How to handle docker deprecation after eks 1.24 upgrade?

Question

As dockershim has been deprecated in eks 1.24 and containerd has become runtime, and I didn't completely understand the deprecation concept and upgraded my eks cluster (with managed nodegroups) from 1.23 to 1.24.

Since then, any new nodes joining the cluster are having docker daemon inactive, and all of my deployments are failing with an error: "Cannot connect to the Docker daemon at unix:///var/run/docker.sock. Is the docker daemon running?".

Now, what can be done exactly to fix this? Do I need to change all of my deployment related docker files or Can I configure something that the cluster still uses docker daemon?

Answer

Hello There,

As a first step, you could look if you are affected by the removal of dockershim in [this documentation](https://kubernetes.io/docs/tasks/administer-cluster/migrating-from-dockershim/check-if-dockershim-removal-affects-you/).

Considering you are already affected by this post migration and based on your question you probably have some node startup scripts that use docker commands, you can now look at the [guidance in this documentation](https://kubernetes.io/blog/2022/03/31/ready-for-dockershim-removal/#i-have-a-docker-dependency-what-now) to migrate away from using docker commands which won't work going forward.

There is also [this very good blog post](https://aws.amazon.com/blogs/containers/all-you-need-to-know-about-moving-to-containerd-on-amazon-eks/?utm_content=bufferc8d6f&utm_medium=social&utm_source=linkedin.com&utm_campaign=buffer) detailing all the necessary step to be containerd ready on amazon EKS.

Please let me know in case of any further queries.

Thanks,
Manish

How to handle docker deprecation after eks 1.24 upgrade?

相关内容