we are having a problem with Win2019 REFS repository and Veeam Backup V10 (all up to date).
We are having about 10TB VM data to backup (incremental 8 times every day, 1x week full), backup and everything is working OK, but the problem is deduplication on that backup repository (RAID5+0, DELL 740xd, with 24x 1,6TB SAS drivers).
Dedup is just 1%. We tried "everything", different compression level in veeam job, different dedup type, even tried on NTFS ... for 35TB backups we are having 300GB dedup savings.
If you stored the Veeam backups as "uncompressed", then you should see higher deduplication values. The only exception that comes into my mind is encrypted backups. But even there, you should see more than 1%
For REFS: do you use Veeam with default settings, or with the experimental reg key?
How did you configure Windows dedup feature? Does it run in background or on a fixed schedule?
The option 'Deduplicate files older than' to 0 or a different number?
Best compression level for Veeam backups would be No compression.
You are writing 9 times a day into your repos as you said that you take 8 daily incremental backups. That would also affect Windows dedup feature performance.
If you encrypt your backups, you won't get good dedup results.
Did you try the powershell dedup commands to check the dedup jobs' results?
