O'Reilly logo

Learning Cascading by Victoria Loewengart, Michael Covert

Stay ahead with the world's most comprehensive technology and business learning platform.

With Safari, you learn the way you learn best. Get unlimited access to videos, live online training, learning paths, books, tutorials, and more.

Start Free Trial

No credit card required

MapReduce execution framework

MapReduce jobs are orchestrated by two primary server types—the JobTracker and the TaskTracker. There is one JobTracker, and it uses one or more TaskTrackers running on slave nodes where it distributes the work (the mappers and reducers).

The JobTracker

The JobTracker is the manager of all the jobs that are submitted to Hadoop and it performs many tasks. It queues jobs into an input queue, and then determines the order they are allowed to run in. Hadoop starts with a very basic approach here, which is to run jobs in the order in which they are received, or first come first served (FCFS). Clearly, this can be very inefficient, so Hadoop allows this to be customized by providing several other types of schedulers that ...

With Safari, you learn the way you learn best. Get unlimited access to videos, live online training, learning paths, books, interactive tutorials, and more.

Start Free Trial

No credit card required