Prioritize Jenkins Build via Groovy Postbuild - jenkins

I want to schedule a job to be at the top of the build queue in Jenkins via a Groovy Postbuild.
For example, for Build X
If previously the build queue looks like this:
Generic Build 1
Generic Build 2
Generic build 3
Now it should look like:
Build X
Generic Build 1
Generic Build 2
Generic Build 3
So far I've been able to schedule the job using
def waitingItem = Jenkins.get().getQueue().schedule(job, 0)
if (waitingItem == null) {
manager.listener.logger.println "Error scheduling ${job}!"
} else {
manager.listener.logger.println "${job.name} was scheduled!"
}
but now I also want waitingItem to be at the top of the queue.
Any help is much appreciated.

Ok so after countless hours of browsing the web I've been able to come up with a Groovy Postbuild script that does the following: once it finishes the build, it triggers another build (let's call it buildToBeTriggered) and buildToBeTriggered automatically gets pushed at the front of the queue.
The code is below:
import hudson.model.AbstractProject
import hudson.model.Queue
import hudson.model.queue.QueueSorter
import jenkins.model.Jenkins
def job = (AbstractProject)Jenkins.get().getItem('gcimpoies-toTrigger')
def isSuccess = manager.getResult() == 'SUCCESS'
def isRelease = manager.getEnvVariable('RELEASE') == 'true'
def secondsToWait = 20
if (isSuccess) {
def waitingItem = Jenkins.get().getQueue().schedule(job, 0)
if (waitingItem == null) {
manager.listener.logger.println "Error scheduling ${job}!"
} else {
manager.listener.logger.println "${job.name} was scheduled!"
}
Thread.sleep(secondsToWait * 1000)
//replace the original queue sorter with one that will place our project build first in the queue
QueueSorter originalQueueSorter = Jenkins.get().getQueue().getSorter()
AcceleratedBuildNowSorter acceleratedBuildNowSorter = new AcceleratedBuildNowSorter(job, originalQueueSorter)
Jenkins.get().getQueue().setSorter(acceleratedBuildNowSorter);
// we sort the queue so that our project is next to be built on the list
Jenkins.get().getQueue().getSorter().sortBuildableItems(Jenkins.getInstance().getQueue().getBuildableItems())
}
class AcceleratedBuildNowSorter extends QueueSorter {
private final AbstractProject project
private final QueueSorter originalQueueSorter
private final AcceleratedBuildNowComparator comparator
AcceleratedBuildNowSorter(AbstractProject project, QueueSorter originalQueueSorter) {
this.project = project
this.originalQueueSorter = originalQueueSorter
comparator = new AcceleratedBuildNowComparator(this.project)
}
#Override
void sortBuildableItems(List<Queue.BuildableItem> buildables) {
if (this.originalQueueSorter != null) {
this.originalQueueSorter.sortBuildableItems(buildables)
}
Collections.sort(buildables, comparator)
}
}
class AcceleratedBuildNowComparator implements Comparator<Queue.BuildableItem> {
private final AbstractProject mostPriorityProject;
AcceleratedBuildNowComparator(AbstractProject mostPriorityProject) {
this.mostPriorityProject = mostPriorityProject;
}
int compare(Queue.BuildableItem buildableItem0, Queue.BuildableItem buildableItem1) {
AbstractProject<?, ?> project0 = (AbstractProject<?, ?>) buildableItem0.task
AbstractProject<?, ?> project1 = (AbstractProject<?, ?>) buildableItem1.task
if (project0.equals(mostPriorityProject)) {
return -1
}
if (project1.equals(mostPriorityProject)) {
return 1
}
return 0
}
}

