Details
-
Bug
-
Resolution: Done
-
Major
-
2.0.0.CR5
-
None
Description
The impl of ManagementRequestContext executeAsync catches RejectedExecutionException and doesn't notify the calling thread. The handling itself seems ok (call failed on the result handler and send a failure response to the client), but not notifying the caller is problematic. There are a number of cases where the caller thread waits for a latch to be tripped, with the async task tripping. If the latch never trips, the caller thread will block forever. The testsuite hang at http://brontes.lab.eng.brq.redhat.com/viewLog.html?buildId=71233&buildTypeId=WildFlyCore_MasterLinux&tab=buildLog#_focus=13620 looks to be a case of this, with a server not stopping due to this:
"Remoting "master:main-one:MANAGEMENT" task-13" #55 prio=5 os_prio=0 tid=0xc6c3bc00 nid=0x19d9 waiting on condition [0xc215c000] java.lang.Thread.State: WAITING (parking) at sun.misc.Unsafe.park(Native Method) - parking to wait for <0xddbb4aa0> (a java.util.concurrent.CountDownLatch$Sync) at java.util.concurrent.locks.LockSupport.park(LockSupport.java:175) at java.util.concurrent.locks.AbstractQueuedSynchronizer.parkAndCheckInterrupt(AbstractQueuedSynchronizer.java:836) at java.util.concurrent.locks.AbstractQueuedSynchronizer.doAcquireSharedInterruptibly(AbstractQueuedSynchronizer.java:997) at java.util.concurrent.locks.AbstractQueuedSynchronizer.acquireSharedInterruptibly(AbstractQueuedSynchronizer.java:1304) at java.util.concurrent.CountDownLatch.await(CountDownLatch.java:231) at org.jboss.as.controller.remote.TransactionalProtocolOperationHandler.sendResponse(TransactionalProtocolOperationHandler.java:540) at org.jboss.as.controller.remote.TransactionalProtocolOperationHandler$ExecuteRequestContext.failed(TransactionalProtocolOperationHandler.java:377) - locked <0xddbb2300> (a org.jboss.as.controller.remote.TransactionalProtocolOperationHandler$ExecuteRequestContext) at org.jboss.as.protocol.mgmt.ActiveOperationSupport$ActiveOperationImpl$2.handleFailed(ActiveOperationSupport.java:350) at org.jboss.threads.AsyncFutureTask$Reg.run(AsyncFutureTask.java:60) at org.jboss.as.protocol.mgmt.ActiveOperationSupport$1.execute(ActiveOperationSupport.java:55) at org.jboss.threads.AsyncFutureTask.safeExecute(AsyncFutureTask.java:169) at org.jboss.threads.AsyncFutureTask.setFailed(AsyncFutureTask.java:162) at org.jboss.as.protocol.mgmt.ActiveOperationSupport$ActiveOperationImpl.access$400(ActiveOperationSupport.java:295) at org.jboss.as.protocol.mgmt.ActiveOperationSupport$ActiveOperationImpl$1.failed(ActiveOperationSupport.java:315) at org.jboss.as.protocol.mgmt.AbstractMessageHandler$2.executeAsync(AbstractMessageHandler.java:333) at org.jboss.as.protocol.mgmt.AbstractMessageHandler$2.executeAsync(AbstractMessageHandler.java:315) at org.jboss.as.controller.remote.TransactionalProtocolOperationHandler$ExecuteRequestHandler.handleRequest(TransactionalProtocolOperationHandler.java:144) at org.jboss.as.protocol.mgmt.AbstractMessageHandler.handleMessage(AbstractMessageHandler.java:270) at org.jboss.as.protocol.mgmt.AbstractMessageHandler.handleMessage(AbstractMessageHandler.java:252) at org.jboss.as.protocol.mgmt.AbstractMessageHandler.handleMessage(AbstractMessageHandler.java:123) at org.jboss.as.protocol.mgmt.ManagementChannelReceiver$1.handleMessage(ManagementChannelReceiver.java:56) at org.jboss.as.protocol.mgmt.ManagementChannelReceiver.handleMessage(ManagementChannelReceiver.java:85) at org.jboss.remoting3.remote.RemoteConnectionChannel$5.run(RemoteConnectionChannel.java:463) at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1142) at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:617) at java.lang.Thread.run(Thread.java:745)
The RejectedExecutionException indicates a thread pool has been shutdown before the response has gone out. Ideally we would prevent that (for which I've recently filed a JIRA) but in any case we should make this more robust.
I think having executeAsync return a boolean should suffice.
