In an Active / Backup scenario, run the following steps:

1. Configure the # address-settings like this:

<address-setting match="#">
...
<auto-create-queues>true</auto-create-queues>
<auto-create-addresses>true</auto-create-addresses>
<auto-delete-queues>true</auto-delete-queues>
<auto-delete-addresses>true</auto-delete-addresses>
<auto-delete-queues-message-count>-1</auto-delete-queues-message-count>
<auto-delete-queues-delay>3600000</auto-delete-queues-delay>
<auto-delete-created-queues>true</auto-delete-created-queues>
</address-setting>

2. The above configuration stays that all auto created address/queue should be auto delete after 1 day if it doesn't have any consumers, doesn't matter the number of messages on it.
3. Using a MQTT client, send a message, with the retained flag equals true, to a TEST queue.
4. Go to the console, look at the created address/queue $sys.mqtt.retain.TEST. As expected by the MQTT specification.
5. Shutdown the active node.
6. Look the logs in the backup node:

INFO [org.apache.activemq.artemis.core.server] AMQ224112: Auto removing Queue $sys.mqtt.retain.TEST with queueID=780 on address=$sys.mqtt.retain.TEST
INFO [org.apache.activemq.artemis.core.server] AMQ224113: Auto removing Address $sys.mqtt.retain.TEST

7. When failing over, the backup node ignored the auto-delete-queues-delay and immediately deleted the queue.