Tivoli Monitoring for Databases: DB2 Agent


Version 6.1.0

User’s Guide

SC32-9449-00
 Tivoli Monitoring for Databases: DB2 Agent
®


Version 6.1.0

User’s Guide

SC32-9449-00
 Note
Before using this information and the product it supports, read the information in Appendix E, “Notices,” on page 207.

First Edition (November 2005)

This edition applies to version 6.1 of the IBM Tivoli Monitoring for Databases: DB2 Agent (5724-B96) and to all
subsequent releases and modifications until otherwise indicated in new editions.
© Copyright International Business Machines Corporation 2005. All rights reserved.
US Government Users Restricted Rights – Use, duplication or disclosure restricted by GSA ADP Schedule Contract
with IBM Corp.
Contents
Tables . . . . . . . . . . . . . . . v Database . . . . . . . . . . . . . . 20
Locking Conflict workspace . . . . . . . . 24
About this guide . . . . . . . . . . vii System Overview . . . . . . . . . . . 24
Table Space workspace . . . . . . . . . 25
Who should read this guide . . . . . . . . . vii
What this guide contains . . . . . . . . . . vii
Publications . . . . . . . . . . . . . . viii Chapter 5. Attributes reference . . . . 27
Prerequisite publications . . . . . . . . . viii About attributes . . . . . . . . . . . . . 27
Related publications . . . . . . . . . . ix More information about attributes . . . . . . . 27
Accessing terminology online . . . . . . . ix Attribute groups and attributes for the Monitoring
Accessing publications online . . . . . . . ix Agent for DB2 . . . . . . . . . . . . . 28
Ordering publications . . . . . . . . . . ix Application attributes (group 00) . . . . . . 28
Accessibility . . . . . . . . . . . . . . x Application attributes (Unicode group 00) . . . 46
Tivoli technical training . . . . . . . . . . . x Application attributes (group 01) . . . . . . 65
Support information . . . . . . . . . . . . x Buffer Pool Data attributes . . . . . . . . 68
Conventions used in this guide . . . . . . . . x Database Information attributes (group 00) . . . 75
Typeface conventions . . . . . . . . . . x Database Information attributes (group 01) . . . 92
Operating system-dependent variables and paths xi Locking Conflict attributes . . . . . . . . 105
System Overview attributes. . . . . . . . 110
Chapter 1. Overview of the Monitoring Tablespace attributes . . . . . . . . . . 120
Disk capacity planning for historical data . . . . 128
Agent for DB2 . . . . . . . . . . . . 1
IBM Tivoli Monitoring overview . . . . . . . . 1
Features of the Monitoring Agent for DB2 . . . . 1
Chapter 6. Situations reference . . . . 131
Monitoring Agent for DB2 components . . . . . 2 About situations . . . . . . . . . . . . 131
User interface options . . . . . . . . . . . 2 More information about situations . . . . . . 131
Predefined situations . . . . . . . . . . . 132
UDB_Agent_WaitToken_High situation . . . . 133
Chapter 2. Requirements and UDB_Agents_Stolen_High situation . . . . . 133
configuration for the monitoring agent . 3 UDB_Appl_BP_Hit_Ratio_Low situation . . . 133
Requirements for the monitoring agent . . . . . 3 UDB_Appl_CatCache_Hit_Low situation . . . 134
Configuration settings for the monitoring agent . . 4 UDB_Appl_HJoinOflws_High situation . . . . 134
DB2 properties for agent deployment . . . . . . 4 UDB_Appl_HJoinSmOflw_High situation . . . 134
tacmd addSystem command parameter for the UDB_Appl_Lock_Warning situation . . . . . 134
monitoring agent . . . . . . . . . . . . . 5 UDB_Appl_PkgCache_Hit_Low situation . . . 134
Starting and stopping the Monitoring Agent for DB2 UDB_Appl_SQL_Fail_High situation . . . . 134
on UNIX operating systems . . . . . . . . . 5 UDB_Appl_Wait_Lock situation . . . . . . 134
Background information . . . . . . . . . 5 UDB_BP_DrtyPg_Steal_Clns situation . . . . 134
Procedure . . . . . . . . . . . . . . 5 UDB_BP_DrtyPg_thresh_Clns situation . . . . 135
Creating an ODBC data source . . . . . . . . 6 UDB_BP_Hit_Ratio_Low situation . . . . . 135
UDB_Buff_Max_Used_Pct_Crit situation . . . 135
Chapter 3. How to use the Monitoring UDB_Buff_Max_Used_Pct_Warn situation . . . 135
Agent for DB2 . . . . . . . . . . . . 9 UDB_Buff_Used_Pct_Crit situation . . . . . 135
View real-time data about DB2 Universal Database . 9 UDB_Buff_Used_Pct_Warn situation . . . . . 135
Investigate an event . . . . . . . . . . . 10 UDB_Ce_Max_Used_Pct_Crit situation . . . . 135
Recover the operation of a resource . . . . . . 10 UDB_Ce_Max_Used_Pct_Warn situation . . . 135
Customize your monitoring environment . . . . 11 UDB_Ce_Used_Pct_Crit situation . . . . . . 135
Monitor with custom situations that meet your UDB_Ce_Used_Pct_Warn situation . . . . . 135
requirements . . . . . . . . . . . . . . 12 UDB_Database_Lock_Warning situation . . . 136
Collect and view historical data . . . . . . . 13 UDB_DB_BP_Hit_Ratio_Low situation . . . . 136
UDB_DB_Cat_Cache_Hit_Ratio_Crit situation 136
UDB_DB_Cat_Cache_Hit_Ratio_Warn situation 136
Chapter 4. Workspaces reference . . . 15
UDB_DB_Cur_Cons_Pct_Crit situation . . . . 136
About workspaces . . . . . . . . . . . . 15
UDB_DB_Cur_Cons_Pct_Warn situation . . . 136
More information about workspaces . . . . . . 15
UDB_DB_Dlk_Rb_Pct_For_Int_Crit situation . . 136
Predefined workspaces . . . . . . . . . . 15
UDB_DB_Dlk_Rb_Pct_For_Int_Warn situation 137
Application . . . . . . . . . . . . . 16
UDB_DB_File_Closed_High situation . . . . 137
Buffer Pool workspace . . . . . . . . . . 20

© Copyright IBM Corp. 2005 iii

 UDB_DB_Int_Deadlock_Rb_Pct_Crit situation 137 More information about policies . . . . . . . 153
UDB_DB_Int_Deadlock_Rb_Pct_Warn situation 137 Predefined policies . . . . . . . . . . . 153
UDB_DB_Invalid_Pkgs_Crit situation . . . . 137
UDB_DB_Invalid_Pkgs_Warn situation . . . . 137 Appendix A. IBM Tivoli Monitoring for
UDB_DB_Invalid_Sys_Pkgs_Crit situation . . . 137 DB2 Workspaces . . . . . . . . . . 155
UDB_DB_Invalid_Sys_Pkgs_Warn situation . . 137
Application Top Ten Summary workspace . . . . 155
UDB_DB_Invalid_Triggers_Crit situation . . . 137
Scenario 1: Appropriate value for LOCK
UDB_DB_Invalid_Triggers_Warn situation . . . 137
TIMEOUT parameter . . . . . . . . . . 155
UDB_DB_Lock_Waits_Pct_Crit situation . . . 138
Scenario 2: Establishing an appropriate amount
UDB_DB_Lock_Waits_Pct_Warn situation . . . 138
of sort memory . . . . . . . . . . . . 156
UDB_DB_Pool_Hit_Idx_Pct_Crit situation . . . 138
Buffer Pool workspace . . . . . . . . . . 156
UDB_DB_Pool_Hit_Idx_Pct_Warn situation . . 138
Scenario: Monitoring buffer pool efficiency . . 156
UDB_DB_Pool_Hit_Ratio_Pct_Crit situation . . 138
Database Overview workspace . . . . . . . 156
UDB_DB_Pool_Hit_Ratio_Pct_Warn situation 138
Scenario 1: Determining the maximum number
UDB_DB_Pri_Log_Used_Pct_Crit situation . . 138
of open files . . . . . . . . . . . . . 157
UDB_DB_Pri_Log_Used_Pct_Warn situation . . 138
Scenario 2: Setting the value of the
UDB_DB_Sec_Log_Used_Pct_Crit situation . . 138
LOCKTIMEOUT parameter. . . . . . . . 157
UDB_DB_Sec_Log_Used_Pct_Warn situation . . 138
UDB_DB_Sort_Overflow_High situation . . . 139
UDB_DB_SQL_Fail_High situation . . . . . 139 Appendix B. IBM Tivoli Enterprise
UDB_DB_SQL_Stmts_Fail_Pct_Crit situation . . 139 Console event mapping . . . . . . . 159
UDB_DB_SQL_Stmts_Fail_Pct_Warn situation 139
UDB_Ma_Max_Used_Pct_Crit situation . . . . 139 Appendix C. Problem determination 179
UDB_Ma_Max_Used_Pct_Warn situation . . . 139 Gathering product information for IBM Software
UDB_Max_Agent_Overflows_High situation . . 139 Support . . . . . . . . . . . . . . . 179
UDB_Pip_Sort_Hit_Ratio_Pct_Crit situation . . 139 Built-in problem determination features . . . . 179
UDB_Pip_Sort_Hit_Ratio_Pct_Warn situation 139 Problem classification. . . . . . . . . . . 180
UDB_Piped_Sorts_Rej_Pct_Crit situation . . . 139 Trace logging . . . . . . . . . . . . . 180
UDB_Piped_Sorts_Rej_Pct_Warn situation . . . 140 Overview of log file management . . . . . 180
UDB_Post_Threshold_Sorts_High situation . . 140 Examples of trace logging . . . . . . . . 181
UDB_Rb_Max_Used_Pct_Crit situation . . . . 140 Principal trace log files . . . . . . . . . 181
UDB_Rb_Max_Used_Pct_Warn situation . . . 140 Setting RAS trace parameters . . . . . . . 184
UDB_Rb_Used_Pct_Crit situation . . . . . . 140 Problems and workarounds . . . . . . . . 186
UDB_Rb_Used_Pct_Warn situation . . . . . 140 Installation and configuration problem
UDB_Status_Warning situation . . . . . . 140 determination . . . . . . . . . . . . 187
UDB_TS_Sp_Used_DMS_Tab_Pct_Crit situation 140 Agent problem determination . . . . . . . 192
UDB_TS_Sp_Used_DMS_Tab_Pct_Warn Tivoli Enterprise Portal problem determination 195
situation . . . . . . . . . . . . . . 140 Problem determination for remote deployment 195
UDB_TS_Status_Warn situation . . . . . . 140 Workspace problem determination . . . . . 196
Situation problem determination . . . . . . 196
Chapter 7. Take Action commands Take Action command problem determination 200
reference . . . . . . . . . . . . . 143 Support information . . . . . . . . . . . 200
About Take Action commands . . . . . . . . 143 Searching knowledge bases . . . . . . . . 201
More information about Take Action commands 143 Obtaining fixes . . . . . . . . . . . . 201
Predefined Take Action commands . . . . . . 143 Receiving weekly support updates . . . . . 202
Backup Database action . . . . . . . . . 144 Contacting IBM Software Support . . . . . 202
Rebind All Packages action . . . . . . . . 145
Rebind Package action . . . . . . . . . 146 Appendix D. Accessibility . . . . . . 205
Reorg Table action . . . . . . . . . . . 146 Navigating the interface using the keyboard . . . 205
Run Statistics action . . . . . . . . . . 147 Magnifying what is displayed on the screen . . . 205
Start DB2 action . . . . . . . . . . . 149
Stop DB2 action . . . . . . . . . . . 149 Appendix E. Notices . . . . . . . . 207
Update Database Configuration action . . . . 150 Trademarks . . . . . . . . . . . . . . 209
Update DB Manager Configuration action . . . 150
Index . . . . . . . . . . . . . . . 211
Chapter 8. Policies reference. . . . . 153
About policies . . . . . . . . . . . . . 153

iv IBM Tivoli Monitoring for Databases: DB2 Agent: User’s Guide

Tables
1. Monitoring agent requirements . . . . . . 3 15. General problems and solutions for
2. Viewing real-time data about DB2 Universal uninstallation . . . . . . . . . . . 191
Database . . . . . . . . . . . . . . 9 16. Agent problems and solutions on DB2 192
3. Investigating an event . . . . . . . . . 10 17. Agent problems and solutions . . . . . . 193
4. Recovering the operation of a resource . . . 11 18. Tivoli Enterprise Portal problems and
5. Customizing your monitoring environment 11 solutions . . . . . . . . . . . . . 195
6. Monitoring with custom situations . . . . . 13 19. Remote deployment problems and solutions 195
7. Collecting and viewing historical data . . . 14 20. Workspace problems and solutions . . . . 196
8. Capacity planning for historical data 129 21. Specific situation problems and solutions 197
9. Overview of event slots to event classes 160 22. Performance Impact by attribute group 198
10. Information to gather before contacting IBM 23. Problems with configuring situations that you
Software Support . . . . . . . . . . 179 solve in the Situation Editor . . . . . . 198
11. Trace log files for troubleshooting agents 182 24. Problems with configuration of situations that
12. Problems and solutions for installation and you solve in the Workspace area . . . . . 199
configuration for agents that run on UNIX 25. Problems with configuration of situations that
systems . . . . . . . . . . . . . 187 you solve in the Manage Tivoli Enterprise
13. Problems and solutions for installation and Monitoring Services window . . . . . . 200
configuration on Windows . . . . . . . 189 26. Take Action commands problems and
14. Problems and solutions for installation and solutions . . . . . . . . . . . . . 200
configuration of DB2 . . . . . . . . . 190

© Copyright IBM Corp. 2005 v

vi IBM Tivoli Monitoring for Databases: DB2 Agent: User’s Guide
About this guide
The IBM Tivoli Monitoring for Databases: DB2 Agent User’s Guide provides
information about using IBM® Tivoli® Monitoring for Databases: DB2® Agent.

Use the requirements and configuration chapter in this guide along with the IBM
Tivoli Monitoring Installation and Setup Guide to install and set up the software.

Use the information in this guide along with the IBM Tivoli Monitoring User’s Guide
to monitor DB2 Universal Database™ (DB2 UDB; hereafter referred to as DB2 in
this document).

Who should read this guide

This guide is for system administrators who install and use the Monitoring Agent
for DB2 to monitor and manage DB2 resources.

Readers should be familiar with the following topics:

v Tivoli Enterprise™ Portal interface
v IBM Tivoli Monitoring application software
v IBM Tivoli Enterprise Console® (optional)
v DB2
v Microsoft® Windows® operating systems
v AIX® operating systems
v Solaris operating systems
v HP-UX operating systems

What this guide contains

This guide contains the following chapters:
v Chapter 1, “Overview of the Monitoring Agent for DB2,” on page 1
Provides an introduction to the Monitoring Agent for DB2.
v Chapter 2, “Requirements and configuration for the monitoring agent,” on page
3
Provides requirements and configuration information specific to the Monitoring
Agent for DB2.
v Chapter 3, “How to use the Monitoring Agent for DB2,” on page 9
Provides a list of tasks to perform when using the monitoring agent, a list of
procedures for completing each task, and references for where to find
information about the procedures. After completing installation and
configuration and becoming familiar with the information in Chapter 1 of this
guide, use this chapter to see how you can use the monitoring agent.
v Chapter 4, “Workspaces reference,” on page 15
Provides an overview of workspaces, references to additional information about
workspaces, and descriptions of predefined workspaces in this monitoring agent.
v Chapter 5, “Attributes reference,” on page 27

© Copyright IBM Corp. 2005 vii

 Provides an overview of attributes, references to additional information about
attributes, descriptions of the attribute groups and attributes in this monitoring
agent, and disk space requirements for historical data.
v Chapter 6, “Situations reference,” on page 131
Provides an overview of situations, references to additional information about
situations, and descriptions of the predefined situations in this monitoring agent.
v Chapter 7, “Take Action commands reference,” on page 143
Provides detailed information about the Take Action commands, references to
additional information about Take Action commands, and descriptions of the
Take Action commands provided in this monitoring agent.
v Chapter 8, “Policies reference,” on page 153
Provides an overview of policies, references for detailed information about
policies, and descriptions of the predefined policies included in this monitoring
agent.
v Appendix A, “IBM Tivoli Monitoring for DB2 Workspaces,” on page 155
Provides additional information about the predefined workspaces provided with
this monitoring agent.
v Appendix B, “IBM Tivoli Enterprise Console event mapping,” on page 159
Provides an overview of the IBM Tivoli Enterprise Console event mapping
information for this monitoring agent.
v Appendix C, “Problem determination,” on page 179
Provides information about troubleshooting the various components of the
Monitoring Agent for DB2, information about log files and messages, and
information about your options for obtaining software support.
v Appendix D, “Accessibility,” on page 205
Provides information about the accessibility features in the Monitoring Agent for
DB2.
v Appendix E, “Notices,” on page 207
Provides IBM and Tivoli notices and trademark information as it applies to the
Monitoring Agent for DB2.

Publications
This section lists publications relevant to the use of the Monitoring Agent for DB2.
It also describes how to access these publications online and how to order these
publications.

Prerequisite publications
To use the information in this guide effectively, you must have some knowledge of
IBM Tivoli Monitoring products, which you can obtain from the following
documentation:
v IBM Tivoli Monitoring Administrator’s Guide
v IBM Tivoli Monitoring Installation and Setup Guide
v IBM Tivoli Monitoring Problem Determination Guide
v IBM Tivoli Monitoring User’s Guide
v Introducing IBM Tivoli Monitoring Version 6.1

viii IBM Tivoli Monitoring for Databases: DB2 Agent: User’s Guide
Related publications
The following documents also provide useful information:
v IBM Tivoli Enterprise Console Adapters Guide
v IBM Tivoli Enterprise Console Event Integration Facility User’s Guide
v IBM Tivoli Enterprise Console Reference Manual
v IBM Tivoli Enterprise Console Rule Developer’s Guide

Accessing terminology online

The Tivoli Software Glossary includes definitions for many of the technical terms
related to Tivoli software. The Tivoli Software Glossary is available at the following
Tivoli software library Web site:

http://publib.boulder.ibm.com/tividd/glossary/tivoliglossarymst.htm

The IBM Terminology Web site consolidates the terminology from IBM product
libraries in one convenient location. You can access the Terminology Web site at the
following Web address:

http://www.ibm.com/ibm/terminology

Accessing publications online

IBM posts publications for this and all other Tivoli products, as they become
available and whenever they are updated, to the Tivoli software information center
Web site. Access the Tivoli software information center by first going to the Tivoli
software library at the following Web address:

http://www.ibm.com/software/tivoli/library

Scroll down and click the Product manuals link. In the Tivoli Technical Product
Documents Alphabetical Listing window, click M to access all of the IBM Tivoli
Monitoring product manuals.

Note: If you print PDF documents on other than letter-sized paper, set the option
in the File → Print window that allows Adobe Reader to print letter-sized
pages on your local paper.

Ordering publications
You can order many Tivoli publications online at the following Web site:

http://www.elink.ibmlink.ibm.com/public/applications/
publications/cgibin/pbi.cgi

You can also order by telephone by calling one of these numbers:

v In the United States: 800-879-2755
v In Canada: 800-426-4968

In other countries, contact your software account representative to order Tivoli

publications.

About this guide ix

Accessibility
Accessibility features help users with a physical disability, such as restricted
mobility or limited vision, to use software products successfully. With this product,
you can use assistive technologies to hear and navigate the interface. You can also
use the keyboard instead of the mouse to operate most features of the graphical
user interface.

For additional information, see Appendix D, “Accessibility,” on page 205.

Tivoli technical training

For Tivoli technical training information, refer to the following IBM Tivoli
Education Web site:

http://www.ibm.com/software/tivoli/education/

Support information
“Support information” on page 200 describes the following options for obtaining
support for IBM products:
v “Searching knowledge bases” on page 201
v “Obtaining fixes” on page 201
v “Contacting IBM Software Support” on page 202

Conventions used in this guide

This guide uses several conventions for special terms and actions, and operating
system-dependent commands and paths.

Typeface conventions
This guide uses the following typeface conventions:
Bold
v Lowercase commands and mixed case commands that are otherwise
difficult to distinguish from surrounding text
v Interface controls (check boxes, push buttons, radio buttons, spin
buttons, fields, folders, icons, list boxes, items inside list boxes,
multicolumn lists, containers, menu choices, menu names, tabs, property
sheets), labels (such as Tip:, and Operating system considerations:)
v Keywords and parameters in text
Italic
v Words defined in text
v Emphasis of words
v New terms in text (except in a definition list)
v Variables and values you must provide
Monospace
v Examples and code examples
v File names, programming keywords, and other elements that are difficult
to distinguish from surrounding text
v Message text and prompts addressed to the user

x IBM Tivoli Monitoring for Databases: DB2 Agent: User’s Guide

 v Text that the user must type
v Values for arguments or command options

Operating system-dependent variables and paths

The direction of the slash for directory paths might vary in this documentation. No
matter which type of slash you see in the documentation, use the following
guidelines for a slash:
v If using UNIX®, use a forward slash (/).
v If using Windows, use a backslash (\).

The names of environment variables are not always the same in Windows and
UNIX. For example, %TEMP% in Windows is equivalent to $TMPDIR in UNIX.

For environment variables, use the following guidelines:

v If using UNIX, use $variable.
v If using Windows, use %variable%.

Note: If you are using the bash shell on a Windows system, you can use the UNIX
conventions.

About this guide xi

xii IBM Tivoli Monitoring for Databases: DB2 Agent: User’s Guide
Chapter 1. Overview of the Monitoring Agent for DB2
The Monitoring Agent for DB2 provides you with the capability to monitor DB2,
and to perform basic actions with DB2. This chapter provides a description of the
features, components, and interface options for the Monitoring Agent for DB2.

IBM Tivoli Monitoring overview

IBM Tivoli Monitoring is the base software for the Monitoring Agent for DB2. IBM
Tivoli Monitoring provides a way to monitor the availability and performance of
all the systems in your enterprise from one or several designated workstations. It
also provides useful historical data that you can use to track trends and to
troubleshoot system problems.

You can use IBM Tivoli Monitoring to do the following:

v Monitor for alerts on the systems that you are managing by using predefined
situations or custom situations.
v Establish your own performance thresholds.
v Trace the causes leading to an alert.
v Gather comprehensive data about system conditions.
v Use policies to perform actions, schedule work, and automate manual tasks.

The Tivoli Enterprise Portal is the interface for IBM Tivoli Monitoring products. By
providing a consolidated view of your environment, the Tivoli Enterprise Portal
permits you to monitor and resolve performance issues throughout the enterprise.

See the IBM Tivoli Monitoring publications listed in “Prerequisite publications” on

page viii for complete information about IBM Tivoli Monitoring and the Tivoli
Enterprise Portal.

Features of the Monitoring Agent for DB2

The Monitoring Agent for DB2 offers a central point of management for your DB2
environment. It provides a comprehensive means for gathering exactly the
information you need to detect problems early and to prevent them. Information is
standardized across the system. You can monitor a multitude of servers from a
single workstation. The Monitoring Agent for DB2 lets you easily collect and
analyze specific information, including information on:
v Applications with the highest percentage of failed SQL statements, sort
overflows, lock timeouts and deadlocks, and the lowest buffer pool hit ratio
v Buffer pool hit ratio by buffer pool, buffer pool hit ratio by database, average
read and write times, asynchronous and synchronous I/O activity, extended
store and non-buffer pool I/O activity
v Databases with the highest percentage of failed SQL statements, the lowest
buffer pool hit ratio, and the highest number of connections, lock timeouts, and
deadlocks
v Applications currently waiting for locks and other details about lock resources
v Server key events, the number of server connections, the databases with the
lowest buffer pool hit ratio, and applications with the highest percentage of
failed SQL statements

© Copyright IBM Corp. 2005 1

 v Tablespaces

Monitoring Agent for DB2 components

After you install and set up the Monitoring Agent for DB2, you have an
environment that contains the client, server, and monitoring agent implementation
for IBM Tivoli Monitoring that contains the following components:
v Tivoli Enterprise Portal client with a Java-based user interface for viewing and
monitoring your enterprise.
v Tivoli Enterprise Portal Server that is placed between the client and the Tivoli
Enterprise Monitoring Server and enables retrieval, manipulation, and analysis
of data from the monitoring agents.
v Tivoli Enterprise Monitoring Server, which acts as a collection and control point
for alerts received from the monitoring agents, and collects their performance
and availability data.
v Monitoring agent, the Monitoring Agent for DB2, installed on the systems or
subsystems that you want to monitor. This monitoring agent collects and
distributes data to a Tivoli Enterprise Portal Server.

For both of the IBM Tivoli Monitoring environments (IBM Tivoli Monitoring 5.x
and IBM Tivoli Monitoring 6.1), IBM Tivoli Enterprise Console is an optional
component, which acts as a central collection point for events from a variety of
sources, including those from other Tivoli software applications, Tivoli partner
applications, custom applications, network management platforms, and relational
database systems. You can view these events through the Tivoli Enterprise Portal
(using the event viewer), and you can forward events from IBM Tivoli Monitoring
situations to the IBM Tivoli Enterprise Console component.

User interface options

Installation of the base software and other integrated applications provides the
following interfaces that you can use to work with your resources and data:
Tivoli Enterprise Portal browser client interface
The browser interface is automatically installed with Tivoli Enterprise
Portal. To start Tivoli Enterprise Portal in your Internet browser, enter the
URL for a specific Tivoli Enterprise Portal browser client installed on your
Web server.
Tivoli Enterprise Portal desktop client interface
The desktop interface is a Java-based graphical user interface (GUI) on a
Windows workstation.
IBM Tivoli Enterprise Console
Event management application
Manage Tivoli Enterprise Monitoring Services window
The window for the Manage Tivoli Enterprise Monitoring Services utility is
used for configuring the monitoring agent and starting Tivoli services not
already designated to start automatically.

2 IBM Tivoli Monitoring for Databases: DB2 Agent: User’s Guide

Chapter 2. Requirements and configuration for the monitoring
agent
This chapter contains information about the following topics and procedures
relevant to the installation and configuration of the Monitoring Agent for DB2:
v “Requirements for the monitoring agent”
v “Configuration settings for the monitoring agent” on page 4
v “DB2 properties for agent deployment” on page 4
v “tacmd addSystem command parameter for the monitoring agent” on page 5
v “Starting and stopping the Monitoring Agent for DB2 on UNIX operating
systems” on page 5
v “Creating an ODBC data source” on page 6

Requirements for the monitoring agent

In addition to the requirements described in the IBM Tivoli Monitoring Installation
and Setup Guide, the Monitoring Agent for DB2 has the requirements listed in
Table 1.
Table 1. Monitoring agent requirements
Operating system Windows UNIX
Operating system v Windows 2000 Server (32-bit) v AIX 5.1, 5.2, 5.3 (32/64 bit)
versions
v Windows 2000 Advanced Server v Solaris 8, 9, 10 (32/64 bit)
(32-bit) v HP-UX 11i (32/64 bit)
v Windows 2003 Server SE
(32-bit)
v Windows 2003 Server EE
(32-bit)
Application DB2 UDB version 8.1 with fix DB2 UDB version 8.1 with fix
versions pack 7b, or version 8.2 with fix pack 7b or version 8.2 with fix
pack 9a (32 bit only on Windows) pack 9a (32 and 64 bit on UNIX)
Note: If the Monitoring Agent for
DB2 is being installed on the same
system as a Tivoli Enterprise
Portal Server, you must have DB2
UDB V8 installed.
Memory v 32 MB RAM v 32 MB RAM
v 150 MB virtual memory, plus 5 v 150 MB virtual memory, plus 5
MB for each agent installed MB for each agent installed
Disk space Space is required only for Space is required only for
historical data collection. See historical data collection. See
“Disk capacity planning for “Disk capacity planning for
historical data” on page 128. historical data” on page 128.

© Copyright IBM Corp. 2005 3

 Table 1. Monitoring agent requirements (continued)
Operating system Windows UNIX
Other None. At installation, the Monitoring
requirements Agent for DB2 must match both
the bit level of the DB2 instance
to be monitored and the bit level
of the operating system (agent bit
level = DB2 instance bit level =
OS bit level). Otherwise, it
cannot monitor the instance.

For example, if you have AIX 5.2

(64-bit) and a 64-bit DB2 instance,
install the 64-bit Monitoring Agent
for DB2 to match the DB2 instance
and the OS. If instead you have
AIX 5.2 (32-bit) and a DB2 32-bit
instance, install the 32-bit
Monitoring Agent for DB2.

The Monitoring Agent for DB2

only supports monitoring of the
same bit DB2 instance and OS.
The monitoring agent does not
support a 32-bit DB2 instance on
AIX (64-bit).

Configuration settings for the monitoring agent

Never attempt to start the monitoring agent until you have completed the
configuration steps appropriate to installation of the monitoring agent.

After completing the steps to install the Monitoring Agent for DB2 as described in
the IBM Tivoli Monitoring Installation and Setup Guide, ″Installing monitoring
agents,″ you configure the Monitoring Agent for DB2.

As you configure the Monitoring Agent for DB2, use the following settings.

On Windows:
DB2 instance name
The name of the DB2 instance.

DB2 properties for agent deployment

If you are using the Tivoli Enterprise Portal to deploy the Monitoring Agent for
DB2 from a central location as described in ″Deploying through the portal″ in the
IBM Tivoli Monitoring Installation and Setup Guide, enter information for the
following tabs:
v DB2 Properties
DB2 Instance Name
The name of the DB2 instance.
v Agent
When entering information under Run As, be sure to enter the name of the DB2
instance owner in the Username field.

4 IBM Tivoli Monitoring for Databases: DB2 Agent: User’s Guide

tacmd addSystem command parameter for the monitoring agent
The following parameters are for the Monitoring Agent for DB2 when using the
tacmd addSystem command as described in ″Deploying monitoring across your
environment″ in the IBM Tivoli Monitoring Installation and Setup Guide. Specify these
properties with the -p or -property option available with the tacmd addSystem
option:
INSTANCE
Specifies the name of the DB2 instance. For example: -p
INSTANCE="DB2inst1". This property applies to both Windows and UNIX
environments.
_UNIX_STARTUP_.Username
Specifies the username. This property applies to UNIX environments.

The following is an example of this command for Windows environments:

tacmd addSystem -t UD -n Primary:hostname:NT -p INSTANCE="DB2"

The following is an example of this command for UNIX environments:

tacmd addSystem -t ud -n hostname:KUX -p INSTANCE="db2inst1"
_UNIX_STARTUP_.Username="db2inst1"

Starting and stopping the Monitoring Agent for DB2 on UNIX operating
systems
This section provides additional information relevant to the starting or stopping of
this monitoring agent for a given database instance.

Background information
The Monitoring Agent for DB2 does not require advanced configuration. However,
you must start this monitoring agent while logged on as the DB2 instance owner,
and you must have DB2SysAdmin, SysCtrl, and SysMaint authorities.

You can run multiple copies of this monitoring agent by specifying different
database instance names. Only one process can be started for each database
instance.

Required authorization role

No special authorization role is required.

Procedure
You can start and stop this monitoring agent using the Manage Tivoli Enterprise
Monitoring Services utility or using the itmcmd agent command.

Starting and stopping DB2 agents using Manage Tivoli Enterprise Monitoring
Services window
1. When starting the DB2 agent using Manage Tivoli Enterprise Monitoring
Services, you are prompted for a database instance name.
2. When stopping the DB2 agent using Manage Tivoli Enterprise Monitoring
Services, you are prompted for which database instance name to stop.

Note: Enter only ASCII characters in the fields for this window.

Starting and stopping DB2 agents using the itmcmd agent command

Chapter 2. Requirements and configuration for the monitoring agent 5

 When using the itmcmd agent command to start or stop this monitoring agent,
include the following command option:
-o Specifies the database instance to start or stop. The database instance name
must match the name used for starting the database.

For example:
itmcmd agent -o DB2inst1 start | stop ud
itmcmd agent -o Preface_DB2inst2 start | stop ud
itmcmd agent -o DB2inst3_suffix start | stop ud

If you start the monitoring agent without specifying the -o option, you receive the
following error message: This agent requires the -o option...

The host name of the system is automatically added to the instance name in the
managed system list of the user interfaces.

For more information about using the itmcmd agent command, see the IBM Tivoli
Monitoring Installation and Setup Guide.

Creating an ODBC data source

Some of the database attributes available with this monitoring agent require the
use of an Open Database Connectivity (ODBC) data source to maintain a
connection with the database. If you intend to use any of the following attributes,
you must create an ODBC data source first; otherwise, you will be unable to use
them. Descriptions of these attributes are available in “Database Information
attributes (group 01)” on page 92.

db cap err
db cap lag
db cap prun
db tablespaces
event monitors
invalid pkgs
invalid sys pkgs
invalid triggers
system tablespaces
tables
tablespaces
tablespaces long data
triggers
user indexes
views

The ODBC data source must be registered on the system where the monitoring
agent runs, where DB2 resides. If you do not intend to use these attributes, you do
not need to register an ODBC data source.

To register an ODBC data source for use with this monitoring agent, follow these
steps (or refer to your database documentation for current instructions):
1. In DB2, open the Add Data Source window.
2. Specify how the data source should be registered. Register the data source as a
system data source so that all users on the system will have access to the
database.
3. Specify the alias for the database in the Database alias field.

6 IBM Tivoli Monitoring for Databases: DB2 Agent: User’s Guide

4. (Optional) In the Data source name field, you can specify a more meaningful
name for the data source. The name you enter is mapped to the Database alias.

Note: The data source name must be the same name as the database.
5. For Windows, select Register this database for ODBC if the database.
6. Select Register this database for CLI if the database will be accessed by DB2
CLI applications. In UNIX, the data source is registered to CLI by default.
7. For Windows, specify one of the common applications in the Optimize for
application field. This will optimize the CLI settings for that type of
application. Existing settings will be overwritten where necessary. You should
optimize the settings for an application first, then use the Settings page to
modify specific configuration keywords as required. If you have already set
specific settings then you should back up your db2cli.ini file before using the
optimize feature. You will then have to go and modify any settings that you
specifically want different from the optimized values.
8. Click OK to register the data source and close the window.

Chapter 2. Requirements and configuration for the monitoring agent 7

8 IBM Tivoli Monitoring for Databases: DB2 Agent: User’s Guide
Chapter 3. How to use the Monitoring Agent for DB2
After you have installed and configured the Monitoring Agent for DB2, you can
begin using this monitoring agent to monitor your resources.

This chapter provides information about how to use the Monitoring Agent for DB2
to perform the following tasks:
v “View real-time data about DB2 Universal Database”
v “Investigate an event” on page 10
v “Recover the operation of a resource” on page 10
v “Customize your monitoring environment” on page 11
v “Monitor with custom situations that meet your requirements” on page 12
v “Collect and view historical data” on page 13

For each of these tasks, there is a list of procedures that you perform to complete
the task. For the procedures, there is a cross-reference to where you can find
information about performing that procedure. Information about the procedures is
located in subsequent chapters and appendixes of this user’s guide and in the IBM
Tivoli Monitoring documentation.

View real-time data about DB2 Universal Database

After you install, configure, and start the Monitoring Agent for DB2, the
monitoring agent begins monitoring.

Table 2 contains a list of the procedures for viewing the real-time data about DB2
Universal Database that the monitoring agent collects. The table also contains a
cross-reference to where you can find information about each procedure.
Table 2. Viewing real-time data about DB2 Universal Database
Procedure Where to find information
View the hierarchy of your monitored IBM Tivoli Monitoring User’s Guide:
resources from a system point of view ″Navigating through workspaces″ (in
(Navigator view organized by operating ″Monitoring: real-time and event-based″
platform, system type, monitoring agents, chapter)
and attribute groups).
View the indicators of real or potential
problems with the monitored resources
(Navigator view).
View changes in the status of the resources IBM Tivoli Monitoring User’s Guide: ″Using
that are being monitored (Enterprise workspaces″ (in ″Monitoring: real-time and
Message Log view). event-based″ chapter)
View the status of the agents in the Chapter 4, “Workspaces reference,” on page
managed enterprise that you are monitoring 15 in this guide
(Monitoring Agent Status view).

© Copyright IBM Corp. 2005 9

 Table 2. Viewing real-time data about DB2 Universal Database (continued)
Procedure Where to find information
View the number of times an event has been IBM Tivoli Monitoring User’s Guide: ″Using
opened for a situation during the past 24 workspaces″ (in ″Monitoring: real-time and
hours (Open Situations Count view). event-based″ chapter)

Chapter 4, “Workspaces reference,” on page

15 in this guide

Chapter 6, “Situations reference,” on page

131 in this guide
Manipulate the views in a workspace. IBM Tivoli Monitoring User’s Guide: ″Using
views″ (in ″Monitoring: real-time and
event-based″ chapter)

Investigate an event
When the conditions of a situation have been met, an event indicator is displayed
in the Navigator. When an event occurs, you want to obtain information about that
event so you can correct the conditions and keep your enterprise running
smoothly. The situation must be associated with a Navigator Item in order to
appear.

Table 3 contains a list of the procedures for investigating an event and a

cross-reference to where you can find information about each procedure.
Table 3. Investigating an event
Procedure Where to find information
Determine which situation raised the event IBM Tivoli Monitoring User’s Guide: ″Opening
and identify the attributes that have values the situation event workspace″ (in
that are contributing to the alert. ″Monitoring: real-time and event-based″
chapter, ″Responding to alerts″ section)

Chapter 4, “Workspaces reference,” on page

Review available advice. 15 in this guide

Appendix A, “IBM Tivoli Monitoring for

DB2 Workspaces,” on page 155 in this guide
Notify other users that you have taken IBM Tivoli Monitoring User’s Guide:
ownership of the problem related to an ″Acknowledging a aituation event″ (in
event and are working on it. ″Monitoring: real-time and event-based″
chapter, ″Responding to alerts″ section)
Remove the event from the Navigator. IBM Tivoli Monitoring User’s Guide: ″Closing
the situation event workspace″ (in
″Monitoring: real-time and event-based″
chapter, ″Responding to alerts″ section)

Recover the operation of a resource

When you find out that a resource is not operating as desired, you can control it
manually or automatically using Take Action commands.

Table 4 on page 11 contains a list of the procedures for recovering the operation of
a resource and a cross-reference to where you can find information about each
procedure.

10 IBM Tivoli Monitoring for Databases: DB2 Agent: User’s Guide

 Table 4. Recovering the operation of a resource
Procedure Where to find information
Take an action on a resource manually. IBM Tivoli Monitoring User’s Guide:
v ″Other views″ (in ″Custom workspaces″
chapter, ″Workspace views″ section)
v ″Take action – Reflex automation″ (in
Situations for event-based monitoring″
chapter, ″Event-based monitoring
overview″ section)

Chapter 7, “Take Action commands

reference,” on page 143 in this guide
Take an action on a system condition IBM Tivoli Monitoring User’s Guide:
automatically by setting up a situation to ″Situations for event-based monitoring″
run a Take Action command. chapter
v ″Customizing a situation″
v ″Creating a situation″
v ″Specify an action to take″
v ″Distribute the situation″

Chapter 7, “Take Action commands

reference,” on page 143 in this guide
Take multiple actions on system conditions IBM Tivoli Monitoring User’s Guide: ″Policies
automatically using a policy. for automation″ chapter
v ″Creating a policy″
v ″Maintaining policies″
Take actions across systems, monitoring v ″Workflows window″
agents, or computers using a policy.
Chapter 8, “Policies reference,” on page 153
in this guide

Customize your monitoring environment

You can change how your monitoring environment looks by creating new
workspaces with one or more views in it.

Table 5 contains a list of the procedures for customizing your monitoring

environment and a cross-reference to where you can find information about each
procedure.
Table 5. Customizing your monitoring environment
Procedure Where to find information
Display data in tables or charts (views) in IBM Tivoli Monitoring User’s Guide:
Tivoli Enterprise Portal. v ″Custom workspaces″
v ″Table and chart views″
Display an overview of changes in the status IBM Tivoli Monitoring User’s Guide: ″Message
of the situations for your monitored log view″ (in ″Situation event views:
resources (Message Log View). message log, situation event console,
graphic, and Tivoli Enterprise Console″
chapter)

Chapter 3. How to use the Monitoring Agent for DB2 11

 Table 5. Customizing your monitoring environment (continued)
Procedure Where to find information
Specify which attributes to retrieve for a IBM Tivoli Monitoring User’s Guide: ″Creating
table or chart so you can retrieve only the custom queries″ (in ″Table and chart views″
data you want by creating custom queries. chapter)

Chapter 5, “Attributes reference,” on page 27

in this guide
Build links from one workspace to another. IBM Tivoli Monitoring User’s Guide:
v ″Link from a workspace″ (in ″Custom
workspaces″ chapter)
v ″Link from a table or chart″ (in ″Table and
chart views″ chapter)
Identify which predefined situations started IBM Tivoli Monitoring User’s Guide: ″What
running automatically when you started the the enterprise workspace shows″ (in
Tivoli Enterprise Monitoring Server. ″Monitoring: real-time and event-based″
chapter, ″Using workspaces″ section)

Chapter 6, “Situations reference,” on page

131 in this guide
Determine whether to run situations as Chapter 6, “Situations reference,” on page
defined, modify the values in situations, or 131 in this guide
create new situations to detect possible
problems.

Monitor with custom situations that meet your requirements

When your environment requires situations with values that are different from
those in existing situations, or when you need to monitor conditions not defined
by the existing situations, you can create custom situations to detect problems with
resources in two ways:
v Create an entirely new situation
v Create a situation by copying and editing a predefined situation

Note: When you create and run a situation, an IBM Tivoli Enterprise Console
event is created. For information on how to define event severities from
forwarded IBM Tivoli Monitoring situations and other event information,
see the IBM Tivoli Monitoring Administrator’s Guide.

You can specify the following information for a situation:

v Name
v Attribute group and attributes
v Qualification to evaluate multiple rows when a situation has a multiple-row
attribute group (display item)
v Formula
v Take Action commands
v Run at startup
v Sampling interval
v Persistence
v Severity
v Clearing conditions

12 IBM Tivoli Monitoring for Databases: DB2 Agent: User’s Guide

 v Expert Advice
v When a true situation closes
v Available Managed Systems

Table 6 contains a list of the procedures for monitoring your resources with custom
situations that meet your requirements and a cross-reference to where you can find
information about each procedure.
Table 6. Monitoring with custom situations
Procedure Where to find information
Create an entirely new situation. IBM Tivoli Monitoring User’s Guide: ″Creating
a new situation″ (in ″Situations for
event-based monitoring″ chapter, ″Creating a
situation″ section)

Chapter 5, “Attributes reference,” on page 27

in this guide
Create a situation by copying and editing a IBM Tivoli Monitoring User’s Guide:
predefined situation. ″Customizing a situation″ (in ″Situations for
event-based monitoring″ chapter)

Chapter 6, “Situations reference,” on page

131 in this guide

Chapter 5, “Attributes reference,” on page 27

in this guide
Run a situation on a managed system. IBM Tivoli Monitoring User’s Guide:
″Situations for event-based monitoring″
chapter
v ″Associate situations with navigator
items″
v ″Distribute the situation″ (in ″Customizing
a situation″ section)
v ″Start, stop, or delete a situation″

Collect and view historical data

When you collect historical data, you specify the following configuration
requirements:
v Attribute groups for which to collect data
v Collection interval
v Summarization and pruning of attribute groups
v Roll-off interval to a data warehouse, if any
v Where to store the collected data (at the agent or the Tivoli Enterprise
Monitoring Server)

Table 7 on page 14 contains a list of the procedures for collecting and viewing
historical data and a cross-reference to where you can find information about each
procedure.

Chapter 3. How to use the Monitoring Agent for DB2 13

 Table 7. Collecting and viewing historical data
Procedure Where to find information
Configure and start collecting short-term IBM Tivoli Monitoring User’s Guide:
data (24 hours). ″Historical reporting″ (in ″Table and chart
views″ chapter)
Configure and start collecting longer-term
data (more than 24 hours). IBM Tivoli Monitoring Administrator’s Guide
View historical data in the Tivoli Enterprise
Portal. “Disk capacity planning for historical data”
on page 128
Create reports from historical data using
third-party reporting tools.
Filter out unwanted data to see specific
areas of interest.

14 IBM Tivoli Monitoring for Databases: DB2 Agent: User’s Guide

Chapter 4. Workspaces reference
This chapter contains an overview of workspaces, references for detailed
information about workspaces, and descriptions of the predefined workspaces
included in this monitoring agent.

About workspaces
A workspace is the working area of the Tivoli Enterprise Portal application
window. At the left of the workspace is a Navigator that you use to select the
workspace you want to see.

As you select items in the Navigator, the workspace presents views pertinent to
your selection. Each workspace has at least one view. Some views have links to
workspaces. Every workspace has a set of properties associated with it.

This monitoring agent provides predefined workspaces. You cannot modify or

delete the predefined workspaces, but you can create new workspaces by editing
them and saving the changes with a different name.

More information about workspaces

For more information about creating, customizing, and working with workspaces,
see the document IBM Tivoli Monitoring User’s Guide.

For a list of the predefined workspaces for this monitoring agent and a description
of each workspace, refer to the Predefined workspaces section below and the
information in that section for each individual workspace.

For additional information about workspaces for this monitoring agent, see
Appendix A, “IBM Tivoli Monitoring for DB2 Workspaces,” on page 155.

Predefined workspaces
The Monitoring Agent for DB2 provides the following predefined workspaces,
which are organized by Navigator item:
v Application
– Application Identification
– Application I/O Activity
– Application Lock Activity
– Application Overview
– Application Package and Catalog Cache Activity
– Application Sort and Hash Join Activity
– Application SQL Activity
– Application SQL Statement Text
– Application Summary
– Application Time Information
– Application Top Ten Summary
v Buffer Pool

© Copyright IBM Corp. 2005 15

 – Buffer Pool
– Buffer Pool Detail
v Database
– Database
– Database Bottom Ten Summary by BP Hit Ratio
– Database Identification/Status/Logging Information
– Database I/O Activity
– Database Lock Activity
– Database Pkg/Cat Cache Activity
– Database Sort/Hash Join Activity
– Database SQL Activity
– Database Summary by BP Hit Ratio
v Locking Conflict
– Locking Conflict
v System Overview
– System Overview
– Connection
– General Information
v Table Space
– Table Space

The remaining sections of this chapter contain descriptions of each of these

predefined workspaces. The workspaces are organized by the Navigator item to
which the workspaces are relevant.

Application
The workspaces related to the Application Navigator item are described in this
section.

Application Identification workspace

The Application Identification workspace provides information on status and space
usage for the database. You can view such topics as:
v Application status and identifiers
v Buffer pool/catalog cache hit ratio
v Client information
v SQL activity
v Sort/hash join overflows

The data for this workspace comes from the Application attribute group.

Based on the information that this workspace provides, you can make changes, set
up situations, and verify that changes are improving performance.

Use the Application Identification workspace to monitor your application status

and space usage.

Application I/O Activity workspace

The Application I/O Activity workspace provides information on various I/O
activities. You can view such topics as:

16 IBM Tivoli Monitoring for Databases: DB2 Agent: User’s Guide

v Buffer pool activity
v Extended store activity
v Direct I/O activity
v Application identification and status

The data for this workspace comes from the Application attribute group.

Based on the information that this workspace provides, you can make changes, set
up situations, and verify that changes are improving performance.

Use the Application I/O Activity workspace to monitor the I/O processing
associated with your applications.

Application Lock Activity workspace

The Application Lock Activity workspace provides information on various lock
activities. You can view such topics as:
v Lock waits
v Locks held
v Lock escalations
v Lock timeouts
v Average lock wait time

The data for this workspace comes from the Application attribute group.

Based on the information that this workspace provides, you can make changes, set
up situations, and verify that changes are improving performance.

Application Overview workspace

The Application Overview workspace provides overview information on the top
ten applications experiencing conditions that you might want to investigate further,
and is the default workspace for the Application Navigator item. You can view
such topics as:
v The 10 applications with the highest percentage of failed SQL statements
v The 10 applications with the highest number of sort overflows
v The 10 applications with the lowest buffer pool hit ratio
v The 10 applications with the highest number of lock timeouts
v The 10 applications with the highest number of deadlocks

The data for this workspace comes from the Application attribute group. This
workspace is similar to the Application Top Ten Summary workspace, although
graphs are used instead of tables to present data.

Based on the information that this workspace provides, you can make changes, set
up situations, and verify that changes are improving performance.

