Readme.md

NFS Dynamic Storage Provisioner

❗❗ This is now deprecated in favor of NFS External Storage Provisioner

Introduction

This Provisions an NFS Server inside of your cluster. The twist here is that it deploys it strictly on your master node which is already holding your etcd data anyway. Ultimately, this means than when PVCs are requested, this provisioner creates a directory under /storage/dynamic on your master node and stores whatever data it gets over the network. This makes the storage Node Failure Resistant i.e. even if the Node completely crashes or is completely replaced, you will not lose data. However, the crashing of the master node itself will result in permanent data loss. That said, crashing the master node is virtually the same as crashing the cluster. So this just means you should backup the master node regularly

Installation

Simply run

kubectl apply -k nfs-storage

Taints and Tolerations

The deployment definition uses tolerations to deploy the storage on the master node depending on the configuration of your cluster you might need to change them

The current tolerations are based on rancher generated clusters (rancher or rke) so if the cluster was deployed differently you might need to first find out the taints on the master node. e.g.

You can check the taints on the master node with the following

# Get all nodes
kubectl get nodes

# Describe Master node
kubectl describe node $MasterNode

In the case of rancher you'll see something like

Taints:
    node-role.kubernetes.io/etcd=true:NoExecute
    node-role.kubernetes.io/controlplane=true:NoSchedule

This translates to the following tolerations in the deployment.yaml file

tolerations:
    - key: node-role.kubernetes.io/etcd
      operator: Equal
      value: "true"
      effect: NoExecute
    - key: node-role.kubernetes.io/controlplane
      operator: Equal
      value: "true"
      effect: NoSchedule

So if your setup is different then you will need to change it accordingly