Local K3s Cluster Made Easy With Multipass
Nov 8, 2019 · 4 min read · source
A couple of weeks ago, I stumble upon Multipass, a tool that allows you to spin up Ubuntu VMs in a matter of seconds on a Mac, Linux, or Windows workstation.
Depending upon your OS, Multipass uses Hyper-V, HyperKit, KVM, or VirtualBox natively for the fastest startup time.
In this short article, we’ll set up a K3s Kubernetes cluster on virtual machines created with Multipass.
Getting Multipass
Just head up to Multipass, and download the binary for your platform. Once installed, you can get the list of the available commands. We’ll only use a couple of them in this article.
$ multipass
Usage: multipass \[options\] <command>
Create, control and connect to Ubuntu instances.This is a command line utility for multipass, a
service that manages Ubuntu instances.Options:
-h, --help Display this help
-v, --verbose Increase logging verbosity, repeat up to three times for more
detailAvailable commands:
delete Delete instances
exec Run a command on an instance
find Display available images to create instances from
get Get a configuration option
help Display help about a command
info Display information about instances
launch Create and start an Ubuntu instance
list List all available instances
mount Mount a local directory in the instance
purge Purge all deleted instances permanently
recover Recover deleted instances
restart Restart instances
set Set a configuration option
shell Open a shell on a running instance
start Start instances
stop Stop running instances
suspend Suspend running instances
transfer Transfer files between the host and instances
umount Unmount a directory from an instance
version Show version details
Creation of the VMs
First, we create three VMs, keeping the default configuration of one CPU and 1GB RAM.
$ multipass launch -n node1
$ multipass launch -n node2
$ multipass launch -n node3
Next, we verify the three VMs are in the running state
$ multipass list
Name State IPv4 Image
node3 Running 192.168.64.13 Ubuntu 18.04 LTS
node2 Running 192.168.64.12 Ubuntu 18.04 LTS
node1 Running 192.168.64.11 Ubuntu 18.04 LTS
Note: In this example, I use Multipass on MacOS, so the VMs are run on HyperKit as I can confirm listing the HyperKit-related processes.
luc@saturn:~$ ps aux | grep -i "\[h\]yperkit"
... com.docker.hyperkit ...
... com.canonical.multipass/bin/hyperkit...
... com.canonical.multipass/bin/hyperkit...
... com.canonical.multipass/bin/hyperkit...
Among those processes, the first one is the Docker daemon already running on my machine. The next three are the VMs created above.
Initialize K3s on node1
First, we run a command on node1 (using the exec subcommand of Multipass) to install K3s as defined in the documentation.
$ multipass exec node1 -- \
bash -c "curl -sfL https://get.k3s.io | sh -"
Note: The K3s initialisation took less than 30 seconds … really impressive for a certified Kubernetes distribution.
Then, from the local machine, we gather the elements we’ll need to add additional nodes to the cluster:
- The token to join the cluster
$ TOKEN=$(multipass exec node1 sudo cat /var/lib/rancher/k3s/server/node-token)
- The IP of the API server running on
node1
$ IP=$(multipass info node1 | grep IPv4 | awk '{print $2}')
Join node2 and node3 to the Cluster
From the local machine, we run the following commands to add node2 and node3 to the cluster.
# Joining node2
$ multipass exec node2 -- \
bash -c "curl -sfL https://get.k3s.io | K3S_URL=\"https://$IP:6443\" K3S_TOKEN=\"$TOKEN\" sh -"
# Joining node3
$ multipass exec node3 -- \
bash -c "curl -sfL https://get.k3s.io | K3S_URL=\"https://$IP:6443\" K3S_TOKEN=\"$TOKEN\" sh -"
As we can see, those commands use the TOKEN and IP address we defined above.
Get the Cluster Configuration
Running the following command, we can list all the nodes of the cluster:
$ multipass exec node1 -- sudo kubectl get nodes
NAME STATUS ROLES AGE VERSION
node1 Ready master 5m31s v1.16.2-k3s.1
node3 Ready <none> 21s v1.16.2-k3s.1
node2 Ready <none> 45s v1.16.2-k3s.1
To access the cluster’s API server from our local machine, we only need to get the kubeconfig file created on node1 during the setup.
$ multipass exec node1 sudo cat /etc/rancher/k3s/k3s.yaml > k3s.yaml
We then need to change the server key so it references the remote IP address of node1 instead of the localhost.
sed -i '' "s/127.0.0.1/$IP/" k3s.yaml
Finally, we configure our local kubectl so it uses the kubeconfig file (k3s.yaml) we’ve just retrieved. A simple way is to set the KUBECONFIG environment variable so it points towards the configuration file.
export KUBECONFIG=$PWD/k3s.yaml
We’re all set, we can now communicate with the cluster from your local machine.
$ kubectl get nodes
NAME STATUS ROLES AGE VERSION
node1 Ready master 10h v1.16.2-k3s.1
node2 Ready <none> 10h v1.16.2-k3s.1
node3 Ready <none> 10h v1.16.2-k3s.1
Note: all the steps defined in this post can be run from the following gist:
https://gist.github.com/lucj/5a0e2286b40130d02388a264e6924ed4
Summary
Multipass is a very useful tool. Its integration with low-level hypervisors makes it a good choice to deploy multiple VMs locally and to play with many distributed solutions such as the great K3s Kubernetes distribution.
Responses (4)
Yohannes Russom
11 months ago
2
Thank you ,
I am facing an issue with kubectl command not found` and i guess its becuase the k3s deployment didnt get succeded.
Zulhilmi Zainudin
6 months ago
Hi. I’m curious, why you put “h” in square brackets for this command?
ps aux | grep -i “\[h\]yperkit”
Scott Lett
7 months ago
Nice work!
On my Ubuntu box, the the sed command doesn’t work unless I add “-e” before the substitution expression.
Praveen Shastrula
3 months ago
Very simple and completely automated. Thank you! (I followed a different article first, but it had some missing instruction and crashed my nodes!, but yours works!)