How many CSVs per cluster (4 node and 8 node). A CSV per node of cluster, or a single CSV for the entire cluster (Hyper-V)

This is a specific question regarding setting up a 4 and a separate 8 node Hyper-V cluster.  Specifically, both of these clusters were setup with a single CSV.  In speaking to a vendor who is helping us setup a new backup appliance, trying to troubleshoot why when the backups of the VMs were running, we had missed polls on VMs in the cluster, as well as sometimes having a node of the cluster crash.  This has happened on both of these clusters (again, with a single CSV; which according to the backup people is incorrect, they were very specific when they said we need to have "A CSV PER NODE of the cluster)."  
So when setting up a 4 or 8 node Hyper-V cluster, does each node of the cluster require a CSV ("a CSV PER NODE"), or should there be a single CSV shared by the nodes of the clusters?
I could not find specific information on MS's technet that says exactly that, but a couple of diagrams they showed me from MS (the backup people), it looked like each node of a cluster should have their own CSV, as well as their own documentation which stated very specifically that each node must have it's own CSV to accommodate backing up the VMs through the cluster (rather than the workaround we are using now, backing each VM up as if it were a Physical Machine).
Any advice on architect the cluster correctly would be greatly appreciated.
The 8 node is Windows 2012 R2, and the 4 node is Windows 2008R2.
pogo1800Asked:
Who is Participating?

[Product update] Infrastructure Analysis Tool is now available with Business Accounts.Learn More

x
I wear a lot of hats...

"The solutions and answers provided on Experts Exchange have been extremely helpful to me over the last few years. I wear a lot of hats - Developer, Database Administrator, Help Desk, etc., so I know a lot of things but not a lot about one thing. Experts Exchange gives me answers from people who do know a lot about one thing, in a easy to use platform." -Todd S.

Philip ElderTechnical Architect - HA/Compute/StorageCommented:
What kind of storage and storage fabric is in place between compute and storage?

The answer is, "That depends." ;)

Realistically though, the typical direction we go in depends on VM density.

In smaller 2-node settings we would create two CSVs to distribute I/O between the two nodes evenly.

In a more dense situation as is the case here it really depends on the workloads the fabric carrying I/O to storage.

For the eight node setup we'd run with 4-8 CSVs depending on workload IOPS requirements.

For a higher IOPS requirement we would set up a dedicated LUN/CSV with a smaller block size in the storage stack which is typically 64KB.
For a higher bandwidth requirement we would set up a dedicated LUN/CSV with a larger block size in the storage stack which would typically be 512KB to 1024KB or in some rarer cases even 2048 KB (4K video files being pushed to static storage).

I have two EE articles that provide a lot of information on our methodologies:
Practical Hyper-V Performance Expectations
Some Hyper-V Hardware and Software Best Practices

Experts Exchange Solution brought to you by

Your issues matter to us.

Facing a tech roadblock? Get the help and guidance you need from experienced professionals who care. Ask your question anytime, anywhere, with no hassle.

Start your 7-day free trial
pogo1800Author Commented:
Thank you very much for the information.  I will be reviewing it today and replying and marking solution if appropriate.  Thank you again for your time and expertise.
pogo1800Author Commented:
I have spent some time reading over the articles, thank you very much.  My question was a fairly strightforward one, and you gave me much more, thanks again.  What it really set us straight on was how to calculate the number of CSVs, etc.  The main problem is trying to rearchitect, on the same hardware, a cluster like that.  We may have to use a workaround backup for the time being, until we reach a time at which the hardware will be replaced (sooner for one than the other), and pay closer attention to the role of CSVs when architecting the cluster.  Thank you again.
pogo1800Author Commented:
Answered the main question, and went further in providing the articles that will help not only answer the question, but allow us to properly tune that answer to our specific situation, equipment, and needs.
It's more than this solution.Get answers and train to solve all your tech problems - anytime, anywhere.Try it for free Edge Out The Competitionfor your dream job with proven skills and certifications.Get started today Stand Outas the employee with proven skills.Start learning today for free Move Your Career Forwardwith certification training in the latest technologies.Start your trial today
Virtualization

From novice to tech pro — start learning today.