I have the following code that I'm trying to deploy as an EJB to WebLogic 12c, but I'm getting an error:
"Error deploying the EJB GeopoliticalServiceBean(Application:
campaigner-ejb, EJBComponent: campaigner-service.jar), the JNDI name
java:global/campaigner-ejb/campaigner-service/GeopoliticalServiceBean!com.dr_dee_sw.campaigner.service.GeopoliticalServiceLocal
is already in use. You must set a different JNDI name in the
weblogic-ejb-jar.xml deployment descriptor or corresponding annotation
for this EJB before it can be deployed."
public interface GeopoliticalService
{
...
}
#Local
public interface GeopoliticalServiceLocal extends GeopoliticalService
{
}
#Remote
public interface GeopoliticalServiceRemote extends GeopoliticalService
{
}
#TransactionManagement(value = TransactionManagementType.CONTAINER)
#TransactionAttribute(value = TransactionAttributeType.REQUIRED)
#Stateless
public class GeopoliticalServiceBean implements GeopoliticalServiceLocal,GeopoliticalServiceRemote
{
...
}
More information:
I've reduced the EJB-JAR file, campaigner-service.jar, so that there's only one bean in it, plus the interfaces and exceptions. the EAR file, campaigner-ejb.ear, has only the EJB-JAR in it at the main level. It also has a "lib" directory with supporting libraries, but it only has the DAO and DTO jars in it plus third-party libraries. So, to me, it doesn't seem like a packaging issue.
This is my first app using all annotations, but it still seems fairly straight-forward. What am I missing?
During the migration from Weblogic 10 to Weblogic 12 we faced the same issue.
We could reproduce the issue by deploying the EAR on a fresh server without the datasources configured properly; this would cause a NameNotFoundException during deployment. Every next attempt to deploy the EAR would result in the JDNI name is already in use, even after restarting, undeploying, redeploying.
The only thing that resolved the issue was removing the cache (and most importantly the EJBCompilerCache) and tmp folder of the target server.
WebLogic caches artefacts even if they are undeployed (deleted). The solution is to remove the caches.
Solution
For each domain you should remove the cached artefacts. For example for svc domain, stop all the servers (Admin and managed), then
Execute the following to delete the tmp and EJB caches.
cd /data/weblogic/Oracle/products/Oracle_Home/fusion/user_projects/domains/svc
rm -fr ./servers/svc_srv1/cache/EJBCompilerCache
rm -fr ./servers/Administrator/cache/EJBCompilerCache
rm -fr ./servers/AdminServer/tmp
rm -fr ./servers/svc_srv1/tmp
Otherwise, from weblogic admin console, go to Home >Summary of Servers >YOUR_DOMAIN_NAME and then on EJBs tab tick Force Generation to true
Related
I currently have the following project structure
EAR
|---myapp1.war
|---myapp2.war
|---myapp-ejb.jar
I would like to get rid of the ear and deploy myapp1 and myapp2 on their own. I tried to make myapp-ejb.jar a maven dependency of the two war and everything works fine at compile time. Nevertheless, there are a lot of jndi lookups in the code that fail at deploy time. Is there a way to make this to work?
Theoretically, it is possible to have ejb in a web archive (war).
Although not necessary, you could try to put a 'ejb-jar.xml' file (located in the WAR module’s WEB-INF) to replace the normal auto-discovery mechanism ?
Be also aware that the WAR file must be version 2.5 or later to contain EJB content.
Extract the WAR from the EAR
Scope ejbs as "provided" per https://stackoverflow.com/a/42847318/8528208
Place the ejb.jar in your WAR's WEB-INF/lib per https://stackoverflow.com/a/6968674/8528208
Further documentation at https://www.ibm.com/docs/en/was/8.5.5?topic=modules-ejb-content-in-war
--
Archived edits comments refer to:
Update-
Is the ejb.jar in the build class path of your .war? If your project is maven based, is the ejb.jar a dependency of the .war project? If so, try adding the EJB module as a "provided" dependency per https://stackoverflow.com/a/42847318/8528208
--
"An .ear file is a collection of entities (i.e., .war files), so you can simply extract the .war file from the .ear file and deploy it."
