Uploaded image for project: 'Jenkins'
  1. Jenkins
  2. JENKINS-33865

Double Build on Multibranch Pipeline with open PR


    • Icon: Improvement Improvement
    • Resolution: Duplicate
    • Icon: Major Major
    • None
    • Jenkins v1.652 running on a Ubuntu 14.04

      The multi pipeline plugin is great, but there is one annoying thing that prevents me from actively using it:

      I have a BitBucket Git Repository where I experimentally created feature branches containing a Jenkinsfile with build instructions - it works very well. Whenever I push to the branch, the BitBucket hooks are set up and work, they trigger a new indexing - so far, so good.

      However, if I now open a PR for that branch (which is quite standard), then indexing will do the following:

      Finds the branch -> matches criteria
      Finds the PR -> of course, matches criteria too (that is, it contains a Jenkinsfile, as opposed to other branches)

      This results in everything being doubled: The build itself, along with all notifications, build messages... This is quite a waste of resources as one build takes around 8 Minutes. Also, the doubled builds interfere because they are unfortunately using the same MySql Database.

      Am I missing something here? The Doc says that the indexing scans for open branches and PRs, so this is not a bug. However, this will very often result in the behavior I just described. How can I prevent Jenkins from doubling everything? I would really like to start using this as it is much more clear than the old pipeline currently in place ...

      I would suggest either to give an option to disable scanning for PRs, or make the indexing smart enough to see that the branch for a PR was already scheduled - which would be what I'd prefer.

            amuniz Antonio Muñiz
            lmanzke Lucas Manzke
            4 Vote for this issue
            7 Start watching this issue