It appears there is already provision for checking duplicates (it is enabled in system config by default): https://github.com/jenkinsci/jobConfigHistory-plugin/blob/jobConfigHistory-2.14/src/main/java/hudson/plugins/jobConfigHistory/FileHistoryDao.java#L269

In that case, I think all we have to do is make sure any decoration of branch jobs use BulkChange to ensure there are not any intermediate saves to the configuration during indexing that are different from the original (and ideally final) configuration. I believe this problem is more limited to the Multi-Branch Project Plugin at the moment and not any other multibranch types.

pmr: I attached a screenshot of JobConfigHistory, showing you the option to not save duplicate histories. You find it in the global jenkins configuration.

jochenafuerbacher jglick I have seen this option and it is enabled in my Jenkins Instance. I also don't see any duplicated history, but there are still tons of files in the following directories: `$JENKINS_HOME/config-history-jobs/MULTIBRANCH_JOB/branches/*` (note the branches subdirectory).

This is the real culprit IMO and makes it necessary to exclude all multi-branch jobs from the job config history or disabling the plugin alltogether.

mjdetullio I noticed, that the Multi-Branch Project plugin deactivates all branches before syncing and reactivates them after syncing. This leads to some small changes of the branch config file (what gets tracked by JobConfigHistory).

Is it necessary to do this deactivation / reactivation of the branches?

Fixed in 0.5 release of multi-branch-project-plugin