We have a queue in our production system (on Artemis 2.27.1) that ended up being used before it was defined in our broker.xml.
This meant that the address has the “Auto created: true” attribute, and the queue under it ended up with “Auto delete: true” and “Configuration managed: false” We run in two datacenters with mirroring across. What I’m seeing in production (but haven’t yet been able to produce in my local) is that the cold side of the queue is being auto deleted – frequently: 2023-02-09 18:35:01,189 INFO [org.apache.activemq.artemis.core.server] AMQ224112: Auto removing Queue CadInbound with queueID=274511220 on address=CadInbound 2023-02-09 18:37:01,190 INFO [org.apache.activemq.artemis.core.server] AMQ224112: Auto removing Queue CadInbound with queueID=274540002 on address=CadInbound 2023-02-09 18:38:01,190 INFO [org.apache.activemq.artemis.core.server] AMQ224112: Auto removing Queue CadInbound with queueID=274561685 on address=CadInbound 2023-02-09 18:39:31,190 INFO [org.apache.activemq.artemis.core.server] AMQ224112: Auto removing Queue CadInbound with queueID=274574121 on address=CadInbound I tested in my local adding the queue to broker.xml while the auto created one still exists, and though I see it deployed in the log, in the console it still says auto created, and the queue still gets deleted. Our worst case workaround would be to shut down all the producers, halting production processing, deleting the queue, and then touching broker.xml to restore it properly. Is there any other less impactful way I can force artemis to recognize that and existing auto address/queue is now configured? Stephen E. Baker
