Support deduplication of the MBS ReFS volume directly
Currently the documents for DPM 2019 state:
> DPM does not support deduplication on ReFS disk used for MBS backups.
It would be preferable that this is supported as it will provide better performance (not needing to create artificial fragmentation due to 1TB disks, wasted deduplication time due to empty space in a vhdx, no additional abstraction layer or multiple OSes running). Further, ReFS supports 4TB files for deduplication and would allow the deduplication of datasources larger than 1TB.
Currently I've been running this on a few servers and as far as I can tell there aren't any incompatibilities as backups, restores, and tape backups (to a VTL) seem to be working without issues and deduplication has been performing far better than using the 1TB VHDx method.
Jeff Miles commented
I fully support this idea!