Related

How to change the Git URL in all Jenkins jobs

Does anyone have an updated version of ceilfors answer that works for both AbstractProject and WorkflowJob?
This is the solution I came up with. It was tested on Jenkins 2.355
The test was run from the script console.
For testing purposes, I limited the test to one Freestyle (AbstractProject) and one Pipeline (WorkflowJob) job each. You would need to modify the code below.
I hope others find this useful
import org.jenkinsci.plugins.workflow.job.WorkflowJob;
import hudson.plugins.git.*
import jenkins.*
import jenkins.model.*
def modifyGitUrl(url) {
def updatedUrl = url.toString().replace("git#gitlab", "git#github")
// println "updatedUrl = ${updatedUrl}"
return updatedUrl
}
Jenkins.instance.getAllItems(Job.class).each {
project = it.getFullName()
if(project.toString().equals("PL_Quick_Testing") || project.toString().equals("A_Freestyle_Job")) {
try {
if (it instanceof AbstractProject){
def oldScm = it.scm
def newUserRemoteConfigs = oldScm.userRemoteConfigs.collect {
new UserRemoteConfig(modifyGitUrl(it.url), it.name, it.refspec, it.credentialsId)
}
def newScm = new GitSCM(newUserRemoteConfigs, oldScm.branches, oldScm.doGenerateSubmoduleConfigurations,
oldScm.submoduleCfg, oldScm.browser, oldScm.gitTool, oldScm.extensions)
it.scm = newScm
it.save()
println "Done"
} else if (it instanceof WorkflowJob) {
def oldScm = it.getTypicalSCM()
def definition = it.getDefinition()
String scriptPath = it.getDefinition().getScriptPath()
def newUserRemoteConfigs = oldScm.userRemoteConfigs.collect {
new UserRemoteConfig(modifyGitUrl(oldScm.userRemoteConfigs.url[0]), it.name, it.refspec, it.credentialsId)
}
def newScm = new GitSCM(newUserRemoteConfigs, oldScm.branches, oldScm.doGenerateSubmoduleConfigurations,
oldScm.submoduleCfg, oldScm.browser, oldScm.gitTool, oldScm.extensions)
def newDefinition = new org.jenkinsci.plugins.workflow.cps.CpsScmFlowDefinition(newScm, scriptPath)
it.definition = newDefinition
it.save()
println "Done"
} else {
println("${project} has no SCM")
}
} catch (Exception e) {
// e.printStackTrace()
}
}
}

check if a resource is locked

I usually lock resources in my declarative pipeline with something like:
lock(resource: "MY_RESOURCE") {
// do something
}
but now I have several different resources I could use, is there a way to check if a resource is locked?
I would like to do something like:
myResources = ["RES1", "RES2", "RES3"]
hasResource = false
for (resource in myResources) {
if (hasresource) {
break
}
if (!isLocked(resource)) {
hasresource = true
lock(resource) {
// do something
}
}
}
(sorry if the syntax is wrong, I don't really program in groovy very often)
according to the sources of lock plugin this should work:
import org.jenkins.plugins.lockableresources.LockableResourcesManager as LRM
def myResources = ["RES1", "RES2", "RES3"]
def notLocked = myResources.find{rName->
LRM.get().forName(rName).with{ r-> !r.isLocked() && !r.isQueued() }
}
if(notLocked){
lock(notLocked){
//do smth
}
}
To Check if a particular resource is locked or not in jenkins
def manager = org.jenkins.plugins.lockableresources.LockableResourcesManager
def myResources = manager.get().fromName("test2")
//def myLabels = manager.get().allLabels // if you want to filter based on labels
def checkLocked = myResources.find { r ->
!r.isLocked() && !r.isQueued() }
if (checkLocked) {
def myResource = checkLocked.toString()
println myResource + "is locked"
}
else {
println "Specified Resource is not Locked"
}
Maybe not an optimal solution, but we managed to achieve this with following approach:
waitUntil {
lock(resource: 'RES1', skipIfLocked: true){
// do something with RES1
return true; // exit the waiting loop
}
lock(resource: 'RES2', skipIfLocked: true){
// do something with RES2
return true; // exit the waiting loop
}
lock(resource: 'RES3', skipIfLocked: true){
// do something with RES3
return true; // exit the waiting loop
}
}
We did it this way due to the following error we received when we tried to use the accepted answer:
Scripts not permitted to use staticMethod org.jenkins.plugins.lockableresources.LockableResourcesManager get

