Persistent increase of rabbit queues

asked 2016-08-02 07:49:59 -0500

DanielK gravatar image


We are running openstack liberty on CentOS 7 with RabbitMQ 3.6.2 and Erlang 18.3.3. Over time we can observe a persistent increase of rabbit queues that never go away. We have durable queues set to false and auto_delete to true, and all the queues in question have a consumer attached, but they often remain idle for days after their initial use for data exchange once they get created.

Restarting individual openstack components like neutron, nova, cinder or heat will get rid of queues used by those services, and restarting rabbitmq itself starts the queue numbers from zero. They then increase rapidly over a short time, and then the amount of queues goes into a slow but continuous path upwards.

I am worried that this can eventually cause resource problems with the rabbitmq server, and I was wondering whether this is working as intended or whether there is a possibility to influence this behaviour.

Is tere a way to do garbage collection on idle queues which have a consumer attached? All I found so far in the rabbitmq documentation is only about queues with no consumer.

Thank you in advance.

edit retag flag offensive close merge delete


As a matter of fact, I have seen today that the queues begin to affect the rabbitmq management dashboard itself. At almost 20K active queues it began to complain about too many database entries to process and warned about increasing memory load when doing so.

DanielK gravatar imageDanielK ( 2016-08-08 08:40:27 -0500 )edit

1 answer

Sort by ยป oldest newest most voted

answered 2016-08-03 03:55:43 -0500

yprokule gravatar image

I can just agree with your findings. U cannot remove idle queue having consumer attached, without affecting the consumer.

edit flag offensive delete link more

Get to know Ask OpenStack

Resources for moderators

Question Tools



Asked: 2016-08-02 07:49:59 -0500

Seen: 127 times

Last updated: Aug 03 '16