Home
 

User login

 
 

Navigation

 
 

Events

« May 2008
SunMonTueWedThuFriSat
123
45678910
11121314151617
18192021222324
25262728293031
 

Cape Clear Server 7.5.1 Release Notes

This page lists the known issues in Cape Clear 7.5.1 and the issues fixed since 7.5.0

Known Issues in Cape Clear Server 7.5.1

The following table lists the issues known issues in Cape Clear Server 7.5.1:

 


Issue Summary:

The ccarchiver tool is not production-ready
ID: ESB-53
Status: Unresolved
Component: Documentation - Server Guide
Description: The ccarchiver command-line tool is not production-ready and should not be used for archiving your BPEL database in a production environment.

 


Issue Summary:

The filechecker property description for the file-in transport is incorrectly documented
ID: SVR-3312
Status: Unresolved
Component: Documentation - Server Guide
Description: The documentation for the file-in transport incorrectly describes the filechecker property as:

Specify the path to a file-checker Java class that scans files in the input directory. Alternatively, add a custom transport that references a Java class that scans the files (see "custom-in").

However, this property should not specify a path. The correct description should be:

Specify the fully qualified name of a Java class that scans files in the input directory. This class must implement the capeclear interface com.capeclear.capeconnect.transport.file.server.NewFileChecker.


Issue Summary:

The file-in transport endpoint URLs are incorrectly documented
ID: SVR-3165
Status: Unresolved
Component: Transports - File Drop
Description: The file-in documentation does not specify that the input-directory is a path-specifier to a directory, for example: C:\tmp\myapp_in

The description for the file-in transport should be:

A required attribute of type String, which indicates the input
directory for files picked up by the transport. The value of this
attribute should be the path to a directory on the file system, for
example: "C:\tmp\myapp_in" (on Windows) or "/tmp/my_app_in" (on Linux or Solaris).

For the file-out transport, the endpoint property should be:

Specify the directory to write files to, using three forward slashes to support UNIX systems. For example, file:///c:/tmp would specify the tmp directory on the C:/ drive. For Windows only, file:/C:/tmp is an acceptable format.


Issue Summary:

The attribute inbound-capacity setting on the file-in transport has been removed
ID: SVR-3151
Status: Unresolved
Component: Mediation
Description: This capacity is now managed by the Cape Clear Server itself. It is not recommended that it be changed by developers and so has been removed as a configuration option.


Issue Summary:

The "failed-directory" setting on the file-in transport has been removed
ID: SVR-3143
Status: Unresolved
Component: Mediation
Description: The file-in transport attribute "failed-directory" has been removed. This functionality was never provided and so should have no effect on assemblies developed for 7.5.0


Issue Summary:

Updating the web.xml file for a running server leads to initialization exceptions during the dynamic reload of ccx war
ID: SVR-2489
Status: Unresolved
Component: Integrations - JBoss
Description: The ccx web.xml file should not be updated when a server is running. The server needs to be
shutdown before any updates to that file are attempted. Forcing a dynamic reload of the server in this way will lead to JMX and hibernate initialization exceptions during the reload and will leave the server in an unusable state.


Issue Summary:

Increased PermGen Size Required for WebLogic
ID: SVR-2437
Status: Unresolved
Component: Documentation - Server Guide
Description: When running Cape Clear Server deployed within WebLogic, a "java.lang.OutOfMemoryError: PermGen space" failure may occur. To ensure that the JVM permanent generation space will not be exhausted, configure a JVM startup parameter of -XX:MaxPermSize=512m in the WebLogic $CAPECLEAR_DOMAIN_HOME/bin/setDomainEnv.cmd file (for Windows) or $CAPECLEAR_DOMAIN_HOME/bin/setDomainEnv.sh (for UNIX platforms).


Issue Summary:

The Adapters Template page in Cape Clear Manager displays JMS adapter templates, which are Retired Features.
ID: SVR-2239
Status: Unresolved
Component: Administration - Manager
Description: The Adapters Template page in Cape Clear Manager displays JMS adapter templates, which are Retired Features. The following items should be removed from the Cape Clear Manager JSPs:

* JMS JBossJMS@localhost JMS provider JBoss 4.0 JBoss 4.0 channel configuration

* JMS WebLogicJMS@localhost WebLogic JMS Provider 9.1 WebLogic JMS example configuration for localhost

* JMS WebSphereMQ@localhost


Issue Summary:

Configuring WS-RM
ID: SVR-2213
Status: Unresolved
Component: Transports - WS-RM
Description: When configuring the WS-RM Source transport in Server, ensure that the values selected for the following configuration properties are the same:

* wsrm.WSRMSequenceProcessor.proxy.pool.size

* msgstore.transport.ThreadCount

If the value set for msgstore.transport.ThreadCount is less than that set for wsrm.WSRMSequenceProcessor.proxy.pool.size, under heavy load conditions, it is possible that sequences will not be marked as terminated at the WS-RM destination, and so will not be cleaned-up until they expire. This can happen when a thread, attempting to send a terminate sequence message to the WS-RM destination, is unable to obtain a proxy from the pool to handle the request.


