[rabbitmq-discuss] Doubt regarding rabbit servers in clustered environment

[Kshitiz Garg]

Thanks Marek,

We will do our homework on these lines and will post our doubts again.

Thanks & Regards,
Kshitiz Garg

On Mon, Nov 22, 2010 at 3:55 PM, Marek Majkowski <majek04 at gmail.com> wrote:

> On Sat, Nov 20, 2010 at 05:01, Kshitiz Garg <stephanion2002 at gmail.com>
> wrote:
> > Since our 8 queues are going to handle lot of data, we can cluster 8
> > RabbitMQ nodes, each to serve a single queue and its contents. But there
> are
> > two concerns:
>
> Remember that any modification of shared database must be agreed between
> all the nodes in the cluster. 8 nodes is a lot and might result in a
> latency for commands like exchange.declare or queue.bind. We rarely
> see a cluster of more than a few nodes.
>
> > 1. In the case a node (say, containing queue a) is down, how will the
> queue
> > a's consumers will get their messages?
>
> They won't.
>
> > 2. A load balancer like HA Proxy in front of these nodes will not be of
> any
> > use since nodes are not similar.
>
> It would help to spread the connections equally between nodes.
> That is pretty useful if you have a lot of connections.
>
> > I am afraid that since we are neither getting "Load balancing" nor "High
> > availability",  we will have to consider another broker.
>
> For HA we're usually suggesting Active-Passive failover:
>  http://www.rabbitmq.com/pacemaker.html
> We're currently working on Active-Active HA, but it will take us some time.
>
> Clustering does help in "Load balancing" but only in some situations.
> For example it may be pretty useful for the 'fanout' cases, when you
> have many queues.
>
> But if you have a single queue, clustering won't help you.
> It's a bit like with database sharding, if you have one monolitic table -
> sharding won't help.
>
> Cheers,
>   Marek
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.rabbitmq.com/pipermail/rabbitmq-discuss/attachments/20101122/90f4d29a/attachment.htm>