[rabbitmq-discuss] RabbitMQ receiving data, but no messages delivered

[Pedro Werneck]

We've been running RabbitMQ 3.0.2 with Erlang R14B04 for months, with no 
issues at all. Recently, our whole AWS infrastructure crashed and we had to 
rebuild some of it from scratch. Since then, we're having some strange 
problems with RabbitMQ. We have a main broker used by four applications, 
let's call them A, B, C and D, using Python 2.7 and Celery 3.0.18. 
Everything works fine when webservice instances are publishing messages and 
when worker instances are consuming messages, however, messages published 
by a worker instance of A and B never reach the queue. 

We can log the calls being made, we can sniff the network and see the 
packets leaving the worker instance, we can see the packets arriving at the 
RabbitMQ instance, and if we open the given Connection in the RabbitMQ 
management plugin we even see data arriving at a rate of 25kb/s, but no 
messages are delivered. The exchange and bindings are correct, since they 
work for other cases, and we can see the exchange and routing key are not 
corrupted or anything when sniffing the packets. If the message is sent by 
the exact same code and settings from a webservice instance or a 
shellscript, everything is fine, but when it's sent by the worker 
instance,  RabbitMQ seems to be ignoring it completely.

Any ideas on what's wrong?


-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.rabbitmq.com/pipermail/rabbitmq-discuss/attachments/20131223/d8b6b690/attachment.html>