Documente Academic
Documente Profesional
Documente Cultură
General Information:
8/24/2012
1. Demo Story:
Its a technical demo to show E2E Hadoop HANA integration process using Data Services. This demo is shorter version of Real Time Big Data Retail POS HANA HADOOP Integration Scenario but focused on end to end technical process of Hadoop Map/Reduce job and integration with HANA using Data Services 4.1.Instead of 90TB of weblogs we deal with 130MB weblogs in this technical demo.
Two parts for the demo: i. Hadoop Map/Reduce Job converting the weblogs to structured data in HDFS ii. DS 4.1 which loads the data from HDFS to HANA
3. HANA studio Revision 26 recommended for this demo. Required to show the data in HANA.
ii.
iii.
HDFS(require FileZila):
Host Name: usphlvm1939.phl.sap.corp Username: <mailed on request> Password: <mailed on request> Port: 22222
4. Demo Preparation:
Yu may need to delete only the data in the HANA table using studio before your execute the DS job.
Tip: Set SYSTEM as Filters for Catalog and ITEM_SESSIONS2 as Filters for table before the demo. Right click on the table ITEM_SESSIONS2 table from SYSTEM schema > Delete. On the pop up select Delete All Rows
2. MapReduce job-The script to run it is (this script is on the Linux file system): /usr/local/mrJob/run.sh Login to Hadoop server using putty to execute this job. This Hadoop job converts the unstructured data (access.log) and creates a output in HDFS.
3. Output in HDFS this file is on the HDFS, not on the Linux file system: 475340 structured records /user/i040723/sessionLogDemo/sessionItems/affinity/part-r-00000
1. Launch DS Management Console http://ideshana04:8080/DataServices/launch/logon.do?LOGOUT=true 2. User Name/ Password : hadoop/welcome -> Log On
3. Click on Administrator
7. Click here to check logs and see if the job is completed successfully
11. This data can be easily consumed by BO client tools for Analysis.