When you install Spark using Ambari, the hive-site.xml file is automatically populated with the Hive metastore location.

If you move Hive to a different server, edit the SPARK_HOME/conf/hive-site.xml file so that it contains only the hive.metastore.uris property. Make sure that the host name points to the URI where the Hive metastore is running, and that the Spark copy of hive-site.xml contains only the hive.metastore.uris property.

<configuration>
  <property>
  <name>hive.metastore.uris</name>
  <!-- hostname must point to the Hive metastore URI in your cluster -->
    <value>thrift://hostname:9083</value>
    <description>URI for client to contact metastore server</description>
  </property>
</configuration>