使用Jenkins Ansible Docker Swarm实现自动化编译部署

自动化部署在项目部署过程中很重要，一旦自动化部署完成，我们就可以减轻我们手动的操作的步骤以及出错的概率。下面介绍一种通用的自动化部署。从打包编译到上线，一条命令就完成所有操作。简单而高效

1、Jenkins部署这里不在赘述，直接从新建项目开始。

项目截图，如下：增加两个变量如图，

• BranchTobuild 默认是master分支，在编译的时候也可以指定分支。

• PushDockerImage 默认布尔值是选中，直接把镜像推到镜像仓库。

增加Pipeline配置，主要是分了几个步骤。所以用pipeline，也可以不用这种方式，直接脚本也行

所有执行步骤写入Jenkinsfile中，并把文件放入项目的根目录，这样才能调用。

Jenkinsfile内容如下：包括编译、打包、推送docker镜像到仓库，ansible部署。

#cat Jenkinsfile
pipeline {

   //定义了执行jenkins的机器
    agent {
        label ‘master’
    }
    stages {
        stage (‘Prepare’) {
            steps {
                sh “echo Prepare”
                //准备阶段，没有暂不处理
            }
        }

        //build打包过程，最后会生产apk包
        stage (‘build’) {
            steps {
                sh ‘echo “start Build”‘
                script {
                    if (params.buildDebug) {
                        sh ‘echo “build debug version”‘
                        sh ‘chmod +x ./gradlew’
                        sh “./gradlew clean”
                        sh “./gradlew build –debug”
                    } else {
                        sh ‘echo “build release version”‘
                        sh ‘chmod +x ./gradlew’
                        sh “./gradlew clean”
                        sh “./gradlew build”
                    }
                }
            }
        }
        stage (‘Test’) {
            steps {
                sh “echo Test”
                //测试阶段，没有暂不处理
            }
        }

       //发布阶段，项目中包括了gradle.properties文件，几个变量：版本、项目、产品，按照这几个名词生产镜像的名
        stage (‘Deploy’) {
            steps {
                sh ‘echo “Start Docker Build And Push Images”‘
                script {
                    if (params.PushDockerImage) {
                        def props = readProperties  file: ‘gradle.properties’
                                    def VERSION = props[‘version’]
                                    def PRODUCT = props[‘product’]
                                    def ARTIFACT = props[‘artifact’]
                        sh ‘echo “start Build”‘
                                    //开始修改Dockerfile
                                    sh “sed -i ‘s#\${PRODUCT}#${PRODUCT}#g’ Dockerfile”
                                    sh “sed -i ‘s#\${VERSION}#${VERSION}#g’ Dockerfile”
                                    sh “sed -i ‘s#\${ARTIFACT}#${ARTIFACT}#g’ Dockerfile”
                        sh “docker build -t registry.sreop.com:5000/${PRODUCT}/${ARTIFACT}:${VERSION} -f Dockerfile .”
                        sh ‘echo “Publish Images To registry.leautolink.com”‘
                        sh “docker push registry.leautolink.com:5000/${PRODUCT}/${ARTIFACT}:${VERSION}”

                        //ansible playbook 部署到线上或者测试环境
                        sh “sudo /usr/bin/ansible-playbook  /data/base-docker-compose/product/light/prod/playbook.yml”
                    }
                }
            }
        }
    }
}

2、每个项目需要一个Dockefile

cat Dockerfile
# 基础镜像
FROM registry.sreop.com:5000/alpine-java:8u121b13_jdk_unlimited
# 维护者信息
MAINTAINER test@test.com
# 镜像操作命令
RUN mkdir -p /data/bin/${PRODUCT}/${ARTIFACT}
# 指定后续命令的执行目录
WORKDIR /data/bin/${PRODUCT}/${ARTIFACT}
# 对外连接端口号
EXPOSE 12429
# 向镜像中增加文件
ADD ./build/libs/${ARTIFACT}-${VERSION}.jar .
# 容器启动命令
CMD java -Djava.security.egd=file:/dev/./urandom -jar ${ARTIFACT}-${VERSION}.jar –spring.profiles.active=prod

3、写ansible playbook，为推送到Docker集群平台，并部署到线上或者测试平台。

注意：docker-compose.yml要拷贝到每个docker节点机器上。功能是：获取各个变量，拉取镜像。最后生产docker services

– name: light Pull Image
  hosts: docker-swarm-prod
  remote_user: root
  tasks:
    – name: get product var 
      shell: grep image /data/base-docker-compose/product/light/prod/docker-compose.yml |head -n1|awk -F’/’ ‘{print $2}’
      register: product
    – name: get artifact var
      shell: grep image /data/base-docker-compose/product/light/prod/docker-compose.yml|head -n1 |awk -F’/’ ‘{print $3}’|sed ‘s#\(.*\):\(.*\)#\1#g’
      register: artifact
    – name: get version var
      shell: grep image /data/base-docker-compose/product/light/prod/docker-compose.yml|head -n1 |awk -F’/’ ‘{print $3}’|sed ‘s#\(.*\):\(.*\)#\2#g’
      register: version
    – name: docker pull image
      shell: docker pull registry.leautolink.com:5000/”{{ product.stdout }}”/”{{ artifact.stdout }}”:”{{ version.stdout }}” || /bin/true

– name: Remove Docker Old Verison
  hosts: docker-swarm-prod-lead 
  remote_user: root
  tasks:
    – name: get product var
      shell: grep image /data/base-docker-compose/product/light/prod/docker-compose.yml|head -n1|awk -F’/’ ‘{print $2}’
      register: product
    – name: get artifact var
      shell: grep image /data/base-docker-compose/product/light/prod/docker-compose.yml|head -n1 |awk -F’/’ ‘{print $3}’|sed ‘s#\(.*\):\(.*\)#\1#g’
      register: artifact
    – name: get version var
      shell: grep image /data/base-docker-compose/product/light/prod/docker-compose.yml|head -n1 |awk -F’/’ ‘{print $3}’|sed ‘s#\(.*\):\(.*\)#\2#g’
      register: version
    – name: remove light
      shell: docker stack rm {{ artifact.stdout }} || /bin/true

– name: Start New verison
  hosts: docker-swarm-prod-lead 
  remote_user: root
  tasks:
    – name: get product var
      shell: grep image /data/base-docker-compose/product/light/prod/docker-compose.yml|head -n1|awk -F’/’ ‘{print $2}’
      register: product
    – name: get artifact var
      shell: grep image /data/base-docker-compose/product/light/prod/docker-compose.yml|head -n1 |awk -F’/’ ‘{print $3}’|sed ‘s#\(.*\):\(.*\)#\1#g’
      register: artifact
    – name: get version var
      shell: grep image /data/base-docker-compose/product/light/prod/docker-compose.yml|head -n1 |awk -F’/’ ‘{print $3}’|sed ‘s#\(.*\):\(.*\)#\2#g’
      register: version
    – name: start light
      shell: docker stack deploy -c /data/base-docker-compose/product/light/prod/docker-compose.yml {{ artifact.stdout }} || /bin/true 

4、swarm集群部署compose.yml

# cat docker-compose.yml 
version: “3”
services:
  config-server:
    image: registry.sreop.com:5000/leradio/light:1.0.1-RELEASE
    command: java -Djava.security.egd=file:/dev/./urandom -jar light-1.0.1-RELEASE.jar –spring.cloud.config.profile=prod –spring.profiles.active=prod
    ports:
      – 12429:12429
    deploy:
      replicas: 2
      update_config:
        parallelism: 2
      restart_policy:
        condition: on-failure
networks:
  frontend: