There seems to be no stacktrace as it probably gets caught somewhere and not rethrown. There is only this warning which is why this problem initially was so perplexing. I just overlooked it because it was hidden between other debug logs.
2025-07-29 20:13:53,341 WARN [org.springframework.jms.JmsListenerEndpointContainer#0-2] o.s.j.l.DefaultMessageListenerContainer: Setup of JMS message listener invoker failed for destination 'my.jms.topic' - trying to recover. Cause: Cannot invoke "java.lang.Boolean.booleanValue()" because the return value of "org.apache.activemq.artemis.api.core.client.ClientSession$QueueQuery.isConfigurationManaged()" is null Philip On Tuesday, July 29th, 2025 at 19:43, Justin Bertram <jbert...@apache.org> wrote: > Do you happen to have the stack-trace from the NPE? I'd like to add it to > the Jira for posterity's sake. > > > Justin > > On Tue, Jul 29, 2025 at 11:02 AM Philip philip.pi...@protonmail.com.invalid > > wrote: > > > You are indeed correct that I am using an old broker (Wildfly 10). > > > > I am unable to provide a complete demo project as this would be more > > complex than it is worth. > > I have created a GitHub repository in which the general JMS setup is > > duplicated with a description of the rest of the setup and the issue at > > hand. > > https://github.com/PinkMoustacheMan/spring-artemis-issue-test > > > > I hope this explains the issue better than my previous mails. > > > > Philip > > > > On Monday, July 28th, 2025 at 06:14, Justin Bertram jbert...@apache.org > > wrote: > > > > > Do you have a sample project or anything that we can use to reproduce > > > what > > > you're seeing? At this point, the only way I can see how you might be > > > getting null there is if you were using a newer client with an older > > > broker, but it's certainly possible I'm not seeing all the possibilities. > > > Regardless, it looks like you're hitting a bug and I need a way to > > > reproduce it in order to fix it. Even a clear, concise description of the > > > steps necessary to reproduce the issue (i.e. no code) would go a long way > > > here. > > > > > > Justin > > > > > > On Sat, Jul 26, 2025 at 11:33 PM Philip philip.pi...@protonmail.com > > > .invalid > > > > > > wrote: > > > > > > > I should have been more clear about my exact setup. > > > > > > > > I am trying to create multiple consumers that either listen to a JMS > > > > queue > > > > or a JMS topic. > > > > Each consumer listening to a queue has no client ID as it is not > > > > needed. > > > > Each consumer listening to a topic has both a client ID (the same one > > > > for > > > > all consumers) > > > > and a subscription name (unique per consumer) configured. > > > > > > > > The queues and topics are all hosted within an external application > > > > server > > > > that I am connecting to. > > > > All of this works perfectly until I try to change the subscription > > > > name of > > > > one of the topic consumers. > > > > > > > > I believe this issue stems from the fact that they are durable > > > > subscribers, which leads my > > > > application server to store them persistently. > > > > After changing the subscription name, the ActiveMQSession is aware of > > > > the > > > > fact that it has changed > > > > and I run into the second part of this if-statement because no > > > > short-circuiting occurs. > > > > https://github.com/apache/activemq-artemis/blob/98d792d135dacb8fc567bc514f42a40d91c0657b/artemis-jms-client/src/main/java/org/apache/activemq/artemis/jms/client/ActiveMQSession.java#L1002 > > > > > > I am trying to understand why the choice was made to have this > > > > `configurationManaged` property be > > > > nullable as it leads to an issue where the null value is attempted to > > > > be > > > > inverted which leads to an exception. > > > > > > > > I am using ActiveMQ Artemis 2.40.0 > > > > > > > > Philip > > > > > > > > On Saturday, July 26th, 2025 at 20:18, Justin Bertram > > > > jbert...