Introduction

The Dell EMC Data Computing Appliance provides a ready-made platform that strives to accommodate the majority of customer workloads. One of Pivotal Greenplum’s strongest value propositions is its ability to run on practically any modern-day hardware platform. More and more, Pivotal Engineering is seeing cases where customers elect to build a cluster that satisfies a specific requirement or purpose.

Pivotal Platform Engineering publishes this framework as a resource for assisting customers in this effort.

Objectives

This guide can be used for:

  • A clear understanding of what characterizes a recommended platform for running Pivotal Greenplum Database
  • A review of the two most common topologies with supporting recommended architecture diagrams
  • Pivotal recommended reference architecture that includes hardware recommendations, configuration, hard disk guidelines, network layout, installation, data loading, and verification
  • Extra guidance with real-world Greenplum cluster examples (see Pivotal Cluster Examples)

This document does:

  • provide recommendations for building a well-performing Pivotal cluster using the hardware guidelines presented
  • provide general concepts without specific tuning suggestions

This document does not:

  • promise Pivotal support for the use of third party hardware
  • assume that the information herein applies to every site, but is subject to modification depending on a customer’s specific local requirements
  • provide all-inclusive procedures for configuring Pivotal Greenplum. A subset of information is included as it pertains to deploying a Pivotal cluster.

Greenplum Terms to Know

master
A server that provides entry to the Greenplum Database system, accepts client connections and SQL queries, and distributes work to the segment instances.

segment instances
Independent PostgreSQL databases that each store a portion of the data and perform the majority of query processing.

segment host
A server that typically executes multiple Greenplum segment instances.

interconnect
Networking layer of the Greenplum Database architecture that facilitates inter-process communication between segments.

Feedback and Updates

Please send feedback and/or questions regarding content to cluster-build@pivotal.io.