vSAN deduplication and compression shows no savings

[jottesen]

Hello,

We have a new vSAN cluster running on version 6.6.1, and the cluster shows no savings in compression or dedup.

I have compression & dedup turned on, disks are set to manual, and the diskgroups are all flash. Any advise?

Attached screenshot of capacity overview:

Attached screenshot of Storage policy:

Thanks in advance.

[TheBobkin]

Hello jottesen,

Dedupe and compression overview will show as 1.00x if a sub 1.00x saving is occurring, we can see that the overhead is using far more than is being saved here so this is why it shows as 1.00x.

Are you using Thick Objects/VMs here? These will not benefit from Deduplication nor compression.

Deduplication and Compression works on a per-Disk-Group level and thus savings are not likely to be seen until you put much more similar data on this cluster, it is currently very low utilization so start adding more VMs to this before expecting to see a decent ratio.

Bob

[jottesen]

Hello,

Thanks for your reply.

We are no using any thick objects/VMs, everything is thin.

All VMs are running Windows server 2016 and should have a lot of similar data (OS files).

[TheBobkin]

Hello jottesen,

Again it depends on how much data is similar AND on the same disk-group - OS disks are generally relatively small in comparison to data-disks so YMMV.

Add a lot more VMs to this clusters datastore and you should see the ratio increase - if this cluster has more than 3 nodes (4 if using RAID5 etc.) you can force the data onto a smaller number of nodes/disk-groups by putting some hosts in Maintenance Mode (Full Evacuation) so they are not targets for storing new data and the data gets moved off them, this should have a higher deduplication ratio.​

Bob

[Jasemccarty]

You said you are all thin... So the Object Space Reservation rule is 0% or simply not present in your applied Storage Policy, correct?

Also how many VM's/and how much RAM is being reserved in a vSwap file for each VM?

10 VM's w/4GB of RAM each will consume 80GB of vSAN capacity.

vSwap files are currently mirrored by default, and have a 100% space reservation. That's 10x4GBx2 or 80GB.

If you enable Sparse Swap, that 80GB would no longer be reserved, and space would only be consumed for 2x the amount of vSwap you need to use. (2x because of protection).

Check out this URL ( https://code.vmware.com/samples/1665/ ) for a PowerCLI script to enable sparse swap on vSAN 6.2 or higher.

[7POD]

Hi,

Do you use same storage policy on all vm's and on the vSan do you use a specific policy?

We had the same issue at first then we changed the vm policy and the storage policy.

PS we use vSan in our VDI environment. 

[jottesen]

Hello,

We use the same policy on all vm's and vsan.

I have tried creating New policies, and moved all vm's to the New, but there was no difference.

[jottesen]

Hello,

Thats correct, Object Space reservation is set to 0%.

The total Space consumed by swap files is 3,19 GB of 2,14 TB (total used Space)

Thanks for Your answer. I will check out sparse swap.