Bdtyktl

I have a shared git repository used by 12 people. I want to rebuild every branch after the branch itself gets updated or a PR gets merged into developer.

I configured a multibranch pipeline and wrote a Jenkinsfile that achieves this goal.

This Jenkinsfile is in developer, and it's now been included into more than 60 branches. This was OK at first, but now that we have so many branches, over 60 builds get queued up every time we merge a PR. It can take a very long time for all of those builds to complete. Also, the builds use up a lot of disk space, causing my disk quota on the Uni computer to be exceeded.

To solve those problems, I would like to do the following:

• make Jenkins stop rebuilding the old branches. E.g. ignore branches that were last updated more than 2 weeks ago.




• Make Jenkins run "deleteDir()" in the post phase of all builds.

I can easily update my Jenkinsfile to achieve these goals. But then I have to put the new Jenkinsfile into every single branch I have! So that's where I'm stuck. I really don't want to have to update every single old branch by hand.

I've had some ideas for solutions, but none of them satisfy me. I could e.g.:

• Delete the old branches, but we can't do that, because this is a university module, and all branches should be kept in the repository as a kind of proof that one has been practicing writing code.




• Write a script to add the new Jenkinsfile to every single branch in our repo. This would be a kind of weird look. I don't want my name to be in every single branch.




• Write a script that fetches all branches from the main repo into a secondary repo, which I would then point the multibranch pipeline at. Then write another script that regularly deletes obsolete branches from this secondary repo. This sounds like so much work!! And it does not solve the issue of running deleteDir() at the end of every build.

Basically, I would like to find a way of reconfiguring my pipeline inside of Jenkins without individually updating the Jenkinsfile in every branch. Is there a way to do this?

Here is my Jenkinsfile, for reference. Could it be that I have to adapt it to be used in a "normal pipeline" rather than a "multibranch pipeline"?

def commitHash

pipeline {

    agent any

    triggers {

        // Every branch should be rebuilt after developer is built

        // This doesn't cause an infinite loop, because 'Developer builden' is a job that only specifically

        // gets triggered after a commit is pushed to developer.  The event will not be fired a second time

        // when developer is built as a result of this upstream trigger.

        upstream(upstreamProjects: 'Developer builden', threshold: hudson.model.Result.FAILURE)

    }



    stages {

        stage('Build') {

            steps {

                echo 'checkout scm'

                checkout scm

                script {

                    commitHash = sh(returnStdout: true, script: "git log -n 1 --pretty=format:'%H'").trim()

                }

                echo 'merging into origin/developer'

                sh 'git merge origin/developer'

                echo 'Building..'

            }

        }

        stage('Test') {

            steps {

                sh 'mvn clean package'

                echo 'Testing..'

            }

        }

        stage('Deploy') {

            steps {

                echo 'Deploying....'

            }

        }

    }



    post {

        always {

            script {

                currentBuild.result = currentBuild.result ?: 'SUCCESS'

                echo "build result: ${currentBuild.result}"

                echo "commitSha1: ${commitHash}"

                notifyBitbucket commitSha1: commitHash,

                        credentialsId: '<redacted>',

                        disableInprogressNotification: false,

                        considerUnstableAsSuccess: false,

                        ignoreUnverifiedSSLPeer: true,

                        includeBuildNumberInKey: false,

                        prependParentProjectKey: false,

                        projectKey: '',

                        stashServerBaseUrl: '<redacted>'

            }

            deleteDir()  // This line is, unfortunately, not present in most of our older branches.

        }

    }

}