[rabbitmq-discuss] Clustering and client reconnection on node failure

Thu Jun 23 21:56:58 BST 2011

I know there are several topics on this in the list already but still
not totally clear on this. If I am only using non-durable queues, and
I have a 2 node cluster, if node A fails, a client connected to queues
on node A will (I assume) get a connection failure of some kind. At
that time, is my correct recovery path to reconnect to the other node,
and re-declare my queues and bindings ? Seems simple enough, I just
need to maintain a list somewhere of all hosts and their availability
status.