My Jenkins pipeline is as follow:
pipeline {
triggers {
cron('H */5 * * *')
}
stages {
stage('Foo') {
...
}
}
}
The repository is part of a Github Organization on Jenkins - every branch or PR pushed results in a Jenkins job being created for that branch or PR.
I would like the trigger to only be run on the "main" branch because we don't need all branches and PRs to be run on a cron schedule; we only need them to be run on new commits which they already do.
Is it possible?
yes - it's possible. To schedule cron trigger only for a specific branch you can do it like this in your Jenkinsfile:
String cron_string = (scm.branches[0].name == "main") ? 'H */5 * * *' : ''
pipeline {
triggers {
cron(cron_string)
}
// whatever other code, options, stages etc. is in your pipeline ...
}
What it does:
Initialize a variable based on a branch name. For main branch it sets requested cron configuration, otherwise there's no scheduling (empty string is set).
Use this variable within pipeline
Further comments:
it's possible to use it also with parameterizedCron (in a case you'd want / need to).
you can use also some other variables for getting branch name, e.g: env.BRANCH_NAME instead of scm.branches[0].name. Whatever fits your needs...
This topic and solution is discussed also in Jenkins community: https://issues.jenkins.io/browse/JENKINS-42643?focusedCommentId=293221#comment-293221
EDIT: actually a similar question that leads to the same configuration - here on Stack: "Build Periodically" with a Multi-branch Pipeline in Jenkins
You can simply add a when condition to your pipeline.
when { branch 'main' }
I've been trying to construct multiple jobs from a list and everything seems to be working as expected. But as soon as I execute the first build (which works correctly) the parameters in the job disappears. This is how I've constructed the pipelineJob for the project.
import javaposse.jobdsl.dsl.DslFactory
def repositories = [
[
id : 'jenkins-test',
name : 'jenkins-test',
displayName: 'Jenkins Test',
repo : 'ssh://<JENKINS_BASE_URL>/<PROJECT_SLUG>/jenkins-test.git'
]
]
DslFactory dslFactory = this as DslFactory
repositories.each { repository ->
pipelineJob(repository.name) {
parameters {
stringParam("BRANCH", "master", "")
}
logRotator{
numToKeep(30)
}
authenticationToken('<TOKEN_MATCHES_WITH_THE_BITBUCKET_POST_RECEIVE_HOOK>')
displayName(repository.displayName)
description("Builds deploy pipelines for ${repository.displayName}")
definition {
cpsScm {
scm {
git {
branch('${BRANCH}')
remote {
url(repository.repo)
credentials('<CREDENTIAL_NAME>')
}
extensions {
localBranch('${BRANCH}')
wipeOutWorkspace()
cloneOptions {
noTags(false)
}
}
}
scriptPath('Jenkinsfile)
}
}
}
}
}
After running the above script, all the required jobs are created successfully. But then once I build any job, the parameters disappear.
After that when I run the seed job again, the job starts showing the parameter. I'm having a hard time figuring out where the problem is.
I've tried many things but nothing works. Would appreciate any help. Thanks.
This comment helped me to figure out similar issue with my .groovy file:
I called parameters property twice (one at the node start and then tried to set other parameters in if block), so the latter has overwritten the initial parameters.
BTW, as per the comments in the linked ticket, it is an issue with both scripted and declarative pipelines.
Fixed by providing all job parameters in each parameters call - for the case with ifs.
Though I don't see repeated calls in the code you've provided, please check the full groovy files for your jobs and add all parameters to all parameters {} blocks.
I've a multi configuration project on Jenkins(ver. 2.222.1) We use JenkinsDSL(v1.77) to create the free-style jobs from the groovy scripts. Email notifications is with extendedEmail(v2.69)
It seems like trigger mode for notifications cannot be specified in the groovy script but it could be done via the UI as below(ONLY_PARENT or ONLY_CONFIGURATIONS or BOTH)
As I see here, below are the configurations we could specify for extendedEmail in groovy for DSL plugin
job('example') {
publishers {
extendedEmail {
recipientList('me#halfempty.org')
defaultSubject('Oops')
defaultContent('Something broken')
contentType('text/html')
triggers {
beforeBuild()
stillUnstable {
subject('Subject')
content('Body')
sendTo {
developers()
requester()
culprits()
}
}
}
}
}
}
It does not say anything in case of multi configuration project where I need the notifications to be sent only if parent job fails
I believe it exists on hipChatNotifier where we could specify as matrixTriggerMode
Is there a way to do this via extendedEmail publisher as we could do it on the UI?
