IBM iDoctor iSeries job watcher : advanced performance tool / Jim Cook et al.

Format:

Book

Author/Creator:

Cook, Jim.

Contributor:

International Business Machines Corporation. International Technical Support Organization.

Chromey, Sandi.

Edgerton, Tom.

Hatt, Debbie.

Series:

IBM redbooks.

Redbooks

Language:

English

Subjects (All):

Operating systems (Computers).

Client/server computing.

i5/OS.

Physical Description:

1 online resource (304 p.)

Edition:

1st ed.

Other Title:

iDoctor iSeries job watcher

Place of Publication:

Research Triangle Park, N.C. : IBM, International Technical Support Organization, c2005.

Language Note:

English

System Details:

text file

Summary:

This IBM Redbooks publication is intended for those familiar with most of the IBM-provided iSeries performance tools that are generally available through the i5/OS operating systems commands and iSeries Navigator interfaces and the additional-cost Performance Tools for iSeries, 5722-PT1, licensed program. iDoctor for iSeries is a set of software performance analysis tools and associated services that extend your ability to evaluate the health of your iSeries-based system by gathering detailed information and providing automated, graphical analysis of this data. One of these tools, Job Watcher, is the key next-step advanced tool for analyzing detailed performance data. This book: - Gives an overview of Job Watcher and most other IBM-provided iSeries performance measurement and management tools. - Describes the components of performance and how Job Watcher provides access to detailed performance data. - Provides examples of Job Watcher functions and its GUI in three applications: traditional RPG, SQL, and Java. - Provides Job Watcher collected data file and field definitions, and SQL query examples of this data beyond Job Watcher's array of graphical reports and drill-down information. This book's objective is to enhance the performance analyst’s proficiency in using Job Watcher as a key tool in the performance analysis tool kit.

Contents:

