You are viewing a plain text version of this content. The canonical link for it is here.
Posted to commits@tomee.apache.org by "Nuno G. de M (JIRA)" <ji...@apache.org> on 2015/05/15 14:59:00 UTC
[jira] [Comment Edited] (TOMEE-1580) Datasource JNDI Name Context
not available to eclipselink non jta data source
[ https://issues.apache.org/jira/browse/TOMEE-1580?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14545442#comment-14545442 ]
Nuno G. de M edited comment on TOMEE-1580 at 5/15/15 12:58 PM:
---------------------------------------------------------------
SAMPLE APP.
https://drive.google.com/open?id=0B_dEiNBGUsxqaC0zMS0wZVRFaU0&authuser=0
Please download the zip file form my google drive.
Enclosed in the zip file you will have:
(1) A docx that briefly describes the problem
(2) A sample application
The sample application is a Maven Module of the Form:
-- ROOT
---- CORE-EJB.jar
---- MAIN-EJB.jar
---- WAR.war
The CORE-EJB.jar
This stuff has a recursive entity in there, and some EJBs in there to do osme CRUD operations, you do not have to be concerned wit the contest of the .jar file -they are there for a different purpose - for a different issue.
What is relevant is only the fact that there is a tiny entity in there.
That's it.
The MAIN-EJB.jar is just glue. It pretends that you ahve some big project that useds different components and it bundles the persistence.xml used by the project. In terms of code it only has one interest class, db.setup.DBSetup.java
It is nothing but a @Singleton ejb that runs on @Startup and that uses a NON_JTA data source. The persistence unit for this data source has those create db properties that will populate a db schema, that's it.
What is most relevant about this jar file is the persistnece.xml that will be scanned by the container.
We have two bugs here.
The first bug is that for the application to work, the main persistence.unit
needs to say this:
<property name="eclipselink.jdbc.sequence-connection-pool.non-jta-data-source" value="java:openejb/Resource/jdbc/ORCL_NON_XA" />
In weblogic and glassfish the application would work with:
<property name="eclipselink.jdbc.sequence-connection-pool.non-jta-data-source" value="jdbc/ORCL_NON_XA" />
The second bug is that the <jar-file> references need to get ../ for the container to locate the CORE-EJB.jar.
So, if deploy the war file that is provided in the link you will be able to go to:
http://localhost:8080/corba-serilization-infinite-loop-war/index.xhtml
Click on the button and write some entiteis into the db.
You need a DB to run the thing, for example if you are using derby, you could have
My tomee.xml looks as follows:
<Resource id="jdbc/ORCL_XA" type="DataSource" classpath="D:/APPS/apache-tomee-plume-1.7.1/ext/derbyclient.jar" >
#Embedded Derby example
JdbcDriver = org.apache.derby.jdbc.ClientDriver
JdbcUrl = jdbc:derby://localhost:1527/ORCL;create=true
UserName = ORCL
Password = ORCL
InitialSize = 7
MaxIdle = 20
MaxActive = 20
JtaManaged = true
</Resource>
<Resource id="jdbc/ORCL_NON_XA" type="DataSource" classpath="D:/APPS/apache-tomee-plume-1.7.1/ext/derbyclient.jar" >
#Embedded Derby example
JdbcDriver = org.apache.derby.jdbc.ClientDriver
JdbcUrl = jdbc:derby://localhost:1527/ORCL;create=true
UserName = ORCL
Password = ORCL
InitialSize = 1
MaxIdle = 2
MaxActive = 5
JtaManaged = false
</Resource>
<!-- JMS Resource Adapter -->
<Container id="MY_JMS_CONTAINER" ctype="MESSAGE">
ResourceAdapter = ActiveMQResourceAdapter
</Container>
<Resource id="ActiveMQResourceAdapter" type="ActiveMQResourceAdapter">
BrokerXmlConfig = broker:(tcp://localhost:7676)
ServerUrl = tcp://localhost:7676
</Resource>
<!-- Connection Factory. -->
<Resource id="jms/NotificationConnectionFactory" type="javax.jms.ConnectionFactory">
ResourceAdapter = ActiveMQResourceAdapter
PoolMaxSize = 250
PoolMinSize = 1
ConnectionMaxWaitMilliseconds = 10000
ConnectionMaxIdleMinutes = 5
</Resource>
<Resource id="jms/QueueJmsFactory" type="javax.jms.ConnectionFactory">
ResourceAdapter = ActiveMQResourceAdapter
PoolMaxSize = 250
PoolMinSize = 2
ConnectionMaxWaitMilliseconds = 10000
ConnectionMaxIdleMinutes = 5
</Resource>
<!-- Queues -->
<Resource id="queue/TinyQueueA" type="javax.jms.Queue">
destination = queue_TinyQueueA
</Resource>
<Resource id="queue/TinyQueueB" type="javax.jms.Queue">
destination = queue_TinyQueueA
</Resource>
Note the JMS queues and connection factories for the time being are irrelevant.
