Skip to main content

I think a lot of customers will plan to migrate their existing repositories (NTFS, ReFS) to a new XFS repository with immutability in v11. I am curious about the options you see to migrate backup files to a new XFS repository. I have already said goodbye to the idea, block-linked (block clone (ReFS), reflink (XFS)) data can be copied without dehydration. A nice way is to put a new repository into place and let the old one(s) fade out. But how to move long-term backup? 

How do you migrate backup repositories?

  1. Create ScaleOut repository (which you might have already), add XFS Repository.
  2. Put the ReFS repository in maintenance mode and select “evacuate backups”
  3. Cross your fingers and have patiens (a lot) depending on IO performance and network thruput of your systems, while the data is being moved.
  4. Expect that the data will grow, due to the simple copy procedure so any “sparse” synthetic backup will expand to essentialy become a full backup on the XFS storage. Over time this will shrink back when XFS reflink does the same thing as ReFS block clone.

What I stronly dislike here is that the last time we did this, with Veeam B&R 9.5.x, failed all backups during transfer time, that still had data on the repository marked as in maintenance, even thou the data is there and obviously readable.


If you essentially want to continue the job with a new full on a new extent, but keep old data for a period of time then “Seal Mode” is the right option for you. See here how to Switch to Seal Mode.


If you essentially want to continue the job with a new full on a new extent, but keep old data for a period of time then “Seal Mode” is the right option for you. See here how to Switch to Seal Mode.

Help center is so Great !! 


I think a lot of customers will plan to migrate their existing repositories (NTFS, ReFS) to a new XFS repository with immutability in v11. I am curious about the options you see to migrate backup files to a new XFS repository. I have already said goodbye to the idea, block-linked (block clone (ReFS), reflink (XFS)) data can be copied without dehydration. A nice way is to put a new repository into place and let the old one(s) fade out. But how to move long-term backup? 

How do you migrate backup repositories?

Very good Question !! I will read each answer...


  1. Create ScaleOut repository (which you might have already), add XFS Repository.
  2. Put the ReFS repository in maintenance mode and select “evacuate backups”
  3. Cross your fingers and have patiens (a lot) depending on IO performance and network thruput of your systems, while the data is being moved.
  4. Expect that the data will grow, due to the simple copy procedure so any “sparse” synthetic backup will expand to essentialy become a full backup on the XFS storage. Over time this will shrink back when XFS reflink does the same thing as ReFS block clone.

What I stronly dislike here is that the last time we did this, with Veeam B&R 9.5.x, failed all backups during transfer time, that still had data on the repository marked as in maintenance, even thou the data is there and obviously readable.

Thank you @Dejan.Ilic for this hint! I wasn’t aware of failing backups during evacuation of scale-out extent. 


Comment