-
Type:
Bug
-
Status: Closed (View Workflow)
-
Priority:
Minor
-
Resolution: Fixed
-
Component/s: workflow-job-plugin
-
Labels:
-
Environment:all-latest
-
Similar Issues:
I've been using infinite Pipeline for a demo. The job was running for a couple of hours, and then Jenkins master declined to the build due to the StackOverflow.
node {
int i = 0;
while(true) {
sh "echo 'Hello, world ${i}!'"
sh "sleep 5"
i = i + 1;
}
}
Log: See attachment
- relates to
-
-
- Closed
-
- links to
Code changed in jenkins
User: Sam Van Oort
Path:
src/main/java/org/jenkinsci/plugins/workflow/job/WorkflowRun.java
http://jenkins-ci.org/commit/workflow-job-plugin/896703ee33d7c7678272a1cecb7cba1a4c3e1f63
Log:
Merge pull request #91 from svanoort/fix-prefix-bugs
JENKINS-38383 Avoid Stackoverflow and full-flowgraph scan when trying to get branch names for log prefixes
Compare: https://github.com/jenkinsci/workflow-job-plugin/compare/d23e989dc6d5...896703ee33d7
Thomas Johansen That sounds like an issue I'm tracking that sometimes happens with performance-optimized builds with specific combinations of build agents used (noted in a couple quirks by https://issues.jenkins-ci.org/browse/JENKINS-47173) – do you have a sample of your pipeline here to reproduce with?
Sam Van Oort: I tried the snapshot. No StackOverflowException in the log this time, but the build did not resume either (waited almost half an hour). I'm considering to use a polling agent in combination with the input construct to completely avoid these damned loops.
[Pipeline] echo 19:41:58 Iteration 6143 [Pipeline] echo Resuming build at Tue Feb 27 19:42:36 GMT 2018 after Jenkins restart Waiting to resume part of jenkins-pipeline » work/PBLEID-15078 #1: jenkins-slave-2 is offline 19:41:58 Iteration 6143 Waiting to resume part of jenkins-pipeline » work/PBLEID-15078 #1: jenkins-slave-2 is offline Waiting to resume part of jenkins-pipeline » work/PBLEID-15078 #1: jenkins-slave-2 is offline Waiting to resume part of jenkins-pipeline » work/PBLEID-15078 #1: Jenkins doesn’t have label jenkins-slave-2 jenkins-slave-1 doesn’t have label jenkins-slave-2 jenkins-slave-2 is offline Ready to run at Tue Feb 27 19:43:10 GMT 2018
Thomas Johansen I'm not sure exactly what's triggering an NPE there (looks like it somehow got caught in an invalid state) – tried to reproduce but I can't.
But I've got a fix that's verified to remove the StackOverflowException and allow a build to cleanly resume in https://github.com/jenkinsci/workflow-job-plugin/pull/91 – please try the attached snapshot built off of it which should remove your issue.
In general I'd advise to find alternate courses in Pipeline than explicit polling loops – in this case an enhancement to JIRA plugin or simply accept the use of an agent to do polling while needed (perhaps with exponential backoff). workflow-job.hpi
Released with v2.19