-
Bug
-
Resolution: Unresolved
-
Major
We noticed the following strange behaviour on quite a lot of our jobs.
After all build steps and post build steps finished, the build is in some quite inconsistent state for up to exactly 5min:
- the Console is still spinning with no output
- the big ball of the job is still blinking
- starting the job again results in "current job is already building"
ALTHOUGH - the build is shown as finished in the build history of the job
- it has disappeared from the slave it was running on
I especially liked the build history, when you start a new build #128, it says, build #127 is still in progress, and right below you see a quite finished build #127
We double checked enabling and disabling the disk-usage-plugin, and it was clearly causing this issue. The 5min are probably some kind of timeout somewhere.
- is duplicated by
-
JENKINS-23347 Several minutes delay at end of job
-
- Open
-
I have encountered this issue on my Jenkins installation with a subset of the jobs.
The jobs that have this problem are the ones that have a Custom Workspace defined.
After the job finishes, the jenkins process on the master starts using a lot of CPU while the Disk Usage plugin is trying to estimate the size of the workspace on the slave.
While this is happening, the build history for the job shows that it has finished, but the job is still shown in the Build Executor Status and the console log for the build does not show the expected last statement of "Finished: SUCCESS".
This goes on for 5 minutes, which happens to match the configured value of the timeout for calculating the size of the workspace on slaves.
Not sure if this matters, but our jobs are of type: "free-style software project" and the version of Jenkins we use is 1.566
We had version 0.23 of the disk usage plugin, the problem went away by uninstalling the plugin.