Db2

 View Only

Building a Recovery Strategy for an IBM Smart Analytics System Data Warehouse

By Ashley Bassman posted Fri September 30, 2022 09:22 AM

  

Developed by: 

Dale McInnis IBM DB2 Availability Architect
Garrett Fitzsimons IBM Smart Analytics System - Best Practices
Richard Lubell IBM Smart Analytics System - Information Development


Executive summary

This paper is targeted at those involved in planning, configuring, designing, implementing, or administering a data warehouse based on DB2® Database for Linux®, UNIX, and Windows® software. In particular, this paper focuses on the IBM ®Smart Analytics System environment with configurations based on System x® and Power Systems™ servers.

The focus of a backup and recovery plan should be the recovery strategy. Identifying and planning for the recovery scenarios that are most likely to occur are the key factors in determining the speed with which you can recover from data loss or corruption. The goal of the recovery strategy should be to identify what has gone wrong and correct it by using the appropriate granular backup. A backup strategy should concentrate on online table space backups, which are more flexible than the traditionally larger full database backups. The backup strategy should also identify and back up active data, thereby reducing the need to continually back up inactive data.

For maximum protection and efficient recovery of data, begin planning your database recovery strategy at project inception and carry it through each stage of the project: procurement, sizing, configuration, database design, data load design, and development. Using the recommendations in this paper will help you develop a strategy that meets current needs but is also scalable and flexible so that it meets future needs. You can see the impact of a poorly designed backup and recovery solution not only when you require a recovery, but also in the performance of the system as a whole. When you use data warehouses which provide input to strategic decision-making and operational intelligence systems, you must be able to run backup processes concurrently with business workloads, without negatively affecting those workloads.

The IBM Smart Analytics System product incorporates best practices guidelines in building a stable and high-performance data warehouse environment. This paper builds on this approach by recommending that you balance backup processes across all database partitions in your system.

Introduction

This paper covers planning a recovery strategy for an IBM Smart Analytics System data warehouse.1 A backup strategy will be a natural output of this planning. The focus is on recommendations. An IBM Smart Analytics System usage scenario is provided; you might have to modify commands that are used in this scenario for your environment. To use this paper, you should have a working knowledge of DB2 software.

A recovery strategy helps to ensure continuous availability of data. Defining your recovery objectives will dictate your backup processes. Planning backups and recovery should begin early and concentrate on recovery speed while maintaining an efficient, flexible, and balanced backup strategy.

This paper does not cover disaster recovery or operating system recovery; these topics are explored in other resources. Also, disk failures, which you can manage with RAID, and server failures, which you can manage with high availability architecture, are not covered here. This paper also does not cover offline backups, instance-level recoveries, and IBM FlashCopy® as components of an overall strategy.

The first section of the paper describes the concept of recovery objectives and considers the design of the data warehouse and load methods in planning a recovery strategy. This section highlights how decisions that you make early in the data warehouse project affect the recovery strategy and what the best practices guidelines are in this area.

The second section of the paper sets out recommendations on how to use configuration as well as backup and recovery techniques to implement a backup and recovery strategy that is balanced and focuses on the stated recovery objectives of your environment. Following these sections is a sample implementation scenario for a backup and recovery strategy, an overview of monitoring, and appendixes on the configuration of the test system, configuring a DB2 instance for TSM and DB2 utilities.

Planning a recovery strategy
Many environments have a backup strategy that highlights the backup schedule, tape location, archiving, and retention policy without addressing recovery scenarios. A scenario-driven recovery strategy should define your backup planning. This section examines how to plan a recovery strategy and how decisions that you make during the design and development of a data warehouse affect how you back up data and how long a recovery process takes to run. You should review the effect of decisions on the recovery strategy at all stages of the implementation of a data warehouse. These key recommendations are used as input to the recovery strategies that are described throughout the paper:
- Using full table space backups rather thanfull database backups. (Recovering a database from full table space backups became possible with the introduction of the RESTORE REBUILD command in DB2 9.5 software for Linux, UNIX, and Windows.) You can choose the frequency of full table space backups for both active and inactive data. This approach compensates for the added disk space and processor usage of including recovery metadata in the table space backup image.
- Using online table space backups of active data rather than offline database backups. Online table space backups support 24-hour data availability needs.
- Using frequent full table space backups rather than less frequent full backups and incremental backups in conjunction. Frequent full table space backups prioritize recovery speed over backup speed because fewer and smaller backup images and fewer transaction logs are required to restore lost data.
- Using IBM Tivoli® Storage Manager (TSM) software to assist in managing the backup and recovery process. Using TSM is recommended over taking local disk backups because integration with DB2 software reduces the complexity of managing backup images and transaction logs.

Planning a backup and recovery strategy should involve the following considerations:
- Recovery point objectives and recovery time objectives
- Physical data warehouse design
- Data warehouse load method
- Backup size and throughput capacity for recovery

Download the report to get started! 

#Db2
0 comments
13 views

Permalink