A retention policy defines the number of restore points you want to keep on your extents and is configured at the Specify Backup Storage Settings step of the backup job configuration wizard.
The restore points that fall under the retention policy will be removed from both the extents and object storage repositories in the following manner:
- An earliest offloaded restore point (that is, a backup file with metadata) will be removed from the backup chain of the associated extent.
Backup files with metadata are created, as described in SOBR Offload Job.
- Offloaded data blocks that correspond to the restore point that is being removed will be purged from the object storage repository upon the next successful session of the offload job.
Make sure that an object storage repository has not been put into maintenance mode, as this mode prevents the synchronization of the on-premises backup chain state with that of object storage. If the offload job fails to synchronize a backup chain state due to the aforementioned mode being applied, the synchronization attempts will be repeated during subsequent sessions until the backup chain state is synchronized successfully.
- Associated indexes will be updated for consistency purposes.
For more information, see Indexes.