Notes and Problems


	BEA Home \| Events \| Solutions \| Partners \| Products \| Services \| Download \| Developer Center \| WebSUPPORT
	http://www.oracle.com/technology/documentation/index.html \| Site Map \| Search \| Download Documentation \| Contact \| Glossary

\| WebLogic Server Doc Home \| Release Notes \| Previous Topic \| Next Topic \| Contents \| View as PDF

Notes and Problems

The following sections describe known problems with this software release.

Connector Known Problems

Change Request Number	Description
098342	The WebLogic Server 6.1 JCA implementation is not compatible with the IBM CICS resource adapter. BEA's implementation of the javax.resource.spi.ConnectionManager interface is not directly returning the object returned by the getConnection() method of the resource adapter's implementation of the javax.resource.spi.ManagedConnection interface. It returns an instance of a class dynamically generated by the java.lang.reflect.Proxy class instead.

Change Request Number

Description

098342

The WebLogic Server 6.1 JCA implementation is not compatible with the IBM CICS resource adapter. BEA's implementation of the javax.resource.spi.ConnectionManager interface is not directly returning the object returned by the getConnection() method of the resource adapter's implementation of the javax.resource.spi.ManagedConnection interface. It returns an instance of a class dynamically generated by the java.lang.reflect.Proxy class instead.

Console Known Problems

Change Request Number	Description
053277	If you use Netscape to bring up the Administration Console and try to edit a template for an LDAP Realm configuration, using the arrows or the bottom scroll bar to go left or right in the Configuration Data window causes the displayed data to look garbled. If you click the Apply button, no corrupt data is saved in config.xml, but it is difficult to edit the template.

Core Known Problems

Change Request Number	Description
CR093104	In SP04 and SP05, the NT Performance pack causes thread deadlock under some circumstances. The problem was exhibited under light loads during stress tests of a .jsps accessing a database. Thread dumps showed the majority of "default" execute threads stuck in "waiting for monitor entry": "ExecuteThread: '10' for queue: 'default'" daemon prio=5 tid=0x273fb548 nid=0xadc waiting for monitor entry [0x2810f000..0x2810fdc4] at weblogic.socket.NTSocketMuxer.initiateIO(NTSocketMuxer.java:483) at weblogic.socket.NTSocketMuxer.read(NTSocketMuxer.java:474) at weblogic.servlet.internal.MuxableSocketHTTP.requeue(MuxableSocketHTTP.java:244) at weblogic.servlet.internal.ServletResponseImpl.send(ServletResponseImpl.java:1094) at weblogic.servlet.internal.ServletRequestImpl.execute(ServletRequestImpl.java:2364) at weblogic.kernel.ExecuteThread.execute(ExecuteThread.java:139) at weblogic.kernel.ExecuteThread.run(ExecuteThread.java:120) Customers experience such deadlocks should create a .hotspot_compiler file in the directory where you invoke the JVM and put the following line in that file : exclude weblogic/socket/NTSocketMuxer processSockets
CR106616	There is a problem in Netscape Version 4.79 running on Linux AS2.1 and HP-UX. After several operations on the WebLogic Server Administration Console, the browser shuts down with a bus error, sometimes with a core dump.
CR179419	In WebLogic Server 6.1, when a Managed Server is configured to use the pure-Java socket reader implementation, rather than native socket readers, execute threads are allocated to act as socket reader threads. You configure the pure-Java socket reader implementation by setting the the server instance's NativeIOEnabled attribute to false, and the number of threads that act as socket reader threads is a function of the server instance's ThreadPoolPercentSocketReaders value and the execute queue's ThreadCount attribute. Under these circumstances, the value of ThreadCount is not honored. The server instance issues the following warning before the configured thread count is reached: <Apr 22, 2004 8:42:54 PM JST> <Warning> <JavaSocketMuxer> <There are: '3' active sockets, but the maximum number of socket readers allowed by the configuration is: '2', you may want to alter your configuration.> This problem results from an error in thread pool initialization logic. The problem does not occur for Administration Servers. The work-around for this problem is to assign the thread pool size using this command line option when starting the Managed Server: -Dweblogic.ThreadPoolSize=xxx where xxx is the desired thread count. For example: java ... -Dweblogic.ThreadPoolSize=15 ... weblogic.Server

Change Request Number

Description

CR093104

In SP04 and SP05, the NT Performance pack causes thread deadlock under some circumstances. The problem was exhibited under light loads during stress tests of a .jsps accessing a database. Thread dumps showed the majority of "default" execute threads stuck in "waiting for monitor entry":

"ExecuteThread: '10' for queue: 'default'" daemon prio=5 tid=0x273fb548 nid=0xadc waiting for monitor entry [0x2810f000..0x2810fdc4]

at weblogic.socket.NTSocketMuxer.initiateIO(NTSocketMuxer.java:483)

at weblogic.socket.NTSocketMuxer.read(NTSocketMuxer.java:474)

weblogic.servlet.internal.MuxableSocketHTTP.requeue(MuxableSocketHTTP.java:244)

weblogic.servlet.internal.ServletResponseImpl.send(ServletResponseImpl.java:1094)

weblogic.servlet.internal.ServletRequestImpl.execute(ServletRequestImpl.java:2364)

at weblogic.kernel.ExecuteThread.execute(ExecuteThread.java:139)

at weblogic.kernel.ExecuteThread.run(ExecuteThread.java:120)

Customers experience such deadlocks should create a .hotspot_compiler file in the directory where you invoke the JVM and put the following line in that file :

exclude weblogic/socket/NTSocketMuxer processSockets

CR106616

There is a problem in Netscape Version 4.79 running on Linux AS2.1 and HP-UX.

After several operations on the WebLogic Server Administration Console, the browser shuts down with a bus error, sometimes with a core dump.

CR179419

In WebLogic Server 6.1, when a Managed Server is configured to use the pure-Java socket reader implementation, rather than native socket readers, execute threads are allocated to act as socket reader threads. You configure the pure-Java socket reader implementation by setting the the server instance's NativeIOEnabled attribute to false, and the number of threads that act as socket reader threads is a function of the server instance's ThreadPoolPercentSocketReaders value and the execute queue's ThreadCount attribute.

Under these circumstances, the value of ThreadCount is not honored. The server instance issues the following warning before the configured thread count is reached:

This problem results from an error in thread pool initialization logic. The problem does not occur for Administration Servers.

The work-around for this problem is to assign the thread pool size using this command line option when starting the Managed Server:

-Dweblogic.ThreadPoolSize=xxx

where xxx is the desired thread count.

