Podman support

[kdrag0n]

This is for first-class Podman support, similar to Podman Desktop. It's not planned, but I've opened this issue because it's been mentioned a few times. Please vote 👍 if you're interested.

For now, you can create a Linux machine and install Podman in it. Port forwards, bind mounts, etc. should still work just like normal.

If you're interested in this, please comment and share why. Thanks!

[redsnapper2006]

looking forward with this issue

[kdrag0n]

Would anyone interested in this care to share why you want Podman support? I'm not sure I understand how it's useful if Docker is already available.

[brandonzylstra]

Would anyone interested in this care to share why you want Podman support? I'm not sure I understand how it's useful if Docker is already available.

Podman is much more secure and also lighter weight.

[brandonzylstra]

Also, I don't know whether this is true or not, but I have heard it said that Podman is the only one that works on Red Hat.

(Not that I'm a fan of Red Hat, but in some situations I have no choice but to work on Red Hat servers.)

[kdrag0n]

Podman is much more secure and also lighter weight.

Lighter, yes, but I don't see how it could be more secure.

Also, I don't know whether this is true or not, but I have heard it said that Podman is the only one that works on Red Hat.

(Not that I'm a fan of Red Hat, but in some situations I have no choice but to work on Red Hat servers.)

That doesn't sound too likely, but even if it were true, it doesn't apply to OrbStack. Since Podman is meant to be compatible with Docker, you can develop locally with Docker and deploy with Podman.

[kdrag0n]

From what I've seen, most people ask for Podman support because they're already using Podman Desktop as a free alternative to Docker Desktop and have gotten in the habit of running Podman instead, in which case you can just alias podman=docker for the most part.

This might be completely off, so it'd be very helpful if more people could chime in. As it stands, it seems like adding Podman would just increase complexity and app size for no good reason because it does mostly the same thing as Docker but with fewer features (e.g. no BuildKit). There must be something I'm missing if so many people have upvoted the issue, and I'd like to understand the use cases and appeal better.

From OrbStack's perspective: We need to keep Docker around because Podman won't cover all use cases. And if Docker already does more than Podman, then why add a second container manager that has fewer features and otherwise does mostly the same thing? It increases complexity, app size, and may cause confusion for users not familiar with the whole Docker/Podman/container scene.

Yes, Podman is lighter and may use fewer resources, but what's the point if most people won't be using it?

[NobodyXu]

Lighter, yes, but I don't see how it could be more secure.

Podman doesn't need root permission to run (I heard that docker has an experimental feature for this?) and also support fuse overlayfs and userspace networking drivers.

[BigBadBassMan]

Podman allows running containers as a user (i.e. not as root, which is generally more secure), has tighter integration with SystemD (both for updating/starting containers and running SystemD inside a container) and is considered a stepstone to K8s/K3s due to its native export kube yaml feature.

I'm using Podman primarily to replicate and test SaltStack workflows with SystemD inside multiple containers - a feature not possible with Docker.

[zhanglolo]

looking forward to this feature.

[haliliceylan]

looking forward to this feature.

[iMonZ]

I would love seeing this

[timelordx]

Please, pretty please ...

[JPBarringer]

+1.

Unprivileged containers are a big plus. Also ecosystem reasons, including trust & support for CNCF / OCI tooling beyond Docker.