[Version 5.0.2 and later]Installing additional features on a base Application Server

This topic describes how to install additional features to an existing base WebSphere Application Server product.

Before you begin

If you encounter a problem such as not having enough temporary space or not having the right packages on your system, cancel the installation, make the required changes, and restart the installation.

The installation of new features uses the InstallShield for Multiplatforms (ISMP) wizard to perform the installation. You can use the Installation wizard or perform the installation from a command line, using the silent installation method.

You can add features to Integration Server on a stand-alone node or a managed node. See Installing available Integration Server features on a base node and Installing deactivated Integration Server features on a base node for more information.

Installation programs in WebSphere Application Server products

[5.0 only][Version 5.0.2]

Product Link to installation procedure Label on product CD
(Base) WebSphere Application Server Base product installation (which links to this procedure!) Application Server, IBM HTTP Server
Network Deployment Network Deployment installation Deployment Manager
Enterprise Enterprise installation Enterprise Application Server


Why and when to perform this task

You can add features to an unfederated base node or to a deployment manager node at any time, by running the Installation wizard again. Use the following procedure to avoid component regression problems when installing additional features.

If the base product node belongs to a deployment manager cell, it is important that you understand that there is no way to add features to the base node without unfederating it from its cell and losing the existing configuration. The existing configuration for the base node is lost as the configuration is restored back to its original configuration, before the node was federated and before you might have changed the configuration for the node with the administrative console of the deployment manager.

Restoring an original configuration means that you lose the configuration for any additional servers that you created on the base node from the administrative console of the deployment manager. The additional servers no longer exist. You lose any other configuration changes that you made from the administrative console of the deployment manager.

Steps for this task

  1. Log on as the root user on a Linux or UNIX system or as a user in the administrator group on a Windows system.

    You cannot install features correctly as a non-root user on a Linux or UNIX system, or as a non-administrator on a Windows system.

  2. Use the backupConfig command to save the configuration.

    The backupConfig command works well in an unfederated environment, to save the configuration of the base node. Issue the command from the install_root/bin directory of the base node.

    Although you can use the backupConfig command to save the configuration on a federated node, you cannot restore it with the restoreConfig command because the configuration you restore does not include any features that you add to the base node.

    You can use the backupConfig command to save the Network Deployment configuration when adding features to a federated node. Then you can restore it if necessary. You can also refer to the configuration as you use the administrative console of the deployment manager to reconfigure the base node. Issue the command from the install_root/bin directory of the deployment manager node.

    For example, use the following command on a Linux system:

    #./backupConfig.sh /tmp/my_backup_2004-06-03.zip -user user_ID -password password

    See backupConfig command for more information.
  3. If the node is a base product node in a deployment manager cell, unfederate the node with the removeNode command.
    Removing the base node from the cell also removes the configuration for the base node, which resides at the cell level. Issue the removeNode command from the install_root/bin directory of the base Application Server. For example, on a Linux server, issue the following command:
    #./removeNode.sh -user user_ID -password password
    See removeNode command for more information.
  4. Uninstall any interim fixes, starting with the last one and finishing with the first one.
    See Uninstalling interim fixes, cumulative fixes, and fix packs for more information about how to remove an interim fix, a cumulative fix, or a fix pack.
  5. Uninstall any cumulative fixes you installed, starting with the last one and finishing with the first one.
  6. Uninstall any fix packs you installed, starting with the last one and finishing with the first one.
  7. Log off and back on or reboot a Windows system.
  8. Insert the product CD labeled, Application Server, IBM HTTP Server.
    Windows systems and most Linux systems are configured to automatically mount CD-ROM drives.
  9. Open a shell window and mount the CD-ROM drive if necessary.
  10. Start the installation with the ./launchpad.sh command or the launchpad command on a Windows system.
    See the platform-specific installation instructions if you need more information about restarting the installation from the product CD-ROM, such as instructions for mounting the CD drive.