Issue Summary:

"java.lang.OutOfMemoryError: PermGen space failure" when editing Orchestration Manager web.xml
ID: SVR-2171
Status: Unresolved
Component: Administration - Orchestration Manager
Description: Editing the Orchestration Manager web.xml file for a running server will lead to a redeployment of the Orchestration Manager Web application. That redeployment may lead to a "java.lang.OutOfMemoryError: PermGen space failure".

The most common cause for this error is that you do not have enough space to load all your classes. Use the parameter '-XX:MaxPermSize=128m' to increase the memory available.


Issue Summary:

Cluster members must not be started together
ID: SVR-1917
Status: Unresolved
Component: Runtime -BPEL Runtime
Description: To ensure that a cluster starts up successfully, servers should be brought up one by one, and not all at once. This is to allow the application servers and Cape Clear to manage the initial cluster co-ordination in a controlled way and also to ensure that each server in the cluster has started up successully before other members join.


Issue Summary:

Authentication Trusted Endpoint Interceptor Issue
ID: SVR-1902
Status: Unresolved
Component: Interceptors
Description: If an Authentication Trusted Endpoint Interceptor is attached to a Web service using Cape Clear Manager, it is configured with the location of the file containing the list of trusted endpoints and associated caller principals. The default value for this configuration property(endpoint.file.name) is the relative path conf/trustedendpoints.xml. The documentation correctly indicates that the interceptor should be looking for this file relative to the WEB-INF directory of the deployed Server WAR file.
However, the interceptor incorrectly attempts to locate this file relative to the directory $CAPE_HOME/Server. To avoid this problem, the value of the property endpoint.file.name should be the absolute file path to the trustedendpoints.xml file. In the case where Server is deployed in JBoss, this would be
$CAPE_HOME/Server/jboss-4.0.2/server/default_or_all/deploy/ccx.war/WEB-INF/conf/trustedendpoints.xml.


Issue Summary:

JMS transports defined in Cape Clear Manager cannot handle spaces in filenames (classpath issue)
ID: SVR-1885
Status: Unresolved
Component: Transports - JMS
Description: If you configure a JAR for a JMS transport with paths that contain spaces, an error will be
raised. An example error is shown below. The underlying cause is a ClassNotFoundException due to an error in processing the classpath.

[18-Dec-2006 10:51:22.995 GMT] ERROR <RequestReceiverTask> Error in listening to

inbound JMS messages: Cannot instantiate class: com.sun.jndi.fscontext.RefFSContextFactory

javax.naming.NoInitialContextException: Cannot instantiate class:
com.sun.jndi.fscontext.RefFSContextFactory [Root exception is
java.lang.ClassNotFoundException:
com.sun.jndi.fscontext.RefFSContextFactory]

at javax.naming.spi.NamingManager.getInitialContext(NamingManager.java:652)

at javax.naming.InitialContext.getDefaultInitCtx(InitialContext.java:243)

at javax.naming.InitialContext.init(InitialContext.java:219)

at javax.naming.InitialContext.<init>(InitialContext.java:195)

at com.capeclear.capeconnect.transport.jms.common.JmsConnectionPool.createJndiContext(JmsConnectionPool.java:178)

at com.capeclear.capeconnect.transport.jms.common.JmsConnectionPool.reserveJndiContext(JmsConnectionPool.java:122)

at com.capeclear.capeconnect.transport.jms.common.JmsSessionPool.lookupDestination(JmsSessionPool.java:183)

at com.capeclear.capeconnect.transport.jms.server.RequestReceiverTask.runInLoop(RequestReceiverTask.java:115)

at com.capeclear.utils.thread.KeepGoingRunnable.run(KeepGoingRunnable.java:46)

at java.lang.Thread.run(Thread.java:534)

Caused by: java.lang.ClassNotFoundException: com.sun.jndi.fscontext.RefFSContextFactory

at java.net.URLClassLoader$1.run(URLClassLoader.java:199)

at java.security.AccessController.doPrivileged(Native Method)

at java.net.URLClassLoader.findClass(URLClassLoader.java:187)

at java.lang.ClassLoader.loadClass(ClassLoader.java:289)

at java.lang.ClassLoader.loadClass(ClassLoader.java:235)

at java.lang.ClassLoader.loadClassInternal(ClassLoader.java:302)

at java.lang.Class.forName0(Native Method)

at java.lang.Class.forName(Class.java:219)

at com.sun.naming.internal.VersionHelper12.loadClass(VersionHelper12.java:42)

at javax.naming.spi.NamingManager.getInitialContext(NamingManager.java:649)

... 9 more

[18-Dec-2006 10:51:22.995 GMT] INFO <RequestReceiverTask> Will retry receiving without logging further errors.


