[rabbitmq-discuss] high number of unack'ed messages when queue grows large

Wed May 12 10:18:18 BST 2010

mmenti at gmail.com wrote:
> On Fri, May 7, 2010 at 3:05 PM, Matthew Sackman <matthew at lshift.net 
> <mailto:matthew at lshift.net>> wrote:
> 
>     On Fri, May 07, 2010 at 02:59:26PM +0100, mmenti at gmail.com
>     <mailto:mmenti at gmail.com> wrote:
>      > Strangely we already have prefetch set to 1 on the consumers, and
>     still see
>      > this issue.
> 
>     That's bizarre. Are you doing some gets as well? They're not counted in
>     the qos setting. Also, which client are you using?
> 
> 
> We're using MQ, from the tmm1-amqp gem, and no, not doing any gets..

Can you send us the consumer code?

Alternatively, please connect one consumer via the tracer 
(http://www.rabbitmq.com/examples.html#tracer) - preferably when there 
are only a few messages in the queue - and send us the output.

Regards,

Matthias.