Use the Application Overview workspace to monitor the top 10 applications in

terms of failed SQL statements, sort overflow, lock timeout, and deadlocks, and
bottom 10 applications in terms of buffer pool hit ratio.

Application Package and Catalog Cache Activity workspace

The Application Package and Catalog Cache Activity workspace provides
information on various package and catalog cache activities. You can view such
topics as:

Chapter 4. Workspaces reference 17

 v Package/catalog cache hit ratio
v Catalog cache overflows
v Application package/catalog cache activities such as package cache lookups,
package cache inserts, catalog cache lookups, catalog cache inserts, catalog cache
heap full

The data for this workspace comes from the Application attribute group.

Based on the information that this workspace provides, you can make changes, set
up situations, and verify that changes are improving performance.

Use the Application Package and Catalog Cache Activity workspace to monitor
package and catalog cache activities.

Application Sort and Hash Join Activity workspace

The Application Sort and Hash Join Activity workspace provides information on
sort/hash join activities. You can view such topics as:
v Sort overflows
v Hash join overflows
v Total sorts, total sort time, average sort time, sort overflows percent
v Total hash joins, total hash loops, hash join overflows

The data for this workspace comes from the Application attribute group.

Based on the information that this workspace provides, you can make changes, set
up situations, and verify that changes are improving performance.

Use the Application Sort/Hash Join Activity workspace to monitor sort/hash join
activities.

Application SQL Activity workspace

The Application SQL Activity workspace provides information about SQL
statements. You can view such topics as:
v SQL statement counts
v Number of statements and cursors

The data for this workspace comes from the Application attribute group.

Based on the information that this workspace provides, you can make changes, set
up situations, and verify that changes are improving performance.

Use the Application SQL Activity workspace to review information about SQL
statements.

Application SQL Statement Text workspace

The Application SQL Statement Text workspace provides the text of SQL
statements. You can view such topics as:
v Text of an SQL statement
v SQL activity such as total SQL statements, failed SQL statements, internal
rollbacks, internal deadlock rollbacks, and rollback SQL statements
v Application identification and status data such as agent ID, application status,
application name, snapshot time, database name and so on
v Query cost estimate

18 IBM Tivoli Monitoring for Databases: DB2 Agent: User’s Guide

v Query card estimate
v Degree of parallelism requested

The data for this workspace comes from the Application attribute group.

Based on the information that this workspace provides, you can make changes, set
up situations, and verify that changes are improving performance.

Use the Application SQL Statement Text workspace to understand what type of
SQL statements are being executed.

Application Summary workspace

The Application Summary workspace provides information on various application
activities. You can view such topics as:
v The 10 applications with the highest number of failed SQL statements
v The 10 applications with the lowest buffer pool hit ratios
v A summary (by agent ID) of application data such as application status,
snapshot time, application name, locks held, lock waits, deadlocks, pool hit ratio,
statement operation

The data for this workspace comes from the Application attribute group.

Based on the information that this workspace provides, you can make changes, set
up situations, and verify that changes are improving performance.

Use the Application Summary workspace to identify application SQL failures,

applications with low buffer pool hit ratios and other information about the
applications.

Application Time Information workspace

The Application Time workspace provides information on various time-related
activities. You can view such topics as:
v Lock wait time
v Buffer pool and direct I/O time
v Application identification and status
v Application time such as application idle time, amount of CPU time spent in
system calls, amount of CPU time spent in executing database manager code,
application connect time

The data for this workspace comes from the Application attribute group.

Based on the information that this workspace provides, you can make changes, set
up situations, and verify that changes are improving performance.

Use the Application Time workspace to monitor time-related information about

your applications.

Application Top Ten Summary workspace

The Application Top Ten Summary workspace provides summary information on
various application-related activities. You can view such topics as:
v The Application Summary table for the top 10 applications
v The 10 applications with the highest percentage of failed SQL statements
v The 10 applications with the lowest buffer pool hit ratio

Chapter 4. Workspaces reference 19

 The data for this workspace comes from the Application attribute group.

Based on the information that this workspace provides, you can make changes, set
up situations, and verify that changes are improving performance.

Use the Application Top Summary workspace to monitor the top 10 applications in
terms of failed SQL statements and bottom 10 applications in terms of buffer pool
hit ratio.

Buffer Pool workspace

The workspaces related to the Buffer Pool Navigator item are described in this
section.

Buffer Pool workspace

The Buffer Pool workspace provides information on various buffer pool activities.
You can view such topics as:
v Buffer pool hit ratio by individual buffer pool
v The ten databases with the lowest buffer pool hit ratios
v Summary data such as buffer pool name, database name, pool total reads,
average pool read time, and average pool write time

The data for this workspace comes from the Buffer Pool attribute group.

Based on the information that this workspace provides, you can make changes, set
up situations, and verify that changes are improving performance.

Use the Buffer Pool workspace and Buffer Pool Detail workspace to monitor your
DB2 buffer pools.

Buffer Pool Detail workspace

The Buffer Pool Detail workspace provides information on various buffer pool
activities. You can view such topics as:
v Databases with the lowest buffer pool hit ratios
v Average read and write times
v Buffer pool identification data
v Various buffer pool activities such as data logical reads, data physical reads,
index logical reads, index physical reads, pool total reads
v Asynchronous and synchronous I/O activity attributes such as pool
asynchronous data reads, pool asynchronous data writes, pool sync data reads,
pool sync data writes
v Extended store and non-buffer-pool I/O activity such as pool data to estore,
pool index to estore, direct reads, direct writes

The data for this workspace comes from the Buffer Pool attribute group.

Based on the information that this workspace provides, you can make changes, set
up situations, and verify that changes are improving performance.

Use the Buffer Pool Detail workspace to monitor your DB2 buffer pools.

Database
The workspaces related to the Database Navigator item are described in this
section.

20 IBM Tivoli Monitoring for Databases: DB2 Agent: User’s Guide

Database workspace
The Database workspace provides information on status and space usage from a
database perspective. You can view such topics as:
v The 10 databases with the lowest buffer pool hit ratio
v The 10 databases with the highest number of connections
v The 10 databases with the highest number of failed SQL statements
v The 10 databases with the highest number of lock timeouts
v The 10 databases with the highest number of deadlock situations

Because of screen width limitations, you need to scroll to the right to view all of
the columns in a table view.

The data for this workspace comes from the Database attribute group.

Based on the information that this workspace provides, you can make changes, set
up situations, and verify that changes are improving performance.

Use the Database workspace to monitor the databases in the DB2 instances in your
environment.

Database Bottom Ten Summary by BP Hit Ratio workspace

The Database Bottom Ten Summary by BP Hit Ratio workspace provides
information on the bottom 10 databases in terms of buffer pool hit ratio.You can
view such topics as:
v The 10 databases with the lowest buffer pool hit ratio
v The 10 databases with the most connections
v The 10 databases with the highest percentage of failed SQL statements
v The 10 databases with the lowest buffer pool hit ratio (summary)

The data for this workspace comes from the Database attribute group.

Based on the information that this workspace provides, you can make changes, set
up situations, and verify that changes are improving performance.

Use this workspace to monitor the applications with the lowest buffer pool hit
ratio.

Database Identification / Status / Logging Information workspace

The Database Identification / Status / Logging Information workspace provides
information on various database identification, status, and logging activities. You
can view such topics as:
v The 10 databases with the lowest buffer pool hit ratio
v The 10 databases with the highest number of connections
v The 10 databases with the highest percentage of failed SQL statements
v The 10 databases with the highest number of lock timeouts
v The 10 databases with the highest number of deadlocks

The data for this workspace comes from the Database attribute group.

Based on the information that this workspace provides, you can make changes, set
up situations, and verify that changes are improving performance.

Chapter 4. Workspaces reference 21

 Use the Database Identification / Status / Logging Information workspace to
monitor the 10 databases with the highest connections, failed SQL statements, lock
timeouts, deadlocks, and the 10 databases with the lowest buffer pool hit ratio.

Database I/O Activity workspace

The Database I/O Activity workspace provides information on various
database-related I/O activities. You can view such topics as:
v Buffer pool hit ratio for the monitored database
v The number of buffer pool reads and writes to the monitored database
v Counts of various buffer pool activities such as pool read time, pool data logical
reads, pool index physical reads, files closed
v Direct I/O activities such as number of: direct reads, direct writes, direct read
time, direct write time

The data for this workspace comes from the Database attribute group.

Based on the information that this workspace provides, you can make changes, set
up situations, and verify that changes are improving performance.

Use the Database I/O Activity workspace to monitor the input/output operations
of your database.

Database Lock Activity workspace

The Database Lock Activity workspace provides information on various locking
activities from a database perspective. You can view such topics as
v Lock timeouts and deadlocks
v SQL activity such as the number of: internal rollbacks, internal deadlock
rollbacks, failed SQL statements
v Database identification and status data such as database path, database name,
database connect time, last backup
v Number of: locks held, deadlocks, lock escalations, lock timeouts; average lock
time; lock wait time

The data for this workspace comes from the Database attribute group.

Based on the information that this workspace provides, you can make changes, set
up situations, and verify that changes are improving performance.

Use the Database Lock Activity workspace to monitor database locking activities.

Database Pkg/Cat Cache workspace

The Database Package/Catalog Cache workspace provides information on various
package and catalog cache activities. You can view such topics as:
v Package and catalog cache hit ratio
v Catalog cache overflows and catalog cache heap full
v Package cache activity such as package cache lookups
v Catalog cache activity such as catalog cache lookups

The data for this workspace comes from the Database attribute group.

Based on the information that this workspace provides, you can make changes, set
up situations, and verify that changes are improving performance.

22 IBM Tivoli Monitoring for Databases: DB2 Agent: User’s Guide

Use the Database Pkg/Cat Cache workspace to monitor cache activities of your
catalogs and packages.

Database Sort/Hash Join Activity workspace

The Database Sort/Hash Join Activity workspace provides information on various
sort and hash join activities. You can view such topics as:
v Total number of sorts and sort overflows
v Total number of hash joins and hash join overflows
v Sort heap allocated, number of active sorts, average sort time
v Total number of hash loops and hash join small overflows

The data for this workspace comes from the Database attribute group.

Based on the information that this workspace provides, you can make changes, set
up situations, and verify that changes are improving performance.

Use the Database Sort/Hash Join Activity workspace to monitor sort and hash join
activities.

Database SQL Activity workspace

The Database SQL Activity workspace is a predefined workspace that provides
information on various SQL activities from a database perspective. You can view
such topics as:
v SQL activity
v Database identification and status, including:
– Database connection information
– Database logging activity
– Database highest agent information
– Database identification and status
v Row counts
v SQL statement data

The data for this workspace comes from the Database attribute group.

Based on the information that this workspace provides, you can make changes, set
up situations, and verify that changes are improving performance.

Use the Database SQL Activity workspace to monitor SQL activity.

Database Summary by BP Hit Ratio workspace

The Database Summary by BP Hit Ratio workspace provides information on
various database-related activities. You can view such topics as:
v The 10 databases with the lowest buffer pool hit ratio
v The 10 databases with the highest number of connections
v The 10 databases with the highest number of failed SQL statements
v The 10 databases with the lowest buffer pool hit ratio

The data for this workspace comes from the Database attribute group.

Based on the information that this workspace provides, you can make changes, set
up situations, and verify that changes are improving performance.

Chapter 4. Workspaces reference 23

 Locking Conflict workspace
The Locking Conflict workspace provides information on various locking activities
and resources. You can view such topics as:
v The 10 applications with the highest number of lock timeouts
v The lock wait time for each application
v Specific information about lock conflicts such as application identification,
application name, snapshot time, client database alias, locks held, lock wait time

The data for this workspace comes from the Lock Conflict attribute group.

Based on the information that this workspace provides, you can make changes, set
up situations, and verify that changes are improving performance.

Use the Locking Conflict workspace to monitor lock timeouts, wait times, and lock
conflicts.

System Overview
The workspaces related to the System Overview Navigator item are described in
this section.

Connection workspace
The Connection workspace provides information on various server-related
activities. You can view such topics as:
v Server connection data such as remote connections initiated, local connections,
number of local databases with current connections
v Agent information such as agents waiting on token, agents stolen, idle agents

The data for this workspace comes from the Server attribute group.

Based on the information that this workspace provides, you can make changes, set
up situations, and verify that changes are improving performance.

Use the Connection workspace to view information about your database server.

General Information workspace

The General Information workspace provides information on various server
activities. You can view such topics as:
v Server key events such as post threshold sorts, agents waiting on a token, agents
stolen, the number of attempts to exceed the maximum number of agents
v Server connection data such as remote connections initiated, local connections,
number of local databases with current connections
v Data about the status of a DB2 instance such as start time, last reset, snapshot
time, instance name
v Data about sort/hash join such as sort heap allocated, piped sorts requested,
piped sorts accepted

The data for this workspace comes from the Server attribute group.

Based on the information that this workspace provides, you can make changes, set
up situations, and verify that changes are improving performance.

Use the General Information workspace to get information on database server

activities.

24 IBM Tivoli Monitoring for Databases: DB2 Agent: User’s Guide

 System Overview workspace
The System Overview workspace provides information about your DB2 system as
a whole. You can view such topics as:
v Key server events
v Server connections
v The 10 databases with the lowest buffer pool hit ratio
v The 10 applications with the highest percentage of failed SQL statements
v Server status

Because of screen width limitations, you need to scroll to the right to view all of
the columns in a table view.

The data for this workspace comes from the Server attribute group.

Based on the information that this workspace provides, you can make changes, set
up situations, and verify that changes are improving performance.

Use the System Overview workspace to monitor your system activity.

Table Space workspace

The Table Space workspace provides tablespace information at the database level.
You can use this information to monitor page size and usage characteristics.

Paging activity of logical groups of files (tablespaces) is shown in the Paging

Statistics bar chart. This chart provides at-a-glance page counts for free pages,
pending free pages, total pages, usable pages, and used pages for the tablespaces
identified in the Table Space Summary table view.

The data for this workspace comes from the Table Space attribute group.

Based on the information that this workspace provides, you can make changes, set
up situations, and verify that changes are improving performance.

Use the Table Space workspace to monitor your DB2 tablespaces.

Chapter 4. Workspaces reference 25

26 IBM Tivoli Monitoring for Databases: DB2 Agent: User’s Guide
Chapter 5. Attributes reference
This chapter contains information about the following topics:
v Overview of attributes
v References for detailed information about attributes
v Descriptions of the attributes for each attribute group included in this
monitoring agent
v Disk space requirements for historical data

About attributes
Attributes are the application properties being measured and reported by the
Monitoring Agent for DB2, such as the amount of memory usage or the message
ID. Some monitoring agents have fewer than 100 attributes, while others have over
1000.

Attributes are organized into groups according to their purpose. The attributes in a
group can be used in the following two ways:
v Chart or table views
Attributes are displayed in chart and table views. The chart and table views use
queries to specify which attribute values to request from a monitoring agent.
You use the Query editor to create a new query, modify an existing query, or
apply filters and set styles to define the content and appearance of a view based
on an existing query.
v Situations
You use attributes to create situations that monitor the state of your operating
system, database, or application. A situation describes a condition you want to
test. When you start a situation, the Tivoli Enterprise Portal compares the values
you have assigned to the situation attributes with the values collected by the
Monitoring Agent for DB2 and registers an event if the condition is met. You are
alerted to events by indicator icons that appear in the Navigator.

Some of the attributes in this chapter are listed twice, with the second attribute
having a ″(Unicode)″ designation after the attribute name. These Unicode attributes
were created to provide access to globalized data. Use the globalized attribute
names because this is where the monitoring agent is putting the data. If you were
using a previous Candle® OMEGAMON® release of this monitoring agent, you
must run the Application Migration Tool to create globalized attributes for your
customized queries, situations, and policies. See the IBM Tivoli Monitoring
Installation and Setup Guide for more information.

More information about attributes

For more information about using attributes and attribute groups, see the
document IBM Tivoli Monitoring User’s Guide.

For a list of the attributes groups, a list of the attributes in each attribute group,
and descriptions of the attributes for this monitoring agent, refer to the Attribute
groups and attributes section in this chapter.

© Copyright IBM Corp. 2005 27

Attribute groups and attributes for the Monitoring Agent for DB2
This monitoring agent contains the following attribute groups:
v Application (group 00, deprecated application attributes)
v Application (group 00, Unicode application attributes)
v Application (group 01)
v Buffer Pool Data
v Database (group 00)
v Database (group 01)
v Locking Conflict
v System Overview
v Tablespace

Because attribute groups have limits on the number of attributes that it they can
hold, there are multiple application groups for some type of data, such as for
application and database information.

The following sections contain descriptions of these attribute groups, which are
listed alphabetically. Each description contains a list of attributes in the attribute
group.

Application attributes (group 00)

The Application attribute group provides information about application activities.
You can use this information to monitor the performance of your applications and
to identify problem areas for corrective action. All values are integers calculated
from the first application connection, unless otherwise noted (integer values are
32-bit or 0 to 4 294 967 296).

Note: The attributes listed in this attribute group are deprecated. See also
“Application attributes (Unicode group 00)” on page 46 for Unicode
versions of the attributes listed here.

acc curs blk The number of times that a request for an I/O block was accepted.
The value format is integer. Use this attribute with the Rejected Block Cursor
Requests attribute to calculate the percentage of blocking requests that are accepted
and/or rejected.

agent id The application handle, which is a system-wide unique ID for the

application. The value format is integer. On multi-node systems, where a database
is partitioned, this ID is the same on every node where the application can make a
secondary connection.

Use the application handle to uniquely identify an active application (application

handle is synonymous with agent ID).

agent id holding lk The application handle of the agent holding a lock for which
this application is waiting. The value format is integer. The lock monitor group
must be turned on to obtain this information.

This attribute can help you to determine which applications are in contention for
resources.

28 IBM Tivoli Monitoring for Databases: DB2 Agent: User’s Guide

agent sys cpu time The total system CPU time (in seconds and microseconds) used
by the database manager agent process.

This counter includes CPU time for both SQL and non-SQL statements, and any
fenced user-defined functions (UDFs). System CPU represents the time spent in
system calls.

User CPU represents time spent executing database manager code. The value
format is a text string with a maximum length of 12 characters. Use this attribute,
along with the other related CPU-time attributes, to help you understand the level
of activity within an application. It can help to identify applications that could
benefit from additional tuning.

agent usr cpu time The total CPU time (in seconds and microseconds) used by the
database manager agent process.

This counter includes time spent on both SQL and non-SQL statements, and any
fenced user-defined functions (UDFs) or stored procedures executed by the
application. System CPU represents the time spent in system calls.

User CPU represents time spent executing database manager code. The value
format is a text string with a maximum of 12 characters. Use this attribute, along
with the other CPU-time related attributes, to help you identify applications or
queries that consume large amounts of CPU time.

appl con time The date and time that an application started a connection request.

Use this attribute to determine when the application started its connection request
to the database.

appl id The identifier generated when the application connects to the database at
the database manager or when Data Definition Control Support (DDCS) receives a
request to connect to a Distributed Relational Database Architecture™ (DRDA®)
database. The value format is a text string, with a maximum of 32 characters. The
identifier is unique across the network. The application ID appears in various
formats, which depend on the communication protocol between the client and the
server on which the database manager and/or DDCS are running. Each of the
formats consists of three parts separated by periods. Use this ID (known on both
the client and server) to correlate the client and server parts of the application.

appl id holding lk The application ID of the application that is holding a lock on

the object that this application is waiting to obtain. The value format is a text
string with a maximum of 32 characters.

This attribute can help you to determine which applications are in contention for
resources. Specifically, it can help you to identify the application handle (agent ID)
and table ID that are holding the lock.

appl idle time The number of seconds since an application issued a request to the
server. The value format is integer.

This includes applications that have not terminated a transaction (for example,
have not issued a commit or rollback). Use this information to force users that
have been idle for a specified number of seconds.

Chapter 5. Attributes reference 29

 appl name The name of the application running at the client as known to the
database manager or DB2 Connect™. The value format is a text string, with a
maximum of 20 characters. Example: *Local.db2inst1.990212202018.

Use this attribute with the Application ID to relate data items with your
application. In a client/server environment, this name is passed from the client to
the server to establish the database connection. For DRDA-AS connections, this
name is the DRDA external name.

appl status The status of the application being monitored. This attribute can help
you diagnose potential application problems. The value format is a text string with
a maximum of 64 characters. Application status values are as shown:
Backing Up Database
Commit Active
Compiling SQL Stmt
Connect Pending
Connected
Creating Database
Disconnect Pending
I/O Error Waiting
Loading Database
Lock Waiting
Prepared Transaction
Quiescing a Tablespace
Recompiling Plan
Request Interrupted
Restarting Database
Restoring Database
Rollback Active
Trans. heuristically aborted
Trans heuristically committed
Transaction ended
UOW Executing
UOW Waiting in the application
Unloading Database

auth id The authorization ID of the user who invoked the application that is being
monitored. On a DB2 Connect gateway node, this is the user’s authorization ID on
the host. The value format is a text string with a maximum of 20 characters. Use
this attribute to determine who invoked the application.

avg lock waittime The average elapsed time spent waiting for a lock. The value
format is integer. If the average lock wait time is high, look for applications that
hold many locks, or have lock escalations, with a focus on tuning your applications
to improve concurrency, if appropriate. If escalations are the reason for a high
average lock wait time, the values of one or both of the LOCKLIST and
MAXLOCKS configuration parameters might be too low.

avg pool read time The average elapsed time for a read request. The value format
is integer.

This value is derived by dividing the value of Pool Read Time attribute by the
value of Pool Total Reads attribute.

This average is important because it can indicate the presence of an I/O wait,
which in turn might indicate that you need to move data to a different device.

avg pool write time The average elapsed time for a write request. The value
format is integer.

30 IBM Tivoli Monitoring for Databases: DB2 Agent: User’s Guide

This value is derived by dividing the value of Pool Write Time attribute by the
value of Pool Total Writes attribute.

avg sort time The average derived by dividing value of Total Sort Time attribute
by the value of the Total Sorts attribute. The average is expressed as elapsed time
and has an integer value.

At a database or application level, this attribute can indicate whether sorting is a

performance issue. Elapsed times are affected by system load. The more processes
you have running, the higher this elapsed time value will be.

binds precompiles The number of binds and precompiles attempted. The value
format is integer. Use this attribute to gain insight into the current level of activity
within the database manager.

cat cache heap full The number of times that an insert into the catalog cache failed
because of a heap-full condition in the database heap. The value format is integer.
The catalog cache draws its storage dynamically from the database heap. Even if
the cache storage has not reached its limit, inserts into the catalog cache can fail
because of a lack of space in the database heap. If the catalog cache heap full count
is not zero, you can correct the insert failure condition by increasing the database
heap size or by reducing the catalog cache size.

cat cache hit ratio The percentage of catalog sections found in cache. The value
format is integer.

This ratio indicates how well the catalog cache is avoiding catalog accesses. If the
ratio is high (more than 0.8), the cache is performing well. A smaller ratio might
indicate that you need to increase the size of the catalog cache. Expect a large ratio
immediately following the first connection to the database.

cat cache inserts The number of times that the system tried to insert table
descriptor information into the catalog cache. The value format is integer. Table
descriptor information is usually inserted into the cache following a failed lookup
to the catalog cache while processing a table, view, or alias reference in an SQL
statement. The catalog cache inserts value includes attempts to insert table
descriptor information that fail because of catalog cache overflow and heap full
conditions.

cat cache lookups The number of times that the catalog cache was referenced to
obtain table descriptor information. The value format is integer.

This attribute includes both successful and unsuccessful accesses to the catalog
cache. To calculate the catalog cache hit ratio use the following formula:
(1 - (cat_cache_inserts / cat_cache_lookups))

This ratio indicates how well the catalog cache is avoiding catalog accesses. If the
ratio is high (more than 0.8), the cache is performing well. A smaller ratio might
indicate that you need to increase the size of the catalog cache. Expect a large ratio
immediately following the first connection to the database.

cat cache overflows The number of times that an insert into the catalog cache
failed because the catalog cache was full. The value format is integer. If the catalog
cache overflows value is large, the catalog cache might be too small for the
workload. Enlarging the catalog cache might improve its performance. If the
workload includes transactions that compile a large number of SQL statements

Chapter 5. Attributes reference 31

 referencing many tables, views, and aliases in a single unit of work, compiling
fewer SQL statements in a single transaction might improve the performance of the
catalog cache. Or if the workload includes the binding of packages containing
many SQL statements referencing many tables, views or aliases, you might want to
split the packages so that they include fewer SQL statements to improve
performance.

client pid The process ID of the client application that made the connection to the
database. The value format is integer. Use this attribute to correlate monitor
information such as CPU and I/O time to your client application. In the case of a
DRDA AS connection, this element is set to 0.

client platform The operating system on which the client application is running.
Use this attribute to analyze problems for remote applications. The value format is
a text string with a maximum of 20 characters. Client platform values are as
shown:
AIX
AS400 DRDA
DOS
HP
MAC
MVS_DRDA
OS/2
SCO
SGI
SNI
SUN
UNKNOWN DRDA
VM DRDA
VSE DRDA
Windows3.x
Windows95
WindowsNT

client prdid The product and version identifier for the software on the client. The
value format is a text string with a maximum of 20 characters. Example: SQL06010.

client protocol The communication protocol that the client application is using to
communicate with the server. The value format is a text string with a maximum of
12 characters. Use this attribute for problem determination of remote applications.
Client protocol values are as shown:
APPC
APPN
CPIC
IPS/SPX
Local
Named Pipe
Netbios
TCPIP
UNKNOWN

commit sql stmts The total number of SQL COMMIT statements that have been
attempted. The value format is integer. A small rate of change in this counter
during the monitor period can indicate that applications are not doing frequent
commits. The lack of frequent commits can lead to problems with logging and data
concurrency. You can also use this attribute to calculate the total number of units of
work by calculating the sum of the following:

32 IBM Tivoli Monitoring for Databases: DB2 Agent: User’s Guide

commit statements attempted
internal commits
rollback statements attempted
internal rollbacks

conn complete time The date and time that a connection request was granted.

Use this attribute to determine when a connection request to the database was
granted.

corr token The DRDA AS correlation token. The value format is a text string with
a maximum of 32 characters.

Use the DRDA correlation token to correlate the processing between the
application server and the application requester. It is an identifier dumped into
logs when errors arise. As a result, you can use it to identify the conversation that
is in error. In some cases, it will be the LUWID of the conversation. If
communications are not using DRDA, this element returns the appl_id attribute
(see Application ID).

country code The country code of the client application. The value format is
integer.

creator The authorization ID of the user that precompiled the application. The
value format is a text string with a maximum of 20 characters. Use this attribute to
help identify the SQL statement that is processing, in conjunction with the
CREATOR column of the package section information in the catalogs.

cursor name The name of the cursor corresponding to this SQL statement. The
value format is a text string with a maximum of 20 characters.

Use this attribute to identify the SQL statement that is processing. This name is
used on an OPEN, FETCH, CLOSE, and PREPARE of an SQL SELECT statement. If
a cursor is not used, this field is blank.

db name The real name of the database for which information is collected or to
which the application is connected. This is the name the database was given when
created. The value format is a simple text string with a maximum of 20 characters.
Use this attribute to identify the specific database to which the data applies.

ddl sql stmts The number of SQL Data Definition Language (DDL) statements that
were executed. The value format is integer.

Use this attribute to determine the level of database activity at the application or
database level. DDL statements are expensive to run because of their impact on the
system catalog tables. As a result, if the value of this attribute is high, determine
the cause and possibly restrict the identified activity from being performed.

deadlocks The total number of deadlocks that have occurred. The value format is
integer. This attribute can indicate that applications are experiencing contention
problems. To resolve the problem, determine in which applications (or application
processes) the deadlock are occurring. You can then modify the application to
enable it to execute concurrently. Some applications, however, might not be capable
of running concurrently.

Chapter 5. Attributes reference 33

 degree parallelism The degree of parallelism requested when the query was
bound. The value format is integer. Use with the Agents Top attribute to determine
if the query achieved maximum level of parallelism.

direct read reqs The number of requests to perform a direct read of one or more
sectors of data. The value format is integer. Use the following formula to calculate
the average number of sectors that are read by a direct read:
direct reads from database / direct read requests

direct read time The elapsed time (in milliseconds) required to perform the direct
reads. The value format is integer. Use the following formula to calculate the
average direct read time per sector:
direct read time / direct reads from database

A high average time can indicate an I/O conflict.

direct reads The number of read operations that do not use the buffer pool. The
value format is integer. Use the following formula to calculate the average number
of sectors that are read by a direct read:
direct reads from database / direct read requests

When using system monitors to track I/O, this data attribute helps to distinguish
database I/O from non-database I/O on the device.

direct write reqs The number of requests to perform a direct write of one or more
sectors of data. The value format is integer. Use the following formula to calculate
the average number of sectors that are written by a direct write:
direct writes to database / direct write requests

direct write time The elapsed time (in milliseconds) required to perform the direct
writes. The value format is integer. Use the following formula to calculate the
average direct write time per sector:
direct write time / direct writes to database

A high average time can indicate an I/O conflict.

direct writes The number of write operations that do not use the buffer pool. The
value format is integer. Use the following formula to calculate the average number
of sectors that are written by a direct write.
direct writes to database / direct write requests

When using system monitors to track I/O, this data attribute helps to distinguish
database I/O from non-database I/O on the device.

dynamic sql stmts The number of dynamic SQL statements that were attempted.
The value format is integer. Use this attribute to calculate the total number of
successful SQL statements at the database or application level by:
1. Adding together the number of Dynamic SQL Statements Attempted and the
Static SQL Statements Attempted
2. Subtracting the number of Failed Statement Operations
The remainder equals the throughput (the number of successful SQL statements)
during the current monitoring period.

34 IBM Tivoli Monitoring for Databases: DB2 Agent: User’s Guide

execution id The ID that the user specified when logging in to the operating
system. This ID is distinct from Authorization ID, which the user specifies when
connecting to the database. The value format is a text string with a maximum of 20
characters. Use this attribute to determine the operating system user ID of the
individual running the monitored application.

failed sql stmts The number of SQL statements that were attempted, but failed.
The value format is integer. This count includes all SQL statements that received a
negative SQLCODE.

Use this attribute to calculate the total number of successful SQL statements at the
database or application level by:
1. Adding together the number of Dynamic SQL Statements Attempted and the
Static SQL Statements Attempted
2. Subtracting the number of Failed Statement Operations
The remainder equals the throughput (the number of successful SQL statements)
during the current monitoring period. This attribute can also help you to
determine the reasons for poor performance; failed statements indicate time wasted
by the database manager, which results in lower throughput for the database.

failed sql stmts pct The percentage of SQL statements that failed to execute
successfully. The value format is integer. This value is derived by dividing the
value of Failed SQL Statements attribute by the value of Total SQL Statements
attribute.

hash join overflows The number of times that hash join data exceeded the
available sort heap space. The value format is integer.

At the database level, if the percentage of Hash Join Small Overflows is greater
than 10% of this value, consider increasing the sort heap size. You can use values
at the application level to evaluate hash join performance for individual
applications.

hash join small overflows The number of times that hash join data exceeded the
available sort heap space by less than 10%. The value format is integer. If this
value and Hash Join Overflows attribute are high, consider increasing the sort
heap threshold. If this value is greater than 10% of Hash Join Overflows, consider
increasing the sort heap size.

int auto rebinds The number of automatic rebinds (or recompiles) that have been
attempted. The value format is integer. Automatic rebinds are the internal binds
the system performs when a package has been invalidated. Use this attribute to
determine the level of database activity at the application or database level.

Because internal automatic rebinds can have a significant impact on performance,

minimize these where possible.

int commits The total number of commits initiated internally by the database
manager. The value format is integer. An internal commit can occur during any of
the following:
v A reorganization
v An import
v A bind or precompile

Chapter 5. Attributes reference 35

 v An application that ends without executing an explicit SQL COMMIT statement
(on UNIX)
Use this attribute to calculate the total number of units of work by calculating the
sum of the following: commit statements attempted, internal commits, rollback
statements attempted, and internal rollbacks.

int deadlock rollbacks The total number of forced rollbacks initiated by the
database manager because of a deadlock. The value format is integer. The database
manager initiates a rollback for the current unit of work in an application that is
experiencing a deadlock. This attribute shows the number of deadlocks that have
been broken. It can indicate the possibility of concurrency problems. It is also
important because internal rollbacks because of deadlocks can cause performance
degradation.

int rollbacks The total number of rollbacks initiated internally by the database
manager. The value format is integer. An internal rollback occurs when any of the
following cannot complete successfully:
v A reorganization
v An import
v A bind or precompile
v An application that ends as a result of a deadlock situation or lock timeout
situation
v An application that ends without executing an explicit COMMIT or ROLLBACK
statement (on Windows)
Use this attribute to calculate the total number of units of work by calculating the
sum of the following: commit statements attempted, internal commits, rollback
statements attempted, and internal rollbacks.

int rows deleted The number of rows deleted from the database as a result of
internal activity. The value format is integer. This attribute can help to gain insight
into internal activity within the database manager. If this activity is high, evaluate
the table design to determine if the referential constraints or triggers that you
defined on the database are necessary.

int rows inserted The number of rows inserted into the database as a result of
internal activity caused by triggers. The value format is integer. This attribute can
help to gain insight into the internal activity within the database manager. If this
activity is high, evaluate the design to determine if you can alter it to reduce this
activity.

int rows updated The number of rows updated from the database as a result of
internal activity. The value format is integer. This attribute can help to gain insight
into internal activity within the database manager. If this activity is high, evaluate
the table design to determine if the referential constraints that you defined on the
database are necessary.

lock escals The number of times that locks have been escalated from several row
locks to a table lock. The value format is integer. This data item includes a count of
all lock escalations, including exclusive lock escalations. When an application
reaches the maximum number of locks allowed and there are no more locks to
escalate, it uses space in the lock list allocated for other applications. When the
entire lock list is full, an error occurs.

36 IBM Tivoli Monitoring for Databases: DB2 Agent: User’s Guide

lock mode The type of lock being held. The mode can help you determine the
source of contention for resources. The value format is a text string with a
maximum of 32 characters. Valid values are as shown:
Exclusive Lock
Intent None (for Dirty Read)
Intention Exclusive Lock
Intention Share Lock
No Lock
Share Lock
Share with Intention Exclusive Lock
Super Exclusive Lock
Update Lock
Unknown

This attribute indicates one of the following, depending on the type of monitor
information being examined:
v The type of lock another application holds on the object that this application is
waiting to lock (for application-monitoring and deadlock-monitoring levels)
v The type of lock held on the object by this application (for object-lock levels)

lock object type The type of object against which the application holds a lock (for
object-lock-level information), or the type of object for which the application is
waiting to obtain a lock (for application-level and deadlock-level information). The
value format is a text string with a maximum of 16 characters. The valid values are
as shown:
Internal
No Lock
Row
Table
Tablespace
Unknown

Use this attribute to help you determine the source of contention for resources.

lock timeouts The number of times that a request to lock an object timed out
instead of being granted. The value format is integer.

This attribute can help you adjust the setting for the LOCKTIMEOUT database
configuration parameter. If the number of lock time-outs becomes excessive when
compared to normal operating levels, an application might be holding locks for
long durations. In this case, this attribute might indicate that you need to analyze
some of the other attributes related to locks and deadlocks to determine if an
application problem exists.

It is also possible to have too few lock time-outs if the LOCKTIMEOUT database
configuration parameter is set too high. In this case, applications might wait
excessively to obtain a lock.

lock wait start time The date and time that the application started waiting to
obtain a lock on the object that is currently locked by another application. This
attribute can help you to determine the severity of resource contention.

lock wait time The total elapsed time waited for a lock.

At the database level, this is the total amount of elapsed time that all applications
were waiting for a lock within this database. The value format is integer.

Chapter 5. Attributes reference 37

 At the application-connection and transaction levels, this is the total amount of
elapsed time that this connection or transaction has waited for a lock to be
granted.

This attribute can be used in conjunction with the Lock Waits attribute to calculate
the average wait time for a lock. This calculation can be performed at either the
database or the application-connection level.

lock waits The total number of times that applications or connections waited for
locks. The value format is integer. At the database level, this is the total number of
times that applications waited for locks within this database. At the
application-connection level, this is the total number of times that this connection
requested a lock but waited because another connection was already holding a lock
on the data.

Use this attribute with the Lock Wait Time attribute to calculate, at the database
level, the average wait time for a lock. This calculation can be done at either the
database or the application-connection level.

If the average lock wait time is high, look for applications that hold many locks, or
have lock escalations, with a focus on tuning your applications to improve
concurrency, if appropriate. If escalations are the reason for a high average lock
wait time, the values of one or both of the LOCKLIST and MAXLOCKS
configuration parameters might be too low.

locks held The number of locks currently held. The value format is integer.

If the monitor information is at the database level, this number represents the total
number of locks currently held by all applications in the database.

If the information is at the application level, this number represents the total
number of locks currently held by all agents for the application.

node name Indicates the originating name. Value format is a simple text string
with a maximum of 32 characters.

open loc curs The number of local cursors currently open for this application,
including those cursors counted by Open Local Cursors with Blocking attribute.
The value format is integer.

Use this attribute with Open Local Cursors with Blocking attribute to calculate the
percentage of local cursors that are blocking cursors. If the percentage is low, you
might be able to improve performance by improving the row blocking in the
application.

For cursors used by remote applications, see Open Remote Cursors attribute.

open loc curs blk The number of local blocking cursors currently open for this
application. The value format is integer.

Use this attribute with Open Local Cursors attribute to calculate the percentage of
local cursors that are blocking cursors. Rejected Block Cursor Requests and
Accepted Block Cursor Requests attributes provide additional information that can
help you tune your configuration parameters to improve row blocking in your
application.

38 IBM Tivoli Monitoring for Databases: DB2 Agent: User’s Guide

For blocking cursors used by remote applications, see Open Remote Cursors with
Blocking attribute.

open rem curs The number of remote cursors currently open for this application,
including the cursors counted by Open Remote Cursors with Blocking attribute.
The value format is integer.

Use this attribute in conjunction with Open Remote Cursors with Blocking
attribute to calculate the percentage of remote cursors that are blocking cursors. If
the percentage is low, you might be able to improve performance by improving the
row blocking in the application. See Open Remote Cursors with Blocking for more
information.

For the number of open cursors used by applications connected to a local database,
see Open Local Cursors attribute.

open rem curs blk The number of remote blocking cursors currently open for this
application. The value format is integer. Use this attribute in conjunction with
Open Remote Cursors attribute to calculate the percentage of remote cursors that
are blocking cursors. If the percentage is low, you might be able to improve
performance by improving the row blocking in the application:
v Check the precompile options for record blocking for treatment of ambiguous
cursors.
v Redefine cursors to allow for blocking.
Rejected Block Cursor Requests and Accepted Block Cursor Requests attributes
provide additional information that can help you tune your configuration
parameters to improve row blocking in your application.

For the number of open blocking cursors used by applications connected to a local
database see Open Local Cursors with Blocking attribute.

package name The name of the package that contains the SQL statement currently
executing. The value format is a text string with a maximum of 20 characters.

Use this attribute to help you identify the application program and the SQL
statement that is executing.

pkg cache hit ratio The percentage of package sections that were found in cache.
The value format is integer.

This ratio tells you whether the package cache is being used effectively. If the hit
ratio is high (more than 0.8), the cache is performing well. A smaller ratio might
indicate that the package cache needs to be increased.

pkg cache inserts The total number of times that a requested section was not
available for use and had to be loaded into the package cache. The value format is
integer. This count includes any implicit prepares performed by the system.

In conjunction with Package Cache Lookups attribute, you can calculate the
package cache hit ratio using the following formula:
1 - (Package Cache Inserts / Package Cache Lookups)

See Package Cache Lookups for information on using this attribute.

Chapter 5. Attributes reference 39

 pkg cache lookups The number of times that an application looked for a section or
package in the package cache. The value format is integer. At a database level, it
indicates the overall number of references since the database was started, or
monitor data was reset. Note that this counter includes the cases where the section
is already loaded in the cache and when the section has to be loaded into the
cache.

To calculate the package cache hit ratio use the following formula:
1 - (Package Cache Inserts / Package Cache Lookups)

The package cache hit ratio tells you whether the package cache is being used
effectively. If the hit ratio is high (more than 0.8), the cache is performing well. A
smaller ratio might indicate that the package cache needs to be increased.

pool data from estore Number of buffer pool data pages copied from extended
storage. The value format is integer.

Required pages are copied from extended storage to the buffer pool. The copy
process can incur the cost of connecting to the shared memory segment, but it
saves the cost of a disk read.

pool data l reads The number of logical read requests for data pages that have
gone through the buffer pool. The value format is integer. This count includes
accesses to data that is: Already in the buffer pool when the database manager
needs to process the page Read into the buffer pool before the database manager
can process the page.

In conjunction with Pool Data Physical Reads attribute, you can calculate the data
page hit ratio for the buffer pool as follows:
1 - (buffer pool data physical reads / buffer pool data logical reads)

In conjunction with Pool Data Physical Reads, Pool Index Physical Reads, and Pool
Index Logical Reads, you can calculate the overall buffer pool hit ratio as follows:
1 - ((buffer pool data physical reads + buffer pool index physical reads) /
(buffer pool data logical reads + buffer pool index logical reads))

Increasing buffer pool size will generally improve the hit ratio until you reach a
point of diminishing return.

pool data p reads The number of read requests that required I/O to get data pages
into the buffer pool. The value format is integer.

See Pool Data Logical Reads and Pool Async Data Reads attributes for information
about how to use this attribute.

pool data to estore Number of buffer pool data pages copied to extended storage.
The value format is integer.

Pages are copied from the buffer pool to extended storage when they are selected
as victim pages. As a result of the copying process, there is sufficient space for new
pages in the buffer pool.

pool data writes The number of times a buffer pool data page was physically
written to disk. The value format is integer. A buffer pool data page is written to
disk for the following reasons:

40 IBM Tivoli Monitoring for Databases: DB2 Agent: User’s Guide

v To free a page in the buffer pool so another page can be read
v To flush the buffer pool.
If a buffer pool data page is written to disk for a high percentage of the Buffer
Pool Data Physical Reads, performance might improve by increasing the number of
buffer pool pages available for the database.

pool hit ratio The buffer pool hit ratio. The value format is integer. The sum of
Pool Data Logical Reads and Pool Index Logical Reads attributes is divided by the
value of Pool Total Reads attribute to derive the ratio.

Use this attribute to determine whether buffer pool assignment is efficient. If the
pool hit ratio is low, increasing the number of buffer pool pages might improve
performance.

pool index from estore Number of buffer pool index pages copied from extended
storage. The value format is integer.

Required index pages are copied from extended storage to the buffer pool. The
copy process can incur the cost of connecting to the shared memory segment, but
it saves the cost of a disk read.

pool index l reads The number of logical read requests for index pages that have
gone through the buffer pool. The value format is integer. This count includes
accesses to index pages that are:
v Already in the buffer pool when the database manager needs to process the
page
v Read into the buffer pool before the database manager can process the page
In conjunction with Pool Index Physical Reads, you can calculate the index page
hit ratio for the buffer pool as follows:
1 - (buffer pool index physical reads / buffer pool index logical reads)

To calculate the overall buffer pool hit ratio, see Pool Data Logical Reads. If the hit
ratio is low, increasing the number of buffer pool pages might improve
performance.

pool index p reads The number of physical read requests to get index pages into
the buffer pool. The value format is integer. See Pool Index Logical Reads for
information about how to use this element.

pool index to estore Number of buffer pool index pages copied to extended
storage. The value format is integer. Pages are copied from the buffer pool to
extended storage when they are selected as victim pages. As a result of the
copying process, there is sufficient space for new pages in the buffer pool.

pool index writes The number of times a buffer pool index page was physically
written to disk. The value format is integer. If a buffer pool index page is written
to disk for a high percentage of the Pool Index Physical Reads, performance might
improve by increasing the number of buffer pool pages available for the database.
If all applications are updating the database, increasing the size of the buffer pool
might have minimal impact on performance; most pages will contain updated data
that must be written to disk.

Chapter 5. Attributes reference 41

 pool read time The total amount of elapsed time spent processing read requests
that caused data or index pages to be physically read from disk to buffer pool. The
value format is integer.

Use this attribute with Buffer Pool Data Physical Reads and Buffer Pool Index
Physical Reads attributes to calculate the average page-read time. This average is
important because it can indicate the presence of an I/O wait, which in turn might
indicate that you need to move data to a different device.

pool total reads The total number of read requests that required I/O to get data
pages and index pages into the buffer pool. The value format is integer. This
attribute is the total of Pool Data Logical Reads, Pool Index Logical Reads, Pool
Data Physical Reads, and Pool Index Physical Reads attributes. Use this attribute to
determine how busy the DB2 server is in terms of I/O activity.

pool total writes The total number of write requests. The value format is integer.
This attribute is the total of Pool Data Writes and Pool Index Writes attributes. Use
this attribute to determine how busy the DB2 server is in terms of write I/O
activity.

pool write time The total amount of time spent physically writing data or index
pages from the buffer pool to disk. The value format is integer. Use this attribute
with Buffer Pool Data Writes and Buffer Pool Index Writes attributes to calculate
the average page-write time. This average is important because it can indicate the
presence of an I/O wait, which in turn might indicate that you need to move data
to a different device.

prefetch wait time The time an application spent waiting for an I/O server
(prefetcher) to finish loading pages into the buffer pool. The value format is
integer. This attribute can be used to experiment with changing the number of I/O
servers and the I/O server sizes.

prev uow stop time The date and time that the unit of work completed.

Use this attribute with the UOW Stop Time attribute to calculate the total elapsed
time between COMMIT/ROLLBACK points, and with UOW Start Time attribute to
calculate the time spent in the application between units of work.

query card estimate An estimate of the number of rows that will be returned by a
query. The value format is integer. You can compare this estimate by the SQL
compiler with the run time actuals.

query cost estimate Estimated cost, in timerons, for a query, as determined by the
SQL compiler. The value format is integer. This attribute allows correlation of
actual run-time with the compile-time estimates.

rej curs blk The number of times that a request for an I/O block at the server was
rejected and the request was converted to non-blocked I/O. If there are many
cursors blocking data, the communication heap can become full. The value format
is integer. When this heap is full, I/O blocks are not allocated for blocking cursors;
however, an error condition does not alert you to this condition. If cursors are
unable to block data, performance can be affected adversely.

rollback sql stmts The total number of SQL ROLLBACK statements that have been
attempted. The value format is integer. A rollback can result from an application

42 IBM Tivoli Monitoring for Databases: DB2 Agent: User’s Guide

request, a deadlock, or an error situation. This attribute counts only the number of
rollback statements issued from applications.

At the application level, this attribute can help you determine the level of database
activity for the application and the amount of conflict with other applications.

At the database level, it can help you determine the amount of activity in the
database and the amount of conflict between applications on the database.

rows deleted The number of row deletions attempted. The value format is integer.
Use this attribute to gain insight into the current level of activity within the
database manager.

rows inserted The number of row insertions attempted. The value format is
integer. Use this attribute to gain insight into the current level of activity within
the database manager.

rows read The number of rows read from the table. The value format is integer.
This attribute helps to identify tables with heavy usage for which you might want
to create additional indexes.

rows selected The number of rows that have been selected and returned to the
application. The value format is integer. Use this attribute to gain insight into the
current level of activity within the database manager.

rows updated The number of row updates attempted. The value format is integer.
Use this attribute to gain insight into the current level of activity within the
database manager.

rows written The number of rows changed (inserted, deleted or updated) in the
table. The value format is integer. A high value for table-level information indicates
heavy usage of the table. If so, you might want to use the Run Statistics
(RUNSTATS) utility to maintain efficiency of the packages used for this table.

section number The internal section number in the package for the SQL statement
currently processing or most recently processed. The value format is integer.

select sql stmts The number of SQL SELECT statements that were executed. The
value format is integer.

Use this attribute to determine the level of database activity at the application or
database level. You can also use the following formula to determine the ratio of
SELECT statements to the total statements by:
1. Adding together the number of static SQL statements attempted and dynamic
SQL statements attempted
2. Dividing the resulting total by the number of select SQL statements executed

snapshot time The date and time when the database system monitor information
was collected.

sort overflows The total number of sorts that ran out of sort heap and might have
required disk space for temporary storage. The value format is integer. At a
database or application level, use this element in conjunction with Total Sorts
attribute. This attribute can help to determine the source of contention for
resources.

