[ARTEMIS-1186] Consumer.receive hangs if http acceptor with non-zero batch-delay is configured

• Type: Bug
• Status: Closed
• Priority: Major
• Resolution: Fixed
• Affects Version/s: 2.1.0
• Fix Version/s: 2.2.0
• Assignee: Unassigned
• Reporter: Tomas Hofman
• Created: 29/May/17 08:39
• Updated: 25/Jul/17 13:59
• Resolved: 25/Jul/17 13:59

Description

Test scenario:

1. Call consumer.receive(1000) on empty queue.
2. Check whether receive call is unblocked after the 1 second.

Issue: The receive call is not unblocked.

I hit the issue which was initially reported in JBEAP-6443. I found out that the issue is somehow related to http connectors/acceptors and batch-delay parameter. I am able to reproduce it only if a client connects via http-connector with non-zero batch-delay. When I tried classic remote-connector, everything worked with both zero and non-zero batch-delay.

I am able to reproduce the issue with both Artemis 1.1 (EAP 7.0.3) and Artemis 1.4.

From the thread dump, you can see where the client hangs.

Stack trace of thread: Thread[Thread-10,5,main]
---java.lang.Object.wait(Native Method)
---org.apache.activemq.artemis.core.client.impl.ClientConsumerImpl.receive(ClientConsumerImpl.java:258)
---org.apache.activemq.artemis.core.client.impl.ClientConsumerImpl.receive(ClientConsumerImpl.java:391)
---org.apache.activemq.artemis.jms.client.ActiveMQMessageConsumer.getMessage(ActiveMQMessageConsumer.java:198)
---org.apache.activemq.artemis.jms.client.ActiveMQMessageConsumer.receive(ActiveMQMessageConsumer.java:122)
---org.jboss.qa.hornetq.apps.clients.Receiver11.receiveMessage(Receiver11.java:140)
---org.jboss.qa.hornetq.apps.clients.ReceiverAutoAck.run(ReceiverAutoAck.java:75)

Issue Links

• links to: GitHub Pull Request #1297
• links to: GitHub Pull Request #1309
• links to: JBEAP-6570