Browse Prior Art Database

%TC1% Using DB2 Partitioning as a High Availability Solution

IP.com Disclosure Number: IPCOM000028611D
Original Publication Date: 2004-May-24
Included in the Prior Art Database: 2004-May-24
Document File: 2 page(s) / 73K

Publishing Venue

IBM

Abstract

This article details how to use DB2 UDB's partitioning feature as a high availability mechanism.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 53% of the total text.

Page 1 of 2

%TC1% Using DB2 Partitioning as a High Availability Solution

The partitioning feature allows a single DB2 UDB database image to be distributed across multiple physical computers, with portions of each table residing on each system or partition. The key to using this as a high availability solution is the ability to restrict the tables to one partition even though the database is installed across multiple partitions. With a database created on tables restricted to one partition, the data is replicated to a second partition. The data on each partition is distinguished by a unique schema with select access restricted by schema to only the original data. In the case of failure of the primary partition the user access is changed to see only the data that was replicated to the secondary partition.

The only piece of custom code is a detection algorithm used to determine if the primary partition is unavailable. The detection code would be operating system specific.

A simple implementation would proceed as outlined below.

    * Install database software on two physically separate systems (Partition 0 and Partition 1).

    * Create a single instance and database spanning the two systems, with the system catalog tables residing on Partition 0.

    * Create two database partition groups, each restricted to a single partition (DBPart0 and DBPart1).

         o CREATE DATABASE PARTITION GROUP DBPart0 ON DBPARTITIONNUM 0

    * Create tablespaces and tables on Partition 1 in DBPart1 so data is restricted to Partition 1. This becomes the Primary database partition.

         o CREATE TABLESPACE <tablespace name> IN DATABASE PARTITION GROUP DBPart1 ...

* Data is loaded into Partition 1....