Jenkins: View last X builds

I want to view the more recent X builds for several Jenkins jobs.
So if I wanted to show the last 5 builds for jobs 1-5 it would looks something like this:
status build time
-------------------------
pass job1#3 13:54
fail job1#2 13:05
fail job1#1 13:01
pass job5#1 12:17
pass job3#1 11:03
How can I accomplish this?
Notice the builds of the jobs are woven together so if one job has run many builds recently it will show up more than other jobs that haven't run as much.
Executing the following script in the Script Console (Manage Jenkins -> Script Console):
import java.text.SimpleDateFormat
def numHoursBack = 24
def dateFormat = new SimpleDateFormat("HH:mm")
def buildNameWidth = 30
def cutOfTime = System.currentTimeMillis() - numHoursBack * 3600 * 1000
SortedMap res = new TreeMap();
for (job in Jenkins.instance.getAllItems(BuildableItem.class)) {
for (build in job.getBuilds()) {
if (build.getTimeInMillis() < cutOfTime) {
break;
}
res.put(build.getTimeInMillis(), build)
}
}
def format = "%-10s%-${buildNameWidth}s%-10s"
println(String.format(format, "status", "build", "Time"))
for (entry in res.descendingMap().entrySet()) {
def build = entry.getValue()
println(String.format(format, build.getResult(), build.getFullDisplayName(), dateFormat.format(build.getTime())))
}
For me this gives:
status build Time
SUCCESS xxx #107393 17:53
SUCCESS xxx #107392 17:48
SUCCESS xxx #107391 17:43
null yyy #3030 17:38
SUCCESS xxx #107390 17:38
FAILURE zzz #3248 17:37
...
You might need to change the numHoursBack constant, it controls the number of hours back to look for builds. As well as the buildNameWidth which determines the column width of the build column (if you have really long job and build names you might need to extend this).
Here is a bit cleanup version of Jon S Groovy script. Also displaying worst build information.
import hudson.model.*;
import java.text.SimpleDateFormat;
//
// Settings
//
def numHoursBack = 24;
def dateFormat = new SimpleDateFormat("HH:mm");
def cutOfTime = System.currentTimeMillis() - numHoursBack * 3600 * 1000;
/**
* Basic build information.
*/
def printBuildInfo(finalizedBuild) {
String level = "INFO";
String result = finalizedBuild.getResult().toString();
switch (result) {
case "UNSTABLE":
level = "WARNING";
break;
case "FAILURE":
level = "ERROR";
break;
}
// basic info and URL
println(String.format(
"[%s] Build %s result is: %s.",
level,
finalizedBuild.getFullDisplayName(),
result
));
// pipe description from downstream
def description = finalizedBuild.getDescription();
if (description != null && !description.isEmpty()) {
println(description.replaceAll("<br>", ""));
}
return finalizedBuild;
}
/**
* Get recent build items.
*/
def getRencentBuilds(cutOfTime) {
SortedMap res = new TreeMap();
for (job in Jenkins.instance.getAllItems(BuildableItem.class)) {
for (build in job.getBuilds()) {
if (build.getTimeInMillis() < cutOfTime) {
break;
}
res.put(build.getTimeInMillis(), build);
}
}
return res;
}
/**
* Print build items.
*
* minResult - minimum to print
*/
def printBuilds(builds, minResult, dateFormat) {
def format = "%-10s %-8s %s";
Result worstResult = Result.SUCCESS;
def worstBuild = null;
// header
println(String.format(format, "status", "Time", "build"));
// list
for (entry in builds.descendingMap().entrySet()) {
def build = entry.getValue();
Result result = build.getResult();
if (result.isWorseThan(worstResult)) {
worstResult = result;
worstBuild = build;
}
if (result.isWorseOrEqualTo(minResult)) {
println(String.format(
format, build.getResult(), dateFormat.format(build.getTime()), build.getFullDisplayName()
));
}
}
return worstBuild;
}
def builds = getRencentBuilds(cutOfTime);
println ("\n\n----------------------\n Failed builds:\n");
def worstBuild = printBuilds(builds, Result.FAILURE, dateFormat);
println ("\n\n----------------------\n Worst build:\n");
if (worstBuild != null) {
printBuildInfo(worstBuild);
}
println ("\n\n----------------------\n All builds:\n");
printBuilds(builds, Result.SUCCESS, dateFormat);

How to get the changes since the last successful build in jenkins pipeline?

