Ops Center Protector has the ability to replicate data being sent to an on-site repository, to an off-site repository without needing to gather the data from the source machine(s) again. This minimises the load on the source machines and quickly and efficiently transfers the data off-site.
- The user selects which backups are sent to the secondary repository. You may choose to only send a subset of the primary.
- You do not have to use the same schedule. You may for instance back up to the primary repository every hour but to the secondary once a day.
- You can have different retention. You may for example keep backups on the first repository for a week and on the secondary repository for 6 months.
Repository to repository backups should be scheduled such that the secondary (off-site) repository takes the latest completed backup from the primary (on-site) repository. The secondary repository backup is therefore scheduled to run on completion of the primary's backup.
When the policy is first triggered, the on-site repository will be resynchronised with the source. The empty off-site repository will then be synchronised with the on-site repository. Depending the amount of backup data and the bandwidth of the network between the on-site and off-site repositories, this initial synchronisation process can take a considerable time (many hours) to complete. To overcome this, a technique called 'repository seeding' (see How to seed an offsite repository) may be used for efficiently setting up an off-site data store, reducing the time and bandwidth required to load the initial backup into the secondary repository. Once seeded, the amount of data transferred between the on-site and off-site repositories is much reduced and dependent only on the data change rate of the source machine(s).