crun vs Rancher Desktop

crun is a free, self-hosted OCI container runtime for Linux, written in C. It conforms to the OCI Container Runtime specifications and provides CLI operations for creating, running, executing commands in, listing, pausing, resuming, updating, checkpointing and restoring containers. It can generate a configuration for unprivileged use and creates a user namespace when run by a non-root user. The project documents integration with Podman, an Ansible role for installing a statically linked binary and a Lua binding. Build instructions cover Fedora, RHEL/CentOS Stream 9 and 10, Ubuntu, Alpine and openSUSE Tumbleweed; release binaries are listed for Linux amd64 and arm64, including variants without systemd. Official release artifacts are signed, with GPG verification commands provided. The project reports a benchmark of 100 sequential `/bin/true` containers completing in 1.69 seconds with crun versus 3.34 seconds with runc on the author's machine. The manual marks cgroup v1 support as deprecated. Regular patch releases target the latest version; fixes for earlier versions are best effort and not guaranteed.

Rancher Desktop is a free, open-source desktop application for managing local containers and Kubernetes on macOS, Windows, and Linux. Users choose Moby/dockerd with the Docker CLI or containerd with nerdctl to build, push, pull, and run containers. Its built-in Kubernetes uses k3s, and Preferences let users select a Kubernetes version or reset Kubernetes and the container runtime. The graphical interface manages images, containers, settings, and local Kubernetes resources, with controls for networking, virtual-machine resources, and registry access. Installation includes Helm, kubectl, nerdctl, Moby, and Docker Compose. Trivy-based image scanning reports vulnerabilities and configuration issues, and Docker Desktop Extensions can be installed through the UI catalog or rdctl. The application uses a virtual machine on macOS and Linux and WSL 2 on Windows. Windows requires WSL and virtualization capabilities; without the privileged service, exposed container applications are limited to 127.0.0.1. The project is intended for local container and Kubernetes work, rather than managing Kubernetes clusters.