August 2017
Beginner to intermediate
278 pages
6h 40m
English
Like Hadoop, Spark also follows the master/slave architecture, master daemons called Spark drivers, and multiple slave daemons called executors. Spark runs on a cluster and uses cluster resource managers such as YARN, Mesos, or Spark Standalone cluster manager.
Let's walk through each component:

Spark driver is master in the Spark architecture. It is the entry point of Spark application.
Spark driver is responsible for the following tasks:
Spark Context: Spark Context is created in Spark driver. The Context object is also responsible for initializing application configuration.
DAG creation: Spark driver ...
Read now
Unlock full access