I need to open another bug for them, but I have to find the time to make some MdBs and simulate the configuration.
Many thanks for your support,
Kindest regards,
Nuno.
was (Author: ngdem):
BUG 1:
https://drive.google.com/open?id=0B_dEiNBGUsxqaC0zMS0wZVRFaU0&authuser=0
Please download the zip file form my google drive.
Enclosed in the zip file you will have:
(1) A docx that briefly describes the problem
(2) A sample application
The sample application is a Maven Module of the Form:
-- ROOT
---- CORE-EJB.jar
---- MAIN-EJB.jar
---- WAR.war
The CORE-EJB.jar
This stuff has a recursive entity in there, and some EJBs in there to do osme CRUD operations, you do not have to be concerned wit the contest of the .jar file -they are there for a different purpose - for a different issue.
What is relevant is only the fact that there is a tiny entity in there.
That's it.
The MAIN-EJB.jar is just glue. It pretends that you ahve some big project that useds different components and it bundles the persistence.xml used by the project. In terms of code it only has one interest class, db.setup.DBSetup.java
It is nothing but a @Singleton ejb that runs on @Startup and that uses a NON_JTA data source. The persistence unit for this data source has those create db properties that will populate a db schema, that's it.
What is most relevant about this jar file is the persistnece.xml that will be scanned by the container.
We have two bugs here.
The first bug is that for the application to work, the main persistence.unit
needs to say this:
<property name="eclipselink.jdbc.sequence-connection-pool.non-jta-data-source" value="java:openejb/Resource/jdbc/WM6_NON_JTA_DS" />
In weblogic and glassfish the application would work with:
<property name="eclipselink.jdbc.sequence-connection-pool.non-jta-data-source" value="jdbc/WM6_NON_JTA_DS" />
The second bug is that the <jar-file> references need to get ../ for the container to locate the CORE-EJB.jar.
So, if deploy the war file that is provided in the link you will be able to go to:
http://localhost:8080/corba-serilization-infinite-loop-war/index.xhtml
Click on the button and write some entiteis into the db.
You need a DB to run the thing, for example if you are using derby, you could have
My tomee.xml looks as follows:
<Resource id="jdbc/ORCL_XA" type="DataSource" classpath="D:/APPS/apache-tomee-plume-1.7.1/ext/derbyclient.jar" >
#Embedded Derby example
JdbcDriver = org.apache.derby.jdbc.ClientDriver
JdbcUrl = jdbc:derby://localhost:1527/ORCL;create=true
UserName = ORCL
Password = ORCL
InitialSize = 7
MaxIdle = 20
MaxActive = 20
JtaManaged = true
</Resource>
<Resource id="jdbc/ORCL_NON_XA" type="DataSource" classpath="D:/APPS/apache-tomee-plume-1.7.1/ext/derbyclient.jar" >
#Embedded Derby example
JdbcDriver = org.apache.derby.jdbc.ClientDriver
JdbcUrl = jdbc:derby://localhost:1527/ORCL;create=true
UserName = ORCL
Password = ORCL
InitialSize = 1
MaxIdle = 2
MaxActive = 5
JtaManaged = false
</Resource>
<!-- JMS Resource Adapter -->
<Container id="MY_JMS_CONTAINER" ctype="MESSAGE">
ResourceAdapter = ActiveMQResourceAdapter
</Container>
<Resource id="ActiveMQResourceAdapter" type="ActiveMQResourceAdapter">
BrokerXmlConfig = broker:(tcp://localhost:7676)
ServerUrl = tcp://localhost:7676
</Resource>
<!-- Connection Factory. -->
<Resource id="jms/NotificationConnectionFactory" type="javax.jms.ConnectionFactory">
ResourceAdapter = ActiveMQResourceAdapter
PoolMaxSize = 250
PoolMinSize = 1
ConnectionMaxWaitMilliseconds = 10000
ConnectionMaxIdleMinutes = 5
</Resource>
<Resource id="jms/QueueJmsFactory" type="javax.jms.ConnectionFactory">
ResourceAdapter = ActiveMQResourceAdapter
PoolMaxSize = 250
PoolMinSize = 2
ConnectionMaxWaitMilliseconds = 10000
ConnectionMaxIdleMinutes = 5
</Resource>
<!-- Queues -->
<Resource id="queue/TinyQueueA" type="javax.jms.Queue">
destination = queue_TinyQueueA
</Resource>
<Resource id="queue/TinyQueueB" type="javax.jms.Queue">
destination = queue_TinyQueueA
</Resource>
Note the JMS queues and connection factories for the time being are irrelevant.
