I'd like to set an env variable in one Stage and have it available in all subsequent Stages and Steps. Something like this:
pipeline {
stages {
stage('One') {
steps {
sh 'export MY_NAME=$(whoami)'
}
}
stage('Two') {
steps {
sh 'echo "I am ${MY_NAME}"'
}
}
stage('Three') {
steps {
sh 'echo "I am ${MY_NAME}"'
}
}
}
}
Those sh steps seem to be independent of each other, and the exported var is not preserved even for the next Step, let alone Stage.
One way I can think of is to write the var to a shell file, like echo "FOLDER_CONTENT=$(ls -lh)" and then source it a next Step, but again, I'll have to do the sourcing in every next Step, which is suboptimal.
Is there a better way to do that?
Finally was able to achieve it like so:
pipeline {
stages {
stage('One') {
steps {
script {
env.MY_NAME= sh (
script: 'whoami',
returnStdout: true
).trim()
}
}
}
stage('Two') {
steps {
echo "I am ${MY_NAME}"
}
}
stage('Three') {
steps {
sh 'echo "I am ${MY_NAME}"'
}
}
}
}
Related
Let's say I have a declarative pipeline. I want to run a stage only when 'Restart from stage' icon is used ?
Is there a way to do this (a method, a variable...)? I want to run the stage only if "Restart from stage" is used
stage('Test') {
when {
expression {
// An expression to detect if Restart from this stage is used
}
}
steps {
sh 'echo 1'
}
}
You can define a global variable that will hold a Boolean value representing if the pipeline was executed from the beginning or from a specific stage, update it in your first stage and use it later on in the when condition to determine if a restart from stage has occurred.
Something like:
RESTART = true
pipeline {
agent any
stages {
stage('Setup') {
steps {
script{
// signaling pipeline was executed from the beginning (first stage)
RESTART = false
}
// other setup steps
}
}
stage('Test') {
when {
expression { return RESTART }
}
steps {
sh 'echo 1'
}
}
}
}
Another nice option based on #Pamela's answer for using a cause condition, is to use the built in triggeredBy option in the when directive, thus avoiding the need to use getBuildCauses() and the need to filter all causes, and instaed get the condition out of the box.
Something like:
stage('Test') {
when { triggeredBy 'RestartDeclarativePipelineCause' }
steps {
sh 'echo 1'
}
}
You can use currentBuild.getBuildCauses(): https://www.jenkins.io/doc/pipeline/examples/#get-build-cause
Then, in your Test stage add when expression checking the cause of the build matches the one you need.
stage('Test') {
when {
expression {
return currentBuild.getBuildCauses().any { cause ->
cause._class == 'org.jenkinsci.plugins.pipeline.modeldefinition.causes.RestartDeclarativePipelineCause'
}
}
}
steps {
sh 'echo 1'
}
}
I am working on a groovy script for a Jenkins pipeline and am struggling to find how to pass a variable across stages when the variable is obtained from a remote ssh connection.
I found Example 1 and Example 2 on this site and I want to merge them together as seen in "My attempt" below. Note that the output of the file on the remote server is 4. I'm trying pass 4 to a_var.
Example 1: works fine. SSH connection. This reads the file and outputs value to the Jenkins console
def sshCredId = 'myid_cred'
def sshUser = 'myid'
def sshServer = 'myserver'
pipeline {
agent { label 'docker-maven-slave' }
stages {
stage('one') {
steps {
script {
sshagent([sshCredId]){
sh "ssh -o StrictHostKeyChecking=no ${sshUser}#${sshServer} cat /mydir/myfile.csv"
}
}
}
}
stage('two') {
steps {
echo "something"
}
}
stage('three') {
steps {
echo "do stuff"
}
}
}
}
Example 2: works fine. This passes a parameter across stages
pipeline {
agent {
label 'docker-maven-slave'
}
parameters {
string(name: 'a_var', defaultValue: '')
}
stages {
stage("one") {
steps {
script {
tmp_param = sh (script: 'echo something', returnStdout: true).trim()
env.a_var = tmp_param
}
}
}
stage("two") {
steps {
echo "${env.a_var}"
}
}
}
}
**My attempt: stage two output is null. I'm expecting '4' **
def sshCredId = 'myid_cred'
def sshUser = 'myid'
def sshServer = 'myserver'
pipeline {
agent { label 'docker-maven-slave' }
parameters {
string(name: 'a_var', defaultValue: 'nothing')
}
stages {
stage('one') {
steps {
script {
tmp_param=sshagent([sshCredId]){
sh "ssh -o StrictHostKeyChecking=no ${sshUser}#${sshServer} cat /mydir/myfile.csv"
}
env.a_var=tmp_param
}
}
}
stage('two') {
steps {
echo "${env.a_var}"
}
}
stage('three') {
steps {
echo "do stuff"
}
}
}
}
Update the answer based on comments and feedback from MayJoAnneBeth
Try below snippet
sshagent([sshCredId]){
env.a_var = sh (script: "ssh -o StrictHostKeyChecking=no ${sshUser}#${sshServer} cat /mydir/myfile.csv", returnStdout: true).trim()
}
Ｉhave noticed that Jenkins pipeline file -- Jenkinsfile which have two syntax
Declarative
Scripted
I have made Declarative Script work to specify node to run my task. However I don't know how to modify my script to Scripted syntax.
