Complete shutdown of 4.3.3 cluster, change cluster from 4.2 to 4.3 compatibility and upgrade of all hosts to current 4.3.3 with patches has seemingly fixed everything.