Dedupe BTRFS

[Rhubarb]

My Xpenology box is running a single volume (/volume 1) using the BTRFS system which, I understand, supports COW (copy-on-write), such that hyper backup can store multiple versions of files without duplication.  Also, I have verified, via PUTTY, that the 'cp --reflink' command is available on my system (DSM 6.1.3-15152-update 5), an essential element for the COW facility utilised by Hyper Backup (which I am also using).

Elsewhere, I have seen multiple users complain about file duplication created by Photo Station as it copies files to the '/photo' folder from other location on the system that those photos are already stored.

I know that Synology has provided, through Storage Analyzer, a very poor (IMO) facility to identify duplicate files via MD5 checking. But this facility only identifies 200 instances of duplicates on each run, and it requires that duplicates be deleted by the operator (no batch deletion facility).  Not only this, but it is reasonable that both the original files on the Synology/xpenology boxes and the copies uploaded to the /Photo folders are required  in their respective locations.  So a better option is required to deal with this issue: either hard linking of the duplicate files, or 'reflink'ing the copied data to minimize redundancy and maximize disk storage efficiency.

My question is: has anyone managed to implement either file deduplication or reflinking of data on Synology's BTRFS in this way?  If so, what have you achieved? Any advice please on how to proceed.  Is anyone working on this issue?

Apologies if this is expressed inadequately/too simply. 

 

Supermicro M'brd X10SL7-F-0

Xeon E3-1241-V3, 2 x 8 Gb DDR3 DIMMs giving 16Gb ECC DDR3 RAM.

9 x 8TB WD-Red HDDs in RAID 6 of 50.91 TB and 48.87 TB of BTRFS Volume

DSM 6.1.3-15152 Update 5

Edited October 4, 2017 by Rhubarb