Rust 实现的高性能 OCI 容器运行时
youki 是一个用 Rust 语言实现的 OCI 容器运行时,旨在提供高效、低内存消耗的容器环境。它已通过 containerd 端到端测试,并在多个生产环境中得到应用。youki 支持无根容器运行,提供详细文档,并欢迎社区贡献。作为 runc 的替代方案,youki 致力于在资源受限场景中推动容器技术创新。
youki is an implementation of the OCI runtime-spec in Rust, similar to runc.
Your ideas are welcome here.
youki is pronounced as /joʊki/ or yoh-key. youki is named after the Japanese word 'youki', which means 'a container'. In Japanese language, youki also means 'cheerful', 'merry', or 'hilarious'.
[!TIP] You can immediately set up your environment with youki on GitHub Codespaces and try it out.
$ just build $ docker run --runtime youki hello-world $ sudo podman run --cgroup-manager=cgroupfs --runtime /workspaces/youki/youki hello-world
Here is why we are writing a new container runtime in Rust.
Rust is one of the best languages to implement the oci-runtime spec. Many very nice container tools are currently written in Go. However, the container runtime requires the use of system calls, which requires a bit of special handling when implemented in Go. This tricky (e.g. namespaces(7), fork(2)); with Rust too, but it's not that tricky. And, unlike in C, Rust provides the benefit of memory safety. While Rust is not yet a major player in the container field, it has the potential to contribute a lot: something this project attempts to exemplify.
youki has the potential to be faster and use less memory than runc, and therefore work in environments with tight memory usage requirements. Here is a simple benchmark of a container from creation to deletion.
Runtime | Time (mean ± σ) | Range (min … max) | vs youki(mean) | Version |
---|---|---|---|---|
youki | 111.5 ms ± 11.6 ms | 84.0 ms ± 142.5 ms | 100% | 0.3.3 |
runc | 224.6 ms ± 12.0 ms | 190.5 ms ± 255.4 ms | 200% | 1.1.7 |
crun | 47.3 ms ± 2.8 ms | 42.4 ms ± 56.2 ms | 42% | 1.15 |
A command used for the benchmark
hyperfine --prepare 'sudo sync; echo 3 | sudo tee /proc/sys/vm/drop_caches' --warmup 10 --min-runs 100 'sudo ./youki create -b tutorial a && sudo ./youki start a && sudo ./youki delete -f a'
Environment
$ ./youki info Version 0.3.3 Commit 4f3c8307 Kernel-Release 6.5.0-35-generic Kernel-Version #35~22.04.1-Ubuntu SMP PREEMPT_DYNAMIC Tue May 7 09:00:52 UTC 2 Architecture x86_64 Operating System Ubuntu 22.04.4 LTS Cores 16 Total Memory 63870 Cgroup setup unified Cgroup mounts Namespaces enabled mount enabled uts enabled ipc enabled user enabled pid enabled network enabled cgroup enabled Capabilities CAP_BPF available CAP_PERFMON available CAP_CHECKPOINT_RESTORE available
I have fun implementing this. In fact, this may be the most important.
youki has aced real-world use cases, including containerd's e2e test, and is now adopted by several production environments. We have our roadmap.
The User and Developer Documentation for youki is hosted at https://containers.github.io/youki/
Local build is only supported on Linux. For other platforms, please use the Vagrantfile that we have prepared. You can also spin up a fully preconfigured development environment in the cloud with GitHub Codespaces.
To install just
, follow the instruction here.
$ sudo apt-get install \ pkg-config \ libsystemd-dev \ build-essential \ libelf-dev \ libseccomp-dev \ libclang-dev \ glibc-static \ libssl-dev
$ sudo dnf install \ pkg-config \ systemd-devel \ elfutils-libelf-devel \ libseccomp-devel \ clang-devel \ openssl-devel
git clone git@github.com:containers/youki.git cd youki just youki-dev # or youki-release ./youki -h # you can get information about youki command
Let's try to run a container that executes sleep 30
with youki. This tutorial may need root permission.
git clone git@github.com:containers/youki.git cd youki just youki-dev # or youki-release mkdir -p tutorial/rootfs cd tutorial # use docker to export busybox into the rootfs directory docker export $(docker create busybox) | tar -C rootfs -xvf -
Then, we need to prepare a configuration file. This file contains metadata and specs for a container, such as the process to run, environment variables to inject, sandboxing features to use, etc.
../youki spec # will generate a spec file named config.json
We can edit the config.json
to add customized behaviors for container. Here, we modify the process
field to run sleep 30
.
"process": { ... "args": [ "sleep", "30" ], ... }
Then we can explore the lifecycle of a container:
cd .. # go back to the repository root sudo ./youki create -b tutorial tutorial_container # create a container with name `tutorial_container` sudo ./youki state tutorial_container # you can see the state the container is `created` sudo ./youki start tutorial_container # start the container sudo ./youki list # will show the list of containers, the container is `running` sudo ./youki delete tutorial_container # delete the container
Change the command to be executed in config.json
and try something other than sleep 30
.
youki
provides the ability to run containers as non-root user(rootless mode). To run a container in rootless mode, we need to add some extra options in config.json
, other steps are same with above:
$ mkdir -p tutorial/rootfs $ cd tutorial # use docker to export busybox into the rootfs directory $ docker export $(docker create busybox) | tar -C rootfs -xvf - $ ../youki spec --rootless # will generate a spec file named config.json with rootless mode ## Modify the `args` field as you like $ ../youki run rootless-container # will create and run a container with rootless mode
Start the docker daemon.
dockerd --experimental --add-runtime="youki=$(pwd)/youki"
If you get an error like the below, that means your normal Docker daemon is running, and it needs to be stopped. Do that with your init system (i.e., with systemd, run systemctl stop docker
, as root if necessary).
failed to start daemon: pid file found, ensure docker is not running or delete /var/run/docker.pid
Now repeat the command, which should start the docker daemon.
