[UIMA-5768] UIMA-AS: fix race condition leading to NPE while recovering from lost connection to broker

• Type: Bug
• Status: Open
• Priority: Major
• Resolution: Unresolved
• Fix Version/s: 3.0.0AS
• Component/s: Async Scaleout
• Assignee: Jaroslaw Cwiklik
• Reporter: Jaroslaw Cwiklik
• Created: 20/Apr/18 14:42
• Updated: 20/Apr/18 14:42

Description

When UIMA-AS client looses broker connection it goes into a retry mode trying to reconnect. In one particular case the NPE was thrown :

7| | javax.jms.JMSException: java.io.EOFException
7| | at org.apache.activemq.util.JMSExceptionSupport.create(JMSExceptionSupport.java:54)
7| | at org.apache.activemq.ActiveMQConnection.onAsyncException(ActiveMQConnection.java:1948)
7| | at org.apache.activemq.ActiveMQConnection.onException(ActiveMQConnection.java:1967)
7| | at org.apache.activemq.transport.TransportFilter.onException(TransportFilter.java:114)
7| | at org.apache.activemq.transport.ResponseCorrelator.onException(ResponseCorrelator.java:126)
7| | at org.apache.activemq.transport.TransportFilter.onException(TransportFilter.java:114)
7| | at org.apache.activemq.transport.TransportFilter.onException(TransportFilter.java:114)
7| | at org.apache.activemq.transport.WireFormatNegotiator.onException(WireFormatNegotiator.java:173)
7| | at org.apache.activemq.transport.AbstractInactivityMonitor.onException(AbstractInactivityMonitor.java:345)
7| | at org.apache.activemq.transport.TransportSupport.onException(TransportSupport.java:96)
7| | at org.apache.activemq.transport.tcp.TcpTransport.run(TcpTransport.java:219)
7| | at java.lang.Thread.run(Thread.java:811)
7| | Caused by: java.io.EOFException
7| | at java.io.DataInputStream.readFully(DataInputStream.java:208)
7| | at java.io.DataInputStream.readInt(DataInputStream.java:398)
7| | at org.apache.activemq.openwire.OpenWireFormat.unmarshal(OpenWireFormat.java:268)
7| | at org.apache.activemq.transport.tcp.TcpTransport.readCommand(TcpTransport.java:240)
7| | at org.apache.activemq.transport.tcp.TcpTransport.doRun(TcpTransport.java:232)
7| | at org.apache.activemq.transport.tcp.TcpTransport.run(TcpTransport.java:215)
7| | ... 1 more
7| | java.lang.NullPointerException
7| | at org.apache.uima.adapter.jms.client.ActiveMQMessageSender.createSession(ActiveMQMessageSender.java:109)
7| | at org.apache.uima.adapter.jms.client.ActiveMQMessageSender.initializeProducer(ActiveMQMessageSender.java:144)
7| | at org.apache.uima.adapter.jms.client.ActiveMQMessageSender.getMessageProducer(ActiveMQMessageSender.java:164)
7| | at org.apache.uima.adapter.jms.client.BaseMessageSender.run(BaseMessageSender.java:286)
7| | at java.lang.Thread.run(Thread.java:811)

Looks like the above is caused by some race condition. One thread trying to reconnect while another trying to use a connection handle which had been invalidated.