If I have a set of parallel jobs and 1 fails and the other succeeds. I try the resume execution/retry failed nodes, it triggers both the jobs again. Is there any setting in rundeck which can trigger only the failed job and not rerun the entire group again ? Is this a bug ?
Asked
Active
Viewed 298 times
0
-
when you mean "If I have a set of parallel jobs" it's about jobs launched from a parent job via job reference step? (parallelly) – MegaDrive68k Feb 11 '22 at 16:20
-
Just 2 jobs in the entire group, each via job reference. they are set to trigger in parallel. ! fails and 1 succeeds, using the resume should trigger only the failed job but both the jobs get triggerred. any reason why ? – jack brown Feb 11 '22 at 16:33
1 Answers
0
The reason is that job reference steps is running on a single job, by design, rundeck consider that as a parent job exection and not individual executions (child jobs, the parallel jobs). If you want to avoid this, you will run these jobs individually (in a job calling each job using rd cli or Rundeck api) in a inline-script step.
In that way you can retry only the failed execution.
Now, to resume since a failed step, you can use the Job Resume Plugin (only for Rundeck Enterprise). The plugin allows to resume in the job failed step.
MegaDrive68k
- 3,768
- 2
- 9
- 51