How to Configure Persistent Volumes with Kubernetes

Kubernetes persistent volume is an excellent framework for providing data storage options for specific stateful applications. They separate the implementation of a storage system from the way your pods use it. Best cloud services vendors must keep the data locally, on a network share, or in a block storage volume in a persistent volume.

Persistent volumes take care of the difficulties associated with storing persistent data, such as databases and logs in Kubernetes. Pod-running containers have stateless, temporary filesystems. Although the files in your applications’ containers are readable and writable. So be ready for the loss of any changes if the pod restarts or terminates.

In this article, you will learn about persistent volumes, their significance, and how to use them in your cluster. Additionally, you’ll see some typical administration commands for utilizing kubectl to communicate with persistent volumes.

Why Persistent Volume?

Without a particular configuration, you cannot start up a database server in a Kubernetes pod because of the transient nature of container filesystems. If not, you would lose your data immediately after the pod reboots. Additionally, because all the files are in a single container, you wouldn’t be able to scale the database deployment.

You can resolve this problem with persistent volumes. They are constructed on top of the more basic volume system, offering a shared storage space accessible to all pod containers. Kubernetes can restore volumes even after the crash and restart of a single container.

Also, a higher abstraction known as persistent volumes separates storage from the pods that consume it. Furthermore, a persistent volume is shareable by several pods, has its lifespan, and stores data at the cluster or namespace level.

Pods employ persistent volumes, but they never truly belong to them. Even after all pods that reference it has been removed, the volume and its contents will still be accessible in the cluster, allowing it to be reattached to new, upcoming pods.

When To Use Persistent Volumes?

It would help if you utilized persistent volumes with Kubernetes anytime you have data that needs to survive individual pods. It’s typically desirable to store data in a persistent volume unless it’s transient or unique to a particular container.

Here are some everyday use cases:

  • Database: Data in a database should always be present on a persistent volume so that it can survive the server’s container lifecycle. Do not delete the user’s data every time the pod restarts.
  • Logs availability: Writing container log files to a persistent volume ensures they’ll be available after a crash or termination. If they weren’t written to a persistent drive, the collision will erase the logs that would have assisted you in troubleshooting the problem.
  • Protection of data: You can prevent unintentional data erasure with persistent volumes. They contain safeguards that prohibit taking away volumes that are being used by pods right now.
  • Independent data: Persistent volumes make sense when your data is crucial to your cluster. You can manage backups, performance, and storage capacity allocations more easily with their help because they give you the capability to manage data independently of application containers.

Creating a Persistent Volume

There are two ways of creating persistent volumes. You can make it statically and dynamically as well. When you add the volume manually to a cluster, it gives birth to a statically created volume. A dynamically constructed volume is designed as a result when a volume that doesn’t exist is referenced. You’ll learn how to use kubectl for creating a volume with the help of a static method.

For that purpose, in the beginning, you will need a YAML file specifically for the persistent volume:

creating persistent volume

This file defines as a straightforward persistent volume with a 1Gi capacity. A few other configuration settings define the volume’s provisioning and access.

Access Modes

The accessModes field describes which nodes and pods can access the volume:

  • ReadWriteOnce implies that all the pods present in a single mode are capable of reading and writing data.
  • ReadOnlyMany and ReadWriteMany apply to the pods on multiple nodes that have the ability to read-only and read-write access both.
  • ReadWriteOncePod is a new mode introduced in Kubernetes v1.22 that allows read-write access by a single pod on a single node.

Volume Mode

A volumeMode of Filesystem is a preselected and default option that usually comes with the desired behavior. This indicates that the volume will be mounted into each pod’s filesystem as a directory. The volume is shown as a raw block storage device without a pre-configured filesystem when Block is set to its alternative value.

Storage Classes

The storageClassName is an integral aspect of the configuration of persistent volume. The hosting environment for your cluster determines the storage classes you can use. If you are using Google Kubernetes Engine (GKE) to run your cluster, the standard class displayed here is accessible.

For clusters running Microsoft Azure Kubernetes Service (AKS), you can use azurefile-csi or do-block-storage with Digital Ocean Managed Kubernetes. When you supply your installation, you can create a storage class that uses your local discs if you manage your own cluster.

When you establish your persistent volume, an error will be generated if you attempt to utilize a storage class that is not supported by your environment,

Adding Your Volume to Your Cluster

Use kubectl option for adding the new persistent volume to a cluster:

>_$ kubectl apply -f pv.yaml

When running this command, you might see the following error message:

The PersistentVolume “example-pv” is invalid: spec:

Required value: must specify a volume type

This typically happens when the underlying storage class creates your storage using a provisioner. The cloud provider is not allocating storage that is not actively used in your cluster. In this case, you want to employ dynamic volume creation to construct a persistent volume with Kubernetes whenever required instantly. The subsequent section goes over this.

