I've put all my Jenkins logic in a structured pipeline script (aka Jenkinsfile).
If something goes wrong, i m sending mails. For the subject i want to use the displayName of the job and not the jobs id env.JOB_NAME (as they are driven by access control patterns and not readability).
With a normal pipeline job i could use currentBuild.rawBuild.project.displayName but for multibranch pipelines this is just the branch name.
Or is there a even better way to get the userfriendly name, then traversing the rawBuild?
For now i found no convinient public api, so this seems to be the way to go:
String getDisplayName(currentBuild) {
def project = currentBuild.rawBuild.project
// for multibranch pipelines
if (project.parent instanceof WorkflowMultiBranchProject) {
return "${project.parent.displayName} (${project.displayName})"
} else {
// for all other projects
return project.displayName
}
}
I use currentBuild.fullProjectName which is set to multibranch_pipeline_name/branch_name or pipeline_name depending if you are using a multibranch pipeline or normal pipeline.
Related
I have a multibranch pipeline that runs a Jenkinsfile of select branches. Now I need to run that same Jenkinsfile with parameters, so I figured I could use a regular pipeline.
Now all I have to do is to determine whether I run in a multibranch pipeline or not. I could check for any parameters in the build, and when there aren't any I could deduce that I'm in a multibranch pipeline:
def isMultibranchPipeline() {
!params.any()
}
I was searching for a more direct method to know whether the script is running in a multibranch pipeline or not, but couldn't find anything like it.
By getting the current "project" (which is a Jenkins job) you're able to know if it's a multibranch job or not thanks to its class:
import jenkins.model.Jenkins
import org.jenkinsci.plugins.workflow.multibranch.WorkflowMultiBranchProject
def isMultiBranchPipeline() {
return Jenkins.get().getItem(currentBuild.projectName) instanceof WorkflowMultiBranchProject
}
if job is PipelineMultiBranchDefaultsProject previous approach is not working. I'm
checking with scm object
boolean isMultiBranchPipeline() {
try {
if (script.scm)
return true
} catch (Exception e) {
return false
}
}
I would like to set a parameter in Jenkins Declarative Pipeline enabling the user to select one of the jobs defined on Jenkins. Something like:
parameters {
choice(choices: getJenkinsJobs())
}
How can this be achieved?
Background info: I would like to implement a generic manual promotion job with the Pipeline, where the user would select a build number and the job name and the job would get promoted.
I dislike the idea of using the input step as it prevents the job from completing and I can't get e.g. the junit reports on tests.
You can iterate over all existing hudson.model.Job instances and get their names. The following should work
#NonCPS
def getJenkinsJobs() {
Jenkins.instance.getAllItems(hudson.model.Job)*.fullName.join('\n')
}
pipeline {
agent any
parameters {
choice(choices: getJenkinsJobs(), name: 'JOB')
}
//...
}
Use http://wiki.jenkins-ci.org/display/JENKINS/Extended+Choice+Parameter+plugin
and use basic groovy script as a input.
Refer the below URL for how to list the build/jobs.
https://themettlemonkey.wordpress.com/2013/01/29/jenkins-build-number-drop-down/
I have Jenkins Pipeline jobs, where the only difference between the jobs is a parameter, a single "name" value, I could even use the multibranch job name (though not what it's passing as JOB_NAME which is the BRANCH name, sadly none of the envs look suitable without parsing). It would be great if I could set this outiside of the Jenkinsfile, since then I could reuse the same jenkinsfile for all the various jobs.
Add this to your Jenkinsfile:
properties([
parameters([
string(name: 'myParam', defaultValue: '')
])
])
Then, once the build has run once, you will see the "build with parameters" button on the job UI.
There you can input the parameter value you want.
In the pipeline script you can reference it with params.myParam
Basically you need to create a jenkins shared library example name myCoolLib and have a full declarative pipeline in one file under vars, let say you call the file myFancyPipeline.groovy.
Wanted to write my examples but actually I see the docs are quite nice, so I'll copy from there. First the myFancyPipeline.groovy
def call(int buildNumber) {
if (buildNumber % 2 == 0) {
pipeline {
agent any
stages {
stage('Even Stage') {
steps {
echo "The build number is even"
}
}
}
}
} else {
pipeline {
agent any
stages {
stage('Odd Stage') {
steps {
echo "The build number is odd"
}
}
}
}
}
}
and then aJenkinsfile that uses it (now has 2 lines)
#Library('myCoolLib') _
evenOrOdd(currentBuild.getNumber())
Obviously parameter here is of type int, but it can be any number of parameters of any type.
I use this approach and have one of the groovy scripts that has 3 parameters (2 Strings and an int) and have 15-20 Jenkinsfiles that use that script via shared library and it's perfect. Motivation is of course one of the most basic rules in any programming (not a quote but goes something like): If you have "same code" at 2 different places, something is not right.
There is an option This project is parameterized in your pipeline job configuration. Write variable name and a default value if you wish. In pipeline access this variable with env.variable_name
I am looking at limiting the number of concurrent builds to a specific number in Jenkins, leveraging the multibranch pipeline workflow but haven't found any good way to do this in the docs or google.
