JBossDeveloper

Andy Taylor wrote:

 

for some reason your MDB's are being deployed twice each (see below)

 

can you double check you ear or provide buildable src.

 

13:12:50,164 INFO  [EJBContainer] STARTED EJB: org.hornetq.test.mdb.DurableNoSelectorMDB ejbName: DurableNoSelectorMDB
13:12:50,566 INFO  [EJBContainer] STARTED EJB: org.hornetq.test.mdb.DurableNoSelectorMDB ejbName: DurableNoSelector
13:12:51,025 INFO  [EJBContainer] STARTED EJB: org.hornetq.test.mdb.DurableSelectorMDB ejbName: DurableSelectorMDB
13:12:51,330 INFO  [EJBContainer] STARTED EJB: org.hornetq.test.mdb.DurableSelectorMDB ejbName: DurableSelector

The deployment contains a ejb-jar.xml with:

<?xml version="1.0" encoding="ISO-8859-1"?>
<ejb-jar version="3.0"
  xmlns="http://java.sun.com/xml/ns/javaee"
  xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
  xsi:schemaLocation="http://java.sun.com/xml/ns/javaee
http://java.sun.com/xml/ns/javaee/ejb-jar_3_0.xsd">
  <enterprise-beans>
    <message-driven>
      <ejb-name>DurableSelector</ejb-name>
      <ejb-class>org.hornetq.test.mdb.DurableSelectorMDB</ejb-class>
    </message-driven>
    <message-driven>
      <ejb-name>DurableNoSelector</ejb-name>
      <ejb-class>org.hornetq.test.mdb.DurableNoSelectorMDB</ejb-class>
    </message-driven>
  </enterprise-beans>
</ejb-jar>

So the same MDB is being mapped twice to two different names. For example, the MDB class DurableNoSelectorMDB is once deployed as an EJB with the name DurableNoSelectorMDB through annotations and once through the ejb-jar.xml with a different name DurableNoSelector.

Looking at the ejb-jar.xml, I guess Daniel can just get rid of that file altogether.

could u fix your ear and retest and confirm either way if it is still an issue

This is now fixed in trunk.

The issue was that we were recreating the subscription queue every time when we shouldnt only when the selector changed. Feel free to test trunk now.

FYI remember that if you do change the selector this will basically re create the subscription as new and only receive messages from that point in time.

thanks for the heads up on this and providing s a test that we can actually use, be great if everyone did this