Running UltraSearch

Oracle UltraSearch is a web-based search application based on Oracle Text. It can be used to index a wide variety of datasources like websites, databases, files etc. and to make these available for searching. It is available with a database, an application server or a collaboration suite installation. I’ll describe in this article how to get it up and running (as part of a database installation). This is a summary of the instructions on OTN.

When you install the 10g database, UltraSearch is automatically installed too, however it is not available immediately. Perform the following steps:

  1. Ugrade the (10g v1) database to version (patch 4163362)
  2. Make sure that the executable ctxhx (on windows in the Oracle Home bin directory) is on the PATH environment variable. Now you can handle binary files
  3. Enable/unlock the database users wksys and wk_test.
  4. Add the following entry to the ORACLE_HOME/oc4j/j2ee/OC4J_SEARCH/config/data-sources.xml file (make sure to enter the correct url and password):

    This configures the jdbc datasource for the default search sample application
  5. Start the UltraSearch webapp: [Oracle install path]/BIN/searchctl start. This launches oc4j and deploys the admin and the example search webapp.

To perform a search you’ll first have to index your datasource(s). This is done with the admin application: http://[server]:5620/ultrasearch/admin/index.jsp. Check your portnumber in the database install info (portlist.ini), mine is 5620. Login with the wk_test user. During database install the wk_inst instance (as it’s called) has been installed. Select this and press the apply button. Now you can crawl a website or a directory or a database . You’ll find out how :-).
After this is done start the sample query application: http://[server]:5620/ultrasearch/query/search.jsp. Enter a search term and hit the button. Unfortunately, this failed with me 🙁 , but when I upgraded my database it succeeded.

One thought on “Running UltraSearch

  1. Hi Aino

    Many thanks for the hints, this did the trick for me nicely.

    One item to mention is that data source sample in line item 4 of the
    blog entry (p=672) doesn’t render correctly in Firefox 1.0.7 (may be
    something funny with the use of the “code” tag) however, that said, it
    did print out okay (mostly). You might want to check this next time
    you’re updating the blog.

    Regards, Ian.
    PS. And many thanks again for relieving my frustration… 😉

Comments are closed.

Next Post

Hacking web apps

On JavaPolis 2004 Erwin Geirnaert did a very nice presentation on hacking Web Applications. He gives an overview of how this can be achieved and then he gives some very nice examples for WebSphere and Weblogic. But my favorite is the website where he only changes the parameter patientId=12345 to […]
%d bloggers like this: