What is the purpose of Backup Copy Job?

Backup Copy Job allows you to create a secondary copy of your backups, thus helping to implement a 3-2-1 strategy.

How Does Backup Copy Job Work?

Backup Copy Job copies the restore points between source and target repository, without any impact on the VM in the production. Backup Copy Job does not work on “file level” copying the backup files. Instead it works on “VM block level”, reading and copying the necessary blocks from the source to create a restore point with the latest VM state on the target.

Example:

Assume you have 3 source jobs:

• Backup job 1 creates backup for VM1, VM2 and VM3
• Backup job 2 creates backup for VM4, VM5 and VM6
• Backup job 3 creates backup for VM4 and VM5.

Goal: to have a second copy of VM1 and VM4.

Solution: you can configure a Backup Copy Job to copy only blocks of VM1 from Backup job 1 and copy the latest state of VM4 either from Backup job 2 or 3 (whichever has the latest state for that VM).

What Does “Copy every” Setting Do?

Unlike a standard backup job, Backup Copy Job works actively during so-called copy intervals: it creates 1 restore point on the secondary target per copy interval, using the latest state of the VMs it has to process.

In the beginning of the copy interval, Backup Copy Job checks what restore points were created by the source (backup) jobs on the source repository during the period equal to the copy interval.

• If the source jobs have created some restore points during this period, Backup Copy Job will start copying data immediately. Note that it does not matter if source job created 1 or 100 restore points - only the latest state will be transferred to the target repository.
• If restore points have not been created yet, Backup Copy Job will wait for them to appear.

Once all VMs included in Backup Copy Job have their restore points processed, the backup file is finalized and Backup Copy Job is put in the idle state until the next interval.

Example:

Backup Copy Job is configured to process VM1 and VM2; it has a copy interval of 1 day that starts at 12.00.

• Restore points for VM1 are created every 6 hours, so during the previous day 4 points were created.
• No new restore points were created for VM2 during previous day.

At 12.00 Backup Copy Job will start copying the restore points for VM1 immediately – as explained, it will copy only 1 restore point with the latest state (not all 4). After that all other restore points created by the source job for VM1 during this copy interval will be ignored.

As for VM2 - Backup Copy Job will wait until the new restore point appears and will copy it immediately. After that Backup Copy Job will remain in idle state until the next copy interval.

What is Grandfather-Father-Son Retention (GFS)?

Apart from keeping the main forever forward incremental backup chain, Backup Copy Job can create a full backup periodically and keep it for long-term retention. By default, Backup Copy Job creates so-called GFS restore point using synthetic method - that is, using the data on the target repository).

If the “Read entire point…” option is enabled, Backup Copy Job will create GFS restore points using active full method – that is, reading all data from the source repository.

Keep in mind the following differences:

• With synthetic method, GFS point is not created exactly on the day when it is scheduled. Instead, when the time comes to merge the oldest incremental point with full point, the full point is not deleted – but it becomes a GFS point. At the same time, a new full point is generated synthetically and becomes the base for main backup chain.

Example:

Backup Copy Job is set to keep 3 points in the main chain.

GFS is scheduled on Monday.

The actual point will be created on Wednesday.

• With active full method, the GFS point is created exactly on the day when it is scheduled. This also changes the whole backup chain structure to forward incremental: after creation of GFS point Backup Copy Job will start creating increments for this point until next GFS point is created. Eventually, incremental points will be deleted by retention.