Collectives™ on Stack Overflow

Find centralized, trusted content and collaborate around the technologies you use most.

Teams

Q&A for work

Connect and share knowledge within a single location that is structured and easy to search.

I randomically experience a NullPointerException while trying to publish a message using RabbitTemplate.convertAndSend

I tried both spring-amqp:1.7.6 and spring-amqp:1.7.8 and this should be the relevant part of my dependency tree:

[INFO] | +- org.springframework.amqp:spring-rabbit:jar:1.7.4.RELEASE:compile
[INFO] | | +- com.rabbitmq:http-client:jar:1.1.1.RELEASE:compile
[INFO] | | \- com.rabbitmq:amqp-client:jar:4.0.3:compile
this is the stacktrace of the exception
org.springframework.amqp.UncategorizedAmqpException: java.lang.NullPointerException
        at org.springframework.amqp.rabbit.support.RabbitExceptionTranslator.convertRabbitAccessException(RabbitExceptionTranslator.java:83)
        at org.springframework.amqp.rabbit.connection.RabbitAccessor.convertRabbitAccessException(RabbitAccessor.java:113)
        at org.springframework.amqp.rabbit.core.RabbitTemplate.doExecute(RabbitTemplate.java:1461)
        at org.springframework.amqp.rabbit.core.RabbitTemplate.execute(RabbitTemplate.java:1411)
        at org.springframework.amqp.rabbit.core.RabbitTemplate.send(RabbitTemplate.java:712)
        at org.springframework.amqp.rabbit.core.RabbitTemplate.convertAndSend(RabbitTemplate.java:813)
        at org.springframework.amqp.rabbit.core.RabbitTemplate.convertAndSend(RabbitTemplate.java:791)
        at io.reactivex.internal.operators.completable.CompletableFromAction.subscribeActual(CompletableFromAction.java:34)
        at io.reactivex.Completable.subscribe(Completable.java:1635)
        at io.reactivex.internal.operators.completable.CompletableCache.subscribeActual(CompletableCache.java:59)
        at io.reactivex.Completable.subscribe(Completable.java:1635)
        at io.reactivex.internal.operators.completable.CompletableSubscribeOn$SubscribeOnObserver.run(CompletableSubscribeOn.java:64)
        at io.reactivex.Scheduler$DisposeTask.run(Scheduler.java:452)
        at io.reactivex.internal.schedulers.ScheduledRunnable.run(ScheduledRunnable.java:61)
        at io.reactivex.internal.schedulers.ScheduledRunnable.call(ScheduledRunnable.java:52)
        at java.util.concurrent.FutureTask.run(FutureTask.java:266)
        at java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.access$201(ScheduledThreadPoolExecutor.java:180)
        at java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.run(ScheduledThreadPoolExecutor.java:293)
        at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1149)
        at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:624)
        at java.lang.Thread.run(Thread.java:748)
Caused by: java.lang.NullPointerException: null
        at sun.reflect.GeneratedMethodAccessor487.invoke(Unknown Source)
        at sun.reflect.DelegatingMethodAccessorImpl.invoke(DelegatingMethodAccessorImpl.java:43)
        at java.lang.reflect.Method.invoke(Method.java:498)
        at org.springframework.amqp.rabbit.connection.CachingConnectionFactory$CachedChannelInvocationHandler.invoke(CachingConnectionFactory.java:980)
        at com.sun.proxy.$Proxy174.basicPublish(Unknown Source)
        at org.springframework.amqp.rabbit.core.RabbitTemplate.doSend(RabbitTemplate.java:1532)
        at org.springframework.amqp.rabbit.core.RabbitTemplate$3.doInRabbit(RabbitTemplate.java:716)
        at org.springframework.amqp.rabbit.core.RabbitTemplate.doExecute(RabbitTemplate.java:1455)
        ... 19 common frames omitted 
Any clue on what might be causing this?
edit: additional infos 
i didn't notice the difference between spring-rabbit and spring-amqp at first, but looking inside the jar i have these dependencies: 
128356 Tue May 29 17:50:46 CEST 2018 lib/spring-amqp-1.7.8.RELEASE.jar
469867 Mon Sep 11 15:28:04 CEST 2017 lib/spring-rabbit-1.7.4.RELEASE.jar
481919 Wed Jun 07 15:25:06 CEST 2017 lib/amqp-client-4.0.3.jar
synchronized (this.targetMonitor) {
                if (this.target == null) {
                    this.target = createBareChannel(this.theConnection, this.transactional);
                Object result = method.invoke(this.target, args);
                if (this.transactional) {
                    if (txStarts.contains(methodName)) {
                        this.txStarted = true;
                    else if (txEnds.contains(methodName)) {
                        this.txStarted = false;
                return result;
Pay attention to the if (this.target == null) { and into the method.invoke(this.target, args);. In between of these to operations the this.target may become as null. That's the fact for the catch (InvocationTargetException ex) { block below: we do this.target = null; outside of the synchronized (this.targetMonitor) {. So, some other thread may cause a race condition for this target property. 
That's what I mean about the NPE problem and this is kinda a suggestion how to fix.
Now we need to understand how is it even possible that different threads get access to the same ChannelProxy to cause such a race condition... Maybe there is still some flaw around caching for the channels. 
Would be great, of course, if you try your solution with the latest spring-rabbit-1.7.8, because you dependencies still show us it as a 1.7.4.
                i edited my post to add the versions i have inside the jar, i'll update spring-rabbit version and come back to you in the following days.
– davide bubz
                Jul 11, 2018 at 16:50
                You need only spring-rabbit-1.7.8.RELEASE: all others are transitive and will be polled automatically.
– Artem Bilan
                Jul 11, 2018 at 17:10
                i tried upgrading to spring-rabbit 1.7.8 and i still get the nullpointer, maybe if i try another connection factory i can at least bypass  the problem?
– davide bubz
                Jul 12, 2018 at 13:48
                OK. I see. So pity to hear... May you can share with us a simple project on GitHub to reproduce and play? You might perform some physicalClose() from other thread though...
– Artem Bilan
                Jul 12, 2018 at 13:56
                im going to try to reproduce in a simple project, but atm i can't even reproduce in a test in the actual project, it only shows up in the live environment... looking at my code i only directly handle a connection to create a queue, and then i dont close that connection, i'm not sure if this can at some point result in the npe i'm seeing.
– davide bubz
                Jul 12, 2018 at 14:23
I got the same error while publishing randomly in live environment, this is due to creating the channels and not closing them. Rabbitmq has a limit of 5000 channels.
This could not be easily reproducible in test environment as it requires opening 5k channels.
        
Thanks for contributing an answer to Stack Overflow!
Please be sure to answer the question. Provide details and share your research!
But avoid …
Asking for help, clarification, or responding to other answers.
Making statements based on opinion; back them up with references or personal experience.
To learn more, see our tips on writing great answers.