When a partition drops to a single node, the remaining leader gets stuck in an infinite AC_Propose, AC_Nack loop with itself. Need to investigate root cause.but a likely culprit seems to be the TOB scheduled timeout triggering more broadcasts than needed.