Solving Operational Business Intelligence with InfoSphere Warehouse Advanced Edition

296 Solving Operational Business Intelligence with InfoSphere Warehouse Advanced Edition

8.2.4 Data needs of downstream systems

The availability of data to the query workload increases the value of the data

asset. The availability of the Optim High Performance Unload and DB2 Merge

backup software provides many possibilities when it comes to creating data

extracts for downstream systems including disaster recovery, latent replication,

data marts, large report listings, and the population of development, test, and

quality assurance (QA) environments.

Use Optim High Performance Unload to take cuts of data for downstream

systems without affecting the production system. Incorporate these needs into

the development of your backup schedule where possible to take full advantage

of your backup infrastructure.

From this planning and data gathering process, have the information available to

develop and implement a backup strategy that will meet your recovery objectives

and fully utilize the investment in your backup infrastructure.

8.3 Implement a backup and recovery strategy

In an environment where hot, warm, cold, and dormant data exists, it is

impractical from a cost or resources context to continue to perform full database

backups and maintain multiple backups of data that is not going to change. In

addition, restoring a table or table space from a full database backup image will

take significantly longer to complete because the restore process will have to

traverse the entire backup image or images to ensure that all data pages for the

tables specified have been recovered.

Choose a backup strategy that is simple to implement, as explained here:

򐂰 Perform online incremental and delta database backups to local disk or to

TSM using LAN-free technology.

This approach helps ensure that you only back up data that has changed on

the production database, and that changes to data not picked up as part of

your table space backup strategy is captured.

򐂰 Perform frequent online table space backups of individual hot table spaces to

local disk.

The most likely recovery scenario is typically related to active table spaces.

By having the most recent backup image available on disk, you can reduce

recovery time. In addition, the merge database backup can take advantage of

these table space backups.

Chapter 8. Building a corporate backup and recovery strategy 297

򐂰 Perform less frequent online table space backups for groups of warm table

spaces to local disk.

You do not want to have an excessive amount of backup operations

processing at the same time because this increases the resources consumed

by the backup workload. As the temperature of data cools, back it up less

frequently.

򐂰 Use DB2 Merge Backup to recreate full database and table space backup

images after each incremental or delta backup completes. This process is to

be a continual cycle.

Take advantage of the backup infrastructure where possible to complete

I/O-intensive operations away from the production database.

򐂰 Align backup and merge operations with your TSM disk pool to tape migration

policy to avoid tape contention.

The most efficient restore process is one that uses the fewest amount of steps

and references the least number of data objects.

Figure 8-1illustrates the process of using local disk database and table space

backup images to create a new merged full database backup image.

Figure 8-1 Example local disk backup strategy incorporating DB2 Merge Backup

Sun/Mon/Tue

DB delta backup

Wed

DB incr. backup

Thu/Fri

DB delta backup

Daily HOT_TS

Full TS backup

/db2bkfs/delta

/db2bkfs/incr

/db2bkfs/delta

/db2bkfs/hotTS

Recovery

history

DB2

/db2bkfs/full

Last merged backup

/db2bkfs/full

New merged backup

db2mbk-d custdb

(Saturday)