Some of the technical material I’ve read about BeyondRAID claims that the system automatically supports deduplication – files are split into chunks, chunks are hashed, and if two chunks are observed to be identical, only one chunk is stored. [1][2]
I just set up a Drobo 5N and tried to observe this feature: with 678 GiB free, I made a copy of a folder already on the Drobo containing 71 GiB of videos. When the copy finished, reported free space was 607 GiB.
Is the deduplication feature just a myth? Is there something I can do to expose or take advantage of it?
hi, i dont think it does de-duplication,
i think the article is meant to show that each chunk (piece of a file) for Each file (duplicate or not), is split up across each drive in the array.
i havent read it fully (will still need to do that), and it may just be that the article was explaining it in a way which might make it sound as though it does dedupe in a way that you (and i) would expect.
Wikipedia: “With the introduction of the DroboPro, a RAID6 like feature was also introduced. BeyondRAID also has the ability to perform hash-based compression using 160-bit SHA1 hashes to maximize storage efficiency.”
Ars article: “Hashing each of the chunks before they’re committed to disk does three things: it’s an easy way to assign a unique identifier to each chunk so the Drobo can keep track of it; it allows the Drobo to perform some limited compression by storing each unique chunk only once; and it allows the Drobo to know if a chunk has been corrupted, since when a chunk is read it’s compared with its original entry in the object table”