    Corresponding entries in the response file exist for every prompt that is described as part of the wizard. Comments in the response file describe how to customize the options.

    The rest of this procedure assumes that you are using the Installation wizard.

  11. Click Next to continue.
    The license agreement displays.

    The Installation wizard does not support hot keys, such as Alt-N. You can tab to Next and press Enter to select it, for example.

    [5.0 only][Version 5.0.2]The license that the Installation wizard displays can contain characters that display incorrectly in Japanese. For example, the section labeled Part 1 does not show the number 1. These missing characters do not significantly affect the content of the license agreement.

  12. Click the radio button beside the I accept the terms in the license agreement message if you agree to the license agreement and click Next to continue.
    After you accept the licensing terms, the Installation wizard checks for prerequisites and for previous versions, with which it can either migrate or coexist.

    If the wizard finds a previous version of WebSphere Application Server, it prompts you to migrate applications and the configuration from the previous version, or to coexist with it. If it finds more than one previous version, the Installation wizard lists them for you to select which one to migrate.

  13. Choose to install additional features.
    The Installation wizard panel that appears when the installer program detects a previous installation at the same product level has an option for installing additional features. Select the check box labeled, Add features to the existing copy.

    If you did not unfederate a managed node, you see a message that the installation wizard is unable to add components to a federated node. Exit the installation, start from the beginning of this procedure, and unfederate the node this time.

  14. Skip the coexistence panel that lets you select new ports, if it appears.
  15. Select features to install and click Next to continue when performing a custom installation.
    A description of each feature appears at the bottom of the panel when you roll the cursor over the feature.

    Previously installed features are checked and grayed out with the term (Installed) at the end of the feature name. The IBM HTTP Server feature is an exception. You can install more than one instance of the IBM HTTP Server product. It is not grayed out or labeled (Installed). You can install the plug-in feature each time you install IBM HTTP Server, too. Select a new directory for each instance of IBM HTTP Server that you install.

    [5.0 only][Version 5.0.2]Selecting certain features causes the installation of other prerequisite features. The following table shows this feature relationship.

    Features and feature dependencies
    If you select this feature: This feature is also installed: Feature description
    Application Server Admin Scripting These features are the Application Server run time and its Samples.
    Sub-features: Application Server Samples

    Administration

    Getting started with System Administration
    Sub-features: Admin Scripting

    Administrative Console Application Server
    Admin Scripting

    Deployment Tools

    Admin Scripting
    Sub-features: Deploy Tool Application Server
    Admin Scripting
    ANT Utilities

    Embedded Messaging

    Sub-features: Server and Client Application Server
    Client Only

    Message-driven Bean Samples (These samples are the MDBSamples application, which is for the server and not for the client.) Application Server samples
    Admin Scripting
    You can install the MDBSamples application without installing the embedded messaging server and client feature. Perhaps you have already installed the server and client feature. Or perhaps you intend to reconfigure the MDBSamples to use messaging resources that WebSphere MQ software provides.

    In any case, if you install the MDBSamples application and do not have the server and client feature installed, errors occur. The Application Server starts the MDBSamples application at start up. The MDBSamples application cannot start the required messaging resources in the Java Naming and Directory Interface (JNDI).

    Because of the lack of required resources, the message-driven beans fail to start, which causes an exception that the Application Server logs. The Application Server is not affected other than being unable to start the MDBSamples application.

    IBM HTTP Server

    Preparing to install and configure a Web server

    Web Server Plug-ins
    Sub-features: IBM HTTP Server
    Apache Web Server
    Microsoft Internet Information Services (IIS)
    Sun ONE (formerly known as iPlanet) Web Server
    Lotus Domino Web Server
    Performance And Analysis Tools

    Sub-features: Tivoli Performance Viewer
    Dynamic Cache Monitor
    Performance Servlet
    Log Analyzer
    Javadocs

    IBM WebSphere Application Server, Release 5 API Specification in Javadoc format


    Performance considerations

    For better performance in either a development or production environment, do not install the Samples. By omitting the Samples, you can improve Application Server startup time by 60 percent and save 15 percent of disk space. You can save up to 30 percent of process footprint (based on a maximum heap size of 256 MB).

