Go / Terratest / Terraform to Create Dynamic Number of Rancher HA Setups

This setup creates a dynamic amount of HA setups without Route 53, so you can use your own TLD.

Webhook Hardening Alert

If wanting to harden the webhook, please see the bottom of this README.md BEFORE running rke up.

What is the purpose of this Terraform?

For Rancher QA to easily create HA Rancher setups in AWS with RKE1 as the base.

This will create however many folders you specify next to ha_test.go

• high-availability-1
• high-availability-2
• high-availability-3
• etc,
• etc..

These folders will contain a cluster.yml pre-configured to run rke up against with whatever rke version you want.

And also a pre-configured installation script that you can run zsh install.sh against.

How to use it?

All you need to do to make this terraform work is to clone the repository and create a file called tool-config.yml that sits next to the README.md.

How the tool-config.yml file should look like:

local:
  pem_path: "your-local-path-to-the-pem-file-you-use-for-aws"
rancher:
  bootstrap_password: "whatever-bootstrap-password-you-want"
total_has: 4  # Number of HA instances to create
ha_config:  # Single template config for all HA instances
  image: "v2.9.2"
  chart: "2.9.2"
tf_vars:
  aws_access_key: "your-aws-access-key"
  aws_secret_key: "your-aws-secret-key"
  aws_prefix: "aws-prefix-should-only-be-3-characters-like-your-initials"
  aws_vpc: "aws-vpc-you-want-to-use"
  aws_subnet_a: "your-subnet-a"
  aws_subnet_b: "your-subnet-b"
  aws_subnet_c: "your-subnet-c"
  aws_ami: "whatever-ami-you-want-one-with-docker"
  aws_subnet_id: "your-subnet-id"
  aws_security_group_id: "whatever-security-group-you-want"
  aws_pem_key_name: "your-aws-pem-key-name-in-aws-no-file-extension"

Then you just need to run this function in ha_test.go >>> TestHaSetup

There is also a cleanup function that you can run in ha_test.go >>> TestHACleanup

How Long Does it Take to Run?

Completes TestHaSetup in ~4 minutes

Harden Webhook Guide (TODO: Automate)

1. Add the following to the cluster.yml file DO NOT RUN RKE UP YET:

ssh_key_path: <redacted>
network:
  plugin: calico
services:
  kube-api:
    extra_args:
      admission-control-config-file: "/etc/rancher/admission/admission.yaml"
    extra_binds:
      - "/etc/rancher/admission/admission.yaml:/etc/rancher/admission/admission.yaml"
      - "/etc/rancher/admission/kubeconfig:/etc/rancher/admission/kubeconfig"
      - "/etc/rancher/admission/client.csr:/etc/rancher/admission/client.csr"
      - "/etc/rancher/admission/client.key:/etc/rancher/admission/client.key"
nodes:

2. SSH into each controlplane ec2 nodes and create these directories.

sudo mkdir /etc/rancher
sudo mkdir /etc/rancher/admission
sudo chmod 777 /etc/rancher
sudo chmod 777 /etc/rancher/admission

3. Then exit the SSH session
4. Then have these files on your local machine, and scp them onto the controlplane ec2 nodes.

# /etc/rancher/admission/admission.yaml
apiVersion: apiserver.config.k8s.io/v1
kind: AdmissionConfiguration
plugins:
  - name: ValidatingAdmissionWebhook
    configuration:
      apiVersion: apiserver.config.k8s.io/v1
      kind: WebhookAdmissionConfiguration
      kubeConfigFile: "/etc/rancher/admission/kubeconfig"
  - name: MutatingAdmissionWebhook
    configuration:
      apiVersion: apiserver.config.k8s.io/v1
      kind: WebhookAdmissionConfiguration
      kubeConfigFile: "/etc/rancher/admission/kubeconfig"

# /etc/rancher/admission/kubeconfig
apiVersion: v1
kind: Config
users:
- name: 'rancher-webhook.cattle-system.svc'
  user:
    client-certificate: /etc/rancher/admission/client.csr
    client-key: /etc/rancher/admission/client.key

5. Run the command: openssl req -newkey rsa:2048 -nodes -keyout client.key -out client.csr -x509 -days 365

6. Now run these scp commands:

• scp -i ~/your/pem-file.pem admission.yaml [email protected]:/etc/rancher/admission/
• scp -i ~/your/pem-file.pem kubeconfig [email protected]:/etc/rancher/admission/
• scp -i ~/your/pem-file.pem client.csr [email protected]:/etc/rancher/admission/
• scp -i ~/your/pem-file.pem client.key [email protected]:/etc/rancher/admission/
• Run these 4 commands against every controlplane node

sudo chmod 777 /etc/rancher/admission/admission.yaml
sudo chmod 777 /etc/rancher/admission/kubeconfig
sudo chmod 777 /etc/rancher/admission/client.csr
sudo chmod 777 /etc/rancher/admission/client.key

7. Run rke up against the cluster.yml file that this repository creates. After adding the additional lines to the cluster.yml file.

8. Install Rancher via helm

9. Install Calico CTL On Your Local Machine: https://docs.tigera.io/calico/latest/operations/calicoctl/install

10. Find the IPs needed for the network.yaml.

11. Run the following: calicoctl get node --allow-version-mismatch -o yaml

12. If you have 3 controlplane nodes, like the default for this repository sets up, find 3 of these lines. ipv4IPIPTunnelAddr: <redacted>

13. Take these 3 IPs and add them to the network.yaml file. Replacing <redacted> with the IP.

apiVersion: crd.projectcalico.org/v1
kind: NetworkPolicy
metadata:
  name: allow-k8s
  namespace: cattle-system
spec:
  selector: app == 'rancher-webhook'
  types:
    - Ingress
  ingress:
    - action: Allow
      protocol: TCP
      source:
        nets:
          - <redacted>/32
          - <redacted>/32
          - <redacted>/32
      destination:
        selector:
          app == 'rancher-webhook'

14. Then apply the network.yaml with k apply -f network.yaml

15. Create a values.yaml and delete the rancher-config

16. base64 encode the client.csr from earlier.

17. create a values.yaml file

# values.yaml
auth:
  clientCA: <base64-string-goes-here>

18. delete the rancher-config with k delete configmap rancher-config -n cattle-system
19. recreate it with kubectl --namespace cattle-system create configmap rancher-config --from-file=rancher-webhook=values.yaml
20. you can check that it was picked up with helm get values rancher-webhook -n cattle-system