Local minikube installation

Here is a step-by-step guide on how to deploy the UMH stack using minikube, a tool to quickly spin up a Kubernetes cluster in your local machine. This method is quicker than using a virtual machine with k3os, but less reliable. We suggest using minikube only for small tests.

Bear in mind that this way of setting up the UMH stack is not recommended. For testing and development purposes we recommend a k3d installation.

If you are using a different way to install the software, make sure to install the following versions. Specifically installing minikube directly from the website results in a new version, which is not supported yet. Older versions are also not supported anymore.

• minikube 1.26.1
• Kubernetes 1.24.3
• Helm 3.10

Requirements

• Operating system (64-bit):
  • Windows 10 version 2004 and higher or Windows 11
• Hardware:
  • CPU: 2 cores or more
  • RAM: 8GB or more
  • Storage: 10GB free

Dependencies

1. Verify that you have WSL installed.
   • Open Powershell and type wsl --status. If you get an error you probably need to install it. Refer to the Microsoft documentation on how to do that.
2. Docker Desktop
   • You can follow the Docker documentation here.
3. Chocolatey
   • Read how to do that here.

Installation

1. Open a PowerShell terminal. Be sure to use the option Run as Administrator.

2. Install minikube

   choco install minikube

   when using Ubuntu you can also use the following command

   curl -LO https://storage.googleapis.com/minikube/releases/v1.26.1/minikube-linux-amd64 && sudo install minikube-linux-amd64 /usr/local/bin/minikube

3. Install helm

   choco install kubernetes-helm

4. Install kubectl

   choco install kubernetes-cli

5. Set the default driver for minikube. We recommend Docker, but you can use a variety of drivers. Read the Minikube docs for further information.

   minikube config set driver docker

6. Start an instance of minikube with 2 CPUs and 4GB of RAM. You may need more if you decide to customize the services with the development.yaml file.

   minikube start --cpus 2 --memory 4096

7. Create a namespace in Kubernetes

   kubectl create namespace united-manufacturing-hub

8. Add the UMH repo helm repo add united-manufacturing-hub https://repo.umh.app/

9. Ensure it is up-to-date with helm repo update

10. Install the stack

    helm install united-manufacturing-hub united-manufacturing-hub/united-manufacturing-hub -n united-manufacturing-hub

11. Get the Helm config to import into UMHLens

    kubectl config view

12. Open UMHLens on your device. You can get UMHLens for free from this GitHub Repository.

13. Sometimes UMHLens automatically creates a minikube-Cluster. Click on Browse-Clusters to check. If there is no cluster you have to create it manually (follow step 14-18) otherwise continue with step 19.

14. Click the three horizontal lines in the upper left corner and choose files → preferences

15. Click on Kubernetes and select “Add custom repo”, type in https://repo.umh.app as the URL and decide on a name for the repository.

16. Click again file → Add Cluster

17. Paste the clipboard (which you got from kubectl config view) into the kubectl prompt of UMHLens.

18. Click on Add Cluster.

19. Click on Browse Clusters in Catalog, then connect to the Cluster.

20. Click on Helm -> Releases and change the namespace from default to united-manufacturing-hub in the upper right corner

    

21. Click on the Release united-manufacturing-hubto be able to inspect the values.yaml file, which holds the configurations of all microservices used in the cluster.

Enable minikube ingress

1. minikube addons enable ingress
2. minikube tunnel

Now it should look like this:

![Untitled](/images/knowledge-base/local-minikube-installation/Untitled.png)

Any you can access Node-RED e.g., via localhost:1880/nodered

This tunnel needs to be always started using minikube tunnel when you want to access these services.

What’s next

You have successfully deployed the UMH stack using minikube.

The next step will be to gather some kind of data, and since you did a local installation you probably want to simulate it. We have a guide to generate data with an MQTT Simulator.

You could also directly connect NodeRED without having data. Follow this guide to see how.

If you are not sure on what to do you can get back to the Getting Started page.

Advanced: run multiple clusters

You can run multiple clusters of minikube by simply using minikube start -p newcluster. This creates a new cluster (called “Profile” in minikube) with the name newcluster.

To switch between the profiles you can use minikube profile <profilename>.

kubectl will automatically change its scope whenever you switch profile.

Learn more about profiles in the minikube documentation.