On Sun, Mar 7, 2010 at 12:34 PM, Tracy Snell <tsn...@gmail.com> wrote:
>
> On Mar 7, 2010, at 1:00 PM, Bruce Snyder wrote:
>
>> Yes, this is Rob's recommendation. Below is a depiction of how it should
>> work:
>>
>> brokerA-old --(network connection)--> brokerB-new
>>
>> * brokerA-old is configured to use the AMQ store
>> * brokerB-new is configured to use the KahaDB store
>>
>> Message consumers connect to brokerB-new to consume messages from a
>> destination that already exists on brokerA-old. When messages are
>> forwarded from brokerA-old to brokerB-new, they are extracted from
>> brokerA-old's AMQ store and inserted into brokerB-new's KahaDB store.
>
> Won't the messages stay on BrokerA until a consumer on BrokerB asks for them?
> Then wouldn't it be pretty transient. The message would be removed as it was
> consumed right?
That's exactly what I said above, 'Message consumers connect to
brokerB-new to consume messages'. When demand exists on brokerB for a
destination that already exists and has persisted messages on brokerA
(and the two brokers are networked), then the messages will be
forwarded appropriately from brokerA to brokerB and on to the
consumer.
> Could BrokerB could be setup to consume the messages and store them local
> using <jmsTopicConnector> or <jmsQueueConnector>? That would pull them local
> to BrokerB and stored in KahaDB. Can the same thing be achieved with a
> network connection?
Anything that creates a JMS consumer subscription (and. therefore,
creates demand for the messages) will cause them to be forwarded.
E.g., a pure JMS consumer, a Camel JMS consumer, a JMS-to-JMS bridge,
etc.
Bruce
--
perl -e 'print
unpack("u30","D0G)u8...@4vyy9&5R\"F)R=6-E+G-N>61E<D\!G;6%I;\"YC;VT*"
);'
ActiveMQ in Action: http://bit.ly/2je6cQ
Blog: http://bruceblog.org/
Twitter: http://twitter.com/brucesnyder