-per https://stackoverflow.com/a/26658071
If this doesn't help, can you add a lookup attribute in your #EJB annotations such as
#EJB(lookup="java:global/mwf_ejb/UserManager")
Similar to this answer?
https://stackoverflow.com/a/10156279/8528208
I'm writing an answer to my own question, since I solved the issue. It is perfectly fine to use EJBs with a WAR only (no EAR) as stated here https://docs.oracle.com/cd/E19798-01/821-1841/gippi/index.html.
As for the lookups calls, the correct way to do it is described here https://docs.oracle.com/cd/E19798-01/821-1841/girgn/index.html.
The issue I was facing wasn't really related to the lookups, as I was thinking, but it was due to the way most of the EJBs were initialized.
I noticed that in most of the classes there was some nasty initialization code like the following:
#Stateless
public class FooResource
{
FooEjb fooEjb = lookupFooEjb();
private FooEjb lookupFooEjb()
{
javax.naming.Context c = new InitialContext();
return (FooEjb) c.lookup("java:global/FooApplication/FooModule/FooEJB!FooInterface");
}
}
This works fine with the EAR because the EJB module is loaded before the WAR archives, so the lookups do not fail.
My guess is that when you package the EJBs with the WAR, it loads the EJBs as they are needed, computing the dependencies based on the #EJB annotation, so that kind of initialization fails since the EJB might be not loaded yet. To make it work, I just removed the lookup and added the annotation #EJB.
#Stateless
public class FooResource
{
#EJB
FooEjb fooEjb;
}
I get this error in a Java maven project. The weird thing is, it doesn't appear on every machine so I assume it has something to do with a configuration issue.
The class RoleKeyCacheImpl is a #Startup #Singleton:
#Startup
#Singleton
public class RoleKeyCacheImpl implements RoleKeyCache { ... }
That's the error Wildfly triggers when deploying the service.
Caused by: java.lang.IllegalArgumentException: WFLYEE0040: A component
named 'RoleKeyCacheImpl' is already defined in this module at
org.jboss.as.ee.component.EEModuleDescription.addComponent(EEModuleDescription.java:167)
at
org.jboss.as.ejb3.deployment.processors.EJBComponentDescriptionFactory.addComponent(EJBComponentDescriptionFactory.java:58)
I've tried:
installing a new Wildfly (V10, V13) on the same machine -> doesn't help
installing a completely new Eclipse on this machine -> doesn't help
cleaning & rebuilding all related projects
making sure the deployments-folder is empty and doesn't contain old versions of the same WAR
read the related question here which also didn't help (they use Spring): A component named 'XXX' is already defined in this module in JBoss 7.1.1
read and tried this q&a: Wrong dependencies with EJB in JBoss Wildfly (server-clean) -> doesn't help
deleted and rebuilt the local maven rep (".m2") -> no effect
checking out the same source on another computer -> does work on one machine, on another it gives the same error
I have absolutely no clue what the issue is or even could be. On one machine, we check it out and it runs without errors. On others, the exact same error happens.
Does anybody have an idea?
I had this same issue multiple times with EAP 7.1 and now again with WildFly 21.0.0. I know by experience this is an issue caused by Eclipse who tries to deploy automatically to a configured WildFly instance. During the deployment (or undeployment) some concurrent file issue arises and files who should be removed, are still on the filesystem, causing this error that a component is already defined.
In fact it is not already defined, it is just WildFly that is confused because it finds in his temporary directories some old files which shouldn't be there and reference your exact same component.
Solution: remove in the WildFly standalone directory the content in the 'deployments' directory and the 'tmp' directory. Rest assured, all what is there is okay to remove safely. Reboot and the error message will be gone ;-)
You should pay attention to not have two #Stateless EJB annotations on top of two classes with the same name - in the same module.
