云原生之旅 - 4）基础设施即代码 使用 Terraform 创建 Kubernetes

前言

上一篇文章我们已经简单的入门Terraform， 本篇介绍如何使用Terraform在GCP和AWS 创建Kubernetes 资源。

Kubernetes 在云原生时代的重要性不言而喻，等于这个时代的操作系统，基本上只需要建这个资源，就可以将绝大多数的应用跑在上面，包括数据库，甚至很多团队的大数据处理例如 Spark, Flink 都跑在Kubernetes上。

• GCP Kubernetes = GKE
• AWS Kubernetes = EKS
• Azure Kubernetes = AKS

本篇文章主要介绍前两者的Terraform 代码实现，现在使用官方的 module 要比以前方便太多了，哪怕是新手都可以很快的将资源建起来，当然如果要更多的了解，还是需要慢慢下功夫的。

关键词：IaC, Infrastructure as Code, Terraform, 基础设施即代码，使用Terraform创建GKE，使用Terraform创建EKS

 

环境信息:

* Terraform 1.2.9

* Google Cloud SDK 397.0.0

* aws-cli 2.7.7

 

使用Terraform创建GKE

准备一个GCS bucket

# valid LOCATION values are `asia`, `eu` or `us`
gsutil mb -l $LOCATION gs://$BUCKET_NAME
gsutil versioning set on gs://$BUCKET_NAME

 

准备如下tf文件

backend.tf

terraform {
  backend "gcs" {
    bucket = "sre-dev-terraform-test"
    prefix = "demo/state"
  }
}

providers.tf

terraform {
  required_version = ">= 1.2.9"

  required_providers {
    google = {
      source  = "hashicorp/google"
      version = "~> 4.0"
    }
    google-beta = {
      source  = "hashicorp/google-beta"
      version = "~> 4.0"
    }
  }
}

provider "google" {
  project = local.project.project_id
  region  = local.project.region
}

provider "google-beta" {
  project = local.project.project_id
  region  = local.project.region
}

　　

使用 terraform google module 事半功倍，代码如下

gke-cluster.tf 

data "google_compute_zones" "available" {
  region = "us-central1"
  status = "UP"
}

resource "google_compute_network" "default" {
  project                 = local.project.project_id
  name                    = local.project.network_name
  auto_create_subnetworks = false
  routing_mode            = "GLOBAL"
}

resource "google_compute_subnetwork" "wade-gke" {
  project       = local.project.project_id
  network       = google_compute_network.default.name
  name          = local.wade_cluster.subnet_name
  ip_cidr_range = local.wade_cluster.subnet_range
  region        = local.wade_cluster.region

  secondary_ip_range {
    range_name    = format("%s-secondary1", local.wade_cluster.cluster_name)
    ip_cidr_range = local.wade_cluster.secondary_ip_range_pods
  }

  secondary_ip_range {
    range_name    = format("%s-secondary2", local.wade_cluster.cluster_name)
    ip_cidr_range = local.wade_cluster.secondary_ip_range_services
  }

  private_ip_google_access = true

}

resource "google_service_account" "sa-wade-test" {
  account_id   = "sa-wade-test"
  display_name = "sa-wade-test"
}


module "wade-gke" {
  source = "terraform-google-modules/kubernetes-engine/google//modules/beta-private-cluster"
  version = "23.1.0"

  project_id = local.project.project_id
  name       = local.wade_cluster.cluster_name

  kubernetes_version     = local.wade_cluster.cluster_version
  region                 = local.wade_cluster.region
  network                = google_compute_network.default.name
  subnetwork             = google_compute_subnetwork.wade-gke.name
  master_ipv4_cidr_block = "10.1.0.0/28"
  ip_range_pods          = google_compute_subnetwork.wade-gke.secondary_ip_range.0.range_name
  ip_range_services      = google_compute_subnetwork.wade-gke.secondary_ip_range.1.range_name

