You should be able to share FC luns between two different clusters. However you will need to create dedicated LVM VG slices for each cluster, so that they dont step on each other. Additionally you will be limited to LVM thick, which means no snapshots.
Other option would be dedicated LUNs per cluster/member and Ceph, the performance implications of data protection amplification could be very significant.
Another option is to front FC with a Software Defined solution that can be extremely flexible in slicing and dicing the global pool, ie Blockbridge.