[SPARK-8061] Document how to use "hadoop provided" builds

• Type: Improvement
• Status: Resolved
• Priority: Major
• Resolution: Duplicate
• Component/s: Documentation
• Target Version/s: 1.4.1, 1.5.0
• Assignee: Unassigned
• Reporter: Patrick Wendell
• Created: 03/Jun/15 02:14
• Updated: 09/Jun/15 19:07
• Resolved: 09/Jun/15 19:07

Description

Since we want users to use the "Hadoop provided" builds, we should explain how to use these since it's much less obvious than just downloading Spark with the built-in hadoop libraries.

My thought is to have a section in the docs for different environments (Apache Hadoop, MapR, HDP, Cloudera, etc) that shows how to modify the classpath to include Hadoop client libraries.

Issue Links

• duplicates: SPARK-6511 Publish "hadoop provided" build with instructions for different distros