You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@activemq.apache.org by "Justin Bertram (Jira)" <ji...@apache.org> on 2021/04/27 19:31:00 UTC

[jira] [Commented] (ARTEMIS-3268) Make Duplicate Detection Optional on all protocols

    [ https://issues.apache.org/jira/browse/ARTEMIS-3268?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17333506#comment-17333506 ] 

Justin Bertram commented on ARTEMIS-3268:
-----------------------------------------

It's often helpful to implement duplicate detection on the client-side, but of course that requires extra work on whoever is implementing the client. Most clients do not, in fact, implement duplicate detection which then pushes the requirement onto application developers who may not have the resources for such work. It makes sense to consolidate this work on the broker, but of course flexibility is great so the ability to disable it selectively is worth implementing.

It's worth noting that the broker itself uses duplicate detection on cluster bridges to implement high-performance, asynchronous once-and-only-once delivery without XA so care should be taken *not* to disable duplicate detection on the store-and-forward resources.

> Make Duplicate Detection Optional on all protocols
> --------------------------------------------------
>
>                 Key: ARTEMIS-3268
>                 URL: https://issues.apache.org/jira/browse/ARTEMIS-3268
>             Project: ActiveMQ Artemis
>          Issue Type: Improvement
>          Components: Configuration
>    Affects Versions: 2.17.0
>            Reporter: Gary Tully
>            Priority: Major
>
> In ARTEMIS-2581 support for server/broker side duplication detection was made optional. It makes sense b/c there is an overhead and thread contention issue with detection at the address level. It would be nice to be able to disable this feature, independent of the presence of a duplicate id header in the message, which is optionally placed by a message producer.
> duplicate detection is often best implemented at the application layer which negates the need for the server/broker to do any work.



--
This message was sent by Atlassian Jira
(v8.3.4#803005)