[rabbitmq-discuss] Clustering - communication

[Dan Dart]

Hi all,

I'm having trouble getting two nodes talking. For reference they're
called backend and back2 - and each is in each other's hosts file
(with and without a FQDN).

Each can connect to each other on ports 4369 and 5672 - and have the
same .erlang.cookie file (had to hexedit to get it exact).
They're not connecting through the same network but they're not far
away (<1ms ping)
When I try to cluster (taking the node down using stop_app first) - I
get the error:

Error: {no_running_cluster_nodes,[rabbit at backend],[rabbit at backend]}

When I try to use rabbitmqctl -n rabbit at back2 status, from backend, I get:

Status of node rabbit at back2 ...
Error: unable to connect to node rabbit at back2: nodedown
diagnostics:
- nodes and their ports on back2: [{rabbit,48765}]
- current node: rabbitmqctl18870 at localhost
- current node home dir: /var/lib/rabbitmq
- current node cookie hash: (the same each time)

and vice versa.

Can anyone give me some pointers? Do both servers need to be
stop_app'd at the same time for this to succeed or is it something
else?
Also, will stop_app kill any messages? I'm looking not to.