Jenkins automates deployment of Spring Cloud services
January 8, 2024
by Kevin Rodriguez
No Comments
An overview of the
Jenkins is an automatic build tool, which can automatically pull code from Git after configuring some application servers on Jenkins server
The project is then deployed and started into the application. The purpose of this article is to go through how to configure the entire deployment environment.
I’ve been doing some work on how the company’s real environment is configured. For reference only, the company is a small company, if there is any inadequacy of the place welcome to point out, I good improvement ~
The deployment plan
Jenkins construction method
The node way
The node information configured by Jenkins was used for construction, which will be described in details later.
The Jenkins server is easy to configure, but the application server needs git and Maven installed. Because you build on the application server
Publish plug-in mode
Build by installing the plug-in Publish Over SSH, as described below.
The Jenkisn server is complex to configure, but the application server only needs to install the JDK
$ http://yourdomain:8080/exit# stop service$ http://yourdomain:8080/restart # restart service$ http://yourdomain:8080/reload Reload the configuration informationCopy the code
Initialize the Jenkins
Visit Jenkins at http://localhost:8080
The replication password is required for the first unlock:
The figure above shows the location of the password in the server. If installed using Docker, the password file is under the mounted directory
# vim /data/jenkins/secrets/initialAdminPassword
8079fb5d3e3f444d857fa5c8e7add6e9
Copy the code
Click “Continue” and click “Install recommended plug-ins”.
The plug-in is being installed, wait a moment
System management
After logging in Jenkins, there is a “System Administration” menu in the right menu, click it
The red boxes in the following figure are mandatory Settings for deployment. The rest depends:
So let’s set these guys up.
System Settings
Here is some basic information by default
Configuring Access Information
To configure the pulish over SSH plug-in, you need to configure the user name and password, create directory/WMS /services/ Jenkins on the application server 192.168.1.233, and click Test Configuration to verify.
The directory/WMS /services/ Jenkins in SSH Server node is very important.
Two of the later build tasks are based on this directory.
Of course, it doesn’t matter if the SSH Server node configuration is not configured if it is not built through the SSH Publishers plug-in.
192.168.1.112 The application Server is built in Publish mode, and the SSH Server information configured here is associated:
Global Tool Configuration
The Jenkisn server needs maven, JDK, and Git installed. I am pre-installed, you can also try to check automatic installation here.
Plug-in management
Ensure that the Git and Publish Over SSH plug-ins are installed successfully
Node management
Reviewing the server planning above, 120.79.xx. XX and 121.40.xx. XX are built in node mode, so nodes need to be configured separately here:
Taking 121.40.xx. XX as an example, click the right gear and then click Configure from node to view the node to be configured:
After saving, return to the node list. Click on the upper right corner of the allow automatic refresh, found normal connection, the effect is like the node list in the morning.
Begin to build
After the system administration is set up to start building, I’ve summarized two ways to build
Construction mode 1: Node mode
This mode needs to be configured in the system management configuration node, and then packaged and started in the remote management.
General Basic Configuration
Specify the running node and configure the node in System Management node Management
Source code management
Use SSH to specify the URL and authentication information of Git
Git basic information is specified in System Administration global Tool Configuration
The build
Invoke the top-level Maven target
The same as SSH.
Maven version
Maven1 is the name of the added Maven environment. The Maven environment is configured in System Settings
The target
Package submodules of the parent module by using maven command
clean package -pl zhirui-zuul-server/ -am -Dmaven.test.skip=true
Copy the code
If there are submodules under submodules, the package command is as follows:
clean package -pl zhirui-zmgj-business/zhirui-zmgj-server/ -am -Dmaven.test.skip=true
Copy the code
Execution of the shell
Because it runs through the designated node, there is no need to deploy through the SSH plug-in. [System Management] [Node Management] Node configuration:
Based on the above node configurationRemote working directory, JenkinsRemote working directory + workspace + task nameLet’s compile. So the jar path also needs to spell git project path.
