System information

About This Guide

SUSE Linux Enterprise Server is used for a broad range of usage scenarios in enter-

prise and scientific data centers. SUSE has ensured SUSE Linux Enterprise Server

is set up in a way that it accommodates different operation purposes with optimal

performance. However, SUSE Linux Enterprise Server must meet very different de-

mands when employed on a number crunching server compared to a file server, for

example.

Generally it is not possible to ship a distribution that will by default be optimized for

all kinds of workloads. Due to the simple fact that different workloads vary substan-

tially in various aspects—most importantly I/O access patterns, memory access pat-

terns, and process scheduling. A behavior that perfectly suits a certain workload might

t reduce performance of a completely different workload (for example, I/O intensive

databases usually have completely different requirements compared to CPU-intensive

tasks, such as video encoding). The great versatility of Linux makes it possible to con-

figure your system in a way that it brings out the best in each usage scenario.

This manual introduces you to means to monitor and analyze your system. It describes

methods to manage system resources and to tune your system. This guide does not

offer recipes for special scenarios, because each server has got its own different de-

mands. It rather enables you to thoroughly analyze your servers and make the most out

of them.

General Notes on System Tuning

Tuning a system requires a carefully planned proceeding. Learn which steps are

necessary to successfully improve your system.

PartII, “System Monitoring” (page7)

Linux offers a large variety of tools to monitor almost every aspect of the system.

Learn how to use these utilities and how to read and analyze the system log files.

PartIII, “Kernel Monitoring” (page69)

The Linux kernel itself offers means to examine every nut, bolt and screw of the

system. This part introduces you to SystemTap, a scripting language for writing

kernel modules that can be used to analyze and filter data. Collect debugging in-

formation and find bottlenecks by using kernel probes and use perfmon2 to ac-

cess the CPU's performance monitoring unit. Last, monitor applications with the

help of Oprofile.