In this section, we show how to run Hive on MR3 on Hadoop. We assume that the user has already installed Hive on MR3 on Hadoop by following the instruction in Installing on Hadoop and that a Hadoop cluster meeting the requirements is available.

Running a component of Hive on MR3, such as Metastore and HiveServer2, involves the following steps.

  • The user updates to set additional environment variables.
  • The user updates hive-site.xml, mr3-site.xml, and tez-site.xml under a configuration directory (e.g., conf/tpcds).
  • The user executes a corresponding script under the directory hive such as hive/ and hive/

Usually updating and hive-site.xml is enough to get started as the default values in mr3-site.xml and tez-site.xml work well in common settings.

Since Hive on MR3 can run only with Metastore, the user can start by following the instruction in Running Metastore. Then the user can try to run HiveServer2 by following the instruction in Running HiveServer2. The user can quickly test Hive on MR3 by running Beeline as explained in Running Beeline. Later the user can try different configurations by following Advanced Guide.