activemq-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Rob Davies (JIRA)" <>
Subject [jira] Assigned: (AMQ-1661) Duplicate messages and extra network traffic with duplex connections
Date Mon, 14 Apr 2008 10:39:43 GMT


Rob Davies reassigned AMQ-1661:

    Assignee: Rob Davies

> Duplicate messages and extra network traffic with duplex connections
> --------------------------------------------------------------------
>                 Key: AMQ-1661
>                 URL:
>             Project: ActiveMQ
>          Issue Type: Bug
>          Components: Broker, Connector
>    Affects Versions: 5.1.0
>            Reporter: Mike Forrest
>            Assignee: Rob Davies
>         Attachments: duplex.png
> (see attached diagram)
> I have two brokers A and B, where A establishes a connection to B with the options: 
> duplex="true" networkTTL="3" dynamicOnly="true" 
> I have two clients connecting to broker A, and one to B.  One of the clients on A is
subscribing to a queue, the other is publishing to it.  The client on B is not using that
queue at all.  When the publishing client sends a message, the subscriber receives it as expected.
 But if I watch the network with a sniffer, I can see the message being sent to broker B and
back again. 
> I also set up a topic that all of the clients subscribe to.  If a client on A sends a
message to the topic, the client on B receives it normally, but the two clients on A receive
the message twice.  I can see with the sniffer that the message is being sent from A to B
and back. 
> If I change the connection so that it's not duplex (i.e. I set up matching simplex connections
on both brokers) all of this unexpected behavior goes away.  In the first example, the messages
are no longer forwarded to B since there are no subscribers on that broker.  In the second
example, the message is received once by every subscriber. 
> I have tested on various 5.1 snapshots up through 4/8/08.

This message is automatically generated by JIRA.
You can reply to this email to add a comment to the issue online.

View raw message