For example:

java ... -Dweblogic.ThreadPoolSize=15 ... weblogic.Server

EJB Known Problems

Change Request Number	Description
035884	The maximum number of Enterprise JavaBeans (EJBs) that can be deployed on one WebLogic Server instance is 390. This is due to a Sun bug and is a limitation caused by the 1.3 JDK; it may be fixed by the 1.3.1 release of the JDK. The Sun Bug ID is: 4390238.
051543	If you try to deploy a completely incorrect EJB, the EJB container throws an exception from its deploy method. However, if you click on that EJB in the console, it still has the Deployed box checked.
061938	In certain conditions, EJB QL queries can return spurious duplicates that are the results of SQL cross products. This can occur under the following conditions: The EJB-QL query contains path expressions that navigate relationships; this generates multiple tables in the generated SQL SELECT clause. The WHERE clause contains OR operands which navigate relationships, and not all of the path expression in the OR operands map to all of the tables in the generated SQL SELECT clause for the query; this may cause a cross product to show up in the results for that OR operand. The following example illustrates the problem: EJB QL: SELECT OBJECT(c) FROM CustomerBean AS c, IN(c.accounts)accts WHERE c.name = '100' OR c.accts.bal = 300 DATA: customer '100' exists but has no accounts EXPECTED RESULT: customer '100' from clause #1 ACTUAL RESULT: customer '100' X number of accts SQL: DROP TABLE thorick_customers; CREATE TABLE thorick_customers ( cust_name VARCHAR(10), cust_interests VARCHAR(10), cust_rating INTEGER, acct_id INTEGER, PRIMARY KEY (cust_name)); DROP TABLE thorick_accounts; CREATE TABLE thorick_accounts (acct_id INTEGER, bal FLOAT, PRIMARY KEY (acct_id)); INSERT INTO thorick_accounts VALUES (100, 100.0); INSERT INTO thorick_accounts VALUES (200, 200.0); INSERT INTO thorick_accounts VALUES (300, 300.0); INSERT INTO thorick_accounts VALUES (400, 400.0); INSERT INTO thorick_accounts VALUES (500, 500.0); INSERT INTO thorick_customers VALUES('100', 'jazz', 2, null); INSERT INTO thorick_customers VALUES('900', 'punk', 3, 400); SELECT WL0.cust_name, WL0.cust_interests, WL0.acct_id FROM thorick_Customers WL0, thorick_Accounts WL1 WHERE WL0.cust_name = '100' OR ( wl1.bal = 300 AND wl0.acct_id=wl1.acct_id ); CUST_NAME CUST_INTER ACCT_ID ---------- ---------- ---------- 100 jazz 100 jazz 100 jazz 100 jazz 100 jazz
062927	Requesting an EJB served by WLE 5.1 from a JSP served by WLS 6.1 may give the following exception: java.lang.NoClassDefFoundError Add the location of the WLE 5.1 wlej2eecl.jar file to the CLASSPATH of your WLS 6.1 startWeblogic.cmd or startWeblogic.sh script.
079539	WebLogic Server 6.1 sp3 is not compatible with Application Integration in WLI 2.1 sp1. Please contact BEA Customer Support for more information.
088461	ejbc behavior varies between WebLogic Server 6.1 SP02 and SP03. In SP02, running ejbc on a bean before deploying it to the server worked as expected, while in SP03, the same action on the bean sometimes fails with an OutOfMemory exception if hot deployed without precompiling with ejbc. The workaround is to use: java -verbosegc weblogic.ejbc -compiler javac -J-mx128m entityBeans.jar final.jar
089847	Many of the WebLogic Server EJB deployment properties have default values that are optimized for performance. In some cases, these default values are not compliant with the EJB specification. Should you wish to make WebLogic Server compliant with the EJB specification, you must set the following properties: In weblogic-ejb-jar.xml set "enable call by reference" to False. In weblogic-cmp-jar.xml, set "include-updates" to True In weblogic-cmp-jar.xml, set "check exists on method" to True.
097323	There is a known issue with deploying or redeploying EJBs to a single server instance in a cluster—referred to as pinned deployment—if the .jar file contains contain uncompiled classes and interfaces. During deployment, the uncompiled EJB is copied to each server instance in the cluster, but it is compiled only on the server instance to which it has been deployed. As a result, the server instances in the cluster to which the EJB was not targeted lack the classes generated during compilation that are necessary to invoke the EJB. When a client on another server instance tries to invoke the pinned EJB, it fails, and an Assertion error is thrown in the RMI layer. Workaround: If you are deploying or redeploying an EJB to a single server instance in a cluster, compile the EJB with appc or ejbc before deploying it, , to ensure that the generated classes copied to all server instances available to all nodes in the cluster.
111022	When using a non-transaction-enabled DataSource, a "java.sql.SQLException: ORA-01002" error will occur when creating a CMP entity bean which has a field mapped to Oracle CLOB/BLOB. The EJB container inserts a row which has CLOB/BLOB columns as follows: insert into the table initializing the LOB using EMPTY_BLOB() or EMPTY_CLOB(). retrieve the LOB locator using select for update statement. modify the LOB. Although the connection from a non-transaction-enabled DataSource is in auto-commit mode, Oracle LOB locators that are used to write cannot span the transactions. So the transaction is closed at step 1 and then ORA-1002 occurs at step 2. Workaround: You need to use a transaction DataSource(TxDataSource) when using Oracle CLOB/BLOB with CMP.

Examples and Pet Store Demo Known Problems

Change Request Number	Description
043156	The ejbmanagedclient.jsp relies on a variable that is set to t3://localhost:7001 and does not get the port number dynamically. Thus, when the page is requested, it cannot find the server. The exception is thrown, although the stack trace is not printed out. The page shows up as: --------------UnexpectedError-------------- Customers running this example on a different port then 7001, will need to edit the jsp file. The variable you must edit is in this section: <!-- Here, we declare a class method --> <%! String pagetitle = "JSP example using EJBean-managed persistence"; String url = "t3://localhost:7001"; Edit the URL to point to the correct server/port number that has been installed.
044641	There is some missing information in examples/ejb/sequence/userDesignated/package-summary.html. Make sure that you use the table.ddl file, included in the example's package, as the source for the table that needs to be built for this example. Also, make sure that table is named "NAMED_SEQUENCE_TABLE" not "NAMED_SEQUENCE-TABLE".
047654	In the dbKona examples, a pool connection cannot be made and an error is thrown when trying to boot the example server.
047928	When running the Web Application security example, there is no link in the console for creating a group and adding members to it. Also, on the group window, clicking on Name or Members causes a 500 Internal Server Error.
048358	The security/rdbmsrealm/rdbmsrealm.ddl and utils/ddl/demo.ddl examples have problems with improper comment symbols.
052556	The WebLogic Tuxedo Connector CORBA simpappcns example fails when run in the examples domain after running the WLEC examples. The problem is a conflict with the Simpapp client stubs left in the SERVER_CLASSES directory from the WLEC examples. Before running the simpappcns example, check the SERVER_CLASSES environment variable. If the WLEC examples have been run, delete all simpapp client stubs from SERVER_CLASSES.
053356	The example samples/examples/wtc/corba/simpappcns does not have a run.sh UNIX script and will not work on UNIX machines. This problem was solved in WebLogic Server 6.1 SP04. See "053356".
062545	WebLogic Server domain running Solaris simpappcns can't call Tuxedo Corba Server domain running on NT. This problem was solved in WebLogic Server 6.1 SP03. See "062545".
062799	The documentation for three of the WIRELESS examples contains an error in Step 7 of the instructions. The examples in question are helloWorld, stockDemo and travelDemo. Step 7 should read: Path=./config/examples/helloWorld rather than: Path=./config/examples If uncorrected, the examples will not find the deployed applications.
064001	examples.webservices.message.package-summary.html contains an incorrect instruction. Step 13 says "Right-click the Destinations node and choose Configure a new <JMSTopic>" from the drop-down list." JMSTopic should read JMSQueue.
078441	Some examples fail when a data source is re-targeted. If, after re-targeting a data source, an example fails with the error message, "Cannot obtain connection after X seconds", make sure to restart the examples server in order for the change to take effect.
CR102138	In WebLogic Server 6.1 SP05, the JSSE version of the SSLClient sample code example may not work because the jssecacerts file does not contain the correct ca.pem file. The work-around for this problem is to add the ca.pem file provided in the WebLogic Server disribution to the jssecacerts file. To add ca.pem to the jssecacerts file: Open a command window and change directory to: WL_HOME\samples\examples\security\sslclient> At the prompt, enter: keytool -v -keystore jssecacerts -trustcacerts -import -file %WL_HOME%\config\examples\ca.pem When prompted for a password, enter: changeit
CR103521	The documentation for the dbkona examples omit the following additional steps, which are necessary to compile and run the examples: Select the appropriate weblogicoci37.dll based on the version of Oracle you are using. Add the path to the correct DLL to your PATH environment variable. Run the following command to load the database: java utils.Schema "jdbc:weblogic:oracle" weblogic.jdbc.oci.Driver -s oracle_server -u scott -p tiger -verbose utils\ddl\demo.ddl Before compiling the code for an example, edit all of the example's Java files to change the server property to the value of oracle_server that you specified in the previous step.
CR103547	dbkona examples fail to start on HP and Solaris. When a dbkona example is launched, this error message results: Starting Loading jDriver/Oracle ..... Trouble while executing example java.sql.SQLException: System.loadLibrary(weblogicoci37) threw java.lang.UnsatisfiedLinkError: /space/bt/ weblogic/load4/bea/wlserver6.1/lib/hpux11/oci817_8/libweblogicoci37.sl: specified library, or one of its dependencies, does not exist. at weblogic.jdbc.oci.Driver.loadLibraryIfNeeded (Driver.java:243 at weblogic.jdbc.oci. Driver. connect (Driver.java:79)at examples.dbkona.query. main(query.java:118) The problem occurred because jDriver is looking for libclntsh.so.8.0 file, which should be located in the ORACLE_HOME/lib directory. jDriver did not find the file because ORACLE_HOME was not defined. The example instructions do not include instructions to: ensure that the ORACLE_HOME environment variable is defined ensure that the ORACLE_HOME/lib directory contains the libclntsh.so.8.0 file ensure that the ORACLE_HOME/lib directory is in the LD_LIBRARY_PATH. These steps are required to run the dbkona examples
CR127253	The examples.security.acl example does not trap the error that results when the user does provide the appropriate command line options. Instead, a stack trace is displayed. The package summary for the examples.security.acl example says: "user and password are the username and password for joeuser. If the -user and -pass command line options are not specified, the JNDI Initial_Context defaults to the user "guest" with the password "guest". In this case, the Altclient client will fail because the aclexample ACL does not assign permissions to the user "guest." Running the example without following these instructions resulted in this error stack trace: C:\610sp6\wlserver6.1\samples\examples\security\acl>java -Dweblogic.security.SSL .ignoreHostnameVerification=true examples.security.acl.AltClient t3s://localhost :7002 -sslCert demokey.pem;democert.pem -servername weblogic.bea.com javax.naming.AuthenticationException. Root exception is java.lang.SecurityException: Authentication for user null denied in realm weblogic Start server side stack trace: java.lang.SecurityException: Authentication for user null denied in realm weblog ic at weblogic.security.acl.Realm.authenticate(Realm.java:212) at weblogic.security.acl.Realm.getAuthenticatedName(Realm.java:233) at weblogic.security.acl.internal.Security.authenticate(Security.java:17 1) at weblogic.common.internal.RMIBootServiceImpl.authenticate(RMIBootServi ceImpl.java:47) at weblogic.common.internal.RMIBootServiceImpl_WLSkel.invoke(Unknown Sou rce) at weblogic.rmi.internal.BasicServerRef.invoke(BasicServerRef.java:360) at....

