[rabbitmq-discuss] Per Connection Flow Control Issue

Brendan Hay brendan at soundcloud.com
Thu Oct 25 15:41:36 BST 2012


Simon: list_connections snippet here -
https://gist.github.com/8056314cece0d44ab21c
Not every channel belonging to those connections marked by 'flow' is
'idle', a seemingly
random distribution are 'active' also, but with no ingress bytes

Michael: No way to reproduce outside of the production environment yet,
will reply
with an example when I can isolate it.

On Thu, Oct 25, 2012 at 4:27 PM, Michael Klishin <
michael.s.klishin at gmail.com> wrote:

> 2012/10/25 Brendan Hay <brendan at soundcloud.com>
>
>> On the client machines, netstat shows the peer sockets to have huge (and
>> growing) send queues.
>>
>> Does anybody have any experience with similar problems? I am yet to catch
>> it in the act,
>> but it appears flow control must kick in at some point and the connection
>> simply does not
>> resume reading the pending bytes off the socket after it is issued new
>> credits.
>>
>> Posted this here as I'm not sure who/what to point the proverbial finger
>> at, suggestions appreciated.
>>
>
> Brendan,
>
> Because amqp gem does not do any implicit or automatic flow control, I'd
> expect the root cause
> to be in EventMachine. Do you have a way to reproduce this issue?
> --
> MK
>
> http://github.com/michaelklishin
> http://twitter.com/michaelklishin
>
>
> _______________________________________________
> rabbitmq-discuss mailing list
> 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/20121025/943d2f09/attachment.htm>


More information about the rabbitmq-discuss mailing list