Installation

Impala is comprised of a set of components that you install and run on a single node or on multiple nodes in a cluster. To run Impala in your cluster, install the required Impala packages on designated nodes. The Impala packages contain the following Impala components:

  • Impala daemon
  • Impala statestore
  • Impala catalog
  • Impala binaries

The following table lists the Impala packages and their descriptions:

Package

Description

mapr-impala

A package that contains all of the Impala binaries, including the Impala server, impala-shell, statestore, and catalog.

mapr-impala-server

The role package that installs the Impala daemon role on the node. This package enables Warden to manage the service. It is recommended (not required) that you install the Impala daemon on a node with the MapR fileserver installed.

mapr-impala-statestore

The role package that installs the Impala statestore role on the node. This package enables Warden to manage the service.

mapr-impala-catalog

The role package that installs the Impala catalog role on the node. This package enables Warden to manage the service.
NOTE: The default Impala memory setting is high, which can result in conflict between Impala and other frameworks running in the cluster. To avoid memory conflicts with Impala, modify the Impala and MapR-FS FileServer settings to ensure that each framework has enough memory to run jobs to completion. See the installation instructions for more information.

Verify that your system meets the prerequisites and then install Impala.