Chapter 5. Attributes reference 43

 sort overflows pct The percentage of sorts that ran out of sort heap and might
have required disk space for temporary storage. The value format is integer. This
percentage is calculated by dividing the value of Sort Overflows attribute by the
value of the Total Sorts attribute. At a database or application level, use this
attribute to evaluate the percentage of sorts that required overflow to disk. If this
percentage is high, you might want to adjust the database configuration by
increasing the value of sortheap.

static sql stmts The number of static SQL statements that were attempted. The
value format is integer.

Use this attribute to calculate the total number of successful SQL statements at the
database or application level by:
1. Adding together the number of Dynamic SQL Statements Attempted and the
Static SQL Statements Attempted
2. Subtracting the number of Failed Statement Operations
The remainder equals the throughput (the number of successful SQL statements)
during the current monitoring period.

stmt operation The statement operation currently being processed or most recently
processed (if none is currently running). The value format is a text string with a
maximum of 20 characters. The valid values are
CLOSE
DESCRIBE
EXECUTE
EXECUTE IMMEDIATE
FETCH
OPEN
PREPARE
STATIC COMMIT
STATIC ROLLBACK
UNKNOWN

Use this attribute to determine the operation that is executing or recently finished.

stmt start The date and time that the most recent SQL statement operation started.

stmt stop The date and time that the most recent SQL statement operation
stopped. The value format is a text string with a maximum of 16 characters. If the
statement is still running, this field is 0 (zero). Use this attribute with the Statement
Start attribute to calculate the elapsed execution time for the statement operation.

stmt text This is the text of the dynamic SQL statement. For application snapshots,
the statement text helps you identify what the application was executing when the
snapshot was taken, or most recently processed if no statement was being
processed at the time the snapshot was taken. For dynamic SQL statements, this
attribute identifies the SQL text associated with a package. The value format is text
string with a maximum of 2000 characters.

stmt type The type of SQL statement processed. The value format is a text string
with a maximum of 32 characters. The valid values are:
DYNAMIC
NON-STATEMENT OPERATION
STATIC
UNKNOWN STMT TYPE

Use this attribute to determine the type of statement that is executing.

44 IBM Tivoli Monitoring for Databases: DB2 Agent: User’s Guide

table name The name of the table. The value format is a text string with a
maximum of 20 characters. Along with Table Schema attribute, this attribute can
help to determine the source of contention for resources.

table schema The schema of the table. The value format is a text string with a
maximum of 20 characters. Along with Table Name attribute, this attribute can
help to determine the source of contention for resources.

tablespace name The name of a tablespace that the application is waiting to lock.
The value format is a text string with a maximum of 20 characters. This attribute
can help you to determine the source of contention for resources.

total hash joins The total number of hash joins executed. The value format is
integer.

At the database or application level, use this value in conjunction with Hash Join
Overflows attribute and Hash Join Small Overflows attribute to determine if a
significant percentage of hash joins would benefit from modest increases in the sort
heap size.

total hash loops The total number of times that a single partition of a hash join
was larger than the available sort heap space. The value format is integer. Values
for this attribute indicate inefficient execution of hash joins. This might indicate
that the sort heap size is too small or the sort heap threshold is too small. Use this
value in conjunction with the other hash join variables to tune the sort heap size
(sortheap) and sort heap threshold (sheapthres) configuration parameters.

total sort time The total elapsed time (in milliseconds) for all sorts that have been
executed. The value format is integer. At a database or application level, use this
element with Total Sorts attribute to calculate the average sort time. This average
can indicate whether sorting is a performance concern.

total sorts The total number of sorts that have been executed. The value format is
integer. At a database or application level, use this value with the Sort Overflows
attribute to calculate the percentage of sorts that need more heap space. You can
also use it with Total Sort Time attribute to calculate the average sort time. If the
number of sort overflows is small with respect to the total sorts, increasing the sort
heap size might have little impact on performance, unless this buffer size is
increased substantially.

total sql stmt The number of SQL statements that have been attempted since the
latter of: application start up, database activation, or last reset. The value format is
integer. Use this attribute to measure database activity at the database or
application level. To calculate the SQL statement throughput for a given period,
divide this attribute by the elapsed time between two snapshots.

uid sql stmts The number of SQL UPDATE, INSERT, and DELETE statements that
were executed. The value format is integer.

Use this attribute to determine the level of database activity at the application or
database level. You can also use the following formula to determine the ratio of
UPDATE, INSERT and DELETE statements to the total number of statements:
1. Adding together the number of static SQL statements attempted and the
dynamic SQL statements attempted

Chapter 5. Attributes reference 45

 2. Dividing the number of UPDATE/INSERT/DELETE SQL statements executed
by the sum derived in step 1

uow comp status The completion status of the previous UOW. The value format is
a text string with a maximum of 32 characters. Use this attribute to determine if
the unit of work ended because of a deadlock or abnormal termination.

Valid values are:

UOW Commit
UOW RB - Lock Timeout
UOW RB due to Abend
UOW RB due to Deadlock
UOW Rolled Back

uow lock wait time The total amount of elapsed time this unit of work has spent
waiting for locks. The value format is integer. This attribute can help you to
determine the severity of the resource contention problem.

uow log space used The amount of log space (in bytes) used in the current unit of
work of the monitored application. The value format is integer. Use this attribute
to understand the logging requirements at the unit-of-work level.

uow start time The date and time that the unit of work first required database
resources. This resource requirement occurs at the first SQL statement execution for
the unit of work.

uow stop time The date and time that the most recent unit of work completed,
which occurs when database changes are committed or rolled back.

x lock escals The number of times that locks have been escalated from several row
locks to one exclusive table lock, or the number of times an exclusive lock on a
row caused the table lock to become an exclusive lock. The value format is integer.
A lock is escalated when the total number of locks held by an application reaches
the maximum amount of lock list space available to the application. The amount of
lock list space available is determined by the LOCKLIST and MAXLOCKS
configuration parameters. Other applications cannot access data held by an
exclusive lock.

Because exclusive locks can impact the concurrency of your data, it is important to
track them. When an application reaches the maximum number of locks allowed
and there are no more locks to escalate, it uses space in the lock list allocated for
other applications. When the entire lock list is full, an error occurs.

Application attributes (Unicode group 00)

The Application attribute group provides information about application activities.

Note: This attribute group contains Unicode versions of attributes listed in

Application attributes group 00, and therefore has a similar list of attributes
as that group. There is a limit to the number of attributes that can be
contained in an application group; therefore, this group holds additional
attributes that could not be contained by the Application attribute group 00
such as the Unicode versions of those attributes. Unicode versions of
attributes are denoted by a Unicode in parenthesis at the end of the attribute
name (for example, appl name (Unicode)).
You can use this information to monitor the performance of your applications and

46 IBM Tivoli Monitoring for Databases: DB2 Agent: User’s Guide

to identify problem areas for corrective action. All values are integers calculated
from the first application connection, unless otherwise noted. (Integer values are
32-bit or 0 to 4 294 967 296.)

acc curs blk The number of times that a request for an I/O block was accepted.
The value format is integer. Use this attribute with the Rejected Block Cursor
Requests attribute to calculate the percentage of blocking requests that are accepted
and/or rejected.

agent id The application handle, which is a system-wide unique ID for the

application. The value format is integer. On multi-node systems, where a database
is partitioned, this ID is the same on every node where the application can make a
secondary connection.

Use the application handle to uniquely identify an active application (application

handle is synonymous with agent ID).

agent id holding lk The application handle of the agent holding a lock for which
this application is waiting. The value format is integer. The lock monitor group
must be turned on to obtain this information.

This attribute can help you to determine which applications are in contention for
resources.

agent sys cpu time The total system CPU time (in seconds and microseconds) used
by the database manager agent process.

This counter includes CPU time for both SQL and non-SQL statements, and any
fenced user-defined functions (UDFs). System CPU represents the time spent in
system calls.

User CPU represents time spent executing database manager code. The value
format is a text string with a maximum length of 12 characters. Use this attribute,
along with the other related CPU-time attributes, to help you understand the level
of activity within an application. It can help to identify applications that might
benefit from additional tuning.

agent usr cpu time The total CPU time (in seconds and microseconds) used by the
database manager agent process.

This counter includes time spent on both SQL and non-SQL statements, and any
fenced user-defined functions (UDFs) or stored procedures executed by the
application. System CPU represents the time spent in system calls.

User CPU represents time spent executing database manager code. The value
format is a text string with a maximum of 12 characters. Use this attribute, along
with the other CPU-time related attributes, to help you identify applications or
queries that consume large amounts of CPU time.

appl con time The date and time that an application started a connection request.

appl id The identifier generated when the application connects to the database at
the database manager or when DDCS receives a request to connect to a DRDA
database. The value format is a text string, with a maximum of 32 characters. The
identifier is unique across the network. The application ID appears in various
formats, which depend on the communication protocol between the client and the

Chapter 5. Attributes reference 47

 server on which the database manager and/or DDCS are running. Each of the
formats consists of three parts separated by periods. Use this ID (known on both
the client and server) to correlate the client and server parts of the application.

appl id holding lk (Unicode) The application ID of the application that is holding

a lock on the object that this application is waiting to obtain. The value format is a
text string with a maximum of 96 bytes. This attribute is globalized.

This attribute can help you to determine which applications are in contention for
resources. Specifically, it can help you to identify the application handle (agent ID)
and table ID that are holding the lock.

appl idle time The number of seconds since an application issued a request to the
server. The value format is integer.

This includes applications that have not terminated a transaction (for example,
have not issued a commit or rollback). Use this information to force users that
have been idle for a specified number of seconds.

appl name (Unicode) The name of the application running at the client as known
to the database manager or DB2 Connect. The value format is a text string, with a
maximum of 60 bytes. Example: *Local.db2inst1.990212202018. This attribute is
globalized.

Use this attribute with the Application ID to relate data items with your
application. In a client/server environment, this name is passed from the client to
the server to establish the database connection. For DRDA-AS connections, this
name is the DRDA external name.

appl status The status of the application being monitored. This attribute can help
you diagnose potential application problems. The value format is a text string with
a maximum of 64 characters. Application status values are as shown:
Backing Up Database
Commit Active
Compiling SQL Stmt
Connect Pending
Connected
Creating Database
Disconnect Pending
I/O Error Waiting
Loading Database
Lock Waiting
Prepared Transaction
Quiescing a Tablespace
Recompiling Plan
Request Interrupted
Restarting Database
Restoring Database
Rollback Active
Trans. heuristically aborted
Trans heuristically committed
Transaction ended
UOW Executing
UOW Waiting in the application
Unloading Database

auth id (Unicode) The authorization ID of the user who invoked the application
that is being monitored. On a DB2 Connect gateway node, this is the user’s

48 IBM Tivoli Monitoring for Databases: DB2 Agent: User’s Guide

authorization ID on the host. The value format is a text string with a maximum of
60 bytes. This attribute is globalized. Use this attribute to determine who invoked
the application.

avg lock waittime The average elapsed time spent waiting for a lock. The value
format is integer. If the average lock wait time is high, look for applications that
hold many locks, or have lock escalations, with a focus on tuning your applications
to improve concurrency, if appropriate. If escalations are the reason for a high
average lock wait time, the values of one or both of the LOCKLIST and
MAXLOCKS configuration parameters might be too low.

avg pool read time The average elapsed time for a read request. The value format
is integer.

This value is derived by dividing the value of Pool Read Time attribute by the
value of Pool Total Reads attribute.

This average is important because it can indicate the presence of an I/O wait,
which in turn might indicate that you need to move data to a different device.

avg pool write time The average elapsed time for a write request. The value
format is integer.

This value is derived by dividing the value of Pool Write Time attribute by the
value of Pool Total Writes attribute.

avg sort time The average derived by dividing value of Total Sort Time attribute
by the value of the Total Sorts attribute. The average is expressed as elapsed time
and has an integer value.

At a database or application level, this attribute can indicate whether sorting is a

performance issue. Elapsed times are affected by system load. The more processes
you have running, the higher this elapsed time value will be.

binds precompiles The number of binds and precompiles attempted. The value
format is integer. Use this attribute to gain insight into the current level of activity
within the database manager.

cat cache heap full The number of times that an insert into the catalog cache failed
because of a heap-full condition in the database heap. The value format is integer.
The catalog cache draws its storage dynamically from the database heap. Even if
the cache storage has not reached its limit, inserts into the catalog cache can fail
because of a lack of space in the database heap. If the catalog cache heap full count
is not zero, you can correct the insert failure condition by increasing the database
heap size or by reducing the catalog cache size.

cat cache hit ratio The percentage of catalog sections found in cache. The value
format is integer.

This ratio indicates how well the catalog cache is avoiding catalog accesses. If the
ratio is high (more than 0.8), the cache is performing well. A smaller ratio might
indicate that you need to increase the size of the catalog cache. Expect a large ratio
immediately following the first connection to the database.

cat cache inserts The number of times that the system tried to insert table
descriptor information into the catalog cache. The value format is integer. Table

Chapter 5. Attributes reference 49

 descriptor information is usually inserted into the cache following a failed lookup
to the catalog cache while processing a table, view, or alias reference in an SQL
statement. The catalog cache inserts value includes attempts to insert table
descriptor information that fail because of catalog cache overflow and heap full
conditions.

cat cache lookups The number of times that the catalog cache was referenced to
obtain table descriptor information. The value format is integer.

This attribute includes both successful and unsuccessful accesses to the catalog
cache. To calculate the catalog cache hit ratio use the following formula:
(1 - (cat_cache_inserts / cat_cache_lookups))

This ratio indicates how well the catalog cache is avoiding catalog accesses. If the
ratio is high (more than 0.8), the cache is performing well. A smaller ratio might
indicate that you need to increase the size of the catalog cache. Expect a large ratio
immediately following the first connection to the database.

cat cache overflows The number of times that an insert into the catalog cache
failed because the catalog cache was full. The value format is integer. If the catalog
cache overflows value is large, the catalog cache might be too small for the
workload. Enlarging the catalog cache might improve its performance. If the
workload includes transactions that compile a large number of SQL statements
referencing many tables, views, and aliases in a single unit of work, compiling
fewer SQL statements in a single transaction might improve the performance of the
catalog cache. Or if the workload includes the binding of packages containing
many SQL statements referencing many tables, views or aliases, you might want to
split the packages so that they include fewer SQL statements to improve
performance.

client pid The process ID of the client application that made the connection to the
database. The value format is integer. Use this attribute to correlate monitor
information such as CPU and I/O time to your client application. In the case of a
DRDA AS connection, this element will be set to 0.

client platform The operating system on which the client application is running.
Use this attribute to analyze problems for remote applications. The value format is
a text string with a maximum of 20 characters. Client platform values are as
shown:
AIX
AS400 DRDA
DOS
HP
MAC
MVS_DRDA
OS/2
SCO
SGI
SNI
SUN
UNKNOWN DRDA
VM DRDA
VSE DRDA
Windows3.x
Windows95
WindowsNT

50 IBM Tivoli Monitoring for Databases: DB2 Agent: User’s Guide

client prdid The product and version identifier for the software on the client. The
value format is a text string with a maximum of 20 characters. Example: SQL06010.

client protocol The communication protocol that the client application is using to
communicate with the server. The value format is a text string with a maximum of
12 characters. Use this attribute for problem determination of remote applications.
Client protocol values are as shown:
APPC
APPN
CPIC
IPS/SPX
Local
Named Pipe
Netbios
TCPIP
UNKNOWN

commit sql stmts The total number of SQL COMMIT statements that have been
attempted. The value format is integer. A small rate of change in this counter
during the monitor period can indicate that applications are not doing frequent
commits. The lack of frequent commits can lead to problems with logging and data
concurrency. You can also use this attribute to calculate the total number of units of
work by calculating the sum of the following:
commit statements attempted
internal commits
rollback statements attempted
internal rollbacks

conn complete time The date and time that a connection request was granted.

corr token (Unicode) The DRDA AS correlation token. The value format is a text
string with a maximum of 96 bytes. This attribute is globalized.

Use the DRDA correlation token to correlate the processing between the
application server and the application requester. It is an identifier dumped into
logs when errors arise. As a result, you can use it to identify the conversation that
is in error. In some cases, it will be the LUWID of the conversation. If
communications are not using DRDA, this element returns the appl_id attribute
(see Application ID).

country code The country code of the client application. The value format is
integer.

creator (Unicode) The authorization ID of the user that precompiled the

application. The value format is a text string with a maximum of 60 bytes. This
attribute is globalized. Use this attribute to help identify the SQL statement that is
processing, in conjunction with the CREATOR column of the package section
information in the catalogs.

cursor name (Unicode) The name of the cursor corresponding to this SQL
statement. The value format is a text string with a maximum of 60 bytes. This
attribute is globalized.

Use this attribute to identify the SQL statement that is processing. This name will
be used on an OPEN, FETCH, CLOSE, and PREPARE of an SQL SELECT
statement. If a cursor is not used, this field will be blank.

Chapter 5. Attributes reference 51

 db name (Unicode) The real name of the database for which information is
collected or to which the application is connected. This is the name the database
was given when created. The value format is a simple text string with a maximum
of 60 bytes. This attribute is globalized. Use this attribute to identify the specific
database to which the data applies.

ddl sql stmts The number of SQL Data Definition Language (DDL) statements that
were executed. The value format is integer.

Use this attribute to determine the level of database activity at the application or
database level. DDL statements are expensive to run because of their impact on the
system catalog tables. As a result, if the value of this attribute is high, determine
the cause and possibly restrict the identified activity from being performed.

deadlocks The total number of deadlocks that have occurred. The value format is
integer. This attribute can indicate that applications are experiencing contention
problems. To resolve the problem, determine in which applications (or application
processes) the deadlock are occurring. You can then modify the application to
enable it to execute concurrently. Some applications, however, are not capable of
running concurrently.

degree parallelism The degree of parallelism requested when the query was
bound. The value format is integer. Use with the Agents Top attribute to determine
if the query achieved maximum level of parallelism.

direct read reqs The number of requests to perform a direct read of one or more
sectors of data. The value format is integer. Use the following formula to calculate
the average number of sectors that are read by a direct read:
direct reads from database / direct read requests

direct read time The elapsed time (in milliseconds) required to perform the direct
reads. The value format is integer. Use the following formula to calculate the
average direct read time per sector:
direct read time / direct reads from database

A high average time can indicate an I/O conflict.

direct reads The number of read operations that do not use the buffer pool. The
value format is integer. Use the following formula to calculate the average number
of sectors that are read by a direct read:
direct reads from database / direct read requests

When using system monitors to track I/O, this data attribute helps to distinguish
database I/O from non-database I/O on the device.

direct write reqs The number of requests to perform a direct write of one or more
sectors of data. The value format is integer. Use the following formula to calculate
the average number of sectors that are written by a direct write:
direct writes to database / direct write requests

direct write time The elapsed time (in milliseconds) required to perform the direct
writes. The value format is integer. Use the following formula to calculate the
average direct write time per sector:
direct write time / direct writes to database

52 IBM Tivoli Monitoring for Databases: DB2 Agent: User’s Guide

A high average time can indicate an I/O conflict.

direct writes The number of write operations that do not use the buffer pool. The
value format is integer. Use the following formula to calculate the average number
of sectors that are written by a direct write.
direct writes to database / direct write requests

When using system monitors to track I/O, this data attribute helps to distinguish
database I/O from non-database I/O on the device.

dynamic sql stmts The number of dynamic SQL statements that were attempted.
The value format is integer. Use this attribute to calculate the total number of
successful SQL statements at the database or application level by:
1. Adding together the number of Dynamic SQL Statements Attempted and the
Static SQL Statements Attempted
2. Subtracting the number of Failed Statement Operations
The remainder equals the throughput (the number of successful SQL statements)
during the current monitoring period.

execution id (Unicode) The ID that the user specified when logging in to the
operating system. This ID is distinct from Authorization ID, which the user
specifies when connecting to the database. The value format is a text string with a
maximum of 60 bytes. This attribute is globalized. Use this attribute to determine
the operating system user ID of the individual running the monitored application.

failed sql stmts The number of SQL statements that were attempted, but failed.
The value format is integer. This count includes all SQL statements that received a
negative SQLCODE.

Use this attribute to calculate the total number of successful SQL statements at the
database or application level by:
1. Adding together the number of Dynamic SQL Statements Attempted and the
Static SQL Statements Attempted
2. Subtracting the number of Failed Statement Operations
The remainder equals the throughput (the number of successful SQL statements)
during the current monitoring period. This attribute can also help you to
determine the reasons for poor performance; failed statements indicate time wasted
by the database manager, which results in lower throughput for the database.

failed sql stmts pct The percentage of SQL statements that failed to execute
successfully. The value format is integer. This value is derived by dividing the
value of Failed SQL Statements attribute by the value of Total SQL Statements
attribute.

hash join overflows The number of times that hash join data exceeded the
available sort heap space. The value format is integer.

At the database level, if the percentage of Hash Join Small Overflows is greater
than 10% of this value, consider increasing the sort heap size. You can use values
at the application level to evaluate hash join performance for individual
applications.

hash join small overflows The number of times that hash join data exceeded the
available sort heap space by less than 10%. The value format is integer. If this

Chapter 5. Attributes reference 53

 value and Hash Join Overflows attribute are high, consider increasing the sort
heap threshold. If this value is greater than 10% of Hash Join Overflows, consider
increasing the sort heap size.

int auto rebinds The number of automatic rebinds (or recompiles) that have been
attempted. The value format is integer. Automatic rebinds are the internal binds
the system performs when a package has been invalidated. Use this attribute to
determine the level of database activity at the application or database level.

Because internal automatic rebinds can have a significant impact on performance,

they need to be minimized where possible.

int commits The total number of commits initiated internally by the database
manager. The value format is integer. An internal commit can occur during any of
the following conditions:
v A reorganization
v An import
v A bind or precompile
v An application that ends without executing an explicit SQL COMMIT statement
(on UNIX)
Use this attribute to calculate the total number of units of work by calculating the
sum of the following: commit statements attempted, internal commits, rollback
statements attempted, and internal rollbacks.

int deadlock rollbacks The total number of forced rollbacks initiated by the
database manager because of a deadlock. The value format is integer. The database
manager initiates a rollback for the current unit of work in an application that is
experiencing a deadlock. This attribute shows the number of deadlocks that have
been broken. It can indicate the possibility of concurrency problems. It is also
important because internal rollbacks because of deadlocks can cause performance
degradation.

int rollbacks The total number of rollbacks initiated internally by the database
manager. The value format is integer. An internal rollback occurs when any of the
following cannot complete successfully:
v A reorganization
v An import
v A bind or precompile
v An application that ends as a result of a deadlock situation or lock timeout
situation
v An application that ends without executing an explicit COMMIT or ROLLBACK
statement (on Windows)
Use this attribute to calculate the total number of units of work by calculating the
sum of the following: commit statements attempted, internal commits, rollback
statements attempted, and internal rollbacks.

int rows deleted The number of rows deleted from the database as a result of
internal activity. The value format is integer. This attribute can help to gain insight
into internal activity within the database manager. If this activity is high, evaluate
the table design to determine if the referential constraints or triggers that you
defined on the database are necessary.

54 IBM Tivoli Monitoring for Databases: DB2 Agent: User’s Guide

int rows inserted The number of rows inserted into the database as a result of
internal activity caused by triggers. The value format is integer. This attribute can
help to gain insight into the internal activity within the database manager. If this
activity is high, evaluate the design to determine if you can alter it to reduce this
activity.

int rows updated The number of rows updated from the database as a result of
internal activity. The value format is integer. This attribute can help to gain insight
into internal activity within the database manager. If this activity is high, evaluate
the table design to determine if the referential constraints that you defined on the
database are necessary.

lock escals The number of times that locks have been escalated from several row
locks to a table lock. The value format is integer. This data item includes a count of
all lock escalations, including exclusive lock escalations. When an application
reaches the maximum number of locks allowed and there are no more locks to
escalate, it uses space in the lock list allocated for other applications. When the
entire lock list is full, an error occurs.

lock mode The type of lock being held. The mode can help you determine the
source of contention for resources. The value format is a text string with a
maximum of 32 characters. The valid values are as shown:
Exclusive Lock
Intent None (for Dirty Read)
Intention Exclusive Lock
Intention Share Lock
No Lock
Share Lock
Share with Intention Exclusive Lock
Super Exclusive Lock
Update Lock
Unknown

This attribute indicates one of the following, depending on the type of monitor
information being examined:
v The type of lock another application holds on the object that this application is
waiting to lock (for application-monitoring and deadlock-monitoring levels)
v The type of lock held on the object by this application (for object-lock levels)

lock object type The type of object against which the application holds a lock (for
object-lock-level information), or the type of object for which the application is
waiting to obtain a lock (for application-level and deadlock-level information). The
value format is a text string with a maximum of 16 characters. The valid values are
as shown:
Internal
No Lock
Row
Table
Tablespace
Unknown

Use this attribute to help you determine the source of contention for resources.

lock timeouts The number of times that a request to lock an object timed-out
instead of being granted. The value format is integer.

Chapter 5. Attributes reference 55

 This attribute can help you adjust the setting for the LOCKTIMEOUT database
configuration parameter. If the number of lock time-outs becomes excessive when
compared to normal operating levels, an application might be holding locks for
long durations. In this case, this attribute might indicate that you need to analyze
some of the other attributes related to locks and deadlocks to determine if an
application problem exists.

It is also possible to have too few lock time-outs if the LOCKTIMEOUT database
configuration parameter is set too high. In this case, applications might wait
excessively to obtain a lock.

lock wait start time The date and time that the application started waiting to
obtain a lock on the object that is currently locked by another application.

lock wait time The total elapsed time waited for a lock.

At the database level, this is the total amount of elapsed time that all applications
were waiting for a lock within this database. The value format is integer.

At the application-connection and transaction levels, this is the total amount of

elapsed time that this connection or transaction has waited for a lock to be
granted.

This attribute can be used in conjunction with the Lock Waits attribute to calculate
the average wait time for a lock. This calculation can be performed at either the
database or the application-connection level.

lock waits The total number of times that applications or connections waited for
locks. The value format is integer. At the database level, this is the total number of
times that applications waited for locks within this database. At the
application-connection level, this is the total number of times that this connection
requested a lock but waited because another connection was already holding a lock
on the data.

Use this attribute with the Lock Wait Time attribute to calculate, at the database
level, the average wait time for a lock. This calculation can be done at either the
database or the application-connection level.

If the average lock wait time is high, look for applications that hold many locks, or
have lock escalations, with a focus on tuning your applications to improve
concurrency, if appropriate. If escalations are the reason for a high average lock
wait time, the values of one or both of the LOCKLIST and MAXLOCKS
configuration parameters might be too low.

locks held The number of locks currently held. The value format is integer.

If the monitor information is at the database level, this number represents the total
number of locks currently held by all applications in the database.

If the information is at the application level, this number represents the total
number of locks currently held by all agents for the application.

node name Indicates the originating name. Value format is a simple text string
with a maximum of 32 characters.

56 IBM Tivoli Monitoring for Databases: DB2 Agent: User’s Guide

open loc curs The number of local cursors currently open for this application,
including those cursors counted by Open Local Cursors with Blocking attribute.
The value format is integer.

Use this attribute with Open Local Cursors with Blocking attribute to calculate the
percentage of local cursors that are blocking cursors. If the percentage is low, you
might be able to improve performance by improving the row blocking in the
application.

For cursors used by remote applications, see Open Remote Cursors attribute.

open loc curs blk The number of local blocking cursors currently open for this
application. The value format is integer.

Use this attribute with Open Local Cursors attribute to calculate the percentage of
local cursors that are blocking cursors. Rejected Block Cursor Requests and
Accepted Block Cursor Requests attributes provide additional information that can
help you tune your configuration parameters to improve row blocking in your
application.

For blocking cursors used by remote applications, see Open Remote Cursors with
Blocking attribute.

open rem curs The number of remote cursors currently open for this application,
including the cursors counted by Open Remote Cursors with Blocking attribute.
The value format is integer.

Use this attribute in conjunction with Open Remote Cursors with Blocking
attribute to calculate the percentage of remote cursors that are blocking cursors. If
the percentage is low, you be able to improve performance by improving the row
blocking in the application. See Open Remote Cursors with Blocking for more
information.

For the number of open cursors used by applications connected to a local database,
see Open Local Cursors attribute.

open rem curs blk The number of remote blocking cursors currently open for this
application. The value format is integer. Use this attribute in conjunction with
Open Remote Cursors attribute to calculate the percentage of remote cursors that
are blocking cursors. If the percentage is low, you be able to improve performance
by improving the row blocking in the application:
v Check the precompile options for record blocking for treatment of ambiguous
cursors.
v Redefine cursors to allow for blocking.
Rejected Block Cursor Requests and Accepted Block Cursor Requests attributes
provide additional information that help you tune your configuration parameters
to improve row blocking in your application.

For the number of open blocking cursors used by applications connected to a local
database see Open Local Cursors with Blocking attribute.

package name (Unicode) The name of the package that contains the SQL statement
currently executing. The value format is a text string with a maximum of 60 bytes.
This attribute is globalized.

Chapter 5. Attributes reference 57

 Use this attribute to help you identify the application program and the SQL
statement that is executing.

pkg cache hit ratio The percentage of package sections that were found in cache.
The value format is integer.

This ratio tells you whether the package cache is being used effectively. If the hit
ratio is high (more than 0.8), the cache is performing well. A smaller ratio indicates
that the package cache needs to be increased.

pkg cache inserts The total number of times that a requested section was not
available for use and had to be loaded into the package cache. The value format is
integer. This count includes any implicit prepares performed by the system.

In conjunction with Package Cache Lookups attribute, you can calculate the
package cache hit ratio using the following formula:
1 - (Package Cache Inserts / Package Cache Lookups)

See Package Cache Lookups for information on using this attribute.

pkg cache lookups The number of times that an application looked for a section or
package in the package cache. The value format is integer. At a database level, it
indicates the overall number of references since the database was started, or
monitor data was reset. Note that this counter includes the cases where the section
is already loaded in the cache and when the section has to be loaded into the
cache.

To calculate the package cache hit ratio use the following formula:
1 - (Package Cache Inserts / Package Cache Lookups)

The package cache hit ratio tells you whether the package cache is being used
effectively. If the hit ratio is high (more than 0.8), the cache is performing well. A
smaller ratio indicates that the package cache needs to be increased.

pool data from estore Number of buffer pool data pages copied from extended
storage. The value format is integer.

Required pages are copied from extended storage to the buffer pool. The copy
process incur the cost of connecting to the shared memory segment, but it saves
the cost of a disk read.

pool data l reads The number of logical read requests for data pages that have
gone through the buffer pool. The value format is integer. This count includes
accesses to data that is: Already in the buffer pool when the database manager
needs to process the page Read into the buffer pool before the database manager
can process the page.

In conjunction with Pool Data Physical Reads attribute, you can calculate the data
page hit ratio for the buffer pool as follows:
1 - (buffer pool data physical reads / buffer pool data logical reads)

In conjunction with Pool Data Physical Reads, Pool Index Physical Reads, and Pool
Index Logical Reads, you can calculate the overall buffer pool hit ratio as follows:
1 - ((buffer pool data physical reads + buffer pool index physical reads) /
(buffer pool data logical reads + buffer pool index logical reads))

58 IBM Tivoli Monitoring for Databases: DB2 Agent: User’s Guide

Increasing buffer pool size generally improves the hit ratio until you reach a point
of diminishing return.

pool data p reads The number of read requests that required I/O to get data pages
into the buffer pool. The value format is integer.

See Pool Data Logical Reads and Pool Async Data Reads attributes for information
about how to use this attribute.

pool data to estore Number of buffer pool data pages copied to extended storage.
The value format is integer.

Pages are copied from the buffer pool to extended storage when they are selected
as victim pages. As a result of the copying process, there is sufficient space for new
pages in the buffer pool.

pool data writes The number of times a buffer pool data page was physically
written to disk. The value format is integer. A buffer pool data page is written to
disk for the following reasons:
v To free a page in the buffer pool so another page can be read
v To flush the buffer pool.
If a buffer pool data page is written to disk for a high percentage of the Buffer
Pool Data Physical Reads, performance improve by increasing the number of
buffer pool pages available for the database.

pool hit ratio The buffer pool hit ratio. The value format is integer. The sum of
Pool Data Logical Reads and Pool Index Logical Reads attributes is divided by the
value of Pool Total Reads attribute to derive the ratio.

Use this attribute to determine whether buffer pool assignment is efficient. If the
pool hit ratio is low, increasing the number of buffer pool pages improve
performance.

pool index from estore Number of buffer pool index pages copied from extended
storage. The value format is integer.

Required index pages are copied from extended storage to the buffer pool. The
copy process incur the cost of connecting to the shared memory segment, but it
saves the cost of a disk read.

pool index l reads The number of logical read requests for index pages that have
gone through the buffer pool. The value format is integer. This count includes
accesses to index pages that are:
v Already in the buffer pool when the database manager needs to process the
page
v Read into the buffer pool before the database manager can process the page
In conjunction with Pool Index Physical Reads, you can calculate the index page
hit ratio for the buffer pool as follows:
1 - (buffer pool index physical reads / buffer pool index logical reads)

To calculate the overall buffer pool hit ratio, see Pool Data Logical Reads. If the hit
ratio is low, increasing the number of buffer pool pages improve performance.

Chapter 5. Attributes reference 59

 pool index p reads The number of physical read requests to get index pages into
the buffer pool. The value format is integer. See Pool Index Logical Reads for
information about how to use this element.

pool index to estore Number of buffer pool index pages copied to extended
storage. The value format is integer. Pages are copied from the buffer pool to
extended storage when they are selected as victim pages. As a result of the
copying process, there is sufficient space for new pages in the buffer pool.

pool index writes The number of times a buffer pool index page was physically
written to disk. The value format is integer. If a buffer pool index page is written
to disk for a high percentage of the Pool Index Physical Reads, performance
improve by increasing the number of buffer pool pages available for the database.
If all applications are updating the database, increasing the size of the buffer pool
have minimal impact on performance; most pages will contain updated data that
must be written to disk.

pool read time The total amount of elapsed time spent processing read requests
that caused data or index pages to be physically read from disk to buffer pool. The
value format is integer.

Use this attribute with Buffer Pool Data Physical Reads and Buffer Pool Index
Physical Reads attributes to calculate the average page-read time. This average is
important because it indicate the presence of an I/O wait, which in turn indicate
that you need to move data to a different device.

pool total reads The total number of read requests that required I/O to get data
pages and index pages into the buffer pool. The value format is integer. This
attribute is the total of Pool Data Logical Reads, Pool Index Logical Reads, Pool
Data Physical Reads, and Pool Index Physical Reads attributes. Use this attribute to
determine how busy the DB2 server is in terms of I/O activity.

pool total writes The total number of write requests. The value format is integer.
This attribute is the total of Pool Data Writes and Pool Index Writes attributes. Use
this attribute to determine how busy the DB2 server is in terms of write I/O
activity.

pool write time The total amount of time spent physically writing data or index
pages from the buffer pool to disk. The value format is integer. Use this attribute
with Buffer Pool Data Writes and Buffer Pool Index Writes attributes to calculate
the average page-write time. This average is important because it indicate the
presence of an I/O wait, which in turn indicate that you need to move data to a
different device.

prefetch wait time The time an application spent waiting for an I/O server
(prefetcher) to finish loading pages into the buffer pool. The value format is
integer. This attribute can be used to experiment with changing the number of I/O
servers and the I/O server sizes.

prev uow stop time The date and time that the unit of work completed.

query card estimate An estimate of the number of rows that will be returned by a
query. The value format is integer. You can compare this estimate by the SQL
compiler with the run time actuals.

60 IBM Tivoli Monitoring for Databases: DB2 Agent: User’s Guide

query cost estimate Estimated cost, in timerons, for a query, as determined by the
SQL compiler. The value format is integer. This attribute allows correlation of
actual run time with the compile-time estimates.

rej curs blk The number of times that a request for an I/O block at the server was
rejected and the request was converted to non-blocked I/O. If there are many
cursors blocking data, the communication heap become full. The value format is
integer. When this heap is full, I/O blocks are not allocated for blocking cursors;
however, an error condition does not alert you to this condition. If cursors are
unable to block data, performance can be affected adversely.

rollback sql stmts The total number of SQL ROLLBACK statements that have been
attempted. The value format is integer. A rollback can result from an application
request, a deadlock, or an error situation. This attribute counts only the number of
rollback statements issued from applications.

At the application level, this attribute can help you determine the level of database
activity for the application and the amount of conflict with other applications.

At the database level, it can help you determine the amount of activity in the
database and the amount of conflict between applications on the database.

rows deleted The number of row deletions attempted. The value format is integer.
Use this attribute to gain insight into the current level of activity within the
database manager.

rows inserted The number of row insertions attempted. The value format is
integer. Use this attribute to gain insight into the current level of activity within
the database manager.

rows read The number of rows read from the table. The value format is integer.
This attribute helps to identify tables with heavy usage for which you want to
create additional indexes.

rows selected The number of rows that have been selected and returned to the
application. The value format is integer. Use this attribute to gain insight into the
current level of activity within the database manager.

rows updated The number of row updates attempted. The value format is integer.
Use this attribute to gain insight into the current level of activity within the
database manager.

rows written The number of rows changed (inserted, deleted or updated) in the
table. The value format is integer. A high value for table-level information indicates
heavy usage of the table. If so, you want to use the Run Statistics (RUNSTATS)
utility to maintain efficiency of the packages used for this table.

section number The internal section number in the package for the SQL statement
currently processing or most recently processed. The value format is integer.

select sql stmts The number of SQL SELECT statements that were executed. The
value format is integer.

Use this attribute to determine the level of database activity at the application or
database level. You can also use the following formula to determine the ratio of
SELECT statements to the total statements by:

Chapter 5. Attributes reference 61

 1. Adding together the number of static SQL statements attempted and dynamic
SQL statements attempted
2. Dividing the resulting total by the number of select SQL statements executed

snapshot time The date and time when the database system monitor information
was collected.

sort overflows The total number of sorts that ran out of sort heap and have
required disk space for temporary storage. The value format is integer. At a
database or application level, use this element in conjunction with Total Sorts
attribute. This attribute can help to determine the source of contention for
resources.

sort overflows pct The percentage of sorts that ran out of sort heap and have
required disk space for temporary storage. The value format is integer. This
percentage is calculated by dividing the value of Sort Overflows attribute by the
value of the Total Sorts attribute. At a database or application level, use this
attribute to evaluate the percentage of sorts that required overflow to disk. If this
percentage is high, you want adjust the database configuration by increasing the
value of sortheap.

static sql stmts The number of static SQL statements that were attempted. The
value format is integer.

Use this attribute to calculate the total number of successful SQL statements at the
database or application level by:
1. Adding together the number of Dynamic SQL Statements Attempted and the
Static SQL Statements Attempted
2. Subtracting the number of Failed Statement Operations
The remainder equals the throughput (the number of successful SQL statements)
during the current monitoring period.

stmt operation The statement operation currently being processed or most recently
processed (if none is currently running). The value format is a text string with a
maximum of 20 characters. The valid values are
CLOSE
DESCRIBE
EXECUTE
EXECUTE IMMEDIATE
FETCH
OPEN
PREPARE
STATIC COMMIT
STATIC ROLLBACK
UNKNOWN

Use this attribute to determine the operation that is executing or recently finished.

stmt start The date and time that the most recent SQL statement operation started.

stmt stop The date and time that the most recent SQL statement operation
stopped. The value format is a text string with a maximum of 16 characters. If the
statement is still running, this field is 0 (zero). Use this attribute with the Statement
Start attribute to calculate the elapsed execution time for the statement operation.

62 IBM Tivoli Monitoring for Databases: DB2 Agent: User’s Guide

stmt text (Unicode) This is the text of the dynamic SQL statement. For application
snapshots, the statement text helps you identify what the application was
executing when the snapshot was taken, or most recently processed if no statement
was being processed at the time the snapshot was taken. For dynamic SQL
statements, this attribute identifies the SQL text associated with a package. The
value format is text string with a maximum of 2000 bytes. This attribute is
globalized.

stmt type The type of SQL statement processed. The value format is a text string
with a maximum of 32 characters. The valid values are:
DYNAMIC
NON-STATEMENT OPERATION
STATIC
UNKNOWN STMT TYPE

Use this attribute to determine the type of statement that is executing.

table name (Unicode) The name of the table. The value format is a text string with
a maximum of 60 bytes. This attribute is globalized. Along with Table Schema
attribute, this attribute can help to determine the source of contention for
resources.

table schema (Unicode) The schema of the table. The value format is a text string
with a maximum of 60 bytes. This attribute is globalized. Along with Table Name
attribute, this attribute can help to determine the source of contention for
resources.

tablespace name (Unicode) The name of a tablespace that the application is

waiting to lock. The value format is a text string with a maximum of 60 bytes. This
attribute is globalized. It can help you to determine the source of contention for
resources.

total hash joins The total number of hash joins executed. The value format is
integer.

At the database or application level, use this value in conjunction with Hash Join
Overflows attribute and Hash Join Small Overflows attribute to determine if a
significant percentage of hash joins would benefit from modest increases in the sort
heap size.

total hash loops The total number of times that a single partition of a hash join
was larger than the available sort heap space. The value format is integer. Values
for this attribute indicate inefficient execution of hash joins. This might indicate
that the sort heap size is too small or the sort heap threshold is too small. Use this
value in conjunction with the other hash join variables to tune the sort heap size
(sortheap) and sort heap threshold (sheapthres) configuration parameters.

total sort time The total elapsed time (in milliseconds) for all sorts that have been
executed. The value format is integer. At a database or application level, use this
element with Total Sorts attribute to calculate the average sort time. This average
can indicate whether sorting is a performance concern.

total sorts The total number of sorts that have been executed. The value format is
integer. At a database or application level, use this value with the Sort Overflows
attribute to calculate the percentage of sorts that need more heap space. You can
also use it with Total Sort Time attribute to calculate the average sort time. If the

Chapter 5. Attributes reference 63

 number of sort overflows is small with respect to the total sorts, increasing the sort
heap size have little impact on performance, unless this buffer size is increased
substantially.

total sql stmt The number of SQL statements that have been attempted since the
latter of: application start up, database activation, or last reset. The value format is
integer. Use this attribute to measure database activity at the database or
application level. To calculate the SQL statement throughput for a given period,
divide this attribute by the elapsed time between two snapshots.

uid sql stmts The number of SQL UPDATE, INSERT, and DELETE statements that
were executed. The value format is integer.

Use this attribute to determine the level of database activity at the application or
database level. You can also use the following formula to determine the ratio of
UPDATE, INSERT and DELETE statements to the total number of statements:
1. Adding together the number of static SQL statements attempted and the
dynamic SQL statements attempted
2. Dividing the number of UPDATE/INSERT/DELETE SQL statements executed
by the sum derived in step 1

uow comp status The completion status of the previous UOW. The value format is
a text string with a maximum of 32 characters. Use this attribute to determine if
the unit of work ended because of a deadlock or abnormal termination.

Valid values are:

UOW Commit
UOW RB - Lock Timeout
UOW RB due to Abend
UOW RB due to Deadlock
UOW Rolled Back

uow lock wait time The total amount of elapsed time this unit of work has spent
waiting for locks. The value format is integer. This attribute can help you to
determine the severity of the resource contention problem.

uow log space used The amount of log space (in bytes) used in the current unit of
work of the monitored application. The value format is integer. Use this attribute
to understand the logging requirements at the unit-of-work level.

uow start time The date and time that the unit of work first required database
resources. This resource requirement occurs at the first SQL statement execution for
the unit of work.

uow stop time The date and time that the most recent unit of work completed,
which occurs when database changes are committed or rolled back.

x lock escals The number of times that locks have been escalated from several row
locks to one exclusive table lock, or the number of times an exclusive lock on a
row caused the table lock to become an exclusive lock. The value format is integer.
A lock is escalated when the total number of locks held by an application reaches
the maximum amount of lock list space available to the application. The amount of
lock list space available is determined by the LOCKLIST and MAXLOCKS
configuration parameters. Other applications cannot access data held by an
exclusive lock.

64 IBM Tivoli Monitoring for Databases: DB2 Agent: User’s Guide

 Because exclusive locks can impact the concurrency of your data, it is important to
track them. When an application reaches the maximum number of locks allowed
and there are no more locks to escalate, it uses space in the lock list allocated for
other applications. When the entire lock list is full, an error occurs.

Application attributes (group 01)

Use these attributes to obtain information about the applications running in the
DB2 environment. With this information, you can determine the efficiency of the
applications and identify any problem areas for corrective action.

Unless otherwise noted, integer values are 32-bit or 0 to 4 294 967 296.

agents stolen The number of times that agents are stolen from an application. The
value format is integer. When another application requires a new subagent and has
no subagents in its associated agent pool, it steals subagents from the agent pools
of other applications.

If the number of agents stolen from this application is high compared to normal
operating levels, the number of pool agents might be too low. When the agent pool
size is too small, one application might fill the pool with associated subagents.
When another application requires a new subagent and has no subagents in its
associated agent pool, it steals subagents from the agent pools of other
applications.

app work load The ratio of the maximum number of subagents associated with
this application to the number of agents that are stolen from the application by
DB2 to work on a different application. The value format is integer.

Use the returned value to evaluate the load that this application places on the
system. An agent working for an application is associated with that application.
After the agent completes the work for the application, it is placed in the agent
pool as an idle agent, but it remains associated with the application. When the
application requires an agent again, DB2 searches the agent pool for an agent
already associated with the application and assigns work to the associated agent.

appl name The name of the application running at the client as known to the
database manager or DB2 Connect. The value format is a text string, with a
maximum of 20 characters. Example: *Local.db2inst1.990212202018.

appl section inserts The number of inserts of SQL sections by an application from
its SQL work area. The value format is integer. The working copy of any
executable section is stored in a unique SQL work area. The returned value is a
count of how many times a copy was not available and had to be inserted.

associated agents top The maximum number of associated agents. The value
format is integer.

avg sect read per direct read The average number of sectors that are read in a
direct read by this application. The value format is integer. Direct reads result in
poor performance because the data is physically read from disk each time instead
of being read from the buffer pool. If you are using system monitors to track input
and output for the device, this value helps you distinguish database input and
output from non-database input and output.

Chapter 5. Attributes reference 65

 avg sect written per direct read The average number of sectors that are written in
a direct read by this application. The value format is integer.

binds precompiles The number of binds and precompiles attempted by an

application. The value format is integer. Use the returned value to determine the
current level of activity in the database manager.

db name The real name of the database for which information is collected or to
which the application is connected. This is the name the database was given when
created. The value format is a simple text string with a maximum of 20 characters.
Use this attribute to identify the specific database to which the data applies.

db name (Unicode) The real name of the database for which information is
collected or to which the application is connected. This is the name the database
was given when created. The value format is a simple text string with a maximum
of 60 bytes. This attribute is globalized. Use this attribute to identify the specific
database to which the data applies.

ddl sql pct for int The percentage of total SQL statements that were SQL DDL
statements executed by the application during the monitoring interval. The value
format is integer.

Because of the high activity in the system catalog tables, try to keep DDL
statement activity to a minimum. If the returned value is high compared to normal
operating levels, determine the activity causing it to be high and restrict it from
being performed. Examples of DDL statements are CREATE TABLE, CREATE
VIEW, ALTER TABLE, and DROP INDEX. You can also use the returned value to
refine the package cache hit ratio for this application. DDL statements can also
impact the package cache by invalidating sections that are stored there and causing
additional system overhead because of section recompilation.

deadlocks for int The total number of deadlocks that occurred for the application
during the monitoring interval. The value format is integer. Use the returned value
to determine if the application is experiencing contention problems. Modify the
application to better enable it to execute concurrently.

lock escalation for int The total number of lock escalations for the application
during the monitoring interval. The value format is integer. Exclusive lock
escalations are included in this number. Use the returned value to help you
evaluate the settings of the LOCKLIST and MAXLOCKS configuration parameters.
Lock escalations can result in a decrease in concurrency between applications
connected to a database.

node name Indicates the originating name. Value format is a simple text string
with a maximum of 32 characters.

pkg cache hit pct The application’s package cache hit ratio (as a percentage) for
the last monitoring interval. The value format is integer. The package cache hit
ratio is the ratio of the difference between the package cache lookups and the
package cache inserts to all package cache lookups.

This percentage tells you whether the package cache is being used efficiently by
this application. If the hit ratio is high (greater than 80%), the package cache is
performing well. A smaller percentage can indicate that the package cache needs to
be increased. However, it is not always worthwhile to increase the size of the

66 IBM Tivoli Monitoring for Databases: DB2 Agent: User’s Guide

package cache for an application that executes only once a day. The size of the
package cache is set by the pckcachesz configuration parameter.

pool index hit ratio pct for int The application’s index page hit ratio (as a
percentage) for the buffer pool during the monitoring interval. The value format is
integer.

