Upstream / downstream jobs stuck in queue on CloudBees Jenkins Enterprise 1.609.1.1
We have the following job setup:
Trigger job - triggers 4 downstream test jobs
Test job 1-4 - triggered by upstream trigger job
All jobs have blockBuildWhenDownstreamBuilding = true and blockBuildWhenUpstreamBuilding=true
Previously the trigger job would sit in queue waiting for all the test jobs to finish. Now somehow both trigger and test jobs get stuck in queue (there are slaves available). Previously the mouseover (queue item description?) would indicate the trigger job was waiting for the test jobs to finish.
It looks like the test jobs were triggered by trigger #2, and should run (but they don’t, seem to be stuck in queue). Trigger #3 is already in the queue, and should wait on the previous test jobs (this seems correct).
When I cancel the queued trigger job, the other jobs seem to behave correctly. One started on an available executor, the other now indicated they were waiting for an available executor
We’ve recently upgraded from 1.580 to 1.609.1.1, and we believe the issue started since then.
- CloudBees Jenkins Enterprise 1.609.1.1
In Jenkins 1.609.1 LTS there were many significant changes , which allowed to increase the performance of Jenkins Queue. These changes caused several regressions, which have been fixed in further LTS patches. Most probably you hit JENKINS-28926.
1.609.1.1 is not recommended for use now, because 1.609.3.1 is available. This release also includes the fix for JENKINS-28926. So the best resolution for the issue would be to perform the core upgrade.