Kubernetes - NFS mount example

By default, containers are stateless meaning they're not storing any data besides the image itself on disk.

This behavior can be quite unfortunate when running workloads that require working with data such as databases. One of the easiest ways is to mount an NFS share to the container(s) and let them store everything on the NFS share so here's an example.

First, prepare the NFS Share:

 [root@nfssrv ~]# echo '/var/nfs/test 172.31.10.0/24(rw,secure,sync,no_root_squash)' >> /etc/exports  
 [root@nfssrv ~]# exportfs -rav  

Create the manifest for the Persistent Volume:

 ---  
 apiVersion: v1  
 kind: PersistentVolume  
 metadata:  
   name: var-nfs-test  
 spec:  
   capacity:  
     storage: 512Mi  
   volumeMode: Filesystem  
   accessModes:  
     - ReadWriteMany  
   persistentVolumeReclaimPolicy: Recycle  
   storageClassName: nfs  
   mountOptions:    
     - vers=4.2  
     - namlen=255  
     - proto=tcp  
     - timeo=30  
     - retrans=3  
   nfs:  
     path: /var/nfs/test  
     server: nfssrv.archyslife.lan  

Create the manifest for the Persistent Volume Claim:

 ---  
 apiVersion: v1  
 kind: PersistentVolumeClaim  
 metadata:  
   name: nfs-mount  
 spec:  
   accessModes:  
     - ReadWriteMany  
   storageClassName: nfs  
   resources:  
     requests:  
       storage: 512Mi  
   volumeName: var-nfs-test  

Create the manifest for the deployment:

 ---  
 apiVersion: apps/v1  
 kind: Deployment  
 metadata:  
   name: nfspods  
   labels:  
     app: nfstest  
   namespace: default  
 spec:  
   replicas: 2  
   selector:  
     matchLabels:  
       app: nfstest  
   strategy:  
     rollingUpdate:  
       maxSurge: 50%  
       maxUnavailable: 50%  
     type: RollingUpdate  
   template:  
     metadata:  
     labels:  
       app: nfstest  
     spec:  
       containers:  
         - name: nfstest  
           image: alpine:latest  
           command: ["/bin/sh"]  
           args: ["-c", "while true; do echo $(date +%H:%M:%S) $(hostname -f) >> /mnt/$(hostname -f); sleep 1; done"]  
           volumeMounts:  
             - mountPath: /mnt  
               name: nfs-volume  
       volumes:  
         - name: nfs-volume  
           persistentVolumeClaim:  
             claimName: nfs-mount  

Log in to the NFS server and check the contents of the file:

 [root@nfssrv ~]# tail -n 10 /var/nfs/test/*                                 
 ==> /var/nfs/test/nfspods-5d6ffbcc87-chw8h <==  
 17:32:08 nfspods-5d6ffbcc87-chw8h  
 17:32:09 nfspods-5d6ffbcc87-chw8h  
 17:32:10 nfspods-5d6ffbcc87-chw8h  
 17:32:11 nfspods-5d6ffbcc87-chw8h  
 17:32:12 nfspods-5d6ffbcc87-chw8h  
 17:32:13 nfspods-5d6ffbcc87-chw8h  
 17:32:14 nfspods-5d6ffbcc87-chw8h  
 17:32:15 nfspods-5d6ffbcc87-chw8h  
 17:32:16 nfspods-5d6ffbcc87-chw8h  
 17:32:17 nfspods-5d6ffbcc87-chw8h  
 ==> /var/nfs/test/nfspods-5d6ffbcc87-vd8rm <==  
 17:32:08 nfspods-5d6ffbcc87-vd8rm  
 17:32:09 nfspods-5d6ffbcc87-vd8rm  
 17:32:10 nfspods-5d6ffbcc87-vd8rm  
 17:32:11 nfspods-5d6ffbcc87-vd8rm  
 17:32:12 nfspods-5d6ffbcc87-vd8rm  
 17:32:13 nfspods-5d6ffbcc87-vd8rm  
 17:32:14 nfspods-5d6ffbcc87-vd8rm  
 17:32:15 nfspods-5d6ffbcc87-vd8rm  
 17:32:16 nfspods-5d6ffbcc87-vd8rm  
 17:32:17 nfspods-5d6ffbcc87-vd8rm  

Seems like everything worked as expected.

Deleting the objects you created can be done by referencing the files:

 [archy@workstation ~]$ kubectl -n default delete -f deployment.yml  
 deployment.apps "nfspods" deleted  
 [archy@workstation ~]$ kubectl -n default delete -f pvc.yml  
 persistentvolumeclaim "nfs-mount" deleted  
 [archy@workstation ~]$ kubectl -n default delete -f pv.yml  
 warning: deleting cluster-scoped resources, not scoped to the provided namespace  
 persistentvolume "var-nfs-test" deleted  

The warning above is perfectly normal as Persistent Volumes are not bound to a specific namespace and can be accessed by the whole cluster.

Feel free to comment and / or suggest a topic.