    If your applications do not use Java Message Service (JMS) messaging, do not install embedded messaging. This recommendation is especially true if your system has 256 MB or less of physical memory.

    Not installing the embedded messaging feature saves about 36 MB of message queuing process memory and can improve application startup time up to 5 percent and save an additional 72 MB of disk space. (This was tested on Windows machines. Results on a Linux platform might differ slightly.)

    In addition, if you install the embedded messaging client feature you do not need to install the message-driven Bean Samples. The message-driven Bean Samples are for the embedded messaging server in the server and client feature.

    Web Server plug-in feature considerations

    Select the IBM HTTP Server feature to install and configure IBM HTTP Server on the same, or on a different machine than the WebSphere Application Server product. You can migrate plug-ins from an earlier version of WebSphere Application Server to access the current WebSphere Application Server product.

    Install IBM HTTP Server on a separate machine using the CD-ROM labeled, Application Server, IBM HTTP Server. After installing IBM HTTP Server, you can install the plug-in by installing the base WebSphere Application Server product and clearing all features except the plug-in for IBM HTTP server.

    See Preparing to install and configure a Web server .

    The Installation wizard automatically installs and configures IBM HTTP Server and the IBM HTTP Server plug-in on the same machine as the WebSphere Application Server if you choose the full installation. No further configuration is necessary. Installing the IBM HTTP Server product with the WebSphere Application Server product also includes it in the WebSphere Application Server uninstaller program that the Installation wizard creates. Choose the custom type of installation to clear the selection of the IBM HTTP Server feature, if you want to install and uninstall it separately.

    You can run the uninstaller program to remove all installed features.

  16. Specify a destination directory for the IBM HTTP Server feature, if selected. Click Next to continue.
    The destination directory for the base product is grayed out. You cannot change it.

    Specify a target directory for the IBM HTTP Server feature, if you are installing the feature. You cannot change the default installation directory for the embedded messaging feature, if you are installing the feature.

    Deleting the default target location and leaving an installation directory field empty stops you from continuing the installation process. The Installation wizard does not proceed when you click Next. Enter the required target directory to proceed to the next panel. Non-ASCII special characters are not supported in the name of the installation directory. Spaces are also not supported in the name of the installation directory.

    If you have problems accessing the administrative console after installation, check the installAdminConsole.log file for a failure indication. Clean up the /tmp space and reinstall the administrative console using the wsadmin scripting facility.

    If you select the embedded messaging feature and prerequisites are missing, the Installation wizard displays the mq_prereq.log error log and takes you back to the Installation type panel. Choose Custom installation and clear the checkbox for the embedded messaging feature. The mq_prereq.log file is in the system temp directory.

  17. Specify target directories for configuration files for any selected Web server plug-ins. Click Next to continue.

    If you are installing the IBM HTTP Server, you do not need to specify a location for its plug-in configuration file. The wizard uses the installation path you specified for the Web server to derive the location.

    If you have previously installed the IBM HTTP Server product on the same machine as the WebSphere Application Server, and are now installing just the plug-in, enter a configuration file location of IHS_DIR/conf/httpd.conf, where IHS_DIR is the directory where the IBM HTTP Server product is installed.

  18. Review the features summary information and click Next to install the product code for the additional features or click Back to change your specifications.
    When the installation is complete, the wizard displays the install_root\logs\mq_install.log installation log if you selected the embedded messaging feature and errors occur with its installation.
  19. Review the mq_install.log installation log if it displays. Click Next to continue.
    The wizard displays the Registration panel.
  20. Click Next to register the product, or clear the check box and click Next to register at a later time.
    The Registration panel is in English when the machine locale uses a double byte character set (DBCS) language.

