Issue Summary
Every time a Jira node is properly stopped, the schedules for this job is removed: com.atlassian.greenhopper.service.sprint.services.SprintAutoStartStopJob
While this makes sense for disabling plugins from the UPM UI, this may break important the functionality for those jobs whenever a node is stopped and not started again.
Steps to Reproduce
- Set up a Jira DC with 2 nodes or more;
- Check "Parallel Sprints" in Jira Software Configuration
- Stop a node.
Expected Results
Everything remains working as normal
Actual Results
The job schedule for com.atlassian.greenhopper.service.sprint.services.SprintAutoStartStopJob is removed:
We can see Jira doing it in the logs:
2024-08-13 11:34:31,850 FelixShutdown INFO [com.atlassian.greenhopper.Launcher] Unscheduling Jira Agile jobs... 2024-08-13 11:34:31,850 FelixShutdown INFO [service.sprint.services.SprintAutoStartStopScheduler] [SPRINT-AUTO-START-STOP] Starting to unschedule a SprintAutoStartStopJob... 2024-08-13 11:34:31,853 FelixShutdown INFO [service.sprint.services.SprintAutoStartStopScheduler] [SPRINT-AUTO-START-STOP] SprintAutoStartStopJob has been unscheduled.
Workaround
- When stopping a node permanently, stop it using
kill -9 <pid>
- Or restart one remaining node after the removal. Jira will always create the schedules again when Jira Software starts up
- Disabling "Parallel Sprints" and re-enabling it recreates the job
- has a derivative of
-
JSDSERVER-15502 Assets stale progress cleanup job (AssetsProgressCleanupJob) is unscheduled on clustered Data Center if one of the nodes is stopped
- Gathering Impact
- blocks
-
ACE-6010 Loading...
- causes
-
ACE-5884 Loading...
- is resolved by
-
JSWDC-63 Loading...
- relates to
-
JSWDC-47 Loading...