[rabbitmq-discuss] Puka client and publisher confirms performance
Aaron Westendorf
aaron at agoragames.com
Tue Mar 20 12:52:03 GMT 2012
I looked at the changes and in addition to those, I recommend using
the Python deque class for your buffers. I noticed a significant
improvement in performance when I switched to them in haigha.
cheers,
Aaron
On Tue, Mar 20, 2012 at 6:41 AM, Marek Majkowski <majek04 at gmail.com> wrote:
> On Wed, Mar 14, 2012 at 18:59, Marek Majkowski <majek04 at gmail.com> wrote:
>> On Wed, Mar 14, 2012 at 18:42, Marek Majkowski <majek04 at gmail.com> wrote:
>>> On Wed, Mar 14, 2012 at 12:31, Christos Stavrakakis
>>> <stavr.chris at gmail.com> wrote:
>>>> I am using Puka client [1] in order to send messages with guaranteed
>>>> delivery using publisher confirms.
>>>>
>>>> Trying to send 200.000 messages I see that waiting for the confirms, takes a
>>>> lot of time.
>>>>
>>>> To be more specific:
>>>>
>>>> promises = []
>>>> t0 = time.time()
>>>> for i in range(0,200000):
>>>> promise = client.basic_publish(exchange='test_exchange',
>>>> routing_key='test',
>>>> body="Hello
>>>> world!")
>>>> promises.append(promise)
>>>>
>>
>> Right, there is a performance problem with handling pubacks. The complexity
>> grows with n^2 of unacked messages
>>
>> Good news: the problem is in this line:
>> https://github.com/majek/puka/blob/master/puka/machine.py#L147-148
>>
>> Bad news: I have no clue how to fix it.
>>
>> I've rised an issue on github:
>> https://github.com/majek/puka/issues/12
>
> Christos,
>
> Thanks for your report and for the pull request!
>
> I just uploaded puka 0.0.4 to pypi, which solves the problem.
>
> Cheers,
> Marek
> _______________________________________________
> rabbitmq-discuss mailing list
> rabbitmq-discuss at lists.rabbitmq.com
> https://lists.rabbitmq.com/cgi-bin/mailman/listinfo/rabbitmq-discuss
--
Aaron Westendorf
Senior Software Engineer
Agora Games
359 Broadway
Troy, NY 12180
Phone: 518.268.1000
aaron at agoragames.com
www.agoragames.com
More information about the rabbitmq-discuss
mailing list