Migrate V6.1 portal search collections

When you migrate or upgrade WebSphere Portal to a later version, the data storage format and index structure of Portal Search is not backward compatible between the different versions. If you migrate the portal to a later version and want to continue using search collections, you need to preserve them before you migrate the portal and import them into the upgraded portal after the migration.

This applies to migrating between versions of WebSphere Portal. Migrating search collections might also be required when you install an interim fix. Refer to the instructions in the ReadMe file provided with the interim fix.
Notes:

  1. If you do not want to migrate search collections, the migration process deletes them automatically and creates new search collections. Content is indexed in the next scheduled crawl after migration is complete.

  2. Export and import portal site search collections into portal v7.0 is not supported.

Search collections that were created in v6.0.1.x require manual migration; use the Import or Export Collection option of the Manage Search portlet to export and import the search collections as described below. For more details about these tasks and the Manage Search portlet, see the portlet help.

To migrate search collections...

  1. Before you migrate the portal to a later version, export search collections. This step exports the configuration data and all document URLs of search collections.

    1. Before you export a collection, verify the portal application process has write access to the target directory location. Otherwise you might get an error message, such as File not found.

    2. When you specify the target directory location for the export, be aware that the export overwrites files in that directory.

  2. For each collection, document the following data:

    • The target file names and directory locations to which you export the collection.

    • Location, name, description, and language.

    • Settings for the Specify collection language and Remove common words from queries options.

  3. Delete the search collections from existing portal. Otherwise they can be corrupted by the import step that follows later. If you are upgrading from one portal version to another, you do not need to delete the search collections, as the new collections are stored under a different directory path location.

  4. Upgrade WebSphere Portal as required.

  5. Create empty search collections that you can use later to hold the imported collections.

      Fill in the following fields and select the following options according to the information that you documented in step 2 above:
      Location of Collection

        The location can match the old setting, but does not have to match it.

      Name of Collection

        The name can match the old setting, but does not have to match it.

      Description of Collection

        The description can match the old setting, but does not have to match it.

      Specify Collection Language

        Select this to match the old setting as documented in step 2.

      Select Categorizer

        The value is overwritten by the import process.

      Select Summarizer

        The value is overwritten by the import process.

      Remove common words from queries (for example. in, of, on, etc.)

        Check or clear this setting to match the old setting as documented in step 2.

      You do not have to add content sources or documents, as that is completed by the import process.

  6. Check that the target search collections that you created in step 5 are empty.

      Do not import collection data into a target collection that already contains sources or documents.

  7. Import the search collection data into the portal. For the import source information, use documented file names and directory locations to which you exported the collections before the portal upgrade.

      When you import a collection, a background process fetches, crawls, and indexes all documents which are listed by URL in the previously exported file. Keep in mind that this process can require a large amount of memory and an extended amount of time.

  8. When importing Web Content Manager collections from portal v6.0.1.x, modify the target URL for the affected content sources to use the IP address of the new portal server. To do this, click the content source, select the General parameters tab, and modify the field Collect documents linked from this URL.


Notes:

  1. When you import search collection data into a collection, most of the configuration data (for example, content sources, schedulers, filters, and language settings) are also imported. If you configured such settings when creating the new collection, they are overwritten by the imported settings.

  2. If you migrate a remote search service, use the updated file PseLibs.zip. This file is compiled by Java 5, therefore you need to have the portal running under IBM WAS v7.0. You do not need to replace the SOAP and EJB applications.

  3. If you migrate to WebSphere Portal v7.0, you can migrate only web collections. You cannot migrate the portal search collections.


Parent

Migrate V6.1 search components

 


+

Search Tips   |   Advanced Search