Installer Known Problems

Change Request Number	Description
039773	The following misleading error message occurs when there is not enough space in /temp to run the installer: Preparing to install... The included VM could not be extracted. Please try to download the installer again and make sure that you download using 'binary' mode. Please do not attempt to install this currently downloaded copy. By default the installer tries to unpack all the installer files in the /temp directory. If there isn't enough temporary space there, you cannot install and you get this message. Make room in the /temp directory for the installer files.
040353	Installation problems can occur on HP-UX and Solaris machines because shared libraries do not have execute permission. File permission changes made by the WebLogic Server installer are reverted by installer imposed standard attributes. A workaround is to issue the chmod +x command on all the shared library objects (under lib/hpux11 and lib/solaris).
043357	If you are getting the following error on Solaris 2.6 or 2.7 UNIX systems: UnsatisfiedLinkError in AWT applications (libmawt.so: open failed: No such file for directory). The policytool will not open. The Java Plug-in Control Panel will not open. Usually you might workaround this problem by modifying the LD_LIBRARY_PATH. For example: setenv LD_LIBRARY_PATH /usr/j2se/jre/lib/i386/motif12:$LD_LIBRARY_PATH But, in some cases with problems with the libmawt.so files, you may require a patch from Sun: see http://java.sun.com/j2se/1.3/install-solaris-patches.html for Sun's patch information.
053469 & 055598	When using the Administration Console, add the line, acl.modify.weblogic.admin=Administrators, to the filerealm.properties to grant permissions to users and/or groups. This ACL was in WebLogic Server 6.0, 6.0 Service Pack 1 and 6.0 Service Pack 2.
054058	When encountering an unsupported browser, beaexec.exe may result in WebLogic Server crashing.
087287	When installing on Solaris, if $DISPLAY is not set, the installer throws the following exception: Invocation of this Java Application has caused an InvocationTargetException. This application will now exit. To work around this problem, set $DISPLAY the IP address of the host machine.
091420	WebLogic Server 6.1 SP04 failed to start if the "Enable Post-Bind UID" option for a Unix Machine was checked. (This attribute can be configured in the Machines node of the Administration Console; its purpose is to return the Unix UID a server instance running a UNIX machine will run under after it has carried out all privileged startup actions). This problem is resolved in WLS 6.1 SP05. See CR091420.
092326	When installing WebLogic Server SP04 using the upgrade installer from WebLogic Server 6.1 SP03, an earlier version of beasvc.exe is installed than if you use the full installer for WebLogic Server SP04. To work around this problem, use the full installer for WebLogic Server SP04.
102815	There is a problem with the Install Anywhere installer that requires WebLogic Server 6.1 to be installed using Windows 2000 compatibility mode. For information, see Installing on Windows XP.

JCA Known Problems

Change Request Number	Description
049344	The current BEA WebLogic J2EE Connector architecture implementation does not provide the ability to detect the errors discussed in section 6.8 (Scenarios: Local Transaction Management) of the J2EE Connector Specification, Version 1.0, Proposed Final Draft 2.

JDBC and jDrivers Known Problems

Change Request Number	Description
033423	dbKona problems exist with QueryDataSet and the OCI805_8 driver. The problems are fixed on Windows NT, but this is still a problem on HPUX and Solaris.
035301	There are problems using the RMI driver with Sybase connection pools.
035530	Problems with enabling OS authentication and using Oracle 8.1.6/oci8 with the jDriver are fixed on NT and HP-UX, but are still an issue on Solaris. An error message that reports this problem: OS level authentication is not currently supported due to a defect in OCI 8 libraries.
040413	Cluster hanging problems have been observed in some instances on Solaris machines. This appears to be caused by a libthread deadlock between SIGALRM and SIGLWP. The deadlock manifests itself when the server is attempting a JDBC connection. To help avoid the problem, use the Administration Console to set the InitialCapacity attribute equal to the MaxCapacity attribute in the JDBCConnectionPool entry. A similar problem was previously logged as a Sun Microsystems Bug (ID 4141709 libthread deadlock between SIGALRM and SIGLWP), against Solaris 2.6. There is currently no mention of the problem for Solaris 2.7. For Solaris 2.6, the patch for this problem is 105569-16.
043224	Oracle supports of its own class of Array (oracle.sql.Array) through its ThinDriver. If you attempt to use these classes in an application that gets its connection using WebLogic Server JDBC, they will not work.
048219	Flush doesn't seem to "commit" with RMI Clob character write. This only seems to affect flushing when you are using these write methods: write(char[] cbuf) write(char[] cbuf, int off, int len) write(String str) write(String str, int off, int len) It works fine with: write(int c) No problems with flushing occur when using an Ascii Stream with a Clob. Also, to commit your write, you have to explicitly close the stream. This may only be a problem if you wanted to read back from the Clob before you were ready to close out your write stream.
049519	Due to a bug in Oracle's thin driver, the following method is not compliant with the java.sql.PreparedStatement API: PreparedStatement.setTimeStamp(int param index, Timestamp x, Calendar cal) The timestamp value is not adjusted according to the TimeZone specified in the Calendar object. The default TimeZone is always used. Oracle may fix this bug in future releases of the thin driver.
127348	When using the Oracle 8.1.7 Thin driver included with WebLogic Server 6.1SP6, the following exception may appear intermittently: java.sql.SQLException: Io exception: Protocol violation This may be caused by an error in the driver. The exception does not occur with a later version of the driver available from Oracle. WebLogic Server 6.1SP6 includes the 8.1.7.0.0 version of the driver. Version 8.1.7.1.0 is currently available from Oracle. Workaround: Get the new version of the driver from Oracle and add it (classes12.zip) to your classpath in front of weblogic.jar.

