The following document describes the deployment of a self-bootstrapping, reliable, multi-node Redis on Kubernetes. It deploys a master with replicated slaves, as well as replicated redis sentinels which are use for health checking and failover.

This example assumes that you have a Kubernetes cluster installed and running, and that you have installed the kubectl command line tool somewhere in your path. Please see the getting started for installation instructions for your platform.

This makes use of a StorageClass, either create a storage class with the name of "ssd" or update the StatefulSet to point to to the correct StorageClass.

To get your cluster up and running simple run:

kubectl apply -Rf k8s

The cluster will automatically bootstrap itself.

Your pods may not show up in the dashboard. This is because we automatically add additional labels to the pods to recognize the master. To see the pods within the dashboard you should look at the redis-nodes service instead.