Kubernetes cluster-api-provider-linode

This repository hosts a concrete implementation of a provider for Linode for the cluster-api project.

Project Status

This project is currently Work-in-Progress and may not be production ready. There is no backwards-compatibility guarantee at this point. Checkout the Features portion of the README for details about the project status.

Getting Started

Prerequisites

In order to create a cluster using this cluster-api implementation, you need the following tools installed on your local machine. These are installation insturctions for macOS. For installation instructions on other platforms, visit the respective links.

• Go toolchain

brew install go

• This Cluster-API implementation

go get github.com/asauber/cluster-api-provider-linode

• kubectl

brew install kubernetes-cli

• minikube version 0.28.1

curl -Lo minikube https://storage.googleapis.com/minikube/releases/v0.28.1/minikube-darwin-amd64 && chmod +x minikube && sudo cp minikube /usr/local/bin/ && rm minikube

• virtualbox on macOS or kvm2 on Linux.

brew cask install virtualbox

• Linode API Access Token generated and set as the LINODE_TOKEN environment variable

export LINODE_TOKEN="<Your Linode API Token>"
echo "export LINODE_TOKEN=<Your Linode API Token>" >> ~/.bash_profile

• The Linode CLI

pip install linode-cli --upgrade
linode-cli

• kustomize

brew install kustomize

Creating a Cluster

To create a cluster using cluster-api-provider-linode, you

1. Start minikube
2. Deploy a collection of Kubernetes resources to minikube that implement the cluster-api, the "provider components"
3. Deploy a collection of Kubernetes resources to minikube that represent your new cluster

First, start minikube

minikube start
# As of 2018-11-06 there is a memory leak in the cluster master components that
# may require more resources of minikube. You can start minikube with more
# resources as follows.
minikube --memory 8192 --cpus 2 start

Next, use kustomize to render the templates for the provider components.

kustomize build config/default/ > cmd/examples/linode/provider-components.yaml
echo "---" >> cmd/examples/linode/provider-components.yaml
kustomize build vendor/sigs.k8s.io/cluster-api/config/default/ >> cmd/examples/linode/provider-components.yaml

This generates a YAML file which will be used to create the namespaces, custom resource defintions, roles, rolebindings, and services which run the Linode cluster-api controllers and the upstream cluster-api controllers.

Deploy these resources to minikube.

kubectl create -f cmd/examples/linode/provider-components.yaml

Next, generate manifests for your cluster. You must provide a path to a public key that will be used to access your cluster.

cd cmd/examples/linode
./generate-yaml.sh $HOME/.ssh/id_rsa.pub
cd ../../..

This generates a YAML file which will be used to create a namespace, cluster, machine, and secrets for the new cluster. Feel free to edit this manifest, for example, to add some additional machines with the "Node" Role. If you would like to use the manifest to add an additional cluster, you will need to edit the namespace used by the all of the resources. Note that only one Master machine and one Cluster resource per namespace is currently supported.

Create these cluster resources

kubectl create -f cmd/examples/linode/cluster.yaml

You can observe the logs of the controllers while the cluster is being created

kubectl logs -f cluster-api-provider-linode-controller-manager-0 -n cluster-api-provider-linode-system

Interacting With Your New Cluster

Download the admin kubeconfig for your cluster. The first argument should be namespace used for your cluster resources.

hack/get_kubeconfig.sh cluster0

You can now interact with the cluster via kubectl. Note that it may take a while before all services are running on the cluster.

kubectl --kubeconfig kubeconfig get nodes
kubectl --kubeconfig kubeconfig get all --all-namespaces

The cluster is fully functional when at least the following services are runnning: kube-apiserver, kube-controller-manager, kube-proxy, kube-scheduler, etcd, coredns, CNI (Calico)

Deleting a cluster

Use the following script to delete a cluster deployed with this API. The first argument must be a namespace used for cluster resources.

hack/delete_cluster.sh cluster01

Development

When using a kustomize-rendered provider-components.yaml, the cluster-api-provider-linode controllers are deployed from a container registry (currently asauber/cluster-api-provider-linode on Docker Hub). To work on development, you can use a make target which compiles and runs the controllers on your local machine.

Follow the above instructions above to deploy the CRDs, upstream controllers, and other required resources to your cluster.

Delete the service and statefulset for the provider controllers, because you'll be running it locally.

kubectl delete statefulset cluster-api-provider-linode-controller-manager -n cluster-api-provider-linode-system
kubectl delete services cluster-api-provider-linode-controller-manager-service -n cluster-api-provider-linode-system

Use the make target to compile and run the controller locally. This will use the kubeconfig at ~/.kube/config

make run

You will immediately see Info-level logging output on your terminal if you have deployed Cluster or Machine resources.