Working with K8s: Shared storage scenarios

Introduction

Your application will have different storage requirements based on problem being solved. I will not be discussing storage options, storage classes as there is a lot of information available. Rather I will discuss about storage usage scenarios

• Your container needs dedicated storage
• Two or more containers needs to share storage
• Storage is shared between containers and external users. This is typically used when end user is required to inject files for processing

Container dedicated storage

This is pretty standard stuff and can be achieved using creating a PV and PVC. The PVC can then be specified in your pod specification.

Shared storage between Containers

Sharing non-persistent storage

Here is a nice example of how to share non-persistant data between containers https://www.stratoscale.com/blog/kubernetes/kubernetes-how-to-share-disk-storage-between-containers-in-a-pod/#:~:text=To%20share%20data%20between%20such,the%20pod%20that%20encloses%20it.

Sharing persistent storage

You would start by defining a PVC first

kind: PersistentVolumeClaim
apiVersion: v1
metadata:
  name: my-pvc
spec:
  accessModes:
    - ReadWriteMany
  storageClassName: myvolume
  resources:
    requests:
      storage: 1Gi

You can then mount this on multiple pods to share data

pod-1

apiVersion: v1
kind: Pod
metadata:
  name: my-app1
spec:
  containers:
...
  volumeMounts:
	- mountPath: /data
      name: data
      subPath: app1
  volumes:
    - name: data
      persistentVolumeClaim:
        claimName: 'my-pvc'

pod-2

apiVersion: v1
kind: Pod
metadata:
  name: my-app2
spec:
  containers:
...
  volumeMounts:
    - mountPath: /data
      name: data
      subPath: app2
  volumes:
    - name: data
      persistentVolumeClaim:
        claimName: 'my-pvc'

Shared storage between Containers and external users

You would follow the same approach as in “Sharing persistent storage”, but ensure that the persistent storage supports network filesystem like NFS, GlusterFS. External entitoes then can connect to NFS or GlusterFS using appropriate clients or linbraries.

NFS

From https://kubernetes.io/docs/concepts/storage/volumes/#nfs

An nfs volume allows an existing NFS (Network File System) share to be mounted into your Pod. Unlike emptyDir, which is erased when a Pod is removed, the contents of an nfs volume are preserved and the volume is merely unmounted. This means that an NFS volume can be pre-populated with data, and that data can be “handed off” between Pods. NFS can be mounted by multiple writers simultaneously.

Here is an example of NFS PV

apiVersion: v1
kind: PersistentVolume
metadata:
  name: my-nfs-pv
spec:
  capacity:
    storage: 5Gi
  accessModes:
    - ReadWriteMany
  mountOptions:
    - hard
    - nfsvers=4.1
  nfs:
    server: 172.17.0.2
    path: /data

Here is an example of PVC to use NFS server

apiVersion: v1
kind: PersistentVolumeClaim
metadata:
  name: my-nfs-pvc
spec:
  accessModes:
    - ReadWriteMany
  volumeName: my-nfs-pv
  resources:
    requests:
      storage: 1Mi

You can omit the volumeName: my-nfs-pv if you wish to let Openshift pick up PV dynamically based on selection criteria like request storage, access mode, labels etc.

Here is a nice example of how to run NFS within k8s and directly mount NFS share within pod without using PV/PVC

https://matthewpalmer.net/kubernetes-app-developer/articles/kubernetes-volumes-example-nfs-persistent-volume.html

If you like this, please follow me and give a clap.