Anyone have a Jenkins Pipeline script that can stuff all the changes since the previous successful build in a variable? I'm using git and a multibranch pipeline job.
Well I managed to cobble something together. I'm pretty sure you can iterate the arrays better but here's what I've got for now:
node('Android') {
passedBuilds = []
lastSuccessfulBuild(passedBuilds, currentBuild);
def changeLog = getChangeLog(passedBuilds)
echo "changeLog ${changeLog}"
}
def lastSuccessfulBuild(passedBuilds, build) {
if ((build != null) && (build.result != 'SUCCESS')) {
passedBuilds.add(build)
lastSuccessfulBuild(passedBuilds, build.getPreviousBuild())
}
}
#NonCPS
def getChangeLog(passedBuilds) {
def log = ""
for (int x = 0; x < passedBuilds.size(); x++) {
def currentBuild = passedBuilds[x];
def changeLogSets = currentBuild.rawBuild.changeSets
for (int i = 0; i < changeLogSets.size(); i++) {
def entries = changeLogSets[i].items
for (int j = 0; j < entries.length; j++) {
def entry = entries[j]
log += "* ${entry.msg} by ${entry.author} \n"
}
}
}
return log;
}
Based on the answer from CaptRespect I came up with the following script for use in the declarative pipeline:
def changes = "Changes:\n"
build = currentBuild
while(build != null && build.result != 'SUCCESS') {
changes += "In ${build.id}:\n"
for (changeLog in build.changeSets) {
for(entry in changeLog.items) {
for(file in entry.affectedFiles) {
changes += "* ${file.path}\n"
}
}
}
build = build.previousBuild
}
echo changes
This is quite useful in stage->when->expression parts to run a stage only when certain files were changed. I haven't gotten to that part yet though, I'd love to create a shared library from this and make it possible to pass it some globbing patterns to check for.
EDIT: Check the docs btw, in case you want to delve a little deeper. You should be able to convert all the object.getSomeProperty() calls into just entry.someProperty.
This is what I've used:
def listFilesForBuild(build) {
def files = []
currentBuild.changeSets.each {
it.items.each {
it.affectedFiles.each {
files << it.path
}
}
}
files
}
def filesSinceLastPass() {
def files = []
def build = currentBuild
while(build.result != 'SUCCESS') {
files += listFilesForBuild(build)
build = build.getPreviousBuild()
}
return files.unique()
}
def files = filesSinceLastPass()
There's the Changes Since Last Success Plugin that could help you with that.
For anyone using Accurev here is an adaptation of andsens answer. andsens answer can't be used because the Accurev plugin doesn't implement getAffectedFiles. Documentation for the AccurevTransaction that extends the ChangeLogSet.Entry class can be found at here.
import hudson.plugins.accurev.*
def changes = "Changes: \n"
build = currentBuild
// Go through the previous builds and get changes until the
// last successful build is found.
while (build != null && build.result != 'SUCCESS') {
changes += "Build ${build.id}:\n"
for (changeLog in build.changeSets) {
for (AccurevTransaction entry in changeLog.items) {
changes += "\n Issue: " + entry.getIssueNum()
changes += "\n Change Type: " + entry.getAction()
changes += "\n Change Message: " + entry.getMsg()
changes += "\n Author: " + entry.getAuthor()
changes += "\n Date: " + entry.getDate()
changes += "\n Files: "
for (path in entry.getAffectedPaths()) {
changes += "\n " + path;
}
changes += "\n"
}
}
build = build.previousBuild
}
echo changes
writeFile file: "changeLog.txt", text: changes
In order to return the changes as a list of strings, instead of just printing them, you may use this function (based on #andsens answer):
def getChangesSinceLastSuccessfulBuild() {
def changes = []
def build = currentBuild
while (build != null && build.result != 'SUCCESS') {
changes += (build.changeSets.collect { changeSet ->
(changeSet.items.collect { item ->
(item.affectedFiles.collect { affectedFile ->
affectedFile.path
}).flatten()
}).flatten()
}).flatten()
build = build.previousBuild
}
return changes.unique()
}

