[rabbitmq-discuss] Federation and Clustering...

[David Buckingham]

I apologize for the long-winded-ness, I hope that I've painted a good picture of what I'm trying to do...

Currently, I have a Federated Exchange setup between 2-nodes, let's call them rabbit1 and rabbit2.  When I publish messages to the exchange on rabbit2, they can be received in a queue on rabbit1.  Fantastic!

Now I'd like to add clustering into the mix.  I created a new node, let's call it rabbit3, and created a cluster between rabbit2 and rabbit3.  I've updated the local-nodename federation parameter on both rabbit2 and rabbit3.  I've added rabbit3 as an upstream for the federated exchange on rabbit1.

With all 3 nodes running, if I publish to the upstream exchange on either rabbit2 or rabbit3, the message properly appears on rabbit1.  If I shutdown rabbit2 (which was the original upstream node), and I try to publish to the upstream exchange on rabbit3, I received "Message published, but not routed." (I'm testing via the Management UI).

If I look at the Federation Status on rabbit1, it appears that the State of the connection to rabbit2 toggles between "error" and "shutdown", while the State of the connection to rabbit3 toggles between "starting" and "shutdown".

Further investigation shows that the internal upstream queue does not exist on rabbit3, but does exist on rabbit2.

I'm not quite sure where I may have gone wrong.  By adding the new clustered node as an upstream on rabbit1, I guess that I expected the same configuration would be pushed-to/created-on rabbit3 that was previously created on rabbit2.

My intentions are to publish messages to a clustered exchange which are then pushed up to the federated exchange.

Any ideas?
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.rabbitmq.com/pipermail/rabbitmq-discuss/attachments/20121127/5ecc4bb3/attachment.htm>