I am using Jenkins version 1.644 and trying to deploy a web application to Websphere 8.5 application server. Jenkins job completed successfully and application is visible through admin console. After the first install, i manually configured Three application configurations namely,
1. Virtual Host
2. Context Root and
3 Modules
after these setup application comes up fine.
Now when i run the Jenkins Job again (option used is Install/Update application), it overrides all the configurations.
Please Let me know how to keep the configurations after each build from Jenkins.
Websphere Plugin Configuration
You can create a build deploy job which will call wsadmin tool and there you can pass parameter in key value pairs
Here is an article which talks about how to build job with parameterized configuration.
http://www.touchdownconsulting.nl/2011/03/building-and-deploying-websphere-applications-with-jenkins-ci/
I have not tried this but looks like it suits your requirement.
Hope this helps!
Current Jenkins Websphere deploy plugin (1.3.4) version does not allow to pass
1. Virtual Host
2. Context Root and
3. Modules
I created a Jython script using AdminApp WAS Utility and updated these parameters
AdminApp.edit("appname", ['-MapWebModToVH', [["appname", "appname.war,WEB-INF/web.xml", "api_host"]]])
AdminApp.edit("appname", ['-CtxRootForWebMod', [["appname", "appname.war,WEB-INF/web.xml", "/appname"]]])
AdminApp.edit("appname",['-MapModulesToServers', [["appname","appname.war,WEB-INF/web.xml","WebSphere:cell=appcell01,node=node12v,server=web2+WebSphere:cell=Cell01,node=node11v,server=web1+WebSphere:cell=Cell01,cluster=api-cluster"]]])
AdminConfig.save()
Used Jenkins Remote SSH Plugin to invoke this script.
Related
I want to deploy a Spring Boot application on an OpenShift cluster that I want to monitor with elastic-apm, therefore, with the JAVA elastic agent.
I managed to deploy in a project an Elasticsearch instance, a Kibana instance and an apm-server.
Next to that, I also managed to deploy my Spring Boot application. For this I used the web console. I imported my project from GitLab, and chose the Java 8 image builder. However, using this method, I didn't find a way to launch my application by associating the java-agent elastic-apm-agent.
Locally, I run this command to start my application:
mvn package && java -javaagent:elastic-apm-agent/elastic-apm-agent-1.26.0.jar \
-Delastic.apm.service_name=ms-salarie \
-Delastic.apm.server_urls=http://localhost:8200 \
-Delastic.apm.secret_token= \
-Delastic.apm.environment=development \
-Delastic.apm.application_packages=com.leanerp.salarie \
-Delastic.apm.config_file=elastic-apm-agent/elasticapm.properties \
-jar target/salarie-1.1.3-SNAPSHOT.jar
Is there a way to override the command launched by the container of my application? Or another solution allowing me to use the elastic-apm-agent?
I am a newbie on OpenShift, so I don't fully understand all the concepts.
Ok, so the answer was adding this environment variable :
JAVA_OPTS_APPEND=-javaagent:{{path_to_elastic_apm_agent}}
this command allows you to launch your java application with options.
The Java agent allows multiple ways to configure it, one of which are command line system properties. Others include packaging an elasticapm.properties resource file or setting environment variables.
Check out the docs. Small excerpt:
Properties file: The elasticapm.properties file is located in the same folder as the agent jar, or provided through the config_file option. dynamic config.
Environment variables: All configuration keys are in uppercase and prefixed with ELASTIC_APM_.
Different option sources have different priority and precedence.
To attach the agent to a running JVM process (from within your application), you can use the API to self-attach.
I am currently building a java web application (with netbeans).
I use Jenkins to create a release version with the following pipeline (for Jenkins):
Build -> Test -> Deploy (to a remote test webserver)
Build and Test are OK but I have a question about the deploy job.
The deploy job is currently taking my previously generated .war file and simply transfer it to a remote web server (with the "Deploy to container" plugin).
But I would like to change the database parameters of my web application first ! (in order to use another remote test database).
I would be glad to modify the java file with shell command but I can't because my .war is only composed of the compiled .class java.
So how could I change some of my web application java code (for database credentials) from the .war file before deploying it to the remote web server ?
If you have multiple environments which have different databases, then the best way to handle this would be application with command line parameters. You can modify your java application to read the command line parameters and use these parameters in application.
For example --dburl = <database url> --dbusername= <db username>
And the another way will be take these paramters from environment variable. And define these variables in the system where you are deploying the applications.
The setup is as follows:
Step 1: I am developing on a windows 10 machine
Step 2: Building artifacts with Maven (using powershell)
Step 3: Go to (semi-local) Jenkins which is running in a Linux VM on my machine & trigger a Jenkins job that syncs and deploys my artifacts to the application server running on the VM.
Q: Is it possible to automate/merge Steps 2 and 3 in this scenario?
(Building and deploying each takes several minutes)
Yes, you can.
Create a custom goal which will trigger the Jenkins with an URL.
But, ideally, I would suggest to build and deploy using Jenkins by fetching the code from the repository when there is a change
Yes you can.
