Click Here!
home account info subscribe login search My ITKnowledge FAQ/help site map contact us


 
Brief Full
 Advanced
      Search
 Search Tips
To access the contents, click the chapter and section titles.

Oracle Performance Tuning and Optimization
(Publisher: Macmillan Computer Publishing)
Author(s): Edward Whalen
ISBN: 067230886x
Publication Date: 04/01/96

Bookmark It

Search this book:
 
Previous Table of Contents Next


Configure for the Disk Array

As mentioned several times in the first parts of this chapter, you must abandon some of the configuration guidelines you have used in the past to take maximum advantage of disk array technology.

With traditional disks, the database administrator tries to segment data and indexes over different disks to achieve concurrent access to both data and indexes. Furthermore, tablespaces are made to span multiple disks to balance the load on the disks.

This balancing act can be very difficult and sometimes impossible. The outcome usually results in some disks being overworked while others are underutilized. As access to the database changes over time, the “hot spots” can migrate, making the load balancing worse.

When you use a disk array, abandon the old concepts and guidelines and configure the database differently. You should still isolate the sequential and random I/O as you do with traditional disks, but now you should put all the random data and indexes on the same logical volume.

The properties of a disk array allow concurrent access to the drives. In fact, the more I/O requests you make to a disk array, the better. Up to the point at which the disks are saturated, you benefit from more requests. These I/O requests are internally sorted by disk drives and then queued and sorted by the controller to achieve maximum throughput on the disk drives.

The only way to get more performance from a randomly accessed disk is to add more disks. Therefore, the most effective way to improve system-wide I/O performance is to supply as many disks as possible for your random requests. Do this by putting all the random data on the same logical volume and putting as many disks as possible on that logical volume.

The exception to this rule is the DSS or batch processing system that can benefit from the sequential nature of loads to both the data and index files. By separating the data and indexes, you can greatly reduce the load time.

RAID Comparison

This chapter has described each of the RAID levels available, how each level works, and what performance can be expected from each level. However, there is a tradeoff involved in fault tolerance; the tradeoff involves performance, fault tolerance, and economics. Table 15.5 compares the different RAID levels.

Table 15.5 Fault Tolerance Comparison

RAID Level Performance Fault Tolerance Economics

RAID-0 (no fault tolerance) Best Worst Best
RAID-1 (mirroring) Good Best Worst
RAID-5 (data guarding) Worst Good Good

As you can see, the choice is not easy. It is up to you to decide which type of fault tolerance is best for your particular requirements.

Summary

This chapter looked at how software and hardware disk arrays operate and how you can characterize the performance of disk arrays. The chapter also looked at how to tune the disk arrays for optimal performance by isolating the sequential I/Os and spreading out the random I/Os as much as possible.

The choice of which RAID level you use may not be an easy one. You must determine the number of I/Os per second that your system must support based on your application and number of users. Use this data and your budget to determine the type of fault tolerance that fits your needs.

Probably the most important topic discussed in this chapter was the difference between how you configure your system with a disk array and how you configure your system with traditional disks. When you use a disk array, you should abandon the configuration rules you have used in the past and put all your randomly accessed data files and indexes on the same logical volume. (This recommendation is the opposite of what you have heard in the past.)

Part III of this book, “Configuring the System,” looks at some specific configurations and examples of how disk arrays can be used in each of those configurations.


Previous Table of Contents Next


Products |  Contact Us |  About Us |  Privacy  |  Ad Info  |  Home

Use of this site is subject to certain Terms & Conditions, Copyright © 1996-2000 EarthWeb Inc.
All rights reserved. Reproduction whole or in part in any form or medium without express written permission of EarthWeb is prohibited.