OK, here's something interesting (at least to me):

This may be working as designed, but seems strange to me. Suppose I
have the following setup:
3 node SLES/OES2 cluster (volumes are all clustered NSS)
NodeA hosting volA
NodeB hosting volB
NodeC hosting volC
virtualIP_A --> volA
virtualIP_B --> volB
virtualIP_C --> volC

If I connect to virtualIP_A from a Mac and login, volA mounts seamlessly
without displaying the "Select volumes to mount" dialog.

If I migrate or fail, say, volC to nodeA (so now nodeA is hosting both
volA & volC), when I connect to either virtualIP_A or virtualIP_C, I get
the "Select volumes to mount" dialog, and must choose to mount either
volA or volC. This behavior I consider undesirable, because it can
potentially serve to confuse users as volumes migrate from node to node.

Shouldn't the clustered volume follow the virtual IP and just mount as
if the node was hosting just that volume?