Chapter 9: Effective Disk Management

Chapter 9: Effective Disk Management

Effective Disk Management

Poor disk performance can have a dramatically negative impact on efficient CPU usage. Since the executing images will be waiting for the disk I/O to complete in many instances, it is imperative that this potential throttle point be resolved. Often users or programmers will not realize the negative effects that an ill-conceived design within the data management area can have on overall application and system performance.

This chapter is about the Record Management Services (RMS) file structure. The new spiral log file system announced with Version 7 is beyond the scope of this chapter. However, I have seen various aspects of this new system and would endorse its utilization for optimizing sequentiaI I/O processing.

Finding Disk Bottlenecks

The key to effective disk usage is keeping the average response time as low as possible. The command MONITOR DISK can show average Operation Rate and Request Queue Lengths for each disk. Using the DECPS performance package can offer insight into all facets of I/O performance from DSSI through HSCs and HSJs. A high response time at the disk completion level could indicate that a disk is a good candidate for improvement. However, if the Operation Rate is low, it may indicate brief spurts of high activity, which may, or may not, be a problem that can be corrected.

If the average disk queue length is over one-half for long periods of time, the disk is a possible candidate for improvement. If this average remains...

< Previous Excerpt Next Excerpt >

Purchase This Book

High Performance Cluster Configuration System Management

TABLE OF CONTENTS

Chapter 9: Effective Disk Management

Contact Preferences

This is embarrasing...

Customize Your GlobalSpec Experience

Select Your Free Newsletters

Industry Newsletters

Select Your Free Product Alerts

This is embarrasing...