I need to open another bug for them, but I have to find the time to make some MdBs and simulate the configuration.
Many thanks for your support,
Kindest regards,
Nuno.
> Datasource JNDI Name Context not available to eclipselink non jta data source
> -----------------------------------------------------------------------------
>
> Key: TOMEE-1580
> URL: https://issues.apache.org/jira/browse/TOMEE-1580
> Project: TomEE
> Issue Type: Bug
> Affects Versions: 1.7.1
> Environment: Windows 7, 64 bit, JDK 7 Sun Hotspot
> Reporter: Nuno G. de M
> Priority: Minor
> Labels: JNDI, eclipselink, naming-exception, openejb, persistence.xml, tomee
> Fix For: 1.7.2
>
>
> Hi,
> I am trying to experiement running an application of relative large size on tomcat ee. The application deploys both to Glassfish 3.1.2.X and weblogic 12.1.2. - Just to have a reference point as to application compatibility.
> When deploying to tomee plume version 1.7.1 one of the problems I have encountered has to do with the JNDI Naming contexts visible to the application during deployment.
> Namely, the persistence.xml will be telling eclipse link to manage the generation of new entity sequence numbers using a non JTA data source.
> For example,
> <persistence-unit name="MY_PU" transaction-type="JTA">
> <provider>org.eclipse.persistence.jpa.PersistenceProvider</provider>
> <jta-data-source>jdbc/JTA_DS</jta-data-source>
> <jar-file>../some.jar</jar-file> <------- this is the first problem. Was forced to have the ../ added to the jar file
>
> <exclude-unlisted-classes>false</exclude-unlisted-classes>
> <shared-cache-mode>DISABLE_SELECTIVE</shared-cache-mode>
> <properties>
> <property name="eclipselink.logging.level" value="INFO" />
> <property name="eclipselink.logging.level.sql" value="INFO" />
> <property name="eclipselink.logging.parameters" value="true" />
> <property name="eclipselink.jdbc.sequence-connection-pool" value="true" />
> <property name="eclipselink.jdbc.sequence-connection-pool.non-jta-data-source" value="jdbc/NON_JTA_DS" /> <------------- Eclipse link will get a naming exception here
> </properties>
> </persistence-unit>
> Ok.
> Now with the above "sample" persistence.xml in mind, what happens is the following.
> PROBLEM 1:
> The <jar-file> defitnition on the persistence.xml was forced to be tuned with ../ - whereas in weblogic and glassfish this was not necessary.
> The application is bundled in a War file that in the lib/ folder contians all the WAR dependencies (ejbs and such). One of these jar files - let us call it my-main.jar is the one that in the classes/META-INF/persistence.xml holds the persistence.xml referring to all other jars interesting in the /lib/ of the war file.
> Under Tomee this the search for the mentioned jar files that are relevant to the persistence unit seems to bugy. Since tomee is blowing up with file not found exception:
> my-main.jar/some.jar
> And of course he is right. The path he built to find the JAR was not correct and therefore he could not fild the <jar-file>.
> Therefore, here I was forced to put in the <jar-file>../some.jar<jar-file>
> to overcome the file not found exception.
> PROBLEM 2:
> Now the second problem, this is a bit more important, is in the JNDI naming context themsevels.
> Using:
> tomee-webaccess-1.7.1/scripting
> And the following script:
> var pairs = ctx.lookup('java:openejb/Resource/jdbc').list(''); <----- This is key.
> while(pairs.hasMore()) {
> println('[entry] ' + pairs.next().getName());
> }
> var user = ctx.lookup('java:comp/EJBContext').getCallerPrincipal().getName();
> println("Logged user: '" + user + "'");
> I am able to verify that the two data sources I defined in the tomee.xml are part of the openejb Resource naming context.
> Which is OK.
> So the list of names under jdbc/ shows me both the JTA_DS and the NON_JTA_DS.
> And indeed the @PersistenceContext inject for the JTA_DS seems to be working fine.
> But eclipse link will blow up when it tries to the JNDI lookup for jdbc/NON_JTA_DS as listed in the following exception snippet.
> Internal Exception: javax.naming.NameNotFoundException: Name [jdbc/NON_JTA_DS] is not bound in this Context. Unable to find [jdbc].
