If I want users to restore from the local Snapshot, the user can just cd to volume "xyz" .snapshot directory and copy data from it. This can be done by users themself.
What if the user wants to restore data from SV on the remote filer, I need to export the volume "xyz_sv" on the secondary to the same client, and then mount "xyz_sv", and then copy data under "xyz_sv/.snapshot" to xyz on the host. Am I correct? In essence, the volume on the secondary has to be different than primary one
Assuming this is for NFS I would suggest you use the automounter and then we've started placing a symlink in the primary volume (qtree in our case) that points to the autofs mount point of the snapvault volume. As long as you follow a standard naming convention on both sides of course..