Hello all, regarding migrating from NFSv3 datastores mounted to VMware, to either NFSv4.1 or iSCSI (vmfs 6+). Up front we either lose Datastore DRS in VMware with NFS4.1, or we lose auto-size with iSCSI LUNs.
I'm looking for potential alternatives using OnTap features (ontap 9.4p5 and higher) with the goal of keeping capacity balancing between volumes and / or aggregates using fewer mount points to VMware (simulating a DRS cluster), and using NFSv4.1 to achieve encryption and keep volume auto-size. Any ideas would be appreciated.
I'm a little unclear what you're trying to achieve here. Usually we recommend storage design discussions go through the account team, even for existing setups. They aren't just there to sell you stuff, but also help set it up and bring in architects as needed, etc. They help you ensure you get the best out of your existing purchase too.
Regarding which protocol to use, IIRC NFSv4.1 on ESX is still kinda new. I heard in the most recent ESX builds it's gotten a lot better, but NFSv3 is still probably the most stable, or iSCSI/FCP. What are you needing exactly, and why are you wanting to change?
Thanks Paul. This is coming from the need to implement encryption in transit between storage & hypervisor, which NFSv3 cannot do. We may end up going iSCSI with CHAP, especially if NFSv4.1 is less stable than v3.
I've never worked with pNFS or flexgroup and was hoping we might be able to simulate DRS functionality with either of those, but so far it doesn't appear so.
If you are seeking inflight-encryption, iSCSI with CHAP is not the answer. CHAP is for secure authentication between host and storage to make sure only hosts that can provide the right credentials can access the storage. It does not encrypt the data.