nfs

Use NFS Server Provisioner on AKS

We already have managed NFS service solution on AKS: Use Azure NetApp Files with AKS. This page shows a cost effective NFS solution(unamanged) on AKS.

Introduction

NFS Server Provisioner is an out-of-tree dynamic provisioner for Kubernetes. You can use it to quickly & easily deploy shared storage that works almost anywhere. This doc shows how to set up NFS Server Provisioner on AKS using NFS Server Provisioner helm chart, the NFS server data is stored on single Azure managed disk. To get high availability, you may use Rook, step-by-step doc: Using Rook/Ceph with PVCs on AKS or Azure NetApp Files with AKS - static provisioning

1. Install nfs-server-provisioner helm chart

• following example would provision 100GB storage(one data disk) on an agent node, serving as a NFSv3 server

helm repo add stable https://kubernetes-charts.storage.googleapis.com/
helm install stable/nfs-server-provisioner --generate-name --set=persistence.storageClass=default,persistence.enabled=true,persistence.size=100Gi

2. After installation successfully, a new storage class nfs created

# kubectl describe sc nfs
Name:                  nfs
IsDefaultClass:        No
Annotations:           <none>
Provisioner:           cluster.local/nfs-server-provisioner-1587007822
Parameters:            <none>
AllowVolumeExpansion:  True
MountOptions:
  vers=3
ReclaimPolicy:      Delete
VolumeBindingMode:  Immediate
Events:             <none>

Example#1: create a statefulset with NFS volume mount

kubectl create -f https://raw.githubusercontent.com/Azure/kubernetes-volume-drivers/master/nfs/statefulset-nfs.yaml

• enter the pod container to validate

# kubectl exec -it statefulset-nfs-0 bash
root@statefulset-nfs-0:/# df -h
Filesystem                                                    Size  Used Avail Use% Mounted on
overlay                                                        97G   11G   87G  11% /
tmpfs                                                          64M     0   64M   0% /dev
tmpfs                                                         3.4G     0  3.4G   0% /sys/fs/cgroup
10.0.212.68:/export/pvc-c08bb76e-6d45-452a-8333-53b13bd01000   99G   60M   99G   1% /mnt/nfs
/dev/sda1                                                      97G   11G   87G  11% /etc/hosts
...

Example#2: set up wordpress based on NFS server

helm repo add bitnami https://charts.bitnami.com/bitnami
helm install --set persistence.storageClass="nfs" --generate-name bitnami/wordpress

• check wordpress pods status

# kubectl get po
NAME                                    READY   STATUS    RESTARTS   AGE
nfs-server-provisioner-1587040611-0     1/1     Running   0          15m
wordpress-1587040914-5659b7f8db-xl44v   1/1     Running   1          10m
wordpress-1587040914-mariadb-0          1/1     Running   0          10m

• nfs server performance in writing small files scenario

time ( wget -qO- https://wordpress.org/latest.tar.gz | tar xvz -C /mnt/nfs )
real    0m16.286s
user    0m0.429s
sys     0m1.214s