Fixing Kubernetes Imagepullbackoff

If you have been working with Kubernetes for a long time, you have probably encountered the ImagePullBackOff condition. If you are unfamiliar with this problem, it can be frustrating. So in this article, you will lead you through the basics of this issue, how to troubleshoot it, what some typical reasons are, and where to begin if you face it.

What is the ImagePullBackOff Error?

The ImagePullBackOff issue is caused by your Kubernetes container runtime being unable to fetch the image from either a public or private container registry. Kubernetes will constantly pull the image with a growing backoff delay, as indicated by the Backoff component. With each try, Kubernetes will increase the delay until it meets the five-minute restriction.

It may seem like a broad statement to suggest that container runtime (whether Docker, container, or something else) fails to retrieve the image from the registry, but let’s look at the various causes that you can find in the next section.

The preceding sections will go over the various reasons why your pod can be in the ImagePullBackOff state when you start your container. You will also learn how to troubleshoot and resolve this dreaded error.

What Causes the ImagePullBackOff Error to Occur?

The following are some of the reasons why your pod may be stuck in the ImagePullBackOff state:

  • Image not available
  • The name or tag for the image is incorrect.
  • Private image is used, and there is a problem with authentication.
  • There is a difficulty with the network.
  • The name of the registry is inaccurate.
  • Rate limits for container registries
  • The pod does not have access to the image because it lacks the necessary credentials.
  • Limits on registry rates

How to Troubleshoot ImagePullBackOff?

Let’s look into a few of the probable causes listed in the bulleted list.

1. The container image is not available, or the name used is inaccurate
The problem is usually generated if there is a typo or the fact that the image pushed to the container registry is failed, but you’re referring to an image that is not there. Let’s attempt recreating this by making a pod with a fictitious image name. The following command achieves this.

$ kubectl run newapp --image=my_image/my_image:latest

As you can see, the pod is created.

If we try to get the details of the pod with the get pod command as you can see below.

$ Kubectl get pod

Here, it is shown that the image is not there and we are unable to pull it.

You can make use of the kubectl describe command for the purpose of discovering the root cause and finding more information about this problem. Because the command produces a lot of output, we will just display the sections that are pertinent to our discussion. The real error message is seen in the following output under Events in the Message column:

$ Kubectl describe pod newapp

Some sections of the produced result are as follows after executing the describe command.

2. Tag Doesn’t Exist
It is possible that the image tags you are trying to get has been retired, or that you typed the erroneous tag name. In some circumstances, your pod will become stuck in the ImagePullBackOff state once more, as shown in the code sample below. To reproduce this issue, we intentionally used the erroneous tag name, lates instead of latest.

$ kubectl run apptwo --image=nginx:lates

The above command has created the pod with the name that you have given.

After that, we get the details of the pod with the get pod command.

$ kubectl get pod

As a result, the image pulls fail.

Now, we are again using the describe command to understand the cause of this status.

$  kubectl describe pod apptwo

In this section of events, you can see the reason for the ImagePullBackOff error.

The reason is shown clearly here for your better understanding.

3. Wrong Credentials & Private Image Registry

Here, we are trying to reproduce the problem and for that, we ate spinning up a pod that tries to pull an image from a private registry.

$ kubectl run appthree --image=docker.io/hiyou/nameofimage

The above command gives the following result.

After that, we have executed the describe command.

The described command shows the overall details of the pod and also mentions the reasons behind the ImagePullBackOff error.

We have not added a secret to Kubernetes or included a reference to it in the pod definition. The pod will become stuck in the ImagePullBackOff state once more, and the notification verifies that access to the registry is denied:

You can create a secret with the kubectl command below to fix this error. The kubectl command is used then to create a secret for a Docker registry (private).

4. Registry Rate Limits

If you verify some of your credentials such as registry URL, details, and tag name, you may obtain ImagePullBackOff because of registry rate limits. You can now only pull 100 containers every six hours on Docker Hub. If you provide your login details, this will climb to 200 pulls every six hours. In a lively cluster with numerous frequently deployed Pods, that limit could be reached quickly.

You will require to wait till the cap is reached after a specific time limit. Kubernetes should now be able to successfully pull the image and start your Pods.

Consider using you are in-cluster registry along with a proxy to cache your relevant images. This can help you remain inside the rate constraints by reducing the number of times you hit Docker’s servers.


When a node fails to pull an image, Kubernetes Pods go into the ImagePullBackOff state. Kubelet will attempt the pull regularly, so temporary problems won’t require any manual intervention. This article discussed ImagePullBackOff and three potential sources of the problem. Although there may be several causes, reading the error message can quickly reveal the true cause of the problem. If you examine and follow the procedures above, fixing this problem should be simple.

About the author

Kalsoom Bibi

Hello, I am a freelance writer and usually write for Linux and other technology related content