[rabbitmq-discuss] Thread per channel and partitioning by channel (java client)

Fri Feb 15 21:57:30 GMT 2013

On Fri, Feb 15, 2013 at 1:19 PM, Matthias Radestock
<matthias at rabbitmq.com>wrote:

> Jonathan,
>
>
> On 15/02/13 20:53, Jonathan Halterman wrote:
>
>> What I was reading was from here:
>>
>> http://www.rabbitmq.com/api-**guide.html#consuming<http://www.rabbitmq.com/api-guide.html#consuming>
>>
>> "Each Channel has its own dispatch thread"
>>
>
> That isn't meant to imply that there is one thread per channel, though I
> can see how it could be read that way.
>
> Consumer processing happens on the threads in the Consumer thread pool See
> www.rabbitmq.com/api-guide.**html#consumer-thread-pool<http://www.rabbitmq.com/api-guide.html#consumer-thread-pool>.
> But processing is constrained s.t. all messages arriving on a channel are
> processed serially, i.e. at any point in time at most one thread in the
> pool is processing consumer work for a given channel.
>
> I am cc'ing my colleague Steve, who may be able to shed more light on this
> and what he meant by "dispatch thread" above.

I understand, thanks for the clarification. It was just the wording of the
docs led me to believe there was a separate thread per channel in addition
to the consumer threadpool. Glad to know there is not!

>
>
> Matthias.
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.rabbitmq.com/pipermail/rabbitmq-discuss/attachments/20130215/36c97403/attachment.htm>