> at org.eclipse.persistence.exceptions.EntityManagerSetupException.deployFailed(EntityManagerSetupException.java:229)
> ... 58 more
> Caused by: Exception [EclipseLink-7060] (Eclipse Persistence Services - 2.4.2.v20130514-5956486): org.eclipse.persistence.exceptions.ValidationException
> Exception Description: Cannot acquire data source [jdbc/NON_JTA_DS].
> Internal Exception: javax.naming.NameNotFoundException: Name [jdbc/NON_JTA_DS] is not bound in this Context. Unable to find [jdbc].
> at org.eclipse.persistence.exceptions.ValidationException.cannotAcquireDataSource(ValidationException.java:502)
> at org.eclipse.persistence.sessions.JNDIConnector.connect(JNDIConnector.java:109)
> at org.eclipse.persistence.sessions.DatasourceLogin.connectToDatasource(DatasourceLogin.java:162)
> at org.eclipse.persistence.internal.databaseaccess.DatasourceAccessor.connectInternal(DatasourceAccessor.java:330)
> at org.eclipse.persistence.internal.databaseaccess.DatabaseAccessor.connectInternal(DatabaseAccessor.java:295)
> at org.eclipse.persistence.internal.databaseaccess.DatasourceAccessor.connect(DatasourceAccessor.java:418)
> at org.eclipse.persistence.sessions.server.ConnectionPool.buildConnection(ConnectionPool.java:216)
> at org.eclipse.persistence.sessions.server.ExternalConnectionPool.startUp(ExternalConnectionPool.java:146)
> at org.eclipse.persistence.internal.sequencing.ServerSessionConnectionHandler.onConnect(ServerSessionConnectionHandler.java:27)
> at org.eclipse.persistence.internal.sequencing.SequencingManager.onConnectInternal(SequencingManager.java:770)
> at org.eclipse.persistence.internal.sequencing.SequencingManager.onConnect(SequencingManager.java:700)
> at org.eclipse.persistence.internal.sessions.DatabaseSessionImpl.initializeSequencing(DatabaseSessionImpl.java:189)
> at org.eclipse.persistence.internal.sessions.DatabaseSessionImpl.initializeDescriptors(DatabaseSessionImpl.java:583)
> at org.eclipse.persistence.internal.sessions.DatabaseSessionImpl.initializeDescriptors(DatabaseSessionImpl.java:579)
> at org.eclipse.persistence.internal.sessions.DatabaseSessionImpl.initializeDescriptors(DatabaseSessionImpl.java:534)
> at org.eclipse.persistence.internal.sessions.DatabaseSessionImpl.postConnectDatasource(DatabaseSessionImpl.java:782)
> at org.eclipse.persistence.internal.sessions.DatabaseSessionImpl.loginAndDetectDatasource(DatabaseSessionImpl.java:727)
> at org.eclipse.persistence.internal.jpa.EntityManagerFactoryProvider.login(EntityManagerFactoryProvider.java:215)
> at org.eclipse.persistence.internal.jpa.EntityManagerSetupImpl.deploy(EntityManagerSetupImpl.java:554)
> ... 56 more
> Caused by: javax.naming.NameNotFoundException: Name [jdbc/NON_JTA_DS] is not bound in this Context. Unable to find [jdbc]. <---------------- OK so eclipse link is operating in some sort of container context that does no see the JDBC name
> at org.apache.naming.NamingContext.lookup(NamingContext.java:819)
> at org.apache.naming.NamingContext.lookup(NamingContext.java:153)
> at org.apache.naming.SelectorContext.lookup(SelectorContext.java:133)
> at javax.naming.InitialContext.lookup(InitialContext.java:415)
> at org.eclipse.persistence.sessions.JNDIConnector.connect(JNDIConnector.java:103)
> ... 73 more
>
> To me this looks like a serious bug, but you will know better than me the standard itself and the expectations that an application is allowed to have on the available naming contexts in the container.
> But I was under the impression that regardless of the application module, the jdbc name should always be visible at the root of the initial naming context.
> I am not sure if it will succeed or not, but for the time being I will try to work around this issue by using the following JNDI name:
> java:openejb/Resource/jdbc/NON_JTA_DS
> Since i can use the scripting tool the validate that this name can bee lookedup.
> I am not satisfied with the solution - i think of it is a hammer pattern approach and injection using the easy name : jdbc/NON_JTA_DS should just work.
> It works for weblogic and glassfish 3.1.2 at least so I would assume it should work as well for tomee.
> Kindest regards,
> And thanks for the support.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)