foreword

Deployment is a repetitive and boring thing. I believe everyone has heard of CI/CD continuous integration and continuous delivery. This article is based on Gitlab CI/CD and Docker to quickly realize the automatic deployment of the project. This solution can be closely integrated with Git and can realize automatic building and deployment after code submission without additional tools.

Preparation

First of all, please install the following three softwares. It is recommended to install them in order. For the software installation, please refer to the reference links given below:

main process

Our main process is to manage the code through Gitlab, and then use the Gitlab CI/CD function that comes with Gitlab to register and bind with GitlabRunner. After the binding is successful, every time the code is submitted, GitlabRunner can be triggered to execute the corresponding script program to realize automatic deployment.
insert image description here

GitlabRunner registration

Open the Gitlab warehouse that needs to be automatically deployed, and select Setting-"CI/CD-"Runners.
As shown in the figure below, you can see the two most critical information, which will be used when GitlabRunner registers later, please remember:
GitlabRunner Registration
Command Guide: gitlab-runner registerRegister, gitlab-runner listview registered Runners, gitlab-runner unregister --all-runnerscancel all registrations.
First, we register one first, and then name a Runner (whatever), here I name it: hello, spring boot!, and then enter the tags of the runner, we can use tags to distinguish different stages of a project for different runners to execute, enter here : maven; Finally, enter the script you want to execute, here use the shell.

After the registration is successful, let's refresh the Gitlab page just now, and you can see that the registration has been successful!

production script

After submitting the code, how should we use Shell scripts to control the build and deployment? Gitlab has provided us with a template, which is to create a .gitlab-ci.ymlfile named: in the root directory of the warehouse. I execute a SpringBoot project here, so the content is as follows:

stages:
  - build
  - deploy_test
#  - clean
# 打包镜像
build:
  stage: build
  only:
    - master
  script:
    - $MAVEN_HOME/bin/mvn clean package -U '-Dmaven.test.skip=true'
    - cp target/GitlabRunnerTest-0.0.1-SNAPSHOT.jar docker/
    - cd docker
    - docker build -t test/gitlabrunner:v1.0.0 .
  tags:
    - maven
# 部署测试服务器
deploy_test:
  stage: deploy_test
  only:
    - master
  script:
    - cd docker
    - docker-compose down
    - docker-compose up -d
  tags:
    - maven
#清理虚悬镜像
#clean:
#  stage: clean
#  only:
#    - master
#  script:
#    - docker rmi $(docker images -q -f dangling=true)
#  tags:
#    - maven

Then create a Docker folder in the root directory, create a Dockerfile:

FROM openjdk:11
COPY . /usr/src/myapp
WORKDIR /usr/src/myapp

Create docker-compose.yml

version: '3.1'
services:
  education-course:
    container_name: test-gitlabrunner
    restart: always
    image: test/gitlabrunner:v1.0.0
    network_mode: "host"
    command: java -jar -Xms1024m -Xmx1024m -Duser.timezone=GMT+08 GitlabRunnerTest-0.0.1-SNAPSHOT.jar --spring.profiles.active=test

The overall project directory structure is as follows:
insert image description here

verify

After it is done, we can upload the code, and then we can check whether the execution is successful in the CI/CD Pipelines.
insert image description here
As you can see, the build is successful, let's visit:

Congratulations, it has been successfully deployed. You can try to modify the GIt submission in the code to see the effect of automatic deployment.

Rapid implementation of Spring project Gitlab automatic deployment based on Docker