On 09/12/2013 10:14 AM, Christian Fromme wrote:
> Hi Gordon,
>
> thanks for your quick reply.
>
> On Wed, Sep 11, 2013 at 5:33 PM, Gordon Sim <gsim@redhat.com> wrote:
>
>>> qmf.default.topic topic 1 44.3k 153 44.1k 0
>>> qpid.management topic 0 44.1k 0 44.1k 0
>>>
>> No, its not the heartbeats, it looks like it is from the qmf.default.topic
>> and qpid.management exchanges which each have dropped 44k messages. Using
>> qpid-stat itself will generate quite a few messages. Also I suspect the high
>> number of drops may be events that are generated but have no-one interested
>> in receiving them.
>
> The increasing message count in "total-enqueues" and "total-dequeues"
> might be caused by qpid-stat then, I guess.
>
> No I am wondering about the high message dropping in qmf.default.topic
> and qpid.management. As far as this application is concerned, we only
> send messages to amq.topic exchanges (between several qpid deamons
> across hosts using routes). Except for application startup or fail
> over, where we create routes in between the qpid daemons via commands
> to qmf.default.direct/broker. Could that be the cause?
>
> It bugs me that we seem to have something in our app constantly
> sending messages, and I don't know what it is. ;-)
I think the drops are caused by management events. The broker will
generate events and route them to the management exchanges. However if
there is no-one subscribed to receive those events, then they will just
be dropped.
Using qpid-stat will itself trigger the raising of events on the broker
(connection opened and closed, queues created and destroyed etc).
Furthermore I believe there is some periodic 'publishing' of current
object stats which again may well just be dropped.
Ideally these stats related to management itself would be hidden by
default of course.
---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe@qpid.apache.org
For additional commands, e-mail: users-help@qpid.apache.org
|