When are jobs purged from jobsdb?

After digging around the EMC support documents, I managed to come up with the following criteria on how the NetWorker server manages the size of jobs database

• Every 12 hours a process runs to check for entries to be purged.
• If a record's lifetime has passed the time limit it will be deleted. This means that the run of a savegroup and all its associated savefs/save records will be removed as a single unit.
• Once the time based check has completed, the size of the database is checked. Records are then deleted from oldest to newest in 6hr 'blocks'. After every block another check is performed to see whether the target size (90% of the maximum size) has been achieved.