The retention process consists of:
- cloning backup to clone devices(if defined in disc media pool definition).
- erasing backups if older than retention time and if cloned (if cloning was defined).
The client does not know anything about cloning (or existence of backups in their original device), backup server does, and it logs all cloning processes in Hiback\log\svc.log.
Clone targets may be of type disc, tape (single tape) or al (autoloader) and can be located on the local or a remote backup server.
When a client requests a restore from the backup server, it will request the disc backup (the original backup). The request consists of a disc device name, a backup ID and the restore file set.
The backup server then searches the disc device directory (for example E:\disc.0001) for the respective backup.
If he finds it, the restore request will be satisfied directly.
If the backup server does not find the disc backup requested by the client, it looks into Hiback\log\svc.log to see if there are one or more clones for this backup.
If there is an entry for this backup in Hiback\log\svc.log, the backup server hands the information (clone server name, clone device name) back to the client and says goodbye.
The backup client now knows that it has to start a second approach and contacts the clone server to request the restore. The clone server acts like the backup server: it looks if it can satisfy the request from the given device and if it is not possible it will search its own Hiback\log\svc.log in order to look for clones.
This procedure works for clones up to the 2nd degree (backup -> clone (backup) -> clone (clone (backup))) and it is hidden to the user.