Seems that this isn't supported by the Job DSL plugin, yet.
You can work around by using a custom configure block. For example
publishers {
extendedEmail {
...
}
}
configure { project -> project / 'publishers' / 'hudson.plugins.emailext.ExtendedEmailPublisher' << {
matrixTriggerMode('ONLY_PARENT')
}
}
Note the << to append a node to the publisher configuration instead of overwriting it.
I have a Jenkinsfile that I've set up with a cron for a pipelineTriggers parameter. I can't seem to figure out how to disable the job from building from a merge to the master branch of the repo. Is there a way in the Jenkinsfile to disable the automatic build from an scm change?
If you're using a Multibranch Pipeline, you should be able to do this on the job's Configure page:
Scroll down to "Branch Sources"
Under "Property strategy", choose "Named branches get different properties"
Click "Add exception", enter "master" as the branch name
Click "Add property", choose "Suppress automatic SCM triggering"
Save
That would prevent changes to the master branch from triggering a build of the corresponding job.
For declarative pipelines, use the when directive with a triggeredBy condition, e.g.
when { triggeredBy 'TimerTrigger' }
With the multibranch pipeline, I could not figure out a way to prevent the next build being triggered. As a workaround, I added the following code to my Jenkinsfile (using scripted syntax), to abort the following build if the only changes contain "[ci-skip]" in the commit message:
def abortBuildIfTriggeredBySkippableCommit() {
def changeSetCount = 0;
def ciSkipCount = 0;
if (currentBuild.changeSets != null) {
for (changeSetList in currentBuild.changeSets) {
for (changeSet in changeSetList) {
changeSetCount++;
if (changeSet.msg.contains('[ci-skip]')) {
ciSkipCount++;
}
}
}
}
if (changeSetCount > 0 && changeSetCount == ciSkipCount) {
currentBuild.result = 'NOT_BUILT'
error("Stopping to prevent auto trigger. All commits contained [ci-skip]")
}
}
Note that this code assumes you are using the git plugin, and that the objects in currentBuild.changeSets will be GitChangeSetList.
In a jenkins job you can navigate to advanced source code management
Select behavior Dont trigger build on commit notification
This disables the Started by an SCM change
For people still looking for a solution, go to configuration for the multi branch pipeline, under Property Strategy, choose "Suppress Automatic SCM Triggering".
Note: This is available on cloudbees version of Jenkins. I am not sure, if it matters.
https://support.cloudbees.com/hc/en-us/articles/360026953651-Preventing-builds-from-getting-triggered-when-creating-a-new-multibranch-Pipeline-or-Organization-Folder?page=29
This is what I came up with. I was hoping for something less messy, but this does seem to work:
I have this as the build's properties:
properties([
pipelineTriggers([cron('H H 7 * *')])
])
I then have this function that defines the source of the build:
// check if the job was started by a timer
#NonCPS
def jobStartedByWhat() {
def startedByWhat = ''
try {
def buildCauses = currentBuild.rawBuild.getCauses()
for ( buildCause in buildCauses ) {
if (buildCause != null) {
def causeDescription = buildCause.getShortDescription()
echo "shortDescription: ${causeDescription}"
if (causeDescription.contains("Started by timer")) {
startedByWhat = 'timer'
}
if (causeDescription.contains("Started by user")) {
startedByWhat = 'user'
}
}
}
} catch(theError) {
echo "Error getting build cause: ${theError}"
}
return startedByWhat
}
def startedByWhat = jobStartedByWhat()
I can then evaluate the function at runtime so that if a build gets triggered because of a merge to master, it will not actually run:
node {
try {
checkout scm
if (env.BRANCH_NAME == 'master') {
if (startedByWhat == 'timer' || startedByWhat == 'user') {
..... RUN THE BUILD .....
} else {
.... EXIT WITHOUT RUNNING THE BUILD ....
I stumbled upon this as well. IMO an acceptable solution would be a filter for commit messages when checking out source code - this feature exists for regular Jobs but is missing for multibranch pipelines, see https://issues.jenkins-ci.org/browse/JENKINS-48296.
