Application server fails to start in zh_TW.EUC locale on Solaris

If you have set the locale to zh_TW.EUC on Solaris, and are using WebSphere MQ as a JMS provider, you can encounter problems that stop application servers starting up.

If you intend using WebSphere MQ as a JMS provider on Solaris, do not set the LANG and LC_ALL variables to zh_TW.EUC (Traditional Chinese locale) to avoid problems when starting application servers. Set the LANG and LC_ALL variables to zh_TW instead of zh_TW.EUC.

Server memory consumption and java.lang.OutOfMemoryError exception when processing JMS messages

Intensive processing of JMS messages using the default JMS provider (for example, significant concurrent processing of large messages) can cause a java.lang.OutOfMemoryError exception and cause the application server to terminate.

Processing of JMS messages by the default provider is performed by a messaging engine within the application server process, and therefore consumes memory from the application server's JVM heap. This is in contrast with Version 5 where the support for the embedded JMS provider run in a separate process.

If the amount of memory available to the application server's JVM heap has not been configured large enough to handle the effect of the number of concurrent producers or consumers of messages and the message size, then a java.lang.OutOfMemoryError exception is thrown and the application server terminates.

Solution: When preparing to deploy applications that process JMS messages using the default messaging provider, you should plan for the potential consumption of the application server's memory for message processing. You should take into account the potential number of concurrent processors or consumers of messages and the message size, then set the size of the application server's JVM heap to handle the effect.

For example, when preparing to deploy a message-driven bean that is to be used to process messages concurrently, you should plan for the potential consumption of the application server's memory by concurrent endpoints. Each endpoint that is concurrently processing a message request adds at least two times the message size to the server's JVM heap and can add more, especially if a two-phase transaction is in place.

You can configure the amount of memory available to the application server's JVM heap by setting the Initial Heap Size and Maximum Heap Size properties of the application server. For example, on the WebSphere administrative console panel: Servers > Application servers > server_name > Java and Process Management > Process Definition > Java Virtual Machine .

You can configure the number of concurrent MDB endpoints that can process messages by setting the Maximum concurrent endpoints property of the activation specification used to deploy the message-driven bean. For example, on the WebSphere administrative console panel: Resources > JMS Providers > Default messaging > JMS activation specification > activationspec_name .

TopicConnectionFactory attributes clash error when using "Basic" WebSphere MQ broker (MA0C SupportPac broker) When creating a JMS topic subscriber using the WebSphere MQ messaging provider, the following error message occurs in the SystemOut.log file:

"WSVR0017E: Error encountered binding the J2EE resource, TopicConnectionFactory, as <JNDI_NAME> from file:<RESOURCES_FILE> com.ibm.ws.runtime.component.binder.ResourceBindingException: invalid configuration passed to resource binding logic. REASON: Failed to create connection factory: Error raised constructing AdminObject, error code:  TopicConnectionFactory attributes clash  :  TopicConnectionFactory attributes clash"  

This problem is causes by the configuration of the JMS topic connection factory used to create the subscriber, which specifies a broker version of "Basic" and a message selection value of "Broker". The "Basic" WebSphere MQ broker (MA0C SupportPac broker) does not support "Broker" message selection. -

Solution: Change the JMS topic connection factory to specify a message selection value of "Client", which is the only supported value for the WebSphere MQ Basic broker (MA0C SupportPac broker).

Message WSEC5061E: The SOAP Body is not signed is issued when running a secured
Web services application using JMS transport and WebSphere MQ When running a secured Web services application using JMS transport using the WebSphere MQ messaging provider, the following error message occurs in the SystemOut.log file:

[9/7/04 12:10:02:895 GMT-06:00] 00000039 enterprise    I   TRAS0014I: The following exception was logged WebServicesFault  faultCode: {http://schemas.xmlsoap.org/ws/2003/06/secext}FailedCheck  faultString: WSEC5061E: The SOAP Body is not signed.; null  faultActor: null  faultDetail:         stackTrace: com.ibm.wsspi.wssecurity.SoapSecurityException: WSEC5061E: The SOAP Body is not signed.; null ... 

The problem scenario has WebSphere global security enabled, and one Web services application configured with web services security has attempted and failed to use the JMS transport to send SOAP requests to the target Web service. The JMS resource is configured with WebSphere MQ using a remote WebSphere MQ server. The queue manager exists on this WebSphere MQ server.

The cause of this problem is another identical application is also running from a different application server but using the same queue manager and same queue name in the same WebSphere MQ server. The request sent from the original application has been processed through the same queue, but to the different application server where global security may not been enabled. Solution: To avoid this problem, complete the following steps:

1. Create a unique queue manager with a unique port in the WebSphere MQ server.

2. Reconfigure the JMS resources to use the new queue manager and port; for example, by using the WebSphere administrative console to change the properties of the WebSphere MQ queue connection factory, as described in Configuring a JMS queue connection factory, for WebSphere MQ.

3. Rerun the application.

Message MQJMS1006: invalid value for tempQPrefix is issued when trying to use a V5.1 client with a V5 Default Messaging queue connection factory on a V6 application server When trying to use a V5.1 application client to connect to a queue connection factory defined as a "V5 Default Messaging" resource on a V6 application server. the following message appears:

com.ibm.websphere.naming.CannotInstantiateObjectException: Exception occurred while the JNDI NamingManager was processing a javax.naming.Reference object.   Root exception is com.ibm.websphere.naming.CannotInstantiateObjectException: Exception occurred while the JNDI NamingManager was processing a javax.naming.Reference object.   Root exception is javax.jms.JMSException: MQJMS1006: invalid value for tempQPrefix: 

Cause: The application client is using WebSphere MQ JMS client CSD 04 JAR files. WebSphere Application Server Version 6 sets tempQprefix to blank, which cannot be handled by the CSD 04 release of the method setTempqPrefix.

Solution If the application client uses WebSphere embedded messaging JAR files, then apply the WebSphere Embedded Messaging interim fixes for WebSphere Application Server V5.1. If the client uses external WebSphere MQ JMS client JAR files, than apply CSD05.

When you use WebSphere MQ as an external JMS provider, messages sent within a user-managed transaction arrive before the transaction commits

When you use WebSphere MQ as an external JMS provider, messages sent within a user-managed transaction can arrive before the transaction commits. This occurs only when you use WebSphere MQ as an external JMS provider, and you send messages to a WebSphere MQ queue within a user-managed transaction. The message arrives on the destination queue before the transaction commits.

The cause of this problem is that the WebSphere MQ resource manager has not been enlisted in the user-managed transaction.

The solution is to use a container-managed transaction.

---

Related tasks
Troubleshooting WebSphere messaging
Troubleshooting service integration messaging