You may differentiate them by using the name attribute in the annotation and put different values in each class
Looks like the class already exists. Check if it does...you may have to rewrite that part of EEModuleDescription to use its own private methods (which would be what you would write) rather than overriding methods in RoleKeyCacheImpl. If the class actually does not exist then right-click on the project -> Maven 2 Tools -> Generate Eclipse Artifacts (Check for Updates). That will regenerate all of the dependencies that the project uses. Also please be sure that you have not added any new projects to the classpath by mistake as that may also cause this error.
I just ran into this today when a colleague added a maven dependency.
Turns out this dependency was a jar with a nasty classpath entry or "../" in the manifest.
I edited the jar's manifest.mf that was cached in my local maven repository using 7-zip and removed the "../" classpath entry.
Then re-packaged my war file (maven clean install) and bingo, it works!
In my case it was caused by org.libreoffice jurt version 5.4.2 (but other versions I checked also have the classpath nastiness).
Unfortunately I was lucky we pinpointed it to a dependency, YMMV!
I use GlassFish 4 web profile and I have the following interface and class.
#Local
public interface SomeService {
...
}
#Singleton
public class SomeServiceBean implements SomeService {
...
}
When I put interface and class in .war archive (that is in domain1/autodeplay) everything works fine. However, when I put interface and class in separate .jar archive (that is in domain1/lib) then deploying war application I get:
java.lang.RuntimeException: Cannot resolve reference Local ejb-ref name=com.temp.MyServlet/someService,Local 3.x interface =com.temp.SomeService,ejb-link=null,lookup=,mappedName=,jndi-name=,refType=Session
at com.sun.enterprise.deployment.util.ComponentValidator.accept(ComponentValidator.java:374) ~[dol.jar:na]
at com.sun.enterprise.deployment.util.DefaultDOLVisitor.accept(DefaultDOLVisitor.java:78) ~[dol.jar:na]
at com.sun.enterprise.deployment.util.ComponentValidator.accept(ComponentValidator.java:123) ~[dol.jar:na]
at com.sun.enterprise.deployment.util.ApplicationValidator.accept(ApplicationValidator.java:152) ~[dol.jar:n
...
I don't use any xml descriptors. So, is it possible to have EJBs in domain1/lib and if yes, how to make EJB container find them? P.S. I tried in GF 4 full - result is the same.
EJBs cannot be added as a library to GlassFish, libraries are just added to the classpath and any annotations on them are ignored and they do not go through the EJB container. If you do want your EJBs as a seperate JAR, they can be deployed just like a WAR or EAR file.
In the Glassfish reference manual for the add-library command it says that it "adds the library to the class loader directory", while for the deploy command it says that "Applications can be...EJB modules".
Also by looking at the source code for Glassfish it can be worked out that all libraries are simply added to the Classloader either at launch (See here and here) or if in applibs then when the application is deployed (See here).
I've written some code to create and run webservice client using CXF. I used JaxWsClientFactoryBean (not sure it's the best solution) to create client from .wsdl file.
The goal here was to do this programmatically avoiding Spring etc. Just pure code with Java and CXF.
JaxWsClientFactoryBean cfb = new JaxWsClientFactoryBean();
cfb.setAddress(getServiceProperty(intClass, PROPERTY_KEY_URL_SUFFIX));
cfb.setServiceClass(intClass);
cfb.setOutInterceptors(getOutInterceptors(intClass));
cfb.setServiceName(SERVICE_NAME);
cfb.setWsdlURL("classpath:wsdl/" + intClass.getSimpleName() + ".wsdl");
cfb.setEndpointName(ENDPOINT_NAME);
Client client = cfb.create();
ClientProxy cp = new ClientProxy(client);
I intService = (I)
Proxy.newProxyInstance(intClass.getClassLoader(), new Class[] { intClass }, cp);
I'm really not sure if this is done correctly, but it works when I run this code locally and when I deploy it on Tomcat.
