DLM will reclaim any backup set that is expired, with the exception that if all copies of a given backup set are expired, by default the last recovery chain will not be reclaimed, even if it is expired (this restriction can be disabled).
Excepting when using removable (RDX style) disk storage, these reclaims occur in the background on a regular cycle and are not triggered by the activity of backup jobs, however reclaims can only take place against backup sets held in online storage devices.
During normal activity where the disk storage remains online, older backup sets are reclaimed (deleted from the system) not long after they expire, and as such, by design there would be no expectation of being able to restore from online older expired sets.
If a customer is detaching the disk storage device from Backup Exec, with the intention of physically managing a longer retention by limiting access to the storage device, then as long as the disk storage is maintained as an offline device inside the storage configuration of Backup Exec then after 2 weeks of being offline for USB, local disk, SAN or NAS drives OR after 4 weeks of being offline for removable/RDX style cartridges, the device or cartridge in question will be marked as "Limited to Read Only" and will not allow reclaims when reconnected. Note: the 2 week/4 weeks timings are defaults and can be adjusted within the storage settings of Backup Exec.
If however the disk device is detached / made unavailable to the Backup Exec server and also deleted from inside the Backup Exec Console, OR moved to another Backup Exec server in a DR situation, then because that installation of Backup Exec is not managing the device in question as an offline device, reclaims can take place against any expired sets when the device is added into Backup Exec and Inventory (or in some cases Inventory and Catalog) operations are run.
As such situations that can cause this deletion of older backup sets are:
- The device in question is not RDX and has been detached and deleted from Backup Exec. Enough time has passed for the backup sets in question to have passed their expiration date and then the device is attached to the original server and configured as a 'new' disk storage device. An inventory (or inventory and catalog) is run
- The device in question is not RDX. Enough time has passed for the backup sets in question to have passed their expiration date and then the device is attached to a different server (or a rebuilt version of the same server where the BEDB has not been recovered). An Inventory and catalog is run (or an inventory only if the catalog files were copied/recovered)
- The backup set in question is already expired on an RDX cartidge and the cartridge is used in a different server (or a rebuilt version of original server without the original BEDB). Inventory and catalog operations are run, however before a restore is attempted a new backup job writes data to the cartridge (which for RDX triggers the reclaim operation)