why does changing nfs.enable-ino32 on a live Gluster NFS volume cause problems with clients?

Issue

• After enabling volume option nfs.enable-ino32 on with the command on Red Hat Storage Server side on a live volume causes client side to observe Stale File Handles

# gluster volume set <volname> nfs.enable-ino32 on 

• Volume accepts setting, but several client receives "Stale NFS handle" As a result the application cannot start and a remount is necessary to get things back to normal.

• Ideally this setting should be applied live without errors allowing no remounts on the client side