Linking Persistent Volumes to Pods

A persistent volume claim connects persistent volumes to pods. A claim is a request from a pod to read and write files from a specific volume.

Persistent volume claims are stand-alone objects. Below is an example of an old volume:

linking persistent volume to pods - image 1

The volumeName field implies an older version of persistent volume. The pod will have access to the example-pv volume when you link this claim to a pod. The storageClassName field is empty, so use the persistent volume’s definition’s storage class in the claim instead.

Instead of referencing previous volumes, persistent volume claims may implicitly construct new ones. You should include information about the volume in the claim’s specification. The above-mentioned dynamic volume production technique is:

Kubernetes persistent volume

The storageClassName and accessModes fields in the claim add to allow for volume configuration. The parameter specifies the volume’s storage capacity. Please be aware that this format differs slightly from that of a standalone persistent volume.

You can also apply the same claim for the cluster with the help of kubectl:

>_$ kubectl apply -f pvc.yaml

persistentvolumeclaim/example-pvc created

Even if the stand-alone volume creation failed with an error, the claim creation should succeed, assuming that you specified a storage class that is accessible in your cluster. The storage class dynamically provisions for new persistent volumes with Kubernetes that will meet the claim.

Finally, you can link the claim to your pods using the volumes and volumeMount fields in the pod manifest:

Kubernetes Persistent Volumes

Then add the pod to your cluster:

>_$ kubectl apply -f pvc-pod.yaml

pod/pod-with-pvc created

Additionally, the pod’s spec. Volumes field makes reference to the persistent volume claim. This, however, creates a pod volume with the name pv that can be added to the manifest’s containers section and is mounted to /pv-mount.

Demonstrating Persistence

To verify this behavior and form, we will use the following example.

For an already created pod, fetch a shell:

>_$ kubectl exec –stdin –tty pod-with-pvc — sh

Now start by writing a file to the /pv-mount directory, which was for mounting the persistent volume:

>_$ echo “This file is persisted” > /pv-mount/demo

Then detach from the container:

>_$ exit

Use kubectl to delete the pod:

>_$ kubectl delete pods/pod-with-pvc

pod “pod-with-pvc” deleted

Reconstruct the pod using the YAML manifest again:

>_$ kubectl apply -f pvc-pod.yaml

pod/pod-with-pvc created

Introduce a shell in the new pod of a container and then try to read the file from /pv-mount/demo:

>_$ kubectl exec –stdin –tty pod-with-pvc — sh

$ cat /pv-mount/demo

This file persisted.

Also, the deletion of the first pod had no impact on the persistent volume’s content. Therefore, everything that has already been written is kept anytime it is installed into fresh pods.

Managing Persistent Volumes:

Kubectl allows you to retrieve a list of all the persistent volumes:

managing persistent volume kubectl

Similarly, you can view all your persistent volume claims:

managing persistent volume claims

It usually means that the storage class is still supplying storage for the volume if a volume or claim has Pending status. Using the describe command to inspect the object’s event history will allow you to determine what’s delaying the operation:

kubectl describe pve example-pvc

For editing the volumes along with claims, the best way to move forward is by modifying the YAML file and then reapplying it to the cluster with the help of kubectl:

>_$ kubectl apply -f changed-file.yaml

This automatically recognizes and applies the modifications you make using the declarative API model of Kubernetes. Run the edit command to launch your editor and open the object’s YAML if you use imperative commands instead. When you save and close the document, changes will apply consequently:

>_$ kubectl edit pvc example-pvc

Volume attributes such as access mode and storage class are also unchangeable. Whereas other fields, like the volume’s capacity, rely on the implementation, while most significant storage types support dynamic resizing, not all do. If unsure, go to the documentation provided by your Kubernetes provider.

Also, manually adding a persistent volume claim does not change dynamically produced persistent volume objects. Edit the claim’s properties instead.

But, with the help of the delete command, you can remove the volume along with a claim:

kubectl delete command

So, by doing this, your clear out and remove the provider’s supplied storage. Unless you perform separate backups, the data contained in the volume will not be recoverable. Hence, avoid deleting volumes that were dynamically provided by a storage class. Moreover, the storage class will handle the persistent volume object for you.


Using persistent volumes, you can also store data independently on the pods in your Kubernetes cluster. Moreover, storage classes that interact with the various local disc, cloud storage, and network share types support them. Similarly, the way that pods access and uses the volume is always separate from these implementation specifics.

We also discussed in this article the use cases for persistent volumes with Kubernetes, how they differ from standard volumes and the procedures for adding and utilizing persistent volumes and persistent volume claims within your Kubernetes cluster. Additionally, you saw a few kubectl utility commands that can communicate with your volumes. With them, you may deploy stateful applications in Kubernetes without worrying about potential data loss due to container restarts.

Let's make it happen

We love fixing complex problems with innovative solutions. Get in touch to let us know what you’re looking for and our solution architect will get back to you soon.