Issue Summary:

Changing log settings using Cape Clear Manager causes parsing error at Server startup
ID: SVR-1873
Status: Unresolved
Component: Administration - Manager
Description: In the Cape Clear Manager Log Settings page for Server, when you click Save, changes are saved but if special characters are present in the log4j.xml file, these characters are not escaped. For example, the character &lt; that is in the original log4j.xml file is saved as "<" and "&gt;" is saved as ">". Therefore, the next line if in the original log4j.xml file:

<param name="ConversionPattern" value="[%d{dd-MMM-yyyy HH:mm:ss.SSS z}] %p &lt;%c{1}&gt; %m%n"/>

is changed to:

<param name="ConversionPattern" value="[%d{dd-MMM-yyyy HH:mm:ss.SSS z}] %p <%c{1}> %m%n"/>

If you want to save a new version of the file, you need to ensure that the file is valid XML, which includes escaping the necessary characters.


Issue Summary:

Topic Transports fail to read all messages when there is high load
ID: SVR-1144
Status: Unresolved
Component: Runtime -BPEL Runtime
Description: We do not recommended the use of the JMS server transport to subscribe to JMS topics. The existing support is not production strength. JMS server transports that read from JMS queues are safe.


Issue Summary:

Keystore password is "capeconnect" instead of the password set in the Cape Clear Server installer
ID: SVR-381
Status: Unresolved
Component: Command Utilities
Description: If using the sample keystore and certificate provided in
$CAPE_HOME/Server/jboss-4.0.2/server/default/conf/certificates (not recommended for production use), when running the gen-key keytool command, you are prompted for a keystore password. The password is "capeconnect", regardless of what was specified as the admin password during installation.

Issue Summary:

The Admin password cannot be blank when deploying Cape Clear in WebSphere
ID: SVR-3290
Status: Unresolved
Component: Installer
Description: It is possible to set a blank Admin password when installing Cape Clear Server on Websphere. However, when running Cape Clear Server on WebSphere, if you have not specified a password, you will be unable to connect to the Cape Clear Manager and Orchestration Manager consoles. The following exception is produced by Cape Clear Manager:

"Authentication failed for user 'admin' from host 'localhost'. Reason: Username and/or password is null."

You must provide a password at installation time and then specify this password later on when attempting to log in to the management consoles.


Issue Summary:

Setting Up a User Account Repository for WebSphere Administrative Security
ID: SVR-3313
Status: Unresolved
Component: Integrations - WebSphere
Description:
  1. When setting up your user account repository, use a Standalone Custom Registry. For more details, see the FileRegistrySample.java file in the WebSphere Information Center.
  2. In particular, ensure that the users and groups files are in the correct format.
    For more details, see users.props file and groups.props file in the WebSphere Information Center.
    For example, following are sample users and groups configuration files:

    Example 1: myWebSphereUsers.txt

    # 5639-D57, 5630-A36, 5630-A37, 5724-D18
    # (C) COPYRIGHT International Business Machines Corp. 1997, 2005
    # All Rights Reserved * Licensed Materials - Property of IBM
    #
    # Format:
    # name:passwd:uid:gids:display name
    # where name = userId/userName of the user
    # passwd = password of the user
    # uid = uniqueId of the user
    # gid = groupIds of the groups that the user belongs to
    admin:admin:111:123:Admin User

    Example 2: myWebsphereGroups.txt

    # 5639-D57, 5630-A36, 5630-A37, 5724-D18
    # (C) COPYRIGHT International Business Machines Corp. 1997, 2003
    # All Rights Reserved * Licensed Materials - Property of IBM
    #
    # Format:
    # name:gid:users:display name
    # where name = groupId of the group
    # gid = uniqueId of the group
    # users = list of all the userIds that the group contains
    # display name = a (optional) display name for the group.
    admins:123:admin:Administrators

  3. Next, you need to configure the Websphere Administrative console with the values that you have just set up in your user repository. To
    do so, in the console, select Security>Secure administration, applications, and infrastructure, then from the drop-down list in the table at the bottom of the page select Standalone custom registry, then click Configure.
  4. In the Standalone custom registry, specify the following settings:
    • Primary administrative username: for example, admin.
    • Server identity that is stored in the repository: true.
    • Server user ID or administrative user on a Version 6.0.x node: for example, admin.
    • Password: for example, admin.
    • Custom registry class name: com.ibm.websphere.security.FileRegistrySample.

