activemq-users mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Matt Pavlovich <mattr...@gmail.com>
Subject Re: Activemq Clustering and failover
Date Thu, 02 Feb 2012 21:17:50 GMT
Abimael-

It might be helpful to understand that the client-side fail-over is 
handled by the ActiveMQ client-side library.  The end-user just needs to 
code the consumer program as if it was talking to one broker.

I suggest adding DEBUG, or TRACE level logging on the client-side.  You 
may find more useful information there.

Can you paste your ActiveMQ config and client-side URL's to 
http://pastebin.com, or other and share the link?  After reviewing the 
config, we may be able to help identify the root cause.

Thanks,
Matt Pavlovich

On 2/2/12 1:46 PM, Abimael wrote:
> Oh no! No it does not work.. I did another test to check  and it does not
> work even with Durable SUbscriber .
> Scenario :
> 2 brokers>: master and slave
> 1 producer  with failover uri on the code to connect and provide messages;
> 1 consumer   with failover uri on the code to connect and consume messages;
>
> When master went down, slave got control.
> I am able to see that producer got a "little pause", then a message appeared
> :
> 18243 [ActiveMQ Task-2] INFO
> org.apache.activemq.transport.failover.FailoverTransport - Successfully
> reconnected to tcp://sim1660:61616
> (sim1660 is the slave)
>
> However, the consumer had FINISHED .
> I know that if I restart consumer it will be able to get the old messages
> (as I am using durable topic)  and continue with the new BUT, I do not know
> why it went down .
>
> Regards
>
>
>
> --
> View this message in context: http://activemq.2283324.n4.nabble.com/Activemq-Clustering-and-failover-tp4326849p4352571.html
> Sent from the ActiveMQ - User mailing list archive at Nabble.com.

Mime
View raw message