Nginx with k0s, Terraform, and AWS

Bootstrap a k0s cluster in AWs with terraform and install Nginx.

Pre-requisite

Along with boundless CLI, the following tools will also be required:

• AWS - used to create VMs for running the cluster
• terraform - used setup VMs in AWS
• k0sctl - required for installing a k0s distribution
• kubectl - used to forward ports to the cluster

Setting up VMs in AWS

Refer to the example Terraform scripts for creating VMs in AWS.

1. Change to the directory containing the Terraform scripts.
2. Copy the terraform.tfvars.example to terraform.tfvars and change the content to be similar to:

   cluster_name = "example-boundless"
   controller_count = 1
   worker_count = 1
   cluster_flavor = "m5.large"
3. terraform init
4. terraform apply
5. terraform output --raw bop_cluster > ./VMs.yaml

Setting up the blueprint

Download the example blueprint for Nginx.

Modify the blueprint so that the spec.kubernetes.infra.hosts section matches your AWS VMs’ IP address, username, SSH port, and SSH credentials. The values can be passed as environment variables or replaced with your own values. For example, the hosts section should match the output from terraform output --raw bop_cluster. For example:

spec:
  kubernetes:
    provider: k0s
    version: 1.27.4+k0s.0
    infra:
      hosts:
        - ssh:
            address: 52.91.89.114
            keyPath: ./example/aws-tf/aws_private.pem
            port: 22
            user: ubuntu
          role: controller
        - ssh:
            address: 10.0.0.2
            keyPath: ./example/aws-tf/aws_private.pem
            port: 22
            user: ubuntu
          role: worker

You can also modify the spec.components.addons.chart.values section. For complete configuration options, see the Nginx documentation.

Apply the blueprint

Apply the blueprint using using bctl:

bctl apply -f k0s-in-aws-with-tf.yaml

It will take a few moments before the Nginx pods are ready. You can monitor the progress with:

kubectl get pods -w

Cleanup

To remove the cluster, run:

bctl reset -f k0s-in-aws-with-tf.yaml

This will remove all resources created by the blueprint but leave the k0s cluster.