kubeslice-controller

README

kubeslice-controller

kubeslice-controller uses Kubebuilder, a framework for building Kubernetes APIs using custom resource definitions (CRDs).

Getting Started

The KubeSlice Controller orchestrates the creation and management of slices on worker clusters. It is strongly recommended to use a released version. Follow the instructions provided in this document.

Building and Deploying kubeslice-controller in a Local Kind Cluster

For more information, see getting started with kind clusters.

Prerequisites

• Docker installed and running in your local machine
• A running kind cluster
• kubectl installed and configured
• Follow the getting started from above to install kubeslice-controller and worker-operator.

Setting up Your Helm Repo

If you have not added avesha helm repo yet, add it.

helm repo add avesha https://kubeslice.github.io/charts/

Upgrade the avesha helm repo.

helm repo update

Build Your Docker Image

Latest docker image - kubeslice-controller

1. Clone the latest version of kubeslice-controller from the master branch.

git clone https://github.com/kubeslice/kubeslice-controller.git
cd kubeslice-controller

1. Adjust image name variable IMG in the Makefile to change the docker tag to be built. Default image is set as IMG ?= aveshasystems/kubeslice-controller:latest. Modify this if required.

make docker-build

Running Local Image on Kind Clusters

1. Loading kubeslice-controller image into your kind cluster (kind). If needed, replace aveshasystems/kubeslice-controller with your locally built image name in the previous step.

• Note: If you use a named cluster, you must specify the name of the cluster you wish to load the images into. See loading an image into your kind cluster.

kind load docker-image aveshasystems/kubeslice-controller --name cluster-name

Example

kind load docker-image aveshasystems/kubeslice-controller --name kind

1. Check the loaded image in the cluster. Modify node name if required.

• Note: kind-control-plane is the name of the Docker container. Modify if needed.

docker exec -it kind-control-plane crictl images

Deploying in a Cluster

1. Create chart values file yourvaluesfile.yaml. Refer to values.yaml on how to adjust this and update the kubeslice-controller image to the local build image.

From the sample:

kubeslice:
---
---
   controller:
   ---
   ---
      image: aveshasystems/kubeslice-controller
      tag: 0.1.1

Change it to:

kubeslice:
---
---
   controller:
   ---
   ---
      image: <my-custom-image> 
      tag: <unique-tag>

1. Deploy the updated chart.

   make chart-deploy VALUESFILE=yourvaluesfile.yaml
   

Verify if the Operator is Running

kubectl get pods -n kubeslice-controller

Sample output to expect

NAME                                            READY   STATUS    RESTARTS   AGE
kubeslice-controller-manager-5b548fb865-kzb7c   2/2     Running   0          102s

Uninstalling the kubeslice-controller

For more information, see uninstalling KubeSlice.

make chart-undeploy

License

Apache License 2.0