Compiling Drill from Source

To develop Drill, you compile Drill from source code and then set up a project in Eclipse for use as your development environment. To review or contribute to Drill code, you must complete the steps required to install and use the Drill patch review tool.

Starting in Drill 1.15, you can build and deploy Apache Drill on the MapR platform.

Prerequisites

Apache Drill requires:

  • Apache Maven 3.6.3 or later
  • Oracle or OpenJDK 8

Apache Drill on the MapR platform also requires:

  • MapR 6.x
  • RedHat/CentOS 7.x

Run the following commands to verify that the correct versions of Maven and JDK are installed:

java -version
mvn -version

1. Clone the Repository

git clone https://gitbox.apache.org/repos/asf/drill.git

2. Compile the Code

Change to the drill directory:

cd drill

Build Apache Drill:

mvn clean install -DskipTests

If you want to deploy Apache Drill on the MapR platform, include the -Pmapr option to build Drill under the MapR profile.

mvn clean install -DskipTests -Pmapr

Introduced in release: 1.20

Drill updated its Hadoop client libraries to Hadoop 3 in release 1.17 leaving Hadoop 2 users unable to upgrade beyond Drill 1.16. As of Drill 1.20 it once again becomes possible to build Drill for deployment in an Hadoop 2 environment, by including the -Phadoop-2 option to select the Hadoop 2 profile.

mvn clean install -DskipTests -Phadoop-2

A tarball is built and appears in the distribution/target directory. Move the tarball to another directory for unpacking. Unpack the tarball and then connect to Drill and query sample data, or connect Drill to your data sources.