[jboss-jira] [JBoss JIRA] Commented: (JBMESSAGING-1887) Messages lost during orderly shutdown of JBoss Messaging server

[Yong Hao Gao (JIRA)]

    [ https://issues.jboss.org/browse/JBMESSAGING-1887?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12613935#comment-12613935 ] 

Yong Hao Gao commented on JBMESSAGING-1887:
-------------------------------------------

Actually on second thought, that might not be helpful for a publisher because a publisher cannot re-sent the message to some of the subscribers to the topic. It only can resent the message to the topic. But doing so all subscribers will receive the message again -- including ones that have successfully received it before.


> Messages lost during orderly shutdown of JBoss Messaging server
> ---------------------------------------------------------------
>
>                 Key: JBMESSAGING-1887
>                 URL: https://issues.jboss.org/browse/JBMESSAGING-1887
>             Project: JBoss Messaging
>          Issue Type: Bug
>          Components: Messaging Core
>    Affects Versions: 1.4.0.SP3.CP13, 1.4.7.GA, 1.4.8.SP1
>         Environment: JBoss EAP 5.1.0
>            Reporter: Tom Ross
>            Assignee: Yong Hao Gao
>             Fix For: 1.4.0.SP3.CP14, 1.4.8.SP2
>
>
> During orderly shutdown of JBoss EAP server/Messaging server messages are being lost.  
> In the above test case a publisher is publishing messages to a topic (which has several durable subscribers) on a JBM server while the user uses control-c to shutdown the server or runs shutdown.sh script.
> The server will first shutdown the topic service (subscribers) and then the server peer and post office. When the subscribers has been just shutdown but the server peer and post office is still alive, if a message are sent in to the post office and post office tries to route the message to subscribers, the subscribers are already deactivated so it didn't get route.
> In this case the post office does not throw any exceptions back to the publisher. Instead it returns without error.
> Following message will apear in the log file:
> (WorkerThread#1[192.168.1.110:56860]) Queue[37225610/2/2952-guest.mdbOne] is not active, returning null delivery for Reference[41396777574629398]:RELIABLE

--
This message is automatically generated by JIRA.
For more information on JIRA, see: http://www.atlassian.com/software/jira