activemq-users mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Christian Posta <christian.po...@gmail.com>
Subject Re: Networked brokers dropping/suppressing messages
Date Fri, 24 May 2013 17:08:17 GMT
don't use the failover directly like that... use masterslave: as discussed
at Torsten's excellent blog entry:

http://tmielke.blogspot.com/2011/09/activemq-network-bridge-to-masterslave.html


On Fri, May 24, 2013 at 8:15 AM, talavera <atalavera@gmail.com> wrote:

> Did you find a solution?
>
> I've just run into the problem too, this what I get:
> 2013-05-23 17:47:04,927 | DEBUG | suppressing duplicate message send
> [ID:corerec2-53231-1368512827283-0:385:1:1:1] with producerSequenceId [1]
> less than last stored: 9 |
> org.apache.activemq.broker.ProducerBrokerExchange
> | ActiveMQ Transport: tcp:///10.25.144.85:41356@61616
>
> Looking into the code [1], it seems this particular log indicates the
> connection from which is received belongs to a local producer, but I've
> tracked the message ID back to a different broker, which says it's bridging
> the message.
> The log prints a transport originating in that remote broker, anyways.
>
> I wonder if this shouldn't be recognized as a networkProducer and then
> "demultiplex" the last stored Id, instead of using a unique index for the
> entire network connection (I'm using conduit subscriptions, the default
> config)
>
> This behaviour seems to happen after a network connection is broken (a
> firewall deciding it's been too long) and the originating broker restarts
> it.
> I haven't found anything interesting during the connection restore phase in
> logs.
> Only restarting the destination broker helps after the broker starts
> marking
> messages as duplicated.
>
> I'll try to get a network capture of the restablishment of the network
> connection, I'd like to see if there's a BrokerInfo command coming to the
> destination broker, which seems to be what helps identifying a network
> connection in destination.
> Not sure of this one though, any hints would be helpful.
>
> I'm running AMQ 5.7.0, using two standalone brokers A and B, with a network
> connection A->B which has a failover to a third C broker, which is a slave
> of A.
> Broker C have been down during the tests, only the network connection uses
> the failover protocol.
>
> This is the config in broker A:
>          <networkConnectors>
>             <networkConnector
> uri="static:(failover:(tcp://B:61616,tcp://C:61617)?randomize=false)"/>
>         </networkConnectors>
>
> [1]
>
> https://svn.apache.org/viewvc/activemq/trunk/activemq-broker/src/main/java/org/apache/activemq/broker/ProducerBrokerExchange.java?revision=1406629&view=markup
>
>
>
> --
> View this message in context:
> http://activemq.2283324.n4.nabble.com/Networked-brokers-dropping-suppressing-messages-tp4662747p4667469.html
> Sent from the ActiveMQ - User mailing list archive at Nabble.com.
>



-- 
*Christian Posta*
http://www.christianposta.com/blog
twitter: @christianposta

Mime
  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message