[rabbitmq-discuss] Troubleshooting handleCancel()
srinath.c at gmail.com
Sat Apr 12 16:25:16 BST 2014
Just checking if the logs provide any insights into the probable causes?
On Wed, Apr 9, 2014 at 8:34 PM, Srinath C <srinath.c at gmail.com> wrote:
> Hi MK,
> Thanks for your inputs. Below are my replies to your queries:
> a, I can say for sure that there is part of this deployment which attempts
> to explicit delete a queue or an exchange.
> b, Queue master node failure could be a possibility. But I'm not sure if
> the failed node was the master. How do you find it out by the way?
> c, I have attached the rabbitmq logs for the surviving node. I have
> cropped the logs to a few minutes before and after the handleCancel was
> d, I can say for sure that publisher was connected to the surviving node.
> The consumer - I believe was also connected to the surviving node because
> of the fact that lyra would have otherwise recovered the queue and bindings.
> e, Sure, will try out 3.3 to see if this issue occurs again.
> Thanks again,
> On Wed, Apr 9, 2014 at 1:31 PM, Michael Klishin <mklishin at gopivotal.com>wrote:
>> On 9 April 2014 at 05:36:42, Srinath C (srinath.c at gmail.com) wrote:
>> > > Scenario:
>> > One of the rabbitmq nodes were brought down for testing.
>> > It was observed that no more messages were being consumed
>> Note that RabbitMQ 3.3.0 will automatically re-register consumers
>> on a queue that had master failure. So I suggest giving your scenario
>> a try with 3.3.
>> Software Engineer, Pivotal/RabbitMQ
>> rabbitmq-discuss mailing list
>> rabbitmq-discuss at lists.rabbitmq.com
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the rabbitmq-discuss