containerd
An open-source container runtime originally developed by Docker and later donated to the Cloud Native Computing Foundation (CNCF). It serves as a core container runtime adopted by Kubernetes and other projects, managing the complete container lifecycle including image transfer, storage, and execution.
Core Facts
Timeline (last 90 days)
Container runtimes (such as containerd and CRI-O) pull images, create namespaces, and mount filesystems so that processes inside the container believe they have a machine to themselves.
Registering the nvidia runtime in the containerd template requires setting runtime_type to "io.containerd.runc.v2" and BinaryName to "/usr/bin/nvidia-container-runtime".
K3s's containerd configuration file is located at /var/lib/rancher/k3s/agent/etc/containerd/config.toml, and editing it directly will be overwritten on K3s restart, so the .tmpl template extension mechanism should be used.
The role of the NVIDIA Container Toolkit is to enable containerd to inject host GPU devices into the OCI container spec.
The containerd runtime built into K3s does not, by default, expose the host's PCI hardware (such as an NVIDIA GPU) to containers.
Open-source container runtimes such as Podman and containerd are also actively adapting to the WSL environment.
runc is the reference implementation of the OCI Runtime Spec, and containerd and CRI-O wrap image management and lifecycle management capabilities on top of it.
The containerd runtime underlying Docker has been widely adopted by projects such as Kubernetes.
containerd was originally split out from the Docker architecture, is currently hosted by CNCF and has graduated; CRI-O is led by Red Hat and does not support Docker-exclusive APIs.
containerd has become a graduated project of the CNCF (Cloud Native Computing Foundation) and is widely adopted by orchestration systems such as Kubernetes
9 more timeline events
All Facts (20)
容器技术（以Docker为代表）利用Linux内核的namespace机制实现进程、网络、文件系统隔离，借助cgroups限制CPU、内存等资源用量
80%UnverifiedContainer runtimes (such as containerd and CRI-O) pull images, create namespaces, and mount filesystems so that processes inside the container believe they have a machine to themselves.
50%UnverifiedK3s's containerd configuration file is located at /var/lib/rancher/k3s/agent/etc/containerd/config.toml, and editing it directly will be overwritten on K3s restart, so the .tmpl template extension mechanism should be used.
50%UnverifiedRegistering the nvidia runtime in the containerd template requires setting runtime_type to "io.containerd.runc.v2" and BinaryName to "/usr/bin/nvidia-container-runtime".
50%UnverifiedThe containerd runtime built into K3s does not, by default, expose the host's PCI hardware (such as an NVIDIA GPU) to containers.
50%UnverifiedThe role of the NVIDIA Container Toolkit is to enable containerd to inject host GPU devices into the OCI container spec.
50%UnverifiedOpen-source container runtimes such as Podman and containerd are also actively adapting to the WSL environment.
50%Unverifiedrunc is the reference implementation of the OCI Runtime Spec, and containerd and CRI-O wrap image management and lifecycle management capabilities on top of it.
50%UnverifiedThe containerd runtime underlying Docker has been widely adopted by projects such as Kubernetes.
50%Unverifiedcontainerd was originally split out from the Docker architecture, is currently hosted by CNCF and has graduated; CRI-O is led by Red Hat and does not support Docker-exclusive APIs.
50%UnverifiedDocker's underlying runtime has migrated from the original LXC to its self-developed containerd and runc
50%Unverifiedcontainerd has become a graduated project of the CNCF (Cloud Native Computing Foundation) and is widely adopted by orchestration systems such as Kubernetes
50%UnverifiedThe Docker image format has been standardized by OCI (Open Container Initiative), and images built with Docker can run on alternative container runtimes such as Podman and containerd.
50%UnverifiedDocker's core container runtime was ultimately replaced by containerd, which is hosted by the CNCF foundation.
50%UnverifiedDocker底层依赖containerd作为高级运行时，再由containerd调用runc完成容器创建
50%Unverified现代容器运行时分为高级运行时（如containerd、CRI-O）和低级运行时（如runc）两层
50%Unverified2015年Docker将其核心运行时组件拆分捐赠给CNCF，形成独立的containerd项目
50%Unverifiedrunc本身约10MB，containerd约50MB，完整的Docker引擎超过200MB
50%Unverified容器运行时通常分为高级运行时（如containerd、CRI-O）和低级运行时（如runc）两层，低级运行时直接调用Linux内核的namespace和cgroup机制实现进程隔离和资源限制
50%UnverifiedDocker 镜像遵循 OCI 标准，任何兼容 OCI 的运行时（如 containerd、CRI-O）都可以运行 Docker 构建的镜像
50%