Ask HN: In 2024, become an expert in Docker or Podman?

[profwalkstr]

I'm a DevOps and need to up my container game. I plan to become an expert in Docker but Podman is a serious contender. It looks like Podman might have a bright future since Red Hat is investing a lot of money in it while Docker is still scratching their heads looking for a business model that the community won't reject. Podman seems to have a better financial future outlook since it's subsidized by Red Hat's OpenShift strategy.

Although I know Docker is the "standandard" and Podman sometimes has incompatibilities with Docker which it shouldn't have.

HN folks seem to prefer Podman instead of Docker.

What's the smartest choice here? Invest my time to become an expert in Docker or Podman?

[seabass-labrax]

I don't think the financial incentives are particularly relevant; it's the idea of containerization which is here to stay now. All across the different software stacks certain modifications are being made so that applications can be containerized. For instance, FreeBSD can now run OCI containers in FreeBSD Jails. OCI came from Docker, but FreeBSD Jails has no shared history with Docker.

If you look further, you can see similar trends to support this. Distributed computing is more relevant with projects like Kubernetes, and Erlang is having a revival. Likewise with functional programming languages: because you can't directly have long-lasting state inside a container, being able to create a pipeline of immutable containers is valuable.

All that is to say that looking at either Docker or Podman itself might be missing out on a lot of really exciting developments that aren't necessarily as easy to spot. Here's one of my favourites from 2023, dynamic resource control for containerized applications: https://archive.fosdem.org/2023/schedule/event/container_cgr...

I wish you the best of luck!

[wmf]

The concepts are the same and I'm not sure deep knowledge of any particular tool is that valuable. The future definitely seems to be containerd or Podman.

[brudgers]

HN folks seem to prefer Podman instead of Docker.

To the degree that's true, the preferences of HN folks probably isn't the best North Star for a career...it encourages the expression of contrarian opinions for the sake of expressing contrarian opinions and by design, HN fosters an endless september.

Out in the real world, Docker has better brand recognition and though there are arguments for Podman, who wants to have an argument? Sure some people want that, but it might not be a good substitute for getting paid.

Finally, no amount of study is going to bring the expertise that comes with a bit of on-the-job experience. So if you are going to study one, study the one that is most likely to be used on the job. And the one most likely to get you a job.

Compare:

  I don't know Podman, but I know a little Docker.

  I don't know Docker, but I know a little Podman.

Which would be better in an interview?

Good luck.

[idontwantthis]

From my experience, Podman is basically the same as Docker except for the stuff it’s still missing. Either one will teach you how to use containers but you may run into some walls with Podman. I’m not sure if it fully supports compose yet, for example.

[gtirloni]

You will barely spend any time switching between one or the other. The tool doesn't matter.

Learn deeply how cgroups, namespaces, etc, work and transfer you knowledge about the building blocks to anything new that shows up.

[mattpallissard]

This. Understanding what the kernel does and the behaviors maps to any technology. What happens when you hit memory pressure? What causes it? What things are global or local to your cgroup?

Being able to answer questions like that transcends your container runtime. At that point it's just reading the docs, looking for the the label on the knob you already know needs tweaking.

[samwho]

Maybe a silly question but what are you hoping to achieve by investing time in becoming an expert in a specific container technology?