We have single SVM with 1 LIF and SVM contains 6 NFS volumes exported through single export rule.
Now, we are planning to create one more LIF (since we have 2x 10Gig physical ethernet ports) and export 3 NAS volumes to 192.168.* segment and remaining 3 NAS volumes needs to be exported through 172.20.* segment. Can this is possible through single export-policy rule which contains all client addresses in both IP segements?
or do we need to create one more SVM and present 3 NAS volumes in each SVM through 1 dedicated LIF?
I understand that, after SVM creation we have to move 3 NAS volumes from exisitng SVM to new one. Please note that, we have only one aggregate and new SVM will also be placed in that. Will Volume migration from one SVM to another require any downtime? How long migration will take to complete?
I think you can accomplish this multiple ways... you could use the same SVM. Export policy rules are one host or subnet per entry (netgroup is an exception for multiples). If exporting by subnet, you could create 2 rules and be set in the same svm... or separate into a new SVM.
There is no supported capability to reassign a volume between SVMs at this time. You can however snapmirror between SVMs to get the volumes there...but in this case I'm not sure you really need a separate SVM. You will net route create the 0.0.0.0/0 default gateway for the new network in the same or new SVM anyway.
Regarding LIFs... even if one SVM, I prefer at least 1 IP per node, then round robin or dns delegation to get the hosts load balanced...if home directory type data. If a datastore, then direct access is best practice but still would have a lif per datastore and more than one IP... this lets you load balance networking.