[rabbitmq-discuss] Scaling Throughput with Highly Available Queues

Tue Mar 19 23:10:25 GMT 2013

That's what I was thinking - the master queue processes are spread across the nodes in the cluster. I will have to go back and take a look at what it takes to ensure that happens...

Thanks!

From: rabbitmq-discuss-bounces at lists.rabbitmq.com [mailto:rabbitmq-discuss-bounces at lists.rabbitmq.com] On Behalf Of Richard Raseley
Sent: Tuesday, March 19, 2013 4:07 PM
To: Discussions about RabbitMQ
Subject: Re: [rabbitmq-discuss] Scaling Throughput with Highly Available Queues

Ron,

You have a couple options (that I know of) if your requirements include a *single* cluster (which was not my case):

1) Increase the resources (CPU, disk, memory, network, etc.) which are available to the nodes which are hosting the master queue processes.

2) Ensure that the master queue processes are spread out evenly across the available nodes.

Regards,

Richard

On Tue, Mar 19, 2013 at 3:48 PM, Cordell, Ron <Ron.Cordell at relayhealth.com<mailto:Ron.Cordell at relayhealth.com>> wrote:
Sorry for jumping in here, but I'm a little confused on how to scale throughput in the HA scenario. If I assume that all queue are mirrored queues on a single cluster then what are my options to adding more throughput should I need to?

From: rabbitmq-discuss-bounces at lists.rabbitmq.com<mailto:rabbitmq-discuss-bounces at lists.rabbitmq.com> [mailto:rabbitmq-discuss-bounces at lists.rabbitmq.com<mailto:rabbitmq-discuss-bounces at lists.rabbitmq.com>] On Behalf Of Richard Raseley
Sent: Tuesday, March 19, 2013 3:36 PM
To: Discussions about RabbitMQ
Subject: Re: [rabbitmq-discuss] Scaling Throughput with Highly Available Queues

Matthias,

I am glad that you were able to anticipate where I was heading - that means I am not completely insane (at least as it relates to this). =]

Yes, I understand that they are actually two separate queues (I am treating them as one logical one for purposes of discussion). In this case the lack of order preservation and visibility from one cluster to the other isn't an issue for us. We decided not to go with two queues for the transparency reasons you mentioned.

Thanks so much for your time today.

Regards,

Richard

_______________________________________________
rabbitmq-discuss mailing list
rabbitmq-discuss at lists.rabbitmq.com<mailto:rabbitmq-discuss at lists.rabbitmq.com>
https://lists.rabbitmq.com/cgi-bin/mailman/listinfo/rabbitmq-discuss

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.rabbitmq.com/pipermail/rabbitmq-discuss/attachments/20130319/bff7462d/attachment.htm>