    You can register the product now by using the link on this panel or later, by using the Web address in the install_root/prt/readme_prt.txt file. The Web address in the file is encoded for your locale. A known problem for some locales corrupts the Web address so that it is similar to the following example. (The Web address is all one line in the file. The following example shows the address on two lines for clarity.)

    http%3A%2F%2Fwww.ibm.com%2Fsoftware%2Fwebprt%2Fregister%2Fprodreg.nsf%
       2Fcontact%3FOpenForm%26lang%3DENU%26pid%3D45P8842%26q2%3DWindows+2000

    Use one of the following Web addresses if your install_root/prt/readme_prt.txt file is corrupt:

    The Installation wizard starts the First Steps tool. See firststeps command for more information.

  21. Click Finish to close the Installation wizard.
  22. Verify the success of the installer program by examining the Exit summary panel and the install_root/logs/log.txt for installation status.
    ISMP records a success message in the install_root/logs/log.txt file: "INSTFIN: The installation is complete." The log is the only source of status information for a silent installation.

    [Version 5.0.2]Verify the success of the installer by querying the response code on Linux and UNIX-based platforms:

    echo $? 
    On Linux and UNIX-based platforms, the return code from the installer is "1" to indicate success; any other response code indicates failure.

    Look for severe errors that the installer records in the install_root/logs/log.txt file in the installation root directory to verify that no file system or other unusual errors occurred during installation.

    If the install_root/logs/log.txt file does not contain a record of any problems but problems exist, verify or troubleshoot the installation, as described in Troubleshooting the installation and in Installation component troubleshooting tips.

    If problems exist, correct them, uninstall the product, as described in Uninstalling the product, log off as root and back on, and reinstall.

    When installing WebSphere Application Server with embedded messaging from the CD-ROM, the following warning messages are issued on some platforms despite a successful installation:

    MQSERIES.MQM-CL-HPUX: A test command parameter is not valid." and "* 
    "hostname:/cdrom/hpux/messaging/images":  Cannot open the
      logfile on this target or source.  Possibly the media is
      read-only or there is a permission problem." on mq_install.log.  
    You can ignore the warning messages, because the problem does not affect the successful installation of WebSphere Application Server with embedded messaging.

  23. Install the most current fix pack for the WebSphere Application Server product.
    See Recommended updates for WebSphere Application Server for information about downloading and installing the upgrades. See Installing interim fixes, cumulative fixes, and fix packs for information about using the update installer program.
  24. Install the most current cumulative fix for the WebSphere Application Server product.
  25. Install any interim fixes to bring the node back to its previous fix level.
  26. Linux and UNIX only: Issue write permissions for the new files to your non-root user group (wasgroup, for example), if you run the Application Server from a non-root user.

    See Running an Application Server and node agent from a non-root user and Running an Application Server from a non-root user ID and the nodeagent from root for more information.

  27. If the node was part of a cell, rejoin the node to the cell with the addNode command or the deployment manager administrative console.
    This action synchronizes the master cell configuration with changes you make to the Application Server configuration during the installation of the new features.
  28. Reconfigure the base node with the Network Deployment administrative console.

Results

You have now successfully installed additional features for the WebSphere Application Server product.


Related tasks
Installing the product
Using the Launchpad to start the installation
Installing silently
Automatically restarting WebSphere processes
Related reference
Platform-specific tips for installing and migrating
Tips for installing the embedded messaging feature



Searchable topic ID:   featuresb
Last updated: Jun 21, 2007 8:07:48 PM CDT    WebSphere Business Integration Server Foundation, Version 5.0.2
http://publib.boulder.ibm.com/infocenter/wasinfo/index.jsp?topic=/com.ibm.wasee.doc/info/ee/ae/tins_featuresb.html

Library | Support | Terms of Use | Feedback