Affects 1.2.0.Final in current WildFly master (just post 8.0.0.Final release).

A server hung during shutdown during the domain testsuite:

http://brontes.lab.eng.brq.redhat.com/viewLog.html?buildId=10974&tab=buildResultsDiv&buildTypeId=WF_MasterIgnoreLinux

I'll attach the thread dump. The thread state is reminiscent of MSC-139 but not exactly the same.

Controller Boot Thread is waiting to acquire monitor of ServiceControllerImpl 0xb0d66b90.

MSC service thread 1-6 holds that monitor. It is waiting in StabilityMonitor.awaitCleanupCompletion; wait for Object 0xa45cf130.

MSC service thread 1-1, 1-2, 1-3 and 1-4 are also all waiting in StabilityMonitor.awaitCleanupCompletion; wait for Object 0xa45cf130.

MSC service thread 1-5 is blocked waiting to lock ServiceControllerImpl 0xb0e19f48.

MSC service thread 1-4 holds the monitor for ServiceControllerImpl 0xb0e19f48 that is needed by 1-5.

The reason MSC 1-1/2/3/4/6 are waiting in StabilityMonitor.awaitCleanupCompletion is because StabilityMonitor.cleanupInProgress == true. The reason it's true is because Controller Boot Thread is executing the clear() method and has set it to true. CBT can't set it to 'false' because MSC 1-6 is holding the monitor.