Qpid cluster wont start after synchronous shutdown if cluster-size option used

Issue

Having a 2node cluster with cluster-size=2 set, there are 2 scenarios where starting fully stopped cluster wont succeed.

1) Whole cluster was stopped by "qpid-cluster -k" in one shot, marking both stores as clean. Subsequent startup of 2nd node fails with "error Exchange already created", while the first node waits for second node startup forever to start operating (due to cluster-size=2).

2) Stopping node A and then node B makes stores dirty on node A and clean on node B. Subsequent startup of node A first and then starting node B stalls forever (or timeouts if run as a daemon).