kubespray/contrib/terraform/gcp/README.md

4.5 KiB

Kubernetes on GCP with Terraform

Provision a Kubernetes cluster on GCP using Terraform and Kubespray

Overview

The setup looks like following

                           Kubernetes cluster
                        +-----------------------+
+---------------+       |   +--------------+    |
|               |       |   | +--------------+  |
| API server LB +---------> | |              |  |
|               |       |   | | Master/etcd  |  |
+---------------+       |   | | node(s)      |  |
                        |   +-+              |  |
                        |     +--------------+  |
                        |           ^           |
                        |           |           |
                        |           v           |
+---------------+       |   +--------------+    |
|               |       |   | +--------------+  |
|  Ingress LB   +---------> | |              |  |
|               |       |   | |    Worker    |  |
+---------------+       |   | |    node(s)   |  |
                        |   +-+              |  |
                        |     +--------------+  |
                        +-----------------------+

Requirements

  • Terraform 0.12.0 or newer

Quickstart

To get a cluster up and running you'll need a JSON keyfile. Set the path to the file in the tfvars.json file and run the following:

terraform apply -var-file tfvars.json -state dev-cluster.tfstate -var gcp_project_id=<ID of your GCP project> -var keyfile_location=<location of the json keyfile>

To generate kubespray inventory based on the terraform state file you can run the following:

./generate-inventory.sh dev-cluster.tfstate > inventory.ini

You should now have a inventory file named inventory.ini that you can use with kubespray, e.g.

ansible-playbook -i contrib/terraform/gcs/inventory.ini cluster.yml -b -v

Variables

Required

  • keyfile_location: Location to the keyfile to use as credentials for the google terraform provider
  • gcp_project_id: ID of the GCP project to deploy the cluster in
  • ssh_pub_key: Path to public ssh key to use for all machines
  • region: The region where to run the cluster
  • machines: Machines to provision. Key of this object will be used as the name of the machine
    • node_type: The role of this node (master|worker)
    • size: The size to use
    • zone: The zone the machine should run in
    • additional_disks: Extra disks to add to the machine. Key of this object will be used as the disk name
      • size: Size of the disk (in GB)
    • boot_disk: The boot disk to use
      • image_name: Name of the image
      • size: Size of the boot disk (in GB)
  • ssh_whitelist: List of IP ranges (CIDR) that will be allowed to ssh to the nodes
  • api_server_whitelist: List of IP ranges (CIDR) that will be allowed to connect to the API server
  • nodeport_whitelist: List of IP ranges (CIDR) that will be allowed to connect to the kubernetes nodes on port 30000-32767 (kubernetes nodeports)

Optional

  • prefix: Prefix to use for all resources, required to be unique for all clusters in the same project (Defaults to default)
  • master_sa_email: Service account email to use for the control plane nodes (Defaults to "", auto generate one)
  • master_sa_scopes: Service account email to use for the control plane nodes (Defaults to ["https://www.googleapis.com/auth/cloud-platform"])
  • master_preemptible: Enable preemptible for the control plane nodes (Defaults to false)
  • master_additional_disk_type: Disk type for extra disks added on the control plane nodes (Defaults to "pd-ssd")
  • worker_sa_email: Service account email to use for the worker nodes (Defaults to "", auto generate one)
  • worker_sa_scopes: Service account email to use for the worker nodes (Defaults to ["https://www.googleapis.com/auth/cloud-platform"])
  • worker_preemptible: Enable preemptible for the worker nodes (Defaults to false)
  • worker_additional_disk_type: Disk type for extra disks added on the worker nodes (Defaults to "pd-ssd")

An example variables file can be found tfvars.json

Known limitations

This solution does not provide a solution to use a bastion host. Thus all the nodes must expose a public IP for kubespray to work.