JMS Known Problems

Change Request Number	Description
037683	The JMS Topic configuration tab in the console accepts any alphabetic characters as valid values for a Multicast Address. It should validate the value that is entered against an Internet Address format.
049229	You may experience problems with JMS if you run the server with the unsupported JDK 1.3.0. This problem occurs when restarting the server when there are existing messages in the JMS store -- these messages are ignored. It is recommended that you use the 1.3.1 JDK that is bundled with the server.
083538	In a multi-server configuration, an application or EJB on one server cannot lookup and get a connection from a TxDataSource on another server. The application or EJB must use a TxDataSource on the same server. This is a system limitation.

JSP Known Problems

Change Request Number	Description
049854	The ejb2 JSP tool does not support EJB 2.0 local interfaces.
075671	When using Java classes such as FileInputStream or ArrayList in JSPs, you will not receive a NoClassDefFoundError, even if you do not import these classes into the JSP.
104828	Struts 1.02 recommends moving all JSPs below the WEB-INF directory of the Web application to prevent direct access. This ensures that JSPs can only be forwarded to from Action classes. Any file located inside the WEB-INF folder cannot be accessed directly. Configuration: WebLogic Server 6.1 SP2 or WebLogic Server 6.1 SP4. Solution S-12864 suggests a workaround using servlet forward.
CR127701	There is a known security vulnerability in JDK 1.3.1 and 1.4.1. This vulnerability could cause a JVM to crash when attempting to parse malformed XML or XSLT, resulting in a Denial of Service on the web or application server. BEA recommends installing the patch provided by Sun Microsystems to resolve this known security vulnerability. With the patch, more rigorous JSP parsing is performed. As a result, code that compiled under JDK 1.3.1 and 1.4.1 may not compile after installation of the patch. For information about how JSPs in the Petstore sample application were upgraded to avoid compilation errors, see WebLogic Server 8.1 Upgrade Guide at http://download.oracle.com/docs/cd/E13222_01/wls/docs81/upgrade/upgrade6xto81.html#1062978.

JVM Known Problems

Change Request Number	Description
CR099314	There is a problem with JDK 1.3.1_04 and the nohup command, as described at http://developer.java.sun.com/developer/bugParade/bugs/4755829.html. The nohup command starts failing; the WebLogic Server process exits when the parent shell exits, regardless of the invocation of nohup. The problem was confirmed for ksh (the default Solaris 8 shell), and may exist for other shells. The problem does not occur with the bash shell. Workaround: Use "(startWeblogic >output 2>&1 &)" instead of "nohup startWeblogic &".
CR100564	Due to an interoperability problem between WebLogic Server 6.1 SP04 and WebLogic Server 8.1 calls from an RMI client on WebLogic Server 6.1 SP04 fails to get the correct exception thrown by an RMI object on WebLogic Server 8.1. This error resulted on WebLogic Server 6.1 SP04: <Mar 8, 2003 12:36:35 PM PST> <Error> <NT Performance Pack> <failure in processSockets() - GetData: 'weblogic.socket.NTSocketMuxer$GetData - native pointer: '272329000', numBytes: '0'' java.lang.ClassCastException: weblogic.iiop. LocateRequestMessage at weblogic.iiop.EndPoint Impl.cleanupPendingResponses(EndPointImpl.java:659) at weblogic.iiop.ConnectionManager.got ExceptionReceiving(ConnectionManager.java:273) at weblogic.iiop.MuxableSocketIIOP.endOfStream (MuxableSocketIIOP.java:681) at weblogic.socket. NTSocketMuxer.processSockets(NTSocketMuxer.java:657) at weblogic.socket.SocketReaderRequest. execute(SocketReaderRequest.java:24) at weblogic. kernel.ExecuteThread.execute(ExecuteThread.java:139) at weblogic.kernel.ExecuteThread.run (ExecuteThread.java:120) > <Mar 8, 2003 12:41:35 PM PST> <Error> <NT Performance Pack> <failure in processSockets() - GetData: 'weblogic.socket. NTSocketMuxer$GetData - native pointer: '272089304', numBytes: '0'' java.lang.ClassCastException: weblogic.iiop. LocateRequestMessage at weblogic. iiop.EndPointImpl.cleanupPendingResponses(EndPointImpl.java:659) at weblogic.iiop.Connection Manager.gotExceptionReceiving(ConnectionManager.java:273) at weblogic.iiop.MuxableSocketIIOP. endOfStream(MuxableSocketIIOP.java:681) at weblogic.socket.NTSocketMuxer.processSockets(NTSocketMuxer.java:657) at weblogic.socket.Socket ReaderRequest.execute(SocketReaderRequest.java:24) at weblogic.kernel.ExecuteThread.execute (ExecuteThread.java:139) at weblogic.kernel. ExecuteThread.run(ExecuteThread.java:120)

Change Request Number

Description

CR099314

There is a problem with JDK 1.3.1_04 and the nohup command, as described at http://developer.java.sun.com/developer/bugParade/bugs/4755829.html.

The nohup command starts failing; the WebLogic Server process exits when the parent shell exits, regardless of the invocation of nohup. The problem was confirmed for ksh (the default Solaris 8 shell), and may exist for other shells. The problem does not occur with the bash shell.

Workaround:

Use "(startWeblogic >output 2>&1 &)" instead of "nohup startWeblogic &".

CR100564

Due to an interoperability problem between WebLogic Server 6.1 SP04 and WebLogic Server 8.1 calls from an RMI client on WebLogic Server 6.1 SP04 fails to get the correct exception thrown by an RMI object on WebLogic Server 8.1. This error resulted on WebLogic Server 6.1 SP04:

