Best Practice for creating a Windows 2012 R2 file server VM with a C: (OS) and D: (DATA) partition?

[quantum_it]

I am in process of deploying a Windows 2012 R2 Standard server, as a primary file server for our environment.  I have create an 80GB VM, which is acting as the C: drive, and is solely for the OS and logs.

My questions is, for the 😧 partition/volume (used only for data sharing/collection) does it make more sense from an administrative standpoint to create this as another virtual machine and then mount that as the 😧 drive, or is it just simpler to create a LUN from my storage array and mount it to the virtual machine using iSCSI?  I'm using a Nimble Storage array for all VMs and iSCSI LUNs.

The initial size of the 😧 volume will be 1.5TB, and I want the flexibility to be able to dynamically grow it, when it starts reaching capacity.

[BrianNorth8]

There are many advantages to leaving everything inside of the vmfs partitions that I would recommend that solution as a primary choice.  If you are using storage specific features like replication or shared disks for clustering then I would consider using the iSCSI mount option.

Depending the sizing of your LUNS you may also choose to set a smaller vmdk size, like 2, 750 GB vmdk and using windows dynamic disks to make the 1.5 TB d drive.  This gives you the option of moving the vmdks between multiple luns independently.  I use 4 TB LUNS in my environment; therefore try to avoid having vmdk larger than 1 TB each.