[ 
https://issues.apache.org/jira/browse/ARTEMIS-3090?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17279996#comment-17279996
 ] 

Justin Bertram commented on ARTEMIS-3090:
-----------------------------------------

Do you have a test-case which reproduces what you've described? I ask because 
I've tried to reproduce this and I can't. 

If you want to "unset" the {{expiry-address}} for a particular address you 
should be able to use {{<expiry-address/>}}. I wrote a quick test of this 
scenario and everything worked as expected.

Also, there is already [logic in 
place|https://github.com/apache/activemq-artemis/blob/2.16.0/artemis-server/src/main/java/org/apache/activemq/artemis/core/server/impl/QueueImpl.java#L2396]
 to prevent expired messages from going back to the configured 
{{expiry-address}}.

> expiry-queue cannot be unset
> ----------------------------
>
>                 Key: ARTEMIS-3090
>                 URL: https://issues.apache.org/jira/browse/ARTEMIS-3090
>             Project: ActiveMQ Artemis
>          Issue Type: Bug
>          Components: Broker
>    Affects Versions: 2.16.0
>            Reporter: Erwin Dondorp
>            Priority: Major
>
> By default, expiry-queue is set to "ExpiryQueue".
> But it is not possible to replace the value with "xsd:nil" since the 
> expiry-queue setting is not nillable as specified in the corresponding XSD.
> However, 
> src/main/java/org/apache/activemq/artemis/core/server/impl/QueueImpl.java 
> implements a use-case to drop the message when no expiry-queue is set.
> This use-case is hard to reach as one must then not set the expiry-queue for 
> the default address (#) and then set the expiry-queue for all other addresses.
> I have 2 use-cases for which not expiry-queue should be set. the first is for 
> destinations that carry low-value data, and second for the ExpiryQueue 
> itself, to make sure that messages will eventually be discarded. The default 
> ExpiryQueue cannot expire messages since its expiry-queue points to itself.
> I suggest that expiry-queue is made nillable, so that it can simply be 
> applied to individual addresses.



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

Reply via email to