Managed File Replication

[wreber]

Is there a good write up on Veeam Managed File Replication? I need to understand how it works and if and how it integrates with backup storage device replication.

[Andreas Neufert]

Are you asking on how we integrate into DataDomain Managed File Replication or how we copy our backup data arround ?

[wreber]

Probably some of both would help. I was asked by a customer how InfiniGuard would work with Managed File Replication. His concern is with rehydration of the data at replication time. He made it sound like MFR was a Veeam feature but I haven't found anything about it (other than DD).

[foggy]

MFR is a Data Domain feature, in Veeam B&R we have backup copy jobs that replicate backup files between repositories. If the customer's concern is whether Veeam B&R uses storage replication technology in the background, then it is not the case for InfiniGuard - data is rehydrated and copied via Veeam B&R data mover components.

[wreber]

Thanks

[Andreas Neufert]

Veeam best practices is to backups (maybe minimum amount) on a non deduped storage system and do from there Backup Copy Jobs to both sides with the dedup devices. Potentially use Veeam WAN Accelerator to replicate data over WAN links.

[benthomas]

Coming back to this slightly older thread rather than opening a duplicate thread, are there any plans these days now that there is more integration with Data Domains to look at integrating with Managed File Replication? This would allow Veeam to be aware of the native replication between devices.
I'm aware Veeam recommends their own Backup Copy Jobs, so maybe these could be the point at which it orchestrations the DD native functionality.
It would reduce the complexity of using scripting and mtree replication, or trying to deploy WAN accelerators and dealing with rehydration issues.

[Andreas Neufert]

Hi Ben,
we have researched this point already. Veeam's backup file format is self describing and we are performing therefore IO operations on the backup files that are not compatible with the data reduction optimization from the DataDomain replication. We go back in the backup file updating metadata and header as well force flushes to ensure consistency of our metadata. It is not done with "just" enabling the replication. We would have to write a new way of writing backup files or completely introduce a new backup chain (daily clone) for this. This will for sure take us a while. The mtree replication can be used. See the discussion here for configuration changes that you need to make or what software to use for it: vmware-vsphere-f24/data-domain-replicat ... 94829.html

As well if you follow the Veeam Best Practices, the primary backup target is a non deduplicated disk system and then you can perform with Veeam Backup Copy Jobs to Datadomain the copies.