Konubinix' site

How to Share Data Using a Local Kubernetes Cluster

Evergreen

how to share data using a local kubernetes cluster, using a custom PersistentVolume.

I will use the clk extension to bootstrap easily a k3d cluster.

clk parameter set k8s --distribution k3d
clk parameter set k8s.create-cluster --volume /tmp/k3d:/tmp/k3d

clk k8s flow

This basically gives the parameters --volume /tmp/k3d:/tmp/k3d to the clk k8s create-cluster command that in turn, will provide it to k3d cluster create.

Then, say that you are debugging a deployement that contain a PersistentVolumeClaim that is named my-claim in the default namespace.

The PersistentVolume to create is simply.

apiVersion: v1
kind: PersistentVolume
metadata:
  name: shared-volume
spec:
  accessModes:
  - ReadWriteOnce
  claimRef:
    name: my-claim
    namespace: default
  capacity:
    storage: 1Gi
  hostPath:
    path: /tmp/k3d/myvolume/
    type: DirectoryOrCreate
  nodeAffinity:
    required:
      nodeSelectorTerms:
      - matchExpressions:
        - key: kubernetes.io/hostname
          operator: In
          values:
          - k3d-k3s-default-server-0
  persistentVolumeReclaimPolicy: Delete
  storageClassName: local-path
  volumeMode: Filesystem

The nodeAffinity is not strictly needed as in the standard k3d cluster, you have only one node, but it might help remembering that you put this persistent volume explicitly on this node.

Then, appreciate the fact that the claim is correctly bound.

kubectl get pvc my-claim
NAME STATUS VOLUME CAPACITY ACCESS MODES STORAGECLASS AGE
my-claim Bound shared-volume 1Gi RWO local-path 2m31s

Then, everything that uses this volume can be seen and played with in the path /tmp/k3d.

Notes linking here