My Declarative Script
pipeline {
agent none
stages {
stage('Build') {
agent { label 'my-label' }
steps {
echo 'Building..'
sh '''
'''
}
}
stage('Test') {
agent { label 'my-label' }
steps {
echo 'Testing..'
sh '''
'''
}
}
stage('Deploy') {
agent { label 'my-label' }
steps {
echo 'Deploying....'
sh '''
'''
}
}
}
}
I have tried to use in this way:
node('my-label') {
stage 'SCM'
git xxxx
stage 'Build'
sh ''' '''
}
But it seems Jenkins cannot find my node to run.
How about this simple example?
stage("one") {
node("linux") {
echo "One"
}
}
stage("two") {
node("linux") {
echo "two"
}
}
stage("three") {
node("linux") {
echo "three"
}
}
Or the below answer, this way you are guaranteed to have the stages run on the same node if there are multiple nodes with the same label and run interrupted by another job.
The above example will release the node after every stage, the below example will hold the node for all three stages.
node("linux") {
stage("one") {
echo "One"
}
stage("two") {
echo "two"
}
stage("three") {
echo "three"
}
}
I have strange problem during build. This is my initial Jenkinsfiles:
pipeline {
agent none
environment {
MAVEN_ARGS = "${HOST}"
}
stages {
stage('Test step') {
agent {
docker {
image 'maven:3-alpine'
}
}
steps {
echo "${HOST}"
echo "${env.HOST}"
echo "${MAVEN_ARGS}"
}
}
}
}
Why first two echos prints correct value of HOST variable and last echo prints null?
What is interesting when I delete stage agent section:
pipeline {
agent any
environment {
MAVEN_ARGS = "${HOST}"
}
stages {
stage('Test step') {
steps {
echo "${HOST}"
echo "${env.HOST}"
echo "${MAVEN_ARGS}"
}
}
}
}
Every single echo is printing correctly HOST variable - this is obvious.
Thanks for help :)
I am creating a sample jenkins pipeline, here is the code.
pipeline {
agent any
stages {
stage('test') {
steps {
sh 'echo hello'
}
}
stage('test1') {
steps {
sh 'echo $TEST'
}
}
stage('test3') {
if (env.BRANCH_NAME == 'master') {
echo 'I only execute on the master branch'
} else {
echo 'I execute elsewhere'
}
}
}
}
this pipeline fails with following error logs
Started by user admin
org.codehaus.groovy.control.MultipleCompilationErrorsException: startup failed:
WorkflowScript: 15: Not a valid stage section definition: "if (env.BRANCH_NAME == 'master') {
echo 'I only execute on the master branch'
} else {
echo 'I execute elsewhere'
}". Some extra configuration is required. # line 15, column 9.
stage('test3') {
^
WorkflowScript: 15: Nothing to execute within stage "test3" # line 15, column 9.
stage('test3') {
^
But when i execute the following example from this url, it executes successfully and print the else part.
node {
stage('Example') {
if (env.BRANCH_NAME == 'master') {
echo 'I only execute on the master branch'
} else {
echo 'I execute elsewhere'
}
}
}
The only difference i can see is that in the working example there is no stages but in my case it has.
What is wrong here, can anyone please suggest?
your first try is using declarative pipelines, and the second working one is using scripted pipelines. you need to enclose steps in a steps declaration, and you can't use if as a top-level step in declarative, so you need to wrap it in a script step. here's a working declarative version:
pipeline {
agent any
stages {
stage('test') {
steps {
sh 'echo hello'
}
}
stage('test1') {
steps {
sh 'echo $TEST'
}
}
stage('test3') {
steps {
script {
if (env.BRANCH_NAME == 'master') {
echo 'I only execute on the master branch'
} else {
echo 'I execute elsewhere'
}
}
}
}
}
}
you can simplify this and potentially avoid the if statement (as long as you don't need the else) by using "when". See "when directive" at https://jenkins.io/doc/book/pipeline/syntax/. you can also validate jenkinsfiles using the jenkins rest api. it's super sweet. have fun with declarative pipelines in jenkins!
It requires a bit of rearranging, but when does a good job to replace conditionals above. Here's the example from above written using the declarative syntax. Note that test3 stage is now two different stages. One that runs on the master branch and one that runs on anything else.
stage ('Test 3: Master') {
when { branch 'master' }
steps {
echo 'I only execute on the master branch.'
}
}
stage ('Test 3: Dev') {
when { not { branch 'master' } }
steps {
echo 'I execute on non-master branches.'
}
}
If you wanted to create a condition to execute only a stage based on expression you can use keyword when
stage ('test3'){
when { expression { return env.BRANCH_NAME == 'master'} }
steps {
echo 'I only execute on the master branch.'
}
}
}
With the expression key word you can add any condition.
e.g. if stage is dependent on generated file in workspace.
stage ('File Dependent stage'){
when { expression { return fileExists ('myfile') } }
steps {
echo "file exists"
}
}
}
if ( params.build_deploy == '1' ) {
println "build_deploy 是 ${params.build_deploy}"
jobB = build job: 'k8s-core-user_deploy', propagate: false, wait: true, parameters: [
string(name:'environment', value: "${params.environment}"),
string(name:'branch_name', value: "${params.branch_name}"),
string(name:'service_name', value: "${params.service_name}"),
]
println jobB.getResult()
}