For those not using the git plugin, this method is a workaround for scripted pipelines (inspired by scarswell's answer):
def abortBuildIfTriggeredBySkippableCommit() {
lastCommitMessage = sh(
script: "${gitBinary} --no-pager log -1 --pretty=%B",
returnStdout: true
)
if (lastCommitMessage != null &&
lastCommitMessage.toString().contains('[maven-release-plugin]')) {
currentBuild.result = 'ABORTED'
error("Stopping build, it was triggered by the maven release plugin")
}
}
For declarative pipelines, there is a much more simple answer now. From the docs:
overrideIndexTriggers
Allows overriding default treatment of branch indexing triggers. If branch indexing triggers are disabled at the multibranch or organization label, options { overrideIndexTriggers(true) } will enable them for this job only. Otherwise, options { overrideIndexTriggers(false) } will disable branch indexing triggers for this job only.
It's a little backwards conceptually, but assuming your jobs are triggering off github webhooks by default, you set overrideIndexTriggers(false) to disable the automatic triggering.
If you are using Pipeline script from SCM then comment out the triggers section(either SCMPoll/BuildPeriodically option ) in Jenkins file as shown below.
//triggers {cron ('H/15 * * * *')}
//pipelineTriggers([pollSCM('H/15 * * * *')])
If you are using Pipeline script then disable the PollSCM/Build periodically(whichever is used) option.
One could disable the scm build trigger by disabling the webhook notification from git.
if (currentBuild.getBuildCauses().toString().contains('BranchIndexingCause') || currentBuild.getBuildCauses().toString().contains('Branch event')) {
print "INFO: Build skipped due to trigger being Branch Indexing"
currentBuild.result = 'ABORTED' // optional, gives a better hint to the user that it's been skipped, rather than the default which shows it's successful
return
}
timestamps{
node
{
stage('Getting Build Parameters')
{
print('build job')
}
}
}
if (currentBuild.getBuildCauses().toString().contains('BranchIndexingCause') || currentBuild.getBuildCauses().toString().contains('Branch event')) {
print "INFO: Build skipped due to trigger being Branch Indexing"
currentBuild.result = 'ABORTED' // optional, gives a better hint to the user that it's been skipped, rather than the default which shows it's successful
return
}
if (currentBuild.getBuildCauses().toString().contains('BranchIndexingCause') || currentBuild.getBuildCauses().toString().contains('Branch event')) {
print "INFO: Build skipped due to trigger being Branch Indexing"
currentBuild.result = 'ABORTED' // optional, gives a better hint to the user that it's been skipped, rather than the default which shows it's successful
return
}
I have a project, named Demo, which doesn't do anything in particular.
I have a DSL script, like the following:
def gitUrl = 'GIT_URL'
job('unit-tests') {
scm {
git(gitUrl)
}
triggers {
buildResult('H/* * * * *') {
combinedJobs()
triggerInfo('Demo', BuildResult.SUCCESS, BuildResult.UNSTABLE)
}
}
}
Now what I'm wanting to do, is that when the Demo project runs successfully (it checks out a PHP application from Github), I want the unit-tests job to run.
Currently, when the Demo project is built, the unit-tests job never gets run.
I'm guessing my DSL script is incorrect, but I'm not sure why
I can reproduce your problem. The check box is not set when running the seed job for the first time. But it's set after running the seed job a second time. Must be a problem in the BuildResultTrigger plugin. Please file a bug report in the Jenkins JIRA: https://issues.jenkins-ci.org/projects/JENKINS
But you do not necessarily need to use the BuildResultTrigger plugin. You can use the built-in "Build after other projects are built" option, see https://jenkinsci.github.io/job-dsl-plugin/#path/job-triggers-upstream.
job('unit-tests') {
triggers {
upstream('Demo', 'UNSTABLE')
}
}
Use upstream which adds the "Build after other projects are built" trigger. see https://jenkinsci.github.io/job-dsl-plugin/#path/job-triggers-upstream
def gitUrl = 'GIT_URL'
job('unit-tests') {
scm {
git(gitUrl)
}
triggers {
buildResult('H/* * * * *') {
upstream('Demo', 'UNSTABLE')
}
}
}