We have several thousand connecting to a single volume via NFS. And we do this for multiple volumes, all hosted on multiple A700 nodes.
There are practical issues you may run into - for example, if you ever need to unmount the volume, you obviously need solid automation procedures.
There are real limits on how much I/O you can do to a single volume before you simply tip over. High I/O on a single volume anywhere in a cluster can impact the entire cluster. Even in an all-SSD environment.
As the age-old saying goes "just because you can, doesn't mean you should".