[rabbitmq-discuss] Creating exchanges on demand.
Camilo Lopez
camilol at gmail.com
Fri Jun 5 20:41:52 BST 2009
OK, I am mostly worried about the possible performance/memory hit, I
don't want to put some code in production and realize I just run out
of memory or becoming unresponsive. So I undestand that's not the
case.
On Fri, Jun 5, 2009 at 3:35 PM, Alexis Richardson
<alexis.richardson at gmail.com> wrote:
> Creating a lot of exchanges is fine, but many designs naturally fit
> just a few. There is no 'one' answer.
>
>
>
> On Fri, Jun 5, 2009 at 8:30 PM, Camilo Lopez<camilol at gmail.com> wrote:
>> Thanks for the link ,
>>
>> No I had not seen it , I'm not using Java (using ruby btw) so I did
>> not really looked into the Java client specifics. I just checked and
>> one of the ruby libraries does implement an RPC convenience.
>>
>> I guess this will be a better idea that what I'm doing now, but in
>> genral is it a bad idea do create a bunch of Exchanges?
>>
>>
>> On Fri, Jun 5, 2009 at 3:03 PM, Arvind Jayaprakash <work at anomalizer.net> wrote:
>>> On Jun 05, Camilo Lopez wrote:
>>>>Hello list,
>>>>
>>>>I'm pretty new to AMQP and Rabbit. I'm trying to solve a problem where
>>>>I have a queue of incoming jobs and a pool of worker process. Some of
>>>>the jobs have to be processed syncrhonically.
>>>>
>>>>There is one producer process feeding jobs to a incoming queue (q1),
>>>>every job has a unique id encoded in the message. All the workers read
>>>>from q1, do the processing and put back the result in a outgoing
>>>>queue (q2) using the unique id as routing key.
>>>>
>>>>When a job is synchronous, the producer will subscribe to q2, waiting
>>>>for a result tagged with the unique id assigned when it was put in
>>>>q1.
>>>>
>>>>Now, in order to wait for a specific message the producer creates a
>>>>new exchange for every synchronous job with the unique id as key and
>>>>binds it to the queue before subscribing.
>>>>
>>>>Now the question/s:
>>>>
>>>>Is it OK to create a large number of exchanges? is it going to impact
>>>>performance?
>>>> or Am I doing something terribly wrong here? is there a best
>>>>practice I can refer to? if so any advice on the correct way to do
>>>>this would be appreciated.
>>>
>>> Have you looked at this?
>>> http://www.rabbitmq.com/api-guide.html#rpc
>>>
>>> _______________________________________________
>>> rabbitmq-discuss mailing list
>>> rabbitmq-discuss at lists.rabbitmq.com
>>> http://lists.rabbitmq.com/cgi-bin/mailman/listinfo/rabbitmq-discuss
>>>
>>
>> _______________________________________________
>> rabbitmq-discuss mailing list
>> rabbitmq-discuss at lists.rabbitmq.com
>> http://lists.rabbitmq.com/cgi-bin/mailman/listinfo/rabbitmq-discuss
>>
>
More information about the rabbitmq-discuss
mailing list