B Altering the Crawler Java Classpath

The Oracle Ultra Search crawler is a pure Java application that runs in a Java virtual machine. A Java virtual machine uses the Java classpath to find classes during runtime. When Oracle Ultra Search is installed, the default crawler classpath is stored in the database. Whenever a new Oracle Ultra Search instance is created, this default classpath is copied and used as the crawler classpath for that specific instance.

B.1 Reasons for Altering the Crawler Java Classpath

Usually, you do not need to alter the crawler Java classpath. However, there are certain reasons for you to do so. One reason could be to replace the JavaMail reference implementation with a third party JavaMail implementation.

B.2 Difference Between the Crawler Classpath and the Remote Crawler Classpath

The crawler classpath is the classpath of a crawler that runs on the same host as the Oracle Ultra Search backend. However, Oracle Ultra Search allows remote crawlers to be run on other hosts for scalability.

Remote crawler activation uses Java remote method invocation (RMI) technology. As a result, the classpath setting of a remote crawler is inherited from the classpath settings of the RMI registry and RMI daemon.

See Also:

"Using the Remote Crawler"

B.3 Altering the Crawler Java Classpath on the Oracle Ultra Search Server Host

Log on to the host where the Oracle Ultra Search backend is installed. Locate the file $ORACLE_HOME/ultrasearch/admin/wk0addcpath.sql.
Using SQL*Plus, run the wk0addcpath.sql script as the WKSYS super-user or as a database user that has been granted the super-user privileges. (This script only updates the CRAWLER_CONFIG_DEFAULT table. You also need to reconfigure your crawlers to get the WK$CRAWLER_CONFIG table updated correctly.)
When prompted, specify whether you want to alter the default classpath or an instance-specific classpath. Altering the default classpath causes all subsequently created instances to use that classpath. Existing instances are not modified.
When prompted, enter the Oracle Ultra Search instance name if you are attempting to modify an instance-specific classpath. If you are modifying the default classpath, then you do not need enter anything here.
When prompted, specify whether you want to update the entire classpath or append to it. Appending to a classpath adds entries to the beginning of it. Usually, earlier entries in the classpath override later entries in the case of duplicate classes.
When prompted, enter the new classpath if updating the entire classpath. If you are appending one or more directories or library files to the classpath, then enter these separated by the classpath separator for the platform where the Oracle Ultra Search backend is installed (the colon on UNIX platforms, or the semicolon on Windows).

B.4 Altering the Crawler Java Classpath on a Remote Crawler Host

Log on to the remote crawler host where the Oracle Ultra Search middle tier is installed. On a UNIX computer, locate and open the file $ORACLE_HOME/ultrasearch/tools/remotecrawler/scripts/unix/define_env. On a Windows computer, locate and open the file $ORACLE_HOME/ultrasearch/tools/remotecrawler/scripts/winnt/define_env.bat.
The define_env file specifies all environment settings used by the RMI subsystem. To alter the classpath, use a text editor to modify the APPLICATION_CLASSPATH variable.
Restart the RMI subsystem for these changes to take effect.

See Also:
"Using the Remote Crawler" for more details on starting up the RMI subsystem