Multiple Shares vs Datasets...or both

[alpaca]

Use case: NFS share(s) as file storage for media archiving attached to load-balanced web servers.

Am in the process of completing a move of some archives I host over to NFS shares. I have been creating a separate dataset per each "client/user" and also a unique NFS share for each. Each linux web server then has the multiple NFS shares mounted, symlinked to the appropriate web-root directory. It is important to keep each archive on it own dataset for quota purposes BUT, do I really need to have separate NFS shares? Looking to see if anyone has seen a pro or con in production to multiple shares or just linking to the specific dataset path?

Currently= 4 datasets and 4 NFS shares

Any benefit to = 4 datasets but 1 NFS share?

Thanks

 

[cyberjock]

That is a question that is so wide that the answer is really only something you'd know about. It's like asking if a truck provides any benefit over a compact car. Well, it provides both benefits and drawbacks.

 

[alpaca]

Thanks for the quick reply. That is a good enough answer and sort of what I expected. Will have to see what happens under real-world load and go from there. It was a little bit of a fishing question, but thanks.