Server 2012 DeDuplication

Question

Greetings Everyone,
Enabled Dedupe on a volume containing 6 Hyper-V VMs migrated from Server 2008 R2. I enabled dedupe Friday afternoon, then set it to dedupe any files older than a day. When I checked the status today (Monday), it says the dedupe ratio is only 3%. So out of
6 VMs (4 x Server 2008 R2, 1 x SBS 2011, 1 x WHS) amounting to ~150GB of data, it only deduped ~5GB.
Being such a new feature, I haven't seen much in the way of real world data on how effective dedupe is, but I would have thought VMs would be ideal, as all four 2008 R2 VMs should have at least 8GB of identical data. Does dedupe work differently/better with
vhdx vs vhd?
Unfortunately, I did not run dedupe eval before enabling the feature, so I don't have any sense of what it "should" be. However, with 4 VMs running the same OS, I expected more...

Answer 1

I am curious as to the reason you say Hyper-V hosts are not good candidates for dedup. I just converted 68 ESXi VMs from our lab environment to a new HP G7 with 2012R2 core and a 10TB volume. Initially, all the VMs totaled 6.8TB. I wanted to test Dedup
out, so I enabled it on the 10TB VM volume. It took about 36 hours to complete, but those 68 VMs dropped from 6.8TB to 800GB. Two weeks have passed, and we have had no issues whatsoever with this. I did this only after I noticed there is a "-Type Hyper-V"
mode.
Again, why do you say Hyper-V is not a good candidate?
Because you confuse two absolutely different scenarios:
1) So-called VHD library (set of VMs files from VMs that are NOT running all the time). This scenario is a good candidate for MSFT off-line dedupe and MSFT does support it.
...and
2) Running VMs. MSFT dedupe needs closed file handle to kick in and start space optimization process. As VM is running and file never gets closed results no space reduction. So this scenario is NOT supported by MSFT.
What can you do? Well... Many ways to go. You can enable dedupe inside your VM. Obviously there would be no overlapping between multiple VMs but for a single file server it's a valid approach (there are tricks to make it work but everything is doable, discussed
here tons of times so use search). Also you can use third-party software that runs either on storage side or on host to use in-line deduplication. Tons of a free offerings here.
StarWind VSAN [Virtual SAN] clusters Hyper-V without SAS, Fibre Channel, SMB 3.0 or iSCSI, uses Ethernet to mirror internally mounted SATA disks between hosts.