Daniel Martin created ARTEMIS-4301:
--------------------------------------
Summary: Resource leaking / ignored configuration
Key: ARTEMIS-4301
URL: https://issues.apache.org/jira/browse/ARTEMIS-4301
Project: ActiveMQ Artemis
Issue Type: Bug
Affects Versions: 2.28.0
Reporter: Daniel Martin
I have an MQTT client connecting to the ActiveMQ Artemis broker. The client
connects with a {{sessionExpiryInterval}} of 60 seconds. Among my
{{<address-settings>}} in {{{}broker.xml{}}}, I have
{{{}<auto-delete-queues>true</auto-delete-queues>{}}}. When I restart the MQTT
client process, I can see in the broker's logs that the disconnect message is
received and, after 60 seconds, the "session state" is removed.
However, every time I restart the client, the queue count in a few addresses
increases by one, and never decreases after that. Inspecting the queues, they
all have a message count of 0, and all but the one corresponding to the
currently-active client have a consumer count of 0. Furthermore, when
inspecting the queue's attributes, they are all marked as "Auto delete: false",
ignoring the setting in {{{}broker.xml{}}}.
A secondary but still important issue: I had to manually turn queue
auto-deletion on, as it was explicitly disabled by default, contrary to what is
claimed in the documentation
(https://activemq.apache.org/components/artemis/documentation/latest/address-settings.html).
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
