CI 自动化部署

GitLab Runner CI 自动化部署

在部署服务器上安装gitlab-runner

yum install gitlab-runner

向GitLab-CI注册gitlab-runner

找到你的gitlab上存放的项目，找到settings–>CI/CD–>Runner

部署服务器上进行runner注册gitlab-runner register

1. url：gitlab所在的测试服务器地址

2. token：项目的token，用于关联runner和项目

3. name：runner的名字，用于区分runner

4. tags：用于匹配任务（jobs）和执行任务的设备（runners），此处不填直接回车时，默认Can run untagged jobs值为true

准备gitlab-ci.yml文件

在gitlabci-service根目录下：

在ci目录下创建.gitlab-ci.yml文件

include:   
	- project: "ci-devops/gitlabci-service"     
	ref: main     
	file: "be/java_build.yml"   
	- project: "ci-devops/gitlabci-service"     
	ref: main     
	file: "be/dingTalk_ci.yml"

	variables:   
	TIMESTAMP: 'date -d $CI_COMMIT_TIMESTAMP +%Y%m%d%H%M'   BUILD_SHELL: "mvn clean package -T 2C -Dmaven.test.skip=true -Dmaven.compile.fork=true"   
	ARTIFACT_REPOSITORY_NAME: "java"   
	TARGET_FILE_PATH: "$CI_PROJECT_NAMESPACE/$CI_PROJECT_NAME/$CI_COMMIT_REF_NAME-$CI_COMMIT_SHORT_SHA-$CI_PIPELINE_ID"   
	TARGET_ARTIFACT_NAME: "$CI_PROJECT_NAME-$CI_COMMIT_REF_NAME"   #IMAGE_TAG: "$HARBOR_URL/$CI_PROJECT_NAME/$CI_JOB_NAME:`$TIMESTAMP`"  stages:   
	- build_image   
	- dingTalk  
auto-deploy-demo:   
	stage: build_image   
	extends: .data-platform-common-template  

dingTalk:   
	rules:     
	- if: '$CI_COMMIT_REF_NAME == "dev/auto-deploy"'   
	stage: dingTalk   
	extends: .dingTalk

在buildfile目录下创建${projectName}.Dockerfile文件，用于docker构建镜像

FROM openjdk:17 
WORKDIR /app COPY demo2/target/*.jar /app/demo2.jar 
EXPOSE 8090 CMD ["java", "-jar", "demo2.jar"]

在deployfile目录下创建deploy-${projectName}.yml文件，用于执行k8s相关命令

apiVersion: apps/v1 
kind: Deployment 
metadata:   
	labels:     
	app: demo2   
	name: demo2   
	namespace: default 
spec:   
	replicas: 1   
	selector:     
		matchLabels:       
			app: demo2   
	template:     
		metadata:       
			labels:         
				app: demo2     
spec:       
	containers:        
	-  image: APP_IMAGE           
		imagePullPolicy: IfNotPresent           
		name: demo2           
		ports:             
		- containerPort: 8090               
		protocol: TCP

在ci目录下创建java_build.yml文件，用于处理在自动化部署中，打包、推包等流程

.data-platform-common-template:   
stage: build_image   
tags:     # 指定runner     
- server   
rules:     # 指定分支     
- if: '$CI_COMMIT_REF_NAME == "dev/auto-deploy"'       
changes:         
- $CI_JOB_NAME/**/*   
script:    
 - |     
- echo "====================进入$CI_BUILD_NAME======================="     
- echo "========定义变量:$CI_JOB_NAME========="     #打印所有的环境变量，用于调试     
- env     
- echo "====================编译======================="     
- java -version     
- mvn -version     # 如果是多层目录，需要进入当前项目所在目录     
- cd $CI_JOB_NAME/     #    - mvn clean     #    - mvn compile  
- echo "====================开始打包======================="     
- $BUILD_SHELL     
- echo "====================打包完毕======================="     
- cd target/     
- ls     
- pwd     
- echo "=================================================="     # 退出到根目录     
- cd ..     
- cd ..     
- ls     # 拉取Dockerfile文件，用于docker build     
- wget -O  Dockerfile http://192.168.6.246:70/ci-devops/gitlabci-service/-/raw/main/buildfile/$CI_JOB_NAME.Dockerfile     # 打印Dockerfile     
- ls && cat Dockerfile     # 构建镜像     
- docker build . -t $HARBOR_URL/demo/$CI_JOB_NAME:$($TIMESTAMP)     # 推送镜像至仓库     
- docker push $HARBOR_URL/demo/$CI_JOB_NAME:$($TIMESTAMP)     
- docker rmi  $HARBOR_URL/demo/$CI_JOB_NAME:$($TIMESTAMP)     
- IMAGES=$HARBOR_URL/demo/$CI_JOB_NAME:$($TIMESTAMP)     
- echo $IMAGES     # 拉去deploy.yml，用于执行k8s相关命令     
- wget -O deploy.yml http://192.168.6.246:70/ci-devops/gitlabci-service/-/raw/main/deployfile/deploy_$CI_JOB_NAME.yml     
- ls     # 替换变量     
- sed -i "s@APP_IMAGE@$IMAGES@g" deploy.yml     
- cat deploy.yml     # 暂时关闭     
- kubectl apply -f deploy.yml