Merging oldest incremental backup file into full backup file

[BackupDude]

Greetings,

I have a question to the collective. We currently have a Veeam server located across a WAN link (100Mbps) from the ESXi environment it backs up. The location has an onsite repository and all jobs point to this repository for local storage. In addition, we have a backup copy job that moves data from this local repository to an offsite location.

The jobs are configured as incremental (forever), and maintain 31 restore points. No syntenic or active fulls are configured. I noticed one of the jobs is stuck at 99% on "Merging oldest incremental backup file into full backup file".

What I notice is that the our WAN connection is saturated with data from this job. The veeam server appears to be pulling in the data from the local repository, merging the data, and then sending it back to the local repository.

Is this standard for the Veeam server to process the data itself, instead doing the work on the backup storage device?

[Shestakov]

Hello Fabio,

I noticed one of the jobs is stuck at 99% on "Merging oldest incremental backup file into full backup file".

Does the job end finally?

What I notice is that the our WAN connection is saturated with data from this job. The veeam server appears to be pulling in the data from the local repository, merging the data, and then sending it back to the local repository. Is this standard for the Veeam server to process the data itself, instead doing the work on the backup storage device?

Correct, it`s a design of the Forever forward incremental backup method. The procedure of the merge is described here. Also here is a great post about Forever forward incremental backup method explaining pros and cons. Please take a look.
Is that what you were looking for?
Thank you!

[BackupDude]

Thanks Shestahov for the reply.

I understand the method. What i don't understand is why the Veeam server is doing the actual manipulation of the data instead of manipulating the data directly on the backup repository?

[alanbolte]

I think Nikita may have missed that point - the data transfer should only occur between the data mover process (on the repository server) and the storage. Is the storage configured as CIFS? If so, you may need to set a gateway server in the repository settings, or instead use a Windows/Linux repository.