Contents I. Note II. Oracle Database server setup III. Oracle client setup IV. Oracle connectivity, cx_Oracle installation TODO: config/setup TODO: data load/migrate I. Note Due to licensing, we cannot distribute Oracle client binaries, and of course we will not distribute Oracle Database. Please refer to these guidelines if you need to use Oracle as a reference database for the Random Query Generator. II. Oracle Database server setup There are lots of choices for Oracle Database. What follows are two suggestions: http://www.oracle.com/technetwork/database/database-technologies/express-edition/downloads/index.html https://hub.docker.com/r/wnameless/oracle-xe-11g/ Follow their instructions for installation and setup. If you can use sqlplus or some other shell to successfully "select 1 from dual;", then you're good to go here. III. Oracle client setup Use cx_Oracle. http://cx-oracle.sourceforge.net/ Note: Installing this via pip or similar means won't work until you have first installed prerequisites. 1. Install Oracle instant-client http://www.oracle.com/technetwork/database/features/instant-client/index-097480.html After you choose your architecture, you need 1. the basic client 2. the SDK You will likely have to register on the Oracle site to download the instant client binaries. As of this writing, install instructions are at the bottom of the page once you make your architecture choice. 2. Additional dependent packages in distribution It's possible either when trying to install cx_Oracle, import it, or use it, you will get some linker error. I cannot predict what libraries you do and don't have. Here's a list of Linux distributions and known package dependencies: Ubuntu 14.04.2 LTS libaio1 3. Ensure impala-python3 is working. It may be enough to simply run: $ impala-python3 Getting it working is out of scope of this document. 4. Run these commands to install cx_Oracle into the impala-python3 virtual environment: $ source "${IMPALA_HOME}"/infra/python/env-gcc${IMPALA_GCC_VERSION}/bin/activate (env)$ pip install cx_Oracle==5.2.1 (env)$ deactivate If pip install fails, double check the following: * ORACLE_HOME environment variable * LD_LIBRARY_PATH environment variable BUILD.txt packaged with cx_Oracle has more instructions for installing cx_Oracle. IV. Oracle connectivity, cx_Oracle installation Verify cx_Oracle can connect to your Oracle Database. Edit appopriately and then run: $ "${IMPALA_HOME}"/tests/comparison/util/verify-oracle-connection.py