Transcript
Page 1: Ambari Meetup: 2nd April 2013: Teradata Viewpoint Hadoop Integration with Ambari

Steve RatayChief Architect, Teradata ViewpointApril 2, 2013

TERADATA VIEWPOINTHADOOP INTEGRATION WITH AMBARI

Page 2: Ambari Meetup: 2nd April 2013: Teradata Viewpoint Hadoop Integration with Ambari

2 Copyright Teradata 2012

• Combination of Teradata Database, Aster, and Hadoop in a single appliance• Connectors between the different systems• Monitor the entire appliance with Teradata Viewpoint

Teradata Unified Data Architecture

DISCOVERY PLATFORM

CAPTURE | STORE | REFINE

INTEGRATED DATA WAREHOUSE

Aster Connector for Hadoop Teradata Connector for Hadoop

Aster Teradata Connector

SQL-H SQL-H

Page 3: Ambari Meetup: 2nd April 2013: Teradata Viewpoint Hadoop Integration with Ambari

3 Copyright Teradata 2012

Viewpoint Architecture

TD 1

DataCollectionService

PostgresDatabase

ViewpointPortal

Viewpoint Managed Server

ActiveMQ

AlertService

TD n

Aster 1

Aster n

HDP 1

HDP nEmail SNMP Portlet

Page 4: Ambari Meetup: 2nd April 2013: Teradata Viewpoint Hadoop Integration with Ambari

4 Copyright Teradata 2012

• Different technologies and formats:> Ganglia> Nagios> JMX> Screen scraping HTML

• Connection to multiple nodes:> Connectivity to all nodes> Access to several ports> Location of each service, especially when using high availability

Monitoring Hadoop without Ambari

Page 5: Ambari Meetup: 2nd April 2013: Teradata Viewpoint Hadoop Integration with Ambari

5 Copyright Teradata 2012

• RESTful APIs• JSON response format• Single server and port

Ambari to the Rescue!

Page 6: Ambari Meetup: 2nd April 2013: Teradata Viewpoint Hadoop Integration with Ambari

6 Copyright Teradata 2012

• Viewpoint uses Java and Spring• Used Spring’s RestTemplate class:

> Makes HTTP call to specified URL> Binds JSON results to Java model objects

• Data is collected every minute and stored in Viewpoint’s Postgres database> Use of Rewind to view previous state of system> Easy generation of trending charts and graphs

Accessing Ambari APIs

Page 7: Ambari Meetup: 2nd April 2013: Teradata Viewpoint Hadoop Integration with Ambari

7 Copyright Teradata 2012

System Health

Page 8: Ambari Meetup: 2nd April 2013: Teradata Viewpoint Hadoop Integration with Ambari

8 Copyright Teradata 2012

Metric Analysis

Page 9: Ambari Meetup: 2nd April 2013: Teradata Viewpoint Hadoop Integration with Ambari

9 Copyright Teradata 2012

Node Monitor

Page 10: Ambari Meetup: 2nd April 2013: Teradata Viewpoint Hadoop Integration with Ambari

10 Copyright Teradata 2012

Hadoop Services

Page 11: Ambari Meetup: 2nd April 2013: Teradata Viewpoint Hadoop Integration with Ambari

11 Copyright Teradata 2012

Hadoop Alerts

Page 12: Ambari Meetup: 2nd April 2013: Teradata Viewpoint Hadoop Integration with Ambari

12 Copyright Teradata 2012

• Viewpoint delivered comprehensive Hadoop monitoring solution in a couple of months• Java and web developers focused on collecting data and

building UI• Didn’t waste time dealing with intricacies of Hadoop and

Hadoop monitoring

Benefits of Using Ambari


Top Related