You can use youki in a different terminal to start the container.
docker run -it --rm --runtime youki busybox
Afterwards, you can close the docker daemon process in other the other terminal. To restart normal docker daemon (if you had stopped it before), run:
systemctl start docker # might need root permission
Go and node-tap are required to run integration tests. See the opencontainers/runtime-tools README for details.
git submodule update --init --recursive just test-oci
You can try youki on platforms other than Linux by using the Vagrantfile we have prepared. We have prepared two environments for vagrant, namely rootless mode and rootful mode
git clone git@github.com:containers/youki.git cd youki # If you want to develop in rootless mode, and this is the default mode vagrant up vagrant ssh # or if you want to develop in rootful mode VAGRANT_VAGRANTFILE=Vagrantfile.root vagrant up VAGRANT_VAGRANTFILE=Vagrantfile.root vagrant ssh # in virtual machine cd youki just youki-dev # or youki-release
Please refer to our community page.
Thanks to all the people who already contributed!
<a href="https://github.com/containers/youki/graphs/contributors"> <img src="https://contributors-img.web.app/image?repo=containers/youki" /> </a>AI数字人视频创作平台
Keevx 一款开箱即用的AI数字人视频创作平台,广泛适用于电商广告、企业培训与社媒宣传,让全球企业与个人创作者无需拍摄剪辑,就能快速生成多语言、高质量的专业视频。
一站式AI创作平台
提供 AI 驱动的图片、视频生成及数字人等功能,助力创意创作
AI办公助手,复杂任务高效处理
AI办公助手,复杂任务高效处理。办公效率低?扣子空间AI助手支持播客生成、PPT制作、网页开发及报告写作,覆盖科研、商业、舆情等领域的专家Agent 7x24小时响应,生活工作无缝切换,提升50%效率!
AI辅助编程,代码自动修复
Trae是一种自适应的集成开发环境(IDE),通过自动化和多 元协作改变开发流程。利用Trae,团队能够更快速、精确地编写和部署代码,从而提高编程效率和项目交付速度。Trae具备上下文感知和代码自动完成功能,是提升开发效率的理想工具。
AI小说写作助手,一站式润色、改写 、扩写
蛙蛙写作—国内先进的AI写作平台,涵盖小说、学术、社交媒体等多场景。提供续写、改写、润色等功能,助力创作者高效优化写作流程。界面简洁,功能全面,适合各类写作者提升内容品质和工作效率。
全能AI智能助手,随时解答生活与工作的多样问题
问小白,由元石科技研发的AI智能助手,快速准确地解答各种生活和工作问题,包括但不限于搜索、规划和社交互动,帮助用户在日常生活中提高效率,轻松管理个人事务。
实时语音翻译/同声传译工 具
Transly是一个多场景的AI大语言模型驱动的同声传译、专业翻译助手,它拥有超精准的音频识别翻译能力,几乎零延迟的使用体验和支持多国语言可以让你带它走遍全球,无论你是留学生、商务人士、韩剧美剧爱好者,还是出国游玩、多国会议、跨国追星等等,都可以满足你所有需要同传的场景需求,线上线下通用,扫除语言障碍,让全世界的语言交流不再有国界。
一键生成PPT和Word,让学习生活更轻松
讯飞智文是一个利用 AI 技术的项目,能够帮助用户生成 PPT 以及各类文档。无论是商业领域的市场分析报告、年度目标制定,还是学生群体的职业生涯规划、实习避坑指南,亦或是活动策划、旅游攻略等内 容,它都能提供支持,帮助用户精准表达,轻松呈现各种信息。
深度推理能力全新升级,全面对标OpenAI o1
科大讯飞的星火大模型,支持语言理解、知识问答和文本创作等多功能,适用于多种文件和业务场景,提升办公和日常生活的效率。讯飞星火是一个提供丰富智能服务的平台,涵盖科技资讯、图像创作、写作辅助、编程解答、科研文献解读等功能,能为不同需求的用户提供便捷高效的帮助,助力用户轻松获取信息、解决问题,满足多样化使用场景。
一种基于大语言模型的高效单流解耦语音令牌文本到语音合成模型
Spark-TTS 是一个基于 PyTorch 的开源文本到语音合成项目,由多个知名机构联合参与。该项目提供了高效的 LLM(大语言模型)驱动的语音合成方案,支持语音克隆和语音创建功能,可通过命令行界面(CLI)和 Web UI 两种方式使用。用户可以根据需求调整语音的性别、音高、速度等参数,生成高质量的语音。该项目适用于多种场景,如有声读物制作、智能语音助手开发等。
最新AI工具、AI资讯
独家AI资源、AI项目落地
微信扫一扫关注公众号