Create and Mount a Persistent Volume Claim

Cloud applications often require a place to persist data. Containers inside pods have ephemeral filesystems that are lost when a pod restarts or terminates. Persistent volumes solve this problem by allowing data to persist beyond a pod's lifetime. This tutorial explains how to create a Persistent Volume Claim, mount it in a pod, and then demonstrate that it works across two replicas of a pod.

Make the Claim

First, here is the yaml the makes a claim to a volume. It will be dynamically created once it is mounted.

• The size of my volume below is set to 30Mi.
• The access mode for a volume intended to be shared between many pods should be set to ReadWriteMany.
• The volume mode is set to Filesystem, which mounts it as a directory in the pod's filesystem.

pvc-claim.yaml

apiVersion: v1
kind: PersistentVolumeClaim
metadata:
  name: pvc-nginx
  namespace: default
spec:
  resources:
    requests:
      storage: 30Mi
  volumeMode: Filesystem
  accessModes:
    - ReadWriteMany # access by many pods

Apply and check the PersistentVolumeClaim resource.

$ kubectl apply -f pvc-claim.yaml
persistentvolumeclaim/pvc-nginx created
$ kubectl get pvc
NAME        STATUS    VOLUME   CAPACITY   ACCESS MODES   STORAGECLASS   AGE
pvc-nginx   Pending                                                     14m
$ 

Mount It

Next, we'll mount the claim in a simple NGINX deployment. This will cause the volume to be created dynamically at this time. Things to note:

• We've specifically asked for 2 replicas so that we can demonstrate access to the volume from each replica.
• We've asked for the volume to be mounted in our pod's filesystem at /tmp/nginx-data.

pvc-mount.yaml

apiVersion: apps/v1
kind: Deployment
metadata:
  name: nginx
  namespace: default
spec:
  selector:
    matchLabels:
      app: nginx
  replicas: 2
  template:
    metadata:
      labels:
        app: nginx
    spec:
      containers:
      - name: nginx
        image: nginx
        ports:
        - containerPort: 80
        volumeMounts:
          - name: nginx-data
            mountPath: /tmp/nginx-data/
        resources:
          requests:
            cpu: 150m
            memory: 100Mi
          limits:
            cpu: 150m
            memory: 100Mi
      volumes:
        - name: nginx-data
          persistentVolumeClaim:
            claimName: pvc-nginx

Apply and check the deployment resource.

$ kubectl apply -f pvc-mount.yaml 
deployment.apps/nginx created
$ kubectl get pods -o wide
NAME                     READY   STATUS              RESTARTS   AGE   IP              NODE        NOMINATED NODE   READINESS GATES
nginx-65bbb4d46c-2rwl8   0/1     ContainerCreating   0          72s   <none>          nyc-adljs   <none>           <none>
nginx-65bbb4d46c-4hp4s   1/1     Running             0          72s   10.244.9.148    sfo-gwuie   <none>           <none>
nginx-65bbb4d46c-dng7q   1/1     Running             0          72s   10.245.160.39   nyc-adljs   <none>           <none>
nginx-65bbb4d46c-mg9pr   1/1     Running             0          71s   10.245.182.89   sfo-gwuie   <none>           <none>
$ 

Test It

Now, let's exec into a pod running in the NYC node. We'll place a file there, and then demonstrate that it is also available in the other pod in the NYC node.

$ kubectl exec -it nginx-65bbb4d46c-dng7q -- sh
# cd /tmp/nginx-data
# echo hi there > file.txt
# ls -l
total 4
-rw-r--r-- 1 nobody nogroup 9 Jan  4 22:09 file.txt
# exit
$

Let's now exec into the other pod running in the NYC node and check for that same file.

$ kubectl exec -it nginx-65bbb4d46c-2rwl8 -- sh
# cd /tmp/nginx-data
# ls -l
total 4
-rw-r--r-- 1 nobody nogroup 9 Jan  4 22:09 file.txt
# cat file.txt
hi there
# exit
$

Success!

Finally, in order to shut everything down, we first need to delete the dployment, so that no pods have the volume mounted. Then we need to delete the PVC itself.

$ kubectl delete deploy nginx
deployment.apps "nginx" deleted
$ kubectl delete pvc
persistentvolumeclaime/pvc-nginx deleted
$

Alternately, you can delete those resources from the Kubernetes dashboard as well.

And that's it!

Next, try installing Postgres on your PVC.