[rabbitmq-discuss] rabbitmq Error:{cannot_discover_cluster, "the nodes provided are either offline or not running"}

Emile Joubert emile at rabbitmq.com
Wed Jun 26 10:22:40 BST 2013


On 25/06/13 18:38, Luis Rodriguez wrote:
> I am running a windows host machine along with three vm's with centos
> 5.9 , rabbitmq-server-3.0.3, and erlang installed on all three nodes.

Make sure all nodes are using the same version of Erlang.

> (I am able to ping between nodes fine)

A number of TCP ports need to be open between cluster nodes, which ping
doesn't guarantee. See


> the cookie hash however, doesn't change regardless of how many times
> I stop (rabbitmqctl stop) and start.. I don't really know what I'm
> missing.

If you change the .erlang.cookie and the cookie hash doesn't change then
you are probably copying to the wrong directory. You can discover a
running node's own idea of where its home directory is by executing

rabbitmqctl eval 'init:get_argument(home).'

Also take note that the cookie and the cookie hash will not be the same.
The hash is displayed in diagnostic output instead of the cookie as a
security precaution. If you have all the cookies synchronised correctly
then the hashes will also be the same.

> when I do rabbitctl join_cluster I get: rabbitmq
> Error:{cannot_discover_cluster,"the nodes provided are either offline
> or not running"}

As per the troubleshooting guide


if that message is accompanied by a broker logfile entry "Connection
attempt from disallowed node" then the erlang cookies are not set up
correctly. Otherwise the problem is network related.


More information about the rabbitmq-discuss mailing list