Unfortunatelly I need to run this code on Weblogic and this results in strange exception:
Caused by: javax.wsdl.WSDLException: WSDLException: faultCode=PARSER_ERROR: org.w3c.dom.DOMException: HIERARCHY_REQUEST_ERR: An attempt was
made to insert a node where it is not permitted.
at org.apache.cxf.wsdl11.WSDLManagerImpl.loadDefinition(WSDLManagerImpl.java:235)
at org.apache.cxf.wsdl11.WSDLManagerImpl.getDefinition(WSDLManagerImpl.java:186)
at org.apache.cxf.wsdl11.WSDLServiceFactory.<init>(WSDLServiceFactory.java:92)
... 26 more
Caused by: org.w3c.dom.DOMException: HIERARCHY_REQUEST_ERR: An attempt was made to insert a node where it is not permitted.
at com.sun.org.apache.xerces.internal.dom.ParentNode.internalInsertBefore(ParentNode.java:356)
at com.sun.org.apache.xerces.internal.dom.ParentNode.insertBefore(ParentNode.java:284)
at com.sun.org.apache.xerces.internal.dom.CoreDocumentImpl.insertBefore(CoreDocumentImpl.java:399)
at com.sun.org.apache.xerces.internal.dom.NodeImpl.appendChild(NodeImpl.java:235)
at org.apache.cxf.staxutils.StaxUtils.readDocElements(StaxUtils.java:1019)
at org.apache.cxf.staxutils.StaxUtils.readDocElements(StaxUtils.java:939)
at org.apache.cxf.staxutils.StaxUtils.read(StaxUtils.java:866)
at org.apache.cxf.wsdl11.WSDLManagerImpl.loadDefinition(WSDLManagerImpl.java:226)
... 28 more
This happens during application deployment. It looks like there is something wrong with .wsdl file, but wait... It was working on Tomcat!
I think that there could be some difference in com.sun.org.apache.xerces.* classes implementation within Weblogic with its JRockit VM and standard JVM, but I have no idea how to solve it.
I spent many hours trying differend ways of client creation. Most of them worked locally and in Tomcat, but none on WebLogic.
Any hints what to try next? I'm kinda tired of this topic :D
I agree with your suspicion that the problem is related to the used version of Xerces. The stacktrace shows that the Sun implementation of Xerces which is derivative of the Apache Xerces is used in your case.
Please check the Apache CFX Application Server Configuration Guide instructions related to WebLogic.
WebLogic ClassLoading
In WebLogic Server, any .jar file present in the system classpath is loaded by the WebLogic Server system classloader. All applications running within a server instance are loaded in application classloaders which are children of the system classloader. In this implementation of the system classloader, applications cannot use different versions of third-party jars which are already present in the system classloader. Every child classloader asks the parent (the system classloader) for a particular class and cannot load classes which are seen by the parent.
For example, if a class called com.foo.Baz exists in both $CLASSPATH as well as the application EAR, then the class from the $CLASSPATH is loaded and not the one from the EAR. Since weblogic.jar is in the $CLASSPATH, applications can not override any WebLogic Server classes.
In order to use an alternate version of Xerces you have to create a FilteringClassLoader.
Usage of FilteringClassLoader
The FilteringClassLoader provides a mechanism for you to configure deployment descriptors to explicitly specify that certain packages should always be loaded from the application, rather than being loaded by the system classloader. This allows you to use alternate versions of applications such as Xerces and Ant.
The FilteringClassLoader sits between the application classloader and the system. It is a child of the system classloader and the parent of the application classloader. The FilteringClassLoader intercepts the loadClass(String className) method and compares the className with a list of packages specified in weblogic-application.xml file.
In conclusion, check the steps included in the Apache CFX Application Server Configuration Guide and take care to explicitly specify that the org.apache.xerces.* package is loaded from the application, rather than being loaded from the system classloader.
