[ 
https://issues.apache.org/jira/browse/ARTEMIS-3759?focusedWorklogId=762889&page=com.atlassian.jira.plugin.system.issuetabpanels:worklog-tabpanel#worklog-762889
 ]

ASF GitHub Bot logged work on ARTEMIS-3759:
-------------------------------------------

                Author: ASF GitHub Bot
            Created on: 27/Apr/22 13:56
            Start Date: 27/Apr/22 13:56
    Worklog Time Spent: 10m 
      Work Description: gemmellr commented on code in PR #4054:
URL: https://github.com/apache/activemq-artemis/pull/4054#discussion_r859833745


##########
artemis-server/src/main/resources/schema/artemis-configuration.xsd:
##########
@@ -2442,6 +2442,13 @@
             </xsd:documentation>
          </xsd:annotation>
       </xsd:attribute>
+      <xsd:attribute name="address" type="xsd:string" use="optional">
+         <xsd:annotation>
+            <xsd:documentation>
+               name of the address this mirror connection applies to

Review Comment:
   I dont think the use case for this configuration is all that different, in 
this case its about what things you want to mirror from, vs the others being 
about what you want to send from or recieve to. An address 'match' for a mirror 
definition certainly doesnt need to mean 'create a seperate mirror for every 
address' just because it might mean that for e.g a sender. I also presume it 
couldnt really mean that given each mirror uses a specific queue named based on 
its parent broker-connections name.
   
   Using *address-match*, *address-pattern(s)*, *address-filter(s)*, or 
anything to similar effect, all seem equally reasonable and far more 
appropriate to me than just "address" does for a configuration that generally 
seems to be expected _not_ to just have a single address in it.
   
   I'm not saying the config has to be the same, but I am saying it seems more 
obvious, and I dont as yet see any compelling reason it should be entirely 
different when its basically for doing entirely equivalent things, configuring 
applicable addresses of interest.
   
   In any case how it is configured and what it does should be documented.





Issue Time Tracking
-------------------

    Worklog Id:     (was: 762889)
    Time Spent: 1h 10m  (was: 1h)

> Allow for Mirroring (Broker Connections) to specify a specific set of 
> addresses to send events, as is done for a cluster connection  
> -------------------------------------------------------------------------------------------------------------------------------------
>
>                 Key: ARTEMIS-3759
>                 URL: https://issues.apache.org/jira/browse/ARTEMIS-3759
>             Project: ActiveMQ Artemis
>          Issue Type: Improvement
>          Components: AMQP
>    Affects Versions: 2.19.1, 2.21.0
>            Reporter: Mikhail Lukyanov
>            Priority: Major
>         Attachments: ImageAddressSyntax.png, ImageInternalQueues.png
>
>          Time Spent: 1h 10m
>  Remaining Estimate: 0h
>
> If a target broker of mirroring is in a cluster, then mirroring of the 
> broker's internal queues also occurs, and often messages accumulate in such 
> queues. In theory, internal cluster queues should not be mirrored, this does 
> not make much sense. 
> Therefore, it would be convenient to allow you to configure for which 
> addresses and their queues mirroring will be performed, events will be sent 
> (message-acknowledgements, queue-removal, queue-creation). The syntax that is 
> used to specify the *_addresses_* of the *_cluster connection_* is well 
> suited for this. 
> Mirrored internal cluster queues
>  !ImageInternalQueues.png! 
> Address syntax
>  !ImageAddressSyntax.png! 



--
This message was sent by Atlassian Jira
(v8.20.7#820007)

Reply via email to