The path where the actual deployment is running
Execute the following script in shell:
#! /bin/bash
#Jenkins pushes the package to the directory on the application server
jenkins="/root/.jenkins/workspace/zhirui-zmgj-server-pro/zhirui-zmgj-business/zhirui-zmgj-server/target"
#Publish directory (change to the corresponding publish directory)
deployDir="/wms/services/zhirui-zmgj-server"
#Jar name (change to the corresponding JAR name)JarName = "zhirui ZMGJ - server - 0.0.1 - the SNAPSHOT. Jar"
#Fixed configuration
jarLib=$deployDir"/lib"
jarBin=$deployDir"/bin"
jarBak=$deployDir"/bak"
jarLog=$deployDir"/logs"
#Directory to buildif [ ! -d ${jarLib}"/" ]; then mkdir ${jarLib} -p fi if [ ! -d ${jarBin}"/" ]; then mkdir ${jarBin} -p fi if [ ! -d ${jarBak}"/" ]; then mkdir ${jarBak} -p fi if [ ! -d ${jarLog}"/" ]; ${jarLog} then mkdir -p fi echo "* * * * * * * * * * * * * * * * * * * * * * basic configuration information * * * * * * * * * * * * * * * * * * * * * * *" Java - version echo "publish project: "${jarName} echo "Post directory: "${jardir} echo "bin directory: "${jarBin} echo" Lib directory: "${jarLib} echo "bak directory: "${jarBak} echo "Log directory: "${jarLog} echo "*******************************************************"
#Check whether new reports have been pushed to the local PC. If yes, perform subsequent operations
if [ "$(ls ${jenkins}/${jarName} 2> /dev/null | wc -l)" != "0" ]; then
#Backup fileOldJarName = "` ls ${jarLib} | grep. Jar | sort - nr | ` head - 1" echo "start backup jar package: "${jarLib}"/"${oldJarName} bakJarName=${oldJarName}$(date +%s) echo "Backup" ${jarLib} "/" ${oldJarName} "to" ${jarBak} "/" ${bakJarName} CD ${deployDir} if [" $(ls "lib/" *. Jar 2 > / dev/null | wc -l)" ! = "0"); Then mv ${jarLib}"/"${oldJarName} ${jarBak}" else echo" CD ${Jenkins} echo "${jarLib} cp ${Jenkins}"/"${jarName} ${jarLib} echo" #Copy start.sh and restart.sh from /var to the bin directory of the projectCp /var/env.start. sh. cp /var/env.restart. sh. chmod 775 * sh restart.sh else echo" ${jenkins}/${jarName}" exit 1 fiCopy the code
Construction method 2: Publish plug-in
This build is based on the plugin Publish Over SSH, which I call the Publish plugin approach. In this way, Jenkins server is packaged into JAR and then pushed to template server for running.
Click New Task
General Basic Configuration
SSH specifies the node to be run as master
Source code management
You need to go to Git and pull the code and compile it, and you need to add identity information to access Git
build
Invoke the top-level Maven target
Maven version
Maven2 is the name of the added Maven environment. The Maven environment is configured in System Settings
The target
Build individual modules using maven commands
clean package -pl zhirui-zuul-server/ -am -Dmaven.test.skip=true
Copy the code
If it is a secondary submodule, the build command is as follows:
clean package -pl zhirui-wms-business/zhirui-handle-file-server/ -am -Dmaven.test.skip=true
Copy the code
Configuration files and global configuration
Select the default Maven Settings, that is, the Maven environment configured in System Settings global Tools Configuration.
Send files or execute commands over SSH
The Jenkins plugin publishes Over SSH. It is used to configure the operation after the plug-in is pushed to the template application server. Several paths are easy to confuse, understand a little, and then try it yourself.
Before performing this step, you need to add the following SSH server configurations in System Management system Settings:
The instructions are as follows:
A: IP address of the SSH Server service Server
B and D: The Jenkins server pushes the JAR to the business server with the user account password information
C: The remote base directory is stored in the JAR on the SSH Server. I will refer to it as the remote base directory
Associate the following SSH servers with the SSH servers configured in System Management and System Settings:
Illustrated by the red box numbers in the figure above:
SSH Server name, which indicates the Server on which the SSH Server is executed
Based on Jenkins in system Management and System SetupHome directory ` ` / var/lib/JenkinsUnder theworkspace, the compiled JAR package is located in the Jenkins server workspace. For example, the full name of the directory is actually/ var/lib/Jenkins/workspace/zhirui - ZMGJ - request - service/zhirui - request - service/target/zhirui - request - service - 0.0.1 - the SNAPSHOT. The jar. Jenkins did by default/var/lib/jenkins/workspace/zhirui-zmgj-oauth-service/, includingzhirui-zmgj-oauth-serviceThis is the buildThe name of the task.
Based on the directory in 2, willtargetGet rid of that prefix.
Remote DirectoryIs a relative path required and for this buildThe name of the taskTo be consistent, Maven compiles the JAR to the target server (i.e192.168.1.112The server)Remote base directory + The name of the taskThe jar will not be found in subsequent shell scripts. For example, the full path to which the JAR is actually copied is/wms/services/jenkins/zhirui-zmgj-oauth-service/target, which is the path at position 6 in the red box below.
Directory where the application server deployment is running
The full path of the JAR under the relative path in 4. Copy to 5 for deployment.
The script for Exec Command is as follows:
Script description:
Modify several according to the actual situationRelease directory.The name of the project.Jenkins push package catalog.
jarBinThe directory needs to be placedStart a shell scriptStart. Sh and restart.sh in /var/env.Click to view the startup shell script.