Some docs say this can be accomplished using concurrency in the stage step of a Jenkinsfile but I've also read elsewhere that that is a deprecated way of doing it.
It looks like there was something released fairly recently for limiting concurrency via Job Properties but I couldn't find documentation for it and I'm having trouble following the code. The only thing I found a PR that shows the following:
properties([concurrentBuilds(false)])
But I am having trouble getting it working.
Does anybody know or have a good example of how to limit the number of concurrent builds for a given, multibranch project? Maybe a Jenkinsfile snippet that shows how to limit or cap the number of multibranch concurrent builds?
Found what I was looking for. You can limit the concurrent builds using the following block in your Jenkinsfile.
node {
// This limits build concurrency to 1 per branch
properties([disableConcurrentBuilds()])
//do stuff
...
}
The same can be achieved with a declarative syntax:
pipeline {
options {
disableConcurrentBuilds()
}
}
Limiting concurrent builds or stages are possible with the Lockable Resources Plugin (GitHub). I always use this mechanism to ensure that no publishing/release step is executed at the same time, while normal stages can be build concurrently.
echo 'Starting'
lock('my-resource-name') {
echo 'Do something here that requires unique access to the resource'
// any other build will wait until the one locking the resource leaves this block
}
echo 'Finish'
As #VadminKotov indicated it is possible to disable concurrentbuilds using jenkins declarative pipelines as well:
pipeline {
agent any
options { disableConcurrentBuilds() }
stages {
stage('Build') {
steps {
echo 'Hello Jenkins Declarative Pipeline'
}
}
}
}
disableConcurrentBuilds
Disallow concurrent executions of the Pipeline. Can be useful for
preventing simultaneous accesses to shared resources, etc. For
example: options { disableConcurrentBuilds() }
Thanks Jazzschmidt, I looking to lock all stages easily, this works for me (source)
pipeline {
agent any
options {
lock('shared_resource_lock')
}
...
...
}
I got the solution for multibranch locking too, with de lockable-resources plugin and the shared-libs here is :
Jenkinsfile :
#Library('my_pipeline_lib#master') _
myLockablePipeline()
myLockablePipeline.groovy :
call(Map config){
def jobIdentifier = env.JOB_NAME.tokenize('/') as String[];
def projectName = jobIdentifier[0];
def repoName = jobIdentifier[1];
def branchName = jobIdentifier[2];
//now you can use either part of the jobIdentifier to lock or
limit the concurrents build
//here I choose to lock any concurrent build for PR but you can choose all
if(branchName.startsWith("PR-")){
lock(projectName+"/"+repoName){
yourTruePipelineFromYourSharedLib(config);
}
}else{
// Others branches can freely concurrently build
yourTruePipelineFromYourSharedLib(config);
}
}
To lock for all branches just do in myLockablePipeline.groovy :
call(Map config){
def jobIdentifier = env.JOB_NAME.tokenize('/') as String[];
def projectName = jobIdentifier[0];
def repoName = jobIdentifier[1];
def branchName = jobIdentifier[2];
lock(projectName+"/"+repoName){
yourTruePipelineFromYourSharedLib(config);
}
}
I am trying to do a poc of jenkins pipeline as code. I am using the Github organization folder plugin to scan Github orgs and create jobs per branch. Is there a way to explicitly define the names for the pipeline jobs that get from Jenkinsfile? I also want to add some descriptions for the jobs.
You need to use currentBuild like below. The node part is important
node {
currentBuild.displayName = "$yournamevariable-$another"
currentBuild.description = "$yourdescriptionvariable-$another"
}
Edit: Above one renames build where as Original question is about renaming jobs.
Following script in pipeline will do that(this requires appropriate permissions)
item = Jenkins.instance.getItemByFullName("originalJobName")
item.setDescription("This description was changed by script")
item.save()
item.renameTo("newJobName")
I'm late to the party on this one, but this question forced me in the #jenkins chat where I spent most of my day today. I would like to thank #tang^ from that chat for helping solve this in a graceful way for my situation.
To set the JOB description and JOB display name for a child in a multi-branch DECLARATIVE pipeline use the following steps block in a stage:
steps {
script {
if(currentBuild.rawBuild.project.displayName != 'jobName') {
currentBuild.rawBuild.project.description = 'NEW JOB DESCRIPTION'
currentBuild.rawBuild.project.setDisplayName('NEW JOB DISPLAY NAME')
}
else {
echo 'Name change not required'
}
}
}
This will require that you approve the individual script calls through the Jenkins sandbox approval method, but it was far simpler than anything else I'd found across the web about renaming the actual children of the parent pipeline. The last thing to note is that this should work in a Jenkinsfile where you can use the environment variables to manipulate the job items being set.
I tried to used code snippet from accepted answer to describe my Jenkins pipeline in Jenkinsfile. I had to wrap code snippet into function with #NonCPS annotation and use def for item variable. I have placed code snippet in root of Jenkinsfile, not in node section.
#NonCPS
def setDescription() {
def item = Jenkins.instance.getItemByFullName(env.JOB_NAME)
item.setDescription("Some description.")
item.save()
}
setDescription()