We have some jobs set up which share a workspace. The workflow for the various branches is:
Build a big honking C++ project called foo.
Execute several downstream tests, each of which uses the workspace of foo.
We accomplish this by assigning the Use custom workspace field of the downstream jobs to the build workspace.
Recently, we took one branch and assigned it to be build on a Jenkins slave machine rather than on the master. I was surprised to find that on master, the foo repository was cloned to $JENKINS_JOBS_PATH/FOO/workspace/foo_repo - while on the slave, the repository was cloned to $JENKINS_JOBS_PATH/FOO/foo_repo.
Is this by design, or have we somehow configured master and slave inconsistently?
Older versions of Jenkins put the workspace under the ${JENKINS_HOME}/jobs/JOB/workspace directories. After upgrading, this pattern stays with the Jenkins instance. New versions put the workspaces in ${JENKINS_HOME}/workspace/. I suspect the slaves don't need to follow the old pattern (especially if it is a newer slave), so the directories may not be consistent across machines.
You can change the location of the workspaces on the master in Jenkins -> Configure Jenkins -> Advanced.
I think the safe way to handle this... If you are going to use a custom workspace, you should use that for all of your jobs, including the first one that builds the big honking c++ project.
If you did this all in a pipeline, you can run all of this in a single job and have more control over where all the files are, and you have the option of stash and unstash, but if the files are huge, stash may not be the way to go.
You can omit 'Use custom workspace' option for each job and instead change master and/or slave workspace paths and use
%WORKSPACE%/../foo_repo path
or (that equal)
./../foo_repo path
In that case
%WORKSPACE% = [master or slave node workspace]/[job name]
and
%WORKSPACE%/../ = [master or slave node workspace]
Related
Who does the actual cloning of the project, is it the master or the agent node? If it is the master, then how does the agent node actually execute the job. If it is the agent node, how can we view the workspace in the browser?
When people ask "where is the workspace" the answer is usually a path, but I am more interested in where that path is, on the master or the agent node? Or maybe it is both?
Edit1
Aligned terminology to this: https://jenkins.io/doc/book/glossary/ in order to avoid confusion.
In a Jenkins set up all the machines are considered nodes. The master node connects to one or more agent nodes. Executors can run both on the master or agent nodes.
In my scenario, no executors run on the master. They are run only on the agent nodes.
The answer is: it depends !
First of all, although it is not a good practice IMO, some installation let the master be an actual worker and run jobs. In this case, the workspace will be on the master.
If you configured the master not to accept jobs, there are still occasion when a workspace can be created on the master. A good example is when your job is a "pipeline script from SCM". In this case, the master will create a workspace for the job, clone the target repo, read the pipeline, and start needed jobs on whatever slave is targeted, creating a workspace to run the actions themselves. If the pipeline targets multiple slaves, there will be a workspace on each of them.
In simple situation (e.g. maven or freestyle job), the workspace will only be on the targeted slave.
I needed to dig a bit deeper to understand this.
I ran a brand new instance of Jenkins and I attached a single agent node. I used SSH and I set the remote (agent) root directory to: /home/igorski/jenkins
As soon as I attached the node the remoting folder and remoting.jar showed up in that root directory.
I ran a basic Gradle Java pipeline job (Jenkinsfile in the project).
The workspace showed up on the slave. Not on the master.
From the Jenkins GUI I can access the workspace and see it's contents.
At the moment I kill the agent machine I can no longer view the workspace in Jenkins.
My guess is that the remoting.jar somehow does a live sync.
I also ran a freestyle project and I can confirm the same. As soon as the agent is killed I can no longer open the Workspace and I get an error stack trace:
hudson.remoting.Channel$CallSiteStackTrace: Remote call to JenkoOne
This was much more obvious with the Pipeline job though. There you get a link to the agent that you need to click in order to see the contents. As soon as the agent is gone the link is disabled. And you know exactly on which agent the node is. With freestyle jobs, you just get a Workspace link. There is no indication on what agent it is or if the agent is accessible at the moment.
So, both Zeitounator and fabian were correct.
I have a bunch of pipeline jobs, yet when executed, workspaces of some get preserved, some are deleted. How does jenkins make these decisions?
Based on my findings so far:
All jobs executed on nodes will have their workspace persisted, e.g. /home/ec2-user/workspaces/some-job
Some works on master keep their workspaces but some others' workspaces disappear after the job has finished. For example, after my build job succeeded, if I ssh in I can see the its workspace directory; but all my e2e jobs have no workspace.
