[rabbitmq-discuss] Memory leak in STOMP connector or federation?

[Tomas Doran]

On May 17, 2014, at 5:28 AM, Michael Klishin <mklishin at gopivotal.com> wrote:

> On 17 May 2014 at 02:55:13, Tomas Doran (bobtfish at bobtfish.net) wrote:
>>> However I’m guessing not, as no re queues - it’s all mcollective,  
>> so there is a persistent server process that listens to client  
>> messages:
>> 
>> all messages to a ‘broadcast’ queue
>> messages with a specific routing key to a ‘direct’ queue
>> 
>> and replies to a specific ‘replies’ queue that the client uses  
>> it’s name as a routing key on.
> 
> Tomas,
> 
> We are not mcollective experts.

Ack - thanks for taking the time to look at this given my non-specific problem description!

> Can you please post a code snippet
> (in any language) that demonstrates what exactly mcollective does for request/reply
> over STOMP?
> 
> This would be a great way for us to actually reproduce the problem.


This is not completely trivial to do (a snippet), as I’m not sure if it’s the server to client or client to server messaging which is causing the leak (or something else, like reconnections).

However, the source for the connector I’m using is:

https://github.com/puppetlabs/marionette-collective/blob/master/plugins/mcollective/connector/rabbitmq.rb

Every one of the messaging patterns in there gets used in our setup, except that notably, we are setting the ‘rabbitmq.use_reply_exchange’ setting (https://github.com/puppetlabs/marionette-collective/blob/master/plugins/mcollective/connector/rabbitmq.rb#L335) to true.

I think that the default pattern  is to use temporary queues for server => client responses, which would cause the effect you’re describing - however with rabbitmq.use_reply_exchange = true, my reply paths are /exchange/mcollective_reply/…, which wouldn’t suffer from the problem mentioned in this thread?

I’m trying to work up a Vagrant box which reproduces my setup faithfully enough to replicate the leak I’m seeing - I’ll share my results as soon as I can get there, but whilst I have a very similar setup running, I haven’t (yet) successfully been able to reproduce the leak.

Cheers
Tom