This course covers the essentials of deploying and managing an Apache™ Hadoop® cluster. The course is lab intensive with each participant creating their own Hadoop cluster using either the CDH (Cloudera's Distribution, including Apache Hadoop) or Hortonworks Data Platform stacks. Core Hadoop services are explored in depth with emphasis on troubleshooting and recovering from common cluster failures. The fundamentals of related services such as Ambari, Zookeeper, Pig, Hive, HBase, Sqoop, Flume, and Oozie are also covered. The course is approximately 60% lecture and 40% labs.
Prerequisites:
Qualified participants should be comfortable with the Linux commands and have some systems administration experience, but do not need previous Hadoop experience
Supported Distributions:
Red Hat Enterprise Linux 6
Course Outline:
- Hadoop: The Big Picture
- Data Analysis
- Big Data
- Hadoop Core Architecture
- Hadoop Ecosystem
- Hadoop Ecosystem continued
- Running Commands on Multiple Systems
- Running Commands on Multiple Hosts
- Preparing to Install Hadoop
- HDFS
- Design Goals
- Design
- Blocks
- Block Replication
- Namenode Daemon
- Secondary Namenode Daemon
- Datanode Daemon
- Accessing HDFS
- Permissions and Users
- Adding and Removing Datanodes
- Balancing
- Single Node HDFS
- Multi-node HDFS
- Files and HDFS
- Managing and Maintaining HDFS
- MapReduce
- MapReduce
- Terminology and Data Flow
- MapReduce Daemons
- YARN
- MapReduce Essential Configuration
- Failure and Recovery
- MapReduce
- MapReduce Schedulers
- Working with Jobs
- Scheduling Concepts
- FIFO Scheduler
- Fair Scheduler
- Fair Scheduler - Configuration
- MapReduce Schedulers
- Installing Hadoop with Ambari Lab Tasks
- Install Ambari