Backup and Replication Strategy

Disbelief5920

Like everyone, we are currently evaluating XCP-NG to replace our existing ESXi/Veeam infrastructure.

What would the best way to accomplish our backup and replication requirements and can it be accomplished with XO backup?

Requirements:

Backup VMs from production cluster in datacenter to separate backup storage in same datacenter. (Delta Backup. Retention <X> Monthly, <X> Weekly, <X> daily).
Copy the backup file from production datacenter backup storage to offsite DR datacenter backup storage. (Retention possibly different than production <Y> Monthly, <Y> Weekly, <Y> daily).
copy the backup from production datacenter or offsite DR datacenter to cloud storage (Backblaze)
Continuous replication from production cluster in datacenter to offsite DR datacenter (1-2x daily for example).

Is it possible to achieve the above by only backing up the production VM one time so it does not need to do 2x snapshots and 2x data copies? (1 for the backup and 1 for the replication jobs?
Is it possible to achieve the above with only 1 transfer across the WAN instead of once for the continuous replication and once for the delta backup?

For example, is there a scenario that would first create a backup from production to onsite backup storage, then replicate across the WAN to the backup storage, then restore the VM from the backup storage to a DR cluster? This was the production NAS/XCP-NG nodes would only have to process/copy the data once, only 1 copy would transfer across the WAN and only 1 restore from offsite backup storage to DR cluster.

Thanks for your help.

DustinB

@Disbelief5920

Backup VMs from production cluster in datacenter to separate backup storage in same datacenter. (Delta Backup. Retention <X> Monthly, <X> Weekly, <X> daily).

Easily, through the scheduling system you can have the system perform Continuous Replication or a standard Delta job.

Copy the backup file from production datacenter backup storage to offsite DR datacenter backup storage. (Retention possibly different than production <Y> Monthly, <Y> Weekly, <Y> daily).

I think what you'd be looking at would be the "Mirror Backup" job. This'll copy your backup repo to another available repo.

copy the backup from production datacenter or offsite DR datacenter to cloud storage (Backblaze)

Yes, this can be completed in XO's interface or from your NAS/SAN solutions (or client side server software).

Continuous replication from production cluster in datacenter to offsite DR datacenter (1-2x daily for example).

I've had CR jobs running as quickly as every 5 minutes between two datacenters. Obviously bandwidth matters, once or twice a day should be perfectly doable.

Is it possible to achieve the above by only backing up the production VM one time so it does not need to do 2x snapshots and 2x data copies? (1 for the backup and 1 for the replication jobs?

You'd use the Delta job format for the bulk of these.

Is it possible to achieve the above with only 1 transfer across the WAN instead of once for the continuous replication and once for the delta backup?

No, CR is very expressly any changes that have happened, it's continuous by it's nature

@Disbelief5920 said in Backup and Replication Strategy:

For example, is there a scenario that would first create a backup from production to onsite backup storage, then replicate across the WAN to the backup storage, then restore the VM from the backup storage to a DR cluster? This was the production NAS/XCP-NG nodes would only have to process/copy the data once, only 1 copy would transfer across the WAN and only 1 restore from offsite backup storage to DR cluster.

I'm not 100% on this scenario.... so I don't think I can readily help explain it.