Questionable processing of interruptions in TcpTransport::doStop
----------------------------------------------------------------
Key: AMQ-2526
URL: https://issues.apache.org/activemq/browse/AMQ-2526
Project: ActiveMQ
Issue Type: Bug
Components: Transport
Affects Versions: 5.3.0, 5.2.0
Reporter: Nick
Imagine you are processing a few jobs by a thread pool. A timeout is set for
the whole batch. A job should send a JMS message. If the timeout expires before
all the jobs are completed the pool will interrupt still running jobs.
Most of the time the interruption will be caught and processed deep inside of
ActiveMQ TCP transport classes. While I'm not entirely convinced it's a good
idea to shut down and reopen the connection to the ActiveMQ server if a client
thread is merely interrupted what really seems ugly is:
15:12:53,745 ERROR [org.apache.activemq.transport.tcp.TcpTransport] Could not
stop service: tcp:///x.x.x.x:61616. Reason: java.lang.InterruptedException
java.lang.InterruptedException
at
java.util.concurrent.locks.AbstractQueuedSynchronizer.tryAcquireSharedNanos(AbstractQueuedSynchronizer.java:1200)
at java.util.concurrent.CountDownLatch.await(CountDownLatch.java:245)
at
org.apache.activemq.transport.tcp.TcpTransport.doStop(TcpTransport.java:482)
at org.apache.activemq.util.ServiceSupport.stop(ServiceSupport.java:69)
at
org.apache.activemq.transport.tcp.TcpTransport.stop(TcpTransport.java:499)
at
org.apache.activemq.transport.InactivityMonitor.stop(InactivityMonitor.java:113)
at
org.apache.activemq.transport.TransportFilter.stop(TransportFilter.java:64)
at
org.apache.activemq.transport.WireFormatNegotiator.stop(WireFormatNegotiator.java:87)
at
org.apache.activemq.util.ServiceSupport.dispose(ServiceSupport.java:43)
at
org.apache.activemq.transport.failover.FailoverTransport.handleTransportFailure(FailoverTransport.java:201)
at
org.apache.activemq.transport.failover.FailoverTransport.oneway(FailoverTransport.java:471)
at
org.apache.activemq.transport.MutexTransport.oneway(MutexTransport.java:40)
at
org.apache.activemq.transport.ResponseCorrelator.oneway(ResponseCorrelator.java:60)
at
org.apache.activemq.ActiveMQConnection.doAsyncSendPacket(ActiveMQConnection.java:1214)
at
org.apache.activemq.ActiveMQConnection.asyncSendPacket(ActiveMQConnection.java:1208)
at org.apache.activemq.ActiveMQSession.send(ActiveMQSession.java:1643)
at
org.apache.activemq.ActiveMQMessageProducer.send(ActiveMQMessageProducer.java:227)
at
org.apache.activemq.ActiveMQMessageProducerSupport.send(ActiveMQMessageProducerSupport.java:241)
and the reason for it is that the await call on the CountDownLatch in
TcpTransport::doStop will throws an InterruptedException if the calling thread
is already interrupted. No attempt is made (in both 5.2 or 5.3) to gracefully
process InterruptedException, the exception itself is logged as ERROR with a
rather menacing message and the log file gets full of meaningless stack traces
although no real harm was done.
Calling latch.await(1,TimeUnit.SECONDS) in a try block seems like a no-brainer
but there could be even smarted approaches to processing InterruptedExceptions
differently than, say, IOEs and other genuine problems.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
