[ 
https://issues.apache.org/jira/browse/ARTEMIS-4196?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17696888#comment-17696888
 ] 

Gary Tully commented on ARTEMIS-4196:
-------------------------------------

interesting observation, the routing type on a message serves to split the 
address model into any cast or multicast, to support the use case where there 
are jms queues and topics with the same name. The jms destination on the 
producer being used to set the message routing type, and the post office 
choosing either the anycast or multicast bindings collection (the logic is not 
that clean cut, with queue bindings rejecting multicast messages reality). 
Without it it, there is pollution b/c there is only a single address name space 
in Artemis. in 5.x they are separate.

I wonder if the clashing of names between queues and topics is present in the 
real world?

use case at: 
[https://github.com/apache/activemq-artemis/commit/279383a7985dd9b92ab89584bf7eda7029343be8#diff-3eaaeaac1574c2130a6cd5a5e924048e307addd5c1dafce87cf51c13d0c53cf8R735]

in the absence of a routing type, all bindings seem to get the messages.

> MQTT cluster message distribution is broken with OFF and 
> OFF_WITH_REDISTRIBUTION loadbalancing types
> ----------------------------------------------------------------------------------------------------
>
>                 Key: ARTEMIS-4196
>                 URL: https://issues.apache.org/jira/browse/ARTEMIS-4196
>             Project: ActiveMQ Artemis
>          Issue Type: Bug
>          Components: clustering, MQTT
>            Reporter: Gary Tully
>            Assignee: Gary Tully
>            Priority: Major
>
> The issue relates to the lack of a routingType on the messages from the mqtt 
> publish in the mqtt protocol head.
> With OFF or OFF_WITH_REDISTRIBUTION we assert that the routing type is 
> multicast to allow clustering, but without a routing type on the message that 
> fails and we don't forward to any remote bindings in error.



--
This message was sent by Atlassian Jira
(v8.20.10#820010)

Reply via email to