Project Icon

docker-plugin

Docker容器动态配置Jenkins构建节点插件

Docker-plugin是一款Jenkins云插件,用于动态配置Docker容器作为构建节点。支持SSH、JNLP和附加连接模式,可集成多个Docker主机或Swarm集群。插件提供灵活配置,支持自定义镜像,并可通过Groovy脚本或JCasC设置。适用于需要扩展Jenkins构建能力的场景。

Docker plugin for Jenkins

Note: This plugin is officially "up for adoption". It would benefit from having a new maintainer who uses it "for real work" and is thus able to test things "for real" before release instead of relying purely on the unit tests.

Overview

This plugin allows containers to be dynamically provisioned as Jenkins nodes using Docker. It is a Jenkins Cloud plugin for Docker.

The aim of this docker plugin is to be able to use a Docker host to dynamically provision a docker container as a Jenkins agent node, let that run a single build, then tear-down that node, without the build process (or Jenkins job definition) requiring any awareness of docker.

The Jenkins administrator configures Jenkins with knowledge of one or more docker hosts (or swarms), knowledge of one or more "templates" (which describe the labels/tags that this template provides, the docker image, how to start it, etc) and Jenkins can then run docker containers to provide Jenkins (agent) Nodes on which Jenkins can run builds.

See also


Note: There is more than one docker plugin for Jenkins. While this can be confusing for end-users, it's even more confusing when end users report bugs in the wrong place. e.g. if you are using Jenkins pipeline / workflow / Jenkinsfile builds with code including terms like docker.withDockerRegistry or docker.image etc then you're using the docker-workflow plugin and should go to its repository instead of this one.


Note: This plugin does not use the OS's native docker client; it uses docker-java. You do not need to install a docker client on Jenkins or on your agents to use this plugin.


Note: This plugin does not provide a Docker daemon; it allows Jenkins to use a docker daemon. i.e. Once you've installed docker somewhere, this plugin will allow Jenkins to make use of it.


Setup

A quick setup is :

  1. get a docker environment running
  2. follow the instructions for creating a docker image that can be used as a Jenkins Agent or use one of the pre-built images e.g. jenkins/inbound-agent

Docker Environment

Follow the installation steps on the docker website.

If your Jenkins instance is not on the same OS as the docker install, you will need to open the docker TCP port so that Jenkins can communicate with the docker daemon. This can be achieved by editing the docker config file and setting (for example)

DOCKER_OPTS="-H tcp://0.0.0.0:2376 -H unix:///var/run/docker.sock"

The docker configuration file location will depend your system, but it is likely to be /etc/init/docker.conf , /etc/default/docker or /etc/default/docker.io.

Multiple Docker Hosts

If you want to use more than just one physical node to run containers, you can use Docker Swarm Standalone or you can define more than one docker "cloud". The docker engine swarm mode API is not supported (at present; enhancement contributions would be welcomed).

To use the standalone swarm, follow docker swarm standalone instructions and configure Jenkins with the swarm's API endpoint.

Jenkins Configuration

Docker plugin is a "Cloud" implementation. You'll need to edit Jenkins system configuration (Jenkins -> Manage -> System configuration) and add a new Cloud of type "Docker".

Configure Docker (or Swarm standalone) API URL with required credentials. The test button lets you check the connection.

Then configure Agent templates, assigning them labels that you can use so your jobs select the appropriate template, and set the docker container to be run with whatever container settings you require.

Creating a docker image

You need a docker image that can be used to run Jenkins agent runtime. Depending on the launch method you select, there's some prerequisites for the Docker image to be used:

Launch via SSH

  • sshd server and a JDK installed. You can use jenkins/ssh-agent as a basis for a custom image.
  • a SSH key (based on the unique Jenkins instance identity) can be injected in container on startup, you don't need any credential set as long as you use standard openssl sshd. When using the jenkins/ssh-agent Docker image, ensure that the user is set to jenkins. For backward compatibility or non-standard sshd packaged in your docker image, you also have option to provide manually configured ssh credentials
  • Note: If the docker container's host SSH key is not trusted by Jenkins (usually the case) then you'll need to set the SSH host key verification method to "non-verifying".

Launch via JNLP

  • a JDK installed. You can use jenkins/inbound-agent as a basis for a custom image.
  • Jenkins controller URL has to be reachable from container.
  • container will be configured automatically with agent's name and secret, so you don't need any special configuration of the container.

Launch attached

  • a JDK installed. You can use jenkins/agent as a basis for a custom image.

To create a custom image and bundle your favorite tools, create a Dockerfile with the FROM to point to one of the jenkins/*-agent reference images, and install everything needed for your own usage, e.g.

