[rabbitmq-discuss] Odd behavoiur with mcollective and RabbitMQ 3.1.3

Niclas Meier niclas at herolabs.com
Fri Jul 19 08:09:58 BST 2013

Hello everyone,

we are using mcollective and a RabbitMQ (and the STOMP connector) to manage 
our servers. Unfortunately we updated to RabbitMQ 3.1.3 and now we have a 
strange behaviour. When I ping a server with mco ping I only get a response 
from a server only once. When I restart mcollective on the server I will 
get an answer the next time and when I restart the MQ all server will 
answer one time.

I checked the MQ with the management console and I see a STOMP connection 
for the mcollective server and every time I use mco ping a new STOMP 
connection appears. The STOMP connection created with the ping seems to 
stay, even when the ruby process is already terminated. 

Fortunately I found an "old" 3.0.1 Rabbit MQ installation on one of our 
servers and on this version mcollecive works like before (the setup works 
unchanged since January 2013). The setup consists of:

   - Ubuntu 12.04 LTS
   - mcollective 2.2.4
   - Stomp Gem 1.2.10 (tested 1.2.9 and 1.2.8)
   - Rabbit MQ 3.1.3 with the embedded Rabbit Stomp Adapter (works on 3.0.1)
   - Ruby 1.8.7

Additionally I have stuff like this in the sasl log:

=SUPERVISOR REPORT==== 19-Jul-2013::07:57:23 ===
     Supervisor: {<0.570.0>,rabbit_stomp_client_sup}
     Context:    shutdown
     Reason:     reached_max_restart_intensity
     Offender:   [{pid,<0.572.0>},
                                <<" ->">>,

Any Ideas?


-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.rabbitmq.com/pipermail/rabbitmq-discuss/attachments/20130719/f930e78a/attachment.htm>

More information about the rabbitmq-discuss mailing list