Lun configuration for HV cluster

Question

Hi,

We have about 110 Hyper-V virtual machines.

Currently they are hosted on about 20 hyper-V host servers, some clustered, some are not.

We are looking to migrate/consolidate all 110 virtual machines, to a single 6 node Hyper-V failover cluster.

I am trying to figure out the best way to configure the cluster shared volumes..... I think I have figured out a way that should work, but would like others opinions.

Currently we have a VNX5300, we just got the FAST suite and will be using configuring a storage pool with a NL-SAS, SAS, and FLASH tier with FAST Cache as well. So the luns I will be carving will be out of that storage pool (roughly 30 NL sas, 30 SAS, and 2 flash for cache, one spare flash that I wasn't sure if it should go in the pool or sit as a hot spare.

I'm estimating about 40-50gb per system for the OS x 110 vm's = 5500 GB I was thinking about provisioning this on a thin lun, but a little unsure on that...

So for the cluster shared volume for OS drives, 6TB seems a little large... and as a thin lun I know there will be some kind of performance hit, but I don't have any hands on experience with that yet.

Then I was thinking I would create individual luns for Data/Log drives per SQL server, or to support whatever the needs may be.

Roger_Wu · Answer

You can read the following whitepapers before making the plan,

EMC Infrastructure for Microsoft Private Cloud

Virtualizing SQL Server 2008 Using EMC VNX Series and Microsoft Windows Server 2008 R2 Hyper-V Reference Architecture

Using EMC CLARiiON with Microsoft Hyper-V Server

EMC VNX Virtual Provisioning

You might find the recommendations like below from the above whitepapers. But at this point Thin is a tradeoff with performance, so you're right, it's not recommended for critical, high IO VMs.

In addition, if you DO like to use Thin (Hyper-V thin, Dynamic VHD) on Thin (storage thin LUN), use Hyper-V or 3rd party usage reports in conjunction with array-level reports, and set thresholds with notification and automated action on both the hypervisor layer (and the array level (if you array supports that). Why? Thin provisioning needs to carefully manage for “out of space” conditions, since you are oversubscribing an asset which has no backdoor. When you use Thin on Thin - this can be very efficient, but can “accelerate” the transition to oversubscription. You can refer to the following blog for this point:

Thin on Thin? Where should you do Thin Provisioning vSphere 4.0 or Array-Level? - Virtual Geek

VNX

Lun configuration for HV cluster

Was this post helpful?