Handle missing guids on startup

When two or more nodes start at the same time, we have to use the list
of running mnesia nodes instead of the fully running rabbit nodes. When
a node is in the node_monitor's init stage, it isn't counted as a fully
running rabbit node. This means that its GUID isn't sent to the other
node starting up. This race condition means that the node_monitor will
not do anything with a node_down event, because the down node will be
missing from the GUID map.

Author: olikasg

deps/rabbit/src/rabbit_node_monitor.erl

@@ -413,7 +413,7 @@ handle_call(_Request, _From, State) ->
     {noreply, State}.
 
 handle_cast(notify_node_up, State = #state{guid = GUID}) ->
-    Nodes = rabbit_nodes:list_running() -- [node()],
+    Nodes = rabbit_nodes:list_reachable() -- [node()],
     gen_server:abcast(Nodes, ?SERVER,
                       {node_up, node(), rabbit_db_cluster:node_type(), GUID}),
     %% register other active rabbits with this rabbit