Openshift/K8S: How to create an NFS Persistent Volume Claim for your Containers

Unlike virtualization, in modern containerization technologies Kubernetes or Openshift doesn’t include embedded storage with containers. Altough a standard vm contains embedded storage due to requirement of operation systems, already comes with virtual disks and it’s storage space.

In Kubernetes, Openshift and other variants, as an Administrator, you need to preconfigure the storage for for your containers/pods. You or your administrator needs to provide your node to persistent storage space(s). This spaces can be traditional block storage (iscsi, fc), file storage (cifs,nfs) or cloud storage (s3,azure files,ebs).

Whatever technology you use, you need a Persistent Volume Claim (PVC) to provide persistent storage space to containers.

As a first task, I recommend you to create a resource quota for your PVC to limit exhausting storage space. I’m going to tell openshift to limit “the project” or “namespace” to use total 500GB storage space and maximum allowed PVC count to limit 3.

[ozgurkkisa@workstation ~]$oc create quota storage \
  --hard=requests.storage=500G,persistentvolumeclaims=3

In other words, an Admin or a Developer can create max 3 PvC and total 500 GB space for projects containers. So they would be create space for their containers, for instance pvc1 = 150 GB, pvc2 = 150 GB, pv3 = 200 GB.

At first, create an empty file with .yml extension. I named my file as pvc.yml:

[ozgurkkisa@workstation ~]$ touch pvc.yml

Open file that you created with your favorite text editor. For example nano or vim. Then write or paste the content below :

apiVersion: v1
kind: PersistentVolumeClaim
metadata:
  name: Apache-www
spec:
  accessModes:
    - ReadWriteMany
  storageClassName: nfs-storage
  resources:
    requests:
      storage: 300Gi

In this example, I created a pvc named Apache-www, storage space as 300 GB and access mode as ReadWriteMany.

After you save your yml file, activate it by typing :

[student@workstation ~]$ oc apply -f pvc.yml
persistentvolumeclaim/Apache-www created

Tip:If you made a mistake or simply delete the pvc that you create use :

oc delete persistentvolumeclaim Apache-www
persistentvolumeclaim "Apache-www" deleted

After applying change you may check your new PvC to by typing this command:

[student@workstation ~]$ oc get persistentvolumeclaims
NAME         STATUS  VOLUME   CAPACITY  ACCESS MODES  STORAGECLASS  AGE
Apache-www  Bound   pvc-...  300Gi       RWX           nfs-storage  15s

Our PvC is now ready! So we can use it in any container in project or namespace. To defining a PvC for container/pod, you have to put required info in yaml file of container application. For example, let’s say we have an Apache application named httpd. We want to put our web site/application files to persistent storage. We have to provide the additional info in provisioning file :

apiVersion: apps/v1
kind: Deployment
metadata:
   name: httpd
spec:
   replicas: 1
   selector:
      app: httpd
   template:
      metadata:
      labels:
      app: httpd
   spec:
      containers:
      - image: registy.access.redhat.com/rhscl/httpd-24-rhel7
        imagePullPolicy : Always
        name: httpd
        ports:
        - containerport: 8080
          protocol: TCP
        - containerPort: 8443
          protocol: TCP
        VolumeMounts:
        - mountPath: /var/www/html
          name: html
        restartPolicy: Always
        volumes: 
        - name: html
          persistenVolumeClaim:
             Apache-www

Be aware that marked with yellow lines. These are volume spesific data that bind persistent storage to our container/pod. In the sample above, we simply add persistent volume claim named Apache-www and it’s related space to containers /var/www/html directory.

You can save the file as apache.yml. After required info added and saved to file you need to create/redeploy your apache instance by running command :

[ozgurkkisa@workstation ~]$ oc apply -f apache.yml

After a while, your new/existing pods would be created or redeployed. Wher your pods crashed or stopped, your data would be persisted in the storage.

Hope to see in new articles and tips soon.

With my best wishes 😀