You can use the instructions in this document to install the latest version of Apache Drill on MapR.
MapR categorizes certain releases of Drill as "production-ready" and certain releases as "developer previews." For more information about developer preview releases, see MapR Developer Preview.
The latest production-ready version of Drill is 1.4.
Before you install Apache Drill, make sure that the node meets the following prerequisites:
- Member of a MapR 4.x or 5.0.0 cluster
Refer to the Apache Drill Release Notes for information about known issues. Do not add HADOOP_HOME to Drill’s classpath. In fact, you do not have to set HADOOP_HOME in current releases of Drill.
Hive and HBase Support
Installation of Hive is optional. Support differs, depending on the release:
- Apache Drill 1.0 supports Hive 0.13.
- Apache Drill 1.1.0 supports Hive 1.0.
- The latest version of Drill on the MapR Distribution for Hadoop 1.2.0 supports all Hive versions available on the corresponding MapR Platform.
Installation of a supported HBase version is optional.
Configuring the maprtech.repo File
To install Drill on Red Hat, CentOS, Ubuntu or SUSE, a repository file must be configured for the Drill package. The repository for the Drill package is different from the repository for the MapR package. The following sections briefly describe how to set up the repositories on each platform. For more information about setting up repositories, see “MapR Repositories and Package Archives.”
RHEL or CentOS
On RedHat Enterprise Linux (RHEL) or CentOS, a
maprtech.repo file should be in the following directory:
To install the developer preview release of Apache Drill 1.3.0 on MapR instead of Apache Drill 1.4.0 on MapR, use the following baseurl in the maprecosystem properties:
On Ubuntu, a
sources.list file should be in the following directory:
sources.list file should contain a repository specification for installing MapR and it might also contain the specification for the ecosystem repository that you need for installing the latest version of Apache Drill. Open the file and check for the specifications. The following example shows the specifications for a MapR 5.0.0 and version 5.x ecosystem. The first deb specification installs MapR 5.0.0. The second deb specification installs an ecosystem component, such as the production-ready version of Drill:
To install the developer preview release of Apache Drill 1.3.0 on MapR instead of Apache Drill 1.4.0 on MapR, use the following URL in the second deb specification:
As root user, or using sudo, issue the following command to configure the repository for MapR ecosystem packages, which includes the latest Apache Drill package, on a MapR 5.0.0 cluster:
To install the developer preview release of Apache Drill 1.3.0 on MapR instead of Apache Drill 1.4.0 on MapR, use the following URL and change maprecosystem to maprdrill:
Using the repositories previously described, complete the following steps to install the latest version of Apache Drill:
Issue the following command to install the mapr-drill package on a node:
Code Block title RedHat or CentOS language text
$ sudo yum install mapr-drill
Code Block title Ubuntu language text
$ sudo apt-get install mapr-drill
Code Block title SUSE language text
$ sudo zypper install mapr-drill
- Run configure.sh to refresh the node configuration.
Code Block title Example language text
$ /opt/mapr/server/configure.sh -R
- Start or restart the warden to make configuration changes effective:
Code Block language text
service mapr-warden restart
- Verify that the Drillbit service is running on the node. It might take a minute or so for the Drillbit to start after starting warden. You can issue the following command to verify the status of the Drillbit service from the command line:
Code Block language text
Alternatively, login to the MCS at
https://<host name>:8443to verify the status of the Drillbit service:
- Repeat the installation process on any other nodes that you want to run Apache Drill.
Starting the Drillbit Service Manually
You can start the Drillbit service on one or more nodes using MCS or the following command:
Use the host name if possible. Using host names instead of IP addresses is a best practice. Replace
start with restart to
restart the service.
Accessing Drill Logs
You can access log files for Apache Drill in
When you have Apache Drill installed and running on selected nodes, you can perform any of the following tasks:
- Review the Apache Drill Architectural Overview.
- Connect Drill to data sources.
- Start the Drill shell and query sample data.
- Install the MapR Drill ODBC Driver to connect to Drill from BI tools.
- Install the JDBC driver and connect to Drill from SQuirreL on Windows.
- Visit the Apache Drill documentation web site for more information.