I suggest you use webhooks (github, gitlab, bitbucket,...)
Steps will be :
Install some plugin in your jenkins in order to expose a job as public url.
Configure your webhook in git provider adding url created in step 1
Push some changes to your code
Your git provider detect this event and execute the url created in step 1
Your jenkins job is launched. Put your mvn commands and deploy commands in this job. I suggest you use jenkins pipeline job.
Contact me if you need some help.
Post step by step :
https://jrichardsz.github.io/devops/devops-with-git-and-jenkins-using-webhooks
In our CI environment, we currently have one build server (based on Atlassian Bamboo) and two SonarQube instances (versions 6.0 and 6.5). Initially, our CI server was configured to communicate with the 6.0 SonarQube instance. This has been configured in the /home/bamboo/.gradle/gradle.properties file on our CI server like this:
systemProp.sonar.host.url=<http url of SonarQube 6.0 instance>
systemProp.sonar.login=<username here>
systemProp.sonar.password=<password here>
Now we have another Gradle-based project running on our CI server which shall talk to the new SonarQube 6.5 instance. I tried configuring this but failed all the time.
Things I have done so far:
Added commandline arguments to gradle wrapper command:
I have tried adding -Dsonar.host.url=, -Dsonar.login=, -Dsonar.password= to the Gradle command. As this didn't seem to work, I have also tried to set commandline arguments as SonarQube system properties using -DsystemProp.sonar.host.url=, -DsystemProp.sonar.login=, -DsystemProp.sonar.password=. This didn't work either.
Added properties to the build.gradle file
- Added properties to the build.gradle file like this:
sonarqube {
properties {
property "sonar.host.url", "<http url of SonarQube 6.0 instance>"
property "sonar.login", "<username here>"
property "sonar.password", "<password here>"
...<other SonarQube analysis settings here>...
}
}
In all cases, the CI server talked to the wrong SonarQube instance (6.0). My question is, whether it is possible to configure a single project to talk to another SonarQube instance. As you have seen, we use Gradle 3.2.1 as a build tool. And we are using the org.sonarqube Gradle plugin too.
Thank you for any help.
André
Your first try did not work, because you set the system properties from the commandline, but setting it from the project properties later on resets the system properties to the configured values.
Your second try did not work, because the systemProp.sonar.login syntax is only suppored in gradle.properties files, not via -P commandline project properties.
Your third try did not work because the SonarQube scanner prefers the system property values over the value configured via the DSL, so that one can change what is configured in the build script with the help of local configuration.
You need to set the system properties in your build script manually, this then overwrite what was automatically set from the project property. Using the project gradle.properties file does not work as the user file overwrite the project file. So you need something like System.properties.'sonar.login' = '...' in your build script. You can either hard-code it there, or then use project properties that you can set in your gradle.properties file or via -P parameters.
Besides that, I'd never depend on having any configuration in Gradle User dir on a build server. Most buildservers use build agents that might run on distributed machines, so you would always have to make sure that all build agents are configured the same and so on. I'd always configure in the build setup of the build server the according configuration, either by setting system properties, or environment properties or commandline arguments.
Just my 2ct.
I am trying to create a deploy job in Jenkins for my Spring Boot - Maven project. These are the steps i have followed until now:
Click New Item
Give a name, and choose Maven project
Configuration page opens
In Source Code Management, i have given my Git repository URL
In Build section under Goals and options i have written "spring-boot:run"
Saved and applied
Now when i click "Build Now", it checks out the project, builds and deploys. But the job does not end. This is the last line in the console output screen:
: Started Application in 4.737 seconds (JVM running for 16.297)
What is my problem? I need a simple step by step guidance since i do not have any Jenkins experience.
EDIT: I do not know what post build action is which is mentioned in the other post. I need a simple step by step guidance since i do not have any Jenkins experience.
The job doesn't end because your application is up and running on embeeded Tomcat. It's normal behaviour and you should be able to use your application at this point.
If you want to make another build, you should stop current build before starting a new one. To do this automaticaly (e.g. when your master branch is changed) you can create another one job.
Example of a job that triggers build of a main job:
Create a 'Freestyle project' in Jenkins;
Fill 'Source Code Management': your Repository URL and Credentials, Branches to build - */master;
Fill 'Build Triggers': select 'Poll SCM' and type H/2 * * * * in 'Shedule' to poll repository every 2 minutes;
Fill 'Build': add build step 'Execute shell' and type next commands to stop current build and start a new one (don't forget to replace JENKINS_URL and JOB_NAME with your values) -
curl -X POST http://JENKINS_URL/job/JOB_NAME/lastBuild/stop;
curl -X POST http://JENKINS_URL/job/JOB_NAME/build;
Save your job :)
I think that if you are going to deploy the spring-boot application in another server different to jenkins server you need to build your app mvn clean package and to move the jar created from jenkins server to the new server then you can use another step to start the app using java -jar myapp.jar. But, if you are going to deploy in tomcat server then you can use tomcat-api to deploy remotely.