OF_SUBNETWORKS"
  subnetwork {
    name                    = google_compute_subnetwork.subnet.id
    source_ip_ranges_to_nat = ["ALL_IP_RANGES"]
  }
}


#resource "google_compute_router" "router2" {
#  name    = var.ROUTER_NAME2
#  region  = google_compute_subnetwork.subnet2.region
#  network = google_compute_network.vpc.id
#}

#resource "google_compute_address" "r2-address" {
#  name   = "r2-nat-address"
#  region = google_compute_subnetwork.subnet2.region
#}

#resource "google_compute_router_nat" "nat_manual2" {
#  name   = var.NAT_NAME2
#  router = google_compute_router.router2.name
#  region = google_compute_router.router2.region

#  nat_ip_allocate_option = "MANUAL_ONLY"
#  nat_ips                = google_compute_address.r2-address.*.self_link
#  enable_dynamic_port_allocation      = true
#  enable_endpoint_independent_mapping = false
#  min_ports_per_vm                    = 4096
#  max_ports_per_vm                    = 65536 
#  source_subnetwork_ip_ranges_to_nat = "LIST_OF_SUBNETWORKS"
#  subnetwork {
#    name                    = google_compute_subnetwork.subnet2.id
#    source_ip_ranges_to_nat = ["ALL_IP_RANGES"]
#  }
#}

#Firewall Rule
#resource "google_compute_firewall" "datalake-prod-composer-internal-ingress" {
#  name    = "datalake-prod-composer-internal-ingress"
#  network = google_compute_network.prod-vpc.name

#  allow {
#    protocol = "tcp"
#  }
#  allow {
#    protocol = "udp"
#  }
#  priority = 1000
#  source_ranges = var.datalake_prod_composer_ingress_allow_source_ip
#  target_tags = var.datalake_prod_composer_ingress_target_tag
#}

$ nano provider.tf
terraform {
  required_providers {
    google = {
      source  = "hashicorp/google"
      version = "4.40.0"
    }
  }

  required_version = ">= 0.14"
}

provider "google" {
  project = var.PROJECT_ID
  region  = var.REGION
}

$ terraform init

$ terraform plan

$ terraform apply

$ gcloud container clusters get-credentials k1 --zone asia-east1-b --project $(gcloud config get-value project 2> /dev/null)

$ kubectl get nodes
NAME                                STATUS   ROLES    AGE     VERSION
gke-k1-default-pool-e0781567-7ljm   Ready    <none>   5m39s   v1.32.4-gke.1106006

3. 清理環境 (窮到只剩錢的人可跳過此步驟)

不做的話，殘留在 Google Cloud 上的物件會一直產生費用

# 將所有建立的物件清除
$ terraform destroy

# 移除 billing account 與 Project 的連結
$ gcloud billing projects unlink $GOOGLE_CLOUD_PROJECT
billingAccountName: ''
billingEnabled: false
name: projects/gketest-462015/billingInfo
projectId: gketest-462015

# 刪除 project
$ gcloud projects delete -q $GOOGLE_CLOUD_PROJECT
Deleted [https://cloudresourcemanager.googleapis.com/v1/projects/gketest-462015].

You can undo this operation for a limited period by running the command below.
    $ gcloud projects undelete gketest-462015

See https://cloud.google.com/resource-manager/docs/creating-managing-projects for information on shutting down projects.

# 確認 project 已被刪除
$ gcloud projects describe $GOOGLE_CLOUD_PROJECT
createTime: '2025-06-05T15:39:13.542189Z'
lifecycleState: DELETE_REQUESTED
name: gketest
projectId: gketest-462015
projectNumber: '876476350521'

Bash Script Debug mode 設定 Alertmanager 觸發告警時去執行 Bash Script