在 macOS 和 Linux 上一条命令就能搭好容器运行时环境，省去手动配置虚拟机和 Docker 的麻烦，还支持 Kubernetes、containerd 等多种后端。
Container runtimes on macOS (and Linux) with minimal setup
brew install colima
brew install --HEAD colima
docker run hello-world
docker ps这个项目的说明还在生成中，先参考下面「最近发版」和 README。
This is an hotfix release to address a regression for Kubernetes runtime.
Full Changelog: https://github.com/abiosoft/colima/compare/v0.10.2...v0.10.3
This is an incremental stability release with bug fixes and features.
--downloader flag to retain curl for downloads. Supports curl or native (default).colima stop --force no longer attempts to stop the container runtime before stopping VM.krunkit vm type.colima delete without --force) is now properly supported on Incus runtime, and container data would be auto-restored on recreation.COLIMA_CACHE_HOME environment variable. The cache directory stores the downloaded assets used by Colima.krunkit vm type.colima list for stopped instances.--force-disk-image flag in addition to --disk-image flag. The image must be a debian based distro with systemd. Note that this is for power users and not officially supported.NOTE: container runtime versions can be updated manually by running the colima update command.
v29.5.2v2.3.1v7.1This is an incremental release with support for Docker Model Runner and improvements to the AI functionality introduced in v0.10.0.
Colima now supports Docker Model Runner as an AI model runner backend. Docker Model Runner is now the default due to its simpler requirements. However, Ramalama can still be used.
# run a model (uses docker runner by default)
colima model run gemma3
# serve a model, chat interface available at localhost:8080
colima model serve gemma3
# explicitly specify the runner
colima model run gemma3 --runner docker
colima model run gemma3 --runner ramalama
# set the runner at start time
colima start --model-runner ramalama
The runner can be configured via:
--runner flag on colima model commands--model-runner flag at colima startmodelRunner in the configuration fileDOCKER_CONFIG environment variable is now respected.Full Changelog: https://github.com/abiosoft/colima/compare/v0.10.0...v0.10.1
This release introduces AI support and other updates.
Colima now has AI support.
By leveraging Krunkit and Ramalama, Colima is able to provide the ideal platform to run confined, isolated and secure gpu-powered AI workloads on Apple Silicon devices.
# run a model
colima model run gemma3
# serve a model, chat interface would be availabe at localhost:8080
colima model serve gemma3
# for more
colima model --help
krunkit virtual machine type with gpu support. colima start --vm-type krunkit.colima start --network-address.CONTAINERD_* and NERDCTL_* environment variables on the host.--port-forwarder=none to colima start.--mount=none to colima start.curl and shasum on the host.after-boot and ready provision modes for provision scripts.NOTE: container runtime versions can be updated manually by running the colima update command.
v29.2.0v2.2.1v6.21v1.35.0+k3s1.This is a hotfix release to address disk error issues for a subset of users.
Check v0.9.0 release notes for the main release notes.
Full Changelog: https://github.com/abiosoft/colima/compare/v0.9.0...v0.9.1
还没有人点评。第一个说出真实体验的人最有价值。
Website & Documentation: colima.run | colima.run/docs
Support for Intel and Apple Silicon macOS, and Linux
Colima is available on Homebrew, MacPorts, Nix and Mise. Check here for other installation options.
# Homebrew
brew install colima
# MacPorts
sudo port install colima
# Nix
nix-env -iA nixpkgs.colima
# Mise
mise use -g colima@latest
Or stay on the bleeding edge (only Homebrew)
brew install --HEAD colima
Start Colima with defaults
colima start
For more usage options
colima --help
colima start --help
Or use a config file
colima start --edit
On initial startup, Colima initiates with a user specified runtime that defaults to Docker.
Docker client is required for Docker runtime. Installable with brew install docker.
colima start
docker run hello-world
docker ps
You can use the docker client on macOS after colima start with no additional setup.
colima start --runtime containerd starts and setup Containerd. You can use colima nerdctl to interact with
Containerd using nerdctl.
colima start --runtime containerd
nerdctl run hello-world
nerdctl ps
It is recommended to run colima nerdctl install to install nerdctl alias script in $PATH.
kubectl is required for Kubernetes. Installable with brew install kubectl.
To enable Kubernetes, start Colima with --kubernetes flag.
colima start --kubernetes
kubectl run caddy --image=caddy
kubectl get pods
For Docker runtime, images built or pulled with Docker are accessible to Kubernetes.
For Containerd runtime, images built or pulled in the k8s.io namespace are accessible to Kubernetes.
Requires v0.7.0
Incus client is required for Incus runtime. Installable with brew brew install incus.
colima start --runtime incus starts and setup Incus.
colima start --runtime incus
incus launch images:alpine/edge
incus list
You can use the incus client on macOS after colima start with no additional setup.
Note: Running virtual machines on Incus is only supported on m3 or newer Apple Silicon devices.
Requires v0.10.0, Apple Silicon and macOS 13+
Colima supports GPU accelerated containers for AI workloads using the krunkit VM type.
Note: To use krunkit with colima, ensure it is installed. Please follow their installation instructions
Setup and use a model.
colima start --runtime docker --vm-type krunkit
colima model run gemma3
Colima supports two model runner backends:
The default registry is the Docker AI Registry. Models can be run by name without a prefix:
colima model run gemma3
colima model run llama3.2
# HuggingFace (Docker Model Runner)
colima model run hf.co/microsoft/Phi-3-mini-4k-instruct-gguf
# Ollama (requires ramalama runner)
colima model run ollama://gemma3 --runner ramalama
See the AI Workloads documentation for more details.
The default VM created by Colima has 2 CPUs, 2GiB memory and 100GiB storage.
The VM can be customized either by passing additional flags to colima start.
e.g. --cpu, --memory, --disk, --runtime.
Or by editing the config file with colima start --edit.
NOTE: Disk size can be increased after the VM is created.
create VM with 1CPU, 2GiB memory and 10GiB storage.
colima start --cpu 1 --memory 2 --disk 10
modify an existing VM to 4CPUs and 8GiB memory.
colima stop
colima start --cpu 4 --memory 8
create VM with Rosetta 2 emulation. Requires v0.5.3 and macOS >= 13 (Ventura) on Apple Silicon.
colima start --vm-type=vz --vz-rosetta
To provide container runtimes on macOS with minimal setup.
Colima means Containers on Lima.
Since Lima is aka Linux Machines. By transitivity, Colima can also mean Containers on Linux Machines.
The logo was contributed by Daniel Hodvogner. Check this issue for more.
Check here for Frequently Asked Questions, or visit the online FAQ for a searchable version.
Check here for the instructions on contributing to the project.
#colima channel in the CNCF Slack
MIT
If you (or your company) are benefiting from the project and would like to support the contributors, kindly sponsor.
生产级的容器编排与调度平台，用来自动化部署、扩缩容和管理跨主机的容器化应用。
自托管的网站与服务可用性监控面板，支持 HTTP、TCP、Ping 等多种探活方式，通过实时图表和告警通知掌握服务宕机情况。
基于 Spring Boot + MyBatis 的完整电商系统，含前台商城和后台管理，覆盖商品、订单、会员、促销等模块，用 Docker 容器化部署。
构建容器化系统的开源协作项目，提供可拼装的组件，用来搭建和定制自己的容器平台，也是 Docker 的上游基础。