Deploy development k0s cluster

Deploy

Deployment of the cluster consist of two main phases:

1. Deployment of infra and k0s on top of this infra.
2. Deployment of applications on top of k0s cluster.

AWS Prereqs

In order to deploy the cluster on AWS:

1. terraform and k0sctl to be installed.
2. Obtain user credentials and set them as environment variables.
3. Clone this repo to your machine.
4. Set infra variable according to your environment and use case in infra/terraform/aws/terraform.tfvars (it can be done easily by copying infra/terraform/aws/terraform.tfvars.example in the same folder).

Deployment process

To deploy infra and k0s, execute following commands:

cd $GIT_ROOT/infra/aws
terraform apply -auto-approve

Here you need to wait a bit for domain name of newly created LB for the cluster to be propagated, otherwise you'll face connectivity issue. If you face this issue, just try to execute k0sctl related commands one more time.

terraform output -raw k0sctl_yaml | k0sctl apply --config -
terraform output -raw k0sctl_yaml | k0sctl kubeconfig --config - > kubeconfig
export KUBECONFIG=$GIT_ROOT/infra/aws/kubeconfig

Once k0s cluster is deployed, we can proceed to applications deployment

cd $GIT_ROOT
kubectl apply -k apps/argocd
kubectl apply -k apps/utility

Here you need to wait again for ArgoCD pods to be up and running (you can check it by executing kubectl get pods -n argocd). Then, when all ArgoCD pods are healthy, execute:

k apply -f apps/apps.yaml

Destroy

AWS Destroy process

To destroy the cluster, run:

GIT_ROOT=$(git rev-parse --show-toplevel)
kubectl delete -f $GIT_ROOT/apps/apps.yaml
cd $GIT_ROOT/infra/aws
terraform destroy -auto-approve