For example the weblogic-application.xml file in the META-INF should look like:
<?xml version="1.0" encoding="UTF-8"?>
<weblogic-application xmlns="http://www.bea.com/ns/weblogic/90">
<application-param>
<param-name>webapp.encoding.default</param-name>
<param-value>UTF-8</param-value>
</application-param>
<prefer-application-packages>
<package-name>javax.jws.*</package-name>
<package-name>org.apache.xerces.*</package-name>
</prefer-application-packages>
</weblogic-application>
I hope this helps.
I have a project with a EJB implementing a JWS Webservice, like this:
#Stateless
#Remote(WebServiceTest.class)
#WebService(
serviceName="TestService",
name="TestName"
)
public class WebServiceTestImpl implements WebServiceTest {
#Override
#WebMethod(operationName="hello")
public String hello() {
return "Hello World!";
}
}
I deploy and test this perfectly on WebLogic 10.3 using a simple EAR project. Now I need to use Hibernate on my project, so from previous experiences I know that I have to use Antrl from Hibernate and not from the container, so I create a weblogic-application.xml in the EAR project:
<?xml version="1.0" encoding="UTF-8"?>
<wls:weblogic-application ...>
<wls:prefer-application-packages>
<wls:package-name>antlr.*</wls:package-name>
</wls:prefer-application-packages>
</wls:weblogic-application>
Now when I deploy the EAR I get this error:
Unable to deploy EJB: WebServiceTestImpl from test-1.0.0-SNAPSHOT.jar:
***** ASSERTION FAILED *****
at weblogic.ejb.container.deployer.EJBModule.prepare(EJBModule.java:467)
at weblogic.application.internal.flow.ModuleListenerInvoker.prepare(ModuleListenerInvoker.java:199)
at weblogic.application.internal.flow.DeploymentCallbackFlow$1.next(DeploymentCallbackFlow.java:507)
at weblogic.application.utils.StateMachineDriver.nextState(StateMachineDriver.java:41)
at weblogic.application.internal.flow.DeploymentCallbackFlow.prepare(DeploymentCallbackFlow.java:149)
Truncated. see log file for complete stacktrace
Caused By: java.lang.ClassNotFoundException: test.WebServiceTestImpl_zd33dy_WSOImpl
at weblogic.utils.classloaders.GenericClassLoader.findLocalClass(GenericClassLoader.java:280)
at weblogic.utils.classloaders.GenericClassLoader.findClass(GenericClassLoader.java:253)
at java.lang.ClassLoader.loadClass(ClassLoader.java:307)
at java.lang.ClassLoader.loadClass(ClassLoader.java:248)
at weblogic.utils.classloaders.GenericClassLoader.loadClass(GenericClassLoader.java:177)
Truncated. see log file for complete stacktrace
How can I solve this? If I don't put Antlr in wls:prefer-application-packages Hibernate will not work, and if I put it I cannot deploy the webservice...
Weblogic somehow keeps precompiled versions of EJBs even though you do a new/update deployment.
this works for me with wls 10.3.5:
within the server directory of your domain (e.g. in my case /home/myUser/apps/bea/domains/myDomain/servers/AdminServer) there is a subfolder "cache/EJBCompilerCache". Stop your weblogic instance and remove the contents of this subfolder. Restart weblogic and everything will work :-)
I'm not sure if this is the same bug, but there is a known bug in Weblogic 10.3.0-2 that causes a problem when using JaxWS with an EJB v 2.x. But I'm getting a near identical error when I put the #WebService annotation on an EJB v 3.0 on Weblogic 10.3.1. The error is loading what looks like the supposedly generated local stubs that Weblogic creates, and its happening on Weblogic 10.3.0, 10.3.1, 10.3.2, and 10.3.3. (yes, I really have them all deployed ;)
I had this problem deploying session bean and then trying to convert it to web service. WebLogic seems to cache session beans somehow.
There are several solution to this:
- Rename slsb.
- Comment #Stateless and #Remote annotations, deploy application, uncomment annotations, redeploy again.
- Bounce WebLogic.