removeFrom* not working and with no errors

I have what I think is a simple problem but have been unable to solve...
For some reason I have a controller that uses removeFrom*.save() which throws no errors but does not do anything.
Running
Grails 1.2
Linux/Ubuntu
The following application is stripped down to reproduce the problem...
I have two domain objects via create-domain-class
- Job (which has many notes)
- Note (which belongs to Job)
I have 3 controllers via create-controller
- JobController (running scaffold)
- NoteController (running scaffold)
- JSONNoteController
JSONNoteController has one primary method deleteItem which aims to remove/delete a note.
It does the following
some request validation
removes the note from the job - jobInstance.removeFromNotes(noteInstance).save()
deletes the note - noteInstance.delete()
return a status and remaining data set as a json response.
When I run this request - I get no errors but it appears that jobInstance.removeFromNotes(noteInstance).save() does nothing and does not throw any exception etc.
How can I track down why??
I've attached a sample application that adds some data via BootStrap.groovy.
Just run it - you can view the data via the default scaffold views.
If you run linux, from a command line you can run the following
GET "http://localhost:8080/gespm/JSONNote/deleteItem?job.id=1&note.id=2"
You can run it over and over again and nothing different happens. You could also paste the URL into your webbrowser if you're running windows.
Please help - I'm stuck!!!
Code is here link text
Note Domain
package beachit
class Note
{
Date dateCreated
Date lastUpdated
String note
static belongsTo = Job
static constraints =
{
}
String toString()
{
return note
}
}
Job Domain
package beachit
class Job
{
Date dateCreated
Date lastUpdated
Date createDate
Date startDate
Date completionDate
List notes
static hasMany = [notes : Note]
static constraints =
{
}
String toString()
{
return createDate.toString() + " " + startDate.toString();
}
}
JSONNoteController
package beachit
import grails.converters.*
import java.text.*
class JSONNoteController
{
def test = { render "foobar test" }
def index = { redirect(action:listAll,params:params) }
// the delete, save and update actions only accept POST requests
//static allowedMethods = [delete:'POST', save:'POST', update:'POST']
def getListService =
{
def message
def status
def all = Note.list()
return all
}
def getListByJobService(jobId)
{
def message
def status
def jobInstance = Job.get(jobId)
def all
if(jobInstance)
{
all = jobInstance.notes
}
else
{
log.debug("getListByJobService job not found for jobId " + jobId)
}
return all
}
def listAll =
{
def message
def status
def listView
listView = getListService()
message = "Done"
status = 0
def response = ['message': message, 'status':status, 'list': listView]
render response as JSON
}
def deleteItem =
{
def jobInstance
def noteInstance
def message
def status
def jobId = 0
def noteId = 0
def instance
def listView
def response
try
{
jobId = Integer.parseInt(params.job?.id)
}
catch (NumberFormatException ex)
{
log.debug("deleteItem error in jobId " + params.job?.id)
log.debug(ex.getMessage())
}
if (jobId && jobId > 0 )
{
jobInstance = Job.get(jobId)
if(jobInstance)
{
if (jobInstance.notes)
{
try
{
noteId = Integer.parseInt(params.note?.id)
}
catch (NumberFormatException ex)
{
log.debug("deleteItem error in noteId " + params.note?.id)
log.debug(ex.getMessage())
}
log.debug("note id =" + params.note.id)
if (noteId && noteId > 0 )
{
noteInstance = Note.get(noteId)
if (noteInstance)
{
try
{
jobInstance.removeFromNotes(noteInstance).save()
noteInstance.delete()
message = "note ${noteId} deleted"
status = 0
}
catch(org.springframework.dao.DataIntegrityViolationException e)
{
message = "Note ${noteId} could not be deleted - references to it exist"
status = 1
}
/*
catch(Exception e)
{
message = "Some New Error!!!"
status = 10
}
*/
}
else
{
message = "Note not found with id ${noteId}"
status = 2
}
}
else
{
message = "Couldn't recognise Note id : ${params.note?.id}"
status = 3
}
}
else
{
message = "No Notes found for Job : ${jobId}"
status = 4
}
}
else
{
message = "Job not found with id ${jobId}"
status = 5
}
listView = getListByJobService(jobId)
} // if (jobId)
else
{
message = "Couldn't recognise Job id : ${params.job?.id}"
status = 6
}
response = ['message': message, 'status':status, 'list' : listView]
render response as JSON
} // deleteNote
}
I got it working... though I cannot explain why.
I replaced the following line in deleteItem
noteInstance = Note.get(noteId)
with the following
noteInstance = jobInstance.notes.find { it.id == noteId }
For some reason the jobInstance.removeFromNotes works with the object returned by that method instead of .get
What makes it stranger is that all other gorm functions (not sure about the dynamic ones actually) work against the noteInstance.get(noteId) method.
At least it's working though!!
See this thread: http://grails.1312388.n4.nabble.com/GORM-doesn-t-inject-hashCode-and-equals-td1370512.html
I would recommend using a base class for your domain objects like this:
abstract class BaseDomain {
#Override
boolean equals(o) {
if(this.is(o)) return true
if(o == null) return false
// hibernate creates dynamic subclasses, so
// checking o.class == class would fail most of the time
if(!o.getClass().isAssignableFrom(getClass()) &&
!getClass().isAssignableFrom(o.getClass())) return false
if(ident() != null) {
ident() == o.ident()
} else {
false
}
}
#Override
int hashCode() {
ident()?.hashCode() ?: 0
}
}
That way, any two objects with the same non-null database id will be considered equal.
I just had this same issue come up. The removeFrom function succeeded, the save succeeded but the physical record in the database wasn't deleted. Here's what worked for me:
class BasicProfile {
static hasMany = [
post:Post
]
}
class Post {
static belongsTo = [basicProfile:BasicProfile]
}
class BasicProfileController {
...
def someFunction
...
BasicProfile profile = BasicProfile.findByUser(user)
Post post = profile.post?.find{it.postType == command.postType && it.postStatus == command.postStatus}
if (post) {
profile.removeFromPost(post)
post.delete()
}
profile.save()
}
So it was the combination of the removeFrom, followed by a delete on the associated domain, and then a save on the domain object.

Resources