Link to home
Create AccountLog in
Avatar of N
N

asked on

Failover Cluster Manager showing uncompressed disk space rather than compressed space from Dell SAN. How do I make them match up??

We have a Hyper-V cluster hooked up to two Dell SANs.  All storage presented from the SANs are Thin and have de-dupe and data reduction turned on.  All hyper-v nodes can see the disks.  


The problem we have is that within Failover Cluster Manager, the disks are showing the uncompressed size rather that the compressed size that the SAN is showing.  


Failover Cluster Manager:User generated imageDell SAN:

User generated image

Dell have confirmed that the SAN is showing the correct sizes etc so it appears to be with how Windows / Failover Cluster Manager sees the disks.  Does anyone have any idea how to make the two match up??  As it stands at the moment, we're wasting storage expanding disks that don't really need to be expanded!

ASKER CERTIFIED SOLUTION
Avatar of Member_2_231077
Member_2_231077

Link to home
membership
Create an account to see this answer
Signing up is free. No credit card required.
Create Account
Avatar of Philip Elder

In an elevated PowerShell:

Get-PhysicalDisk | Format-Table

Open in new window

Does the number presented there line up with Failover Cluster Manager?

I think that they're going to be inherently different. There's no way for the Hyper-V hosts to know how much space gets consumed on the SAN, not with compression and deduplication. If you deduplicate across volumes, how would you even assign the cost to any particular volume?

If you expand the allocation on the volume so that Hyper-V sees more space, but the Dell SAN doesn't need to consume any more blocks, you didn't waste any space. This is exactly what thin provisioning is supposed to do.
Avatar of noci
noci

In general SAN clients are given the "Size" of the disk, which is the upper limit specified in the SAN.
The client needs that to manage allocation bitmaps etc.
HOW the SAN manages (or fails) to provide that space (compression, deduplication, redundancy, ...) is not shown to the outside.   Failures are easy..., give out a total size exceeding the actual size on the SAN and all systems claiming the space as well.