Share Docker Images Without Using the Docker Hub

1. Overview

Let’s suppose we need to share a Docker image that is present locally on our machine. To solve this problem, Docker Hub comes to the rescue.

Docker Hub is a cloud-based central repository where Docker images can be stored. So all we need to do is push our Docker image to the Docker Hub, and later, anyone can pull the same Docker image.

Being a cloud-based repository, Docker Hub requires an additional network bandwidth to upload and download the Docker images. Also, as the image size increases, the time needed to upload/download the image also increases. Hence, this method of sharing the Docker images is not always useful.

In this tutorial, we’ll discuss a way to share Docker images without using the Docker Hub. This approach proves out to be handy when sender and receiver are connected to the same private network.

2. Save Docker Image as a tar Archive

Suppose there is a Docker image baeldung which we need to transfer from machine A to machine B. To achieve this, first, we’ll convert the Docker image to a .tar file using the docker save command:

$ docker save --output baeldung.tar baeldung

The above command will create a tar archive named baeldung.tar. Alternatively, we can also use file redirection to achieve similar results:

$ docker save baeldung > baeldung.tar

The docker save command can create a single tar archive using multiple Docker images:

$ docker save -o ubuntu.tar ubuntu:18.04 ubuntu:16.04 ubuntu:latest

3. Transfer the tar Archive

The tar archive that we created is present on machine A. Let’s now transfer the baeldung.tar file to machine B. We can use the protocols like scp or ftp.

This step is highly flexible and depends significantly on the environment where machine A and machine B are present.

4. Load tar Archive into the Docker Image

So far, we have created the tar archive of the Docker image and moved it to our target machine B.

Now, we’ll create the actual Docker image from the tar archive baeldung.tar using the docker load command:

$ docker load --input baeldung.tar 
Loaded image: baeldung:latest

Again, we can also use redirection from the file to convert the tar archive:

$ docker load < baeldung.tar
Loaded image: baeldung:latest

Let’s now verify whether the image is successfully loaded by running the docker images command:

$ docker images
baeldung                                        latest                            277bcd6563ce        About a minute ago       466MB

Note that if the Docker image, baeldung, is already present on the target machine (machine B in our example), then the docker load command will rename the tag of the existing image to an empty string <none>:

$ docker load --input baeldung.tar 
cfd97936a580: Loading layer [==================================================>]  466MB/466MB
The image baeldung:latest already exists, renaming the old one with ID sha256:
  277bcd6563ce2b71e43b7b6b7e12b830f5b329d21ab690d59f0fd85b01045574 to empty string

5. Drawbacks

Using this approach, we lose the freedom to reuse the cached layers of the Docker image. So, each time we run the docker save command, it’ll create the tar archive of the entire Docker image.

Another drawback is that we need to maintain the Docker image versions manually by saving all the tar archives.

Hence, it is recommended to use this approach in the testing environment or when we have restricted access to Docker Hub.

6. Conclusion

In this tutorial, we learned about the docker save and docker load commands and how to transfer a Docker image using these commands.

We also went through the downsides involved and the ideal situations where this approach could prove out to be efficient.