EJB jars not updating until hot deployment Development mode server is restarted
Adam Sulesky
of EJB jar files on the server. Here are the particulars. The
Weblogic server is running in DEVELOPMENT mode, not production. No
exceptions or errors are being thrown from deployment and it is
successful the first time deployed or when the server is restarted.
If the EJB is updated, either through the console or the
weblogic.deploy utility using the 'update' action, it appears in the
applications directory. The .wlnotdelete directory shows that the
server copied the new JAR and placed it in the directory. The problem
is that the new JAR is not being loaded into the server. If updating
through the console, we un-deploy the EJB and then redeploy it (by
unchecking and checking the Deploy checkbox). The updated JAR isn't
loaded into the active server until the server is restarted.
The logs verify that production mode is turned off.
<Java system property: weblogic.ProductionModeEnabled = false>
What would cause weblogic to not refresh resident memory with the
latest JAR while in development mode?
The jar, for what its worth, is precompiled by the weblogic.ejbc in
the form of Foo-ejb-compiled.jar (noting because I read there was an
existing bug with jar name using dots).
Rob Woollen
errors?
In development mode, you should just be able to copy a new jar into the
applications directory and see the updated version.
-- Rob
Adam Sulesky
there are a few common support classes found in the weblogic
classpath. Some of those common support classes are also found in the
EJB.jar.
There aren't any exceptions when deploying it. Thats whats so
confusing about this. One problem we had was with a class not being
included in the deployed JAR. We created a new EJB, and tried to
redeploy. No luck. It was possible to open the Jar's in applications
and in .wlnotdelete and verify that there was latest copy of the JAR
and it included the missing class, although ClassDefNotFound
exceptions would be thrown. Undeployed, redeployed -- no success.
Restarted the server, it loaded properly.
Rob Woollen <r...@nomail.org> wrote in message news:<3D780EBE...@nomail.org>...
Adam Sulesky
number of exceptions at startup now that we're running SP3.
We aren't sure what the cause of these errors are, or if we should
even be concerned about them. We've done searches, and it appears to
be a popular question on the boards but no one has answered what it
means.
<Sep 6, 2002 11:31:27 AM EDT> <Error> <NT Performance Pack> <failure
in processSockets() - GetData: 'weblogic.socket.Get
Data@2d3df7 - fd: '2660', numBytes: '14''
java.lang.NullPointerException
at weblogic.rjvm.ConnectionManager.getOutputStream(ConnectionManager.java:1316)
at weblogic.rjvm.ConnectionManager.createPeerGoneMsg(ConnectionManager.java:1172)
at weblogic.rjvm.ConnectionManagerServer.handleIdentifyRequest(ConnectionManagerServer.java:228)
at weblogic.rjvm.ConnectionManager.dispatch(ConnectionManager.java:628)
at weblogic.rjvm.t3.T3JVMConnection.dispatch(T3JVMConnection.java:454)
at weblogic.socket.NTSocketMuxer.processSockets(NTSocketMuxer.java:658)
at weblogic.socket.SocketReaderRequest.execute(SocketReaderRequest.java:24)
at weblogic.kernel.ExecuteThread.execute(ExecuteThread.java:139)
at weblogic.kernel.ExecuteThread.run(ExecuteThread.java:120)
>
Rob Woollen <r...@nomail.org> wrote in message news:<3D780EBE...@nomail.org>...
Adam Sulesky
anywhere in the release notes for SP3 (or SP2).
The EJB is being updated and deployed within Weblogic. But WAR's that
have not been redeployed since the EBJ was updated, continue to use
the old version of the EJB.jar. The previous behavior was for the
WAR's to throw an exception due to having a stale/non-existent JNDI
reference to the EJB.
We like this new behavior, just didn't find it documented anywhere.
It would also be helpful if this was configurable.
Rob Woollen <r...@nomail.org> wrote in message news:<3D780EBE...@nomail.org>...
Rajesh Mirchandani
Adam Sulesky wrote:
--
Rajesh Mirchandani
Developer Relations Engineer
BEA Support