<Mar 8, 2003 12:36:35 PM PST> <Error> <NT Performance Pack> <failure in processSockets() - GetData: 'weblogic.socket.NTSocketMuxer$GetData - native pointer: '272329000', numBytes: '0'' java.lang.ClassCastException: weblogic.iiop. LocateRequestMessage at weblogic.iiop.EndPoint Impl.cleanupPendingResponses(EndPointImpl.java:659) at weblogic.iiop.ConnectionManager.got ExceptionReceiving(ConnectionManager.java:273) at weblogic.iiop.MuxableSocketIIOP.endOfStream (MuxableSocketIIOP.java:681) at weblogic.socket. NTSocketMuxer.processSockets(NTSocketMuxer.java:657) at weblogic.socket.SocketReaderRequest. execute(SocketReaderRequest.java:24) at weblogic. kernel.ExecuteThread.execute(ExecuteThread.java:139) at weblogic.kernel.ExecuteThread.run (ExecuteThread.java:120) > <Mar 8, 2003 12:41:35 PM PST> <Error> <NT Performance Pack> <failure in processSockets() - GetData: 'weblogic.socket. NTSocketMuxer$GetData - native pointer: '272089304', numBytes: '0'' java.lang.ClassCastException: weblogic.iiop. LocateRequestMessage at weblogic. iiop.EndPointImpl.cleanupPendingResponses(EndPointImpl.java:659) at weblogic.iiop.Connection Manager.gotExceptionReceiving(ConnectionManager.java:273) at weblogic.iiop.MuxableSocketIIOP. endOfStream(MuxableSocketIIOP.java:681) at weblogic.socket.NTSocketMuxer.processSockets(NTSocketMuxer.java:657) at weblogic.socket.Socket ReaderRequest.execute(SocketReaderRequest.java:24) at weblogic.kernel.ExecuteThread.execute (ExecuteThread.java:139) at weblogic.kernel. ExecuteThread.run(ExecuteThread.java:120)

Plug-Ins Known Problems

Change Request Number	Description
CR172808	WebLogic Server 6.1 and Apache plug-in are certified on Solaris 2.6, but the WebLogic Server 6.1 SP5 plug-in fails to start with the following linkage error: # ./apachectl startssl Syntax error on line 216 of /usr/local/apache3/conf/httpd.conf: Cannot load /usr/local/apache3/libexec/mod_wl_ssl.so into server: ld.so.1: /usr/local/apache3/bin/httpd: fatal: libpthread.so.1: version \QSUNW_1.2' not found (required by file /usr/local/apache3/libexec/mod_wl_ssl.so) ./apachectl startssl: httpd could not be started As of WebLogic Server 6.1 SP5, plug-in binaries require libthread.so.1 (SUNW_1.2), but Solaris 2.6 has an older version of it (SUNW_1.1, SUNW_0.9). WORKAROUND: Statically link libweblogic.a, instead of mod_wl.so/mod_wl_ssl.so. This is a system limitation.

Change Request Number

Description

CR172808

WebLogic Server 6.1 and Apache plug-in are certified on Solaris 2.6, but the WebLogic Server 6.1 SP5 plug-in fails to start with the following linkage error:

# ./apachectl startssl

Syntax error on line 216 of /usr/local/apache3/conf/httpd.conf:

Cannot load /usr/local/apache3/libexec/mod_wl_ssl.so into server:

ld.so.1: /usr/local/apache3/bin/httpd: fatal: libpthread.so.1: version

\QSUNW_1.2' not found (required by file

/usr/local/apache3/libexec/mod_wl_ssl.so)

./apachectl startssl: httpd could not be started

As of WebLogic Server 6.1 SP5, plug-in binaries require libthread.so.1 (SUNW_1.2), but Solaris 2.6

has an older version of it (SUNW_1.1, SUNW_0.9).

WORKAROUND:

Statically link libweblogic.a, instead of mod_wl.so/mod_wl_ssl.so.

This is a system limitation.

RMI Known Problems

Change Request Number	Description
CR100713	WebLogic Server 6.1 SP04 and earlier does not interoperate with WebLogic Server 8.1. Transactions between 6.1 and 8.1 server instances timeout with an ArrayIndexOutOfBoundsException on the 6.1 server instance. There is an RJVM/RMI problem when a 6.1 server instance performs a lookup of the 8.1 server instance's coordinator remote object. When the 6.1 server attempts an ack, the operation fails, the 8.1 coordinator fails to resolves the transaction and throws a commit timeout exception to the client. For example: a client and remote object startup class are deployed on 8.1 and 6.1 servers. The client looks up the remote object on the 8.1 server and invokes a remote method specifying the URL of the 6.1 server. The 8.1 method looks up the remote object on the 6.1 server and invokes a method with the URL of the 8.1 server. The 6.1remote method looks up the remote object on the 8.1 server and invokes a method. The problem occurs when bootstrapping and sending the JVMID. There is no way of telling who the remote peer is. This problem is resolved in WebLogic Server 6.1 SP05.

Change Request Number

Description

CR100713

WebLogic Server 6.1 SP04 and earlier does not interoperate with WebLogic Server 8.1. Transactions between 6.1 and 8.1 server instances timeout with an ArrayIndexOutOfBoundsException on the 6.1 server instance.

There is an RJVM/RMI problem when a 6.1 server instance performs a lookup of the 8.1 server instance's coordinator remote object. When the 6.1 server attempts an ack, the operation fails, the 8.1 coordinator fails to resolves the transaction and throws a commit timeout exception to the client.

For example: a client and remote object startup class are deployed on 8.1 and 6.1 servers. The client looks up the remote object on the 8.1 server and invokes a remote method specifying the URL of the 6.1 server. The 8.1 method looks up the remote object on the 6.1 server and invokes a method with the URL of the 8.1 server. The 6.1remote method looks up the remote object on the 8.1 server and invokes a method. The problem occurs when bootstrapping and sending the JVMID. There is no way of telling who the remote peer is.

This problem is resolved in WebLogic Server 6.1 SP05.

RMI-IIOP Known Problems

Change Request Number	Description
CR126991	There is an intereroperability problem between WebLogic Server 6.1 SP05 and WebLogic Server 6.1 SP06 that may result in a ClassCastException.The error is a result of casting IIOPRemoteRef to a IOR during replaceObject. A patch for WebLogic Server 6.1 SP05 is available.

Security Known Problems

Change Request Number	Description
048159	When a Log Audit Provider is being used, no messages are being sent to the WebLogic Log File notifying you that auditing is turned on.
051215	When entering the system password during startup for WebLogic Server, if you use a CTRL/C sequence to interrupt the WebLogic Server startup, the terminal window stays in no-echo mode and characters that are typed are not echoed. If possible, wait to interrupt the WebLogic Server startup until after you enter the system password. If you do use a CTRL/C sequence when entering the system password, reset the terminal window or close the terminal window and start another window.
051315	The Node Manager in WebLogic Server cannot process the private keys created by the Certificate Request Generator servlet. The wlkeytool converter tool has been provided to solve this problem. Use the wlkeytool converter tools as follows: wlkeytool inputkey.pem outputkey.pem where inputkey.pem is the name of the private key created by the Certificate Request Generator servlet. outputkey.pem is the name of the converted private key. If the private key is password protected (that is, if you specified a password in the Private Key Password field in the Certificate Request Generator servlet), the wlkeytool converter tool prompts you for the password. If the private key is unprotected, press Enter. The wlkeytool converter tool prompts you for a password for the converted private key. A password must be supplied, or the wlkeytool converter tool will not create an unprotected private key. A deployed WebLogic Server does not support the converted private key created by the wlkeytool converter tool, so you should retain the private key created by the Certificate Request Generator servlet and use it with your deployed WebLogic Servers.
053390	If you are using an RDBMSRealm with a Caching Realm enabled, and the DatabasePassword is not specified in the RDBMSRealm tag in the config.xml file, the server halts with the error "<wlsstartuperror1>" and does not startup. Workaround: Open the config.xml file and add the DatabasePassword in clear text to the RDBMSRealm tag.
053409	To use the Java Security Manager with WebLogic Server, specify the -Djava.security.manager property when starting WebLogic Server. The included startup scripts specify a policy file but no security manager. Therefore the policy file ends up getting ignored. You must set a security manager for the policy file to become effective.
078619	Group Membership caching is not using the value set in GroupMembershipCacheTTL. It is instead using the value set in GroupCacheTTLPositive.
078894	When FlatGroup.clearCache() is invoked to refresh the cache, and isMember is invoked for a second time, isMember is returning false. It is still using old cache data.
087225	In WebLogic Server 6.1 SP04 and later, the LDAP realm V2 has issues with the Open LDAP Server when running the getGroups() method for large numbers of groups (more than 300). This problem is due to caching bugs in Open LDAP. Workaround: Do not try to obtain more than 300 groups when using the getGroups() method to retrieve groups from the Open LDAP server.