The index page hit ratio for the buffer pool is the ratio of the number of accesses
to index pages already in the buffer pool to the total number of read requests for
index pages that use the buffer pool. The total number of read requests for index
pages includes both those index pages that must be loaded into the buffer pool
before the request can be fulfilled and those index pages that are already in the
buffer pool.

If the hit ratio is low compared to normal operating levels, increasing the number
of buffer pool pages can improve performance.

Typically, you need to analyze buffer pool usage at the database level because the
buffer pool size is configured at the database level. However, if the buffer pool hit
ratio is low for an application, it can indicate the reason that application has poorer
performance when compared to other applications connected to the same database.

snapshot time The date and time when the database system monitor information
was collected.

sql reqs since commit The number of SQL requests that were submitted by the
application since the last commit. Use the returned value to monitor the progress
of a transaction. The value format is integer.

stmts sorts The total number of times that a set of data was sorted to process the
OPEN operation of the current SQL statement. The value format is integer.

Use the returned value to help identify the need for an index, because indexes can
reduce the need for sorting a set of data. Identify the SQL statement for which this
returned value is providing sort information. Then, analyze this SQL statement to
determine index candidates by looking at columns that are being sorted. For
example, a column used in an ORDER BY clause can be an index candidate.

tot pool io time The total time (in seconds) that an application spent performing
buffer pool input and output operations (reading or writing pages). The value
format is integer. The returned value is an indication of how much time the
application performs input and output operations using the buffer pool.

uid sql pct for int The percentage of total SQL statements that were SQL UPDATE,
INSERT, and DELETE statements executed by the application during the
monitoring interval. The value format is integer.

Use the returned value to determine if the application performs frequent updates.
If the returned value is low compared to normal operating levels, the application is
query-based; otherwise, it is update-based. Knowing what type of applications you
have (query-based or update-based) can aid you in refining the database
configuration parameters.

Chapter 5. Attributes reference 67

 Buffer Pool Data attributes
The Buffer Pool attributes provide information about buffer pool activities. You can
use this information to monitor the performance of your buffer pools and to
identify problem areas for corrective action.

All values are integers calculated from the first application connection, unless
otherwise noted. Integer values are 32-bit or 0 to 4 294 967 296.

avg data page read per async req The average number of pages read for each
asynchronous request. The value format is integer. This value is derived by
dividing the value of Pool Async Data Reads attribute by the value of Pool Async
Data Read Reqs attribute.

avg direct read time The average elapsed time for a direct read request. The value
format is integer. This value is derived by dividing the value of Direct Read Time
attribute by the value of the Direct Reads attribute. This average is important
because it indicate the presence of an I/O wait, which in turn indicate that you
need to move data to a different device.

avg direct write time The average elapsed time for a direct write request. The
value format is integer. This value is derived by dividing the value of the Direct
Write Time attribute by the value of the Direct Writes attribute. This average is
important because it indicate the presence of an I/O wait, which in turn indicate
that you need to move data to a different device.

avg pool read time The average elapsed time for a read request. The value format
is integer. This value is derived by dividing the value of Pool Read Time attribute
by the value of Pool Total Reads attribute. This average is important because it
indicate the presence of an I/O wait, which in turn indicate that you need to move
data to a different device.

avg pool write time The average elapsed time for a write request. The value
format is integer. This value is derived by dividing the value of Pool Write Time
attribute by the value of Pool Total Writes attribute.

avg sync read time The average elapsed time used to perform a synchronous read.
The value format is integer. This value is derived by dividing the value of Pool
Sync Read Time attribute by the value of Pool Sync Read attribute. This average is
important because it indicate the presence of an I/O wait, which in turn indicate
that you need to move data to a different device.

avg sync write time The average elapsed time used to perform a synchronous
write. The value format is integer. This value is derived by dividing the value of
Pool Sync Write Time attribute by the value of Pool Sync Write attribute. This
average is important because it indicate the presence of an I/O wait, which in turn
indicate that you need to move data to a different device.

bp id The internal identifier for the buffer pool. The value format is a text string
with a maximum of 4 characters. This attribute is for IBM internal use only.

bp name The name of the buffer pool. The value format is a text string with a
maximum of 20 characters. A new database has a default buffer pool (named
IBMDEFAULTBP). The size of the default buffer pool is determined by the
platform. Depending on your needs you choose to create several buffer pools, each

68 IBM Tivoli Monitoring for Databases: DB2 Agent: User’s Guide

of a different size, for a single database. The CREATE, ALTER, and DROP
BUFFERPOOL statements allow you to create, change, or remove a buffer pool.

bp name (Unicode) The name of the buffer pool. The value format is a text string
with a maximum of 60 bytes. This attribute is globalized. A new database has a
default buffer pool (named IBMDEFAULTBP). The size of the default buffer pool is
determined by the platform. Depending on your needs you choose to create
several buffer pools, each of a different size, for a single database. The CREATE,
ALTER, and DROP BUFFERPOOL statements allow you to create, change, or
remove a buffer pool.

db name The real name of the database for which information is collected or to
which the application is connected. This is the name the database was given when
created. The value format is a simple text string with a maximum of 20 characters.
Use this attribute to identify the specific database to which the data applies.

db name (Unicode) The real name of the database for which information is
collected or to which the application is connected. This is the name the database
was given when created. The value format is a simple text string with a maximum
of 60 bytes. This attribute is globalized. Use this attribute to identify the specific
database to which the data applies.

db path The full path of the location where the database is stored on the
monitored system. The value format is a text string with a maximum of 256
characters. Use this attribute with the Database Name attribute to identify the
specific database to which the data applies.

db path (Unicode) The full path of the location where the database is stored on the
monitored system. The value format is a text string with a maximum of 768 bytes.
This attribute is globalized. Use this attribute with the Database Name attribute to
identify the specific database to which the data applies.

direct read reqs The number of requests to perform a direct read of one or more
sectors of data. The value format is integer. Use the following formula to calculate
the average number of sectors that are read by a direct read:
direct reads from database / direct read requests

direct read time The elapsed time (in milliseconds) required to perform the direct
reads. The value format is integer. Use the following formula to calculate the
average direct read time per sector:
direct read time / direct reads from database

A high average time indicate an I/O conflict.

direct reads The number of read operations that do not use the buffer pool. The
value format is integer. Use the following formula to calculate the average number
of sectors that are read by a direct read:
direct reads from database / direct read requests

When using system monitors to track I/O, this data attribute helps to distinguish
database I/O from non-database I/O on the device.

direct write reqs The number of requests to perform a direct write of one or more
sectors of data. The value format is integer. Use the following formula to calculate
the average number of sectors that are written by a direct write:

Chapter 5. Attributes reference 69

 direct writes to database / direct write requests

direct write time The elapsed time (in milliseconds) required to perform the direct
writes. The value format is integer. Use the following formula to calculate the
average direct write time per sector:

direct write time / direct writes to database

A high average time indicate an I/O conflict.

direct writes The number of write operations that do not use the buffer pool. The
value format is integer. Use the following formula to calculate the average number
of sectors that are written by a direct write.

direct writes to database / direct write requests

When using system monitors to track I/O, this data attribute helps to distinguish
database I/O from non-database I/O on the device.

files closed The total number of database files closed. The value format is integer.
The database manager opens files for reading and writing into and out of the
buffer pool. The maximum number of database files open by an application at any
time is controlled by the MAXFILOP configuration parameter. If the maximum is
reached, one file will be closed before the new file is opened. Note that the actual
number of files opened not equal the number of files closed.

input db alias The alias of the database provided when calling the snapshot
function. The value format is a simple text string with a maximum of 20
characters.

Use this attribute to help you identify the specific database to which the monitor
data applies. It contains blanks unless you requested monitor information related
to a specific database.

input db alias (Unicode) The alias of the database provided when calling the
snapshot function. The value format is a simple text string with a maximum of 60
bytes. This attribute is globalized.

Use this attribute to help you identify the specific database to which the monitor
data applies. It contains blanks unless you requested monitor information related
to a specific database.

node name Indicates the originating name. Value format is a simple text string
with a maximum of 32 characters.

pool async data read reqs The number of asynchronous read requests. The value
format is integer. To calculate the average number of data pages read per
asynchronous request, use the following formula:
buffer pool asynchronous data reads / buffer pool asynchronous read requests

This average can help to determine the amount of asynchronous I/O done in each
interaction with the prefetcher.

pool async data reads The number of pages read asynchronously into the buffer
pool. The value format is integer.

70 IBM Tivoli Monitoring for Databases: DB2 Agent: User’s Guide

Use this attribute with Buffer Pool Data Physical Reads attribute to calculate the
number of physical reads that were performed synchronously (that is, physical
data page reads that were performed by database manager agents). Use the
following formula:
buffer pool data physical reads - buffer pool asynchronous data reads

By comparing the ratio of asynchronous to synchronous reads, you can gain

insight into how well the prefetchers are working.

pool async data writes The number of times a buffer pool data page was
physically written to disk by an asynchronous page cleaner or by a prefetcher. The
value format is integer. A prefetcher have written dirty pages to disk to make
space for the pages being prefetched.

Use this attribute with Buffer Pool Data Writes to calculate the number of physical
write requests that were performed synchronously (that is, physical data page
writes that were performed by database manager agents). Use the following
formula:
buffer pool data writes - buffer pool asynchronous data writes

pool async index reads The number of index pages read asynchronously into the
buffer pool by a prefetcher. The value format is integer. Asynchronous reads are
performed by database manager prefetchers.

Use this attribute with Buffer Pool Index Physical Reads to calculate the number of
physical reads that were performed synchronously (that is, physical index page
reads that were performed by database manager agents). Use the following
formula:
buffer pool index physical reads - buffer pool asynchronous index reads

By comparing the ratio of asynchronous to synchronous reads, you can gain

insight into how well the prefetchers are working.

pool async index writes The number of times a buffer pool index page was
physically written to disk by an asynchronous page cleaner or a prefetcher. The
value format is integer. A prefetcher have written dirty pages to disk to make
space for the pages being prefetched.

Use this attribute with Buffer Pool Index Writes to calculate the number of physical
index write requests that were performed synchronously. That is, physical index
page writes that were performed by database manager agents. Use the following
formula:
buffer pool index writes - buffer pool asynchronous index writes

By comparing the ratio of asynchronous to synchronous writes, you can gain

insight into how well the buffer pool page cleaners are performing.

pool async read time The total elapsed time spent reading by database manager
prefetchers. The value format is integer. Use this attribute to calculate the elapsed
time for synchronous reading, using the following formula:
total buffer pool physical read time - buffer pool synchronous read time

You can also use this attribute to calculate the average asynchronous read time
using the following formula:
buffer pool asynchronous read time / buffer pool asynchronous data reads

Chapter 5. Attributes reference 71

 These calculations can be used to understand the I/O work being performed.

pool async write time The total elapsed time spent writing data or index pages
from the buffer pool to disk by database manager page cleaners. The value format
is integer. Calculate the elapsed time spent writing pages synchronously by
subtracting the value of Pool Async Write Time attribute from the value of Pool
Physical Write Time attribute. You can also use this attribute to calculate the
average asynchronous read time by:
1. Adding together the value of Pool Async Data Writes and the value of Pool
Async Index Writes.
2. Dividing the value of Pool Async Write Time by the sum from step 1.
These calculations can be used to understand the I/O work being performed.

pool data from estore Number of buffer pool data pages copied from extended
storage. The value format is integer. Required pages are copied from extended
storage to the buffer pool. The copy process incur the cost of connecting to the
shared memory segment, but it saves the cost of a disk read.

pool data l reads The number of logical read requests for data pages that have
gone through the buffer pool. The value format is integer. This count includes
accesses to data that is:
v Already in the buffer pool when the database manager needs to process the
page
v Read into the buffer pool before the database manager can process the page.
In conjunction with Pool Data Physical Reads, you can calculate the data page hit
ratio for the buffer pool as follows:
1 - (buffer pool data physical reads / buffer pool data logical reads)

In conjunction with Pool Data Physical Reads, Pool Index Physical Reads, and Pool
Index Logical Reads, you can calculate the overall buffer pool hit ratio as follows:
1 - ((buffer pool data physical reads + buffer pool index physical reads) /
(buffer pool data logical reads + buffer pool index logical reads))

Increasing buffer pool size will generally improve the hit ratio until you reach a
point of diminishing return.

pool data p reads The number of read requests that required I/O to get data pages
into the buffer pool. The value format is integer. See Pool Data Logical Reads and
Pool Async Data Reads attributes for information about how to use this attribute.

pool data to estore Number of buffer pool data pages copied to extended storage.
The value format is integer.

Pages are copied from the buffer pool to extended storage when they are selected
as victim pages. As a result of the copying process, there is sufficient space for new
pages in the buffer pool.

pool data writes The number of times a buffer pool data page was physically
written to disk. The value format is integer. A buffer pool data page is written to
disk for the following reasons:
v To free a page in the buffer pool so another page can be read
v To flush the buffer pool.

72 IBM Tivoli Monitoring for Databases: DB2 Agent: User’s Guide

If a buffer pool data page is written to disk for a high percentage of the Buffer
Pool Data Physical Reads, performance improve by increasing the number of
buffer pool pages available for the database.

pool hit ratio The buffer pool hit ratio. The value format is integer. The sum of
Pool Data Logical Reads and Pool Index Logical Reads attributes is divided by the
value of Pool Total Reads attribute to derive the pool hit ratio. This attribute can
determine whether buffer pool assignment is efficient. If the pool hit ratio is low,
increasing the number of buffer pool pages improve performance.

pool index from estore Number of buffer pool index pages copied from extended
storage. The value format is integer. Required index pages are copied from
extended storage to the buffer pool. The copy process incur the cost of connecting
to the shared memory segment, but it saves the cost of a disk read.

pool index l reads The number of logical read requests for index pages that have
gone through the buffer pool. The value format is integer. This count includes
accesses to index pages that are:
v Already in the buffer pool when the database manager needs to process the
page
v Read into the buffer pool before the database manager can process the page

In conjunction with Buffer Pool Index Physical Reads attribute, you can calculate
the index page hit ratio for the buffer pool as follows:
1 - (buffer pool index physical reads / buffer pool index logical reads)

To calculate the overall buffer pool hit ratio, see Buffer Pool Data Logical Reads. If
the hit ratio is low, increasing the number of buffer pool pages improve
performance.

pool index p reads The number of physical read requests to get index pages into
the buffer pool. The value format is integer. See Pool Index Logical Reads for
information about how to use this element.

pool index to estore Number of buffer pool index pages copied to extended
storage. The value format is integer. Pages are copied from the buffer pool to
extended storage when they are selected as victim pages. As a result of the
copying process, there is sufficient space for new pages in the buffer pool.

pool index writes The number of times a buffer pool index page was physically
written to disk. The value format is integer. If a buffer pool index page is written
to disk for a high percentage of the Buffer Pool Index Physical Reads, performance
improve by increasing the number of buffer pool pages available for the database.

pool read time The total amount of elapsed time spent processing read requests
that caused data or index pages to be physically read from buffer pool to disk. The
value format is integer.

Use this attribute with Buffer Pool Data Physical Reads and Buffer Pool Index
Physical Reads attributes to calculate the average page-read time. This average is
important because it indicate the presence of an I/O wait, which in turn indicate
that you need to move data to a different device.

Chapter 5. Attributes reference 73

 pool sync data reads The number of physical data page reads that were performed
by database manager agents. The value format is integer. This value is derived by
subtracting the value of Pool Async Data Reads attribute from the Pool Data
Physical Reads attribute.

By comparing the ratio of asynchronous to synchronous reads, you can gain

insight into how well the prefetchers are working.

pool sync data writes The total number of physical write requests that were
performed synchronously (that is, physical data page writes that were performed
by database manager agents). The value format is integer. This value is derived by
subtracting the value of Pool Async Data Writes attribute from the value of Pool
Data Writes attribute.

By comparing the ratio of asynchronous to synchronous writes, you can gain

insight into how well the buffer pool page cleaners are performing.

pool sync index reads The number of index pages read synchronously (that is,
physical index page reads that were performed by database manager agents) into
the buffer pool. The value format is integer. This value is derived by subtracting
the value of Pool Async Index Reads attribute from Pool Index Physical Reads
attribute.

By comparing the ratio of asynchronous to synchronous reads, you can gain

insight into how well the prefetchers are working.

pool sync index writes The number of physical index write requests that were
performed synchronously (that is, physical index page writes that were performed
by database manager agents). The value format is integer. This value is derived by
subtracting the value of Pool Async Index Writes attribute from the value of Pool
Index Writes attribute.

By comparing the ratio of asynchronous to synchronous writes, you can gain

insight into how well the buffer pool page cleaners are performing.

pool sync read The total number of synchronous reads. The value format is
integer. This value is derived by adding together the values of the Pool Sync Data
Reads and Pool Sync Index Reads attributes.

pool sync read time The elapsed time used to perform all synchronous reads. The
value format is integer. This value is derived by subtracting the value of Pool
Async Read Time attribute from the value of Pool Read Time attribute. Use this
attribute to understand the I/O work being performed.

pool sync write The total number of synchronous index writes. The value format
is integer. The value is derived by adding together the values of Pool Sync Data
Writes attribute and Pool Sync Index Writes attribute.

pool sync write time The total elapsed time used to perform all synchronous
writes. The value format is integer. This value is derived by subtracting the value
of Pool Async Write Time attribute from the value of Pool Write Time attribute.

pool total reads The total number of read requests that required I/O to get data
pages and index pages into the buffer pool. The value format is integer. This
attribute is the total of Pool Data Logical Reads, Pool Index Logical Reads, Pool
Data Physical Reads, and Pool Index Physical Reads attributes.

74 IBM Tivoli Monitoring for Databases: DB2 Agent: User’s Guide

 pool total writes The total number of write requests. The value format is integer.
This attribute is the total of Pool Data Writes and Pool Index Writes attributes.

pool write time The total amount of time spent physically writing data or index
pages from the buffer pool to disk. The value format is integer. Use this attribute
with Buffer Pool Data Writes and Buffer Pool Index Writes attributes to calculate
the average page-write time. This average is important because it indicate the
presence of an I/O wait, which in turn indicate that you need to move data to a
different device.

Database Information attributes (group 00)

Use these attributes to obtain information about database activities. With this
information, you can determine the efficiency of the databases and identify any
problem areas for corrective action.

All values are integers calculated from the first database connection, unless
otherwise noted. Integer values are 32-bit or 0 to 4 294 967 296.

active sorts The number of sorts in the database that currently have a sort heap
allocated. The value format is integer.

Use this value in conjunction with Sort Heap Allocated attribute to determine the
average sort heap space used by each sort. If the SORTHEAP configuration
parameter is substantially larger than the average sort heap used, you be able to
lower the value of this parameter.

agents top The maximum number of agents (at one time) associated with
applications that are connected to the monitored database. The value format is
integer. Use this attribute to indicate how well the intra-query parallelism was
realized.

appl section inserts The number of inserts of SQL sections by an application from
its SQL work area. The value format is integer. The working copy of any
executable section is stored in a unique SQL work area. This value represents the
number of times when a copy was not available and therefore was inserted.

appl section lookups The number of lookups of SQL sections by an application

from its SQL work area. The value format is integer. This counter indicates how
many times the SQL work area was accessed by agents for an application. It is a
cumulative total of all lookups on all SQL work heaps for agents working for this
application.

appls cur cons The number of applications currently connected to the monitored
database. The value format is integer. Use this attribute to help you understand the
level of activity within a database and the amount of system resource being used.

avg data page read per async req The average number of pages read for each
asynchronous request. The value format is integer. This value is derived by
dividing the value of Pool Async Data Reads attribute by the value of Pool Async
Data Read Reqs attribute. Use this attribute to determine whether good enough
data pages were read per asynchronous request.

avg lock wait time The average elapsed time spent waiting for a lock. The value
format is integer. If the average lock wait time is high, look for applications that
hold many locks, or have lock escalations, with a focus on tuning your applications

Chapter 5. Attributes reference 75

 to improve concurrency, if appropriate. If escalations are the reason for a high
average lock wait time, the values of one or both of the LOCKLIST and
MAXLOCKS configuration parameters be too low.

avg pool read time The average elapsed time for a read request. The value format
is integer. This value is derived by dividing the value of Pool Read Time attribute
by the value of Pool Total Reads attribute. This average is important because it
indicate the presence of an I/O wait, which in turn indicate that you need to move
data to a different device.

avg pool write time The average elapsed time for a write request. The value
format is integer. This value is derived by dividing the value of Pool Write Time
attribute by the value of Pool Total Writes attribute.

avg sort time The average derived by dividing value of Total Sort Time attribute
by the value of the Total Sorts attribute. The value format is integer. The average is
expressed as elapsed time. At a database or application level, this attribute can
indicate whether sorting is a performance issue. Elapsed times are affected by
system load. The more processes you have running, the higher this elapsed time
value will be.

avg sync read time The average elapsed time used to perform a synchronous read.
The value format is integer. This value is derived by dividing the value of Pool
Sync Read Time attribute by the value of Pool Sync Read attribute. This average is
important because it indicate the presence of an I/O wait, which in turn indicate
that you need to move data to a different device.

avg sync write time The average elapsed time used to perform a synchronous
write. The value format is integer. This value is derived by dividing the value of
Pool Sync Write Time attribute by the value of Pool Sync Write attribute. This
average is important because it indicate the presence of an I/O wait, which in turn
indicate that you need to move data to a different device.

binds precompiles The number of binds and pre-compiles attempted. The value
format is integer. Use this attribute to gain insight into the current level of activity
within the database manager.

cat cache heap full The number of times that an insert into the catalog cache failed
because of a heap-full condition in the database heap. The value format is integer.
The catalog cache draws its storage dynamically from the database heap. Even if
the cache storage has not reached its limit, inserts into the catalog cache fail
because of a lack of space in the database heap. If the catalog cache heap full count
is not zero, you can correct the insert failure condition by increasing the database
heap size or by reducing the catalog cache size.

cat cache hit ratio The percentage of catalog sections found in cache. The value
format is integer. The value is calculated as the catalog found in cache divided by
the catalog cache lookups. This ratio indicates how well the catalog cache is
avoiding catalog accesses. If the ratio is high (more than 0.8), the cache is
performing well. A smaller ratio might indicate that you need to increase the size
of the catalog cache. Expect a large ratio immediately following the first connection
to the database.

cat cache inserts The number of times that the system tried to insert table
descriptor information into the catalog cache. The value format is integer. Table
descriptor information is usually inserted into the cache following a failed lookup

76 IBM Tivoli Monitoring for Databases: DB2 Agent: User’s Guide

to the catalog cache while processing a table, view, or alias reference in an SQL
statement. The catalog cache inserts value includes attempts to insert table
descriptor information that fail because of catalog cache overflow and heap full
conditions.

cat cache lookups The number of times that the catalog cache was referenced to
obtain table descriptor information. The value format is integer. This attribute
includes both successful and unsuccessful accesses to the catalog cache. This
attribute is used in calculating the catalog cache hit ratio. This ratio indicates how
well the catalog cache is avoiding catalog accesses. If the ratio is high (more than
0.8), the cache is performing well. A smaller ratio might indicate that you need to
increase the size of the catalog cache. Expect a large ratio immediately following
the first connection to the database.

cat cache overflows The number of times that an insert into the catalog cache
failed because the catalog cache was full. The value format is integer. If the catalog
cache overflows value is large, the catalog cache be too small for the workload.
Enlarging the catalog cache improve its performance. If the workload includes
transactions that compile a large number of SQL statements referencing many
tables, views, and aliases in a single unit of work, compiling fewer SQL statements
in a single transaction improve the performance of the catalog cache. Or if the
workload includes the binding of packages containing many SQL statements
referencing many tables, views or aliases, you want to split the packages so that
they include fewer SQL statements to improve performance.

catalog node name The network node name of the catalog node. The value format
is a simple text string with a maximum of 32 characters. Use this attribute to
determine the location of a database.

commit sql stmts The total number of SQL COMMIT statements that have been
attempted. The value format is integer. A small rate of change in this counter
during the monitor period indicate that applications are not doing frequent
commits. The lack of frequent commits can lead to problems with logging and data
concurrency. You can also use this attribute to calculate the total number of units of
work by calculating the sum of the following:
commit statements attempted
internal commits
rollback statements attempted
internal rollbacks

connections top The highest number of simultaneous connections to the database

since the database was activated. The value format is integer. You can calculate the
current number of connections at the time the snapshot was taken by adding
together the Remote Connections to Database Manager and Local Connections
attributes. Use this attribute to evaluate the setting of the MAXAPPLS
configuration parameter.

coord agents top The maximum number of coordinating agents working at one
time. The value format is integer. The MAXCAGENTS configuration parameter
determines the number of coordinating agents that can be executing concurrently.
If the peak number of coordinating agents results in a workload that is too high for
this node, you can reduce the MAXCAGENTS configuration parameter.

db conn time The date and time when the first database connection was made.

Chapter 5. Attributes reference 77

 db heap top This data attribute (now maintained for DB2 version compatibility)
measures memory usage, but not exclusively usage by the database heap. It
represents the high water mark for database heap (in bytes). The value format is
integer.

db location The location of the database in relation to the application. The value
format is a text string up to 8 characters. Valid values are:
LOCAL
REMOTE
UNKNOWN

Determine the relative location of the database server with respect to the
application taking the snapshot.

db name The real name of the database for which information is collected or to
which the application is connected. This is the name the database was given when
created. The value format is a simple text string with a maximum of 20 characters.
Use this attribute to identify the specific database to which the data applies.

db name (Unicode) The real name of the database for which information is
collected or to which the application is connected. This is the name the database
was given when created. The value format is a simple text string with a maximum
of 60 bytes. This attribute is globalized. Use this attribute to identify the specific
database to which the data applies.

db path The full path of the location where the database is stored on the
monitored system. The value format is a simple text string with a maximum of 256
characters. Use this attribute with the Database Name attribute to identify the
specific database to which the data applies.

db path (Unicode) The full path of the location where the database is stored on the
monitored system. The value format is a simple text string with a maximum of 768
bytes. This attribute is globalized. Use this attribute with the Database Name
attribute to identify the specific database to which the data applies.

dbase status The status of the database. The value format is a text string with a
maximum of 32 characters. The valid values are as follows:
Active The database is active.
Quiesce pending
The database is in quiesce-pending state. New connections to the database
are not permitted and new units of work cannot be started. Depending on
the quiesce request, active units of work are allowed to complete or roll
back immediately.
Quiesced
The database has been quiesced. New connections to the database are not
permitted and new units of work cannot be started.
Roll Forward
A roll forward is in progress on the database.
Unknown
The status is unknown.

ddl sql stmts The number of SQL Data Definition Language (DDL) statements that
were executed. The value format is integer.

78 IBM Tivoli Monitoring for Databases: DB2 Agent: User’s Guide

Use this attribute to determine the level of database activity at the application or
database level. DDL statements are expensive to run because of their impact on the
system catalog tables. As a result, if the value of this attribute is high, you need to
determine the cause and possibly restrict the identified activity from being
performed. You can also use this attribute to determine the percentage of DDL
activity using the following formula:
divide the number of DDL SQL statements by the total number of statements

deadlocks The total number of deadlocks that have occurred since the first
database connection. The value format is integer. This attribute can indicate that
applications are experiencing contention problems. To resolve the problem,
determine in which applications (or application processes) the deadlock are
occurring. You can then modify the application to enable it to execute concurrently.
Some applications, however, not be capable of running concurrently.

direct read reqs The number of requests to perform a direct read of one or more
sectors of data. The value format is integer. Use the following formula to calculate
the average number of sectors that are read by a direct read:
direct reads from database / direct read requests

direct read time The elapsed time (in milliseconds) required to perform the direct
reads. The value format is integer. Use the following formula to calculate the
average direct read time per sector:
direct read time / direct reads from database

A high average time indicate an I/O conflict.

direct reads The number of read operations that do not use the buffer pool. The
value format is integer. Use the following formula to calculate the average number
of sectors that are read by a direct read:
direct reads from database / direct read requests

When using system monitors to track I/O, this data attribute helps to distinguish
database I/O from non-database I/O on the device.

direct write reqs The number of requests to perform a direct write of one or more
sectors of data. The value format is integer. Use the following formula to calculate
the average number of sectors that are written by a direct write:
direct writes to database / direct write requests

direct write time The elapsed time (in milliseconds) required to perform the direct
writes. The value format is integer. Use the following formula to calculate the
average direct write time per sector:
direct write time / direct writes to database

A high average time indicate an I/O conflict.

direct writes The number of write operations that do not use the buffer pool. The
value format is integer. Use the following formula to calculate the average number
of sectors that are written by a direct write:
direct writes to database / direct write requests

When using system monitors to track I/O, this data attribute helps to distinguish
database I/O from non-database I/O on the device.

Chapter 5. Attributes reference 79

 dynamic sql stmts The number of dynamic SQL statements that were attempted.
The value format is integer. Use this attribute to calculate the total number of
successful SQL statements at the database or application level by:
1. Adding together the number of Dynamic SQL Statements Attempted and the
Static SQL Statements Attempted
2. Subtracting the number of Failed Statement Operations
The remainder equals the throughput (the number of successful SQL statements)
during the current monitoring period.

failed sql stmts The number of SQL statements that were attempted, but failed.
The value format is integer. This count includes all SQL statements that received a
negative SQLCODE. Use this attribute to calculate the total number of successful
SQL statements at the database or application level by:
1. Adding together the number of Dynamic SQL Statements Attempted and the
Static SQL Statements Attempted
2. Subtracting the number of Failed Statement Operations
The remainder equals the throughput (the number of successful SQL statements)
during the current monitoring period. This attribute can also help you to
determine the reasons for poor performance; failed statements indicate time wasted
by the database manager, which results in lower throughput for the database.

files closed The total number of database files closed. The value format is integer.
The database manager opens files for reading and writing into and out of the
buffer pool. The maximum number of database files open by an application at any
time is controlled by the MAXFILOP configuration parameter. If the maximum is
reached, one file will be closed before the new file is opened. Note that the actual
number of files opened not equal the number of files closed.

hash join overflows The number of times that hash join data exceeded the
available sort heap space. The value format is integer.

At the database level, if the percentage of Hash Join Small Overflows is greater
than 10% of this value, consider increasing the sort heap size. You can use values
at the application level to evaluate hash join performance for individual
applications.

hash join small overflows The number of times that hash join data exceeded the
available sort heap space by less than 10%. The value format is integer. If this
value and Hash Join Overflows attribute are high, consider increasing the sort
heap threshold. If this value is greater than 10% of Hash Join Overflows, consider
increasing the sort heap size.

input db alias The alias of the database provided when calling the snapshot
function. The value format is a simple text string with a maximum of 20
characters. Use this attribute to help you identify the specific database to which the
monitor data applies. It contains blanks unless you requested monitor information
related to a specific database.

Because internal automatic rebinds can have a significant impact on performance,

they need to be minimized where possible.

input db alias (Unicode) The alias of the database provided when calling the
snapshot function. The value format is a simple text string with a maximum of 60
bytes. This attribute is globalized. Use this attribute to help you identify the

80 IBM Tivoli Monitoring for Databases: DB2 Agent: User’s Guide

specific database to which the monitor data applies. It contains blanks unless you
requested monitor information related to a specific database.

Because internal automatic rebinds can have a significant impact on performance,

they need to be minimized where possible.

int deadlock rollbacks The total number of forced rollbacks initiated by the
database manager because of a deadlock. The value format is integer. The database
manager initiates a rollback for the current unit of work in an application that is
experiencing a deadlock. This attribute shows the number of deadlocks that have
been broken. It can indicate the possibility of concurrency problems. It is also
important because internal rollbacks because of deadlocks can cause performance
degradation.

int rollbacks The total number of rollbacks initiated internally by the database
manager. The value format is integer. An internal rollback occurs when any of the
following cannot complete successfully:
v A reorganization
v An import
v A bind or precompile
v An application that ends as a result of a deadlock situation or lock timeout
situation
v An application that ends without executing an explicit COMMIT or ROLLBACK
statement (on Windows)
Use this attribute to calculate the total number of units of work by calculating the
sum of the following: commit statements attempted, internal commits, rollback
statements attempted, and internal rollbacks.

last backup The date and time that the latest database backup was completed.

lock escals The number of times that locks have been escalated from several row
locks to a table lock. The value format is integer. This data item includes a count of
all lock escalations, including exclusive lock escalations. When an application
reaches the maximum number of locks allowed and there are no more locks to
escalate, it uses space in the lock list allocated for other applications. When the
entire lock list is full, an error occurs.

lock list in use The total amount of lock list memory (in bytes) that is currently in
use. The value format is integer. This attribute can be used in conjunction with the
locklist configuration parameter to calculate the lock list utilization. If the lock list
utilization is high, you want to consider increasing the size of that parameter.

lock timeouts The number of times that a request to lock an object timed-out
instead of being granted. The value format is integer. This attribute can help you
adjust the setting for the LOCKTIMEOUT database configuration parameter. If the
number of lock time-outs becomes excessive when compared to normal operating
levels, an application be holding locks for long durations. In this case, this attribute
indicate that you need to analyze some of the other attributes related to locks and
deadlocks to determine if an application problem exists.

It is also possible to have too few lock time-outs if the LOCKTIMEOUT database
configuration parameter is set too high. In this case, applications wait excessively
to obtain a lock.

Chapter 5. Attributes reference 81

 lock wait time The total elapsed time waited for a lock in seconds. The value
format is integer.

At the database level, this is the total amount of elapsed time that all applications
were waiting for a lock within this database. At the application-connection and
transaction levels, this is the total amount of elapsed time that this connection or
transaction has waited for a lock to be granted.

Use this attribute in conjunction with the Lock Waits attribute to calculate the
average wait time for a lock. This calculation can be performed at either the
database or the application-connection level. If the average lock wait time is high,
look for applications that hold many locks, or have lock escalations, with a focus
on tuning your applications to improve concurrency, if appropriate. If escalations
are the reason for a high average lock wait time, the values of one or both of the
LOCKLIST and MAXLOCKS configuration parameters be too low.

lock waits The total number of times that applications or connections waited for
locks.

At the database level, this is the total number of times that applications waited for
locks within this database. At the application-connection level, this is the total
number of times that this connection requested a lock but waited because another
connection was already holding a lock on the data.

Use this attribute with Lock Wait Time attribute to calculate, at the database level,
the average wait time for a lock. This calculation can be done at either the
database or the application-connection level. If the average lock wait time is high,
look for applications that hold many locks, or have lock escalations, with a focus
on tuning your applications to improve concurrency, if appropriate. If escalations
are the reason for a high average lock wait time, the values of one or both of the
LOCKLIST and MAXLOCKS configuration parameters be too low.

locks held The number of locks currently held. The value format is integer. If the
monitor information is at the database level, this number represents the total
number of locks currently held by all applications in the database. If the
information is at the application level, this number represents the total number of
locks currently held by all agents for the application.

locks waiting The number of agents that are currently waiting on a lock. The
value format is integer.

When used in conjunction with Applications Currently Connected attribute, this

attribute indicates the percentage of applications waiting on locks. If this number is
high, the applications have concurrency problems. In this case, identify
applications that are holding locks or exclusive locks for long periods of time.

log reads The number of log pages read from disk by the logger. The value format
is integer. Use this attribute with an operating system monitor to quantify the
amount of I/O on a device that is attributable to database activity.

log writes The number of log pages written to disk by the logger. The value
format is integer. Use this attribute with an operating system monitor to quantify
the amount of I/O on a device that is attributable to database activity.

node name Indicates the originating name. Value format is a simple text string
with a maximum of 32 characters.

82 IBM Tivoli Monitoring for Databases: DB2 Agent: User’s Guide

num assoc agents The current number of subagents associated with all
applications that are connected to the monitored database. The value format is
integer. Use this attribute to evaluate the settings for the agent configuration
parameters.

pkg cache hit ratio The percentage of package sections that were found in cache.
The value format is integer. The value is calculated as the packages found in cache
divided by the package cache lookups. This ratio tells you whether the package
cache is being used effectively. If the hit ratio is high (more than 0.8), the cache is
performing well. A smaller ratio indicates that the package cache needs to be
increased.

pkg cache inserts The total number of times that a requested section was not
available for use and had to be loaded into the package cache. The value format is
integer. This count includes any implicit prepares performed by the system. In
conjunction with Package Cache Lookups attribute, you can calculate the package
cache hit ratio using the following formula:
1 - (Package Cache Inserts / Package Cache Lookups)

pkg cache lookups The number of times that an application looked for a section or
package in the package cache. The value format is integer. At a database level, it
indicates the overall number of references since the database was started, or
monitor data was reset. Note that this counter includes the cases where the section
is already loaded in the cache and when the section has to be loaded into the
cache. To calculate the package cache hit ratio use the following formula:
1 - (Package Cache Inserts / Package Cache Lookups)

The package cache hit ratio tells you whether the package cache is being used
effectively. If the hit ratio is high (more than 0.8), the cache is performing well. A
smaller ratio indicates that the package cache needs to be increased.

pool async data read reqs The number of asynchronous read requests. The value
format is integer. To calculate the average number of data pages read per
asynchronous request, use the following formula:
buffer pool asynchronous data reads / buffer pool asynchronous read requests

This average can help to determine the amount of asynchronous I/O done in each
interaction with the prefetcher.

pool async data reads The number of pages read asynchronously into the buffer
pool. The value format is integer. Use this attribute with Buffer Pool Data Physical
Reads attribute to calculate the number of physical reads that were performed
synchronously (that is, physical data page reads that were performed by database
manager agents). Use the following formula:
buffer pool data physical reads - buffer pool asynchronous data reads

By comparing the ratio of asynchronous to synchronous reads, you can gain

insight into how well the prefetchers are working.

pool async data writes The number of times a buffer pool data page was
physically written to disk by an asynchronous page cleaner or by a prefetcher. The
value format is integer. A prefetcher have written dirty pages to disk to make
space for the pages being prefetched.

Chapter 5. Attributes reference 83

 Use this attribute with Buffer Pool Data Writes to calculate the number of physical
write requests that were performed synchronously (that is, physical data page
writes that were performed by database manager agents). Use the following
formula:
buffer pool data writes - buffer pool asynchronous data writes

By comparing the ratio of asynchronous to synchronous writes, you can gain

insight into how well the buffer pool page cleaners are performing.

pool async index reads The number of index pages read asynchronously into the
buffer pool by a prefetcher. The value format is integer. Asynchronous reads are
performed by database manager prefetchers.

Use this attribute with Buffer Pool Index Physical Reads to calculate the number of
physical reads that were performed synchronously (that is, physical index page
reads that were performed by database manager agents). Use the following
formula:
buffer pool index physical reads - buffer pool asynchronous index reads

By comparing the ratio of asynchronous to synchronous reads, you can gain

insight into how well the prefetchers are working.

pool async index writes The number of times a buffer pool index page was
physically written to disk by an asynchronous page cleaner or a prefetcher. The
value format is integer. A prefetcher have written dirty pages to disk to make
space for the pages being prefetched.

Use this attribute with Buffer Pool Index Writes to calculate the number of physical
index write requests that were performed synchronously. That is, physical index
page writes that were performed by database manager agents. Use the following
formula:
buffer pool index writes - buffer pool asynchronous index writes

By comparing the ratio of asynchronous to synchronous writes, you can gain

insight into how well the buffer pool page cleaners are performing.

pool async read time The total elapsed time spent reading by database manager
prefetchers. The value format is integer. Use this attribute to calculate the elapsed
time for synchronous reading, using the following formula:
total buffer pool physical read time - buffer pool synchronous read time

You can also use this attribute to calculate the average asynchronous read time
using the following formula:
buffer pool asynchronous read time / buffer pool asynchronous data reads

These calculations can be used to understand the I/O work being performed.

pool async write time The total elapsed time spent writing data or index pages
from the buffer pool to disk by database manager page cleaners. The value format
is integer. Calculate the elapsed time spent writing pages synchronously by
subtracting the value of Pool Async Write Time attribute from the value of Pool
Physical Write Time attribute. You can also use this attribute to calculate the
average asynchronous read time by:
1. Adding together the Pool Async Data Writes and the Pool Async Index Writes.
2. Dividing the Pool Async Write Time by the sum from step 1.

84 IBM Tivoli Monitoring for Databases: DB2 Agent: User’s Guide

These calculations can be used to understand the I/O work being performed.

pool data from estore Number of buffer pool data pages copied from extended
storage. The value format is integer. Required pages are copied from extended
storage to the buffer pool. The copy process incur the cost of connecting to the
shared memory segment, but it saves the cost of a disk read.

pool data l reads The number of logical read requests for data pages that have
gone through the buffer pool. The value format is integer. This count includes
accesses to data that is:
v Already in the buffer pool when the database manager needs to process the
page
v Read into the buffer pool before the database manager can process the page
In conjunction with Pool Data Physical Reads, you can calculate the data page hit
ratio for the buffer pool as follows:
1 - (buffer pool data physical reads / buffer pool data logical reads)

In conjunction with Pool Data Physical Reads, Pool Index Physical Reads, and Pool
Index Logical Reads, you can calculate the overall buffer pool hit ratio as follows:
1 - ((buffer pool data physical reads + buffer pool index physical reads) /
(buffer pool data logical reads + buffer pool index logical reads))

Increasing buffer pool size will generally improve the hit ratio until you reach a
point of diminishing return.

pool data p reads The number of pool data read requests that have gone through
the buffer pool since the first connection.

pool data to estore Number of buffer pool data pages copied to extended storage.
The value format is integer.

Pages are copied from the buffer pool to extended storage when they are selected
as victim pages. As a result of the copying process, there is sufficient space for new
pages in the buffer pool.

pool data writes The number of times a buffer pool data page was physically
written to disk. The value format is integer. A buffer pool data page is written to
disk for the following reasons:
v To free a page in the buffer pool so another page can be read
v To flush the buffer pool
If a buffer pool data page is written to disk for a high percentage of the Buffer
Pool Data Physical Reads, performance improve by increasing the number of
buffer pool pages available for the database.

pool drty pg steal clns The number of times a page cleaner was invoked because a
synchronous write was needed during the victim buffer replacement for the
database. The value format is integer. Use this attribute, in combination with
others, to evaluate the number of page cleaners that are defined.

pool drty pg thrsh clns The number of times a page cleaner was invoked because
a buffer pool had reached the dirty page threshold criterion for the database. The
value format is integer. When the number of dirty pages in the pool exceeds this
value, the cleaners are triggered. If this value is set too low, pages might be written

Chapter 5. Attributes reference 85

 out too early, requiring them to be read back in. If set too high, too many pages
accumulate, requiring users to write out pages synchronously.

pool hit ratio The buffer pool hit ratio. The value format is integer. The sum of
Pool Data Logical Reads and Pool Index Logical Reads attributes is divided by the
value of Pool Total Reads attribute to derive the pool hit ratio. Use this attribute to
determine whether buffer pool assignment is efficient. If the pool hit ratio is low,
increasing the number of buffer pool pages improve performance.

pool index from estore Number of buffer pool index pages copied from extended
storage. The value format is integer. Required index pages are copied from
extended storage to the buffer pool. The copy process incur the cost of connecting
to the shared memory segment, but it saves the cost of a disk read.

pool index l reads The number of logical read requests for index pages that have
gone through the buffer pool. The value format is integer. This count includes
accesses to index pages that are:
v Already in the buffer pool when the database manager needs to process the
page
v Read into the buffer pool before the database manager can process the page
In conjunction with Buffer Pool Index Physical Reads attribute, you can calculate
the index page hit ratio for the buffer pool as follows:
1 - (buffer pool index physical reads / buffer pool index logical reads)

To calculate the overall buffer pool hit ratio, see Buffer Pool Data Logical Reads. If
the hit ratio is low, increasing the number of buffer pool pages improve
performance.

pool index p reads The number of physical read requests to get index pages into
the buffer pool. The value format is integer. See Pool Index Logical Reads for
information about how to use this element.

pool index to estore Number of buffer pool index pages copied to extended
storage. The value format is integer. Pages are copied from the buffer pool to
extended storage when they are selected as victim pages. As a result of the
copying process, there is sufficient space for new pages in the buffer pool.

pool index writes The number of times a buffer pool index page was physically
written to disk. The value format is integer. If a buffer pool index page is written
to disk for a high percentage of the Buffer Pool Index Physical Reads, performance
improve by increasing the number of buffer pool pages available for the database.
If all applications are updating the database, increasing the size of the buffer pool
have minimal impact on performance; most pages contain updated data that must
be written to disk.

pool lsn gap clns The number of times a page cleaner was invoked because the
logging space used had reached a predefined criterion for the database. The value
format is integer. Use this attribute to evaluate whether you have enough space for
logging, and whether you need more log files or larger log files.

pool read time The total amount of elapsed time spent processing read requests
that caused data or index pages to be physically read from buffer pool to disk. The
value format is integer.

86 IBM Tivoli Monitoring for Databases: DB2 Agent: User’s Guide

Use this attribute with Buffer Pool Data Physical Reads and Buffer Pool Index
Physical Reads attributes to calculate the average page-read time. This average is
important because it indicate the presence of an I/O wait, which in turn indicate
that you need to move data to a different device.

pool sync data reads The number of physical data page reads that were performed
by database manager agents. The value format is integer. This value is derived by
subtracting the value of Pool Async Data Reads attribute from the Pool Data
Physical Reads attribute.

By comparing the ratio of asynchronous to synchronous reads, you can gain

insight into how well the prefetchers are working.

pool sync data writes The total number of physical write requests that were
performed synchronously (that is, physical data page writes that were performed
by database manager agents). The value format is integer. This value is derived by
subtracting the value of Pool Async Data Writes attribute from the value of Pool
Data Writes attribute.

By comparing the ratio of asynchronous to synchronous writes, you can gain

insight into how well the buffer pool page cleaners are performing.

pool sync index reads The number of index pages read synchronously (that is,
physical index page reads that were performed by database manager agents) into
the buffer pool. The value format is integer. This value is derived by subtracting
the value of Pool Async Index Reads attribute from Pool Index Physical Reads
attribute.

By comparing the ratio of asynchronous to synchronous reads, you can gain

insight into how well the prefetchers are working.

pool sync index writes The number of physical index write requests that were
performed synchronously (that is, physical index page writes that were performed
by database manager agents). The value format is integer. This value is derived by
subtracting the value of Pool Async Index Writes attribute from the value of Pool
Index Writes attribute.

By comparing the ratio of asynchronous to synchronous writes, you can gain

insight into how well the buffer pool page cleaners are performing.

pool sync read The total number of synchronous reads. The value format is
integer. This value is derived by adding together the values of the Pool Sync Data
Reads and Pool Sync Index Reads attributes.

pool sync read time The elapsed time used to perform all synchronous reads. The
value format is integer. This value is derived by subtracting the value of Pool
Async Read Time attribute from the value of Pool Read Time attribute.

Use this attribute to understand the I/O work being performed.

pool sync write The total number of synchronous index writes. The value format
is integer. The value is derived by adding together the values of Pool Sync Data
Writes attribute and Pool Sync Index Writes attribute.

Chapter 5. Attributes reference 87

 pool sync write time The total elapsed time used to perform all synchronous
writes. The value format is integer. This value is derived by subtracting the value
of Pool Async Write Time attribute from the value of Pool Write Time attribute.

pool total reads The total number of read requests that required I/O to get data
pages and index pages into the buffer pool. The value format is integer. This
attribute is the total of Pool Data Logical Reads, Pool Index Logical Reads, Pool
Data Physical Reads, and Pool Index Physical Reads attributes.

pool total writes The total number of write requests. The value format is integer.
This attribute is the total of Pool Data Writes and Pool Index Writes attributes.

pool write time The total amount of time spent physically writing data or index
pages from the buffer pool to disk. The value format is integer.

Use this attribute with Buffer Pool Data Writes and Buffer Pool Index Writes
attributes to calculate the average page-write time. This average is important
because it indicate the presence of an I/O wait, which in turn indicate that you
need to move data to a different device.

prefetch wait time The time an application spent waiting for an I/O server
(prefetcher) to finish loading pages into the buffer pool. The value format is
integer. This attribute can be used to experiment with changing the number of I/O
servers and the I/O server sizes.

rollback sql stmts The total number of SQL ROLLBACK statements that have been
attempted. The value format is integer. A rollback can result from an application
request, a deadlock, or an error situation. This attribute counts only the number of
rollback statements issued from applications. At the application level, this attribute
can help you determine the level of database activity for the application and the
amount of conflict with other applications.

