Issue
When using CloudBees CI HA (active/active) and triggering multiple builds of the same Pipeline job with different parameters within milliseconds of each other, intermittently there could be unexpected issues due to build number collisions.
One example of an issue caused by this would be output after the Finished:
line of the log, which should never happen:
[2024-10-07T06:12:36.908Z] Finished: FAILURE [2024-10-07T06:12:38.103Z] Created Pod: ...
Another example of am issue this causes is CloudBees Pipeline Explorer plugin failing to display Pipeline logs, with the error:
Error 500 There were issues loading the console log
To determine if you are impacted by this issue where two replicas assign the same build number to different builds, enable a custom logger for the class hudson.model.Run
at the FINER
level, as per Configure Loggers for Jenkins and CloudBees CI.
Then when the issue happens next time, review the custom logger from each replica for the message new JOB_NAME#BUILD_NUMBER @…
with the same BUILD_NUMBER being printed on two different replicas, then that would be sufficient evidence to confirm this is the issue you are facing.
Resolution
This issue is tracked as a bug, and will be fixed in a future release of CloudBees CI. This article will be updated when the fix is available.
Workaround
If impacted, the workaround is to avoid triggering multiple builds of the same Pipeline job with different parameters within milliseconds of each other. This may require redesign of your Pipeline job, for example introducing a delay between build triggers, or moving work from a downstream job into a stage of the upstream job. Please contact CloudBees Support if you are impacted by this bug.