Dedupe options using JBOD

[Kellen.Flood]

Using a 27x8TB disk JBOD Server for Veeam testing. I’m thinking of ways to get the most effective dedupe given that many of our machines are over 1TB and Windows 2012R2 dedupe doesn’t love that.

Any thoughts on Starwind vSAN or other possible solutions that can get us dedupe in a cost effective manner?

[Gostev]

I don't have an experience with Starwind vSAN, but Windows Server 2016 dedupe has been pretty awesome in our testing - and it works on backup files of up to 4TB in size.

[VladV]

We have been using WS2012R2 deduplication with 1.5TB file for over 2 years know and we haven't encountered problems on normal incremental jobs. The only issues I've seen is related to fragmentation when the large VBK file is modified during reverse incremental jos. Also, be sure to apply https://www.veeam.com/kb2023 to mitigate the issues I've mentioned.

[Kellen.Flood]

Thanks all, we will get to testing and see what numbers we come up with using 2012R2, then see our ability to go to 2016 once released

[Net Runner]

StarWind has in-line deduplication with 4KB blocks size and is made for primary production storage. It currently uses RAM to store hash tables and thus isn’t a perfect candidate for cold data or backup storage. On the day when they will move hash tables to flash (from what I know from their forums they are currently working on this) situation might change.
As for now Windows Deduplication looks like a perfect solution for OP.

[veremin]

And since MS has made the Windows Server 2016 Technical Preview 5 generally available last week; it might be a perfect opportunity to finally install it and play with features it can boast with including improved deduplication. Thanks.

[KOOLER]

It's a pity ReFS still can't do any deduplication. Even in RTM.