  service_account                 = google_service_account.sa-wade-test.email
  master_authorized_networks      = local.wade_cluster.master_authorized_networks
  master_global_access_enabled    = false
  istio                           = false
  issue_client_certificate        = false
  enable_private_endpoint         = false
  enable_private_nodes            = true
  remove_default_node_pool        = true
  enable_shielded_nodes           = false
  identity_namespace              = "enabled"
  node_metadata                   = "GKE_METADATA"
  horizontal_pod_autoscaling      = true
  enable_vertical_pod_autoscaling = false

  node_pools              = local.wade_cluster.node_pools
  node_pools_oauth_scopes = local.wade_cluster.oauth_scopes
  node_pools_labels       = local.wade_cluster.node_pools_labels
  node_pools_metadata     = local.wade_cluster.node_pools_metadata
  node_pools_taints       = local.wade_cluster.node_pools_taints
  node_pools_tags         = local.wade_cluster.node_pools_tags

}

　　

变量 locals.tf

master_authorized_networks 需要改为自己要放行的白名单，只有白名单的IP才能访问 cluster api endpoint。为了安全性，不要用0.0.0.0/0

locals {
  # project details
  project = {
    project_id       = "sre-eng-cn-dev"
    region           = "us-central1"
    network_name     = "wade-test-network"
  }

  # cluster details
  wade_cluster = {
    cluster_name                = "wade-gke"
    cluster_version             = "1.22.12-gke.500"
    subnet_name                 = "wade-gke"
    subnet_range                = "10.254.71.0/24"
    secondary_ip_range_pods     = "172.20.72.0/21"
    secondary_ip_range_services = "10.127.8.0/24"
    region                      = "us-central1"

    node_pools = [
      {
        name               = "app-pool"
        machine_type       = "n1-standard-2"
        node_locations     = join(",", slice(data.google_compute_zones.available.names, 0, 3))
        initial_node_count = 1
        min_count          = 1
        max_count          = 10
        max_pods_per_node  = 64
        disk_size_gb       = 100
        disk_type          = "pd-standard"
        image_type         = "COS"
        auto_repair        = true
        auto_upgrade       = false
        preemptible        = false
        max_surge          = 1
        max_unavailable    = 0
      }
    ]

    node_pools_labels = {
      all = {}
    }

    node_pools_tags = {
      all = ["k8s-nodes"]
    }

    node_pools_metadata = {
      all = {
        disable-legacy-endpoints = "true"
      }
    }

    node_pools_taints = {
      all = []
    }

    oauth_scopes = {
      all = [
        "https://www.googleapis.com/auth/monitoring",
        "https://www.googleapis.com/auth/compute",
        "https://www.googleapis.com/auth/devstorage.full_control",
        "https://www.googleapis.com/auth/logging.write",
        "https://www.googleapis.com/auth/service.management",
        "https://www.googleapis.com/auth/servicecontrol",
      ]
    }

    master_authorized_networks = [
      {
        display_name = "Whitelist 1"
        cidr_block   = "4.14.xxx.xx/32"
      },
      {
        display_name = "Whitelist 2"
        cidr_block   = "64.124.xxx.xx/32"
      },
    ]
  }
}

 

output.tf 

output "cluster_id" {
  description = "GKE cluster ID"
  value       = module.wade-gke.cluster_id
}

output "cluster_endpoint" {
  description = "Endpoint for GKE control plane"
  value       = module.wade-gke.endpoint
  sensitive   = true
}

output "cluster_name" {
  description = "Google Kubernetes Cluster Name"
  value       = module.wade-gke.name
}

output "region" {
  description = "GKE region"
  value       = module.wade-gke.region
}

output "project_id" {
  description = "GCP Project ID"
  value       = local.project.project_id
}

 

tf文件结构如下

 

部署

确保自己的GCP account已经登陆，并且有足够的权限操作GCP Project。

gcloud auth login

gcloud auth list

 

terraform init
terraform plan
terraform apply

 

配置连接GKE集群

### Adding the cluster to your context
gcloud container clusters get-credentials $(terraform output -raw cluster_name) \
--region $(terraform output -raw region) \
--project $(terraform output -raw project_id)

