Data Domain Replication and DDBoost

[sdtuser]

We have Prod and DR site - each with a DD (data domain).
We are using vmware
Backups are done at prod with VBR to prod_DD using ddboost and all seems to work fine
prod_dd uses DD replication to replicate to DR_DD - this too seems to be fine
I can see the RO mtree on the DR_DD
The goal is to be able to access those backups at DR with the DR_VBR server.
I read the post by Oliver.rossi (https://community.veeam.com/blogs-and-p ... -site-3174)
I was just wondering - if I follow this and enable the ddboost on the DR Mtree and then scan in DR_VBR so I can see the backups in DR_VBR - Does that have any affect on the mtree at prod. Can the above process still continue where we backup at prod to the mtree and DD repl continues to replicate to the DR_DD - even though we can also now see and access the DR_DD mtree with VBR at the DR site.

My assumption is that if the above does work - then we would have to rescan at DR with VBR if we want to see the new backups that were taken and prod and replicated to DR???
Thank you

[Andreas Neufert]

Your assumptions are correct.
Replication would work independently.
Target need to be rescanned in order to see the backups.

I would like to mention, that a external storage replication without verification is not seen based on Veeam best practices as a secondary copy of the backups. A secondary chain controlled by Veeam that can check on metadata consistency and handling the chain differently would mean that you would not loose your secondary backup if the first one is somehow broken.

For mtree replication it is best practices to deactivate some of the transport optimizations as Veeam need to flush and write metadata, which forces the mtree replication to throw away the "receipt" and not deactivating it, would slow down the processing. As well it is best practices to deactivate replication until Veeam has written the backups for the same reason. There is no need to change when you are happy with the transport time and data amount transported, it is just an optimization.

[sdtuser]

Andreas
Thank you for the response.
I somewhat follow your response but I do have questions.
If DD replication is not BP then what is the Veeam BP for replicating the Veeam backups from Prod to DR

[sdtuser]

Andreas
In addition - the mtree at DD (where prod is replicating to) is in RO mode.
The article by Oliver.rossi mentions needing to change that so I can access the mtree on the DD side to scan it for backups.
The question is - Is it ok to go thru the process that Oliver Rossi mentions (making the dd mtree available to be scanned) and still have prod backups replicated to the dd mtree?
Will me making the DD mtree scannable affect or break or .... the process that is currently setup and working?

[Andreas Neufert]

It is always best practices to use a Veeam controlled job that creates backup copies.
Backup Copy Job, Scale-out-backup-repository Capacity Tier (for object offloading) and Backup 2 Tape Jobs.
For the Backup Copy Jobs you can implement as well (software based) WAN Accelerators from Veeam for low bandwidth/high latency situations.

Your question about the mtree replication I can not answer fully as this is more a matter for Dell to answer.
If making it RWtable destroys the replication, then I would check for other options like Fast Copies and mount the fast copy RWable, replacing it daily with the latest FastCopy of the production replicated data. Or just perform this step when you need it at restore.

[sdtuser]

Do you have a link to the Veeam BP for what we are talking about? (Prod -> DD copy)

[Andreas Neufert]

https://helpcenter.veeam.com/docs/backu ... ml?ver=120
https://bp.veeam.com/vbr/4_Operations/O ... y_job.html
https://www.veeam.com/blog/321-backup-rule.html
https://community.veeam.com/blogs-and-p ... p-rule-569

[sdtuser]

Thank you

[epaape]

Hey!
As an alternative you can also use Cyber Recovery Manager from Dell. It is a software Solution (the license comes with every DD System; no extra charge) that controls the Replication and that enables you to build some different scenarios.
For example:
We have a Customer that uses Veeam with Data Domains in different locations similar to your setup. There the Cyber Recovery Manager starts and stops Replication between Data Domain Systems. After Replication you can also copy that mtree in a new one. Then you have a seperate mtree with the Backups from day x. Next day same Operation and a new mtree with backups from day y.
This leads to some manual work if you need to restore something, but you do not have to break your whole Replication Context on the Data Domain. Plus: If something happens to the Backups on your Remote Site and the Error is replicated to the Main Site, you still have your stuff from yesterday.
That is a little workaround you can use for Veeam especially to get rid of the Read Only mtree that comes with Replication.
Greetings Eike