-
Bug
-
Resolution: Cannot Reproduce
-
Major
-
None
We were using post build script old version, and just upgraded to the new version which looks great. Thanks!
We're using postBuildScript with Job DSL. Each "postBuildStep" in the postBuildScripts takes 1 full second just to evaluate (note: not the generation, just the evaluation of this single block takes 1s). We are generating around 1500 jobs, each with 3 post build steps, and the Seed jobs are now taking more than an hour to run.
Sample DSL and output below.
println " ${ZonedDateTime.now()} postbuildscript start" postBuildScript { markBuildUnstable(true) buildSteps { postBuildStep { results(Arrays.asList("SUCCESS", "UNSTABLE", "FAILURE", "NOT_BUILT", "ABORTED")) role("SLAVE") buildSteps { scriptBuildStep { buildStepId("[redacted]") scriptBuildStepArgs { defineArgs(false) tokenized(false) } } } } postBuildStep { results(Arrays.asList("SUCCESS")) role("SLAVE") buildSteps { shell { command("[redacted]") } } } postBuildStep { results(Arrays.asList("FAILURE")) role("SLAVE") buildSteps { shell { command("[redacted]") } } } } } println " ${ZonedDateTime.now()} postbuildscript end"
Output:
2018-01-25T23:13:38.246+01:00[Europe/Berlin] postbuildscript start 2018-01-25T23:13:41.355+01:00[Europe/Berlin] postbuildscript end 2018-01-25T23:13:41.597+01:00[Europe/Berlin] postbuildscript start 2018-01-25T23:13:44.180+01:00[Europe/Berlin] postbuildscript end 2018-01-25T23:13:44.561+01:00[Europe/Berlin] postbuildscript start 2018-01-25T23:13:48.403+01:00[Europe/Berlin] postbuildscript end 2018-01-25T23:13:48.592+01:00[Europe/Berlin] postbuildscript start 2018-01-25T23:13:51.224+01:00[Europe/Berlin] postbuildscript end 2018-01-25T23:13:51.405+01:00[Europe/Berlin] postbuildscript start 2018-01-25T23:13:53.573+01:00[Europe/Berlin] postbuildscript end 2018-01-25T23:13:53.688+01:00[Europe/Berlin] postbuildscript start 2018-01-25T23:13:55.221+01:00[Europe/Berlin] postbuildscript end 2018-01-25T23:13:55.333+01:00[Europe/Berlin] postbuildscript start 2018-01-25T23:13:56.841+01:00[Europe/Berlin] postbuildscript end 2018-01-25T23:13:56.982+01:00[Europe/Berlin] postbuildscript start 2018-01-25T23:13:59.251+01:00[Europe/Berlin] postbuildscript end 2018-01-25T23:13:59.410+01:00[Europe/Berlin] postbuildscript start 2018-01-25T23:14:01.732+01:00[Europe/Berlin] postbuildscript end 2018-01-25T23:14:01.886+01:00[Europe/Berlin] postbuildscript start 2018-01-25T23:14:04.231+01:00[Europe/Berlin] postbuildscript end 2018-01-25T23:14:04.375+01:00[Europe/Berlin] postbuildscript start 2018-01-25T23:14:06.656+01:00[Europe/Berlin] postbuildscript end 2018-01-25T23:14:06.797+01:00[Europe/Berlin] postbuildscript start 2018-01-25T23:14:09.167+01:00[Europe/Berlin] postbuildscript end
During this time, Jenkins master is found to use constant 25% of CPU as well. We tested this after putting Jenkins on maintenance mode, so only this job takes 25% of CPU on master.
We have confirmed after multiple attempts that each postBuildStep takes 1 second to evaluate. If we have two postBuildSteps, the logs are printed 2s apart. 1 step and logs take 1s, etc.
Any ideas what can be going wrong?