JIRA CaesiumSchedulerService can insert duplicated records into clusteredjob

Summary

In some cases JIRA CaesiumScheduler can create more then single record in clusteredjob table.
This will lead to problems similar to one in JRA-64325: method getJobDetails expects single uniq value from DB, while clusteredjob table does not ensure this.

Steps to Reproduce

1. TBD

Expected Results

Caesium ensures data integrity

Actual Results

CaesiumScheduler can create more then single record in clusteredjob table.

Notes

Workaround

• Delete duplicated records, see https://confluence.atlassian.com/jirakb/jira-fails-to-start-due-to-too-many-rows-found-for-query-on-clusteredjob-800702971.html