[rabbitmq-discuss] Error: cannot_discover_cluster: The nodes provided are either offline or not running

[Tim Watson]

On 20 Jun 2013, at 18:58, k.madnani84 wrote:
> What modifications do we need in /etc/hosts and should it be in all the
> nodes?
> I was under the impression if i am able to ping the other node and telnet on
> epmd port,thats enough to confirm the connectivity,which i am able to do?
> 

If your nodes are using the hostname in their node names, then both machines need to be able to resolve one another's host name via DNS. For example, if when you start the broker (or run `rabbitmqctl status` against an existing broker) you see that the node name is "rabbit at host1" then any other nodes that wish to cluster with this broker will need to be able to `ping host1` - pinging the IP address isn't enough, since Erlang will use DNS to attempt external host discovery.

When you run `rabbitmqctl join_cluster CLUSTER1 at zldv0432` the local node needs to be able to resolve zldv0432, and a common way to make that happen is editing the hosts file on each machine.

Cheers,
Tim