[rabbitmq-discuss] How can I connect in a cluster using Java library

Matthew Sackman matthew at rabbitmq.com
Thu Mar 17 14:23:02 GMT 2011


On Thu, Mar 17, 2011 at 07:18:52AM -0700, Lucas Souza wrote:
> Anyone can help me?

I'm sure we can!

> On 16 mar, 16:10, Lucas Souza <luca... at gmail.com> wrote:
> > If I use Java Client Library, this behavior is automatic?
> >
> > For example, if the real message was published on node A, and node A
> > downs,
> > and I connect on node B, will I receive the message?

Ok, so it's a bit more complex: each queue resides on a single node. As
a result, if that node fails, then the queue and its contents are lost.
There are a number of ways of working around this, ranging from
publishing to multiple redundant queues, to the pacemaker work
(http://www.rabbitmq.com/pacemaker.html), to experimental work in
development on mirrored queues.

Queues _reside_ on the node to which the client was connected when the
client created the queue initially. The queue is addressable from any
node within the cluster, both from a publishing and consuming pov.

Matthew


More information about the rabbitmq-discuss mailing list