In addition to runc, I'd like to point out an alternative OCI runtime implementation, crun (https://github.com/containers/crun). You can play with both either directly, or through Podman (https://podman.io/)

Useful for cgroups v2 too.

Unfortunately on mobile the zoom is fixed (I can't zoom out, didn't know that was possible) and I can't see the left and right edges of the text.

    javascript:document.querySelector('meta[name=viewport]').setAttribute('content','width=device-width,initial-scale=1.0,maximum-scale=10.0,user-scalable=1');

Are there any runc shims that just use processes (I know, containers are just processes) ignoring network/user/etc namespace isolation and other Linux-specific security features? For example a shim that could run native MacOS processes on MacOS, native FreeBSD binaries on FreeBSD, etc. just by executing the processes directly.

The point of this would be to take advantage of the Docker ecosystem for _scheduling_ particularly in developer environments. Specifically I'd like a "docker-compose for processes" that can run on any system and just handles scheduling multiple processes together but without requiring root access to modify init scripts or systemd services at the system level.

wget directly to /usr/bin. Am I the only one who cringes upon such a pattern? I am probably too old. I recently almost doubled over when I saw that /sbin is now a symlink to /usr/sbin on bullseye. Even worse, /lib/modules is a symlink to /usr/lib/modules. Try $ find /lib -name \mlx5\ and learn how find treats symlinks.