LAB11a: Dynamic Provisioning with StorageClass

LAB 50: Dynamic Provisioning with Storageclass

Let’s craft a clean, end-to-end guide that takes you from:

1. Spinning up an NFS server in Docker

2. Creating NFS-based PersistentVolume (PV) + StorageClass (SC)

3. Enabling Dynamic Provisioning using NFS Subdir External Provisioner

---

##PART 1 — Create NFS Server in Docker

1️ Run an NFS server container

You can use the popular itsthenetwork/nfs-server-alpine image:

mkdir -p /srv/nfs
docker run -d --name nfs-server \
  --privileged \
  -v /srv/nfs:/nfsshare \
  -p 2049:2049 \
  -e SHARED_DIRECTORY=/nfsshare \
  itsthenetwork/nfs-server-alpine:latest

Check container logs:

docker logs nfs-server

Find server IP:

docker inspect -f '{{range.NetworkSettings.Networks}}{{.IPAddress}}{{end}}' nfs-server

Suppose your NFS server IP is: 172.17.0.2.

---

PART 2 — Deploy NFS Subdir External Provisioner

---

2️ Create a Kubernetes namespace

kubectl create namespace nfs-provisioner

---

** 3 Create the NFS Provisioner Deployment**

Create a file: nfs-provisioner.yaml

apiVersion: apps/v1
kind: Deployment
metadata:
  name: nfs-subdir-external-provisioner
  namespace: nfs-provisioner
spec:
  replicas: 1
  selector:
    matchLabels:
      app: nfs-subdir-external-provisioner
  template:
    metadata:
      labels:
        app: nfs-subdir-external-provisioner
    spec:
      containers:
        - name: nfs-subdir-external-provisioner
          image: registry.k8s.io/sig-storage/nfs-subdir-external-provisioner:v4.0.2
          volumeMounts:
            - name: nfs-volume
              mountPath: /persistentvolumes
          env:
            - name: NFS_SERVER
              value: "172.17.0.2"   # Use your Docker NFS server IP
            - name: NFS_PATH
              value: "/nfsshare"
      volumes:
        - name: nfs-volume
          nfs:
            server: 172.17.0.2
            path: /nfsshare

Apply it:

kubectl apply -f nfs-provisioner.yaml

---

PART 3 — Create the StorageClass

4 Create StorageClass

Create file: nfs-storageclass.yaml

apiVersion: storage.k8s.io/v1
kind: StorageClass
metadata:
  name: nfs-client
provisioner: k8s-sigs.io/nfs-subdir-external-provisioner
parameters:
  archiveOnDelete: "true"
reclaimPolicy: Delete
mountOptions:
  - vers=4.1

Apply it:

kubectl apply -f nfs-storageclass.yaml

---

PART 4 — Test Dynamic Provisioning

---

5 Create a PVC

File: test-pvc.yaml

apiVersion: v1
kind: PersistentVolumeClaim
metadata:
  name: nfs-test-pvc
spec:
  storageClassName: nfs-client
  accessModes:
    - ReadWriteMany
  resources:
    requests:
      storage: 1Gi

Apply:

kubectl apply -f test-pvc.yaml
kubectl get pvc

You’ll see STATUS = Bound ✔️

---

6 Create a Pod that uses the PVC

File: test-pod.yaml

apiVersion: v1
kind: Pod
metadata:
  name: nfs-test-pod
spec:
  containers:
    - name: app
      image: busybox
      command: ["sh", "-c", "echo hello-from-nfs > /data/hello.txt && sleep 3600"]
      volumeMounts:
        - name: nfs-vol
          mountPath: /data
  volumes:
    - name: nfs-vol
      persistentVolumeClaim:
        claimName: nfs-test-pvc

Apply:

kubectl apply -f test-pod.yaml

Then check on the NFS server host:

ls /srv/nfs/

You’ll see a shiny new folder named something like:

default-nfs-test-pvc-xxxx/

Inside it:

hello-from-nfs

---

Summary

Component	Purpose
Docker NFS Server	Provides shareable storage
NFS Subdir External Provisioner	Auto-creates folder per PVC
StorageClass	Defines dynamic provisioning rules
PVC	Requests storage dynamically
Pod	Uses that dynamically created storage

Test - Filling up Storage Quota

---

** YAML: storage-filler.yaml**

apiVersion: v1
kind: Pod
metadata:
  name: storage-filler
spec:
  containers:
    - name: filler
      image: busybox
      command: ["/bin/sh", "-c"]
      args:
        - |
          echo "Starting slow storage fill..."
          i=1
          while true; do
            echo "Creating file $i..."
            dd if=/dev/urandom of=/data/file_$i.bin bs=100M count=1 2>&1
            if [ $? -ne 0 ]; then
              echo "❌ Write failed — storage is FULL!"
              break
            fi
            i=$((i+1))
            sleep 5
          done
          echo "Sleeping so pod stays alive..."
          sleep 3600;
      volumeMounts:
        - name: vol
          mountPath: /data
  volumes:
    - name: vol
      persistentVolumeClaim:
        claimName: your-pvc-name

Replace:

your-pvc-name

with your real PVC name (the one that is 1 GB).

---

How to Run This

Apply PVC (example 1 GB)

apiVersion: v1
kind: PersistentVolumeClaim
metadata:
  name: test-pvc
spec:
  storageClassName: nfs-client
  accessModes:
    - ReadWriteMany
  resources:
    requests:
      storage: 1Gi

Apply:

kubectl apply -f pvc.yaml
kubectl apply -f storage-filler.yaml

---

What You Will Observe When PVC Becomes Full

Your cluster will begin telling tales:

1. dd write error

Inside container logs:

dd: writing 'file_11.bin': No space left on device
❌ Write failed — storage is FULL!

2. Events on Pod

kubectl describe pod storage-filler

You may see:

No space left on device
Failed to write
Input/output error

3. Application behavior

• The pod will not crash, but the write loop exits

• Your service may exhibit read-only behavior

• Many apps crash when persistent disk becomes read-only

4. NFS / host logs

Depending on NFS backend, you may see disk pressure warnings.

---

Want the storage to fill even slower?

Change the speed:

Write 50 MB every 10 seconds:

dd if=/dev/urandom of=/data/file_$i.bin bs=50M count=1
sleep 10

Reference

1. BLOG 50: Available StorageClass Provisioner (HackMD) (local blog50-storageclass.md)

2. Storage Reclaim Policy