in NFS datastores, each virtual disk file has its own I/O queue directly managed by the NetApp FAS system.
in LUN datastores, it doesn’t natively allow a storage array to identify the I/O load generated by an individual VM.
Would you please explore these two statements more? I could not fully understand them. In NFS, please explain to me in detail or by example on how I/O queue directly managed by a filer.
Why in LUN datastores the filer cann't identify the I/O load?
Also, in a "best practice" document, it says that virtual disks could be connected to all nodes in the clustter but is only accessed by a single VM, i thought it could be shared which means it could be access by multiple VM's.
Thanks for your inputs.
With an NFS datastore, each VM is a file (a few files) which the Filer hosts. With a LUN, it all appears as a single file to the filer, regardless of how many VMs are in the LUN.
All ESX cluster hosts can see each datastore and therefore every VM. I think you may be confusing virtual disks and datastores.