This repo contains scripts and configuration files needed to bring up a cluster of virtual machines with a running Kubernetes cluster (1 control plane node, 2 worker nodes; scalable).
Use config.yaml to decide how many machines will be brought up. Keep in mind
that we currently support one control plane node.
You will need:
- Vagrant (tested with 2.2.18)
- VirtualBox (tested with 6.1.2 r145957)
- Vagrant vbguest plugin (can be installed with
vagrant plugin install vagrant-vbguest)
As per this configuration, each machine will be named node-X where node-1 is
the control plane node. node-2 and node-3 will be worker nodes and any other
one will get an incremental ID.
The IPs of the nodes are stated in the config.yaml. They are within the
10.0.0.0/24 range. 10.0.0.2 is always expected to be the IP of the control plane
(and the first node).
The Pods IP range is configured by default to 192.168.0.0/16. The cluster is
deployed with Calico as the network provider, and the range was selected to
match it's defaults.
Run vagrant up to bring up the cluster. The first time it will take some time
creating the VMs and provisioning them.
The provision.sh script will be run for all machines, while the master.sh
and worker.sh scripts will run only in the machines with that role (as stated
in the config.yaml file. Feel free to change those to add/remove
configurations.
Run vagrant halt to shut down all machines without deleting it. You can also
halt a single machine by using the host name.
Run vagrant status to get see which virtual machines are currently running.
Run vagrant ssh node-x to ssh into the machine x.
WARNING: This will permanently delete all the nodes and their data.
Run vagrant destroy to permanently delete all virtual machines. You'll be
prompted to confirm deletion for each of them. Alternativelly you can delete a
signle machine using its host name.
Vagrant virtual machines share with the host the directory where
the Vagrantfile is. This direcotry is mounted in /vagrant on every VM.