activemq-users mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Gary Tully <gary.tu...@gmail.com>
Subject Re: conduit subscriptions vs full duplex on network connections
Date Tue, 11 Jan 2011 11:20:08 GMT
The single duplex network connector per broker on a transport
connectors is a side effect of
https://issues.apache.org/jira/browse/AMQ-2774 - an unintended
restriction I think. That needs to be redressed, using the
networkConnector name along with the brokerId or using just the
networkconnector name may be better solution to address unwanted
duplicates after a network partition.

As a work around, you can configure a second transport connector on
the hub, so it has two listener ports and configure each network
connector to use a different transport connector.

Can you raise a jira issue to track the additional restriction (one
duplex network connector per transport connector) introduced by
AMQ-2774

On 10 January 2011 22:30, adam <adam.sussman@gmail.com> wrote:
>
>
> The documentation on networks of brokers suggest that it is more ideal to
> have
> two separate network connections to a peer in order to have different
> "conduitSubscriptions"
> settings for Topics and for Queues.
>
> However, if you want a duplex connection to a peer broker (say for a
> hub&spoke architecture),
> activemq won't let you have more than one.
>
> So is there a safe way to get a duplex connection?  Can you have one
> connection be duplex
> and the other not?  If so, which one should be duplex?
>
> If you can't get one or the other (or both) of the connections to be duplex,
> is the only solution
> to have explicit network connections back from the other end?
>
> -adam
> --
> View this message in context: http://activemq.2283324.n4.nabble.com/conduit-subscriptions-vs-full-duplex-on-network-connections-tp3208138p3208138.html
> Sent from the ActiveMQ - User mailing list archive at Nabble.com.
>



-- 
http://blog.garytully.com
http://fusesource.com

Mime
View raw message