  5. Click the Custom properties link, then add the following new properties, which point to your users and groups configuration files described earlier, for example:
    • groupsFile: C:\CapeClearProfile_Auth_WS_Full\Authn\myWebsphereGroups.txt
    • usersFile: C:\CapeClearProfile_Auth_WS_Full\Authn\myWebsphereUsers.txt

  6. Click Save master configuration.

Other Issues with Running or Restarting WebSphere Administrative Security on the Default Installed Profile

The keystore and certificates in the default profile installed by Cape Clear appear to be unusable. When you start up the Cape Clear profile, an error similar to the following is displayed:

WSKeyStore E CWPKI0033E: The keystore located at "C:/CapeClearProfile_Auth_WS_Full/HOSTNAME/CapeClear/config/cells/CapeClearCell/nodes/CapeClear/ltpa.jceks"
failed to load due to the following error: Keystore was tampered with, or password was incorrect.

As a workaround, copy the keystore, truststore, and ltps.jceks file from the IBM/WebSphere/AppServer/profiles/AppSrv01/config/cells/myhostNode01Cell/nodes/myhostNode01
WebSphere installation directory and place them in the following directory:

<CAPE_CLEAR_PROFILE>/CapeClear/config/cells/<HOSTNAME>CapeClearCell/nodes/CapeClear

Note: The default keystore, truststore, and JCEKS files are provided for development purposes only and should not be used for a production environment. When generating shared keystore (JCEKS) files for WebSphere, ensure that you use the IBM JDK, see SVR-3306.


Issue Summary:

Using Shared Keystores for WS-Security on WebSphere
ID: SVR-3306
Status: Unresolved
Component: Integrations - WebSphere
Description: On WebSphere, you must generate JCEKS keystores using the IBM JDK. Shared keystores generated with Sun's JDK will not work. JKS keystores work on both JDKs.

Issue Summary:

Incorrect Class loader mode specified for WebSphere in post-installation steps.
ID: SVR-3308
Status: Unresolved
Component: Documentation - Server Guide
Description: In step 2 of the
"Completing Your WebSphere 6.1 Configuration>Step 2: Customize the Classpath Used by the Application" section, the class loader mode is incorrect:

2. The Configuration page is displayed as shown below. On this page, for Class loader mode, select Parent Last.

The correct class loader mode is: Application First and not Parent Last.

 

Issues Fixed Since Cape Clear Server 7.5.0

The following table lists the issues fixed in Cape Clear Server 7.5.1:


Issue Summary:

Ignore-wsa-endpoint renamed to ignore-dynamic-endpoint
ID: SVR-3214
Status: Fixed
Component: Mediation
Description: The generic out-transport attribute ignore-wsa-endpoints has been renamed to ignore-dynamic-endpoints. This is more inline with the step set-dynamic-endpoint.


Issue Summary:

Clarify use of prototype vs singleton for custom Spring extensions
ID: SVR-3199
Status: Fixed
Component: Mediation
Description: A clarification of the usage of Spring scopes has been made in 7.5.1 release. Setting a custom spring bean (for example, custom mediation step, custom routing strategy, custom-out transport, custom-scheduled-in transport, or custom error-handler) to have "prototype" scope means that they are single-shot beans. They are used once and then destroyed immediately afterwards.

Setting a custom spring bean to have "singleton" scope means that they are used for the lifetime of the application and are destroyed when the application is undeployed or disabled.

It is important to note that marking custom beans as "singleton" means that they must be thread safe.

Marking beans as "prototype" means that they do not have to be thread-safe, but at the cost of additional object creation and destruction. Also note that such beans can have no memory, since they are created on demand and destroyed immediately after use.

See documentation for further information.


Issue Summary:

Singleton Spring configured beans no longer require lazy-init=true
ID: SVR-3188
Status: Fixed
Component: Mediation
Description: The restriction to have lazy-init="true" for singleton scoped Spring beans has been removed.


Issue Summary:

Project name incorrect in the 'WS-Security within an Assembly' sample help page text
ID: SVR-3039
Status: Fixed
Component: Documentation - Other
Description: Step 2 of the 'WS-Security within an Assembly' sample help page states that you should run the AssemblyWithSOAPBindingTest.java from the BPELHelloWorldSecurityAssemblies project. This is incorrect. The project in which to find this file is called BPELHelloWorldSecurityAssembliesClient. Note that both projects exist in the sample.


Issue Summary:

File-out transport now supports MIME-Encoding Multipart messages
ID: SVR-3008
Status: Fixed
Component: Mediation
Description: The file-out transport now supports MIME-encoding multi-part messages. Previously, the file-out transport would only save the root part of any MediationMessage. Now, if a MediationMessage contains attachments, then it will MIME-encode the message before saving.

If your assemblies took advantage of the previous behavior (only saving the root part), then you will need to insert a <copy input="rootpart" output="message"/> step immediately prior to the file-out transport.