@apache.org > > > > wrote: > > > > > > > > > The simplest way to investigate is with code that reproduces the > > > > > problem. > > > > > Do you have a sample project or anything that we can use to reproduce > > > > > what > > > > > you're seeing? > > > > > > > > > > Also, what version of ActiveMQ Artemis are you using on the broker > > > > > and > > > > > the > > > > > client? > > > > > > > > > > Lastly, can you clarify your use-case? It seems you're trying to > > > > > create > > > > > multiple different durable subscriptions using the same client ID and > > > > > subscription name on the same topic. Is that correct? > > > > > > > > > > Justin > > > > > > > > > > On Fri, Jul 25, 2025 at 12:59 PM Philip philip.pi...@protonmail.com > > > > > .invalid > > > > > > > > > > wrote: > > > > > > > > > > > Hello ActiveMQ users, > > > > > > > > > > > > I am writing this with square eyes and a tired mind. > > > > > > > > > > > > I am trying to do an (assumedly) simple thing within my Spring Boot > > > > > > application, but I keep failing. > > > > > > > > > > > > I want to be able to define JmsListener endpoints that may either > > > > > > be > > > > > > listening to queues or topics. > > > > > > In order to achieve that I want to provide different configurations > > > > > > that > > > > > > e.g. make it so that the underlying connection is durable. > > > > > > > > > > > > All seems to go well until - in the deepest depths of ActiveMQ - a > > > > > > consumer is attempted to be constructed. > > > > > > A request is sent to (presumably) check if the Topic can be reached > > > > > > using > > > > > > the provided configuration (url, username, password). > > > > > > The initial response seems fine, until this dreaded line is reached > > > > > > and my > > > > > > world starts to crumble. > > > > https://github.com/apache/activemq-artemis/blob/98d792d135dacb8fc567bc514f42a40d91c0657b/artemis-jms-client/src/main/java/org/apache/activemq/artemis/jms/client/ActiveMQSession.java#L1002 > > > > > > > > Because as it turns out, this `result.isConfigurationManaged()` > > > > > > statement > > > > > > returns `null` and an Exception is thrown when trying to invert it. > > > > > > > > > > > > I am trying to understand what I need to do for this value to be > > > > > > set > > > > > > (or > > > > > > if this value is not supposed to be set and this is indeed a bug). > > > > > > > > > > > > On another (connected) note, this check seems to only happen > > > > > > because > > > > > > the > > > > > > `selectorChanged` flag is true. > > > > > > I don't quite understand what it has changed from (or at least how > > > > > > the > > > > > > `oldFilterString` is even able to be set to a value in the first > > > > > > place > > > > > > when > > > > > > the consumer that is constructed is the first an only one). > > > > > > > > > > > > I am assuming this is in some way connected to this javadoc > > > > > > comment, > > > > > > but I > > > > > > cannot figure out how to combat this issue. > > > > https://github.com/apache/activemq-artemis/blob/98d792d135dacb8fc567bc514f42a40d91c0657b/artemis-jms-client/src/main/java/org/apache/activemq/artemis/jms/client/ActiveMQSession.java#L983-L988 > > > > > > > > I hope I was able to somewhat explain my disarrayed thoughts. > > > > > > > > > > > > Best Regards,Philip > > > > > > > > --------------------------------------------------------------------- > > > > To unsubscribe, e-mail: users-unsubscr...@activemq.apache.org > > > > For additional commands, e-mail: users-h...@activemq.apache.org > > > > For further information, visit: https://activemq.apache.org/contact > > > > --------------------------------------------------------------------- > > To unsubscribe, e-mail: users-unsubscr...@activemq.apache.org > > For additional commands, e-mail: users-h...@activemq.apache.org > > For further information, visit: https://activemq.apache.org/contact --------------------------------------------------------------------- To unsubscribe, e-mail: users-unsubscr...@activemq.apache.org For additional commands, e-mail: users-h...@activemq.apache.org For further information, visit: https://activemq.apache.org/contact
