[ https://issues.apache.org/jira/browse/AMQNET-610?focusedWorklogId=310300&page=com.atlassian.jira.plugin.system.issuetabpanels:worklog-tabpanel#worklog-310300 ]
ASF GitHub Bot logged work on AMQNET-610: ----------------------------------------- Author: ASF GitHub Bot Created on: 11/Sep/19 04:31 Start Date: 11/Sep/19 04:31 Worklog Time Spent: 10m Work Description: Havret commented on issue #29: AMQNET-610: Race conditions during consumer creation URL: https://github.com/apache/activemq-nms-amqp/pull/29#issuecomment-530123108 Hi Chris, I see where it may come from: https://github.com/apache/activemq-nms-amqp/blob/3c7ecd5c1b11d7211027293e4e4b38f05c21e0d9/src/NMS.AMQP/NmsMessageConsumer.cs#L399-L409 In line 405 we are using AcknowledgementMode property of a session, and there is a check which throws exception when connection is closed. In qpid jms they have this helper method ```isIndividualAcknowledge``` which circumvents the check that sits inside the getter. I remember asking myself during the implementation, why the heck they have this method when they can make the check directly. Apparently I've got my answer. 😂 ``` if (acknowledgementMode == Session.CLIENT_ACKNOWLEDGE) { envelope.getMessage().setAcknowledgeCallback(new JmsAcknowledgeCallback(session)); } else if (session.isIndividualAcknowledge()) { envelope.getMessage().setAcknowledgeCallback(new JmsAcknowledgeCallback(session, envelope)); } ``` https://github.com/apache/qpid-jms/blob/94822aced1380bae7935db6305250bf6bea7a9ac/qpid-jms-client/src/main/java/org/apache/qpid/jms/JmsMessageConsumer.java#L497-L501 @cjwmorgan-sol Could you be kind enough to file an issue, so it could be fixed in a separate PR? ---------------------------------------------------------------- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org Issue Time Tracking ------------------- Worklog Id: (was: 310300) Time Spent: 1h 10m (was: 1h) > Race condition during consumer creation > --------------------------------------- > > Key: AMQNET-610 > URL: https://issues.apache.org/jira/browse/AMQNET-610 > Project: ActiveMQ .Net > Issue Type: Bug > Components: AMQP, NMS > Reporter: Krzysztof Porebski > Priority: Critical > Time Spent: 1h 10m > Remaining Estimate: 0h > > Working on CI I found a race condition that reveals itself on slower machines > and may result in losing messages by consumers. > The race condition is a result of treating the creation of consumer resources > and starting them as an atomic operation: > {code:java} > public async Task Init() > { > await Session.Connection.CreateResource(Info); > await Session.Connection.StartResource(Info); > } > {code} > On slower boxes pending message may be delivered to the provider before above > code returns: > {code:java} > public IMessageConsumer CreateConsumer(IDestination destination, string > selector, bool noLocal) > { > CheckClosed(); NmsMessageConsumer messageConsumer = new > NmsMessageConsumer(consumerIdGenerator.GenerateId(), this, destination, > selector, noLocal); > messageConsumer.Init().ConfigureAwait(false).GetAwaiter().GetResult(); > // here may be a race condition > consumers.TryAdd(messageConsumer.Info.Id, messageConsumer); > > return messageConsumer; > } > {code} > If message arrives before `Init` returns (which means before newly created > consumer is added to the list of available consumers), the message will be > ignored. > {code:java} > public void OnInboundMessage(InboundMessageDispatch envelope) > { > if (consumers.TryGetValue(envelope.ConsumerInfo.Id, out > NmsMessageConsumer messageConsumer)) > { > messageConsumer.OnInboundMessage(envelope); > } > } > {code} >  >  >  -- This message was sent by Atlassian Jira (v8.3.2#803003)