　

使用

 下载安装 kubectl 来管理以及部署资源到集群。

 

 ### 本文首发于博客园 https://www.cnblogs.com/wade-xu/p/16839468.html

 

使用Terraform创建EKS

准备好S3 bucket，更新backend

providers.tf

terraform {
  backend "s3" {
    bucket = "sre-dev-terraform"
    key    = "test/eks.tfstate"
    region = "cn-north-1"
  }
  required_providers {
    aws = {
      source  = "hashicorp/aws"
      version = "~> 4.25.0"
    }
  }
}

provider "aws" {
  region     = local.region
}

# https://github.com/terraform-aws-modules/terraform-aws-eks/issues/2009
provider "kubernetes" {
  host                   = module.wade-eks.cluster_endpoint
  cluster_ca_certificate = base64decode(module.wade-eks.cluster_certificate_authority_data)

  exec {
    api_version = "client.authentication.k8s.io/v1beta1"
    command     = "aws"
    # This requires the awscli to be installed locally where Terraform is executed
    args = ["eks", "get-token", "--cluster-name", module.wade-eks.cluster_id]
  }
}

　　

类似的，使用terraform aws module， 这里有个小插曲，我建的时候提示 cn-north-1d 这个zone没有足够的资源，所以我在data available zone 里面排除了这个zone

data "aws_availability_zones" "available" {

  # Cannot create cluster because cn-north-1d, 
  # the targeted availability zone, does not currently have sufficient capacity to support the cluster.
  exclude_names = ["cn-north-1d"]
}

module "wade-eks" {
  source  = "terraform-aws-modules/eks/aws"
  version = "18.27.1"

  cluster_name    = local.cluster_name
  cluster_version = local.cluster_version

  cluster_endpoint_private_access = true
  cluster_endpoint_public_access  = true

  # api server authorized network list
  cluster_endpoint_public_access_cidrs = local.master_authorized_networks

  cluster_addons = {
    coredns = {
      resolve_conflicts = "OVERWRITE"
    }
    kube-proxy = {}
    vpc-cni = {
      resolve_conflicts = "OVERWRITE"
    }
  }

  vpc_id     = module.vpc.vpc_id
  subnet_ids = module.vpc.private_subnets

  # Extend cluster security group rules
  cluster_security_group_additional_rules = local.cluster_security_group_additional_rules

  eks_managed_node_group_defaults = {
    ami_type     = local.node_group_default.ami_type
    min_size     = local.node_group_default.min_size
    max_size     = local.node_group_default.max_size
    desired_size = local.node_group_default.desired_size
  }

  eks_managed_node_groups = {
    # dmz = {
    #   name = "dmz-pool"
    # }
    app = {
      name                           = "app-pool"
      instance_types                 = local.app_group.instance_types
      create_launch_template         = false
      launch_template_name           = ""
      disk_size                      = local.app_group.disk_size

      create_security_group          = true
      security_group_name            = "app-node-group-sg"
      security_group_use_name_prefix = false
      security_group_description     = "EKS managed app node group security group"
      security_group_rules            = local.app_group.security_group_rules

      update_config = {
        max_unavailable_percentage = 50
      }
    }
  }

  # aws-auth configmap
  # create_aws_auth_configmap = true
  manage_aws_auth_configmap = true

  aws_auth_roles = [
    {
      rolearn  = "arn:aws-cn:iam::9935108xxxxx:role/CN-SRE" # replace me
      username = "sre"
      groups   = ["system:masters"]
    },
  ]

  aws_auth_users = [
    {
      userarn  = "arn:aws-cn:iam::9935108xxxxx:user/wadexu" # replace me
      username = "wadexu"
      groups   = ["system:masters"]
    },
  ]

  tags = {
    Environment = "dev"
    Terraform   = "true"
  }

  # aws china only because https://github.com/terraform-aws-modules/terraform-aws-eks/pull/1905
  cluster_iam_role_dns_suffix = "amazonaws.com"
}

