Implement disk reporting of retrieve jobs requested to be cancelled via tape state change in new Scheduler DB

Currently cancelled retrieve jobs are simply removed from the PENDING_QUEUE. It seems that in some cases reporting back to disk buffer is required. Investigate why do we report these back to disk buffer and if this is really needed. In case it is, then we shall introduce new RETRIEVE_REPORT_QUEUE table in order to avoid hacks such as picking up reporting jobs from 2 tables (PENDING and ACTIVE) or sending the PENDING jobs to ACTIVE table with no mount adjusting the purpose of ACTIVE table to accommodate this use-case.