[
https://issues.apache.org/jira/browse/ARTEMIS-967?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Justin Bertram resolved ARTEMIS-967.
------------------------------------
Resolution: Duplicate
> Graceful shutdown for core bridge
> ---------------------------------
>
> Key: ARTEMIS-967
> URL: https://issues.apache.org/jira/browse/ARTEMIS-967
> Project: ActiveMQ Artemis
> Issue Type: Improvement
> Components: Broker
> Affects Versions: 1.5.2, 2.0.0
> Reporter: Miroslav Novak
> Priority: Major
>
> Scenario:
> 1) Start 2 servers with configured core bridge from queue sourceQueue
> (server1) to queue targetQueue (server2)
> 2) Send 10 large messages to sourceQueue (to server1)
> 3) Receive 10 messages from targetQueue (from server2) a commit them.
> 4) Immediately cleanly shutdown server1
> Result:
> If step 4. (shutdown of server1) was executed fast enough then there are
> still messages in sourceQueue.
> Investigation showed that messages were not acknowledged by bridge on server1
> in time. From trace logs of server1, there is call
> BridgeImpl#sendAcknowledged(message) which should ack the message in the
> journal. This means that server2 send acknowledge that message was
> successfully received. But because server1 is shutting down, bridge is marked
> as not "active" and code which acknowledges the message is skipped. At the
> same time server2 thinks that messages were acknowledge and allows
> consumption of those messages from server2. Result of this is that after
> restart of server1, those message can be consumed from server1 by any
> receiver or resend to by core bridge server2.
> The variable "active" was switched to false by:
> {code}
> 15:36:40,062 main INFO
> [org.apache.activemq.artemis.core.server.cluster.impl.BridgeImpl:374] mnovak
> - stop on bridge was called from:
> java.lang.Thread.getStackTrace(Thread.java:1552)
> org.apache.activemq.artemis.core.server.cluster.impl.BridgeImpl.stop(BridgeImpl.java:369)
> org.apache.activemq.artemis.core.server.cluster.ClusterManager.stop(ClusterManager.java:323)
> org.apache.activemq.artemis.core.server.impl.ActiveMQServerImpl.stopComponent(ActiveMQServerImpl.java:1115)
> org.apache.activemq.artemis.core.server.impl.ActiveMQServerImpl.stop(ActiveMQServerImpl.java:863)
> org.apache.activemq.artemis.core.server.impl.ActiveMQServerImpl.stop(ActiveMQServerImpl.java:813)
> org.apache.activemq.artemis.core.server.impl.ActiveMQServerImpl.internalStop(ActiveMQServerImpl.java:680)
> org.apache.activemq.artemis.core.server.impl.ActiveMQServerImpl.stop(ActiveMQServerImpl.java:673)
> {code}
> The idea is to allow to finish all deliveries in progress and stop all new
> delivery calls by bridge when BridgeImpl.stop() is called.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
For further information, visit: https://activemq.apache.org/contact