 

locals.tf 这里为了安全性，最好给cluster api server endpoint 加好白名单来访问，否则 0.0.0.0/0代表全开

locals {
  cluster_name    = "test-eks-2022"
  cluster_version = "1.22"
  region          = "cn-north-1"

  vpc = {
    cidr = "10.0.0.0/16"
    private_subnets = ["10.0.1.0/24", "10.0.2.0/24"]
    public_subnets  = ["10.0.4.0/24", "10.0.5.0/24"]
  }

  master_authorized_networks =   [
    "4.14.xxx.xx/32",   # allow office 1
    "64.124.xxx.xx/32", # allow office 2
    "0.0.0.0/0"         # allow all access master node
  ]

  # Extend cluster security group rules example
  cluster_security_group_additional_rules = {
    egress_nodes_ephemeral_ports_tcp = {
      description                = "To node 1025-65535"
      protocol                   = "tcp"
      from_port                  = 1025
      to_port                    = 65535
      type                       = "egress"
      source_node_security_group = true
    }
  }

  node_group_default = {
    ami_type     = "AL2_x86_64"
    min_size     = 1
    max_size     = 5
    desired_size = 1
  }

  dmz_group = {
  }

  app_group = {
    instance_types = ["t3.small"]
    disk_size    = 50

    # example rules added for app node group
    security_group_rules = {
        egress_1 = {
          description = "Hello CloudFlare"
          protocol    = "udp"
          from_port   = 53
          to_port     = 53
          type        = "egress"
          cidr_blocks = ["1.1.1.1/32"]
        }
      }
  }
}

　

vpc.tf

module "vpc" {
  source  = "terraform-aws-modules/vpc/aws"
  version = "3.14.2"

  name = "wade-test-vpc"

  cidr = local.vpc.cidr
  azs  = slice(data.aws_availability_zones.available.names, 0, 2)

  private_subnets = local.vpc.private_subnets
  public_subnets  = local.vpc.public_subnets

  enable_nat_gateway   = true
  single_nat_gateway   = true
  enable_dns_hostnames = true

  public_subnet_tags = {
    "kubernetes.io/cluster/${local.cluster_name}" = "shared"
    "kubernetes.io/role/elb"                      = 1
  }

  private_subnet_tags = {
    "kubernetes.io/cluster/${local.cluster_name}" = "shared"
    "kubernetes.io/role/internal-elb"             = 1
  }
}

　　

output.tf

output "cluster_id" {
  description = "EKS cluster ID"
  value       = module.wade-eks.cluster_id
}

output "cluster_endpoint" {
  description = "Endpoint for EKS control plane"
  value       = module.wade-eks.cluster_endpoint
}

output "region" {
  description = "EKS region"
  value       = local.region
}

output "cluster_name" {
  description = "AWS Kubernetes Cluster Name"
  value       = local.cluster_name
}

 

文件结构如下

 ### 本文首发于博客园 https://www.cnblogs.com/wade-xu/p/16839468.html

 

部署

配置aws account key/secret

Option 1: Export AWS access and security to environment variables

export AWS_ACCESS_KEY_ID=xxx

export AWS_SECRET_ACCESS_KEY=xxx

 

Option 2: Add a profile to your AWS credentials file

aws configure
# or
vim ~/.aws/credentials

[default]
aws_access_key_id=xxx
aws_secret_access_key=xxx

 

可以使用如下命令来验证当前用的是哪个credentials

aws sts get-caller-identity 

　　

部署tf资源

terraform init

terraform plan

terraform apply

　　

成功之后有如下输出

  

配置连接EKS集群

#### Adding the cluster to your context
aws eks --region $(terraform output -raw region) update-kubeconfig \
    --name $(terraform output -raw cluster_name)　　

 

使用

同上面，需要下载kubectl 

Example 命令：

kubectl cluster-info

kubectl get nodes

　　

感谢阅读，如果您觉得本文的内容对您的学习有所帮助，您可以打赏和推荐，您的鼓励是我创作的动力