Network and Storage Protocols

NAS SVM - Seperate NAS LIFs per node?

prcpa8w3p

Hi there,

 

We have created a dedicated IP-only NAS SVM. NetApp Best Practices state each SVM shall have 1 LIF/node.

 

Taking that 1 step further, do you recommend separating CIFS and NFS to each have their own dedicated LIFs per node?

 

This is a 2-node cluster comprised of ESX, VIO, and TSM (to be replaced with CommVault later this year.)

 

TIA

1 ACCEPTED SOLUTION

robinpeter

Netapp recommendation is *minimum* one LIF per node for each SVM..

And seperating NAS and SAN SVM's are recommended. 

 

Seperating the NAS LIFs (NFS and CIFS) is depends on the data your cluster is serving.

there are customers who serve same data to NFS and CIFS clients.. in that case there is no advantage in having sperate LIF for each protocol. (unless they are different network)

 

Its always good to have dedicated network (usually non routable) for ESX datastore.

otherwise, keep it simple, one NAS LIF per node per SVM.

 

Robin.

View solution in original post

3 REPLIES 3

robinpeter

Netapp recommendation is *minimum* one LIF per node for each SVM..

And seperating NAS and SAN SVM's are recommended. 

 

Seperating the NAS LIFs (NFS and CIFS) is depends on the data your cluster is serving.

there are customers who serve same data to NFS and CIFS clients.. in that case there is no advantage in having sperate LIF for each protocol. (unless they are different network)

 

Its always good to have dedicated network (usually non routable) for ESX datastore.

otherwise, keep it simple, one NAS LIF per node per SVM.

 

Robin.

View solution in original post

AdvUniMD

I would definitely separate CIFS and NFS interfaces. For example, NFS for ESX should be in a separate Layer-2 subnet and not routed, so this is usually different from CIFS and requires separate LIFs.

 

Also, in larger environments, especially if you use NFSv3 with ESX or Oracle or something, it helps if you actually use one NFS LIF for each Datastore/Mount. Because then if you later decide to move a volume to an Aggregate on a different node ("vol move", for performance reasons or something), you can move the corresponding LIF to the destination node as well, so that the traffic does not go through the cluster backend network

prcpa8w3p

@AdvUniMD - doesn't 1 LIF per volume create significant administrative overhead? Also, what's the ONTAP Maximum LIF limit per node in 8.3.2P2 and 9.1P3?

Announcements
Register for Insight 2021 Digital

INSIGHT 2021 Digital: Meet the Specialists 2

On October 20-22, gear up for a fully digital, totally immersive virtual experience with a downright legendary lineup of world-renowned specialists. Tune in for visionary conversations, solution deep dives, technical sessions and more.

NetApp on Discord Image

We're on Discord, are you?

Live Chat, Watch Parties, and More!

Explore Banner

Meet Explore, NetApp’s digital sales platform

Engage digitally throughout the sales process, from product discovery to configuration, and handle all your post-purchase needs.

NetApp Insights to Action
I2A Banner
Public