Server Known Problems

Change Request Number	Description
039575	If your servlet makes an HTTP connection back to the same server there is the possibility of deadlock under load. To work around the problem, use a separate execute queue for the servlet that returns the DTD.
042556	The weblogic.Admin command now requires the user and password options. For example: java weblogic.Admin -username system -password gumby1234 The username "system" is required for most functions (for example: VERSION). If you do not specify -username system, you will get the following error: Exception in thread "main" java.lang.SecurityException: Authentication for user system denied in realm weblogic <<no stack trace available>>
051980	When using the 1.3 VM with HotSpot, doing a thread dump causes the JVM to exit and the server to crash.
052300	Using Call-by-Reference doesn't work when you deploy your application as an .ear file that is not located in the applications directory and if you have set ServletReloadCheckSecs is set to -1.
052406	Attempting to start Managed Servers in quick succession creates problems. A delay between starting them fixes the problem.
052543	When passing complex objects server to server using RMI-IIOP, there will be many problems using com.sun.jndi.cosnaming.CNCtxFactory when getting an initial context; it is recommended to use weblogic.jndi.WLInitialContextFactory instead.
053328	If you are having problems with out of memory errors while using HP, try increasing the kernel parameter called max_thread_proc to 1024.
053330	The SNMP agent uses an object identifier (OID) for WebLogic in communicating with SNMP management systems. The WebLogic OID for 6.1 is different than the WebLogic OID for 5.1. In 5.1 it was .1.3.6.1.4.140.600, but in 6.1 it is .1.3.6.1.4.140.625.
054504	We have encountered problems on Windows 2000 and NT when using the -hotspot option with the JVM. This is due to a Sun bug (http://developer.java.sun.com/developer/bugParade/bugs/4479571.html) and results in intermittent crashes of the server. To work around this problem, try setting: -XX:MaxPermSize=128m This option allows you to increase the size of the area HotSpot uses for class representations.
057008	The NTSocketMuxer getCPUCount() is returning an incorrect value. A Service Pack from Microsoft that is installed does not recognize all of the CPUs.
058868	Due to a HotSpot virtual machine error, a signal 11 may unexpectedly crash WebLogic Server running on Solaris.
059018, 063937, 063939, 063965	By default, a WebLogic Server 6.1 Service Pack 2 client does not interoperate with a WebLogic Server 6.1 GA or 6.1 Service Pack 1 server. This is true both for a 6.1 SP2 client and a 6.1 SP2 server acting as a client to a 6.1 GA or 6.1 SP1 server. To allow an SP2 client to talk to a 6.1 GA or 6.1 SP1 server, add the following switch to the WebLogic Server 6.1 Service Pack 2 start script: -Dweblogic.61compat=true
065410	Due to a Sun bug, a java lang.OutOfMemory error my occur when starting WebLogic Server if the number of classes being loaded is too high. As a workaround, increase the value of the JVM option -XXMaxPermSize when initializing WebLogic Server, using this syntax: java -XX:MaxPermSize=<value>K where <value> is some number of kilobytes. The JVM uses a default maximum value of 32 MB. This problem has been reported to Sun.
078289	When a browser using Java Plug-in 1.3.1_02 or 1.3.1_03 refreshes a page with applets, the JVM issues simultaneous destroy() and init() calls to the applets. Because the calls are at the same time, applets may not properly close before trying to initialize. As a workaround, use thread.sleep() in the init() method of your applets to delay initialization by approximately 1000 milliseconds. The delay ensures that the destroy() call finishes any required cleanup before the applet attempts to reinitialize.

Servlet and Web Application Known Problems

Change Request Number	Description
041528	Cookies generated by the first release of WebLogic Server 6.0 with no Service Pack applied are not recognized by WebLogic Server 6.1.
046536	HTTPS does not work with Virtual Hosting.
046537	Virtual Hosts should define their own security realms.
048678	Previously, to log out a user from a Web Application you could use session.invalidate(). This now only works if the user has been to just one Web Application. If more than one Web Application is visited, session.invalidate() will not log out the user. This is not a known issue; rather it is a design change due to the implementation of Single-Signon. The server tracks the user at the Web server level and not the Web Application level--this behavior is mandated by the Servlet 2.2 specification. If you wish to log out users who have been to more than one Web Application, use: weblogic.servlet.security.ServletAuthentication.logout(request);
050811	If you modify a filter class, you must redeploy the Web Application to apply the change.
051311	You cannot not put white space in the name (context path) of a Web Application.
051329	The following issue is merely informational and will not be fixed or changed because it is the desired functionality. In the WebLogic Server config.xml file, the Path attribute of an archived Web application should consist only of the path to the archive file, but not the archive file name itself. In contrast, the URI attribute of a Web Application Component should consist only of the archive file name. For example: <Application Name=foo, Path=/apps> <WebAppComponent Name=foo,URI=foo.war> </Application> For backward compatibility, WebLogic Server 6.1 accepts a config.xml that specifies both the path and the archive filename of an application in the Path attribute (for example, Path=/apps/foo.war). However, the XML file is overwritten to specify only the path (/apps).
052686	The Web Application Deployment Descriptor Editor shipped with WebLogic Server 6.1 supports Public Final Draft 1 of the Servlet 2.3 specification. Because of this, there is no means within the editor for configuring the <ejb-local-ref> element of web.xml, which was introduced in Public Final Draft 2. Further, the editor will remove any <ejb-local-ref> that may be present in an existing Web App deployment descriptor.
CR088485	WebLogic Server 6.1 SP03 contained the wrong version of the Jakarta regular expression matching library. It was shipped with jakarta-oro-2.jar, rather than jakarta-oro-2.0.6.jar. This resulted in unexpected results for the customer. Work-around: place the correct version of Jakarta—jakarta-oro-2.0.6.jar—in the WebLogic Server classpath, ahead of weblogic.jar.
089868	An important performance fix introduced in WebLogic Server 6.1 Service Pack 3 (083654) caused multi-byte headers to become garbled; previously the default behavior was that Weblogic sent HTTP headers as raw bytes. This previous behavior was incorrect as the HTTP specification states that headers should be ASCII single byte characters. (BEA is aware that this fix can cause interoperability problems between WebLogic Servers 6.1 with Service Pack 3 and lower and WebLogic Servers with Service Pack 4.) For SP03 and SP04 users experiencing header garbling, the workaround is: Install the CR089868_610sp3.jar patch or CR089868_610sp4.jar, available from BEA Customer Support, and set UseHeaderEncoding="true" in the the <WebServer> element of your config.xml. If set to "true", the server will use the Content-Type character set to encode the Content-Disposition header. Otherwise, and by default, it will use the default platform encoding. Note that the server will not use any encoding if the resulting encoding is an 8-Bit Character set. In SP05, the workaround for the garbling problem is simply to set UseHeaderEncoding="true" as described in the second bullet item above.