At the database level, it can help you determine the amount of activity in the
database and the amount of conflict between applications on the database.

rows deleted The number of row deletions attempted. The value format is integer.
Use this attribute to gain insight into the current level of activity within the
database manager.

rows inserted The number of row insertions attempted. The value format is
integer. Use this attribute to gain insight into the current level of activity within
the database manager.

rows selected The number of rows that have been selected and returned to the
application. The value format is integer. Use this attribute to gain insight into the
current level of activity within the database manager.

rows updated The number of row updates attempted. The value format is integer.
Use this attribute to gain insight into the current level of activity within the
database manager.

sec logs allocated The total number of secondary log files that are currently being
used for the database. The value format is integer.

Use this attribute in conjunction with Secondary Log Used Top and Total Log Used
Top attributes to show the current dependency on secondary logs. If this value is

88 IBM Tivoli Monitoring for Databases: DB2 Agent: User’s Guide

consistently high, you need larger log files, more primary log files, or more
frequent COMMIT statements within your application.

sec log used top The maximum amount of secondary log space (in bytes) that has
been used. The value format is integer.

Use this attribute in conjunction with Secondary Logs Allocated and Total Log
Used Top attributes to show the current dependency on secondary logs. If this
value is high, you need larger log files, more primary log files, or more frequent
COMMIT statements within your application.

select sql stmts The number of SQL SELECT statements that were executed. The
value format is integer. Use this attribute to determine the level of database
activity at the application or database level. You can also use the following formula
to determine the ratio of SELECT statements to the total statements by:
1. Adding together the number of static SQL statements attempted and dynamic
SQL statements attempted
2. Dividing the resulting total by the number of select SQL statements executed

server platform The operating system upon which the database management
system is running. The value format is a text string with a maximum of 16
characters. Use this attribute during problem determination for remote
applications.

snapshot time The date and time when the database system monitor information
was collected.

sort heap allocated The total number of allocated pages of sort heap space for all
sorts at the level chosen (database manager or database) and at the time the
snapshot was taken. The value format is integer. Memory estimates normally do
not include sort heap space. If excessive sorting occurs, add the extra memory
(used for the sort heap) to the base memory requirements for running the database
manager. Generally, the larger the sort heap, the more efficient the sort.
Appropriate use of indexes can reduce the amount of sorting required.

sort overflows The total number of sorts that ran out of sort heap and have
required disk space for temporary storage. The value format is integer. At a
database or application level, use this attribute in conjunction with Total Sorts
attribute to calculate the percentage of sorts that required overflow to disk. If this
percentage is high, you want adjust the database configuration by increasing the
value of sortheap.

sort overflows pct The percentage of sorts that ran out of sort heap and have
required disk space for temporary storage. The value format is integer. This
percentage is calculated by dividing the value of Sort Overflows attribute by the
value of the Total Sorts attribute. At a database or application level, use this
attribute to evaluate the percentage of sorts that required overflow to disk. If this
percentage is high, you want adjust the database configuration by increasing the
value of sortheap.

sql stmts failed pct The percentage of SQL statements that failed to execute
successfully. The value format is integer. This value is derived by dividing the
value of Failed SQL Statements attribute by the value of Total SQL Statements
attribute. Use this attribute to determine whether an application has some design
issues.

Chapter 5. Attributes reference 89

 sql stmts rollback pct The percentage of SQL statements that resulted in a
rollback. The value format is integer. This value is derived by dividing the value of
Rollback SQL Statements attribute by the value of the Total SQL Statements
attribute. Use this attribute to determine whether an application has some design
issues.

static sql stmts The number of static SQL statements that were attempted. The
value format is integer.

Use this attribute to calculate the total number of successful SQL statements at the
database or application level by:
1. Adding together the number of Dynamic SQL Statements Attempted and the
Static SQL Statements Attempted
2. Subtracting the number of Failed Statement Operations

The remainder equals the throughput (the number of successful SQL statements)
during the current monitoring period.

tot log used top The maximum amount of total log space (in bytes) that has been
used. The value format is integer. Use this attribute to evaluate the amount of
primary log space that is allocated. Comparing the value of this attribute with the
amount of primary log space that is allocated can help you to evaluate the
configuration parameter settings.

total cons The number of connections to the database since the first connect,
activate, or last reset (coordinator agents). The value format is integer.

Use this attribute in conjunction with the Database Activation Timestamp and the
Start Database Manager Timestamp attributes to calculate the frequency at which
applications have connected to the database.

The first connect to a database (such as initial buffer pool allocation) causes extra
overhead. If the frequency of connects is low, it be beneficial to activate the
database explicitly using the ACTIVATE DATABASE command before connecting
any other application. As a result, subsequent connects will be processed at a
higher rate.

Note: When you reset this attribute, its value is set to the number of applications
that are currently connected, not to zero.

total hash joins The total number of hash joins executed. The value format is
integer.

At the database or application level, use this value in conjunction with Hash Join
Overflows attribute and Hash Join Small Overflows attribute to determine if a
significant percentage of hash joins would benefit from modest increases in the sort
heap size.

total hash loops The total number of times that a single partition of a hash join
was larger than the available sort heap space. The value format is integer. This
attribute indicate inefficient execution of hash joins (the sort heap size is too small
or the sort heap threshold is too small). Use this value in conjunction with the
other hash join variables to tune the sort heap size (SORTHEAP) and sort heap
threshold (SHEAPTHRES) configuration parameters.

90 IBM Tivoli Monitoring for Databases: DB2 Agent: User’s Guide

total sec cons The number of connections made by a subagent to the database at
the node. The value format is integer.

Use this attribute with the Connects Since Database Activation, Database
Activation Timestamp, and the Start Database Manager Timestamp attributes to
calculate the frequency at which applications have connected to the database.

total sort time The total elapsed time (in milliseconds) for all sorts that have been
executed. The value format is integer. At a database or application level, use this
attribute with Total Sorts to calculate the average sort time, which can indicate
whether sorting is a performance issue. Elapsed times are affected by system load.
The more processes you have running, the higher this elapsed time value will be.

total sorts The total number of sorts that have been executed. The value format is
integer. At a database or application level, use this value with the Sort Overflows
attribute to calculate the percentage of sorts that need more heap space. You can
also use it with Total Sort Time attribute to calculate the average sort time. If the
number of sort overflows is small with respect to the total sorts, increasing the sort
heap size have little impact on performance, unless this buffer size is increased
substantially.

total sql stmts The total number of UPDATE, INSERT, and DELETE SQL
statements. The value format is integer. This value is derived by adding together
the values of the Dynamic SQL Statements and the Static SQL Statements
attributes.

uid sql stmts The number of SQL UPDATE, INSERT, and DELETE statements that
were executed. The value format is integer.

Use this attribute to determine the level of database activity at the application or
database level. You can also use the following formula to determine the ratio of
UPDATE, INSERT and DELETE statements to the total number of statements:
1. Adding together the number of static SQL statements attempted and the
dynamic SQL statements attempted
2. Dividing the number of UPDATE/INSERT/DELETE SQL statements executed
by the sum derived in step 1

x lock escals The number of times that locks have been escalated from several row
locks to one exclusive table lock, or the number of times an exclusive lock on a
row caused the table lock to become an exclusive lock. The value format is integer.
A lock is escalated when the total number of locks held by an application reaches
the maximum amount of lock list space available to the application. The amount of
lock list space available is determined by the LOCKLIST and MAXLOCKS
configuration parameters. Other applications cannot access data held by an
exclusive lock.

Because exclusive locks can impact the concurrency of your data, it is important to
track them. When an application reaches the maximum number of locks allowed
and there are no more locks to escalate, it uses space in the lock list allocated for
other applications. When the entire lock list is full, an error occurs.

See Lock Escals attribute for possible causes and resolutions to excessive exclusive
lock escalations.

Chapter 5. Attributes reference 91

 Database Information attributes (group 01)
Use these attributes to obtain information about the database. With this
information, you can determine the efficiency of the database and identify any
problem areas for corrective action.

Unless otherwise noted, integer values are 32-bit or 0 to 4 294 967 296.

appl ctl heap sz The maximum size (in 4-KB pages) for the application control
heap in the database during the monitoring interval. The value format is integer.

The heap is required to share information among agents working on behalf of the
same application at a node in a massively parallel processing (MPP) or a
symmetric multiprocessor (SMP) system. If complex applications are being run or
the MPP configuration has a large number of nodes, you need to increase the size
of this heap.

In a partitioned database environment, this heap is used to store copies of the

executing section of SQL statements for agents and subagents. However, symmetric
multiprocessor agents (SMP), subagents, and agents in all other environments use
applHeapSz.

applHeapSz The size (in 4-KB pages) of the application heap that is available for
each individual agent in the database during the monitoring interval. The value
format is integer.

Increase the value of the parameter if your application receives an error indicating
that there is not enough storage in the application heap. The heap is allocated
when an agent or subagent is initialized for an application. The amount allocated
is the minimum amount needed to process the request given to the agent or
subagent. When the agent or subagent requires more heap space to process larger
SQL statements, the database manager will allocate memory as needed, up to the
maximum specified by the parameter.

appls in db2 The number of applications currently executing in the database. The
value format is integer.

avg appls The value of the average number of active applications. The value
format is integer.

avg direct read time The average time (in milliseconds) that is used to perform
direct reads to the database. The value format is integer. The value is derived
through this formula: direct read time / direct reads.

avg direct write time The average time (in milliseconds) for performing direct
writes to the database. The value format is integer. A high average time can
indicate the existence of an input and output conflict. The value is derived through
this formula: direct write time / direct write.

avg locks held The average number of locks held by each currently connected
application in the database. The value format is integer. The value is derived
through this formula: 100 * locks held / appls cur cons. If the returned value is
high compared to normal operating levels, it can indicate that one or more
applications is using an excessive number of locks. Refine such applications to
improve performance.

92 IBM Tivoli Monitoring for Databases: DB2 Agent: User’s Guide

avg pages per cleaner for int The average number of pages written per page
cleaner that are invoked for the database during the monitoring interval. The value
format is integer.

Use the returned value to determine how many pages are handled by this
database’s page cleaners. If this value increases over time, you can define more
page cleaners.

avg pool async data reads The average number of buffer pool asynchronous data
reads when compared to the total number of pool reads for the database. The
value format is integer. The value is derived through this formula: pool async
data reads / (pool data p reads + pool index p reads).

Use the returned value to gain insight into how well the prefetchers are working
and to refine the num_ioservers configuration parameter. If the returned value is
low compared to normal operating levels, there might not be enough input and
output servers to prefetch data into the buffer, causing the database manager
agents to spend extra time on physical reads. Increase the number of input and
output servers by increasing the value of num_ioservers configuration parameter. If
too many servers are allocated, system performance is not reduced because the
extra input and output servers are not used.

avg pool async data writes The average number of buffer pool asynchronous data
writes (data and index) when compared to the total number of pool writes for the
database. The value format is integer. The value is derived through this formula:
pool async data writes / (pool data writes + pool index writes).

Use the returned value to gain insight into how well the page cleaners are working
and to refine the num_iocleaners configuration parameter. If the returned value is
low compared to normal operating levels, increase the number of input and output
cleaners by increasing the value of num_iocleaners. If the returned value is high
compared to normal operating levels, you can save system resources by decreasing
the number of input and output cleaners (by decreasing the value of
num_iocleaners).

avg pool io time The average time (in milliseconds) for performing buffer pool
input and output operations (reading or writing) to the database. The value format
is integer.

A high average time can indicate the existence of an input and output conflict. In
this case, you might need to move data to a different device. The returned value
includes the time applied to asynchronous input and output operations (which are
performed by prefetchers and page cleaners).

avg pool writes per read The ratio of total pool writes to pool reads for the
database. The value format is integer. The value is derived through this formula:
(pool data writes + pool index writes) / (pool data p reads + pool index p
reads).

If the returned value is greater than 1, you can improve performance by increasing
the available buffer pool space. A returned value greater than 1 indicates that at
least one write to disk had to occur (either to free a page in the buffer pool, or to
flush the buffer pool) before a page could be read into the buffer pool. You can
increase the available buffer pool space by freeing the space more often or by
increasing the total space for the buffer pool.

Chapter 5. Attributes reference 93

 avg sect read per direct read The average number of sectors that are read by a
direct read for the database. The value format is integer. The value is derived
through this formula: direct reads / direct read reqs. Direct reads do not use
the buffer pool, and so result in poor performance because the data is physically
read from disk each time. If you are using system monitors to track input and
output for the device, this value helps you distinguish database input and output
from non-database input and output.

avg sect written per direct write The average number of sectors that are written
by a direct write to the database. The value format is integer. The value is derived
through this formula: direct writes / direct write reqs. Direct writes do not
use the buffer pool, which results in poor performance because the data is
physically written from disk each time. If you are using system monitors to track
input and output for the device, this value helps you distinguish database input
and output from non-database input and output.

avg sync io time The average time (in milliseconds) to perform synchronous input
and output operations for the database. The value format is integer.

Use the returned value to analyze the input and output work being performed for
the database. Synchronous input and output operations for a database are
performed by database manager agents. Asynchronous input and output
operations are performed by prefetchers (reads) and page cleaners (writes). In
general, asynchronous input and output helps your applications run faster.

buffpage The value (in pages) of the default buffer pool size. The value format is
integer.

In the currently supported releases of DB2, multiple buffer pools be defined in a

single database. For instance, buffer pools can be defined and associated with a
particular tablespace. Each buffer pool created can be given its own individual
size.

The buffpage attribute serves only as a default value for buffer pools created
within a particular database. Therefore, the value of the buffpage attribute is much
less critical to performance in current releases of DB2, because most buffer pools
are given an individual size when created.

Do not use the buffpage attribute to evaluate or tune the performance of DB2
unless it is used as the default value when creating buffer pools in a database.

catalogcache sz The value (in units of 4-KB pages) of the catalog cache size. The
value format is integer.

This value is the maximum amount of space that the catalog cache can use from
the database heap (dbheap). The catalog cache is referenced whenever a table,
view, or alias name is processed during the compilation of an SQL statement. It is
dynamically allocated from dbheap, as required, until the catalog cache size is
reached.

chngpgs thresh The value (in percentage units) of the changed pages threshold.
The value format is integer.

This value sets a limit on how much buffer pool space can be occupied by changed
pages before the asynchronous page cleaners are started, if they are not currently
active. Asynchronous page cleaners write changed pages from the buffer pool to

94 IBM Tivoli Monitoring for Databases: DB2 Agent: User’s Guide

disk before the space in the buffer pool is required by a database agent. This
means that the agents do not need to wait for a changed page to be written out
before being able to read a page, and application transactions run faster.

commit stmts per sec The total number of commits initiated internally by the
database per second. Use the returned value to determine rates of database activity.
The value format is integer.

cur cons pct The percentage of applications currently connected. The value format
is integer.

db cap err The number of errors encountered by the Capture program within the
last 5 minutes. The value format is integer.

Use the returned value to determine whether the Capture program encountered an
error that prevented it from running. If any errors are detected, the Capture
program came down at the time the error occurred. The Capture program might or
might not still be down.

The Capture program is the most critical replication component in the replication
system. If the Capture program is not active, there will be no new change records
to apply to the target systems. If your data concurrency requirements are high and
you want to ensure that the Capture program runs continuously, use this monitor
to determine when the Capture program encounters an error that prevents it from
running.

db cap lag The time difference (in minutes) between the current timestamp and the
last timestamp recorded by the Capture program. The value format is integer. This
time difference is the Capture lag.

Use the returned value to determine whether the Capture program is keeping up
with the DB2 database log. The Capture program uses an interface to the DB2
database log or journal to detect and save changes to the data in the tables
registered for replication.

db cap prun The number of rows in the unit-of-work (UOW) table. The value
format is integer.

Use the returned value to help you determine whether you need to prune the
UOW table or the change data (CD) table.

dbheap The value (in units of 4-KB pages) of the database heap. The value format
is integer.

This value is the maximum amount of memory allowed for a database heap. There
is one database heap for each database. It is used on behalf of all applications
connected to the database. Refining dbheap has minimal impact on performance.
The main function of this parameter is to prevent the database manager from
allocating an excessive amount of space for a particular database.

db name The real name of the database for which information is collected or to
which the application is connected. This is the name the database was given when
created. Value format is a simple text string with a maximum of 20 characters.

Chapter 5. Attributes reference 95

 db name (Unicode) The real name of the database for which information is
collected or to which the application is connected. This is the name the database
was given when created. Value format is a simple text string with a maximum of
60 bytes. This attribute is globalized.

db tablespaces The number of Database Managed Space (DMS) tablespaces in the

database. The value format is integer.

Use this attribute to track database growth over a period of time.

ddl sql pct for int The percentage of total SQL statements that were SQL DDL
statements during the monitoring interval. The value format is integer.

Because of the high activity in the system catalog tables, try to keep DDL
statement activity to a minimum. If the returned value is high compared to normal
operating levels, determine the activity causing it to be high and restrict it from
being performed. Examples of DDL statements are CREATE TABLE, CREATE
VIEW, ALTER TABLE, and DROP INDEX.

You can also use the returned value to refine the package cache hit ratio for this
application. DDL statements can also impact the package cache by invalidating
sections that are stored there and causing additional system overhead because of
section recompilation.

deadlocks for int The number of deadlocks detected in the database during the
monitoring interval. The value format is integer.

Use the returned value to determine whether applications are experiencing conflict
problems in the database. These problems could be caused by the following
situations:
v Lock escalations are occurring for the database.
v An application can be locking tables explicitly when system-generated row locks
are sufficient.
v An application can be using an inappropriate isolation level when binding.
v Database catalog tables are locked for repeatable read.
v Applications are getting the same locks in different orders, resulting in deadlock.

You can resolve the problem by determining in which applications the deadlocks
are occurring. You could then try to modify the applications to better enable them
to execute concurrently.

estore rw ratio for int The ratio (as a percentage) of data and index pages copied
from extended storage to pages copied to extended storage during the monitoring
interval. The value format is integer.

When a page is transferred from extended storage to the buffer pool, you save a
system input and output call. However, you still incur the cost of attaching to the
extended memory segment, copying the page, and detaching from the segment.
Use the returned value to determine if you would benefit from using extended
storage. The higher the ratio, the more likely you are to benefit. In general,
extended storage is particularly useful if input and output activity is very high on
your system.

event monitors The number of event monitors defined in the database. The value
format is integer.

96 IBM Tivoli Monitoring for Databases: DB2 Agent: User’s Guide

Use the returned value to determine how many event monitors are defined for the
database. When you define an event monitor, its definition is stored in the
database system catalog table. You can create any number of event monitors.
However, the maximum number of event monitors that can be active for a
database at any given time is 32.

int auto rebinds The number of automatic rebinds or recompiles that were
attempted in the database. The value format is integer.

Use the returned value to determine the level of database activity. Automatic
rebinds are the internal binds that the system performs when a package is
invalidated. They can have a significant impact on performance and need to be
minimized where possible.

int commits The total number of commits initiated internally by the database. The
value format is integer.

Use the returned value to gain insight into internal activity within the database.
The returned value is also used in calculating the returned value for the ″db:
Committed statements per second″ monitor.

int deadlock rollbacks pct The percentage of the total number of internal rollbacks
because of deadlocks. The value format is integer.

Use the returned value to distinguish those rollbacks caused by internal deadlocks
from rollbacks caused by other situations (for example, incomplete imports). The
returned value is the percentage of internal rollbacks because of internal deadlocks
since the first database connection or the last reset of the database monitor
counters.

int deadlock rollbacks pct for int The percentage of rollbacks that were because of
deadlock during the monitoring interval. The value format is integer.

Use the returned value to distinguish those rollbacks caused by internal deadlocks
from rollbacks caused by other situations (for example, incomplete imports). The
returned value is the percentage of internal rollbacks because of internal deadlocks
since the first database connection or the last reset of the database monitor
counters.

int rows deleted The number of rows deleted from the database as a result of
internal activity. The value format is integer.

Use the returned value to gain insight into internal activity within the database. If
this activity is high compared to normal operating levels, you can evaluate your
table design to determine if the referential constraints or triggers that you defined
on your database are necessary.

Internal row deletions can be a result of either of the following situations:

v Enforcement of an ON CASCADE DELETE referential constraint
v Firing of a trigger

int rows inserted The number of rows inserted into a database as a result of
internal activity caused by triggers. The value format is integer.

Chapter 5. Attributes reference 97

 Use the returned value to gain insight into internal activity within the database. If
this activity is high compared to normal operating levels, you can evaluate your
design to determine if you can alter it to reduce this activity.

int rows updated The number of rows updated in the database as a result of
internal activity. The value format is integer.

Use the returned value to gain insight into internal activity within the database. If
this activity is high compared to normal operating levels, you can evaluate your
table design to determine if the referential constraints that you defined are
necessary.

Internal update activity can be a result of either of the following situations:

v Enforcement of a referential constraint with the ON DELETE SET™ NULL rule
v Firing of a trigger

invalid pkgs The number of all packages that are currently marked invalid in the
database. The value format is integer.

Use the returned value as an indication of the current number of invalid packages.
A package is marked invalid if it depends on an object (for example, a table) and
that object is dropped.

The number of invalid packages can indicate how many automatic rebinds are
necessary in the database. An invalid package is usually automatically rebound the
next time it is accessed, unless it was marked invalid because a trigger was
dropped or because the dropped object was not recreated. Use of automatic
rebinds can significantly lower performance and need to be minimized where
possible.

invalid sys pkgs The number of SYSTEM packages that are currently marked
invalid in the database. The value format is integer.

Use the returned value as an indication of the current number of invalid packages
owned by SYSTEM. A package is marked invalid if it depends on an object (for
example, a table) and that object is dropped.

The number of invalid packages can indicate how many automatic rebinds are
necessary in the database. An invalid package is usually automatically rebound the
next time it is accessed, unless it was marked invalid because a trigger was
dropped or because the dropped object was not recreated. Use of automatic
rebinds can significantly lower performance and need to be minimized where
possible.

invalid triggers The number of triggers that are marked invalid in the database.
The value format is integer.

Use the returned value to determine the number of triggers that must be
revalidated. A trigger is marked invalid if an object on which the trigger depends
is dropped. To revalidate an invalid trigger, you must retrieve its definition from
the database system catalog and submit a new CREATE TRIGGER statement.

locklist The value (in units of 4-KB pages) of the maximum storage for lock lists.
The value format is integer.

98 IBM Tivoli Monitoring for Databases: DB2 Agent: User’s Guide

This value is the amount of storage that is allocated to the lock list. There is one
lock list for each database, and it contains the locks held by all applications
concurrently connected to the database. Too small a value can lead to excessive
lock waits. Too high a value compared to normal operating levels can deprive the
system of resources or memory.

lock list in use pct The percentage of space used in the database’s locklist. The
value format is integer.

Use the returned value to determine how much of the locklist space is free for new
locks to be requested.

lock waits for int The number of times that applications had to wait for locks in
the database during the monitoring interval. The value format is integer.

Use the returned value as an indication of how much time is applied to waiting for
locks during a particular monitoring interval.

lock waits pct The percentage of currently connected applications that are waiting
for a lock in the database. The value format is integer. The value is derived
through this formula: 100 * locks waiting / appls cur cons.

If the returned value is high compared to normal operating levels, the applications
could have concurrency problems. Identify applications that are holding locks or
exclusive locks for long periods of time and determine whether they can release
their locks more often.

logbufsz This value specifies the amount of the database heap (dbheap) to use as a
buffer for log records before writing these records to disk. The value format is
integer. It is important that the log buffer be able to hold the amount of log space
used by an average transaction. Otherwise, logging performance will decrease and
slow the overall system.

log io for int The total amount of log input and output. This amount is the sum of
the number of log pages read and the number of log pages written within the
monitoring interval. The value format is integer.

Use the returned value to determine whether you need to move the log to a
different device. If this input and output is beyond the capabilities of the current
device, you can determine if moving the log (by changing the newlogpath
configuration parameter) will improve performance.

logprimary The number of primary log files. The value format is integer.

maxappls The value of the maximum number of active applications. The value
format is integer.

This value is the maximum number of concurrent applications that can be

connected (both local and remote) to a database. Because each application that
attaches to a database causes some private memory to be allocated, allowing a
large number of concurrent applications will potentially use more memory.

Increasing the value of this parameter without lowering the maxlocks parameter or
increasing the locklist parameter can cause you to reach the database limit on locks
(locklist) rather than the application limit. The result can be pervasive lock
escalation problems.

Chapter 5. Attributes reference 99

 maxlocks The value of the maximum percentage of lock list before escalation. The
value format is integer.

This value specifies the percentage of the lock list that an application can hold
before the database manager performs lock escalation. Lock escalation can increase
contention, which reduces system throughput and increases user response time.
The values for maxlocks and maxappls must satisfy (maxlocks × maxappls) >100,
and each lock uses 32 bytes. It is recommended that you rebind application
packages after changing this parameter.

mincommit The value of the number of commits to group. The value format is
integer.

With this parameter you can delay the writing of log records to disk until a
minimum number of commits have been performed. This delay can help reduce
the overhead associated with writing log records and can improve performance.

The default value for mincommit is 1, which can be too low for your environment.
By sampling the number of transactions per-second throughout the day, you can
determine the peak per second rate and adjust mincommit to accommodate all or
most transactions. This adjustment would minimize the number of log writes
under the heaviest conditions.

As you increase mincommit, you might also need to increase the log buffer size
(logbufsz) to avoid filling the log buffer. Filling the log buffer also forces the
writing of log records to disk. If you change mincommit, you must change the
value for the logbufsz configuration parameter.

newlogpath The current value of the newlogpath configuration parameter. Valid

value is a text string up to 256 characters in length.

You use the newlogpath configuration parameter to specify a new location for the
log files. The specified path does not become the current log path until both of the
following conditions are met:
v The database is in a consistent state.
v All users are disconnected from the database.

When the first new connection is made to the database, the database manager
moves the logs to this location.

newlogpath (Unicode) The current value of the newlogpath configuration

parameter. Valid value is a text string up to 768 bytes. This attribute is globalized.

You use the newlogpath configuration parameter to specify a new location for the
log files. The specified path does not become the current log path until both of the
following conditions are met:
v The database is in a consistent state.
v All users are disconnected from the database.

When the first new connection is made to the database, the database manager
moves the logs to this location.

node name Indicates the originating name. Value format is a simple text string
with a maximum of 32 characters.

100 IBM Tivoli Monitoring for Databases: DB2 Agent: User’s Guide
num iocleaners The current value of the number of asynchronous page cleaners.
The value format is integer.

This parameter specifies the number of asynchronous page cleaners for a database.
Page cleaners monitor the buffer pool and asynchronously write out changed
pages to disk to free space in the buffer pool.

num ioservers The current value of the number of input and output servers. The
value format is integer.

This value specifies the number of input and output servers for a database. Input
and output servers are used on behalf of the database agents to perform
asynchronous input and output operations for utilities such as backup and restore,
and to perform prefetch input and output (in which case, they are called
prefetchers) operations.

Prefetchers read pages from disk into the buffer pool in anticipation of their use. In
most situations, these pages are read just before they are needed. However,
prefetchers can cause unnecessary input and output operations by reading pages
into the buffer pool that might not be used.

For example, an application starts reading through a table, and prefetchers read
consecutive pages into the buffer pool before the pages are required by the
application. Then the application fills the application buffer and stops reading.
Meanwhile, the prefetchers already have performed the input and output
operations for additional pages and the buffer pool is partially taken up with those
pages.

To fully exploit all the input and output devices in the system, a good value for
num_ioservers to use is generally one or two more than the number of physical
devices on which the database resides.

page cleans for interval The number of times a page cleaner was invoked for the
database (for any reason) during the monitoring interval. The value format is
integer.

Use the returned value to determine how often pages are written to disk by the
database’s page cleaners. If this value increases over time, you can define more
page cleaners. The number of page cleaners is determined by the number of I/O
cleaners configured.

pages per prefetch for int The number of data pages read per prefetch request for
the database during the monitoring interval. The value format is integer.

Use the returned value to determine the amount of asynchronous input and output
done in each interaction with the prefetcher. An excessively low returned value
when compared to normal operating levels indicates that you need more input and
output servers. The more input and output servers that you have, the better your
query performance.

pckcachesz The current value (in units of 4-KB pages) of the package cache size.
The value format is integer.

This value controls the amount of application heap memory to be used for caching
a package’s static and dynamic SQL statements.

Chapter 5. Attributes reference 101

 You will need to experiment with the size of the package cache to find the optimal
number for this attribute. For example, you can use a smaller package cache size if
there is no increase in the number of package cache inserts when you decrease the
size of the cache. Decreasing the package cache size frees up system resources for
other work. However, increasing the package cache size can improve overall
system performance if it results in a decrease of package cache inserts.

pool hit ratio index pct for int The database’s index page hit ratio (as a
percentage) for the buffer pool during the monitoring interval. The value format is
integer.

The index page hit ratio for the buffer pool indicates the percentage of index page
requests for which the database manager did not need to load an index page from
disk to service. That is, the index page was already in the buffer pool. The higher
the returned value, the lower the frequency of disk input and output, and the
faster the performance.

If the hit ratio is low compared to normal operating levels, increasing the number
of buffer pool pages can improve performance.

pool hit ratio pct for int The overall buffer pool hit ratio (as a percentage) for the
database during the monitoring interval. The value format is integer. This hit ratio
includes both index and data page activity.

The overall buffer pool hit ratio indicates the percentage of page requests for
which the database manager did not need to load a page from disk to service.
(That is, the page was already in the buffer pool.) The greater the buffer pool hit
ratio, the lower the frequency of disk input and output.

If the hit ratio is low compared to normal operating levels, increasing the number
of buffer pool pages can improve performance. A ratio of zero indicates that pages
needed to be read for every request.

For a large database, increasing the buffer pool size can have a minimal effect on
the buffer pool hit ratio. Such a database can have so large a number of data pages
that the statistical chance of a hit is not increased by an increase of the buffer
pools. However, even though the data might be too large to fit in the buffer pool,
the entire index can fit. In this case, you can refine buffer pool sizes until the
overall buffer pool hit ratio stops increasing, then refine the buffer pool until the
buffer pool index hit ratio no longer increases.

pool io per sec The rate (per second) of buffer pool input and output operations
for the database. The value format is integer. Buffer pool input and output includes
all physical data and index pages that go through the buffer pool when read or
written.

Use the returned value to determine how efficient your data storage device is. A
low value indicates the presence of an input and output wait, in which case you
need to move data to a different device.

pool sync index reads The number of pool index physical reads minus the pool
asynchronous index reads. The value format is integer.

pri log used pct The percentage of total log space used by the primary log. The
value format is integer.

102 IBM Tivoli Monitoring for Databases: DB2 Agent: User’s Guide
Use the returned value to help you evaluate the allocated amount of primary log
space and refine the log buffer size, log file size, and primary log configuration
parameters. The returned value is valid only if circular logging is used.

pri log used top The maximum number of primary logs used. The value format is
integer.

restore pending The RESTORE PENDING status in the database during the last
monitoring interval. The value format is integer.

sec log used pct The percentage of maximum log space used by the secondary log.
The value format is integer.

Use the returned value to show the current dependency on secondary logs.
Secondary logs are used when you have circular logging (log retention off) and the
primary log files are full.

If the returned value is high compared to normal operating levels, you might need
one or more of the following:
v Larger log files
v More primary log files
v More frequent COMMIT statements within your applications.

The returned value is zero if the database does not have any secondary log files (if
none are defined or if log retention is enabled).

select sql pct for int The percentage of total SQL statements that were SQL
SELECT statements during the monitoring interval. The value format is integer.

Use the returned value to determine the level of application activity and
throughput for the database.

seqdetect The current value of the sequential detection flag, which determines if
the database manager needs to perform sequential detection. The value format is
integer.

The database manager can monitor input and output operations. If sequential page
reading is occurring, the database manager can activate input and output
prefetching. This type of sequential prefetch is known as sequential detection.

If this configuration parameter is set to ″no,″ prefetching takes place only if the
database manager determines that it is useful (for example, in table sorts).

snapshot time The date and time when the database system monitor information
was collected.

sortheap The current value (in units of 4-KB pages) of the sort heap size. The value
format is integer.

This value is the maximum amount of memory that can be allocated as sort heap
for each sort within a database. The sort heap is the memory block where data is
sorted.

system tablespaces The number of SMS tablespaces in the database. The value
format is integer.

Chapter 5. Attributes reference 103

 Use the returned value to evaluate the use of SMS tablespaces and their effects on
performance. Table data that is read from disk is normally available in the
database’s buffer pool. Sometimes a data page is freed from the buffer pool before
it is used. For SMS tablespaces, when the database manager requests that data
page from the file system, the data page might still be in the file system’s own
cache. Having the page in the file system’s cache saves an input and output
operation that would otherwise have been required. (For more information, see the
DB2 administration documentation for the version of DB2 that you are using.) If
you have many SMS tablespaces, you can increase the size of the file system cache
to take advantage of this extra buffering.

tables The number of tables in the database. The value format is integer.

Use this attribute to track database growth because of an increased number of

tables over a period of time.

tablespaces The number of tablespaces in the database. The value format is integer.

Use this attribute to track database growth over a period of time.

tablespaces long data The number of tablespaces that store LONG data in the
database. The value format is integer.

Use this attribute to track database growth over a period of time. LONG data can
take up a large amount of space in a database.

total log used The total log space used in bytes. The value format is integer.

total sync io The total number of synchronous reads and writes for both data and
index pages. The value format is integer.

Use the returned value to gain insight into how well the prefetchers and page
cleaners are working and help you refine the number of I/O servers and number
of I/O cleaners configuration. Because synchronous input and output operations
are performed by database agents, a high number of operations can slow
performance.

tot direct io time The total time (in milliseconds) applied to direct reads and writes
for the database. The value format is integer.

The returned value indicates the amount of time that the database performs direct
reads and writes. A high returned value compared to normal operating levels can
indicate the presence of an input and output conflict.

tot pool phys io The total time (in milliseconds) applied to physical I/O for the
database. The value format is integer.

A high returned value (as compared to the total number of physical buffer pool
input and output operations) can indicate the presence of an input and output
wait, which in turn can indicate that you need to move data to a different device.

tot pool phys read The total time (in milliseconds) applied to processing read
requests that caused data or index pages to be physically read from disk to the
buffer pool for the database. The value format is integer. The value is derived
through this formula: pool data p reads + pool index p reads.

104 IBM Tivoli Monitoring for Databases: DB2 Agent: User’s Guide
 The returned value is used to calculate the average pool read time. This average
can indicate the presence of an input and output wait, which in turn can indicate
that you need to move data to a different device.

tot pool phys write The total time (in milliseconds) for buffer pool physical writes
(including asynchronous writes). The value format is integer. The value is derived
through this formula: pool data writes + pool index writes.

The returned value is used to calculate the average pool write time. This average
can indicate the presence of an input and output wait, which in turn can indicate
that you need to move data to a different device.

tot sync io time The total time (in milliseconds) applied to processing requests for
synchronous reads or writes for the database. The value format is integer.

The returned value is the sum of the returned values from the average pool write
time (ms) and average pool read time (ms). This time is the amount of time that
database agents spend doing synchronous reads and writes.

triggers The number of triggers defined in the database. The value format is
integer.

Use this attribute to track the use of triggers in the database. There are benefits to
using triggers, including faster application development, easier maintenance, and
global enforcement of business rules. For more information, see the DB2
administration documentation for the version of DB2 that you are using.

uid sql pct for int The percentage of total SQL statements that were SQL UPDATE,
INSERT, and DELETE statements during the monitoring interval. The value format
is integer.

Use the returned value to determine the level of database data change activity.

user indexes The number of indexes created by users in the database. Indexes
created by SYSIBM are not counted. The value format is integer.

Use this to track the use of indexes in the database. The use of indexes can
improve performance; for example, faster sorting of data. However, indexes can
also have adverse effects on performance; for example, each INSERT or DELETE
operation performed on a table requires additional updating of each index on that
table. For a discussion of this topic, see the DB2 administration documentation for
the version of DB2 that you are using.

views The number of views in the database. The value format is integer.

Use this attribute to track the use of views in the database. Views can be created to
limit access to sensitive data, while allowing more general access to other data.
This provides flexibility in the way your programs and end-user queries can look
at the table data.

Locking Conflict attributes

These attributes provide information about locking conflicts. You can use this
information to identify which applications are waiting for resources and which
applications are holding the resources.

Chapter 5. Attributes reference 105

 Unless otherwise noted, integer values are 32-bit or 0 to 4 294 967 296.

agent id The application handle, which is a system-wide unique ID for the

application. The value format is integer. On multi-node systems, where a database
is partitioned, this ID is the same on every node where the application make a
secondary connection. Use the application handle to uniquely identify an active
application (application handle is synonymous with agent ID).

agent id holding lk The application handle of the agent holding a lock for which
this application is waiting. The value format is integer. The lock monitor group
must be turned on to obtain this information. This attribute can help you to
determine which applications are in contention for resources.

appl id The identifier generated when the application connects to the database at
the database manager or when DDCS receives a request to connect to a DRDA
database. The value format is a text string with a maximum of 32 characters. The
identifier is unique across the network. The application ID appears in various
formats, which depend on the communication protocol between the client and the
server on which the database manager and/or DDCS are running. Each of the
formats consists of three parts separated by periods. Use this ID (known on both
the client and server) to correlate the client and server parts of the application.

appl id (Unicode) The identifier generated when the application connects to the
database at the database manager or when DDCS receives a request to connect to a
DRDA database. The value format is a text string with a maximum of 96 bytes.
This attribute is globalized. The identifier is unique across the network. The
application ID appears in various formats, which depend on the communication
protocol between the client and the server on which the database manager and/or
DDCS are running. Each of the formats consists of three parts separated by
periods. Use this ID (known on both the client and server) to correlate the client
and server parts of the application.

appl id holding lk The application ID of the application that is holding a lock on

the object that this application is waiting to obtain. The value format is a text
string with a maximum of 32 characters.

Use this attribute to determine which applications are in contention for resources.
Specifically, it can help you to identify the application handle (agent ID) and table
ID that are holding the lock. Example: *Local.db2inst1.990212202018

appl id holding lk (Unicode) The application ID of the application that is holding

a lock on the object that this application is waiting to obtain. The value format is a
text string with a maximum of 96 bytes. This attribute is globalized.

Use this attribute to determine which applications are in contention for resources.
Specifically, it can help you to identify the application handle (agent ID) and table
ID that are holding the lock. Example: *Local.db2inst1.990212202018

appl name The name of the application running at the client as known to the
database manager or DB2 Connect. The value format is a text string with a
maximum of 20 characters.

Use this attribute with the Application ID to relate data items with your
application. In a client/server environment, this name is passed from the client to
the server to establish the database connection. For DRDA-AS connections, this
name is the DRDA external name.

106 IBM Tivoli Monitoring for Databases: DB2 Agent: User’s Guide
appl name (Unicode) The name of the application running at the client as known
to the database manager or DB2 Connect. The value format is a text string with a
maximum of 60 bytes. This attribute is globalized.

Use this attribute with the Application ID to relate data items with your
application. In a client/server environment, this name is passed from the client to
the server to establish the database connection. For DRDA-AS connections, this
name is the DRDA external name.

appl status The status of the application being monitored. The value format is a
text string with a maximum of 64 characters. The valid status values are:
Backing Up Database
Commit Active
Compiling SQL Stmt
Connect Pending
Connected
Creating Database
Disconnect Pending
I/O Error Waiting
Loading Database
Lock Waiting
Prepared Transaction
Quiescing a Tablespace
Recompiling Plan
Request Interrupted
Restarting Database
Restoring Database
Rollback Active
Trans. heuristically aborted
Trans. heuristically committed
Transaction ended
UOW Executing
UOW Waiting in the application
Unknown
Unloading Database

auth id The authorization ID of the user who invoked the application that is being
monitored. On a DB2 Connect gateway node, this is the user’s authorization ID on
the host. The value format is a text string with a maximum of 20 characters. Use
this attribute to determine who invoked the application.

auth id (Unicode) The authorization ID of the user who invoked the application
that is being monitored. On a DB2 Connect gateway node, this is the user’s
authorization ID on the host. The value format is a text string with a maximum of
60 bytes. This attribute is globalized. Use this attribute to determine who invoked
the application.

client db alias The alias defined within the database manager where the database
connection request originated. The value format is a text string with a maximum of
20 characters.

Use to identify the actual database that the application is accessing. The mapping
between this name and Database Name could be done by using the database
directories at the client node and the database manager server node.

Because different database aliases can have different authentication types, this
attribute can also help you determine the authentication type.

Chapter 5. Attributes reference 107

 client db alias (Unicode) The alias defined within the database manager where the
database connection request originated. The value format is a text string with a
maximum of 60 bytes. This attribute is globalized.

Use to identify the actual database that the application is accessing. The mapping
between this name and Database Name could be done by using the database
directories at the client node and the database manager server node.

Because different database aliases can have different authentication types, this
attribute can also help you determine the authentication type.

codepage id The codepage or CCSID at the node where the application started.
The value format is integer. For snapshot monitor data, this is the code page at the
node where the monitored application started.

Use this attribute to analyze problems for remote applications. With this
information, you can ensure that data conversion is supported between the
application code page and the database code page (or for DRDA host databases,
the host CCSID).

lock escalation An indicator of whether a lock request was made as part of a lock
escalation. The value format is a text string with a maximum of 4 characters. The
valid values are: No, Yes.

Use this attribute to better understand the cause of deadlocks. If deadlocks occur
that involve applications doing lock escalation, you want to increase the amount of
lock memory or change the percentage of locks that any one application can
request.

lock mode The type of lock being held. The value format is a text string with a
maximum of 32 characters. The mode can help you determine the source of
contention for resources. This attribute indicates one of the following, depending
on the type of monitor information being examined:
v The type of lock another application holds on the object that this application is
waiting to lock (for application-monitoring and deadlock-monitoring levels)
v The type of lock held on the object by this application (for object-lock levels).

Valid values are:

Exclusive Lock
Intent None
Intn Excl Lock
Intn Share Lock
No Lock
Share Lock
Shr Int Ex Lck
Super Excl Lck
U-Lock
Unknown

lock object type The type of object against which the application holds a lock (for
object-lock-level information), or the type of object for which the application is
waiting to obtain a lock (for application-level and deadlock-level information). The
value format is a text string with a maximum of 16 characters. Use this attribute to
help you determine the source of contention for resources.

Valid values are:

108 IBM Tivoli Monitoring for Databases: DB2 Agent: User’s Guide
No Lock
UNKNOWN
INTERNAL
ROW
TABLE
TABLESPACE

lock wait start time The date and time that the application started waiting to
obtain a lock on the object that is currently locked by another application.

lock wait time The total elapsed time waited for a lock, in seconds. The value
format is integer.

At the database level, this is the total amount of elapsed time that all applications
were waiting for a lock within this database. At the application-connection and
transaction levels, this is the total amount of elapsed time that this connection or
transaction has waited for a lock to be granted. This attribute be used in
conjunction with the Lock Waits attribute to calculate the average wait time for a
lock. This calculation can be performed at either the database or the
application-connection level.

locks held The number of locks currently held. The value format is integer. If the
monitor information is at the database level, this is the total number of locks
currently held by all applications in the database. If it is at the application level,
this is the total number of locks currently held by all agents for the application.
Usage of this attribute depends on the level of information being returned from the
database system monitor.

node name Indicates the originating name. Value format is a simple text string
with a maximum of 32 characters.

snapshot time The date and time when the database system monitor information
was collected.

status change time The date and time the application entered its current status.

table name The name of the table. The value format is a text string with a
maximum of 20 characters. Along with Table Schema attribute, use this attribute to
help you determine the source of resource contention.

table name (Unicode) The name of the table. The value format is a text string with
a maximum of 60 bytes. This attribute is globalized. Along with Table Schema
attribute, use this attribute to help you determine the source of resource
contention.

table schema The schema of the table. The value format is a text string with a
maximum of 20 characters. Along with Table Name attribute, use this attribute to
help you determine the source of resource contention.

table schema (Unicode) The schema of the table. The value format is a text string
with a maximum of 60 bytes. This attribute is globalized. Along with Table Name
attribute, use this attribute to help you determine the source of resource
contention.

Chapter 5. Attributes reference 109

 tablespace name The name of the table space against which the application
currently holds a lock. The value format is a text string with a maximum of 20
characters. Use this attribute to help you determine the source of resource
contention.

tablespace name (Unicode) The name of the table space against which the
application currently holds a lock. The value format is a text string with a
maximum of 60 bytes. This attribute is globalized. Use this attribute to help you
determine the source of resource contention.

System Overview attributes

Use the System Overview attributes to monitor general information about the DB2
subsystem in your environment. These attributes provide data on DB2 status, DB2
start time, local and remote connections, sorting, and agent-related high-water
marks.

Unless otherwise noted, integer values are 32-bit or 0 to 4 294 967 296.

agentpri The current value of the priority of agents. The value format is integer.

This value is the priority that the operating system scheduler gives to agent and
other database manager instance processes and threads. This priority determines
how the operating system gives CPU time to the DB2 processes and threads
relative to the other processes and threads running on the system. A value of -1
indicates that no special action is taken and the operating system schedules the
database manager in the normal way that it schedules all processes and threads.
Any other value indicates that the database manager creates its processes and
threads with a static priority set to this value.

agents created empty pool The number of agents created because the agent pool
was empty. The value format is integer. It includes the number of agents started at
DB2 start up. In conjunction with Agents Assigned From Pool, you can calculate
the ratio of Agents Created Empty Pool attribute to Agents From Pool attribute.
See Agents From Pool for information on using this attribute.

agents created empty pool ratio The percentage of agents that are created because
the pool is empty. The value format is integer. This ratio is calculated by dividing
the value of Agents Created Empty Pool attribute by the value of Agents From
Pool attribute. Use this attribute to evaluate how often an agent must be created
because the pool is empty.

agents from pool The number of agents assigned from the pool. The value format
is integer. Use this attribute with Agents Created Empty Pool attribute to
determine how often an agent must be created because the pool is empty.

agents registered The number of agents registered in the database manager

instance that is being monitored (coordinator agents and subagents). The value
format is integer. Use this attribute to help you evaluate your setting for the
MAXAGENTS configuration parameter.

agents registered top The maximum number of agents that the database manager
has ever registered, at the same time, since it was started (coordinator agents and
subagents). The value format is integer. Use this attribute to evaluate the setting of
the MAXAGENTS configuration parameter. The number of agents registered at the
time the snapshot was taken is recorded by the Agents Registered attribute.

110 IBM Tivoli Monitoring for Databases: DB2 Agent: User’s Guide
agents stolen The number of times that agents are stolen from an application. The
value format is integer. Agents are stolen when an idle agent associated with an
application is reassigned to work on a different application. Use this attribute in
conjunction with Maximum Number of Associated Agents attribute to evaluate the
load that this application places on the system

agents waiting on token The number of agents waiting for a token so they can
execute a transaction in the database manager. The value format is integer.

Use this attribute to evaluate your setting for the MAXCAGENTS configuration
parameter. Each application has a dedicated coordinator agent to process database
requests (transactions) within the database manager. Each agent must have a token
in order to execute a transaction. The maximum number of coordinator agents is
limited by the MAXCAGENTS configuration parameter.

agents waiting on token pct The percentage of agents waiting on a token. The
value format is integer. The percentage is calculated by dividing the value of
Agents Waiting on Token attribute by the number of local applications that are
currently connected to a database (Local Cons attribute). Use this attribute to
assess the number of agents.

agents waiting top The highest number of agents waiting on a token, at the same
time, since the database manager was started. The value format is integer.

Use this attribute to evaluate the setting of the MAXCAGENTS configuration

parameter. In contrast, the Agents Waiting on Token attribute records the number
of agents waiting for a token at the time the snapshot was taken.

aslheapsz The current value (in units of 4-KB pages) of the application support
layer heap size. The value format is integer.

This value is the amount of memory that is allocated for the application support
layer heap. This heap is used as a communication buffer between the local
application and its associated agent. In addition, this value is used to determine
the input and output block size when a blocking cursor is opened.

buff free The number of Fast Communication Manager (FCM) buffers that are free
in the partitioned database server during the monitoring interval. The value format
is integer.

Use the returned value to determine the current buffer pool utilization. Use this
information to refine the configuration of the number of FCM buffers.

buff free bottom The minimum number of free FCM buffers. The value format is
integer.

buff max used pct The percentage of maximum FCM buffers used during
processing within the partitioned database server. The value format is integer.

