[rabbitmq-discuss] Troubleshooting handleCancel()

Wed Apr 9 02:33:58 BST 2014

Hi,
    I'm looking for help in troubleshooting a scenario where my consumers
stopped processing messages from the queue they were listening on.

*Setup*:
A 2 node RabbitMQ cluster running 3.2.4. Both are disc nodes.
A topic E (durable=true, auto-delete=false)
A queue Q (auto-delete=true, exclusive=false, durable=false)
Q binds to E using binding key as #
C the creator & consumer of Q is running on a JVM using Lyra to handle
recovery of connections.

*Scenario*:
One of the rabbitmq nodes were brought down for testing.
It was observed that no more messages were being consumed from the exchange.
(This is not always reproducible using the same steps)

*Observations*:
Q existed on the remaining rabbitmq node
Q was not bound to E (hence no message were consumed)

*Question*:
Since the consumer C is using Lyra, it would have re-created the queue and
bindings to exchanges in case of connection failures (it has been
configured to do so). This has been tested several times and it works.

But in this particular case, there were no logs from Lyra indicating a
connection failure and recovery. Indicating that the consumer was connected
to the node that was not brought down.

The only details in the consumer logs were a "handleCancel" called on the
consumer. What are the scenarios where this is called? I can simulate this
scenario by explicitly logging into the console and deleting the queue or
the exchange....

Thanks in advance,
Srinath.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.rabbitmq.com/pipermail/rabbitmq-discuss/attachments/20140409/fcbd8ba9/attachment.html>