Wildfly 41.0.0.Final Sporadic Application hang

[Chris Palmer]

I have an application that has run happily in many wildfly versions, including 40.0.0.Final. I upgraded to 41.0.0.Final, and after 2-3 days light use one instance hung, and a few hours later the other did the same. Nothing more is logged as I try to connect to it. The wildfly administration console is still alive but I can't spot anything obviously wrong. After downgrading to 40.0.0.Final it is fine again, using the same ear file. I have no idea where to look for more info. These are production instances so I can't play with them too much.

An anomaly I can see in the logs is that the following appears in 41.0.0, but not in previous releases:

2026-08-07 14:24:02,922 WARN  [org.jboss.as.connector.services.resourceadapters.ResourceAdapterActivatorService$ResourceAdapterActivator] (MSC service thread 1-5) IJ020017: Invalid archive: activemq-ra
2026-08-07 14:24:02,925 WARN  [org.jboss.as.connector.services.resourceadapters.ResourceAdapterActivatorService$ResourceAdapterActivator] (MSC service thread 1-5) Severity: WARNING
Section: 19.3
Description: A ManagedConnectionFactory must have a default constructor (class must not be final)
2026-08-07 14:24:02,999 INFO  [org.apache.activemq.artemis.ra.ActiveMQRALogger] (MSC service thread 1-5) AMQ151007: Resource adaptor started
2026-08-07 14:24:03,000 INFO  [org.jboss.as.connector.services.resourceadapters.ResourceAdapterActivatorService$ResourceAdapterActivator] (MSC service thread 1-5) IJ020002: Deployed: file://RaActivatoractivemq-ra

There is a message-driven EJB that has very intermittent use, and that does appear to be working ok in both versions.

My test instance has been left running 41.0.0 for some days and has not (yet) failed. But it only gets accessed occasionally.

Any ideas of how to diagnose this would be much appreciated.

Regards, Chris