Install Hue

This topic includes instructions for using package managers to download and install Hue from the MEP repository.

For instructions on setting up the MEP repository, see Step 8: Install Ecosystem Components Manually.
Note: The Hue package, mapr-hue, can be installed on either a MapR cluster node (recommended) or a MapR client node. If you choose to install on a client node, keep in mind that Hue directories are owned by the user who installed Hue.
Execute the following commands as root or using sudo.
  1. Install the Hue packages.

    On Ubuntu:

    apt-get install mapr-hue

    On RedHat/ CentOS:

    yum install mapr-hue

    On SUSE:

    zypper install mapr-hue
    Note: To use Hue with Spark, you also need to install the mapr-hue-livy package.
  2. If the node is a MapR Client node, follow the additional instructions:
    1. To determine who the <INSTALL_USER> is, enter:
      logname
    2. Set the following properties in hue.ini to that user.
      server_user=<INSTALL_USER>
      server_group=<INSTALL_USER>
      default_user=<INSTALL_USER>
    3. Change the default_hdfs_superuser property to the owner of /var on the cluster.
      Warning: The <INSTALL_USER> must exist on the cluster on all nodes. It must also be set as the proxy user in all configuration files listed in Configure Hue, depending on the Hue version you are installing.

Installing Other Components

Based on your requirements, you may also want to install the following components on at least one node in the cluster.

Component Description
HBase or MapR-DB Required for access to HBase tables or MapR-DB tables
HttpFS Required for viewing files in MapR-FS through Hue file browser
Sqoop2 Required to create and submit jobs to transfer bulk data between structured datastores and the MapR-FS.
Spark Required to process data using the Notebook UI.
Hive Required to run queries with HiveServer2.
Oozie Required for Oozie workflows

When you finish installing Hue, go to Configure Hue to learn how to configure Hue.