If the percentage of maximum FCM buffers used is high compared to normal

operating levels, you can increase the number of FCM buffers; if the percentage is
low compared to normal operating levels, you can decrease the value.

buf used pct The percentage of FCM buffers that are used within the partitioned
database server during the monitoring interval. The value format is integer.

Chapter 5. Attributes reference 111

 If the percentage of FCM buffers used is high compared to normal operating
levels, you can adjust the number of FCM buffers.

ce free The number of connection entries that are free in the partitioned database
server during the monitoring interval. The value format is integer.

Use the returned value to help determine the current connection entry utilization.

ce free bottom The minimum number of free connection entries. The value format
is integer.

ce max used pct The maximum percentage of FCM connection entries used during
processing within the partitioned database server. The value format is integer.

If the percentage of maximum FCM connection entries used is high compared to

normal operating levels, you can increase the number of FCM connections; if the
percentage is low compared to normal operating levels, you can decrease the
value.

ce used pct The percentage of FCM connection entries that are used within the
partitioned database server during the monitoring interval. The value format is
integer.

If the percentage of FCM connection entries used is high compared to normal

operating levels, you can increase the number of FCM connections; if the
percentage is low compared to normal operating levels, you can decrease the
value.

comm private mem The amount (in bytes) of private memory that the instance of
the database manager currently has committed at the time of the snapshot. The
value format is integer. Use this attribute to assess the MIN_PRIV_MEM
configuration parameter to ensure that enough private memory is available. This
attribute is returned for all platforms, but tuning can be accomplished only on
platforms where DB2 uses threads (such as OS/2® and Windows NT®).

con local dbases The number of local databases with current connections to the
monitored DB2 instance. The value format is integer. This value gives an indication
of how many database information records to expect when gathering data at the
database level. The applications can be running locally or remotely, and might be
executing a unit of work within the database manager.

connection status The status of the communication connection between the

database partition specified by the DB2 Node Number variable and the database
partition where this monitor executes. The value format is integer.

Two nodes can be active, but the connection between them will remain inactive
unless there is active communication between them.

The connection values are as shown:

0 No current connection
1 Active connection
2 Congested connection

112 IBM Tivoli Monitoring for Databases: DB2 Agent: User’s Guide
coord agents top The maximum number of coordinating agents working at one
time. The value format is integer. The MAXCAGENTS configuration parameter
determines the number of coordinating agents that can be executing concurrently.
If the peak number of coordinating agents results in a workload that is too high for
this node, you can reduce the MAXCAGENTS configuration parameter.

db2 avail The amount of time (in seconds) the instance has been available since a
DB2START command was issued. The value format is integer. The value is derived
through this formula: snapshot time - db2start time.

db2start time The date and time that the database manager was started using the
DB2START command.

db2 status The current status of the DB2 instance. Valid values are string, up to 32
characters, as shown:
Active The DB2 instance is currently active.
Quiesce Pending
A quiesce is pending for the DB2 instance.
Quiesced
The DB2 instance has been quiesced.
Unknown
The status is unknown.

dbpg node status The list of failing nodes in the partition group. The list consists
of integers where each integer represents a failed local node.

Depending on the actual partitions defined in the database partition group, it is

not necessarily true that all the nodes defined in the parallel environment will be
examined. To ensure that all the nodes in the partitioned environment are
examined, define a partition group that contains at least one database partition
from each of the nodes in the partitioned environment.

fcm num anchors The number of FCM message anchors for the DB2 instance
during the monitoring interval. The value format is integer.

Agents use the message anchors to send messages among themselves.

fcm num buffers The number of buffers that are used for internal communications
(messages) among the nodes and within the nodes in a DB2 instance during the
monitoring interval. The value format is integer.

You might need to increase the value of this parameter if you have either of the
following conditions: multiple logical nodes on a processor, or too many users,
nodes, or complex applications that exceed the buffer limit.

fcm num connect The number of FCM connection entries for the DB2 instance
during the monitoring interval. The value format is integer. Agents use connection
entries to pass data among themselves.

Use the results from the fcm_num_rqb attribute to help you refine the
fcm_num_connect attribute.

fcm num rqb The number of FCM request blocks for the DB2 instance during the
monitoring interval. The value format is integer.

Chapter 5. Attributes reference 113

 Request blocks are the media through which information is passed between the
FCM daemon and an agent. The requirement for request blocks will vary according
to the number of users on the system, the number of database partition servers in
the system, and the complexity of queries that are run.

gw cons wait client For host databases being handled by the DB2 Connect
gateway, the current number of connections that are waiting for the client to send a
request. The value format is integer.

Because this value can change frequently, take samples at regular intervals over an
extended period in order to obtain a realistic view of gateway usage.

gw cons wait host For host databases being handled by the DB2 Connect gateway,
the current number of connections that are waiting for a reply from the host. The
value format is integer.

Because this value can change frequently, take samples at regular intervals over an
extended period to obtain a realistic view of gateway usage.

gw cur cons The current number of connections to host databases being handled
by the DB2 Connect gateway. The value format is integer. Use this attribute to help
you understand the level of activity at the DB2 Connect gateway and the
associated use of system resources.

gw total cons The total number of connections attempted from the DB2 Connect
gateway since the last db2start command or the last reset. The value format is
integer. Use this attribute to help you understand the level of activity at the DB2
Connect gateway and the associated use of system resources.

idle agents The number of agents in the agent pool that are currently unassigned
to an application. The value format is integer. Use this attribute to set the
NUM_POOLAGENTS configuration parameter. By having idle agents available to
satisfy requests for agents, you can improve performance.

instance name (Unicode) The name of the monitored DB2 instance. The valid
format is text string with a maximum of 60 bytes. This attribute is globalized.

last reset The date and time that the monitor counters were reset for the
application requesting the snapshot.

local cons The number of local applications that are currently connected to a
database within the database manager instance being monitored. The value format
is integer.

With this number, you can determine the level of concurrent processing occurring
in the database manager. This value changes frequently. As a result, you need to
sample the data at specific intervals over an extended period of time to get a
realistic view of system usage. This number only includes applications that were
initiated from the same instance as the database manager. The applications are
connected, but or not be executing a unit of work in the database.

local cons in exec The number of local applications that are currently connected to
a database within the database manager instance being monitored and are
currently processing a unit of work.

114 IBM Tivoli Monitoring for Databases: DB2 Agent: User’s Guide
With this number, you can determine the level of concurrent processing occurring
in the database manager. This value changes frequently. Sample the data at specific
intervals over an extended period of time to get a realistic view of system usage.
This number only includes applications that were initiated from the same instance
as the database manager.

ma free bottom The minimum number of free message anchors. The value format
is integer.

ma max used pct The maximum number of message anchors used as a percentage.
The value format is integer.

max agent overflows The number of attempts to create a new agent when the
MAXAGENTS configuration parameter had already been reached. The value
format is integer. If requests to create new agents are received after reaching the
MAXAGENTS configuration parameter, the workload for this node be too high.

maxagents The current value of the maximum number of existing agents. The
value format is integer.

This value is the maximum number of database manager agents available at any
given time to accept application requests. It limits the total number of applications
that can connect to all databases in the DB2 instance at a given time.

The value of maxagents is the sum of the values for maxappls in each database
that is allowed to be accessed concurrently. Increasing maxagents can increase
resource use because resources for each agent are allocated when the DB2 instance
is started.

maxcagents The maximum number of database manager coordinator agents that

can concurrently execute a database manager transaction in the DB2 instance
during the monitoring interval. The value format is integer.

When this monitor is used with DB2, the maxcagents value is the default. This
value is the maximum number of database manager agents that can be
concurrently executing a database manager transaction. Use the maxcagents
attributes to control the load on the system during periods of high simultaneous
application activity.

A value of -1 indicates that the limit is equal to the maximum number of agents
(maxagents). The maxcagents parameter does not limit the number of applications
that can have connections to the database.

maxcoordagents The maximum number of database manager coordinating agents

that can exist on a server in a partitioned or nonpartitioned database environment.
The value format is integer.

One coordinating agent is acquired for each local or remote application that
connects to a database or attaches to an instance. Requests that require an instance
attachment include CREATE DATABASE, DROP DATABASE, and Database System
Monitor commands.

mon heap sz The current value (in units of 4-KB pages) of the database system
monitor heap size. The value format is integer.

Chapter 5. Attributes reference 115

 This value is the amount of memory that is allocated for database system monitor
data. A value of zero prevents the database manager from collecting database
system monitor data.

node name Indicates the originating name. Value format is a simple text string
with a maximum of 32 characters.

piped sort hit ratio pct for int The value format is integer. The piped sort hit ratio
(as a percentage) for the last monitoring interval. The piped sort hit ratio is the
ratio of piped sorts accepted to piped sorts requested.

Use the returned value to help you improve sort performance. Every active sort on
the system allocates memory, which can result in the sorting process taking up too
much of the available system memory. Piped sorts help reduce disk input and
output, so increasing the number of accepted piped sorts can improve
performance.

When the returned value is low compared to normal operating levels, you can
improve performance by adjusting either the sort heap size (sortheap) or sort heap
threshold (sheapthres) configuration parameter, or by adjusting both. If piped sorts
are being rejected (resulting in a low-percentage piped sort hit ratio), consider
decreasing your sort heap or increasing your sort heap threshold. A piped sort is
not accepted if the sort heap threshold will be exceeded when the sort heap is
allocated for that piped sort.

Be aware that adjusting sortheap or sheapthres can have consequences: Decreasing

sortheap can result in the need for extra sort merge phases that could slow down
sorts. Increasing sheapthres can result in more memory allocated for sorting,
causing the paging of memory to disk and slowing overall system performance.

piped sorts accepted The number of piped sorts that have been accepted. The
value format is integer. When the number of accepted piped sorts is low compared
to the number requested, you can improve sort performance by adjusting one or
both of the following configuration parameters: SORTHEAP or SHEAPTHRES. If
piped sorts are being rejected, consider decreasing your sort heap or increasing
your sort heap threshold. Be aware of the possible implications of these options.

If you increase the sort heap threshold, more memory remain allocated for sorting.
This could cause the paging of memory to disk. If you decrease the sort heap, an
extra merge phase (which could slow down the sort) be required.

piped sorts accepted pct The percentage of piped sorts that have been accepted.
The value format is integer. The percentage is calculated by dividing the value of
Piped Sorts Accepted attribute by the value of Piped Sorts Requested attribute. Use
this attribute to determine whether the value of Piped Sorts Accepted attribute is
in an acceptable range.

piped sorts rejected for int The total number of piped sorts that were rejected
during the monitoring interval. The value format is integer.

In the return phase of sorting, if the sorted information can return directly through
the sort heap, it is a piped sort. However, even if the optimizer requests a piped
sort, this request is rejected at run time if the total amount of sort heap memory
for all sorts on the database is close to exceeding the sheapthres value.

116 IBM Tivoli Monitoring for Databases: DB2 Agent: User’s Guide
If this returned value is high compared to the total number of sorts requested,
consider decreasing your sort heap (using the sortheap configuration parameter) or
increasing your sort heap threshold (using the sheapthres configuration parameter).
However, be aware of the implications of these options. If you increase the sort
heap threshold, more memory can remain allocated for sorting, causing the paging
of memory to disk. If you decrease the sort heap, you can require an extra merge
phase that might slow down the sort.

piped sorts rejected pct for int The percentage of piped sort requests that were
rejected for the DB2 instance during the monitoring interval. The value format is
integer.

In the sort return phase, if the sorted information can return directly through the
sort heap, it is a piped sort. However, even if the optimizer requests a piped sort,
this request is rejected at run time if the total amount of sort heap memory for all
sorts on the database is close to exceeding the sheapthres value.

If this returned value is high compared to normal operating levels, consider

decreasing your sort heap (using the sortheap configuration parameter) or
increasing your sort heap threshold (using the sheapthres configuration parameter).
However, be aware of the implications of these options. If you increase the sort
heap threshold, more memory can remain allocated for sorting, causing the paging
of memory to disk. If you decrease the sort heap, you can require an extra merge
phase that might slow down the sort.

piped sorts requested The number of piped sorts that have been requested. The
value format is integer.

Because piped sorts reduce disk I/O, allowing more piped sorts can improve the
performance of sort operations and possibly the performance of the overall system.
A piped sort is not accepted if the sort heap threshold will be exceeded by
allocating the requested sort heap.

See Piped Sorts Accepted for more information if piped sorts are being rejected.
The SQL EXPLAIN output shows whether the optimizer requested a piped sort.

post threshold hash joins The total number of times that a hash join heap request
was limited because of concurrent use of shared or private sort heap space. The
value format is integer. If this value is large (for example, greater than 5% of Hash
Join Overflows), consider increasing the sort heap threshold.

post threshold sorts The number of sorts that have requested heaps after reaching
the sort heap threshold. The value format is integer. By modifying the sort heap
threshold and sort heap size configuration parameters, you can improve the
performance of sort operations and the overall system. If this attribute’s value is
high, you can do one of the following:
v Increase the sort heap threshold (sheapthres).
v Adjust applications to use fewer or smaller sorts via SQL query changes.

prdid The product and version that is running on the DB2 instance. Valid format is
a string up to 20 characters in format PPPVVRRM:
PPP Is SQL
VV Identifies a 2-digit version number (with high-order 0 in the case of a
1-digit version)

Chapter 5. Attributes reference 117

 RR Identifies a 2-digit release number (with high-order 0 in the case of a
1-digit release)
M Identifies a 1-digit modification level

query heap sz The maximum amount of memory that could be allocated for the
query heap within the DB2 instance during the monitoring interval. The value
format is integer.

Use a query heap to store each query in the agent’s private memory. Use the
results from the aslheapsz attribute to refine the query heap size.

rb free The number of request blocks that are free in the partitioned database
server during the monitoring interval. The value format is integer.

Use the returned value in conjunction with the fcm_num_rqb attribute to

determine the current request block utilization. You can use this information to
refine fcm_num_rqb.

rb free bottom The minimum number of free request blocks. The value format is
integer.

rb max used pct The percentage of maximum FCM request blocks used during
processing within the partitioned database server. The value format is integer.

If the percentage of maximum FCM request blocks used is high compared to

normal operating levels, you can adjust the fcm_num_rqb attribute.

rb used pct The percentage of FCM request blocks used within the partitioned
database server during the monitoring interval. The value format is integer.

If the percentage of FCM request blocks used is high compared to normal

operating levels, you can adjust the fcm_num_rqb attribute.

rem cons in The current number of connections initiated from remote clients to the
instance of the database manager that is being monitored. The value format is
integer.

This attribute shows the number of connections from remote clients to databases in
this instance. This value changes frequently. Sample the data at specific intervals
over an extended period of time to get a realistic view of system usage. This
number does not include applications that were initiated from the same instance as
the database manager.

rem cons in exec The number of remote applications currently connected to a

database and currently processing a unit of work within the database manager
instance being monitored. The value format is integer.

With this number, you can determine the level of concurrent processing occurring
on the database manager. This value changes frequently. Sample the data at
specific intervals over an extended period of time to get a realistic view of system
usage. This number does not include applications that were initiated from the
same instance as the database manager.

rqrioblk The current value (in byte units) of the client input and output block size.
The value format is integer.

118 IBM Tivoli Monitoring for Databases: DB2 Agent: User’s Guide
This value is the amount of memory that is allocated for the communication buffer
between remote applications and their database agents on the database server.
When a database client requests a connection to a remote database, this
communication buffer is allocated on the client. On the database server, a
communication buffer of 32767 bytes is initially allocated, until a connection is
established and the server can determine the value of rqrioblk at the client. In
addition to this communication buffer, this parameter is also used to determine the
input and output block size at the database client when a blocking cursor is
opened.

server db2 type The type of database manager being monitored. The value format
is a text string with a maximum of 32 characters. Valid values are as shown:
Client/Server
Client with local databases
Host Database Server
MPP
Requestor
Satellite
Standalone
UNKNOWN

sheapthres The current value (in units of 4-KB pages) of the sort heap threshold.
The value format is integer.

This value is the maximum amount of memory that the database manager allocates
for piped sorts. Piped sorts perform better than non-piped sorts and tend to be
used more often. However, their use can impact performance. The value for
sheapthres needs to be at least two times the largest sortheap defined for any
database within the instance.

snapshot time The date and time when the database system monitor information
was collected.

sort heap allocated The total number of allocated pages of sort heap space for all
sorts at the level chosen (database manager or database) and at the time the
snapshot was taken. The value format is integer. Memory estimates normally do
not include sort heap space. If excessive sorting occurs, add the extra memory
(used for the sort heap) to the base memory requirements for running the database
manager. Generally, the larger the sort heap, the more efficient the sort.
Appropriate use of indexes can reduce the amount of sorting required.

total buffers rcvd The total number of FCM buffers received by the database node
where this monitor executes. The value format is integer. The database node is
specified in the DB2_node_number variable.

Use the returned value to measure the level of traffic between the node where this
monitor executes and another node. If the total number of FCM buffers received
from the other node is high compared to normal operating levels, you can
redistribute the database or move tables to reduce the internode traffic.

total buffers sent The total number of FCM buffers sent from the database node
where this monitor executes to the node specified. The value format is integer.

Use the returned value to measure the level of traffic between the current node
where this monitor executes and the node specified. If the total number of FCM
buffers sent to the other node is high compared to normal operating levels, you
can redistribute the database or move tables to reduce the internode traffic.

Chapter 5. Attributes reference 119

 version The version of the server that is returning the data (for example: 6.1 or
7.1). The value format is a text string with a maximum of 8 characters. The data
structures used by the monitor change between releases. As a result, check the
version of the data stream to determine whether your applications can process the
data.

Tablespace attributes
The Tablespace attributes provides tablespace information for the monitored
database. You can use this information to monitor page size and usage
characteristics.

Unless otherwise noted, integer values are 32-bit or 0 to 4 294 967 296.

avg direct read time The time (in milliseconds) for performing the direct reads for
the tablespace. The value format is integer.

avg direct write time The time (in milliseconds) for performing the direct writes
for the tablespace. A high average time can indicate the existence of an input and
output conflict. The value format is integer.

avg pool io time The average time (in milliseconds) for performing buffer pool
input and output operations (reading or writing) for the tablespace. The value
format is integer.

A high average time can indicate the existence of an input and output conflict. In
this case, you might need to move data to a different device.

The returned value includes the time applied to asynchronous input and output
operations (which are performed by prefetchers and page cleaners).

avg pool read time The average time (in milliseconds) for processing read requests
that caused data or index pages to be physically read from disk to buffer pool for
the tablespace. The value format is integer.

A high average time generally indicates the existence of an input and output
conflict. In this case, you might need to move data to a different device.

The returned value includes the time applied to asynchronous read operations that
are performed by prefetchers.

avg pool write time The average time (in milliseconds) for processing write
requests that caused data or index pages to be physically written from buffer pool
to disk for the tablespace. The value format is integer.

A high average time generally indicates the existence of an input and output
conflict. In this case, you might need to move data to a different device.

The returned value includes the time applied to asynchronous write operations
that are performed by page cleaners.

avg sect read The average number of sectors that are read for this tablespace for
each direct read. The value format is integer.

Direct reads do not use the buffer pool, and so result in poor performance because
the data is physically read from disk each time. If you are using system monitors

120 IBM Tivoli Monitoring for Databases: DB2 Agent: User’s Guide
to track input and output for the device, this returned value helps you distinguish
database input and output from non-database input and output.

avg sect written The average number of sectors that are written for this tablespace
for each direct read. The value format is integer.

Direct writes do not use the buffer pool, and so result in poor performance because
the data is physically written from disk each time. If you are using system
monitors to track input and output for the device, this returned value helps you
distinguish database input and output from non-database input and output.

avg sync data read time The average time (in milliseconds) for synchronous data
reads for the tablespace.

Use the returned value to analyze the input and output work being performed for
the tablespace. Synchronous read operations are performed by database manager
agents. Asynchronous reads are performed by prefetchers, which read data pages
from disk into the buffer pool in anticipation of their use. The value format is
integer.

avg sync data write time The average time (in milliseconds) for synchronous data
writes for the tablespace. The value format is integer.

Use the returned value to analyze the input and output work being performed for
the tablespace. Synchronous write operations are performed by database manager
agents. Asynchronous writes are performed by page cleaners, which write out
changed pages to disk and free up space in the buffer pool.

avg sync io time The average time (in milliseconds) for synchronous input and
output operations for the tablespace. The value format is integer.

Use the returned value to analyze the input and output work being performed for
the tablespace. Synchronous input and output operations are performed by
database manager agents. Asynchronous input and output operations are
performed by prefetchers (reads) and page cleaners (writes). In general,
asynchronous input and output helps your applications run faster.

container name Indicates the location of the container. Value format is a simple
text string with a maximum of 256 characters.

container name (Unicode) Indicates the location of the container. Value format is a
simple text string with a maximum of 768 bytes. This attribute is globalized.

db name Represents the real name of the database for which information is
collected or to which the application is connected. This is the name the database
was given when created. Value format is a simple text string with a maximum of
32 characters.

db name (Unicode) Represents the real name of the database for which
information is collected or to which the application is connected. This is the name
the database was given when created. Value format is a simple text string with a
maximum of 96 bytes. This attribute is globalized.

direct read reqs The number of requests to perform a direct read from disk of one
or more sectors of data for the database. The value format is integer.

Chapter 5. Attributes reference 121

 The returned value is used in calculating the returned value for the average
number of sectors read per direct read for the tablespace. Direct reads are
performed in units, the smallest being a 512-byte sector. They are used while the
system is doing any of the following: reading LONG VARCHAR columns, reading
LOB columns, performing a backup.

direct read time The time (in milliseconds) for performing the direct reads for the
tablespace since the first connection. The value format is integer.

The returned value is used in calculations for the average direct read time (ms). A
high average time can indicate an input and output conflict.

direct reads The number of requests to perform a direct read from disk of one or
more sectors of data for the tablespace since the first connection. The value format
is integer.

direct write reqs The number of requests to perform a direct write to disk of one
or more sectors of data for the database. The value format is integer.

The returned value is used in calculating the returned value for the average
number of sectors written per direct write. Direct writes are performed in units, the
smallest being a 512-byte sector. They are used while the system is doing any of
the following: writing LONG VARCHAR columns, writing LOB columns,
performing a restore, or performing a load.

direct write time The time (in milliseconds) for performing the direct writes for the
tablespace since the first connection. The value format is integer.

The returned value is used in calculations for the average direct write time (ms). A
high average time can indicate an input and output conflict.

direct writes The number of direct writes to disk for the tablespace since the first
connection. The value format is integer.

The returned value is used in calculating the returned value for the average
number of sectors written per direct write. Direct writes are performed in units, the
smallest being a 512-byte sector. They are used while the system is doing any of
the following: writing LONG VARCHAR columns, writing LOB columns,
performing a restore, or performing a load.

estore rw ratio The ratio (as a percentage) of pages (data plus index) copied from
extended storage to pages copied to extended storage within the tablespace. The
value format is integer.

When a page is transferred from extended storage to the buffer pool, you save a
system input and output call. However, you still incur the cost of attaching to the
extended memory segment, copying the page, and detaching from the segment.

Use the returned value to determine if you would benefit from using extended
storage. The higher the ratio, the more likely you are to benefit. In general,
extended storage is particularly useful if input and output activity is very high on
your system.

extent size Indicates the extent size (the number of pages of table data that is
written to a container before data is written to the next container). The value
format is integer.

122 IBM Tivoli Monitoring for Databases: DB2 Agent: User’s Guide
files closed The total number of closed files for the tablespace since the first
database connection. The value format is integer.

free pages Represents the number of free pages associated with the database. The
value format is integer.

node name Indicates the originating name. Value format is a simple text string
with a maximum of 32 characters.

num containers Indicates the number of containers used. The value format is
integer.

object id Represents the identifier for the object. The value format is integer.

page size Indicates the page size. The value format is integer.

pending free pages Represents the number of pending free pages associated with
the database. The value format is integer.

pool async data read reqs The number of asynchronous data read requests. The
value format is integer.

pool async data reads The number of data pages read asynchronously into the
buffer pool for the tablespace. The value format is integer.

Compare the returned value with number of synchronous reads to gain insight
into how well the prefetchers are working.

pool async data writes The number of times a buffer pool data page was
physically written asynchronously to disk for the tablespace. The value format is
integer.

Compare the returned value with the number of synchronous writes to gain
insight into how well the page cleaners are working.

pool async index read reqs The number of asynchronous index read requests. The
value format is integer.

pool async index reads The number of index pages read asynchronously into the
buffer pool by a prefetcher within the tablespace. The value format is integer.

By comparing the ratio of asynchronous to synchronous reads, you can determine

how well the prefetchers are working. Asynchronous reads are performed by
database manager prefetchers.

pool async index writes The number of times a buffer pool index page was
written asynchronously to disk for the tablespace. The value format is integer.

Subtract the returned value from the buffer pool index writes to calculate the
number of synchronous index writes. By comparing the number of asynchronous
index writes to synchronous index writes, you can gain insight into how well the
buffer pool page cleaners are performing.

Chapter 5. Attributes reference 123

 By comparing the ratio of asynchronous to synchronous writes using the tbsp: # of
synchronous index reads monitor, you can determine how well the buffer pool
page cleaners are performing. This ratio can be helpful when you are refining the
num_iocleaners configuration parameter.

pool async read time The total time (in milliseconds) that database manager
prefetchers spent reading data into the buffer pool for the tablespace. The value
format is integer.

Compare the returned value to the synchronous read time to understand where
input and output time is being spent.

pool async write time The total time (in milliseconds) that database manager page
cleaners spent writing data or index pages from the buffer pool to disk for the
tablespace. The value format is integer.

Compare the returned value to the synchronous write time to understand where
input and output time is being spent.

pool data from estore The number of buffer pool data pages copied from extended
storage within the tablespace to the buffer pool. The value format is integer.

Required pages are copied from extended storage to the buffer pool if they are not
in the buffer pool but are in extended storage. This copying can incur the cost of
connecting to the shared memory segment but saves the cost of a disk read.

pool data l reads The number of logical read requests for data pages that went
through the buffer pool for the tablespace since the connection occurred. The value
format is integer.

The returned value includes requests for data that is already in the buffer pool or
read from disk into the buffer pool to fulfill the request.

pool data p reads The number of read requests requiring input and output to get
data pages into the buffer pool for the tablespace since the first connection. The
value format is integer.

pool data reads The number of read requests to get data pages into the buffer pool
for the tablespace. The value format is integer.

pool data to estore The number of buffer pool data pages copied to extended
storage for the tablespace. The value format is integer.

pool data writes The number of times that a buffer pool data page was physically
written to disk for the tablespace

pool hit pct The percent buffer pool hit ratio (data plus index). The value format is
integer.

pool idx hit pct for int The percent buffer pool index hit ratio for the monitoring
interval. The value format is integer.

pool index from estore The number of buffer pool index pages copied from
extended storage for the tablespace. The value format is integer.

124 IBM Tivoli Monitoring for Databases: DB2 Agent: User’s Guide
pool index l reads The number of logical read requests for index pages that went
through the buffer pool for the tablespace since the connection. The value format is
integer.

The returned value includes requests for index pages that are already in the buffer
pool or read from disk into the buffer pool to fulfill the request.

pool index p reads The number of physical read requests to get index pages into
the buffer pool for the tablespace. The value format is integer.

pool index to estore The number of buffer pool index pages copied to extended
storage within the tablespace. The value format is integer.

Pages are copied from the buffer pool to extended storage when they are selected
as victim pages. This copying is required to make space for new pages in the
buffer pool.

pool index writes The number of times that a buffer pool index page was
physically written to disk for the tablespace since the first connection. The value
format is integer.

If the returned value is high compared to the buffer pool index physical reads, you
can improve performance by increasing the available buffer pool space.

pool io per sec The rate (per second) for buffer pool input and output for the
tablespace. Buffer pool input and output includes all physical data and index
pages that go through the buffer pool when read or written. The value format is
integer.

Use the returned value to determine how efficient your data storage device is. A
low value when compared to normal operating levels can indicate the presence of
an input and output wait, which can indicate that you need to move data to a
different device.

pool read time The time (in milliseconds) spent reading data from the buffer pool
to disk for the tablespace since the first connection. The value format is integer.

pool sync idx reads The number of buffer pool synchronous index reads. The
value format is integer.

pool sync idx writes The number of buffer pool synchronous index writes. The
value format is integer.

pool sync reads The number of buffer pool synchronous reads. The value format is
integer.

pool sync writes The number of buffer pool synchronous writes. The value format
is integer.

pool write time The time (in milliseconds) spent writing data from the buffer pool
to disk for the tablespace since the first connection. The value format is integer.

prefetch pct for int The percentage of asynchronous read requests that were
satisfied for a tablespace during the last monitoring interval. The value format is
integer.

Chapter 5. Attributes reference 125

 prefetch reqs for int The number of prefetch requests for the tablespace during the
monitoring interval. The value format is integer.

prefetch size Indicates the prefetch size of the tablespace. The value format is
integer.

space used dms table pct The percentage of space used in the Database Managed
Space (DMS) tablespace. The value format is integer.

Use the returned value to determine if the tablespace needs more space.

space used sms table The number of bytes allocated to the System Managed Space
(SMS) tablespace. The value format is integer.

Use the returned value to determine whether the number of bytes used by the SMS
tablespace is excessive in relation to the file system on which the tablespace
resides.

sync index reads The number of physical index pages that were read
synchronously for the tablespace.

By comparing the ratio of asynchronous reads to synchronous reads, you can

determine how well the prefetchers are working.

sync index writes The number of physical index page write requests that were
performed synchronously for the tablespace.

By comparing the ratio of asynchronous writes to synchronous writes, you can

determine how well the buffer pool page cleaners are performing.

sync read time The time (in milliseconds) applied to synchronous reads for the
tablespace. The value format is integer.

Compare the returned value to the buffer pool async read time to understand
where input and output time for this tablespace is used.

sync reads The number of times that physical data pages were read synchronously
into the buffer pool for the tablespace.

Compare the returned value to the buffer pool async data reads to gain insight into
how well the prefetchers are working.

sync writes The number of times that data pages were written synchronously from
the buffer pool to disk for the tablespace.

Compare the returned value with the buffer pool async data writes to gain insight
into how well the page cleaners are working. Because synchronous writes are
performed by database agents, a high number of writes can slow performance.

sync write time The time (in milliseconds) spent synchronously writing data to
disk from the buffer pool for the tablespace. The value format is integer.

Compare the returned value to the value returned by the buffer pool async write
time to understand where input and output time for this tablespace is used.

tablespace id The identifier for the tablespace. The value format is integer.

126 IBM Tivoli Monitoring for Databases: DB2 Agent: User’s Guide
tablespace name The tablespace name of DB2. Value value is a simple text string
with a maximum of 32 characters.

tablespace name (Unicode) The tablespace name of DB2. Value is a simple text
string with a maximum of 96 bytes. This attribute is globalized.

tablespace type The tablespace type of DB2. Value value is a simple text string
with a maximum of 32 characters.

tbsp status The status of the tablespace. The value format is integer.

Value Tablespace State

0x0 Normal
0x1 Quiesced: SHARE
0x2 Quiesced: UPDATE
0x4 Quiesced: EXCLUSIVE
0x8 Load pending
0x10 Delete pending
0x20 Backup pending
0x40 Roll forward in progress
0x80 Roll forward pending
0x100 Restore pending
0x200 Disable pending
0x400 Reorganization in progress
0x800 Backup in progress
0x1000 Storage must be defined
0x2000 Restore in progress
0x4000 Offline and not accessible
0x8000 Drop pending
0x2000000 Storage be defined
0x4000000 Storage definition is in ’final’ state
0x8000000 Storage definition was changed prior to roll forward
0x10000000 DMS rebalancer is active
0x20000000 TBS deletion in progress
0x40000000 TBS creation in progress
0x8 For service use only

total direct io time The total time (in milliseconds) for direct reads to and writes
from the tablespace. The value format is integer.

total io pct The percentage total of I/O. The value format is integer.

total pages The number of total pages available associated with the database. The
value format is integer.

A high time can indicate the presence of an input and output conflict.

Chapter 5. Attributes reference 127

 total pool io time The total time (in milliseconds) for physical buffer pool input
and output operations (including both synchronous and asynchronous reads and
writes) for the tablespace. The value format is integer.

A high returned value (as compared to the total number of physical buffer pool
input and output operations) can indicate the presence of an input and output
wait, which could indicate that you need to move data to a different device.

total pool p read time The total physical read time. The value format is integer.

total pool p write time The total pool physical write time. The value format is
integer.

total sync io The total number of synchronous reads and writes for both data and
index pages for the tablespace. The value format is integer.

Use the returned value to gain insight into how well the prefetchers and page
cleaners are working. Because synchronous input and output operations are
performed by database agents, a high number of operations can slow performance.

total sync io time The total time (in milliseconds) for processing requests for
synchronous reads or writes within the tablespace. The value format is integer.

The returned value is the sum of the returned values from synchronous read time
synchronous write time monitors. This time is the amount of time that database
agents spend doing synchronous reads and writes for the tablespace.

usable pages Represents the number of usable pages associated with the database.
The value format is integer.

used pages Represents the total number of used pages. The value format is integer.

version Indicates the version number of DB2. The value format is integer.

Disk capacity planning for historical data

Disk capacity planning for a monitoring agent is a prediction of the amount of disk
space to be consumed for each attribute group whose historical data is being
collected. Required disk storage is an important factor to consider when you are
defining data collection rules and your strategy for historical data collection.

Calculate expected disk space consumption by multiplying the number of bytes

per instance by the expected number of instances, and then multiplying that
product by the number of samples. Table 8 on page 129 provides the following
information required to calculate disk space for the Monitoring Agent for DB2:
v DB table name is the table name as it would appear in the warehouse database, if
the attribute group is configured to be written to the warehouse.
v Bytes per instance (agent) is an estimate of the record length for each row or
instance written to the agent disk for historical data collection. This estimate can
be used for agent disk space planning purposes.
v Bytes per instance (warehouse) is an estimate of the record length for detailed
records written to the warehouse database, if the attribute group is configured to
be written to the warehouse. Detailed records are those that have been uploaded
from the agent for long-term historical data collection. This estimate can be used
for warehouse disk space planning purposes.

128 IBM Tivoli Monitoring for Databases: DB2 Agent: User’s Guide
 v Bytes per summarized instance (warehouse) is an estimate of the record length for
aggregate records written to the warehouse database, if the attribute group is
configured to be written to the warehouse. Aggregate records are created by the
Summarization agent for attribute groups that have been configured for
summarization. This estimate can be used for warehouse disk space planning
purposes.
v Expected number of instances is a guideline that can be different for each attribute
group, because it is the number of instances of data that the agent will return for
a given attribute group, and depends upon the application environment that is
being monitored. For example, if your attribute group is monitoring each
processor on your machine and you have a dual processor machine, the number
of instances is 2.
The IBM Tivoli Monitoring Installation and Setup Guide contains formulas that can
be used to estimate the amount of disk space used at the agent and in the
warehouse database for historical data collection of an attribute group.
Table 8. Capacity planning for historical data
Bytes per
Bytes per summarized
Attribute Bytes per instance instance Expected number of
Group DB table name instance (agent) (warehouse) (warehouse) instances
Application KUD2649700 3140 3255 4810 Number of applications
(group 00, using the DB2 instance
deprecated
application
attributes)
Application KUD2649900 3668 3783 5386 Number of applications
(group 00, using the DB2 instance
Unicode
application
attributes)
Application KUD2649800 308 316 818 Number of applications
(group 01) using the DB2 instance
Buffer Pool KUD4177600 1512 1560 2425 Number of buffer pools per
Data database (default of 1 )
Database KUD3437500 1820 1936 4007 Number of databases with
(group 00) the DB2 instance
Database KUD3437600 1508 1588 3473 Number of databases with
(group 01) the DB2 instance
Locking KUD5214100 988 1003 1094 Number of applications
Conflict using the DB2 instance
System KUD4238000 1560 1623 3229 One per DB2 instance
Overview
Table space KUDTABSPC 1668 1742 3237 Number of table spaces per
database (default of 4)

For more information about historical data collection, see the IBM Tivoli Monitoring
Administrator’s Guide.

Chapter 5. Attributes reference 129

130 IBM Tivoli Monitoring for Databases: DB2 Agent: User’s Guide
Chapter 6. Situations reference
This chapter contains an overview of situations, references for detailed information
about situations, and descriptions of the predefined situations included in this
monitoring agent.

About situations
A situation is a logical expression involving one or more system conditions.
Situations are used to monitor the condition of systems in your network. You can
manage situations from the Tivoli Enterprise Portal by using the Situation editor.

The IBM Tivoli Monitoring monitoring agents that you use to monitor your system
environment are shipped with a set of predefined situations that you can use as-is
or you can create new situations to meet your requirements. Predefined situations
contain attributes that check for system conditions common to many enterprises.

Using predefined situations can improve the speed with which you can begin
using the Monitoring Agent for DB2. You can examine and, if necessary, change the
conditions or values being monitored by a predefined situation to those best suited
to your enterprise.

Note: The predefined situations provided with this monitoring agent are not
read-only. Do not edit these situations and save over them. Software updates
will write over any of the changes that you make to these situations.
Instead, clone the situations that you want to change to suit your enterprise.

You can display predefined situations and create your own situations using the
Situation editor. The left frame of the Situation editor initially lists the situations
associated with the Navigator item that you selected. When you click a situation
name or create a new situation, the right frame opens with the following tabs:
Formula
Condition being tested
Distribution
List of managed systems (operating systems, subsystems, or applications)
to which the situation can be distributed.
Expert Advice
Comments and instructions to be read in the event workspace
Action
Command to be sent to the system
Until Duration of the situation

More information about situations

The IBM Tivoli Monitoring User’s Guide contains more information about predefined
and custom situations and how to use them to respond to alerts.

For a list of the predefined situations for this monitoring agent and a description
of each situation, refer to the Predefined situations section below and the
information in that section for each individual situation.

© Copyright IBM Corp. 2005 131

Predefined situations
This monitoring agent contains the following predefined situations, which are
organized alphabetically:
v UDB_Agent_WaitToken_High
v UDB_Agents_Stolen_High
v UDB_Appl_BP_Hit_Ratio_Low
v UDB_Appl_CatCache_Hit_Low
v UDB_Appl_HJoinOflws_High
v UDB_Appl_HJoinSmOflw_High
v UDB_Appl_Lock_Warning
v UDB_Appl_PkgCache_Hit_Low
v UDB_Appl_SQL_Fail_High
v UDB_Appl_Wait_Lock
v UDB_BP_DrtyPg_Steal_Clns
v UDB_BP_DrtyPg_thresh_Clns
v UDB_BP_Hit_Ratio_Low
v UDB_Buff_Max_Used_Pct_Crit
v UDB_Buff_Max_Used_Pct_Warn
v UDB_Buff_Used_Pct_Crit
v UDB_Buff_Used_Pct_Warn
v UDB_Ce_Max_Used_Pct_Crit
v UDB_Ce_Max_Used_Pct_Warn
v UDB_Ce_Used_Pct_Crit
v UDB_Ce_Used_Pct_Warn
v UDB_Database_Lock_Warning
v UDB_DB_BP_Hit_Ratio_Low
v UDB_DB_Cat_Cache_Hit_Ratio_Crit
v UDB_DB_Cat_Cache_Hit_Ratio_Warn
v UDB_DB_Cur_Cons_Pct_Crit
v UDB_DB_Cur_Cons_Pct_Warn
v UDB_DB_Dlk_Rb_Pct_For_Int_Crit
v UDB_DB_Dlk_Rb_Pct_For_Int_Warn
v UDB_DB_File_Closed_High
v UDB_DB_Int_Deadlock_Rb_Pct_Crit
v UDB_DB_Int_Deadlock_Rb_Pct_Warn
v UDB_DB_Invalid_Pkgs_Crit
v UDB_DB_Invalid_Pkgs_Warn
v UDB_DB_Invalid_Sys_Pkgs_Crit
v UDB_DB_Invalid_Sys_Pkgs_Warn
v UDB_DB_Invalid_Triggers_Crit
v UDB_DB_Invalid_Triggers_Warn
v UDB_DB_Lock_Waits_Pct_Crit
v UDB_DB_Lock_Waits_Pct_Warn
v UDB_DB_Pool_Hit_Idx_Pct_Crit

132 IBM Tivoli Monitoring for Databases: DB2 Agent: User’s Guide
 v UDB_DB_Pool_Hit_Idx_Pct_Warn
v UDB_DB_Pool_Hit_Ratio_Pct_Crit
v UDB_DB_Pool_Hit_Ratio_Pct_Warn
v UDB_DB_Pri_Log_Used_Pct_Crit
v UDB_DB_Pri_Log_Used_Pct_Warn
v UDB_DB_Sec_Log_Used_Pct_Crit
v UDB_DB_Sec_Log_Used_Pct_Warn
v UDB_DB_Sort_Overflow_High
v UDB_DB_SQL_Fail_High
v UDB_DB_SQL_Stmts_Fail_Pct_Crit
v UDB_DB_SQL_Stmts_Fail_Pct_Warn
v UDB_Ma_Max_Used_Pct_Crit
v UDB_Ma_Max_Used_Pct_Warn
v UDB_Max_Agent_Overflows_High
v UDB_Pip_Sort_Hit_Ratio_Pct_Crit
v UDB_Pip_Sort_Hit_Ratio_Pct_Warn
v UDB_Piped_Sorts_Rej_Pct_Crit
v UDB_Piped_Sorts_Rej_Pct_Warn
v UDB_Post_Threshold_Sorts_High
v UDB_Rb_Max_Used_Pct_Crit
v UDB_Rb_Max_Used_Pct_Warn
v UDB_Rb_Used_Pct_Crit
v UDB_Rb_Used_Pct_Warn
v UDB_Status_Warning
v UDB_TS_Sp_Used_DMS_Tab_Pct_Crit
v UDB_TS_Sp_Used_DMS_Tab_Pct_Warn
v UDB_TS_Status_Warn

The remaining sections of this chapter contain descriptions of each of these

predefined situations. The situations are organized alphabetically.