System Administration Known Problems

Change Request Number	Description
034177	The Administration Console does not set the correct attribute value for application status.
034955	The application directory list used to scan for deployments is not persistent. Multiple application directories are not supported.
035670	When you reset an MBean name, the configuration is not saved correctly. To work around this problem: In the Administration Console, clone the MBean. Rename the clone. Delete the original MBean.
036201	Occasionally, after making modifications in the Administration Console, WebLogic Server cannot be restarted and a Fatal initialization exception Throwable: weblogic.server.ServiceFailureException: Initializing JNDI: javax.naming.ServiceUnavailableException [Root exception is java.net.UnknownHostException: Unknown protocol: 'Protocol: 'unknown''] message is displayed.
038589	If Managed Server(s) are communicating through HTTPS with the Administrative Server, the discovery of Managed Servers does not work. For example, if the Administrative Server is restarted with the -Dweblogic.management.discover=true command line argument and the running managed servers were booted with the -Dweblogic.admin.host=https://<adminhost>:<adminport> argument, then the discovery of managed servers does not work. Workaround: When Managed Servers are booted with an HTTPS connection to the Administrative Server, do not restart the Administrative Server with the -Dweblogic.management.discover=true argument. Instead, do the following: 1. Restart the Administrative server without the discover flag. 2. Once the Administrative Server is booted, force discovery by the Admin Server by running the following command: java weblogic.Admin -url <ADMINHOST>:<ADMINPORT> -username system -password <SYSTEM-PASSWORD> INVOKE -mbean "<YOURDOMAIN>:Type=Domain,Name=<YOURDOMAIN>" -method discoverManagedServers
039033	WebLogic Server complains of an error undeploying an application that does not exist.
039311	When you start a managed server that has been cloned, it throws a ConfigurationException. The workaround is to create a server instead of cloning it.
039532	An .ear file can only be undeployed using the Administration Console once it has been exploded under the applications directory. Also, once the application is undeployed, the only way it can be redeployed is through the Administration Console.
040034	Attempting to start a server as a Managed Server when it is already running as an Administration Server results in the following misleading error: <Dec 8, 2000 7:44:02 PM PST> <Error> <Configuration Management> <Error connecting to admin server and initializing admin home: admin URL: t3://localhost:7001 java.lang.NullPointerException Workaround: This is incorrect usage. The same server cannot be started as an Administration Server and a Managed Server. Define a new server for your Managed Server, and specify its name in -Dweblogic.Name during startup.
040124	While the Administration Server is running, copying an existing .jar or .ear file to an applications directory does not redeploy the application. The following error is displayed in the server log: javax.naming.NameAlreadyBoundException: Can't rebind anything but a replica-aware stub to a name that is currently bound to a replica-aware stub; remaining name '' <Dec 12, 2000 12:09:37 PM PST> <Error> <J2EE> <Error deploying application <YOUR-EJB-NAME>: Could not deploy: '<YOUR-FILE-NAME>': JNDI name in use Make sure that either the guest user or the everyone group is granted lookup permission to weblogic.jndi. The fileRealm.properties file should contain the following: acl.lookup.weblogic.jndi=everyone
041290	If the username property is set from the weblogic.Admin or weblogic.Server command lines, the username must be "system". Note however, that the username syntax is retained to prevent breaking a user's scripts. The username may not be modified in the config.xml file. The weblogic.management.password property value must be the system password.
042232, 055217	It is no longer possible to have a domain with the same name as its Administration Server. This will not be fixed, as the namespace conflicts produce too many problems.
048833	The ServletRuntime mbean contains the method getName(). In previous releases, this method returned the name of the servlet. For all other mbeans, the getName() method returns the name of the mbean. For consistency, in WLS 6.1 the getName() method on the ServletRuntime mbean has been modified to return the name of the mbean. A new method, getServletName(), has been added to the ServletRuntime mbean to return the name of the servlet.
052657	If you set configuration attributes for an application in config.xml, yet the application does not really exist, you will see the application incorrectly displayed as deployed in the Administration Console.
077958	When using NodeManager, the following exception may be thrown: <May 18, 2002 8:53:36 PM PDT> <Error> <NodeManager@localhost:5555> <SocketInputHandler: handshake failed 'FATAL Alert:HANDSHAKE_FAILURE - The handshake handler was unable to negotiate an acceptable set of security parameters.' on socket /172.17.24.148> This exception could be due to one of the following: The JVM version being used to run NodeManager is not supported by WebLogic Server. A JSEE conflict if NodeManager is running on the AIX platform. If NodeManger is running on the AIX platform, see Weblogic AIX Platform Support in Platforms for more information.
CR093687	In WebLogic Server 6.1 SP03 and later, you cannot create a Dynamic Connection Pool when LoadBeforeAppDeployments=true for Startup class. The connection pool is not created and goes to a hung state. There is no exception reported. If but, if LoadBeforeAppDeployments=false, the connection pool is successfully created.
CR097152	In WebLogic Server 6.1 SP03 and SP4, restarting the Administration Server may result in this exception: java.rmi.ConnectException: This RJVM has already been shutdown This problem was not reproduced in In WebLogic Server 6.1 SP05.

Tools Known Problems

Change Request Number	Description
063745	The documentation on the weblogic.refresh tool states that you can specify files using wildcards (e.g. "*.gif") and that you can use a comma-separated list to specify multiple files. However, in WebLogic Server Service Pack 2, weblogic.refresh does not allow the use of wildcards in comma-separated lists.

Web Services Known Problems

Change Request Number	Description
046807	The exception messages generated by the Java client application in the Java client JAR file are not internationalized.
047435	In RPC-style Web services, user-defined exceptions thrown by the EJB are not propagated to the client applications that use the Web services client API to invoke the service. Instead, they are propagated as generic SOAPFault exceptions.
049966	If you use the WebLogic Server Web Service client API (version 6.1) in a WebLogic Server 6.0 component, such as an EJB, you get the error java.rmi.ConnectException: No available router to destination. This is because a 6.0 JNDI URL factory gets called rather than the initial context factory that you specify in your client code. To work around this problem, get rid of the 6.0 JNDI URL factories by adding the following code snippet when you make your initial context: h.put(Context.URL_PKG_PREFIXES, "");
051917	In message-style Web services, JMS exceptions that occur on the target JMS destination are not propagated to the client applications that use the Web services client API to invoke the service.
055062	The path attribute of the clientjar element of the wsgen Ant task does not, as documented, default to client.jar if the path attribute is not specified in the build.xml file.
055096	You cannot pass a JavaBean that has no package name as a parameter to a WebLogic Web service.
055596	Interoperability issue. WebLogic Web services do not support SOAP 1.1 multi-reference compound data types.
056287	Interoperability issue. WebLogic Web services do not support the xsd:timeInstant and xsd:dateTime in the SOAP encoding.
056452	Interoperability issue. WebLogic Web services do not support multi-dimensional array data types.
056634	Interoperability issue. WebLogic Web services do not support SOAP-ENV:Header attributes.
058175	Interoperability issue. WebLogic Web services do not support xsd:ur-type array data types.
058840	Interoperability issue. The WebLogic Web service client API cannot bind to a WSDL that specifies more than one return type.
059782	Interoperability issue. The WebLogic Web service client API does not support invoking a Web service whose WSDL contains a <soap:header> element in its <binding> section.
059858	Interoperability issue. WebLogic Web services use only the 1999 XML Schema URL in the generated SOAP request, as shown in the following SOAP request snippet: xmlns:xsi='http://www.w3.org/1999/XMLSchema-instance' xmlns:xsd='http://www.w3.org/1999/XMLSchema'
061161	Interoperability issue. The WebLogic Web services client API does not support the use="literal" attribute of the <soap:body> element.
061335	Interoperability issue. The WebLogic Web services client API does not support invoking a Web service whose WSDL contains a <part> element with an element attribute inside of a <message> element.
061350	Interoperability issue. By default, the WebLogic Web services client API generates SOAP requests with qualified names. This does not interoperate well with some other non-WLS Web services, such as .NET. The workaround to this problem is to add the following code to your client application: CodecFactory factory = CodecFactory.newInstance(); SoapEncodingCodec codec = new SoapEncodingCodec(); codec.writeQualifiedName( false ); factory.register( codec );
061572	Interoperability issue. The WebLogic Web services client API does not support the style="document" attribute of the <soap:binding> element of WSDL.
061606	When assembling a message-style Web service using the wsgen Ant task, the task adds a literal encoding to the generated WSDL. This forces a programmer invoking the Web service to register the LiteralCodec with the CodecFactory, even when it is not required. The LiteralCodec forces the programmer to have the weblogic.jar file on the client computer.
061929	Interoperability issue. The WebLogic Web services client API always uses qualified names in its generated SOAP requests, even if the client application includes the following code: codec.writeQualifiedName( false );
77911	WebLogic Server correctly maps the built-in Java data type byte[] to xsd:hexBinary. However, the JAX-RPC specification states that byte[] should also have a mapping to xsd:base64Binary. Currently, there is no way to specify this mapping using Version 6.1sp3 of WebLogic Web services.
78062	WebLogic Web services cannot interoperate with .NET Web services that use arrays and JavaBeans as parameters or return values.
78530	The Web Service examples do not build on the Solaris operating system.

