ReFS handling questions

[andreash]

Hi,
I'm evaluating Server 2016 and ReFS v3 for VEEAM Repositories, and I'm severely impressed by the space savings and performance increase! Just for fun I had setup my daily backup to create synthetic fulls every day, and it appears I get daily fulls at the cost of daily incrementals

But my question is about our backup copy repositories. I've also seen impressive space savings there, but I am concerned about the process to move this data to a new partition/drive.
What happens when I have to migrate ~10 GFS restore points from one repository to the other? In my ReFS v3 enabled repository they will probably take up 3~ the space of a full backup, even though they are 10 full backups. When I copy these over to a new partition I will loose the block cloning savings if I'm not mistaken (at least windows won't let me copy the data because the target is to small).
Is there a way to move Files with "shared" blocks from one ReFS filesystem to another?

Regards,
Andreas

[Mike Resseler]

Hi Andreas,

One of the issues you will see is that you cannot simply move the data from one repository to a new one, with ReFS formatted. Simply copying files doesn't align the blocks and therefore the block cloning won't work. For now there is not really a solution as the application that writes the data needs to use the API to align them. Copying or moving the files is something that doesn't do this.

There is also more information in this thread where there is a part that describes backup copy jobs and GFS: veeam-backup-replication-f2/vbr-9-5-refs-t38521-75.html

[Andreas Neufert]

Maybe clone the Job and enable "Map existing backup to avoid active full processing". Disable the "old" jobs. Let the Cloned Jobs fill the new chain on the new storage and delete the old ReFS Repository after you reached the needed restore points at the new location.