Front cover

Contents

Notices

Trademarks

Preface

The team that wrote this redbook

Become a published author

Comments welcome

Chapter 1. iDoctor for iSeries Job Watcher positioning

1.1 Job Watcher view of IBM-provided iSeries performance tools

1.2 Components of performance

1.3 System-level tools

1.3.1 Real-time performance tools

1.3.2 WRKSYSSTS command

1.3.3 WRKACTJOB command

1.3.4 WRKDSKSTS command

1.3.5 WRKSYSACT command

1.3.6 Collection Services

1.3.7 Management Central

1.3.8 Collect performance trace data

1.3.9 Performance Explorer

1.3.10 Database Monitor for iSeries

1.3.11 SQL performance monitors

1.3.12 iDoctor for iSeries suite of products

1.4 Performance tools for previously collected data

1.4.1 Performance Tools for iSeries Licensed Program product (5722-PT1)

1.4.2 Performance Management for iSeries

1.4.3 Performance Explorer

1.5 Using the system available performance tools

1.6 Expanded Job Watcher overview

Chapter 2. Overview of job waits and iDoctor for iSeries Job Watcher

2.1 Simplified view of running and waiting

2.2 Job Watcher terminology

2.3 Level set on wait conditions

2.4 The mysteries of waiting

2.5 Are waits bad?

2.6 Detailing waits

2.7 iDoctor for iSeries Job Watcher

2.8 Waiting point groupings (wait buckets)

2.8.1 Do wait buckets defeat the purpose of many block points?

2.8.2 Job Watcher wait points (ENUMs) and wait buckets

2.9 LIC queuing primitives and more granular wait points

2.9.1 Disclaimer

2.9.2 Bucket 1: Dispatched Time (previously referred to as CPU)

2.9.3 Bucket 2: CPU queuing

2.9.4 Bucket 3: Total block time

2.9.5 Bucket 4: Reserved

2.9.6 Bucket 5: DASD (page faults)

2.9.7 Bucket 6: DASD (non-fault reads)

2.9.8 Bucket 7: DASD space usage contention.

2.9.9 Bucket 8: Idle/waiting for work

2.9.10 Bucket 9: DASD writes

2.9.11 Bucket 10: DASD (other reads or writes)

2.9.12 Bucket 11: DASD operation start contention

2.9.13 Bucket 12: Mutex/Semaphore contention

2.9.14 Bucket 13: Journal serialization

2.9.15 Bucket 14: Machine level gate serialization

2.9.16 Bucket 15: Seize contention

2.9.17 Bucket 16: Database record lock contention

2.9.18 Bucket 17: Object lock contention

2.9.19 Bucket 18: Other waits

2.9.20 Bucket 19: Main storage pool overcommitment

2.9.21 Bucket 20: Java user (including locks)

2.9.22 Bucket 21: Java JVM

2.9.23 Bucket 22: Java (other)

2.9.24 Bucket 23: Socket accepts

2.9.25 Bucket 24: Socket transmits

2.9.26 Bucket 25: Socket receives

2.9.27 Bucket 26: Socket (other)

2.9.28 Bucket 27: IFS pipe

2.9.29 Bucket 28: IFS (other)

2.9.30 Bucket 29: Data queue receives

2.9.31 Bucket 30: MI queue (other)

2.9.32 Bucket 31: MI wait on events

2.9.33 Bucket 32: Abnormal contention

2.10 Management Central's use of wait bucket mapping

Chapter 3. Getting started

3.1 Enhancements to Job Watcher at V5R3M0

3.1.1 The collection engine

3.1.2 GUI

3.1.3 Naming

3.2 How to obtain Job Watcher

3.3 Starting a Job Watcher collection

3.3.1 Connecting to your system

3.3.2 Access Job Watcher

3.3.3 Starting the Job Watcher collection wizard

3.3.4 Data collection options

3.3.5 Job and task options

3.3.6 More on selecting specific jobs and tasks

3.3.7 Ending Options window

3.3.8 Job Watcher Wizard summary

3.3.9 Active Job Watcher

3.4 Stopping a Job Watcher collection

3.5 Viewing Job Watcher data

3.5.1 Job Watcher view

3.5.2 iDoctor preferences

3.5.3 Reports and graphs: Example 1

3.5.4 Graphs and reports: Example 2

3.5.5 Tabs in the Interval Details window.

3.5.6 Additional Job Watcher graphs

Chapter 4. Analysis example: traditional batch ILE RPG application

4.1 Analyzing the problem

4.2 Activation group usage

Chapter 5. SQL, call stack, and journal analysis examples

5.1 Our example application and preliminary performance analysis

5.2 SQL analysis

5.2.1 Getting started

5.2.2 Viewing an active SQL statement using the GUI

5.2.3 Data Viewer

5.2.4 Finding jobs running most SQL statements

5.3 Call stack analysis for task counts

5.3.1 Produce the resolved call stack file

5.3.2 The call stack for jobs with active SQL statements

5.3.3 Finding user programs running active SQL statements

5.3.4 Finding any SQL-related or database-related system activity

5.4 Journal analysis

5.4.1 No journal caching

5.4.2 Journal caching

Chapter 6. Analysis example: Java application

6.1 Two problem scenarios

6.2 CPU "misusage"

6.2.1 Job Watcher

6.3 Heap growth

6.3.1 Heap Analyzer

Appendix A. Installing and uninstalling Job Watcher details

Installation requirements

Installing Job Watcher

Uninstalling Job Watcher

Server side

Client side

Appendix B. Database files created by Job Watcher

Job Watcher terminology

Files created by every Job Watcher collection

Files created only when call stack information is specified in a Job Watcher collection

Files created only when SQL information is specified in a Job Watcher collection

Files created only when activation group information is specified in a Job Watcher collection

Files created only when sockets information is specified in a Job Watcher collection

Description of ENUM to queueing bucket identifier mappings

Description of ENUM to queueing bucket mappings

Getting started writing queries over Job Watcher collection data.

The master file QAPYJWTDE for jobs, threads, and tasks

Status file QAPYJWSTS

Job file QAPYJWPRC

Task dispatchable unit priorities scheme

Appendix C. Querying and graphing tips for Job Watcher

Querying Job Watcher data

Query definition interface

Accessing the query definition interface

Viewing your query

Saving your new query definition

Working with query definitions

SQL Query view

Accessing the SQL Query view

Changing field (column) headings in your query results

Graph views

User-defined graphs

Appendix D. Overview of Job Watcher 5250 commands

Job Watcher 5250 commands

WCHJOB command

WCHJOB parameters

CPYJWCOL command

CPYJWCOL parameters

DLTJWCOL command

DLTJWCOL parameters

FTPJWCOL command

FTPJWCOL parameters

RSTJWCOL command

RSTJWCOL parameters

SAVJWCOL command

SAVJWCOL parameters

ADDPRDACS command

RTVSTKDTA command

tRTVSTKDTA parameters

Automatically submit a Job Watch

Appendix E. Job Watcher advanced topics

Collection specification tips

Interval size

Ensure job names are captured for all jobs on the system

Collect data as fast as possible?

The Job Watcher collection process in detail

Task count state changes

Waits that span multiple intervals

Job run/wait profile: how to fill in blank intervals

Before and after graphs for filled blank interval support

Job Watcher limitations

Glossary

Related publications

IBM Redbooks

Online resources

How to get IBM Redbooks

Using the SQL examples in this redbook

IBM Support and downloads

IBM Global Services

Index

Back cover.

Notes:

"March 2005."

Includes bibliographical references and index.

OCLC:

80244093