/vagrant-k8s-kubeadm

Primary LanguageShell

Vagrant project to create various Kubernetes clusters for learning using kubeadm. It requires VirtualBox and vagrant to be installed.

More about Vagrant can be found here

This project contains Vagrant files and configuration for for 4 different kubernetes clusters

  • cluster1 - Creates a fully functioning kubernetes cluster with 1 master and 2 worker nodes.

  • cluster2 - Vagrantfile creates a cluster that contains a functioning master and 1 worker node. A 2nd worker machine is created so you can practice joining a new node to a cluster.

  • cluster3 - Vagrant file creates two machines configured with the basics, including Docker, but no kubernetes resources. It can be used to practice configuring a new Kubernetes cluster with kubeadm.

  • cluster4 - Vagrant file creates a fully functioning kubernetes cluster with and older version of kubernetes installed. Use it to practice upgrading kubernetes on a running cluster.

The following are useful commands for working with the clusters.

vagrant up - This will provision the machines and run the scripts creating the cluster as necessary.

vagrant suspend- This will shutdown the machines and save their state. They can be restarted by re-running vagrant up

vagrant halt && vagrant delete - This will stop and delete the vms. This would be necessary for clusters 2-4 when practicing Installing or updating kubernetes.

vagrant ssh cluster1-master1 - This allows you to ssh into the vm specified. When inside the vm, the current folder is mounted as /vagrant/ and the scripts can be access from that directory.

** Troubleshooting **

  • Depending on the machine, I've occasionally seen issues with clusterX-worker1 not being able to join the cluster because the kube-apiserver not being fully functional when the join command was executed on the worker. If this occurs, you can wait until k get nodes shows a Ready status for the clusterX-master1 node. The, vagrant ssh clusterX-worker1 and run the join script using sudo /vagrant/scripts/kubeadm-join.sh

** The project makes use of direnv to set the KUBECONFIG environment variable for whichever cluster you're currently working with. If you don't have that installed, you can source the .envrc file manually.