The following commands create a 4 node cluster, get and export the kubeconfig, list the cluster and show the nodes: $ k3d create -workers 3 $ k3d get-kubeconfig -name=’k3s-default’ $ export KUBECONFIG=~/.config/k3d/k3s-default/kubeconfig.yaml $ k3d list $ kubectl get nodes If you’re using k3d for the first time, deploying k3s with k3d is a breeze away, you won’t believe how fast you can get a k3s made k8s cluster running within a few seconds.

To install k3d on Mac or Linux using Homebrew or curl, you need only to run: $ brew install k3d or $ curl -s | TAG=v1.3.4 bash $ k3d -v k3d version v1.3.4 $ k3d -h # help is helpful :-) To overcome this issue Almir Kadric has developed a tuntap support shim installer for Docker on Mac and this post is about how to use this nice implementation with MetalLB on a k3d launched k3s cluster on your machine.ĭocker: you need docker running on your machine: $ docker -version Docker version 19.03.5, build 633a0ea on multipass VMs.īut if it comes to exposing a service with the type LoadBalancer on k3s, we’ll face some problems, since docker desktop on Mac doesn’t support network routing into the host virtual machine which is created using hyperkit on Mac and the main reason is due to the fact that the network interface options used to create the hyperkit instance does not create a bridge interface between the Physical Machine and the Host Virtual Machine. If you want to run RIO on k3s on your machine, you need to have k3s running on VMs, e.g. K3d is the default and in most cases the preferred deployment tool for k3s clusters on our machines, it uses docker to run a k3s cluster within few seconds in docker containers.

In my previous post we could see how to get an external IP for load balancing on a k3s cluster running in multipass VMs and I promised to show you how MetalLB can work with k3d launched k3s clusters on Mac too (Linux users are lucky, MetalLB works out of the box with k3d).

Install docker on mac using brew