Steps for Configuring OLH

Step 1: Installing OLH in the OFSAAI VM

1. Unzip the OLH Package downloaded from the Oracle site in the VM where OFSAAI is installed.
   Location: Inside the Home Directory of the user where OFSAAI is installed.
2. Set OLH_HOME environment variable in the .profile file.
   OLH_HOME contains the directories such as bin, examples, jlib and lib.
   Step 2: Configuring the Property
   1. Set the following property in the jdbcOutput.xml file, which is present in the $FIC_DB_HOME/conf/ location:

      <property>

      <name>oracle.hadoop.loader.connection.defaultExecuteBatch</name>

      <value>100</value>

      </property>

   2. From the DMT Configurations window, set H2T Mode as OLH.
   3. From the Register Cluster tab in the DMT Configurations window, register a cluster with Source Information domain name as the Cluster Name.
   For more information, see the DMT Configurations section in the OFS Analytical Applications Infrastructure User Guide.
   Step 3: Copy Configuration XMLs from Hadoop Cluster

   1. Copy the following files from the Hadoop Cluster to the Configuration File Path given in the Cluster Configurations window of the registered cluster.
      •  core-site.xml
      •  hdfs-site.xml
      •  mapred-site.xml
      •  hive-site.xml
      •  yarn-site.xml

      Note:

      Only Client Configuration Properties are required. If Cloudera Manager is used, the same can be downloaded directly, which will contain only the client properties.
   2. Modify the following property in the mapred-site.xml file in the $FIC_HOME/conf folder:

      <property>

      <name>mapred.child.java.opts</name>

      <value>-Xmx4096m</value>

      </property>

      <property>

      <name>mapreduce.job.outputformat.class</name>

      <value>oracle.hadoop.loader.lib.output.JDBCOutputFormat</value>

      </property>

      <property>

      <name>mapreduce.output.fileoutputformat.outputdir</name>

      <value>(Any temporary directory)</value>

      </property>

      <property>

      <name>oracle.hadoop.loader.defaultDateFormat</name>

      <value>yyyy-MM-dd</value>

      </property>

      Note:

      If proxy user is enabled and the Job is submitted by the same, the user should be created in every node of the Hadoop Cluster. For more information, see CDH4 Security Guide. The version of jars to be copied differs depending upon the version of CDH configured.
   Step 4: Copy the required Jars

   1. Copy commons-httpclient-3.1.jar from the CDH installation libraries into the $OLH_HOME/jlib folder.
   2. If OFSAA is using Apache driver:
   3. Usually jars such as hive-exec-*.jar, libfb303-*.jar, hive-service-*.jar, hive-metastore-*.jar are present in the ext/lib folder and are added to the Classpath. In case of any ClassNotFound Exception, perform the following steps:
      • Edit the oracle.hadoop.loader.libjars property present in the OLH_HOME/doc/oraloader-conf.xml file to accommodate the newly added jars. That is, $FIC_HOME/ext/lib/ hive-exec-*.jar (repeat for each of the mentioned jars)
      • Copy the entire property to the FIC_DB_HOME/conf/dtextInput.xml file.
   4. If OFSAA is using Cloudera Connectors:
      • Note that OLH is not qualified on Cloudera Connectors. Perform the following workaround:
        • Copy the following jars (Apache Drivers) to the OLH_HOME/jlib folder:

          hive-exec-*.jar, libfb303-*.jar, hive-service-*.jar, hive-metastore-*.jar

          Usually these jars are added to the Classpath. In case of any ClassNotFound Exception, perform the following steps:

          • Edit the oracle.hadoop.loader.libjars property present in OLH_HOME/doc/oraloader-conf.xml to accommodate the newly added jars. That is, ${oracle.hadoop.loader.olh_home}/jlib/ hive-exec-*.jar

   • Cop0y the entire property to the FIC_DB_HOME/conf/dtextInput.xml file.

     Note:

     Add the aforementioned jars only if OLH task is to be run. If any other OFSAA task is running, do not keep a copy of the jars in the OLH_HOME/jlib folder.