activemq-users mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From ttmdev <joe.fernan...@ttmsolutions.com>
Subject Re: JMS Bridge retries and transactions
Date Fri, 20 Jun 2008 18:53:36 GMT

A broker automatically retries connections to remote brokers. 

Clients will also automatically retry connections to brokers, but they need
to use the 'failover' transport.

Joe
www.ttmsolutions.com



dlaidlaw wrote:
> 
> 
> OK, so JMS Bridges are not the recommended solution. And I will look at
> camel and how robust I can make it. But in the mean time, what about a
> network of brokers?
> 
> The broker network stuff also moves messages between brokers. In my case
> they will all be ActiveMQ brokers so no issue about foreign broker
> connections. I could configure the network of brokers to only replicate
> certain queues in one direction.
> 
> But is this robust? Will it recover from connection drops when a remote
> broker is stopped and restarted? Or how about at local broker startup
> time, if the remote broker is not up at the time will the connection keep
> trying in the background, waiting for the remote broker to startup?
> 
> Or is this not a recommended solution as well?
> 
> Thanks for your time on this!
> -Don
> 
> 
> rajdavies wrote:
>> 
>> 
>> On 20 Jun 2008, at 16:25, dlaidlaw wrote:
>> 
>>>
>>> I was looking at the source code for JMS Bridges to see how robust the
>>> connections would be. Would they auto-recover, retry delivery, were  
>>> they
>>> transacted, etc.  I noticed some code in there that seemed to be  
>>> aimed at
>>> having retry behavior but no loop for a retry.
>>>
>>> Look at the org.apache.activemq.network.jms.DestinationBridge class,  
>>> the
>>> onMessage method. A local method variable called attempt is created  
>>> and
>>> initialized to 0. There is also a test against maximumRetries, but  
>>> there are
>>> no loops in the code to actually have another attempt. Any idea how  
>>> a second
>>> attempt could ever happen? I can't see it.
>>>
>>> Has anyone ever done an implementation using transactions to ensure  
>>> the
>>> consume/send pair commits as a unit? It would be possible here for  
>>> the send
>>> to succeed and the acknowledge to fail resulting in duplicates being  
>>> sent
>>> eventually.
>>>
>>> Thanks
>>> -Don
>>> -- 
>>> View this message in context:
>>> http://www.nabble.com/JMS-Bridge-retries-and-transactions-tp18031980p18031980.html
>>> Sent from the ActiveMQ - User mailing list archive at Nabble.com.
>>>
>> 
>> Hi Don  - the recommended approach for JMS bridges is to now use Camel  
>> - see http://activemq.apache.org/camel/jms.html
>> 
>> 
>> 
>> 
>> cheers,
>> 
>> Rob
>> 
>> http://open.iona.com/products/enterprise-activemq
>> http://rajdavies.blogspot.com/
>> 
>> 
>> 
>> 
>> 
>> 
> 
> 

-- 
View this message in context: http://www.nabble.com/JMS-Bridge-retries-and-transactions-tp18031980p18035458.html
Sent from the ActiveMQ - User mailing list archive at Nabble.com.


Mime
View raw message