Step 4: Install Cluster Service Packages

The installation process varies based on the location of your packages and the configuration of your cluster.

Install services based on your cluster plan and service layout.

Before Installing Packages

Note these considerations:
  • Review security vulnerabilities: Make sure that you have reviewed the list of known vulnerabilities in Security Vulnerabilities. If a vulnerability applies to your release, contact your support representative for a fix. Apply the fix immediately, if applicable.

List of Packages by Node

The following table lists the core packages to install on cluster nodes:

On These Nodes

Install These Packages

On all [compute] cluster nodes mapr-fileserver

mapr-s3server
On designated cluster nodes mapr-cldb

mapr-zookeeper

mapr-mastgateway

mapr-nfs or mapr-loopbacknfs1

mapr-webserver2

mapr-apiserver2

mapr-gateway

On client machines that run Hadoop commands that are not already part of the cluster mapr-client

1See NFS Considerations.

2For special considerations related to the installation of the mapr-webserver and mapr-apiserver packages, see API Server and Web Server Packages for EEP 8.1.0.

WARNING
This table is a rough guide and does not include the additional non-data-fabric packages required for internal Package Dependencies or Hadoop ecosystem components.

Install the packages based on a thorough plan. For example cluster designs, see Example Cluster Designs.

To install the HPE Ezmeral Data Fabric, select one of the installation methods in the subsequent topics, depending on your operating system.

NFS Considerations

When you install mapr-nfs, NFSv3 is installed. To install NFSv4, you must use the mapr-nfs4server package. NFS is not secure by default. If you wish to configure NFSv4 server to work with Kerberos servers, you must first install Active Directory and Kerberos servers. For more information, see Installing NFS for the HPE Ezmeral Data Fabric and Configuring NFSv4 Server for Kerberos.

Consider installing mapr-loopbacknfs if you need a secure POSIX client. Note that the Installer installs mapr-loopbacknfs on all nodes in the cluster when Enable NFS is not specified. For more information about mapr-loopbacknfs, see POSIX Clients.

Hadoop and YARN Packages

With Release 6.2.0, Hadoop and YARN packages moved into the MEP repository. For more information, see Installing Hadoop and YARN.

Installing from a Repository

Before installing from the repository, change to the root user or use sudo.

  • On RedHat, CentOS, or Oracle Linux, use the yum command to install the services that you want to run on the node.

    Syntax and Example

    yum install <package_name> <package_name> <package_name> 
    yum install mapr-fileserver mapr-webserver
  • On SLES, use the zypper command to install the services that you want to run on the node.

    Syntax and Example

    zypper install <package_name> <package_name> <package_name>
    zypper install mapr-fileserver mapr-webserver
  • On Ubuntu, use the apt-get commands to update the Ubuntu package cache and install the services that you want to run on the node.
    1. Update the Ubuntu package cache: 
      apt-get update
    2. Install the services:

      Syntax and Example

      apt-get install <package_name> <package_name> <package_name> 
      apt-get install mapr-fileserver mapr-webserver

Installing from a Local Repository

Before installing from the repository, change to the root user or use sudo.

  • On RedHat, CentOS, Oracle Linux, or SLES, use rpm command to install the appropriate packages for the node:
    1. Change the working directory to the location where the rpm package files are located.
    2. Install the services:

      Syntax and Example

      yum install <package_file> <package_file> <package_file>
      yum install /path/to/mapr-core-<version>.x86_64.rpm mapr-cldb-<version>.x86_64.rpm \
              mapr-resourcemanager-<version>.x86_64.rpm mapr-webserver-<version>.x86_64.rpm \
      NOTE
      Replace <version> with the exact version string found in the package filename.
  • On Ubuntu, use the dpkg command to install the appropriate packages for the node.
    1. Change the working directory to the location where the deb package files are located.
    2. Install the services:

      Syntax and Example

      dpkg -i <package_file> <package_file> <package_file>
      dpkg -i mapr-core-<version>.x86_64.rpm mapr-cldb-<version>.x86_64.rpm \
              mapr-resourcemanager-<version>.x86_64.rpm mapr-webserver-<version>.x86_64.rpm \
      NOTE
      Replace <version> with the exact version string found in the package filename.

Installing from Package Files

When you install from package files, you must manually pre-install any dependency packages in order for the installation to succeed. Most data-fabric packages depend on the package mapr-core. Similarly, many Hadoop ecosystem components have internal dependencies. For details, see Package Dependencies.