Loading UDXs into AnzoGraph DB

To load new extensions to AnzoGraph DB, you place the UDX C++ .so files (or JVM JAR files) in a specified extension library directory and then restart AnzoGraph DB. Each time AnzoGraph DB starts, it scans the files in the extension directory and loads the extensions in the registration database. This topic provides instructions for configuring the shared object directory and creating new extension points.

On a cluster, perform the steps below on the leader server.

1. If necessary, create a directory in the location where you want to store the UDX C++ .so files (or JVM JAR files), and any other files or libraries on which the extensions are dependent. AnzoGraph DB needs to be able to access the directory. Cambridge Semantics recommend that you create the directory outside of the AnzoGraph DB local file system, or you keep a backup copy of the directory outside of the AnzoGraph DB file system, so that the files are not removed or overwritten any time you upgrade or re-deploy AnzoGraph DB.
2. Edit the AnzoGraph DB system configuration file, settings.conf, to add the following setting and value to the file. For instructions on editing settings.conf, see Changing System Settings.

   extensions_dir=/path/directory_name

   For example, the following line configures AnzoGraph DB to scan shared object files in the /opt/udx/lib directory:

   extensions_dir=/opt/udx/lib

   Any .so files that you store in the directory will be scanned and registered in AnzoGraph DB anytime the database starts.

3. After specifying the extensions_dir value and copying any shared object files (and dependent files) into the directory, restart AnzoGraph DB to apply the new setting and load your extension libraries. For instructions, see Starting and Stopping AnzoGraph DB.