You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@activemq.apache.org by "Jarek Przygódzki (JIRA)" <ji...@apache.org> on 2018/02/05 19:45:00 UTC
[jira] [Comment Edited] (ARTEMIS-1648) Messages from 1.x clients to
2.x cluster are lost when they are load balanced to nodes with matching
consumers
[ https://issues.apache.org/jira/browse/ARTEMIS-1648?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16352828#comment-16352828 ]
Jarek Przygódzki edited comment on ARTEMIS-1648 at 2/5/18 7:44 PM:
-------------------------------------------------------------------
Not really. What happens in current master (19147e49869d23906baf2cc07d98564043b3b0e0) is really messed up - consumer calling {{ActiveMQJMSContext#createQueue}} creates a queue with prefixed name and that messages from 'myQueue' on source node are forwarded to 'jms.queue.myQueue' on node with active consumer where two different queue exists - 'jms.queue.myQueue' and 'myQueue'. Message are no longer lost, but are forwarded to wrong queue!
was (Author: jarek-przygodzki):
Not really. What happens in current master (19147e49869d23906baf2cc07d98564043b3b0e0) is really messed up - consumer calling {{ActiveMQJMSContext#createQueue}} creates a queue with prefixed name and that messages from 'myQueue' on source node are forwarded to 'jms.queue.myQueue' on node with active consumer where two different queue exists - 'jms.queue.myQueue' and 'myQueue'.
> Messages from 1.x clients to 2.x cluster are lost when they are load balanced to nodes with matching consumers
> --------------------------------------------------------------------------------------------------------------
>
> Key: ARTEMIS-1648
> URL: https://issues.apache.org/jira/browse/ARTEMIS-1648
> Project: ActiveMQ Artemis
> Issue Type: Bug
> Components: Broker
> Affects Versions: 2.3.0
> Reporter: Jarek Przygódzki
> Priority: Major
>
> When a message from 1.x client is transfered to node in the cluster with matching consumers (message-load-balancing=ON_DEMAND), it's moved to appropriate replication queue, forwarded to target node and then subsequently dropped because there aren't any matching bindings for message address. This is because if the original message is sent to address with the jms.queue prefix (1.x clients), than the message sent by ClusterConnectionBridge also has the prefixed destination address and this situation is not supported.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)