1a) The default name for the dead letter queue is DLQ. You can use the web console (http://activemq.apache.org/web-console.html) to view the queues the broker has. Note that ActiveMQ will automatically create a destination when the first message is published into it, so if it doesn't exist then it just means that there hasn't yet been a message moved to the DLQ.
1b) The link Tim Bish provided contains instructions for telling which destinations are causing your KahaDB log files to be held. If the DLQ had old messages, it's likely that you would see that queue in the output of that command, telling you that those old DLQ messages were keeping log files from being deleted. 2) db.data is the index. It won't shrink when messages are deleted, but it should stop growing once the broker reaches its max size. So it may not stay small per se (depending on your definition of small), but it should not grow unbounded. The comments on https://issues.apache.org/jira/plugins/servlet/mobile#issue/AMQ-3736 may be useful for understanding the expected behavior of the index file. Tim On Mar 15, 2017 6:08 AM, "Hidekazu" <hi_de...@yahoo.co.jp> wrote: Dear cschneider Thanks for teaching me a good way to avoid increasing kahaDB. I came to know that dead letter queues are one of keys. By the way, I have two questions. (I'm sorry, I'm ActiveMQ newcomer...) 1. How do I see dead letter queues? I want to make sure whether dead letter queue make kahadb bigger. command? filename for dead letter queue? 2.I was able to keep KahaDB small by deleteing pending queues and delete journal files. however, I noticed that db.data kept increasing. I think this is a different problem from managing kahaDB journals. Do you know how you keep db.data small? Thanks Kawai -- View this message in context: http://activemq.2283324.n4.nab ble.com/The-size-of-KahaDB-doesn-t-seem-to-decrease-tp4723669p4723778.html Sent from the ActiveMQ - User mailing list archive at Nabble.com.