I'm not aware of any false positives for the invalid concurrent session usage check. The stack-trace indicates the offense is occurring in the thread used to service an invocation of a MessageListener's onMessage method. Presumably there is another thread using the same Session. Do you have a way to reproduce this?
I do not believe WildFly supports setting suppress-session-notifications. If you're experiencing a large number of these notifications it may indicate a problem with one or more of your clients using an anti-pattern. I can't really comment on the ActiveMQUnblockedException, ActiveMQRemoteDisconnectedException, and ActiveMQNotConnectedException without more details. Justin On Thu, Nov 30, 2023 at 8:47 AM Zoltan Lutor (Nokia) <zoltan.lu...@nokia.com.invalid> wrote: > Hi, > > I've experienced some strange behavior during trying to experiment with > Apache Camel 4.0 SJMS2 and Wildfly 29.0.1-1 and Apache ActiveMQ Artemis: > > 1. <warning> AMQ212051: Invalid concurrent session usage. Sessions are not > supposed to be used by more than one thread concurrently. > java.lang.Exception: trace > mailto:org.apache.activemq.artemis@2.26.0 > //org.apache.activemq.artemis.core.client.impl.ClientSessionImpl.startCall(ClientSessionImpl.java:1587) > mailto:org.apache.activemq.artemis@2.26.0 > //org.apache.activemq.artemis.core.client.impl.ClientSessionImpl.acknowledge(ClientSessionImpl.java:1209) > mailto:org.apache.activemq.artemis@2.26.0 > //org.apache.activemq.artemis.core.client.impl.ClientConsumerImpl.doAck(ClientConsumerImpl.java:1138) > mailto:org.apache.activemq.artemis@2.26.0 > //org.apache.activemq.artemis.core.client.impl.ClientConsumerImpl.acknowledge(ClientConsumerImpl.java:785) > mailto:org.apache.activemq.artemis@2.26.0 > //org.apache.activemq.artemis.core.client.impl.ClientMessageImpl.acknowledge(ClientMessageImpl.java:136) > mailto:org.apache.activemq.artemis@2.26.0 > //org.apache.activemq.artemis.core.client.impl.ClientMessageImpl.acknowledge(ClientMessageImpl.java:38) > mailto:org.apache.activemq.artemis@2.26.0 > //org.apache.activemq.artemis.jms.client.JMSMessageListenerWrapper.onMessage(JMSMessageListenerWrapper.java:136) > mailto:org.apache.activemq.artemis@2.26.0 > //org.apache.activemq.artemis.core.client.impl.ClientConsumerImpl.callOnMessage(ClientConsumerImpl.java:1001) > mailto:org.apache.activemq.artemis@2.26.0 > //org.apache.activemq.artemis.core.client.impl.ClientConsumerImpl$Runner.run(ClientConsumerImpl.java:1154) > > org.apache.activemq.artemis.journal//org.apache.activemq.artemis.utils.actors.OrderedExecutor.doTask(OrderedExecutor.java:56) > ... 5 more > > I've found an age-old thread where they were discussing about quite > similar issue and, based on that, I was afraid of having some bug in SJMS2 > - or SJMS as it does the heavy lifting behind SJMS2 - component of Camel > when is being used with Apache ActiveMQ Artemis (as messaging subsystem in > Wildfly) > https://developer.jboss.org/thread/231892 > > 2. I was not able to figure out how to switch off > suppress-session-notifications via Wildfly settings/CLI. Any idea if it can > be done? There are tons of related notification messages created > (open/close of session) that is not consumed by anyone ATM... > > 3. I see tons of ActiveMQUnblockedException, > ActiveMQRemoteDisconnectedException and ActiveMQNotConnectedException in my > test run (last two exceptions have exactly the same number of occurrence) > when the simple demo application tries to use simple Apache Camel SJMS2 > endpoint with async consumers, exclusive reply-to queues, 4 worker threads > for request and reply queues, respectively with given rate of traffic. In > principle, SJMS2 components should provide connection pooling, Artemis is > connected via in-vm stuff as Wildfly messaging backend... > > Any help/hint is more than welcome and thanks in advance. > > Br, > >
