[ https://issues.apache.org/activemq/browse/AMQ-1301?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel#action_40499 ]
marcusn edited comment on AMQ-1301 at 10/23/07 11:41 AM: ---------------------------------------------------------------- Here is a theory. If this issue is the same as AMQ-961 then it could be that the following happens: * If there are persistent messages, maybe it takes some time for one of the brokers to start, making the transport interrupt. * If there are no persistent messages, both network connectors start reliably. In any case, the fix to AMQ-961 (applied a bit after 4.1.1) will fix the problem, it will restart the bridge whenever the transport resumes. was (Author: marcusn): Here is a theory. If this issue is the same as https://issues.apache.org/activemq/browse/AMQ-961 then it could be that the following happens: * If there are persistent messages, maybe it takes some time for one of the brokers to start, making the transport interrupt. * If there are no persistent messages, both network connectors start reliably. In any case, the fix to https://issues.apache.org/activemq/browse/AMQ-961 (applied a bit after 4.1.1) will fix the problem, it will restart the bridge whenever the transport resumes. > Problems with reconnecting broker to network of brokers after broker shutdown > ----------------------------------------------------------------------------- > > Key: AMQ-1301 > URL: https://issues.apache.org/activemq/browse/AMQ-1301 > Project: ActiveMQ > Issue Type: Bug > Components: Broker > Affects Versions: 4.1.2 > Reporter: Pieter > Fix For: 5.2.0 > > > When using a network of brokers, I get problems in the following scenario. > Consider two brokers, broker1 and broker2, connected to eachother (see config > below). Everything works ok: durable consumers on topics get messages sent to > either server. The messages are all persistent. > The problems occur in this scenario: > 1. Bring down one of the brokers, let's say broker2 (all consumers and > producers connect to the other server, without any problems) > 2. Start the killed broker2 again > 3. The broker2 reconnects to the network as expected > 4. However, messages don't get propagated correctly on the network anymore. > Consumers on broker1 don't get messages sent to broker2 anymore, and messages > from broker2 get sent to broker1 over and over again. This can also be seen > in the webconsole for example, where the number of messages in the queue > grows rapidly. > The only way to fix this is to kill all brokers, and remove the persistent > message stores. THis is unacceptable, since this loses messages. > Config is below: > <beans xmlns="http://www.springframework.org/schema/beans" > xmlns:util="http://www.springframework.org/schema/util"> > > <broker brokerName="broker1" useJmx="true" persistent="true" > xmlns="http://activemq.org/config/1.0"> > > <persistenceAdapter> > <journaledJDBC journalLogFiles="5" > dataDirectory="${activemq.base}/activemq-data"/> > </persistenceAdapter> > > <transportConnectors> > <transportConnector name="openwire" uri="tcp://localhost:61616"/> > <transportConnector name="stomp" uri="stomp://localhost:61627"/> > </transportConnectors> > > <networkConnectors> > <networkConnector name="ha-network" > uri="static://(tcp://localhost:61618)" networkTTL="5"/> > </networkConnectors> > </broker> > <beans xmlns="http://www.springframework.org/schema/beans" > xmlns:util="http://www.springframework.org/schema/util"> > > <broker brokerName="broker2" useJmx="true" persistent="true" > xmlns="http://activemq.org/config/1.0"> > > <persistenceAdapter> > <journaledJDBC journalLogFiles="5" > dataDirectory="${activemq.base}/activemq-data"/> > </persistenceAdapter> > > <transportConnectors> > <transportConnector name="openwire" uri="tcp://localhost:61618"/> > <transportConnector name="stomp" uri="stomp://localhost:61628"/> > </transportConnectors> > > <networkConnectors> > <networkConnector name="ha-network" > uri="static://(tcp://localhost:61616)" networkTTL="5"/> > </networkConnectors> > </broker> -- This message is automatically generated by JIRA. - You can reply to this email to add a comment to the issue online.