Forever Forward plus Backup Copy

[unsichtbarre]

Strategy advice needed:

I have a job where Replication from Backup was chosen as the Replication source. The Backup Repository (source for Replicas) is "remote" on the far side of a VPN from the actual *_replica VMs. There is a desired RPO of 1-3 Hours.

Now here's the problem: In order for a replica to be useful, the backup has to run as/more frequently than the replica RPO. So to attain 1-3 Hours RPO, unless we do Forever Forward Incremental, there are too many restore points (168) between full/synthetic full backups. Also, the synthetic full is proving to be disruptive, taking more than 18 hours.

How about forever forward with a retenmtion policy of 24 and an hourly backup cycle, plus a Backup Copy job to keep a weekly full and any GFS policy we like?
Then we could source the *_replica from the forever forward job and attain 1-hour RPO!

Opinions?

THX!

[Shestakov]

Hello John,
Your plan sounds good.
Another idea is to have 2 sets of backup/backup copy jobs taking turns.

[unsichtbarre]

Follow up:
Using Forever Forward, theoretically the Full Backup (*.vbk) is being "injected" each time a Retention Policy is applied.
We have seen a Replication job where the replication is cancelled by the Backup job it sources.
Could this be because the *.vbk is being grabbed from the Replication job by the Backup Job?
BTW: Replication was working until the Forever Forward job reached saturation and began applying Retention policy?

[Shestakov]

Forever Forward chain merges .vbk and the oldest .vib while replica from backup works with the latest(newest) increments.