Thanks, DinoRS.  That's kind of what I expected.  I suggested the upgrade-in-place option, since it is a cluster and if things went south we could make the 2012 node active...