#! /bin/bash
#Publish directory (change to the corresponding publish directory)
deployDir="/wms/services/zhirui-wms-erp-server"
#Project name (change to corresponding service name)JarName = "zhirui WMS - erp server - 0.0.1 - the SNAPSHOT. Jar"
#Jenkins push package catalog
jenkins="/wms/services/jenkins/zhirui-wms-erp-server/target"
#Fixed configuration
jarLib=$deployDir"/lib"
jarBin=$deployDir"/bin"
jarBak=$deployDir"/bak"
jarLog=$deployDir"/logs"
#Directory to buildif [ ! -d ${jarLib}"/" ]; then mkdir ${jarLib} -p fi if [ ! -d ${jarBin}"/" ]; then mkdir ${jarBin} -p fi if [ ! -d ${jarBak}"/" ]; then mkdir ${jarBak} -p fi if [ ! -d ${jarLog}"/" ]; ${jarLog} then mkdir -p fi echo "* * * * * * * * * * * * * * * * * * * * * * basic configuration information * * * * * * * * * * * * * * * * * * * * * * *" Java - version echo "publish project: "${jarName} echo "Post directory: "${jardir} echo "bin directory: "${jarBin} echo" Lib directory: "${jarLib} echo "bak directory: "${jarBak} echo "Log directory: "${jarLog} echo "*******************************************************"
#Check whether new reports have been pushed to the local PC. If yes, perform subsequent operations
if [ "$(ls ${jenkins}/${jarName} 2> /dev/null | wc -l)" != "0" ]; then
#Backup fileOldJarName = "` ls ${jarLib} | grep. Jar | sort - nr | ` head - 1" echo "start backup jar package: "${jarLib}"/"${oldJarName} bakJarName=${oldJarName}$(date +%s) echo "Backup" ${jarLib} "/" ${oldJarName} "to" ${jarBak} "/" ${bakJarName} CD ${deployDir} if [" $(ls "lib/" *. Jar 2 > / dev/null | wc -l)" ! = "0"); Then mv ${jarLib}"/"${oldJarName} ${jarBak}"/"${bakJarName} echo #Copies the new package to the specified directory${Jenkins} cp ${Jenkins}"/"${jarName} ${jarLib} #Restart the service
cd ${jarBin}
#Copy start.sh and restart.sh from /var to the bin directory of the projectCp /var/env.start. sh. cp /var/env.restart. sh. chmod 775 * sh./restart.sh else echo ": ${jenkins}/${jarName}" exit 1 fiCopy the code
Comparison between server directories and configuration parameters:
I feel that the path configured here is very complicated, I will try to make it clear, if you feel more round, you can skip, more tests can be.
Parameters to be configured:
Description:
1 The red box corresponds to the red box shown in the following figure. The path you want to fill in is based on the home directory +workspace
The red boxes in 2.1 and 2.2 correspond to the red boxes in figure 2. The path of the JAR after it is pushed to the application server is filled in. The 2.1 red box is the path based on the remote base directory. The jar installation 2.1 configuration is pushed to the target server. The result is the 2.2 red box path.
Red box 3 corresponds to red box 3 in the following figure. After the shell script configured in Exec Command is executed, the JAR will be copied to the path in red box 3 for execution, which is the actual running path of the program.
Some directory structure diagrams for the application server and Jinkins server:
Description:
The Jenkins server is packaged into jars and stored inHome directory ` ` workspaceThe position of the bottom.
After being pushed to the service server, based onRemote base directoryPost position.
When the service server runs the JAR, it executes a shell script, copies the script to this location, and starts the jar.
Configure the service server environment
All application servers need to be configured except the Jenkins server
Set the environment variables used by the program
Sh must be created in the /var/env directory. The following startup and restart scripts will load /var/env.sh.
These environment variables are used by our program eureka Server and ConfigServer, so they need to be configured. If not, do not configure it.
Go to one of the application projects and view the directory:
Bak: The directory of the previous JAR is automatically backed up at build time
Bin: Two shell scripts, start.sh and restart.sh, are required. For Jenkins startup, two scripts need to be placed in /var/env. Start. sh and restart.sh in /var/env will be copied to the bin directory. If you manually start the program, add it in bin.
Lib: The actual running JAR, which will be placed here automatically when building
Logs: Used to store logs. Whether to store logs depends on the log path configuration in the application
[root@localhost services]# cd zhirui-zuul-server-zmgj/ [root@localhost zhirui-zuul-server-zmgj]# ll total 4 drwxr-xr-x. 2 root root 4096 Jul 24 14:02 bak drwxr-xr-x. 2 root root 40 Jul 1 23:09 bin drwxr-xr-x. 2 root root 51 Jul 24 14:02 lib drwxr-xr-x. 2 root root 6 Jul 1 23:09 logs [root@localhost zhirui-zuul-server-zmgj]#Copy the code
scripting
Sh and start.sh are as follows:
Create the file directly after the copy and paste can be, the script has loaded the above environment variables:
