Veeam architecture for Backup, Copy and Replication

[Superkikim]

Hi,

I'm currently working on an architecture with Many-to-One backup copy and replication. Hardware is existing, and there might be no way to buy new.

We in branches and in a DRP site: vSphere cluster on hyperconverged storage, and for backup storage DR4x00 storage.

The idea is the following:

BRANCH:

• All VM Backups to local deduplication box (Full/Incremental)
• Copy of all backups to remote site (see below)
• Creation of Replica for selected VMs in DRP site

About copy to remote sites: The problem I face is Dedup boxes are not meant for forever incremental. Hence Full/Incremental for backups.

For backup copies, I have two ways of achieving this:

Solution 1.

local vSphere => Backup to local DR4xxx => Backup copy to MD3xxx on remote site (5 restore points) => Backup archive to DR4xxx (35 restore points and more) => Replication job (source: backup copy on MD3xxx) to remote vSphere => Backup archive to DR Box.

For this purpose, I create a non dedup repository in the MD3xxx at DRP site (I have some limited space available), and I create Backup Copy jobs there. This would be the easiest to manage. Therefore, I would like to have 5 restore points on non dedup storage, and then I would like to archive to dedup storage.

My question here: How can I copy Full/Incremental backups from backup copy restore points from the non dedup storage to the dedup storage as I want to avoid any post processing on the dedup storage (transformation, synthetic full, etc...)


Solution 2.

local vSphere => Backup to local DR4xxx => Replication local DR4xxx to remote DR4xxx (based on DR mechanisms) => Rescan of remote DR from local Veeam => Replication job (source: backups on remote DR) to remote vSphere

This solution I find more complex as I have to manage DR replication + repository rescan before I can generate the Replicas.

Any ideas, insight, recommendations ? comments, questions ?

Thank you for any participation to this

[Vitaliy S.]

Hi,

As you've correctly stated, using dedupe device for the primary backup job target is really a bad idea due to transform/files re-built process, so If I were you I would do the following:

1. Point backup jobs to MD3xxx (short retention just for FLR purposes, usually 7 restore points is more than enough)
2. Create backup copy job for long term archival and point them to DR4xxx. Since this is a dedupe storage, you need to enable this checkbox > Read the entire restore point from source instead of synthesizing it from increments
3. Run replication jobs on the backup server (separate) which will be hosted the DRP site. Here is an existing topic on why it is usually recommended > Veeam Replication Best Practice and Veeam location

It is a bit different from what you've purposed initially, but it is pretty common among our existing customers, so please let me know what you think on this design.

Thanks!