Skip to content

Latest commit

 

History

History
41 lines (23 loc) · 2.28 KB

minikube-setup.md

File metadata and controls

41 lines (23 loc) · 2.28 KB
Raw

Install minikube

On a laptop or VM to run the minikube cluster, the docker engine must be running before further into other steps.

To have the docker engine install, refer to its document.

As an example, on a MacOS laptop, the Docker Desktop should be installed with the instruction. After it is installed, go to its preferences and make some configuration changes similar to follow docker-desktop to give the to-be-created cluster enough CPU resources.

Current Docker Desktop on MacOS has an option to start a one-node Kubernetes cluster, one may enable enable-kubernetes it and use the cluster for the dojo.

Alternatively one should install and set up a minikube cluster. This is also preferred by this dojo. Reason for such choice is obvious as developers working on many cloud related projects are not deploying services directly on their MacOS laptop.

To install minikube, follow this link. It basically includes three steps:

  • install kubectl. For the purpose of this dojo, please choose version v1.15.6 of kubectl to install.

  • install a hypervisor. On Ubuntu and other linux VM, a minikube cluster can run without a hypervisor and so this step can be bypassed.

  • install minikube.

Start minikube

For the purpose of this dojo, we recommend start up a minikube cluster with following command line:

  minikube start --driver=docker --memory=8192MB --cpus=6 --disk-size=30000mb --kubernetes-version=v1.15.6

Note: on Linux VM, you can set the --driver=none to not use a hypervisor.

This configuration takes into account of the memory and cpu resources that are suitable to run a Kubeflow deployment with most of essential applications.

Once the minikube cluster is up and running, you should see some screen output like following minikube.

Like other Kubernetes clusters, you can also view the minikube cluster dashboard with this command:

minikube dashboard&

The dashboard with show on your default web browser like this. dashboard