ReFS Repository size

[Ctek]

Hi,

Just as a quick sanity check, I have deployed new Veeam dedicated repository servers which have 327TB of NET space in RAID60 and 192GB of RAM. Performance is great so far.

Server is formated in ReFS 64K, and has 2 mounted VHDX on this file system, for other backups. What's left is available for Veeam as a ReFS repository.

Now, should I use 100% of the free space as a repository as a whole? Or I can alternatively create locally-mounted VHDX volumes with lower space to segment the repositories? I like segmenting my storage but I was wondering if ReFS performance would be better as a single big repository (300TB+). I tried and it works ok but wanted input.

Thanks

[Gostev]

Let me to some data mining here to see what do other customers do for volume sizes... stay tuned, may take a couple of days.

[Gostev]

First results are in, and I am seeing almost perfectly equal distribution between the following buckets. Apparently, I am awesome at designing these buckets for statistical purposes!

< 10 TB
10-25TB
25-50TB
50-100TB
100-200TB
>200TB

Another point of interest to you, in my sample (1525 installations) there are 100 installations with over 400TB volumes.

[Ctek]

Thanks Gostev