WebLogic Tuxedo Connector Known Problems

Change Request Number	Description
052737	A misconfigured node name in the WebLogic Tuxedo Connector XML configuration file will cause the system to hang. Use the following step to restart your system: 1. Shut down WebLogic Server. 2. Verify your node names in your WebLogic Tuxedo Connector XML configuration file. 3. Update the WebLogic Tuxedo Connector XML configuration file with valid node names. 4. Restart WebLogic Server.
077553	Integer array object passed by value from Tuxedo to WebLogic Server throws weblogic.utils.AssertionError. This problem was resolved in WebLogic Server 6.1 SP04. See "077553".
078774	A tBridge session connection may stop pulling messages from the Tuxedo /Q.
079630	Most users will expect to set the Correlation ID using the setJMSCorrelationID(String) method. This will take a 32 character string and turn it into a 64 byte array. JMS stores the string as UTF-16BE. When the tBridge receives the correlation id from Tuxedo it is 32 characters as 32 bytes. The tBridge then uses the setJMSCorrelationIDAsBytes(byte[]) method to set the message for JMS receiving queue. The characters may be the same in ASCII but the two will not compare due the length difference. When it is necessary to compare the Correlation ID with the ID returned from Tuxedo in the receiving JMS queue, create a byte array containing the hexadecimal values of the Correlation ID. Then use the setJMSCorrelationIDAsBytes() and getJMSCorrelationIDAsBytes()methods to create an ID to compare with the ID returned from Tuxedo in the receiving JMS queue. For example, if the string Correlation ID is "1234567890ABCDEFGHIJKLMNOPQRSTUV": private byte[] coridbyte={0x31,0x32,0x33,0x34,0x35,0x36,0x37, 0x38,0x39,0x30,0x41,0x42,0x43,0x44, 0x45,0x46,0x47,0x48,0x49,0x4a,0x4b, 0x4c,0x4d,0x4e,0x4f,0x50,0x51,0x52, 0x53,0x54,0x55,0x56}; msg.setJMSCorrelationIDAsBytes(coridbyte); corIDasBytes = msg.getJMSCorrelationIDAsBytes(); corIDasBytes contains the correct value to compare with the Correlation ID returned from Tuxedo.

XML Known Problems

Change Request Number	Description
050274	When you use the Administration Console to add either a parser or entity XML Registry entry, and you do not specify a public ID, you cannot subsequently navigate to the entry to edit it. The workaround is to always enter a value in the public ID field when you create an XML registry entry.
052985	The built-in Xalan transformer package weblogic.apache.xalan does not include the Xalan compatibility API. These classes are included in the xalanj1compat.jar file if you were to download Xalan 2.0.1 directly from the Apache Web site. For more information on the compatibility API, see http://xml.apache.org/xalan-j/usagepatterns.html#compat. Because these classes are not included in WebLogic Server, if your application uses code from Xalan 1.X that has been deprecated in Xalan 2.0.1, your application will not work correctly. Modify your code so that it does not use the Xalan 1.x deprecated classes, or rewrite your Xalan code to use JAXP instead.

Change Request Number

Description

050274

When you use the Administration Console to add either a parser or entity XML Registry entry, and you do not specify a public ID, you cannot subsequently navigate to the entry to edit it.

The workaround is to always enter a value in the public ID field when you create an XML registry entry.

052985

The built-in Xalan transformer package weblogic.apache.xalan does not include the Xalan compatibility API. These classes are included in the xalanj1compat.jar file if you were to download Xalan 2.0.1 directly from the Apache Web site. For more information on the compatibility API, see http://xml.apache.org/xalan-j/usagepatterns.html#compat.

Because these classes are not included in WebLogic Server, if your application uses code from Xalan 1.X that has been deprecated in Xalan 2.0.1, your application will not work correctly. Modify your code so that it does not use the Xalan 1.x deprecated classes, or rewrite your Xalan code to use JAXP instead.

ZAC Known Problems

Change Request Number	Description
CR047534	The JRE Update Frequency property set in the ZAC Publishing Wizard is no longer supported, and you need to use one of the following methods to update your JRE when necessary. When a JRE is packaged with an application, it will be reinstalled on the client only if: The entire JRE installation directory is deleted from the client AND either the client executable OR the ZAC bootstrap routine is run; or A new version of the JRE library package is published on the WebLogic Server AND the bootstrap routine (not the client executable) is run from the client. No package versioning takes place on the server (unless the version number is made part of the package name); it is only possible to "revert" a package if it has not already been published.
061699	WebLogic Server does not automatically register a default web application. Because ZAC gets classes through the default web application, users of ZAC must now explicitly configure a default web application. If you fail to explicitly configure a default web application, ZAC will not function, and the server log will contain a message similar to the following: <Nov 2, 2001 2:50:49 PM PST> <Error> <HTTP> <HttpServer(7908535,null default ctx,myserver) found no context for "/drp-exports/ZACHello/index.xml". This request does not match the context path for any installed web applications and there is no default web application configured.> <Nov 2, 2001 2:50:49 PM PST> <Error> <HTTP> <HttpServer(7908535,null default ctx,myserver) found no context for "/drp-publish/ZACHello/". This request does not match the context path for any installed web applications and there is no default web application configured.>

Change Request Number

Description

CR047534

The JRE Update Frequency property set in the ZAC Publishing Wizard is no longer supported, and you need to use one of the following methods to update your JRE when necessary.

When a JRE is packaged with an application, it will be reinstalled on the client only if:

The entire JRE installation directory is deleted from the client AND either the client executable OR the ZAC bootstrap routine is run; or
A new version of the JRE library package is published on the WebLogic Server AND the bootstrap routine (not the client executable) is run from the client.

No package versioning takes place on the server (unless the version number is made part of the package name); it is only possible to "revert" a package if it has not already been published.

061699

WebLogic Server does not automatically register a default web application. Because ZAC gets classes through the default web application, users of ZAC must now explicitly configure a default web application.

If you fail to explicitly configure a default web application, ZAC will not function, and the server log will contain a message similar to the following:

<Nov 2, 2001 2:50:49 PM PST> <Error> <HTTP> <HttpServer(7908535,null default

ctx,myserver) found no context for "/drp-exports/ZACHello/index.xml". This

request does not match the context path for any installed web applications

and there is no default web application configured.>

<Nov 2, 2001 2:50:49 PM PST> <Error> <HTTP> <HttpServer(7908535,null default

ctx,myserver) found no context for "/drp-publish/ZACHello/". This request

does not match the context path for any installed web applications and there

is no default web application configured.>