MSA 2052 and Windows Server 2016

[KIngstone]

Hi guys,

I'm not sure whether to ask this on an HPE or Microsoft forum. Hope I'll get some help.

I'm virtualizing and clustering two HPE servers, and a MSA SAN. When creating Hyper-V VMs, should I host all VMs on the same volume/LUN, or create a volume for each VM created? The later preferably in case i need to restore a specific VM from a snapshot, or there's something I'm missing.

[SUBHAJIT KHANBARMAN_1]

I think you can consider Cluster Shared Volume (CSV)

Please refer the "HPE MSA 2050/2052 storage and Microsoft Windows Server 2016 Implementation Guide",

https://h20628.www2.hp.com/km-ext/kmcsdirect/emr_na-a00038738en_us-1.pdf

Hope this helps!
Regards
Subhajit

I am an HPE employee

If you feel this was helpful please click the KUDOS! thumb below!

****************************************************************************

[AshrafK]

Hi There,

Hyper-V can use different types of storage and methods of storage connectivity and configuration.

While keeping in mind on shared configuration plan, Microsoft Server Hyper-V 2016 running on the HPE Server with MSA SAN 2050 supports shared storage either via Fibre Channel or iSCSI.

Since we are talking on the Hyper-V side, it is essential to think through both HYPER-V Node level storage provisioning and Virtual Machine Level Storage provisioning.

CSV shared volume can host all the VMs running on the Hyper-V Cluster nodes to which the CSVs are presented. But, when deploying VMs and VHDs to CSVs, it is not a best practice to create a single large LUN or logical disk to hold all virtual machines. Besides, It is better to separate the VM System files and the VM Disk files at the LUN level. Create the VHD on the appropriate CSV volume.

Moreover, as you mentioned about restore of VM from snapshots; if it operates at the individual LUN level, and when checkpoint should fill all available space, only the virtual machines sharing that space will be affected.

Other benefits that you achieve doing so will grant you the flexibility to use Storage Live Migration to address changing needs. For example, moving virtual hard disks into a fresh space automatically defragment them.

Furthermore, many hypervisor-aware backup solutions can leverage hardware snapshot providers, which will operate at the LUN level. Separated LUNs means smaller snapshots and quicker operations.

Microsoft makes additional recommendations with Hyper –V and failover clustering. Thus, it is also recommended you adopt all best practices described in the MSA best practice document according to the MSA model in use to ensure supportability and maximum performance.

Hope this information help!

[KIngstone]

Hi Subhajit,

The CSV idea is well understood. But the question is, is it ideal to have a CSV per VM or geenral CSVs for the VM VHDs and config file. THe question is mostly from the idea that in case I restore a CSV snapshot when i have a single failed VM, does it restore all the VMs in this CSV?

[SUBHAJIT KHANBARMAN_1]

Your question more specific to Windows related rather MSA related.

Also this type of query can't be answered just like that because this requires proper consultant work.

It depends upon your business requirement, no of users use, IOPs per VM and many more factors. If you are using all similar type VMs then 100 VMs within single CSV make sense as you can upgrade and do the changes in one shot or similar type of situation.

Sometime it may require that one VM may need to be associated with single MSA volume for heavy workload, memory usage and many other factors like application that will be running.

Coming to CSV snapshot restore, this option itself not good or correct. You are recovering a snapshot - so this node (at the time of recovery) is going back in time to a previous day while all the other nodes exist in today. What if the moment you took the snapshot he was the cluster master?  or he was performing some action on your CSV?  or your configuration changed and suddenly he tells everyone else that there is a configuraiton problem as he thinks that new volume is now offline (and he became the master when he came on line).

It's better to use VSS or any other 3rd party app for backup solution.

You can consider checking these links,

https://docs.microsoft.com/en-us/windows-server/failover-clustering/failover-cluster-csvs

Best think is to involve Solution Consultant and get proper help from design perspective.

Hope this helps!
Regards
Subhajit

I am an HPE employee

If you feel this was helpful please click the KUDOS! thumb below!

************************************************************************

[AshrafK]

Best practice from storage side (HPE)

One VHD for VM System files on one CSV, and another VHD for VM data files on another CSV.

Moreover, Virtual Machine snapshot ( Microsoft terminology it is called “checkpoints” )  is at the VHD level for each VM, as snapshot differencing disks (.AVHDs) get created and stored in a folder under the virtual machine’s snapshot directory.

Thus, only the virtual machine in question for restore to certain snapshot checkpoints will only be restored not all VMs.

[SUBHAJIT KHANBARMAN_1]

Can you please help us to understand if the issue got resolved or not?

If issue got resolved then how?

Also request you to mark the forum as resolved if there is no more outstanding query from your end on this issue.

This will help for everyone who are all following your forum.

Hope this helps!
Regards
Subhajit

I am an HPE employee

If you feel this was helpful please click the KUDOS! thumb below!

***************************************************************************