I think that is the correct approach. Compare
FanoutTransportFactory.createTransport to some of the other transport
factories, FailoverTransportFactory for example, it follows that
pattern.
Possibly open a jira issue and attach your patch and a test case to protect it.

On 22 July 2010 17:51, nondescript1 <jmsig...@gmail.com> wrote:
>
> Greetings.  This is my first post, so let me say thank you up front for a
> great tool in ActiveMQ and for any help you provide.
>
> Is there a correct way to apply transport parameters to discovered brokers?
>
> I'm trying to send a message using the fanout transport and the following
> URI,
>    fanout:(multicast://default?maxReconnectAttempts=1)?minAckCount=1
> As you can see, I'm using multicast broker discovery.  However, in certain
> circumstances, I encounter connection timeout issues when sending the
> message after the brokers are discovered.  In these cases, the default
> connectionTimeout of 30 seconds causes a considerable delay in my
> application, and I would rather have a 3 second delay.  Therefore, I would
> like to apply the parameter connectionTimeout=3000 to any discovered
> brokers.
>
> A colleague was able to accomplish this by modifying
> FanoutTransportFactory.createTransport(URI) to call
> transport.setParameters(parameters) on the DiscoveryTransport that the
> FanoutTransport is wrapped in.  Since DiscoveryTransport parameters are
> applied to discovered services in
> DiscoveryTransport.onServiceAdd(DiscoveryEvent), the following URI now has
> the desired result of timing out after three seconds of not being able to
> send a message to a discovered broker,
>
> fanout:(multicast://default?maxReconnectAttempts=1)?minAckCount=1&connectionTimeout=3000
>
> However, this solution feels like a hack.  Do you agree that this is not a
> correct approach, and is there a best practice for doing this?
>
> Thank you again for your time and help.
> --
> View this message in context: 
> http://old.nabble.com/Applying-Parameters-to-Discovered-Brokers-tp29239157p29239157.html
> Sent from the ActiveMQ - User mailing list archive at Nabble.com.
>
>



-- 
http://blog.garytully.com

Open Source Integration
http://fusesource.com

Reply via email to