So I have a dynamic number of jobs which all have only one build step.
at some ("random") point of its execution, each job run some application which couldn't have more than one instance at a given time.
In general I do want parallel run of the jobs. but I still need some synchronization when two or more jobs trying to run the above application at the same time.
I though about using the Locks and Latches plugin , but I can't see how this will help me in my situation.
Ideas will be more than welcomed!
Run the application from a separate job (let's call it APP_JOB) that you would invoke via Parametrized Trigger Plugin (as a build step, not as a post-build step) from your other jobs, with an option to wait for it enabled. By default Jenkins won't run parallel instances of APP_JOB.
Related
What I'd like to do:
Before/after any build allocates a slave: Reset slave to a pristine state (eg restore/delete files, kill rogue processes, etc. In other words, run some script). I know how to do this in one job, I don't know how to have Jenkins trigger it for every job of every type without modifying them all.
Unsatisfactory Ways I can think of accomplishing this:
Enforce a pre/post step in every AbstractProject (old style jobs). This is only scalable if using the Job DSL Plugin to generate all jobs.
Enforce a pre/post step in every PipelineJob. This is only scalable if using a Jenkins shared library (eg: myBuild() that encapsulates these steps).
Make use of the Global Post Script plugin and run a groovy script after every build. This works, but the script runs after the slave has been released, so another build may have grabbed it already (it's too late). Figuring out what nodes were allocated during the build is also fairly complicated
Switch all slaves to one-shot type (ie Docker) that require no clean-up. This doesn't work for my use cases, but may work for someone else.
Periodically run a job that uses a System Groovy Script to edit all other jobs to add a "clean-up" "pre" build step if not present. (Will not work for Pipeline jobs)
(I currently do #1 and #2)
Ideal Solution
Theoretical options:
A plugin to hook into some event (ie WorkspaceListener.beforeUse() ) and execute something then (unfortunately, WorkspaceListener does not apply to Pipeline Jobs). This should trigger right before a slave is used. It gets a little complicated when a slave has multiple executors (mine don't).
A plugin to enforce execution of some steps, similar to "Execute builders from job X", but in every job. (also doesn't work for Pipeline jobs)
Assuming that I'm using Swarm for all slaves, modify the swarm client to handle this logic (perform a task when slave becomes idle). A poor man's way would be to make the swarm client run in "one-shot" mode, in a bash loop.
Question:
What am I overlooking? Is there a better way?
We have set up a Jenkins instance as a remote testing resource for our developers. Every time a tag is created matching our refspec a job is triggered and the results emailed to the developer.
A job is defined as follows:
1 phase consisting of three jobs (frontend tests, integration tests,
unit tests)
All subjobs are executed, irrespective of success
Email the developer the test results
This setup mostly works except for two issues:
I cannot get the job to run in parallel. The subjobs run in
parallel, but only one instance of the job runs at a time. Is this
something I can configure differently somewhere, or is this inherent
in the way the plugin works?
The main job checks out and occupies one of our build servers for
the duration of the job. Is there a way to do git polling and then
just grab the hashref and release the build server on which the
polling was done before continuing building the subjobs?
In the multi job plugin, everything runs in parallel that is listed in the same "Phase", however the multijob itself needs somewhere to run. If you have a build followed by a test phase, you can add a "Build Phase" prior to the test phase, and only that phase will require a "build server".
There is an option called "Execute concurrent builds if necessary" that will allow multiple jobs of the same name to run simultaneously. This option must be set for the parent job and the subjobs as the default behavior of Jenkins is to only allow one build of a Project (job) to run at a time. Beware: Read the comments as this may have unintended side effects.
Not clear what you mean about polling however if using git, you may want to use webhooks so that pushes to the git repository directly invoke Jenkins. No need to poll.
I'm trying to configure a job in jenkins in a way that the promotion process (Promoted builds plugin) happens in the middle of a run. The point is that there are some annoying tasks like javadocs, sonar integration,... that can be run even after the promotion process, therefore I would like to make the build, run all needed tests and then promote the build automatically. Other tasks can run after the promotion process.
Do you know how I can implement this using the Promoted builds plugin?
Thanks in advance for the help.
No, Promotions only run on completed builds. If you want to run it in the middle of the job, then it's a build step, not a promotion.
You can either configure build steps for your actions (same actions as promotions). You can even call other jobs, and wait for them to complete (or not wait).
Or you can configure your "after promotion" tasks as a second promotion that is executed after the first one.
I have several integration tests within my Jenkins jobs. They run on several application servers, and I want to make sure that only one integration test job is run at the same time on one application server.
I would need something like a tag or variable within my jobs which create a group of jobs and then configure the logic that within that group, only one job may run at the same time.
Could I use the Exclusion plugin for that? Does anyone have experience with it?
Use the Throttle Concurrent Builds Plugin. It replaces the Locks and Latches plugin, and provides the capability to restrict the number of jobs running for specific labels.
For example: you create a project category 'Integration Test Server A' and tie jobs to it with a maximum concurrent count of 1, and a second 'Integration Test Server B' label and tie other jobs to it, both categories will only run a single concurrent build (assuming you've set a max job count of 1), and the other jobs in that category will queue until the 'lock' has cleared.
Using this method, you don't have to restrict the number of executors available on any specific Jenkins instance, and can easily add further slaves in the future without having to reconfigure all your jobs.
If I understand you right, you have a pool of application servers and it doesn't matter on what server your tests run. They only need to be the only test on that server.
I haven't seen a plugin that can do that. However, you can get easily around it. You need to configure a slave for each application server. (1 slave = 1 app server) You need to assign the same label to all slaves and every slave can only have one executor. Then you assign the jobs that run the integration tests, to run on that label. Jenkins will assign the jobs then to the next available slave (or node) that has that label.
Bare in mind that you can have more than one slave running on the same piece of hardware and even a master and a slave can coexist on the same server.
Did you check below parameter in the Jenkins -> Manage Jenkins -> Configure system
# of executors
The above parameter helps you restrict the number of jobs to be executed at a time.
A Jenkins executor is one of the basic building blocks which allow a build to run on a node/agent (e.g. build server). Think of an executor as a single “process ID”, or as the basic unit of resource that Jenkins executes on your machine to run a build. Please see Jenkins Terminology for more details regarding executors, nodes/agents, as well as other foundational pieces of Jenkins.
You can find information on how to set the number of Jenkins executors for a given agent on the Remoting Best Practices page, section Number of executors.
Source - https://support.cloudbees.com/hc/en-us/articles/216456477-What-is-a-Jenkins-Executor-and-how-can-I-best-utilize-my-executors
I have a Jenkins job which compiles and publishes our Java project to a JBoss server. Obviously, the server takes time to start and deploy the new code. I have a second Jenkins job that runs Selenium tests against the running JBoss instance.
I would like to make the second (Selenium) job be performed automatically as a post-build action from the first job (I have already done this), but I want it to be delayed by, say, 2 minutes. The amount of delay time isn't important, but I can't find anywhere that describes how to delay the start of a post-build job. How would I accomplish this?
In the advanced project options of a project configuration, you can set a "quiet period" that does exactly that. Jenkins will wait the specified amount of time after a build has been triggered before actually starting the build.
Alternatively, you could have the JBoss server trigger the build (e.g. by calling a URL) once it's up and running. The advantage of that is what it would take care of cases where the JBoss server doesn't start for some reason.
You might also want to have a look at the Parameterized Trigger Plugin which allows you to run builds of other projects as build steps. This way you could run the Selenium tests as part of the original job and fail if those tests fail.