Timothy Bish updated ARTEMIS-1297:
    Comment: was deleted

(was: While it would be nice to avoid decoding the full message one bit that 
makes this a bit hard is that the server will peek into message annotations for 
each message to check not only for the routing type but also to see if a 
message has a delivery delay specific.  You can avoid the routing type lookup 
by figuring that out ahead of time on the inbound link but the delay bits are 
hard to avoid.  You might be able to avoid that if you knew the delay feature 
as disable as it could be in 5.x but otherwise I'm not sure how we'd factor 
that out.  )

> AMQPMessages cannot be load balanced across a Message Bridge
> ------------------------------------------------------------
>                 Key: ARTEMIS-1297
>                 URL: https://issues.apache.org/jira/browse/ARTEMIS-1297
>             Project: ActiveMQ Artemis
>          Issue Type: Bug
>    Affects Versions: 2.1.0
>            Reporter: Dan Langford
> Currently when an AMQPMessage is sent to a Artemis cluster configured for 
> load balancing the server throws an exception. I would be in favor of 1) 
> allowing an AMQPMessage to be load balanced. but other options include 2) 
> improving the documentation around load balancing AMQP to make users aware of 
> this limitation or 3) more gracefully handling this limitation by still 
> allowing the message to be delivered but not attempt to load balance it. the 
> current behavior is unexpected and disruptive. 
> this was brought up in the users list 
> http://activemq.2283324.n4.nabble.com/Artemis-2-1-0-struggling-to-produce-AMQP1-0-messages-on-a-cluster-server-side-load-balancing-td4728658.html
>  and responded to by [~clebertsuconic]. the error is demonstrated here in a 
> test case https://github.com/danlangford/tq-artemis-example/tree/ARTEMIS-1297
> as a work around the current suggested pattern for load balancing AMQP is to 
> use QPID Dispatch Router

This message was sent by Atlassian JIRA

Reply via email to