Installing Hive on MR3

In order to install Hive on MR3, clone the git repository of MR3.

git clone https://github.com/mr3project/mr3.git
cd mr3/

The MR3 git repository contains scripts and preset configuration files for running Hive on MR3, but it does not contain binary files. In order to download binary files, execute the script install.sh with the address of an MR3 release.

info

For running Hive on MR3 on Kubernetes, it is not necessary to download binary files because we use a pre-built Docker image available at DockerHub.

Hive 4 on MR3 with Java 17

./install.sh https://github.com/mr3project/mr3/releases/download/v2.2/hive4-mr3-2.2.tar.gz

Hive 4 on MR3 with Java 8

./install.sh https://github.com/mr3project/mr3/releases/download/v2.2/hive4-mr3-2.2-java8.tar.gz
git apply update-hive4-java8.diff

Hive 3 on MR3 with Java 17

./install.sh https://github.com/mr3project/mr3/releases/download/v2.2/hive3-mr3-2.2.tar.gz
git apply update-hive3.diff

Hive 3 on MR3 with Java 8

./install.sh https://github.com/mr3project/mr3/releases/download/v2.2/hive3-mr3-2.2-java8.tar.gz
git apply update-hive3-java8.diff

The script copies jar files of Hive on MR3 and creates symbolic links. That's it!

info

For running Hive on MR3 on Hadoop, it suffices to install Hive on MR3 only on the master node (where HiveServer2 is executed).

caution

Hive on MR3 with Java 17 requires Java 17 or later (up to Java 22). Install Java 17+ on your system, except when running Hive on MR3 on Kubernetes.

caution

The MR3 release has a worker capacity limit of 512GB. For example, the user can create up to 8 workers, each with 64GB of memory.

For larger deployments, a commercial license is available. See Pricing for more information.