UDB_Agent_WaitToken_High situation
Issues a warning alert if the DB2 server experiences more than 20 agents waiting
for a token. This situation has the following formula:
agents_waiting_on_token GT 20

UDB_Agents_Stolen_High situation
Issues a warning alert if the DB2 server experiences more than 50 stolen agents.
This situation has the following formula:
agents_stolen GT 50

UDB_Appl_BP_Hit_Ratio_Low situation
(Deprecated) Issues a warning alert if an application experiences a buffer pool hit
ratio below 50%. This situation has the following formula:
pool_hit_ratio LT 50

Chapter 6. Situations reference 133

 UDB_Appl_CatCache_Hit_Low situation
(Deprecated) Issues a warning alert if an application experiences a catalog cache
hit ratio that is lower than 50%. This situation has the following formula:
cat_cache_hit_ratio LT 50

UDB_Appl_HJoinOflws_High situation
(Deprecated) Issues a warning alert if an application experiences more than 20
hash join overflows. This situation has the following formula:
hash_join_overflows GT 20

UDB_Appl_HJoinSmOflw_High situation
(Deprecated) Issues a warning alert if an application experiences more than 20%
hash join small overflows. This situation has the following formula:
hash_join_small_overflows GT 20

UDB_Appl_Lock_Warning situation
(Deprecated) Issues a warning alert if a monitored application experiences one or
more of the following conditions:
v More than 5 deadlocks
v More than 5 lock timeouts
v More than 20 lock waits

This situation has the following formula:

deadlocks GT 5

or
lock_timeouts GT 5

or
lockk_waits GT 20

UDB_Appl_PkgCache_Hit_Low situation
(Deprecated) Issues a warning alert if an application experiences a package cache
hit ratio that is lower than 50%. This situation has the following formula:
pkg_cache_hit_ratio LT 50

UDB_Appl_SQL_Fail_High situation
(Deprecated) Issues a warning alert if an application experiences a failure rate for
SQL statements that is greater than 50% . This situation has the following formula:
failed_sql_stmts_pct GT 50

UDB_Appl_Wait_Lock situation
(Deprecated) Issues a warning alert if an application is waiting for a lock. This
situation has the following formula:
agent_id_holding_lk NE 0

UDB_BP_DrtyPg_Steal_Clns situation
Issues a warning alert if a monitored database invokes an asynchronous page
cleaner more than 20% of the time. This situation has the following formula:
pool_drty_pg_steal_clns GT 20

134 IBM Tivoli Monitoring for Databases: DB2 Agent: User’s Guide
UDB_BP_DrtyPg_thresh_Clns situation
Issues a warning alert if a monitored database invokes an asynchronous page
cleaner more than 20% of the time because the buffer pool reached the dirty page
threshold criterion for the database. This situation has the following formula:
pool_drty_pg_thrsh_clns GT 20

UDB_BP_Hit_Ratio_Low situation
Issues a warning alert if the buffer pool hit atio below 50%. This situation has the
following formula:
pool_hit_ratio LT 50

UDB_Buff_Max_Used_Pct_Crit situation
Issues a critical alert if the percentage maximum FCM buffers used exceeds the
critical threshold. This situation has the following formula:
buff_max_used_pct GT 95

UDB_Buff_Max_Used_Pct_Warn situation
Issues a warning alert if the percentage maximum FCM buffers used exceeds the
warning threshold. This situation has the following formula:
buff_max_used_pct GT 80 and LT 96

UDB_Buff_Used_Pct_Crit situation
Issues a critical alert if the percentage of FCM buffers currently used exceeds the
critical threshold. This situation has the following formula:
buff_used_pct GT 95

UDB_Buff_Used_Pct_Warn situation
Issues a warning alert if the percentage of FCM buffers currently used exceeds the
warning threshold. This situation has the following formula:
buff_used_pct GT 80 and LT 96

UDB_Ce_Max_Used_Pct_Crit situation
Issues a critical alert if the percentage maximum FCM connection entries used
exceeds the critical threshold. This situation has the following formula:
ce_max_used_pct GT 95

UDB_Ce_Max_Used_Pct_Warn situation
Issues a warning alert if the percentage maximum FCM connection entries used
exceeds the warning threshold. This situation has the following formula:
ce_max_used_pct GT 80 and LT 96

UDB_Ce_Used_Pct_Crit situation
Issues a critical alert if the percentage of FCM connection entries currently used
exceeds the critical threshold. This situation has the following formula:
ce_used_pct GT 95

UDB_Ce_Used_Pct_Warn situation
Issues a warning alert if the percentage of FCM connection entries currently used
exceeds the warning threshold. This situation has the following formula:

Chapter 6. Situations reference 135

 ce_used_pct GT 80 and LT 96

UDB_Database_Lock_Warning situation
Issues a warning alert if the monitored database experiences one or more of the
following conditions:
v More than 10 deadlocks
v More than 10 lock timeouts
v More than 20 lock waits

This situation has the following formula:

deadlocks GT 10

or
lock_timeouts GT 10

or
lock_waits GT 20

UDB_DB_BP_Hit_Ratio_Low situation
Issues a warning alert if a database’s buffer pool hit ratio falls below 65%. This
situation has the following formula:
pool_hit_ratio LT 65

UDB_DB_Cat_Cache_Hit_Ratio_Crit situation
Issues a critical alert if the percentage catalog cache hit ratio drops below the
critical threshold. This situation has the following formula:
cat_cache_hit_ratio LT 80

UDB_DB_Cat_Cache_Hit_Ratio_Warn situation
Issues a warning alert if the percentage catalog cache hit ratio drops below the
warning threshold. This situation has the following formula:
cat_cache_hit_ratio LT 95 and GT 79

UDB_DB_Cur_Cons_Pct_Crit situation
Issues a critical alert if the percentage of connections used exceeds the critical
threshold. This situation has the following formula:
cur_cons_pct GT 95

UDB_DB_Cur_Cons_Pct_Warn situation
Issues a warning alert if the percentage of connections used exceeds the warning
threshold. This situation has the following formula:
cur_cons_pct GT 60 and LT 96

UDB_DB_Dlk_Rb_Pct_For_Int_Crit situation
Issues a critical alert if the internal deadlock rollbacks percent for interval exceeds
the critical threshold. This situation has the following formula:
Int_deadlock_rollbacks_pct_for_int GT 95

136 IBM Tivoli Monitoring for Databases: DB2 Agent: User’s Guide
UDB_DB_Dlk_Rb_Pct_For_Int_Warn situation
Issues a warning alert if the internal deadlock rollbacks percent for interval
exceeds the warning threshold. This situation has the following formula:
Int_deadlock_rollbacks_pct_for_int GT 80 and LT 96

UDB_DB_File_Closed_High situation
Issues a warning alert if the number of files closed for a databased exceeds 500
files. This situation has the following formula:
files_closed GT 500

UDB_DB_Int_Deadlock_Rb_Pct_Crit situation
Issues a critical alert if the percentage of internal rollbacks because of internal
deadlocks exceeds the critical threshold. This situation has the following formula:
Int_deadlock_rollbacks_pct GT 95

UDB_DB_Int_Deadlock_Rb_Pct_Warn situation
Issues a warning alert if the percentage of internal rollbacks because of internal
deadlocks exceeds the warning threshold. This situation has the following formula:
Int_deadlock_rollbacks_pct GT 70 and LT 96

UDB_DB_Invalid_Pkgs_Crit situation
Issues a critical alert if the number of all packages exceeds the critical threshold.
This situation has the following formula:
invalid_pkgs GT 20

UDB_DB_Invalid_Pkgs_Warn situation
Issues a warning alert if the number of all packages exceeds the warning threshold.
This situation has the following formula:
invalid_pkgs GT 1 and invalid_pkgs LT 21

UDB_DB_Invalid_Sys_Pkgs_Crit situation
Issues a critical alert if the number of invalid SYSTEM packages exceeds the critical
threshold. This situation has the following formula:
invalid_sys_pkgs GT 20

UDB_DB_Invalid_Sys_Pkgs_Warn situation
Issues a warning alert if the number of invalid SYSTEM packages exceeds the
warning threshold. This situation has the following formula:
invalid_sys_pkgs GT 1 and LT 21

UDB_DB_Invalid_Triggers_Crit situation
Issues a critical alert if the number of invalid triggers exceeds the critical threshold.
This situation has the following formula:
invalid_triggers GT 20

UDB_DB_Invalid_Triggers_Warn situation
Issues a warning alert if the number of invalid triggers exceeds the warning
threshold. This situation has the following formula:
invalid_triggers GT 1 and LT 21

Chapter 6. Situations reference 137

 UDB_DB_Lock_Waits_Pct_Crit situation
Issues a critical alert if the percentage application in lock wait exceeds the critical
threshold. This situation has the following formula:
lock_waits_pct GT 85

UDB_DB_Lock_Waits_Pct_Warn situation
Issues a warning alert if the percentage application in lock wait exceeds the
warning threshold. This situation has the following formula:
lock_waits_pct GT 80 and LT 86

UDB_DB_Pool_Hit_Idx_Pct_Crit situation
Issues a critical alert if the percentage buffer pool hit ratio (index) exceeds the
critical threshold. This situation has the following formula:
pool_hit_ratio_index_pct_for_int LT 80

UDB_DB_Pool_Hit_Idx_Pct_Warn situation
Issues a warning alert if the percentage buffer pool hit ratio (index) exceeds the
warning threshold. This situation has the following formula:
pool_hit_ratio_index_pct_for_int LT 95 and GT 79

UDB_DB_Pool_Hit_Ratio_Pct_Crit situation
Issues a critical alert if the percentage buffer pool hit ratio (data plus index)
exceeds the critical threshold. This situation has the following formula:
pool_hit_ratio_pct_for_int LT 80

UDB_DB_Pool_Hit_Ratio_Pct_Warn situation
Issues a warning alert if the percentage buffer pool hit ratio (data plus index)
exceeds the warning threshold. This situation has the following formula:
pool_hit_ratio_pct_for_int LT 95 and GT 79

UDB_DB_Pri_Log_Used_Pct_Crit situation
Issues a critical alert if the percentage used in primary log exceeds the critical
threshold. This situation has the following formula:
pri_log_used_pct GT 95

UDB_DB_Pri_Log_Used_Pct_Warn situation
Issues a warning alert if the percentage used in primary log exceeds the warning
threshold. This situation has the following formula:
pri_log_used_pct GT 80 and LT 96

UDB_DB_Sec_Log_Used_Pct_Crit situation
Issues a critical alert when the percentage used in secondary log exceeds the
critical threshold. This situation has the following formula:
sec_log_used_pct GT 95

UDB_DB_Sec_Log_Used_Pct_Warn situation
Issues a warning alert when the percentage used in secondary log exceeds the
warning threshold. This situation has the following formula:
sec_log_used_pct GT 80 and LT 96

138 IBM Tivoli Monitoring for Databases: DB2 Agent: User’s Guide
UDB_DB_Sort_Overflow_High situation
Issues a warning alert if a monitored database experiences more than 30% sort
overflows. This situation has the following formula:
sort_overflows_pct GT 30

UDB_DB_SQL_Fail_High situation
Issues a warning alert if a monitored database experiences more than 40% SQL
statement failures. This situation has the following formula:
sql_stmts_failed_pct GT 40

UDB_DB_SQL_Stmts_Fail_Pct_Crit situation
Issues a critical alert if a monitored database experiences more than 95% SQL
statement failures. This situation has the following formula:
sql_stmts_failed_pct GT 95

UDB_DB_SQL_Stmts_Fail_Pct_Warn situation
Issues a warning alert if a monitored database currently exceeds the warning
threshold for the percentage of SQL statement failures. This situation has the
following formula:
sql_stmts_failed_pct GT 80 and LT 96

UDB_Ma_Max_Used_Pct_Crit situation
Issues a critical alert if the percentage maximum FCM message anchors used
exceeds the critical threshold. This situation has the following formula:
ma_max_used_pct GT 95

UDB_Ma_Max_Used_Pct_Warn situation
Issues a warning alert if the percentage maximum FCM message anchors used
exceeds the warning threshold. This situation has the following formula:
ma_max_used_pct GT 80 and LT 96

UDB_Max_Agent_Overflows_High situation
Issues a warning alert if the UDB server experiences more than 50 max agent
overflows. This situation has the following formula:
max_agent_overflows GT 50

UDB_Pip_Sort_Hit_Ratio_Pct_Crit situation
Issues a critical alert if the percentage piped sort hits ratio exceeds the critical
threshold. This situation has the following formula:
piped_sort_hit_ratio_pct_for_int LT 80

UDB_Pip_Sort_Hit_Ratio_Pct_Warn situation
Issues a warning alert if the percentage piped sort hits ratio exceeds the warning
threshold. This situation has the following formula:
piped_sort_hit_ratio_pct_for_int LT 95 and GT 79

UDB_Piped_Sorts_Rej_Pct_Crit situation
Issues a critical alert if the percentage of piped sorts rejected exceeds the critical
threshold. This situation has the following formula:

Chapter 6. Situations reference 139

 piped_sorts_rejected_pct_for_int GT 95

UDB_Piped_Sorts_Rej_Pct_Warn situation
Issues a warning alert if the percentage of piped sorts rejected exceeds the warning
threshold. This situation has the following formula:
piped_sorts_rejected_pct_for_int GT 80 and LT 96

UDB_Post_Threshold_Sorts_High situation
Issues a warning alert if the UDB server experiences more than 20 post-threshold
sorts. This situation has the following formula:
UDB_Post_Threshold_Sorts GT 20

UDB_Rb_Max_Used_Pct_Crit situation
Issues a critical alert if the percentage maximum FCM request blocks used exceeds
the critical thresholds. This situation has the following formula:
rb_max_used_pct GT 95

UDB_Rb_Max_Used_Pct_Warn situation
Issues a warning alert if the percentage maximum FCM request blocks used
exceeds the warning thresholds. This situation has the following formula:
rb_max_used_pct GT 80 and LT 96

UDB_Rb_Used_Pct_Crit situation
Issues a critical alert if the percentage of FCM request blocks currently used
exceeds the critical threshold. This situation has the following formula:
rb_used_pct GT 95

UDB_Rb_Used_Pct_Warn situation
Issues a warning alert if the percentage of FCM request blocks currently used
exceeds the warning threshold. This situation has the following formula:
rb_used_pct GT 80 and LT 96

UDB_Status_Warning situation
Issues a warning alert if the status of the monitored UDB instance is other than
active status. This situation has the following formula:
db2_status NE active

UDB_TS_Sp_Used_DMS_Tab_Pct_Crit situation
Issues a critical alert if the percentage of spaced used in the DMS tablespace
exceeds the critical threshold. This situation has the following formula:
space_used_dms_table_pct GT 95

UDB_TS_Sp_Used_DMS_Tab_Pct_Warn situation
Issues a warning alert if the percentage of spaced used in the DMS tablespace
exceeds the warning threshold. This situation has the following formula:
space_used_dms_table_pct GT 80 and LT 96

UDB_TS_Status_Warn situation
Issues a warning alert if the tablespace status is not normal. This situation has the
following formula:
140 IBM Tivoli Monitoring for Databases: DB2 Agent: User’s Guide
tbsp_status GT 0

Chapter 6. Situations reference 141

142 IBM Tivoli Monitoring for Databases: DB2 Agent: User’s Guide
Chapter 7. Take Action commands reference
This chapter contains an overview of Take Action commands, references for
detailed information about Take Action commands, and descriptions of the Take
Action commands included in this monitoring agent.

About Take Action commands

Take Action commands can be run from the desktop or included in a situation or a
policy.

When included in a situation, the command executes when the situation becomes
true. A Take Action command in a situation is also referred to as reflex automation.
When you enable a Take Action command in a situation, you automate a response
to system conditions. For example, you can use a Take Action command to send a
command to restart a process on the managed system or to send a text message to
a cell phone.

Advanced automation uses policies to perform actions, schedule work, and

automate manual tasks. A policy comprises a series of automated steps called
activities that are connected to create a workflow. After an activity is completed,
Tivoli Enterprise Portal receives return code feedback, and advanced automation
logic responds with subsequent activities prescribed by the feedback.

More information about Take Action commands

For more information about working with Take Action commands, see the IBM
Tivoli Monitoring User’s Guide.

For a list of the Take Action commands for this monitoring agent and a description
of each command, refer to the Predefined Take Action commands section below
and the information in that section for each individual command.

Predefined Take Action commands

This monitoring agent contains the following Take Action commands:
v Backup Database
v Rebind All Packages
v Rebind Package
v Reorg Table
v Run Statistics
v Start DB2
v Stop DB2
v Update Database Configuration
v Update DB Manager Configuration

Note: The user ID of the process running the command is used for authentication
against the DB2 server.

© Copyright IBM Corp. 2005 143

 The remaining sections of this chapter contain descriptions of each of these Take
Action commands, which are listed alphabetically. The following information is
provided about each Take Action command:
Description
Which actions the command performs on the system to which it is sent
Arguments
List of arguments, if any, for the Take Action with a short description and
default value for each one
Destination systems
Where the command is to be executed: on the Managed System
(monitoring agent) where the agent resides or on the Managing System
(Tivoli Enterprise Monitoring Server) to which it is connected
Usage notes
Additional relevant notes for using the Take Actions

Backup Database action

Description
Creates a backup copy of a database.

Arguments
db_instance_name
Specifies the name of the DB2 instance. Enter a text string with a
maximum of 60 bytes.
db_name
Specifies the name of the database to be backed up. This is the name the
database was given when created. Enter a text string with a maximum of
60 bytes.
on_line_switch
Enter YES or NO to specify an online backup.
TSM_switch
Enter YES or NO to indicate whether to use Tivoli Storage Manager
managed output (formerly ADSM) for the backup.
TSM_Sessions
Specifies the number of I/O sessions to be used with the Tivoli Storage
Manager managed output. The default value is 2. If you are not using
managed output, specify any number (it will be ignored).
target_device
Specifies a directory or tape device name to which the backup is written. If
you specify a directory, you must enter the full path on which the
directory resides. The directory must already exist. The default for UNIX
hosts is the /TMPDIR directory. There is no default for Windows hosts.
number_of_buffers
Specifies the number of buffers to use during the backup process. A typical
value is 1.
buffer_size
Specifies the number of pages for the buffer that is used when building the
backup image. The minimum size is 16 pages and the default is 1024.

144 IBM Tivoli Monitoring for Databases: DB2 Agent: User’s Guide
 parallelism
Specifies the number of buffer manipulators to spawn during the restore
process. The default is 1.
libname
Specifies the name of the shared library that contains the vendor backup
and restore input and output functions to be used. This variable can
contain the full path and file name of the library. If the full path is not
specified, the path defaults to where the user exit program resides. This
variable is optional.
table_space
Specifies one or more tablespaces to back up. If no tablespace is specified,
all tablespaces are backed up. To specify more than one tablespace,
separate them using commas.

Destination systems
Managed system

Usage notes
Because this task can take longer to complete than the default time of 60 seconds,
you need to set the timeout value to a larger value, like 600 seconds.

If a database becomes damaged or corrupted, it can be returned to the state of the

backed up copy. If a successfully restored database was enabled for roll-forward
recovery at the time of the backup, it can be returned to the state that it was in
prior to the occurrence of damage. The backup might be directed to hard disk,
diskette, tape, Tivoli Storage Manager utility, or to other vendor products enabled
for DB2.

For more information, see the BACKUP DATABASE CLP command in the DB2
command reference information for the version of DB2 that you are using.

Rebind All Packages action

Description
Lets users re-create all packages stored in the database without needing a bind file.

Arguments
db_instance_name
Specifies the name of the DB2 instance. Enter a text string with a
maximum of 60 bytes.
db_name
Specifies the name of the database. This is the name the database was
given when created. Enter a text string with a maximum of 60 bytes.
log_file_name
Specifies the path (optional) and the file name (mandatory) to be used for
recording errors that result from the package revalidation procedure.

Destination systems
Managed system

Usage notes
The Rebind Package command does not automatically commit the transaction
following a successful rebind. This task uses the CLP REBIND command to

Chapter 7. Take Action commands reference 145

 attempt the revalidation of all packages in a database. You can allow package
revalidation to occur implicitly when the packages are first used.

For more information, see the db2rbind system command in the DB2 command
reference for the version of DB2 that you are using.

Rebind Package action

Description
Lets users re-create a package stored in the database without needing a bind file.

Arguments
db_instance_name
Specifies the name of the DB2 instance. Enter a text string with a
maximum of 60 bytes. .
db_name
Specifies the name of the database. This is the name the database was
given when created. Enter a text string with a maximum of 60 bytes.
package_name_schema
Specifies the qualifier for the package name. For example, if the qualified
name of a package is USERID.PACK1, the schema is USERID. If you
specify _default_, the current user ID is used.
package_name
Specifies the unqualified package name. For example, if the qualified name
of a package is USERID.PACK1, the name is PACK1.

Destination systems
Managed system

Usage notes
The Rebind Package command does not automatically commit the transaction
following a successful rebind. The user must explicitly commit the transaction. This
enables ″what if″ analysis, in which the user updates certain statistics, and then
tries to rebind the package to see what changes. It also permits multiple rebinds
within a unit of work.

The Rebind Package command commits the transaction if auto-commit is enabled.

For more information, see the REBIND CLP command in the DB2 command
reference for the version of DB2 that you are using.

Reorg Table action

Description
Reorganizes a table by reconstructing the rows to eliminate fragmented data and
by compacting information.

Arguments
db_instance_name
Specifies the name of the DB2 instance. Enter a text string with a
maximum of 60 bytes.

146 IBM Tivoli Monitoring for Databases: DB2 Agent: User’s Guide
 db_name
Specifies the name of the database. This is the name the database was
given when created. Enter a text string with a maximum of 60 bytes.
table_schema
Specifies the qualifier for the table name. For example, if the qualified
name of a table is USERID.TABLE1, the schema is USERID.
table_name
Specifies the unqualified table name. For example, if the qualified name of
a table is USERID.TABLE1, the name is TABLE1.
index_schema
Specifies the qualifier for the index name.
index_name
Specifies the unqualified index name.
table_space
Specifies the unqualified tablespace name.

Destination systems
Managed system

Usage notes
Because this task may take longer to complete than the 60-second default time, set
the timeout value to a larger value, such as 600 seconds.

Tables that have been modified so many times that data is fragmented and access
performance is noticeably slow are candidates for reorganization. Be sure to
complete all database operations and release all locks before running the Reorg
Table command. After reorganizing a table, use the Run Statistics Take Action
command to update the table statistics, and the Rebind Take Action command to
rebind the packages that use this table.

If the reorganization is not successful, do not delete temporary files. The database
manager uses these files to recover the database.

For more information, see the REORG TABLE CLP command in the DB2 command
reference information for the version of DB2 that you are using.

Run Statistics action

Description
Updates statistics about the physical characteristics of a table and the associated
indexes.

Arguments
db_instance_name
Specifies the name of the DB2 instance. Enter a text string with a
maximum of 60 bytes.
db_name
Specifies the name of the database. This is the name the database was
given when created. Enter a text string with a maximum of 60 bytes.
table_schema
Specifies the qualifier for the table name. For example, if the qualified
name of a table is USERID.TABLE1, the schema is USERID.

Chapter 7. Take Action commands reference 147

 table_name
Specifies the unqualified table name. For example, if the qualified name of
a table is USERID.TABLE1, the name is TABLE1.
dist_switch
Enter Yes or No to indicate whether you want distribution statistics
calculated.
index_switch
Enter one of the following responses:
Yes Updates statistics on the indexes.
No Does not update statistics on the indexes.
Index Specifies to update statistics on a specific index that is defined in
the index-schema and index-name variables.
index_only_switch
Enter Yes or No to indicate whether you want to update statistics on the
indexes only.
If the index argument is Yes and index_only_switch is No, statistics on
both the table and its indexes are updated.
If index is No and index_only_switch is No, statistics on the table only are
updated.
detail_switch
Enter Yes or No to indicate whether you want extended index statistics
calculated when you are updating statistics on the indexes.
share_switch
Enter Change or Reference to indicate how other users can access the table
while statistics are calculated:
Change
Other users can read from and write to the table while statistics are
calculated.
Reference
Other users can only read from the table while statistics are
calculated.
index_schema
Specifies the qualifier for the index name. This variable is required if you
specify the Index parameter for the index argument.
index_name
Specifies the unqualified index name. This variable is required if you
specify the Index parameter for the index argument.

Destination systems
Managed system

Usage notes
Run this command when a table has had many updates, or after reorganizing a
table. The statistics updated by this task include number of records, number of
pages, and average record length. The optimizer uses these statistics when
determining access paths to the data.

After the command is run, note the following:

v A COMMIT must be issued to release the locks.

148 IBM Tivoli Monitoring for Databases: DB2 Agent: User’s Guide
 v To allow new access plans to be generated, the packages that reference the target
table must be rebound.

After statistics are updated, you can create new access paths to the table by
rebinding the packages.

Collect statistics that apply only to the tables before you create indexes. By running
the table statistics first, you ensure that the indexes statistics are not overlaid by
table statistics.

For more information, see the RUNSTATS CLP command in the DB2 command
reference information for the version of DB2 that you are using.

Start DB2 action

Description
Starts the DB2 instance and allocates resources.

Arguments
db_instance_name
Specifies the name of the DB2 instance. Enter a text string with a
maximum of 60 bytes.

Destination systems
Managed system

Usage notes
After this task starts, the database manager instance runs until you stop it, even if
all programs that were using it have ended.

For more information when you are running UNIX, see the db2start system
command in the DB2 command reference for the version of DB2 that you are
using.

Stop DB2 action

Description
Stops the DB2 instance and releases resources.

Arguments
db_instance_name
Specifies the name of the DB2 instance. Enter a text string with a
maximum of 60 bytes.

Destination systems
Managed system

Usage notes
This task does not stop DB2 if any applications are connected to databases. If there
are no database connections, but there are instance attachments, the Stop DB2 Take
Action command forces the instance attachments and stops DB2.

Chapter 7. Take Action commands reference 149

 Stop DB2 can be run on a DB2 server or on other managed nodes. After DB2 stops,
a successful completion message is sent to the standard output device. If an error
occurs, processing stops, and an error message is sent to the standard output
device.

For more information when you are running UNIX, see the db2stop system
command in the DB2 command reference information for the version of DB2 that
you are using.

Update Database Configuration action

Description
Modifies individual entries in a specific database configuration file.

Arguments
db_instance_name
Specifies the name of the DB2 instance. Enter a text string with a
maximum of 60 bytes.
dbname
Specifies the name of the database. This is the name the database was
given when created. Enter a text string with a maximum of 60 bytes.
keyword1
Select the name of the database configuration parameter you want to
update. You can specify from one to five keyword-value pairs using
keyword1 to keyword5 arguments.
value1 Specifies the new value for the configuration parameter specified in the
corresponding keyword.

Destination systems
Managed system

Usage notes
Changes to the database configuration file become effective only after they are
loaded into memory. All applications must disconnect from the database before
changes can be loaded. If an error occurs, the database configuration file does not
change.

The database configuration file cannot be updated if the checksum is invalid. If

this problem occurs, the database must be restored to reset the database
configuration file.

For more information, see the UPDATE DATABASE CONFIGURATION CLP

command in the DB2 command reference information for the version of DB2 that
you are using.

Update DB Manager Configuration action

Description
Modifies individual entries in the database manager configuration file.

Arguments
db_instance_name
Specifies the name of the DB2 instance. Enter a text string with a
maximum of 60 bytes.

150 IBM Tivoli Monitoring for Databases: DB2 Agent: User’s Guide
keyword1
Specifies the name of the database manager configuration parameter you
want to update. You can specify from one to five keyword-value pairs
using keyword1 to keyword5 arguments.
value1 Specifies the new value for the configuration parameter specified in the
corresponding keyword.

Destination systems
Managed system

Usage notes
Changes to the database manager configuration file become effective only after
they are loaded into memory. For a server configuration, parameter changes are
loaded into memory when the Start DB2 Take Action command is run. For a client
configuration parameter, parameter changes are loaded into memory when the
application is restarted. If an error occurs, the database manager configuration file
does not change.

The database manager configuration file cannot be updated if the checksum is

invalid. If this error occurs, the database manager must be reinstalled to reset the
database manager configuration file.

For more information, see the UPDATE DATABASE MANAGER

CONFIGURATION CLP command in the DB2 command reference information for
the version of DB2 that you are using.

Chapter 7. Take Action commands reference 151

152 IBM Tivoli Monitoring for Databases: DB2 Agent: User’s Guide
Chapter 8. Policies reference
This chapter contains an overview of policies and references for detailed
information about policies.

About policies
Policies are an advanced automation technique for implementing more complex
workflow strategies than you can create through simple automation.

A policy is a set of automated system processes that can perform actions, schedule
work for users, or automate manual tasks. You use the Workflow Editor to design
policies. You control the order in which the policy executes a series of automated
steps, which are also called activities. Policies are connected to create a workflow.
After an activity is completed, Tivoli Enterprise Portal receives return code
feedback and advanced automation logic responds with subsequent activities
prescribed by the feedback.

Note: For monitoring agents that provide predefined policies, predefined policies
are not read-only. Do not edit these policies and save over them. Software
updates will write over any of the changes that you make to these policies.
Instead, clone the policies that you want to change to suit your enterprise.

More information about policies

For more information about working with policies, see the IBM Tivoli Monitoring
User’s Guide.

For information about using the Workflow Editor, see the IBM Tivoli Monitoring
Administrator’s Guide or the Tivoli Enterprise Portal online help.

For a list of the policies for this monitoring agent and a description of each policy,
refer to the Predefined policies section below and the information in that section
for each individual policy.

Predefined policies
There are no predefined policies for this monitoring agent.

© Copyright IBM Corp. 2005 153

154 IBM Tivoli Monitoring for Databases: DB2 Agent: User’s Guide
Appendix A. IBM Tivoli Monitoring for DB2 Workspaces
Additional information about the following predefined workspaces is provided
here:
v Application Top Ten Summary workspace
v Buffer Pool workspace
v Database Overview workspace

Application Top Ten Summary workspace

The following scenarios describe specific uses for the Application Top Summary
workspace and its linked workspaces.

Scenario 1: Appropriate value for LOCK TIMEOUT parameter

The LOCKTIMEOUT configuration parameter specifies the number of seconds that
an application waits to obtain a lock. By specifying a maximum value, you can
avoid global deadlocks for applications.

If you set this parameter to 0, an application does not wait for a lock. In this case,
if a lock is not available at the time of the request, the application receives
notification of this immediately.

If you set this parameter to -1, lock timeout detection is turned off. In this case an
application waits for a lock (if one is not available at the time of the request) until
the lock is granted or a deadlock occurs.

Set the value to detect quickly any waits that are occurring because of an abnormal
situation, such as a transaction that is stalled. Set the LOCKTIMEOUT parameter
high enough that valid lock requests do not time out because of peak workloads.
However, the LOCKTIMEOUT configuration parameter can be set too high, which
causes the system to experience too few lock timeouts. In this case, applications
might wait excessively to obtain a lock.

You can use the Application Top Summary workspace to help track the number of
times an application (connection) experienced a lock timeout. By using the
Application Lock Activity workspace, you can view the Lock_Timeout attribute in
addition to other lock-related attributes. This attribute indicates the number of
times that a request to lock an object timed out instead of being granted. A high
value for the Lock_Timeout attribute can be caused by:
v The value of the LOCKTIMEOUT configuration parameter being too low
v An application (transaction) that is holding one or more locks for an extended
period.
v A concurrency problem, that can be caused by lock escalations (from the
row-level to a table-level).

The Lock_Timeout attribute can help you adjust the setting for the
LOCKTIMEOUT configuration parameter. If the number of lock timeouts becomes
excessive when compared to normal operating levels, you might have an
application that is holding locks for long durations. In this case, this attribute
might indicate that you need to analyze some of the other attributes related to
locks and deadlocks to determine if you have an application problem.

© Copyright IBM Corp. 2005 155

 Scenario 2: Establishing an appropriate amount of sort
memory
Online transaction processing (OLTP) applications should not perform large sort
operations. Large sort operations are very costly in terms of CPU, I/O, and elapsed
time; as a result, sorts can slow down an OLTP application. The default
SORTHEAP size is 1 MB (256 4KB pages), which is adequate for most situations.

You can use the information in the Application Top Summary workspace to help
you track the number of sort overflows. In the Application Summary workspace,
you can view information about the number of sort overflows and the sort
overflow percentage. Additionally, you can use the Application Sort/Hash Join
Activity workspace to find information about the total number of sorts, the
average sort time, the number of sort overflows, and the percentage of sorts that
cause an overflow condition.

Sort overflows indicate that large sorts are occurring. If the number of sort
overflows represents greater than 3% of sorts, an application might experience
serious, unexpected sort problems. You must identify the SQL statements that are
causing the sorts and modify the SQL, indexes, or clustering to reduce the cost of
the sorts.

Buffer Pool workspace

The following scenario describes a specific use for the Buffer Pool workspace.

Scenario: Monitoring buffer pool efficiency

Database performance and tuning always start with buffer pool efficiency. The
buffer pool hit ratio indicates the percentage of time that the database manager did
not need to load a page from disk in order to satisfy a page request. That is, the
page was already in the buffer pool. The greater the buffer pool hit ratio, the lower
the frequency of disk I/O.

If the buffer pool hit ratio is low, the database will experience excessive I/O
activity. If this is the case, consider enlarging the buffer pool size for frequently
accessed tables or placing the indexes into a separate buffer pool. Buffer pools that
are too small result in excessive, unnecessary, physical I/O. Buffer pools that are
too large put a system at risk for operating system paging activity.

You can use the information displayed in the Buffer Pool workspace to evaluate
many of the characteristics of buffer pool activity. In the associated Buffer Pool
Detail workspace, you can evaluate the values of the various attributes related to
buffer pool hit ratios, asynchronous and synchronous I/O activity, and extended
store and non-buffer pool I/O activity.

Armed with this information, you can identify aspects of buffer pool activity that
are outside normal operating levels and take corrective action.

Database Overview workspace

The following scenarios describe specific uses for the Database Overview
workspace and its related workspaces.

156 IBM Tivoli Monitoring for Databases: DB2 Agent: User’s Guide
Scenario 1: Determining the maximum number of open files
The MAXFILOP parameter specifies the maximum number of database files that
any single database agent can have open at the same time. If opening a file might
cause this value to be exceeded, DB2 closes a file already in use by this agent. If
the value of MAXFILOP is too small, DB2 encounters the overhead of opening and
closing files so that the system does not exceed this limit. The overhead can
become excessive and cause performance degradation. SQL response time can slow
considerably. You can monitor the opening and closing of files by using the
Database Overview workspace, Database Summary workspace, and Database I/O
Activity workspace. In the Buffer Pool Activity area, you can determine the value
of the Files Closed attribute.

The Files Closed attribute can help you determine the best value for the
MAXFILOP configuration parameter. If the number of files being closed exceeds
the norm in your environment, consider increasing the value of MAXFILOP
parameter until the opening and closing reaches an acceptable level.

Scenario 2: Setting the value of the LOCKTIMEOUT parameter

The LOCKTIMEOUT configuration parameter specifies the number of seconds that
an application will wait to obtain a lock. By specifying a maximum value, you can
avoid global deadlocks for applications.

If you set this parameter to 0, an application does not wait for a lock. In this case,
if a lock is not available at the time of the request, the application receives
notification of this immediately.

If you set this parameter to -1, lock timeout detection is turned off. In this case an
application waits for a lock (if one is not available at the time of the request) until
the lock is granted or a deadlock occurs.

Set the value to detect quickly any waits that occur because of an abnormal
situation, such as a transaction that is stalled. Set the LOCKTIMEOUT parameter
high enough that valid lock requests do not time out because of peak workloads.
However the LOCKTIMEOUT configuration parameter can be set too high, which
causes the system to experience too few lock timeouts. In this case, applications
might wait excessively to obtain a lock.

You can use the information in the Database Overview workspace to help you
track the number of times that a database detected a timeout situation for all
applications that were connected. Use the Database Summary workspace and
Database Lock Activity workspace. The Database Lock Activity workspace displays
the Lock_Timeout attribute in addition to other lock-related attributes.

The Lock_Timeout attribute indicates the number of times that a request to lock an
object timed-out instead of being granted. A high value for this attribute can be
caused by:
v The value of the LOCKTIMEOUT configuration parameter being too low
v An application (transaction) that is holding one or more locks for an extended
period.
v A concurrency problem, that might be caused by lock escalations (from the
row-level to a table-level).

The Lock_Timeout attribute can help you adjust the setting for the
LOCKTIMEOUT configuration parameter. If the number of lock timeouts becomes

Appendix A. IBM Tivoli Monitoring for DB2 Workspaces 157

 excessive when compared to normal operating levels, one or more applications
might be holding locks for long durations. In this case, the Lock_Timeout attribute
might indicate that you need to analyze some of the other attributes related to
locks and deadlocks to determine if you have an application problem.

158 IBM Tivoli Monitoring for Databases: DB2 Agent: User’s Guide
Appendix B. IBM Tivoli Enterprise Console event mapping
Generic event mapping provides useful event class and attribute information for
situations that do not have specific event mapping defined. Each event class
corresponds to an attribute group in the monitoring agent. For a description of the
event slots for each event class, see Table 9 on page 160. For more information
about mapping attribute groups to event classes, see the IBM Tivoli Monitoring
Administrator’s Guide.

BAROC files are found on the Tivoli Enterprise Monitoring Server in the
installation directory in TECLIB (that is, install_dir/cms/TECLIB for Windows
systems and install_dir/tables/TEMS_hostname/TECLIB for UNIX systems). IBM
Tivoli Enterprise Console event synchronization provides a collection of
ready-to-use rule sets that you can deploy with minimal configuration. Be sure to
install IBM Tivoli Enterprise Console event synchronization to access the correct
Sentry.baroc, which is automatically included during base configuration of IBM
Tivoli Enterprise Console rules if you indicate that you want to use an existing
rulebase. See the IBM Tivoli Monitoring Installation and Setup Guide for details.

Each of the event classes is a child of KUD_Base. The KUD_Base event class can be
used for generic rules processing for any event from the Monitoring Agent for DB2

© Copyright IBM Corp. 2005 159

 Table 9. Overview of event slots to event classes
IBM Tivoli Enterprise Console event class event slots
ITM_KUDINFO00 KUDINFO00 attribute group
v node_name: STRING
v db2_status: STRING
v db2start_time: STRING
v last_reset: STRING
v snapshot_time: STRING
v prdid: STRING
v instance_name: STRING
v version: STRING
v server_db2_type: STRING
v sort_heap_allocated: INTEGER
v post_threshold_sorts: INTEGER
v piped_sorts_requested: INTEGER
v piped_sorts_accepted: INTEGER
v piped_sorts_accepted_pct: INTEGER
v rem_cons_in: INTEGER
v rem_cons_in_exec: INTEGER
v local_cons: INTEGER
v local_cons_in_exec: INTEGER
v con_local_dbases: INTEGER
v agents_registered: INTEGER
v agents_waiting_on_token: INTEGER
v agents_waiting_on_token_pct: INTEGER
v agents_from_pool: INTEGER
v agents_created_empty_pool: INTEGER
v agents_created_empty_pool_ratio:
INTEGER
v coord_agents_top: INTEGER
v max_agent_overflows: INTEGER
v agents_stolen: INTEGER
v agents_registered_top: INTEGER
v agents_waiting_top: INTEGER
v comm_private_mem: INTEGER
v idle_agents: INTEGER
v gw_total_cons: INTEGER
v gw_cur_cons: INTEGER
v gw_cons_wait_host: INTEGER
v gw_cons_wait_client: INTEGER
v post_threshold_hash_joins: INTEGER
v piped_sort_hit_ratio_pct_for_int:
INTEGER
v agentpri: INTEGER
v aslheapsz: INTEGER
v fcm_num_anchors: INTEGER
(Continued on the next page.)

160 IBM Tivoli Monitoring for Databases: DB2 Agent: User’s Guide
Table 9. Overview of event slots to event classes (continued)
IBM Tivoli Enterprise Console event class event slots
ITM_KUDINFO00 v fcm_num_buffers: INTEGER
(Continued)
v fcm_num_connect: INTEGER
v fcm_num_rqb: INTEGER
v maxagents: INTEGER
v max_coordagents: INTEGER
v maxcagents: INTEGER
v mon_heap_sz: INTEGER
v query_heap_sz: INTEGER
v rqrioblk: INTEGER
v piped_sorts_rejected_pct_for_int:
INTEGER
v sheapthres: INTEGER
v connection_status: INTEGER
v buff_free: INTEGER
v ce_free: INTEGER
v rb_free: INTEGER
v buff_free_bottom: INTEGER
v ce_free_bottom: INTEGER
v ma_free_bottom: INTEGER
v rb_free_bottom: INTEGER
v buf_used_pct: INTEGER
v rb_used_pct: INTEGER
v ce_used_pct: INTEGER
v buff_max_used_pct: INTEGER
v ce_max_used_pct: INTEGER
v ma_max_used_pct: INTEGER
v rb_max_used_pct: INTEGER
v total_buffers_rcvd: INTEGER
v total_buffers_sent: INTEGER
v piped_sorts_rejected_for_int: INTEGER
v dbpg_node_status: STRING
v instance_name_u: STRING
v db2_avail: INTEGER

Appendix B. IBM Tivoli Enterprise Console event mapping 161

 Table 9. Overview of event slots to event classes (continued)
IBM Tivoli Enterprise Console event class event slots
ITM_KUDDBASEGROUP00 KUDDBASEGROUP00 attribute group
v node_name: STRING
v db_name: STRING
v input_db_alias: STRING
v db_path: STRING
v dbase_status: STRING
v db_conn_time: STRING
v last_backup: STRING
v total_cons: INTEGER
v appls_cur_cons: INTEGER
v snapshot_time: STRING
v int_rollbacks: INTEGER
v int_deadlock_rollbacks: INTEGER
v rollback_sql_stmts: INTEGER
v failed_sql_stmts: INTEGER
v prefetch_wait_time: INTEGER
v agents_top: INTEGER
v coord_agents_top: INTEGER
v db_location: STRING
v server_platform: STRING
v connections_top: INTEGER
v db_heap_top: INTEGER
v locks_held: INTEGER
v lock_waits: INTEGER
v lock_wait_time: INTEGER
v lock_list_in_use: INTEGER
v deadlocks: INTEGER
v lock_escals: INTEGER
v x_lock_escals: INTEGER
v locks_waiting: INTEGER
v lock_timeouts: INTEGER
v avg_lock_wait_time: INTEGER
v sort_heap_allocated: INTEGER
v total_sorts: INTEGER
v total_sort_time: INTEGER
v sort_overflows: INTEGER
v active_sorts: INTEGER
v avg_sort_time: INTEGER
v sort_overflows_pct: INTEGER
v pool_data_l_reads: INTEGER
v pool_data_p_reads: INTEGER
v pool_data_writes: INTEGER
v pool_index_l_reads: INTEGER
v pool_index_p_reads: INTEGER
v pool_index_writes: INTEGER
(Continued on the next page.)

162 IBM Tivoli Monitoring for Databases: DB2 Agent: User’s Guide
Table 9. Overview of event slots to event classes (continued)
IBM Tivoli Enterprise Console event class event slots
ITM_KUDDBASEGROUP00 v pool_read_time: INTEGER
(Continued)
v pool_write_time: INTEGER
v files_closed: INTEGER
v pool_async_index_reads: INTEGER
v pool_data_to_estore: INTEGER
v pool_index_to_estore: INTEGER
v pool_index_from_estore: INTEGER
v pool_data_from_estore: INTEGER
v pool_async_data_reads: INTEGER
v pool_async_data_writes: INTEGER
v pool_async_index_writes: INTEGER
v pool_async_read_time: INTEGER
v pool_async_write_time: INTEGER
v pool_async_data_read_reqs: INTEGER
v pool_lsn_gap_clns: INTEGER
v pool_drty_pg_steal_clns: INTEGER
v pool_drty_pg_thrsh_clns: INTEGER
v pool_total_reads: INTEGER
v pool_hit_ratio: INTEGER
v avg_pool_read_time: INTEGER
v pool_total_writes: INTEGER
v avg_pool_write_time: INTEGER
v pool_sync_data_reads: INTEGER
v pool_sync_index_reads: INTEGER
v pool_sync_read: INTEGER
v pool_sync_data_writes: INTEGER
v pool_sync_index_writes: INTEGER
v pool_sync_write: INTEGER
v pool_sync_read_time: INTEGER
v avg_sync_read_time: INTEGER
v pool_sync_write_time: INTEGER
v avg_sync_write_time: INTEGER
v avg_data_page_read_per_async_req:
INTEGER
v direct_reads: INTEGER
v direct_writes: INTEGER
v direct_read_reqs: INTEGER
v direct_write_reqs: INTEGER
v direct_read_time: INTEGER
v direct_write_time: INTEGER
v commit_sql_stmts: INTEGER
v dynamic_sql_stmts: INTEGER
v static_sql_stmts: INTEGER
v select_sql_stmts: INTEGER
v ddl_sql_stmts: INTEGER
(Continued on the next page.)

Appendix B. IBM Tivoli Enterprise Console event mapping 163

 Table 9. Overview of event slots to event classes (continued)
IBM Tivoli Enterprise Console event class event slots
ITM_KUDDBASEGROUP00 v uid_sql_stmts: INTEGER
(Continued)
v total_sql_stmts: INTEGER
v sql_stmts_failed_pct: INTEGER
v sql_stmts_rollback_pct: INTEGER
v appl_section_lookups: INTEGER
v appl_section_inserts: INTEGER
v binds_precompiles: INTEGER
v total_hash_joins: INTEGER
v total_hash_loops: INTEGER
v hash_join_overflows: INTEGER
v hash_join_small_overflows: INTEGER
v rows_deleted: INTEGER
v rows_inserted: INTEGER
v rows_updated: INTEGER
v rows_selected: INTEGER
v total_sec_cons: INTEGER
v num_assoc_agents: INTEGER
v catalog_node_name: STRING
v sec_log_used_top: INTEGER
v tot_log_used_top: INTEGER
v sec_logs_allocated: INTEGER
v log_reads: INTEGER
v log_writes: INTEGER
v pkg_cache_lookups: INTEGER
v pkg_cache_inserts: INTEGER
v pkg_cache_hit_ratio: INTEGER
v cat_cache_lookups: INTEGER
v cat_cache_inserts: INTEGER
v cat_cache_overflows: INTEGER
v cat_cache_hit_ratio: INTEGER
v cat_cache_heap_full: INTEGER
v db_name_u: STRING
v input_db_alias_u: STRING
v db_path_u: STRING

164 IBM Tivoli Monitoring for Databases: DB2 Agent: User’s Guide
Table 9. Overview of event slots to event classes (continued)
IBM Tivoli Enterprise Console event class event slots
ITM_KUDDB2APPLGROUP00 KUDDB2APPLGROUP00 attribute group
v node_name: STRING
v agent_id: INTEGER
v appl_id: STRING
v appl_status: STRING
v snapshot_time: STRING
v appl_name: STRING
v auth_id: STRING
v client_prdid: STRING
v db_name: STRING
v execution_id: STRING
v corr_token: STRING
v client_platform: STRING
v client_protocol: STRING
v locks_held: INTEGER
v lock_waits: INTEGER
v lock_wait_time: INTEGER
v lock_escals: INTEGER
v x_lock_escals: INTEGER
v deadlocks: INTEGER
v uow_lock_wait_time: INTEGER
v lock_timeouts: INTEGER
v avg_lock_waittime: INTEGER
v agent_id_holding_lk: INTEGER
v appl_id_holding_lk: STRING
v lock_mode: STRING
v lock_object_type: STRING
v lock_wait_start_time: STRING
v table_name: STRING
v table_schema: STRING
v tablespace_name: STRING
v pool_data_l_reads: INTEGER
v pool_data_p_reads: INTEGER
v pool_data_writes: INTEGER
v pool_index_l_reads: INTEGER
v pool_index_p_reads: INTEGER
v pool_index_writes: INTEGER
v pool_read_time: INTEGER
v pool_write_time: INTEGER
v pool_data_to_estore: INTEGER
v pool_index_to_estore: INTEGER
v pool_index_from_estore: INTEGER
v pool_data_from_estore: INTEGER
v pool_total_reads: INTEGER
v pool_hit_ratio: INTEGER
(Continued on next page.)

Appendix B. IBM Tivoli Enterprise Console event mapping 165

 Table 9. Overview of event slots to event classes (continued)
IBM Tivoli Enterprise Console event class event slots
ITM_KUDDB2APPLGROUP00 v avg_pool_read_time: INTEGER
(Continued)
v pool_total_writes: INTEGER
v avg_pool_write_time: INTEGER
v appl_idle_time: INTEGER
v agent_usr_cpu_time: STRING
v agent_sys_cpu_time: STRING
v appl_con_time: STRING
v conn_complete_time: STRING
v prefetch_wait_time: INTEGER
v direct_reads: INTEGER
v direct_writes: INTEGER
v direct_read_reqs: INTEGER
v direct_write_reqs: INTEGER
v direct_read_time: INTEGER
v direct_write_time: INTEGER
v stmt_type: STRING
v stmt_operation: STRING
v cursor_name: STRING
v creator: STRING
v package_name: STRING
v total_sql_stmt: INTEGER
v failed_sql_stmts: INTEGER
v int_rollbacks: INTEGER
v int_deadlock_rollbacks: INTEGER
v rollback_sql_stmts: INTEGER
v failed_sql_stmts_pct: INTEGER
v total_sorts: INTEGER
v total_sort_time: INTEGER
v sort_overflows: INTEGER
v avg_sort_time: INTEGER
v sort_overflows_pct: INTEGER
v commit_sql_stmts: INTEGER
v dynamic_sql_stmts: INTEGER
v static_sql_stmts: INTEGER
v select_sql_stmts: INTEGER
v ddl_sql_stmts: INTEGER
v uid_sql_stmts: INTEGER
v binds_precompiles: INTEGER
v section_number: INTEGER
v stmt_start: STRING
v stmt_stop: STRING
v uow_log_space_used: INTEGER
v uow_comp_status: STRING
v prev_uow_stop_time: STRING
v uow_start_time: STRING
(Continued on the next page.)
166 IBM Tivoli Monitoring for Databases: DB2 Agent: User’s Guide
Table 9. Overview of event slots to event classes (continued)
IBM Tivoli Enterprise Console event class event slots
ITM_KUDDB2APPLGROUP00 v uow_stop_time: STRING
(Continued)
v int_auto_rebinds: INTEGER
v int_rows_deleted: INTEGER
v int_rows_updated: INTEGER
v int_commits: INTEGER
v int_rows_inserted: INTEGER
v rows_deleted: INTEGER
v rows_inserted: INTEGER
v rows_updated: INTEGER
v rows_selected: INTEGER
v rows_read: INTEGER
v rows_written: INTEGER
v open_loc_curs: INTEGER
v open_loc_curs_blk: INTEGER
v open_rem_curs: INTEGER
v open_rem_curs_blk: INTEGER
v rej_curs_blk: INTEGER
v acc_curs_blk: INTEGER
v client_pid: INTEGER
v country_code: INTEGER
v pkg_cache_lookups: INTEGER
v pkg_cache_inserts: INTEGER
v pkg_cache_hit_ratio: INTEGER
v cat_cache_lookups: INTEGER
v cat_cache_inserts: INTEGER
v cat_cache_overflows: INTEGER
v cat_cache_heap_full: INTEGER
v cat_cache_hit_ratio: INTEGER
v total_hash_joins: INTEGER
v total_hash_loops: INTEGER
v hash_join_overflows: INTEGER
v hash_join_small_overflows: INTEGER
v query_cost_estimate: INTEGER
v query_card_estimate: INTEGER
v degree_parallelism: INTEGER
v stmt_text: STRING

Appendix B. IBM Tivoli Enterprise Console event mapping 167

 Table 9. Overview of event slots to event classes (continued)
IBM Tivoli Enterprise Console event class event slots
ITM_KUDLOCKCONFLICT00 KUDLOCKCONFLICT00 attribute group
v node_name: STRING
v agent_id: INTEGER
v appl_id: STRING
v appl_name: STRING
v appl_status: STRING
v snapshot_time: STRING
v auth_id: STRING
v client_db_alias: STRING
v status_change_time: STRING
v codepage_id: INTEGER
v locks_held: INTEGER
v lock_wait_time: INTEGER
v agent_id_holding_lk: INTEGER
v appl_id_holding_lk: STRING
v table_name: STRING
v table_schema: STRING
v tablespace_name: STRING
v lock_escalation: STRING
v lock_wait_start_time: STRING
v lock_mode: STRING
v lock_object_type: STRING
v appl_id_u: STRING
v appl_name_u: STRING
v auth_id_u: STRING
v client_db_alias_u: STRING
v appl_id_holding_lk_u: STRING
v table_name_u: STRING
v table_schema_u: STRING
v tablespace_name_u: STRING

168 IBM Tivoli Monitoring for Databases: DB2 Agent: User’s Guide
Table 9. Overview of event slots to event classes (continued)
IBM Tivoli Enterprise Console event class event slots
ITM_KUDBUFFERPOOL00 KUDBUFFERPOOL00 attribute group
v node_name: STRING
v bp_id: STRING
v bp_name: STRING
v input_db_alias: STRING
v db_name: STRING
v db_path: STRING
v pool_data_l_reads: INTEGER
v pool_data_p_reads: INTEGER
v pool_data_writes: INTEGER
v pool_index_l_reads: INTEGER
v pool_index_p_reads: INTEGER
v pool_index_writes: INTEGER
v pool_read_time: INTEGER
v pool_write_time: INTEGER
v files_closed: INTEGER
v pool_async_index_reads: INTEGER
v pool_data_to_estore: INTEGER
v pool_index_to_estore: INTEGER
v pool_index_from_estore: INTEGER
v pool_data_from_estore: INTEGER
v pool_async_data_reads: INTEGER
v pool_async_data_writes: INTEGER
v pool_async_index_writes: INTEGER
v pool_async_read_time: INTEGER
v pool_async_write_time: INTEGER
v pool_async_data_read_reqs: INTEGER
v pool_total_reads: INTEGER
v pool_hit_ratio: INTEGER
v avg_pool_read_time: INTEGER
v pool_total_writes: INTEGER
v avg_pool_write_time: INTEGER
v pool_sync_data_reads: INTEGER
v pool_sync_index_reads: INTEGER
v pool_sync_read: INTEGER
v pool_sync_data_writes: INTEGER
v pool_sync_index_writes: INTEGER
v pool_sync_write: INTEGER
v pool_sync_read_time: INTEGER
v avg_sync_read_time: INTEGER
v pool_sync_write_time: INTEGER
v avg_sync_write_time: INTEGER
v avg_data_page_read_per_async_req:
INTEGER
v direct_reads: INTEGER
(Continued on the next page.)

