Tommy Lindgren created ARTEMIS-5205:
---------------------------------------
Summary: Queue messages can become stuck if using noLocal
Key: ARTEMIS-5205
URL: https://issues.apache.org/jira/browse/ARTEMIS-5205
Project: ActiveMQ Artemis
Issue Type: Bug
Affects Versions: 2.37.0
Reporter: Tommy Lindgren
I have STOMP clients using noLocal. There are diverts routing topic messages to
queues and I have observed that queue messages sometimes become stuck. Theory:
# A STOMP client sends a message to a topic with noLocal=true. The __AMQ_CID
header is added to the message.
# There's a divert from that topic to a queue.
# The STOMP client happens to subscribe to that queue, also with noLocal=true.
This client is the only consumer for that queue.
# The message is added to the queue and is now stuck, because the only
consumer has a has selector expression filtering on that message's __AMQ_CID
header.
Not sure if this is an Artemis bug or an application/configuration issue but I
don't believe I saw this behavior with ActiveMQ Classic. I suspect Classic
solved noLocal filtering in a different way or simply didn't respect the
setting for queues (the application filters messages locally as a fallback).
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
For further information, visit: https://activemq.apache.org/contact