Note I didn't use any of clearWs, deleteDir etc in my pipelines.
By the way, the reason I'm looking into workspaces is the disk usage keeps increasing and I want to cleanup. I thought the workspace is overwritten each time a job runs, but yet I get the 'Disk space is too low' warning several times.
Jenkins is creating a new workspace for every build job (= run) per default. You can see that in the path of the ws in your console log: /here/is/the/ws#buildnumber. If you dont want to have that behavior you can set it to an path which is for instance for every repo the same: How to set specific workspace folder for jenkins multibranch pipeline projects
Maybe some of your jobs don't get executed on the Jenkins Master, but on some connected Node (via an agent directive within your Jenkinsfile or Pipeline description). If that's the case you won't see a build directory inside the workspace for this Job on the Jenkins master, but on the connected Node.
You would only get the build results (like artifacts, reports, etc.) under /<JENKINS_HOME>/jobs/My_Job/ on the Master.
Remember that you could trigger a Jenkins build on a node also indirectly if you, for example, run the build within a Dockerfile and have configured (within Jenkins configuration) a specific node label for execution of Docker builds.
I've set up multibranch pipeline to track my repo and automatically build and test for all merge requests. It works wonders, however, I noticed that Jenkins creates a new workspace for each new branch. It is a pretty big project with a heavy build process and a lot of non-tracked cache files, that mostly stay valid from one version to another - so if instead of a fresh git checkout it would re-use previous workspace, it would build much faster (and also not use up so much hard drive space).
How can I configure it to re-use the same workspace for different branches?
After researching the issue, I found out that this is not something I can do with multibranch pipeline, so I switched to using the regular pipeline project. Now every build uses one of the available workspaces, so they end up re-using previous workspaces and the same cache files that really speed up the build.
Jenkins for MultiBranch projects by default uses isolated workspaces for every branch.
Jobs within the same branch use the same workspace.
A possible solution for you is to use ws(path) inside a pipeline.
node("agent_name") {
ws(workspacePath) {
echo '...'
// ..
}
}
I have several Jenkins Pipeline jobs set up on my Jenkins installation all of them with a Jenkinsfile inside the repository.
These pipelines are run for all branches, and contains all steps necessary to build and deploy the branch. However, there are some differences for the different branches with regards to building and deploying them, and I would like to be able to configure different environment variables for the different branches.
Is that possible with Jenkins, or do I need to reevaluate my approach or use another CI system?
#rednax answer works if you're using a branch-per-environment git strategy. But if you're using git-flow (or any strategy where you assume that changes will be propogated up, possibly without human intervention, to master/production) you'll run into headaches where a merge will overwrite scripts/variables.
We use a set of folders which match the environment names : infrastructure/Jenkinsfile contains the common steps, and infrastructure/test/Jenkinsfile contains the steps specific to the test environment (the folders also contain Dockerfiles and cloudformation scripts). You could make that very complex with cascading includes or file merges, or simply have almost-identical copies of each file in each folder.
When configuring the job you can specify for Jenkins to grab the script (Jenkins file) from the branch on which you are running. This mean that technically you can adjust the script on each of your branches to set up parameters there. Or you can grab the script from the same source control location, but commit a configuration file in each of your branches and have the script read that file after the checkout.
This question is similar to the question being asked at How to configure a single Jenkins job to make the release process from trunk or branches? however in this case Perforce is the SCM being used within Jenkins. Currently in Jenkins I have the following:
One release job per branch/trunk.
Each job has a separate Perforce workspace mapping the necessary branch/trunk
Upon running the job, the jenkins-perforce-plugin synchronises the complete workspace and then runs the maven release plugin.
Ideally I would like to have one release job that can point to any branch, synchronise to the code from that branch and carry out a maven release. However, with Perforce workspaces, I will require a view mapping for each branch/trunk. Is there are way to tell the jenkins-perforce-plugin to only synchronise to a particular view in the workspace view? This way I could build the release job with a parameter that passes in the branch path and the jenkins-perforce-plugin synchronises to jobs perforce workspace to this path only and then carries a build from there.
If I were trying to implement this as described I would created a parameterized build where I could hand in a Perforce label name. The Perforce Jenkins plugin can sync to a label; I would create labels for each release specifying the paths that should be synced and with a revision specifier of #head.
Jenkins should then sync just the files you want for that build. The workspace would of course map everything; the labels will specify the files to fetch.