empty server groups should be ignored when deciding if server groups configuration has changed

3 nodes A, B and C install couchbase server community 2.5.1-1082
Create cluster of node ABC
Create bucket default with 1 replica. No RZA feature.
Install couchbase server enterprise 2.5.1-1082 at node D (151 as in capture screen)
Do upgrade from community edition to enterprise edition by swap rebalance.
Add node D into cluster and remove node C. Rebalance done.
Uninstall couchbase server CE 2.5.1-1082 in node C and install couchbase server EE 2.5.1-1082 back to node C
Add node C into cluster and remove node B. Rebalance done.
Uninstall couchbase server CE 2.5.1-1082 in node B and install couchbase server EE 2.5.1-1082 back to node B
Add node B into cluster and remove node A (node 148). Rebalance done.
At the end of rebalance, rebalance button grey out as expected.
Then rebalance button comes back as a clickable button

Log shows the end of swap rebalance (remove node 148 and add node 149)

Haven't heard from a higher priority node or a master, so I'm taking over. mb_master000 ns_1@192.168.171.149 17:45:17 - Fri Mar 21, 2014
Node 'ns_1@192.168.171.149' saw that node 'ns_1@192.168.171.148' went down. Details: [

{nodedown_reason, connection_closed}] ns_node_disco005 ns_1@192.168.171.149 17:45:07 - Fri Mar 21, 2014
Node 'ns_1@192.168.171.150' saw that node 'ns_1@192.168.171.148' went down. Details: [{nodedown_reason, connection_closed}

] ns_node_disco005 ns_1@192.168.171.150 17:45:07 - Fri Mar 21, 2014
Node 'ns_1@192.168.171.151' saw that node 'ns_1@192.168.171.148' went down. Details: [

{nodedown_reason, connection_closed}

] ns_node_disco005 ns_1@192.168.171.151 17:45:07 - Fri Mar 21, 2014
Node 'ns_1@192.168.171.148' is leaving cluster. ns_cluster001 ns_1@192.168.171.148 17:45:07 - Fri Mar 21, 2014