FROM jenkins/inbound-agent
RUN apt-get update && apt-get install XXX
COPY your-favorite-tool-here

Note on ENTRYPOINT

Avoid overriding the docker command, as the SSH Launcher relies on it.

You can use an Entrypoint to run some side service inside your build agent container before the agent runtime starts and establish a connection ... but you MUST ensure your entrypoint eventually runs the passed command:

exec "$@"

Further information

More information can be obtained from the online help built into the Jenkins web UI. Most configurable fields have explanatory text. This, combined with knowledge of docker itself, should answer most questions.

Configuration as code

Jenkins and the docker-plugin can be configured using Groovy code and/or using the JCasC plugin.

If you're unsure which method to use, use JCasC.

JCasC plugin

Install the configuration-as-code plugin and follow its example.

As another alternative, a Docker daemon can listen to requests from remote hosts by following the Docker documentation. The following configuration as code example creates a cloud named "my-docker-cloud" that uses the docker daemon at port 2375 on dockerhost.example.com to run up to 3 containerized agents at a time. Agents run the Jenkins Alpine inbound agent container image with Java 21. Thay use an inbound connection and run as the user ID 1000 with the home directory "/home/jenkins/agent".

jenkins:
  clouds:
  - docker:
      containerCap: 3
      dockerApi:
        connectTimeout: 23
        dockerHost:
          uri: "tcp://dockerhost.example.com:2375"
        readTimeout: 43
      errorDuration: 313
      name: "my-docker-cloud"
      templates:
      - connector:
          jnlp:
            jenkinsUrl: "https://jenkins.example.com/"
            user: "1000"
        dockerTemplateBase:
          cpuPeriod: 0
          cpuQuota: 0
          image: "jenkins/inbound-agent:latest-alpine-jdk21"
        labelString: "alpine jdk21 alpine-jdk21 git-2.43"
        name: "alpine-jdk21"
        pullTimeout: 171
        remoteFs: "/home/jenkins/agent"

Groovy script

For example, this configuration script could be run automatically upon Jenkins post-initialization or through the Jenkins script console. If run, this script will configure the docker-plugin to look for a docker daemon running within the same OS as the Jenkins controller (connecting to Docker service through unix:///var/run/docker.sock) and with the containers connecting to Jenkins using the "attach" method.

项目侧边栏1项目侧边栏2
推荐项目
Project Cover

豆包MarsCode

豆包 MarsCode 是一款革命性的编程助手,通过AI技术提供代码补全、单测生成、代码解释和智能问答等功能,支持100+编程语言,与主流编辑器无缝集成,显著提升开发效率和代码质量。

Project Cover

问小白

问小白是一个基于 DeepSeek R1 模型的智能对话平台,专为用户提供高效、贴心的对话体验。实时在线,支持深度思考和联网搜索。免费不限次数,帮用户写作、创作、分析和规划,各种任务随时完成!

Project Cover

白日梦AI

白日梦AI提供专注于AI视频生成的多样化功能,包括文生视频、动态画面和形象生成等,帮助用户快速上手,创造专业级内容。

Project Cover

有言AI

有言平台提供一站式AIGC视频创作解决方案,通过智能技术简化视频制作流程。无论是企业宣传还是个人分享,有言都能帮助用户快速、轻松地制作出专业级别的视频内容。

Project Cover

讯飞绘镜

讯飞绘镜是一个支持从创意到完整视频创作的智能平台,用户可以快速生成视频素材并创作独特的音乐视频和故事。平台提供多样化的主题和精选作品,帮助用户探索创意灵感。

Project Cover

讯飞文书

讯飞文书依托讯飞星火大模型,为文书写作者提供从素材筹备到稿件撰写及审稿的全程支持。通过录音智记和以稿写稿等功能,满足事务性工作的高频需求,帮助撰稿人节省精力,提高效率,优化工作与生活。

Project Cover

阿里绘蛙

绘蛙是阿里巴巴集团推出的革命性AI电商营销平台。利用尖端人工智能技术,为商家提供一键生成商品图和营销文案的服务,显著提升内容创作效率和营销效果。适用于淘宝、天猫等电商平台,让商品第一时间被种草。

Project Cover

Trae

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。

Project Cover

AIWritePaper论文写作

AIWritePaper论文写作是一站式AI论文写作辅助工具,简化了选题、文献检索至论文撰写的整个过程。通过简单设定,平台可快速生成高质量论文大纲和全文,配合图表、参考文献等一应俱全,同时提供开题报告和答辩PPT等增值服务,保障数据安全,有效提升写作效率和论文质量。

投诉举报邮箱: service@vectorlightyear.com
@2024 懂AI·鲁ICP备2024100362号-6·鲁公网安备37021002001498号