Individual Disks vs Storage Pool

[magnosantos]

I'm planning on implementing Virtual SAN for a 2 Node FailOver Cluster with Hyper-V.
The Servers will have a bunch of disks (equally by each one). What would be the best for performance? One Target by Disk, or creating a Storage Pool with All Disks?

I can dedicate 100GB od SSD for Caching.

VMs will be of differente nature, including FileServer, ERP, and Dev Machines.

[Oleg(staff)]

Hi magnosantos,
What type of disks do you have in the bunch, SSDs and HDDs? And do you have RAID controller installed on the servers?

[sunniegoldie]

For your 2 Node FailOver Cluster with Hyper-V, implementing a Virtual SAN, the best approach for performance would depend on your specific workload requirements and setup. Generally, creating a Storage Pool with all disks can provide better overall performance and easier management compared to configuring one target per disk.

A Storage Pool allows for the aggregation of all available storage resources, enabling features like striping and redundancy which can significantly enhance performance and reliability. With the added 100GB of SSD for caching, the Storage Pool can leverage this for improved read/write speeds, benefiting all your virtual machines, including FileServer, ERP, and Dev Machines.

Using a single Storage Pool simplifies administration and optimizes resource utilization across your cluster, ensuring that your VMs, especially those running resource-intensive applications like ERP software, have the necessary performance and redundancy to operate smoothly.

In summary, a Storage Pool with SSD caching should provide a balanced and high-performance solution for your Virtual SAN setup.

[yaroslav (staff)]

L2 cache improves only reads. StarWind VSAN has reads boosted from the box as the client reads form 2 mirrors (i.e., 2x read boost).