Appendix B. IBM Tivoli Enterprise Console event mapping 169

 Table 9. Overview of event slots to event classes (continued)
IBM Tivoli Enterprise Console event class event slots
ITM_KUDBUFFERPOOL00 v direct_writes: INTEGER
(Continued)
v direct_read_reqs: INTEGER
v direct_write_reqs: INTEGER
v direct_read_time: INTEGER
v direct_write_time: INTEGER
v avg_direct_read_time: INTEGER
v avg_direct_write_time: INTEGER
v bp_name_u: STRING
v input_db_alias_u: STRING
v db_name_u: STRING
v db_path_u: STRING

170 IBM Tivoli Monitoring for Databases: DB2 Agent: User’s Guide
Table 9. Overview of event slots to event classes (continued)
IBM Tivoli Enterprise Console event class event slots
ITM_KUDTABSPACE KUDTABSPACE attribute group
v node_name: STRING
v db_name: STRING
v version: INTEGER
v tablespace_name: STRING
v tablespace_type: STRING
v extent_size: INTEGER
v page_size: INTEGER
v prefetch_size: INTEGER
v num_containers: INTEGER
v container_name: STRING
v object_id: INTEGER
v tablespace_id: INTEGER
v total_pages: INTEGER
v usable_pages: INTEGER
v used_pages: INTEGER
v pending_free_pages: INTEGER
v free_pages: INTEGER
v direct_reads: INTEGER
v direct_writes: INTEGER
v direct_read_reqs: INTEGER
v direct_write_reqs: INTEGER
v pool_data_writes: INTEGER
v pool_async_data_reads: INTEGER
v pool_async_data_writes: INTEGER
v pool_async_index_reads: INTEGER
v pool_async_index_writes: INTEGER
v pool_async_read_time: INTEGER
v pool_async_write_time: INTEGER
v pool_data_l_reads: INTEGER
v pool_data_from_estore: INTEGER
v pool_data_to_estore: INTEGER
v pool_data_p_reads: INTEGER
v pool_index_l_reads: INTEGER
v pool_index_from_estore: INTEGER
v pool_index_to_estore: INTEGER
v pool_index_p_reads: INTEGER
v pool_index_writes: INTEGER
v direct_write_time: INTEGER
v direct_read_time: INTEGER
v files_closed: INTEGER
v pool_read_time: INTEGER
v pool_write_time: INTEGER
v pool_async_data_read_reqs: INTEGER
v pool_data_reads: INTEGER
(Continued on the next page.)

Appendix B. IBM Tivoli Enterprise Console event mapping 171

 Table 9. Overview of event slots to event classes (continued)
IBM Tivoli Enterprise Console event class event slots
ITM_KUDTABSPACE v pool_sync_idx_writes: INTEGER
(Continued)
v pool_sync_reads: INTEGER
v pool_sync_writes: INTEGER
v avg_sect_read: INTEGER
v avg_sect_written: INTEGER
v avg_direct_read_time: INTEGER
v avg_direct_write_time: INTEGER
v avg_pool_io_time: INTEGER
v avg_pool_read_time: INTEGER
v avg_pool_write_time: INTEGER
v avg_sync_data_read_time: INTEGER
v avg_sync_data_write_time: INTEGER
v avg_sync_io_time: INTEGER
v pool_io_per_sec: INTEGER
v estore_rw_ratio: INTEGER
v pool_hit_pct: INTEGER
v pool_idx_hit_pct_for_int: INTEGER
v prefetch_pct_for_int: INTEGER
v total_io_pct: INTEGER
v prefetch_reqs_for_int: INTEGER
v sync_read_time: INTEGER
v sync_write_time: INTEGER
v total_direct_io_time: INTEGER
v total_pool_io_time: INTEGER
v total_pool_p_read_time: INTEGER
v total_pool_p_write_time: INTEGER
v total_sync_io: INTEGER
v total_sync_io_time: INTEGER
v pool_async_index_read_reqs: INTEGER
v space_used_dms_table_pct: INTEGER
v space_used_sms_table: INTEGER
v tbsp_status: INTEGER
v db_name_u: STRING
v tablespace_name_u: STRING
v container_name_u: STRING

172 IBM Tivoli Monitoring for Databases: DB2 Agent: User’s Guide
Table 9. Overview of event slots to event classes (continued)
IBM Tivoli Enterprise Console event class event slots
ITM_KUDDBASEGROUP01 KUDDBASEGROUP01 attribute group
v db_name: STRING
v app_ctl_heap_sz: INTEGER
v applheapsz: INTEGER
v avg_appls: INTEGER
v buffpage: INTEGER
v catalogcache_sz: INTEGER
v chngpgs_thresh: INTEGER
v dbheap: INTEGER
v locklist: INTEGER
v logbufsz: INTEGER
v logprimary: INTEGER
v maxappls: INTEGER
v maxlocks: INTEGER
v mincommit: INTEGER
v newlogpath: STRING
v num_iocleaners: INTEGER
v pckcachesz: INTEGER
v restore_pending: INTEGER
v seqdetect: INTEGER
v sortheap: INTEGER
v appls_in_db2: INTEGER
v deadlocks_for_int: INTEGER
v int_auto_rebinds: INTEGER
v int_commits: INTEGER
v int_rows_deleted: INTEGER
v int_rows_inserted: INTEGER
v int_rows_updated: INTEGER
v log_io_for_int: INTEGER
v cur_cons_pct: INTEGER
v lock_list_in_use_pct: INTEGER
v sec_log_used_pct: INTEGER
v total_log_used: INTEGER
v pri_log_used_pct: INTEGER
v page_cleans_for_interval: INTEGER
v pages_per_prefetch_for_int: INTEGER
v tot_sync_io: INTEGER
v pool_sync_index_reads: INTEGER
v avg_pool_async_data_reads: INTEGER
v avg_pool_async_data_writes: INTEGER
v avg_pool_writes_per_read: INTEGER
v avg_sect_read_per_direct_read: INTEGER
v avg_sect_written_per_direct_write:
INTEGER
v avg_direct_read_time: INTEGER
(Continued on the next page.)

Appendix B. IBM Tivoli Enterprise Console event mapping 173

 Table 9. Overview of event slots to event classes (continued)
IBM Tivoli Enterprise Console event class event slots
ITM_KUDDBASEGROUP01 v avg_direct_write_time: INTEGER
(Continued)
v avg_pool_io_time: INTEGER
v avg_sync_io_time: INTEGER
v avg_pages_per_cleaner_for_int: INTEGER
v pool_io_per_sec: INTEGER
v estore_rw_ratio_for_int: INTEGER
v pool_hit_ratio_pct_for_int: INTEGER
v pool_hit_ratio_index_pct_for_int:
INTEGER
v tot_direct_io_time: INTEGER
v tot_pool_phys_io: INTEGER
v tot_pool_phys_read: INTEGER
v tot_pool_phys_write: INTEGER
v tot_sync_io_time: INTEGER
v db_cap_err: INTEGER
v db_cap_lag: INTEGER
v db_cap_prun: INTEGER
v avg_locks_held: INTEGER
v lock_waits_for_int: INTEGER
v lock_waits_pct: INTEGER
v int_deadlock_rollbacks_pct_for_int:
INTEGER
v db_tablespaces: INTEGER
v event_monitors: INTEGER
v invalid_triggers: INTEGER
v system_tablespaces: INTEGER
v tables: INTEGER
v tablespaces: INTEGER
v tablespaces_long_data: INTEGER
v triggers: INTEGER
v views: INTEGER
v user_indexes: INTEGER
v commit_stmts_per_sec: INTEGER
v ddl_sql_pct_for_int: INTEGER
v select_sql_pct_for_int: INTEGER
v uid_sql_pct_for_int: INTEGER
v invalid_pkgs: INTEGER
v invalid_sys_pkgs: INTEGER
v pri_log_used_top: INTEGER
v int_deadlock_rollbacks_pct: INTEGER
v num_ioservers: INTEGER
v snapshot_time: STRING
v node_name: STRING
v db_name_u: STRING
v newlogpath_u: STRING

174 IBM Tivoli Monitoring for Databases: DB2 Agent: User’s Guide
Table 9. Overview of event slots to event classes (continued)
IBM Tivoli Enterprise Console event class event slots
ITM_KUDDB2APPLGROUP01 KUDDB2APPLGROUP01 attribute group
v snapshot_time: STRING
v appl_name: STRING
v db_name: STRING
v agents_stolen: INTEGER
v app_work_load: INTEGER
v associated_agents_top: INTEGER
v avg_sect_read_per_direct_read: INTEGER
v avg_sect_written_per_direct_read:
INTEGER
v pool_index_hit_ratio_pct_for_int:
INTEGER
v pkg_cache_hit_pct: INTEGER
v tot_pool_io_time: INTEGER
v lock_escalation_for_int: INTEGER
v deadlocks_for_int: INTEGER
v stmts_sorts: INTEGER
v appl_section_inserts: INTEGER
v binds_precompiles: INTEGER
v ddl_sql_pct_for_int: INTEGER
v uid_sql_pct_for_int: INTEGER
v sql_reqs_since_commit: INTEGER
v node_name: STRING
v appl_name_u: STRING
v db_name_u: STRING

Appendix B. IBM Tivoli Enterprise Console event mapping 175

 Table 9. Overview of event slots to event classes (continued)
IBM Tivoli Enterprise Console event class event slots
ITM_KUDDB2APPLGROUP00_U KUDDB2APPLGROUP00_U attribute group
v node_name: STRING
v agent_id: INTEGER
v appl_id: STRING
v appl_status: STRING
v snapshot_time: STRING
v appl_name_u: STRING
v auth_id_u: STRING
v client_prdid: STRING
v db_name_u: STRING
v execution_id_u: STRING
v corr_token_u: STRING
v client_platform: STRING
v client_protocol: STRING
v locks_held: INTEGER
v lock_waits: INTEGER
v lock_wait_time: INTEGER
v lock_escals: INTEGER
v x_lock_escals: INTEGER
v deadlocks: INTEGER
v uow_lock_wait_time: INTEGER
v lock_timeouts: INTEGER
v avg_lock_waittime: INTEGER
v agent_id_holding_lk: INTEGER
v appl_id_holding_lk_u: STRING
v lock_mode: STRING
v lock_object_type: STRING
v lock_wait_start_time: STRING
v table_name_u: STRING
v table_schema_u: STRING
v tablespace_name_u: STRING
v pool_data_l_reads: INTEGER
v pool_data_p_reads: INTEGER
v pool_data_writes: INTEGER
v pool_index_l_reads: INTEGER
v pool_index_p_reads: INTEGER
v pool_index_writes: INTEGER
v pool_read_time: INTEGER
v pool_write_time: INTEGER
v pool_data_to_estore: INTEGER
v pool_index_to_estore: INTEGER
v pool_index_from_estore: INTEGER
v pool_data_from_estore: INTEGER
v pool_total_reads: INTEGER
v pool_hit_ratio: INTEGER
(Continued on the next page.)

176 IBM Tivoli Monitoring for Databases: DB2 Agent: User’s Guide
Table 9. Overview of event slots to event classes (continued)
IBM Tivoli Enterprise Console event class event slots
ITM_KUDDB2APPLGROUP00_U v avg_pool_read_time: INTEGER
(Continued)
v pool_total_writes: INTEGER
v avg_pool_write_time: INTEGER
v appl_idle_time: INTEGER
v agent_usr_cpu_time: STRING
v agent_sys_cpu_time: STRING
v appl_con_time: STRING
v conn_complete_time: STRING
v prefetch_wait_time: INTEGER
v direct_reads: INTEGER
v direct_writes: INTEGER
v direct_read_reqs: INTEGER
v direct_write_reqs: INTEGER
v direct_read_time: INTEGER
v direct_write_time: INTEGER
v stmt_type: STRING
v stmt_operation: STRING
v cursor_name_u: STRING
v creator_u: STRING
v package_name_u: STRING
v total_sql_stmt: INTEGER
v int_deadlock_rollbacks: INTEGER
v rollback_sql_stmts: INTEGER
v failed_sql_stmts_pct: INTEGER
v total_sorts: INTEGER total_sort_time:
INTEGER
v sort_overflows: INTEGER
v avg_sort_time: INTEGER
v ort_overflows_pct: INTEGER
v commit_sql_stmts: INTEGER
v dynamic_sql_stmts: INTEGER
v static_sql_stmts: INTEGER
v select_sql_stmts: INTEGER
v ddl_sql_stmts: INTEGER
v uid_sql_stmts: INTEGER
v binds_precompiles: INTEGER
v section_number: INTEGER
v stmt_start: STRING
v stmt_stop: STRING
v uow_log_space_used: INTEGER
v uow_comp_status: STRING
v prev_uow_stop_time: STRING
v uow_start_time: STRING
v uow_stop_time: STRING
v int_auto_rebinds: INTEGER
(Continued on the next page.)

Appendix B. IBM Tivoli Enterprise Console event mapping 177

 Table 9. Overview of event slots to event classes (continued)
IBM Tivoli Enterprise Console event class event slots
ITM_KUDDB2APPLGROUP00_U v int_rows_deleted: INTEGER
(Continued)
v int_rows_updated: INTEGER
v int_commits: INTEGER
v int_rows_inserted: INTEGER
v rows_deleted: INTEGER
v rows_inserted: INTEGER
v rows_updated: INTEGER
v rows_selected: INTEGER
v rows_read: INTEGER
v rows_written: INTEGER
v open_loc_curs: INTEGER
v open_loc_curs_blk: INTEGER
v open_rem_curs: INTEGER
v open_rem_curs_blk: INTEGER
v rej_curs_blk: INTEGER
v acc_curs_blk: INTEGER
v client_pid: INTEGER
v country_code: INTEGER
v pkg_cache_lookups: INTEGER
v pkg_cache_inserts: INTEGER
v pkg_cache_hit_ratio: INTEGER
v cat_cache_lookups: INTEGER
v cat_cache_inserts: INTEGER
v cat_cache_overflows: INTEGER
v cat_cache_heap_full: INTEGER
v cat_cache_hit_ratio: INTEGER
v total_hash_joins: INTEGER
v total_hash_loops: INTEGER
v hash_join_overflows: INTEGER
v hash_join_small_overflows: INTEGER
v query_cost_estimate: INTEGER
v query_card_estimate: INTEGER
v degree_parallelism: INTEGER
v stmt_text_u: STRING

178 IBM Tivoli Monitoring for Databases: DB2 Agent: User’s Guide
Appendix C. Problem determination
This appendix explains how to troubleshoot the IBM Tivoli Monitoring for
Databases: DB2 Agent. Troubleshooting, or problem determination, is the process
of determining why a certain product is malfunctioning.

Note: You can resolve some problems by ensuring that your system matches the
system requirements listed in Chapter 2, “Requirements and configuration
for the monitoring agent,” on page 3.

This appendix provides agent-specific problem determination information. See the

IBM Tivoli Monitoring Problem Determination Guide for general problem
determination information. Also see “Support information” on page 200 for other
problem-solving options.

Gathering product information for IBM Software Support

Before contacting IBM Software Support about a problem you are experiencing
with this product, gather the following information that relates to the problem:
Table 10. Information to gather before contacting IBM Software Support
Information type Description
Log files Collect trace log files from failing systems. Most logs are located in a logs subdirectory
on the host computer. See “Trace logging” on page 180 for lists of all trace log files and
their locations. See the IBM Tivoli Monitoring User’s Guide for general information about
the IBM Tivoli Monitoring environment.
DB2 Universal Database v Version number and patch level
information v Sample application data file (if monitoring a file)
Operating system Operating system version number and patch level
Messages Messages and other information displayed on the screen
Version numbers for Version number of the following members of the monitoring environment:
IBM Tivoli Monitoring v IBM Tivoli Monitoring. Also provide the patch level, if available.
v IBM Tivoli Monitoring for Databases: DB2 Agent
Screen captures Screen captures of incorrect output, if any.
(UNIX only) Core dump If the system stops on UNIX systems, collect core dump file from install_dir/bin directory,
files where install_dir is the directory path where you installed the monitoring agent.

Upload files for review to the following FTP site: ftp.emea.ibm.com. Log in as
anonymous and place your files in the directory that corresponds to the IBM Tivoli
Monitoring component that you use. See “Contacting IBM Software Support” on
page 202 for more information about working with IBM Software Support.

Built-in problem determination features

The primary troubleshooting feature in the IBM Tivoli Monitoring for Databases:
DB2 Agent is logging. Logging refers to the text messages and trace data generated
by the IBM Tivoli Monitoring for Databases: DB2 Agent. Messages and trace data
are sent to a file.

Trace data captures transient information about the current operating environment
when a component or application fails to operate as designed. IBM Software

© Copyright IBM Corp. 2005 179

 Support personnel use the captured trace information to determine the source of
an error or unexpected condition. See “Trace logging” for more information.

Problem classification
The following types of problems might occur with the IBM Tivoli Monitoring for
Databases: DB2 Agent:
v Installation and configuration
v General usage and operation
v Display of monitoring data
v Take Action commands

This appendix provides symptom descriptions and detailed workarounds for these
problems, as well as describing the logging capabilities of the monitoring agent.
See the IBM Tivoli Monitoring Problem Determination Guide for general problem
determination information.

Trace logging
Trace logs capture information about the operating environment when component
software fails to operate as intended. The principal log type is the RAS (Reliability,
Availability, and Serviceability) trace log. These logs are in the English language
only. The RAS trace log mechanism is available for all components of IBM Tivoli
Monitoring. Most logs are located in a logs subdirectory on the host computer. See
the following sections to learn how to configure and use trace logging:
v “Principal trace log files” on page 181
v “Examples: using trace logs” on page 183
v “Setting RAS trace parameters” on page 184

Note: The documentation refers to the RAS facility in IBM Tivoli Monitoring as
″RAS1″.

IBM Software Support uses the information captured by trace logging to trace a
problem to its source or to determine why an error occurred. The default
configuration for trace logging, such as whether trace logging is enabled or
disabled and trace level, depends on the source of the trace logging. Trace logging
is always enabled.

Overview of log file management

Table 11 on page 182 provides the names, locations, and descriptions of RAS1 log
files. The log file names adhere to the following naming convention:
hostname_product_instance_program_timestamp-nn.log

where:
v hostname is the host name of the machine on which the monitoring component is
running.
v product is the two-character product code. For Monitoring Agent for DB2, the
product code is ud.
v instance is the name of a database instance that is being monitored.
v program is the name of the program being run.

180 IBM Tivoli Monitoring for Databases: DB2 Agent: User’s Guide
 v timestamp is an 8-character hexadecimal timestamp representing the time at
which the program started.
v nn is a rolling log suffix. See “Examples of trace logging” for details of log
rolling.

Examples of trace logging

For example, if a DB2 Universal Database instance ″dbinst02″ is running on
computer ″server01″, the RAS log file for the Monitoring Agent for DB2 might be
named as follows:
server01_ud_dbinst02_kudcma_437fc59-01.log

Note: These examples show log names on Windows, where the program is named
kudcma.
For long-running programs, the nn suffix is used to maintain a short history of log
files for that startup of the program. For example, the kudcma program might have
a series of log files as follows:
server01_ud_dbinst02_kudcma_437fc59-01.log
server01_ud_dbinst02_kudcma_437fc59-02.log
server01_ud_dbinst02_kudcma_437fc59-03.log

As the program runs, the first log (nn=01) is preserved because it contains program
startup information. The remaining logs ″roll." In other words, when the set of
numbered logs reach a maximum size, the remaining logs are overwritten in
sequence. Each time a program is started, a new timestamp is assigned to maintain
a short program history. For example, if the Monitoring Agent for DB2 is started
twice, it might have log files as follows:
server01_ud_dbinst02_kudcma_437fc59-01.log
server01_ud_dbinst02_kudcma_437fc59-02.log
server01_ud_dbinst02_kudcma_437fc59-03.log

server01_ud_dbinst02_kudcma_537fc59-01.log
server01_ud_dbinst02_kudcma_537fc59-02.log
server01_ud_dbinst02_kudcma_537fc59-03.log

Each program that is started has its own log file. For example, the Monitoring
Agent for DB2 would have agent logs in this format:
server01_ud_dbinst02_kudcma_437fc59-01.log

Other logs, such as logs for collector processes and Take Action commands, have a
similar syntax, as in the following example:
server01_ud_dbinst02_kudBackupDatabase_447fc59-01.log

where kudBackupDatabase is the name of a program.

Note: When you communicate with IBM Software Support, you must capture and
send the RAS1 log that matches any problem occurrence that you report.

Principal trace log files

Table 11 on page 182 contains locations, file names, and descriptions of trace logs
that can help determine the source of problems with agents.

Appendix C. Problem determination 181

Table 11. Trace log files for troubleshooting agents
System where log File name and path Description
is located
On the computer The RAS1 log files are named Traces activity of the monitoring agent.
that hosts the hostname_ud_instance_program_timestamp-nn.log Note: Other logs, such as logs for
monitoring agent and are located in the following path: collector processes and Take Action
v On Windows: install_dir\tmaitm6\logs commands (if available), have a similar
See “Definitions of syntax and are located in this directory
variables” on page v On UNIX: install_dir/logs
path.
183 for Note: File names for RAS1 logs include a
descriptions of the hexadecimal time stamp.
variables in the file
names in column Also on UNIX, a log with a decimal time stamp is
two. provided: hostname_ud_timestamp.log and
hostname_ud_timestamp.pidnnnnn in the
install_dir/logs path, where nnnnn is the process
ID number.
The *.LG0 file is located in the following path: A new version of this file is generated
v On Windows: install_dir\tmaitm6\logs every time the agent is restarted. IBM
Tivoli Monitoring generates one backup
v On UNIX: install_dir/logs
copy of the *.LG0 file with the tag .LG1.
View .LG0 to learn the following details
regarding the current monitoring
session:
v Status of connectivity with the
monitoring server.
v Situations that were running.
v The success or failure status of Take
Action commands.
On the Tivoli On UNIX: The candle_installation.log file in the Provides details about products that are
Enterprise install_dir/logs path. > installed.
Monitoring Server Note: Trace logging is enabled by
On Windows: The file in the default. A configuration step is not
See “Definitions of install_dir\InstallITM path. Unlike RAS1 log files, required to enable this tracing.
variables” on page the name of the file displays a decimal time stamp. *
183 for
The Warehouse_Configuration.log file is located in Provides details about the configuration
descriptions of the
the following path on Windows: of data warehousing for historical
variables in the file
install_dir\InstallITM. reporting.
names in column
two. The RAS1 log file is named hostname_ms_timestamp- Traces activity on the monitoring server.
nn.log and is located in the following path:
v On Windows: install_dir\logs
v On UNIX: install_dir/logs
Note: File names for RAS1 logs include a
hexadecimal time stamp

Also on UNIX, a log with a decimal time stamp is

provided: hostname_ms_timestamp.log and
hostname_ms_timestamp.pidnnnnn in the
install_dir/logs path, where nnnnn is the process
ID number.

182 IBM Tivoli Monitoring for Databases: DB2 Agent: User’s Guide
Table 11. Trace log files for troubleshooting agents (continued)
System where log File name and path Description
is located
On the Tivoli The RAS1 log file is named hostname_cq_timestamp- Traces activity on the portal server.
Enterprise Portal nn.log and is located in the following path:
Server v On Windows: install_dir\logs
See “Definitions of v On UNIX: install_dir/logs
variables” for Note: File names for RAS1 logs include a
descriptions of the hexadecimal time stamp
variables in the file
names in column Also on UNIX, a log with a decimal time stamp is
two. provided: hostname_cq_timestamp.log and
hostname_cq_timestamp.pidnnnnn in the
install_dir/logs path, where nnnnn is the process ID
number.
The TEPS_ODBC.log file is located in the following When you enable historical reporting,
path on Windows: install_dir\InstallITM. this log file traces the status of the
warehouse proxy agent.
Definitions of variables for RAS1 logs:
v hostname is the host name of the machine on which the agent is running.
v install_dir represents the directory path where you installed the IBM Tivoli Monitoring component. install_dir can
represent a path on the computer that hosts the monitoring server, the monitoring agent, or the portal server.
v product is the two character product code. For Monitoring Agent for DB2, the product code is ud.
v instance refers to the name of the database instance that you are monitoring.
v program is the name of the program being run.
v timestamp is an eight-character hexadecimal time stamp representing the time at which the program started.
v nn is a rolling log suffix. See “Examples of trace logging” on page 181 for details of log rolling.

See the IBM Tivoli Monitoring Installation and Setup Guide for more information on
the complete set of trace logs that are maintained on the monitoring server.

Examples: using trace logs

Typically IBM Software Support applies specialized knowledge to analyze trace
logs to determine the source of problems. However, you can open trace logs in a
text editor to learn some basic facts about your IBM Tivoli Monitoring
environment.
Example one
This excerpt shows the typical .LG0 log for a failed connection between a
monitoring agent and a monitoring server with the host name server1a:
(Thursday, August 11, 2005, 08:21:30-{94C}kdcl0cl.c,105,"KDCL0_ClientLookup") status=1c020006,
"location server unavailable", ncs/KDC1_STC_SERVER_UNAVAILABLE
(Thursday, August 11, 2005, 08:21:35-{94C}kraarreg.cpp,1157,"LookupProxy") Unable to connect to
broker at ip.pipe:: status=0, "success", ncs/KDC1_STC_OK
(Thursday, August 11, 2005, 08:21:35-{94C}kraarreg.cpp,1402,"FindProxyUsingLocalLookup") Unable
to find running CMS on CT_CMSLIST <IP.PIPE:#server1a>
Example two
The following excerpts from the trace log for the monitoring server show the
status of an agent, identified here as ″Remote node.″ The name of the
computer where the agent is running is SERVER5B:
(42C039F9.0000-6A4:kpxreqhb.cpp,649,"HeartbeatInserter") Remote node SERVER5B:KUD is ON-LINE.
. . .
(42C3079B.0000-6A4:kpxreqhb.cpp,644,"HeartbeatInserter") Remote node SERVER5B:KUD is OFF-LINE.

Key points regarding the preceding excerpt:

v The monitoring server appends the KUD product code to the server
name to form a unique name (SERVER5B:KUD) for this instance of
Appendix C. Problem determination 183
 Monitoring Agent for DB2. This unique name enables you to distinguish
multiple monitoring products that might be running on SERVER5B.
v The log shows when the agent started (ON-LINE) and later stopped
(OFF-LINE) in the environment.
v For the sake of brevity an ellipsis (...) represents the series of trace log
entries that were generated while the agent was running.
v Between the ON-LINE and OFF-LINE log entries, the agent was
communicating with the monitoring server.
v The ON-LINE and OFF-LINE log entries are always available in the
trace log. All trace levels that are described in “Setting RAS trace
parameters” provide these entries.

On Windows, you can use the following alternate method to view trace logs:
1. In the Windows Start menu, choose Program Files > IBM Tivoli Monitoring >
Manage Tivoli Monitoring Service. The Manage Tivoli Enterprise Monitoring
Services window is displayed.
2. Right-click a component and select Advanced > View Trace Log in the pop-up
menu. The program displays the Select Log File window that lists the RAS1
logs for the monitoring agent.
3. Select a log file from the list and click OK. You can also use this viewer to
access remote logs.

Note: The viewer converts time stamps in the logs to a readable format.

Setting RAS trace parameters

Objective
Pinpoint a problem by setting detailed tracing of individual components of the
monitoring agent and modules.

Background Information
Monitoring Agent for DB2 uses RAS1 tracing and generates the logs described in
Table 11 on page 182. The default RAS1 trace level is ERROR.

RAS1 tracing has control parameters to manage to the size and number of RAS1
logs. Use the procedure described in this section to set the parameters.

Note: The KBB_RAS1_LOG parameter also provides for the specification of the
log file directory, log file name, and the inventory control file directory and
name. Do not modify these values or log information can be lost.

Before you begin

See “Overview of log file management” on page 180 to ensure that you understand
log rolling and can reference the correct log files when you managing log file
generation.

After you finish

Monitor the size of the logs directory. Default behavior can generate a total of 45 to
60 MB for each agent that is running on a computer. For example, each database
instance that you monitor could generate 45 to 60 MB of log data. See the
″Procedure″ section to learn how to adjust file size and numbers of log files to
prevent logging activity from occupying too much disk space.

184 IBM Tivoli Monitoring for Databases: DB2 Agent: User’s Guide
Regularly prune log files other than the RAS1 log files in the logs directory. Unlike
the RAS1 log files which are pruned automatically, other log types can grow
indefinitely, for example, the logs in Table 11 on page 182 that include a process ID
number (PID).

Consider using collector trace logs (described in Table 11 on page 182) as an

additional source of problem determination information.

Note: The KDC_DEBUG setting and the Maximum error tracing setting can
generate a large amount of trace logging. Use them only temporarily, while
you are troubleshooting problems. Otherwise, the logs can occupy excessive
amounts of hard disk space.

Procedure
Specify RAS1 trace options in the install_dir\tmaitm6\KUDENV file on Windows or
the install_dir/config/ud.ini file on UNIX systems. Use one of the following
methods to modify trace options:
v Manually edit the configuration file to set trace logging
1. Open the trace options file:
– On Windows, open the install_dir\tmaitm6\KUDENV file.
– On UNIX systems, open the /install_dir/config/ud.ini file.
2. Edit the line that begins with KBB_RAS1= to set trace logging preferences.
For example, if you want detailed trace logging, set the Maximum Tracing
option:
– On Windows,
KBB_RAS1=ERROR (UNIT:kdd ALL) (UNIT:kud ALL) (UNIT:kra ALL)
– On UNIX systems,
export KBB_RAS1=’ERROR (UNIT:kdd ALL) (UNIT:kud ALL) (UNIT:kra ALL)’
3. Edit the line that begins with KBB_RAS1_LOG= to manage the generation
of log files:
– Edit the following parameters to adjust the number of rolling log files and
their size.
- MAXFILES: the total number of files that are to be kept for all startups
of a given program. Once this value is exceeded, the oldest log files are
discarded. Default value is 9.
- LIMIT: the maximum size, in megabytes (MB) of a RAS1 log file.
Default value is 5.
– IBM Software Support might guide you to modify the following
parameters:
- COUNT: the number of log files to keep in the rolling cycle of one
program startup. Default value is 3.
- PRESERVE: the number of files that are not to be reused in the rolling
cycle of one program startup. Default value is 1.

Note: The KBB_RAS1_LOG parameter also provides for the specification of

the log file directory, log file name, and the inventory control file
directory and name. Do not modify these values or log information
can be lost.
4. Restart the monitoring agent so that your changes take effect.
v (Windows only) Alternate method to edit trace logging parameters:
1. Open the Manage Tivoli Enterprise Monitoring Services window.

Appendix C. Problem determination 185

 2. Right-click the icon of the monitoring agent whose logging you want to
modify.
3. Select Advanced > Edit Trace Parms. The Tivoli Enterprise Monitoring
Server Trace Parameters window is displayed.
4. Select a new trace setting in the pull-down menu in the Enter RAS1 Filters
field or type a valid string.
The selections are as follows:
– No error tracing. KBB_RAS1=-none-
– General error tracing. KBB_RAS1=ERROR
– Intensive error tracing. KBB_RAS1=ERROR (UNIT:kud ALL)
– Maximum error tracing. KBB_RAS1=ERROR (UNIT:kud ALL) (UNIT:kra ALL)

Note: As this example shows, you can set multiple RAS tracing options in
a single statement.
5. Modify the value for ″Maximum Log Size Per File (MB)″ to change the log
file size (changes LIMIT value).
6. Modify the value for ″Maximum Number of Log Files Per Session″ to change
the number of logs files per startup of a program (changes COUNT value).
7. Modify the value for ″Maximum Number of Log Files Total″ to change the
number of logs files for all startups of a program (changes MAXFILES
value).
8. (Optional) Click Y (Yes) in the KDC_DEBUG Setting menu to log
information that can help you diagnose communications and connectivity
problems between the monitoring agent and the monitoring server.

Note: The KDC_DEBUG setting and the Maximum error tracing setting can
generate a large amount of trace logging. Use them only temporarily,
while you are troubleshooting problems. Otherwise, the logs can
occupy excessive amounts of hard disk space.
9. Click OK. You see a message reporting a restart of the monitoring agent so
that your changes take effect.

Problems and workarounds

The following sections provide symptoms and workarounds for problems that
might occur with Monitoring Agent for DB2:
v “Installation and configuration problem determination” on page 187
v “Agent problem determination” on page 192
v “Tivoli Enterprise Portal problem determination” on page 195
v “Workspace problem determination” on page 196
v “Problem determination for remote deployment” on page 195
v “Situation problem determination” on page 196
v “Take Action command problem determination” on page 200

Note: You can resolve some problems by ensuring that your system matches the
system requirements listed in Chapter 2, “Requirements and configuration
for the monitoring agent,” on page 3.
This appendix provides agent-specific problem determination information. See the
IBM Tivoli Monitoring Problem Determination Guide for general problem
determination information.

186 IBM Tivoli Monitoring for Databases: DB2 Agent: User’s Guide
 Installation and configuration problem determination
This section provides tables that show solutions for the following types of
installation, configuration, and uninstallation problems:
v Operating system problems
v Problems with database applications
Table 12. Problems and solutions for installation and configuration for agents that run on UNIX systems
Problem Solution
When you upgrade to IBM Tivoli Fixpacks for Candle, Version 350, are delivered as each monitoring agent
Monitoring, you might need to apply is upgraded to IBM Tivoli Monitoring.
fixpacks to Candle, Version 350, agents. Note: The IBM Tivoli Monitoring download image or CD provides
application fixpacks for the monitoring agents that are installed from
that CD (for example, the agents for operating systems such as
Windows, Linux, UNIX, and i5/OS). The upgrade software for other
agents is located on the download image or CDs for that specific
monitoring agent, such as the agents for database applications.

If you do not upgrade the monitoring agent to IBM Tivoli Monitoring,

the agent continues to work. However, you must upgrade to have all the
functionality that IBM Tivoli Monitoring offers.
Presentation files and customized The upgrade from version 350 to IBM Tivoli Monitoring handles export
Omegamon DE screens for Candle of the presentation files and the customized Omegamon DE screens.
monitoring agents need to be upgraded
to a new Linux on z/Series system.
Non-ASCII characters entered into the Enter only ASCII characters into these fields.
configuration window for the monitoring
agent do not show up or are not the
correct characters.
During the command-line installation, You must exit and restart the installation process. You cannot return to
you choose to install a component that is the list where you selected components to install. When you run the
already installed, and you see the installer again, do not attempt to install any component that is already
following warning: installed. >
WARNING - you are about to install
the SAME version of "component"

where component is the name of the

component that you are attempting to
install.
Note: This problem affects UNIX
command-line installations. If you
monitor only Windows environments,
you would see this problem if you
choose to install a product component
(for example, a monitoring server) on
UNIX.
While installing the agent from a CD, the This error is caused by low disk space. Although the install.sh script
following message is displayed and you indicates that it is ready to install the agent software, the script considers
are not able to continue the installation: the size of all tar files, not the size of all the files that are contained
install.sh warning: unarchive of within the tar file.Run the df -k command to check whether the file
"/cdrom/unix/cienv1.tar" may systems have enough space to install agents.
have failed

Appendix C. Problem determination 187

Table 12. Problems and solutions for installation and configuration for agents that run on UNIX systems (continued)
Problem Solution
Installing as root: The product has been Use the CHMOD command to update the privileges for specific
installed as root, which is not directories in the installation path for IBM Tivoli Monitoring. For
recommended. Without re-installing the example, you might use the root user account to install the monitoring
product, how can you change from root agent and attempt to run the agent as the DB2 instance owner. You
to a different user account? would see an error similar to the following:
db2inst3@aix7%> ./bin/itmcmd Agent -o db2inst3 start ud
CandleAgent : installer level 400 / 100.
find: 0652-023 Cannot open file /home/ITM/images.
find: 0652-023 Cannot open file /home/ITM/images.
find: 0652-023 Cannot open file /home/ITM/images.
find: 0652-023 Cannot open file /home/ITM/images.
CandleAgent : running aix513 jre.
Starting agent...
Agent Started...
db2inst3@aix7%>

Note: In the preceding example, the error is happening for only the
images directory.The monitoring agent runs correctly, in spite of the error
message. To avoid getting the Cannot open file... message, run the
chmod -R 777 <directory> command.
Cannot locate the KDCB0_HOSTNAME Go to install_dir/config and edit the corresponding .ini file. Set the
setting. KDCB0_HOSTNAME parameter followed by the IP address. If you use
multiple network interface cards (NICs), give the Primary IP address of
the network interface.
The Monitoring Agent for DB2 You can collect data to analyze this problem as follows:
repeatedly restarts. 1. Access the install_dir/config/ud.ini file, which is described in
“Setting RAS trace parameters” on page 184.
2. Add the following line: KBB_SIG1=trace –dumpoff
Agents in the monitoring environment Configure both the monitoring server and the Warehouse proxy server to
use different communication protocols. accept multiple protocols, as described in the IBM Tivoli Monitoring
For example, some agents have security Installation and Setup Guide.
enabled and others do not.
Creating a firewall partition file: The How it works: When the agents start, they search KDCPARTITION.TXT
partition file enables an agent to connect for the following matches:
to the monitoring server through a v An entry that matches the partition name OUTSIDE.
firewall.
v An entry that also includes a valid external address.
For more information, see the IBM Tivoli Monitoring Installation and Setup
Guide.
You see the following error: Confirm that the password within the Tivoli Enterprise Monitoring
Hub not registered with location Server is correct.
broker. Error-code 1195.
The monitoring agent is started and Confirm that application support for this monitoring agent has been
running but not displaying data in the added to the Tivoli Enterprise Monitoring Server, and confirm that you
Tivoli Enterprise Portal. have configured the Tivoli Enterprise Portal Server as described in the
IBM Tivoli Monitoring Installation and Setup Guide. Also see the Exploring
IBM Tivoli Monitoring book to learn about overall installation and
configuration issues.

188 IBM Tivoli Monitoring for Databases: DB2 Agent: User’s Guide
Table 12. Problems and solutions for installation and configuration for agents that run on UNIX systems (continued)
Problem Solution
You successfully migrate a Candle Copy the attribute files for the upgraded Candle monitoring agent to
monitoring agent to IBM Tivoli install_dir\tmaitm6\attrlib on the computer where you have installed
Monitoring, Version 6.1.0. However, the Warehouse Proxy. The Warehouse Proxy must be able to access the
when you configure historical data short attribute names for tables and columns. That way, if the longer
collection, you see an error message that versions of these names exceed the limits of the Warehouse database, the
includes, Attribute name may be shorter names can be substituted.
invalid, or attribute file not
installed for warehouse agent.
The monitoring agent does not start in a Check the agent configuration to ensure that all the values are correctly
non-ASCII environment. represented. To view these parameters, go to the Manage Tivoli
Enterprise Monitoring Services window, select the agent template, and
choose the Configure using defaults. From the resulting window, select
and edit the database instance to view its parameters.

Table 13. Problems and solutions for installation and configuration on Windows
Problem Solution
When you upgrade to IBM Tivoli Fixpacks for Candle, Version 350, are delivered as each monitoring agent is
Monitoring, you might need to upgraded to IBM Tivoli Monitoring.
apply fixpacks to Candle, Version Note: The IBM Tivoli Monitoring download image or CD provides application
350, agents. fixpacks for the monitoring agents that are installed from that CD (for example,
the agents for operating systems such as Windows, Linux, UNIX, and i5/OS).
The upgrade software for other agents is located on the download image or CDs
for that specific monitoring agent, such as the agents for database applications.

If you do not upgrade the monitoring agent to IBM Tivoli Monitoring, the agent
continues to work. However, you must upgrade to have all the functionality that
IBM Tivoli Monitoring offers.
Presentation files and customized The upgrade from version 350 to IBM Tivoli Monitoring handles export of the
Omegamon DE screens for presentation files and the customized Omegamon DE screens.
Candle monitoring agents need to
be upgraded to a new Linux on
z/Series system.
Diagnosing problems with When you have problems with browse settings, perform the following steps:
product browse settings. 1. Click on Start > Programs > IBM Tivoli Monitoring > Manage Tivoli
Enterprise Monitoring Services. The Manage Tivoli Enterprise Monitoring
Services is displayed.
2. Right-click the Windows agent and select Browse Settings. A text window is
displayed.
3. Click Save As and save the information in the text file. If requested, you can
forward this file to IBM Software Support for analysis.
A message similar to ″Unable to If a message similar to ″Unable to find running CMS on CT_CMSLIST″ is
find running CMS on displayed in the Log file, the agent is not able to connect to the monitoring
CT_CMSLIST″ in the log file is server. Confirm the following points:
displayed. v Is the computer that hosts the monitoring server up and connected to the
network?
v Do multiple network interface cards (NICs) exist on the system?
v If multiple NICs exist on the system, find out which one is configured for the
monitoring server. Ensure that you specify the correct host name and port
settings for communication in the IBM Tivoli Monitoring environment.

Appendix C. Problem determination 189

Table 13. Problems and solutions for installation and configuration on Windows (continued)
Problem Solution
The system is experiencing high Agent process: View the memory usage of the KUDCMA process on Windows
CPU usage. or the kuddb2 process on UNIX. If CPU usage seems to be excessive, recycle the
monitoring agent.

Network Cards: The network card configurations can decrease the performance
of a system. Each of the stream of packets that a network card receives
(assuming it is a broadcast or destined for the under-performing system) must
generate a CPU interrupt and transfer the data through the I/O bus. If the
network card in question is a bus-mastering card, work can be off-loaded and a
data transfer between memory and the network card can continue without using
CPU processing power. Bus-mastering cards are generally 32-bit and are based
on PCI or EISA bus architectures.

The following table shows problems and solutions for IBM DB2.
Table 14. Problems and solutions for installation and configuration of DB2
Problem Solution
The procedure for launching This problem happens when the trace options file is missing. You can correct the
Monitoring Agent for DB2: Trace problem as follows:
Parameters window in “Setting 1. Create a text file with the following path name:
RAS trace parameters” on page
v On Windows systems: install_dir\tmaitm6\logs\KUDRAS1
184 fails.
v On UNIX systems: /install_dir/config/ud.ini
2. Paste the following configuration setting in the file:
v On Windows:
KBB_RAS1=ERROR ^> C:\IBM\ITM\tmaitm6\logs\KUDRAS1.LOG

Note: If you installed the product in a directory path other than the
default, use that directory path instead of C:\IBM\ITM.
v On UNIX systems:
KBB_RAS1=ERROR

The default installation path on UNIX is /opt/IBM/ITM.

1. Save your changes.
2. Repeat the “Setting RAS trace parameters” on page 184 procedure. Now the
Tivoli Enterprise Monitoring Server: Trace Parameters window is
displayed.
After running the agent Consult the list of supported versions in the configuration chapter of the agent
successfully, you reinstall the user’s guide for DB2. Confirm that you are running a valid version of the target
agent software, and collection of application. If you are monitoring a supported version of the database
monitoring data stops. application, gather log files and other information and contact IBM Software
Support, as described in “Gathering product information for IBM Software
Support” on page 179.
Error counts are displayed in the Check the time stamp for the reports. If you have set up historical data
Alert summary report in the collection for Alert summary report, it is recommended that you set up
Tivoli Enterprise Portal, however, historical data collection for the Alert detail report, too.
error messages are not displayed
in the Alert detail report.
An error window is displayed This problem occurs when non-ASCII characters are entered for the DB2
with the message "Service instance name. Delete the configured Tivoli Enterprise Monitoring Server
executable not found. Cannot instance and recreate an instance specifying an instance name with ASCII
configure service." characters only.

190 IBM Tivoli Monitoring for Databases: DB2 Agent: User’s Guide
Table 15. General problems and solutions for uninstallation
Problem Solution
On Windows, uninstallation of Be sure that you follow the general uninstallation process described in the IBM
IBM Tivoli Monitoring fails to Tivoli Monitoring Installation and Setup Guide:
uninstall the entire environment. 1. Uninstall monitoring agents first, as in the following examples:
v Uninstall a single monitoring agent for a specific database.
—OR—
v Uninstall all instances of a monitoring product, such as IBM Tivoli
Monitoring for Databases.
2. Uninstall IBM Tivoli Monitoring.
The way to remove inactive When you want to remove a managed system from the navigation tree,
managed systems (systems whose right-click the item that you want to remove, and select Remove managed
status is OFFLINE) from the system.
Enterprise navigation tree in the
portal is not obvious.

Unique names for monitoring components

IBM Tivoli Monitoring might not be able to generate a unique name for monitoring
components due to the truncation of names that the product automatically
generates.

IBM Tivoli Monitoring automatically creates a name for each monitoring

component by concatenating the subsystem name, host name, and product code
separated by colons (subsystem_name:hostname:UD).

Note: When you monitor a multinode system, such as a database, IBM Tivoli
Monitoring adds a subsystem name to the concatenated name, typically a
database instance name.
The length of the name that IBM Tivoli Monitoring generates is limited to 32
characters. Truncation can result in multiple components having the same
32-character name. If this problem happens, shorten the hostname portion of the
name as follows:
1. Open the configuration file for the monitoring agent, which is located in the
following path:
v On Windows: install_dir\tmaitm6\Kproduct_codeCMA.INI. For example, the
product code for the Monitoring Agent for Windows OS is NT file name for
is KNTCMA.INI.
v On UNIX and Linux: install_dir/tmaitm6/product_code.ini and
product_code.config. For example, the file names for the Monitoring Agent
for UNIX OS is ux.ini and ux.config.
2. Find the line the begins with CTIRA_HOSTNAME=.
3. Type a new name for host name that is a unique, shorter name for the host
computer. The final concatenated name including the subsystem name, new
host name, and UD, cannot be longer than 32 characters.

Note: You must ensure that the resulting name is unique with respect to any
existing monitoring component that was previously registered with the
Tivoli Enterprise Monitoring Server.
4. Save the file.
5. Restart the agent.
6. If you do not find the files mentioned in Step 1, perform the workarounds
listed in the next paragraph.

Appendix C. Problem determination 191

 If you do not find the files mentioned in the preceding steps, perform the
following workarounds:
1. Change CTIRA_HOSTNAME environment variable in the configuration file of
the monitoring agent.
v Find the KUDENV file in the same path mentioned in the preceding row.
v For z/OS agents, find the RKANPAR library.
v For i5/OS agents, find the QAUTOTMP/KMSPARM library in member
KBBENV.
2. If you cannot find the CTIRA_HOSTNAME environment variable, you must
add it to the configuration file of the monitoring agent:
v On Windows: Use the Advanced > Edit Variables option.
v On UNIX and Linux: Add the variable to the config/product_code.ini and to
config/product_code.config files.
v On z/OS: Add the variable to the RKANPAR library, member
Kproduct_codeENV.
v On i5/OS: Add the variable to the QAUTOTMP/KMSPARM library in
member KBBENV.
3. Some monitoring agents (for example, the monitoring agent for MQ Series) do
not reference the CTIRA_HOSTNAME environment variable to generate
component names. Check the documentation for the monitoring agent that you
are using for information on name generation. If necessary, contact IBM
Software Support.

Agent problem determination

This section lists problems that might occur with agents.

This appendix provides agent-specific problem determination information. See the

IBM Tivoli Monitoring Problem Determination Guide for general problem
determination information.
Table 16. Agent problems and solutions on DB2
Problem Solution
When you edit the configuration The original configuration settings might include non-ASCII characters. These
for an existing monitoring agent, values were stored incorrectly and result in the incorrect display. Enter new
the values displayed are not values using only ASCII characters.
correct.
Column names have different Some attributes for the DB2 agent have length restrictions that cause the short
formats in table views for the column name to be used if the modified column name becomes too long.
DB2 agent.

192 IBM Tivoli Monitoring for Databases: DB2 Agent: User’s Guide
Table 16. Agent problems and solutions on DB2 (continued)
Problem Solution
A problem can arise when you You must reconfigure the previously existing agents to restore their
run multiple agents on one communication connection with TEMS1. For example, you can right-click the
computer and want them to row for a specific agent in the Manage Tivoli Enterprise Monitoring Services, and
communicate with multiple select Reconfigure. See the IBM Tivoli Monitoring Installation and Setup Guide for
monitoring servers, as described more information on reconfiguration.
in this example:
v Agents are running on
computer and communicating
with a Tivoli Enterprise
Monitoring Server, called
TEMS1.
v You install a new agent on the
same computer and you want
this agent to communicate
with a different monitoring
server, called TEMS2.
v When you configure the new
agent to communicate with
TEMS2, all the existing agents
are re-configured to
communicate with TEMS2.

Table 17. Agent problems and solutions

Problem Solution
You want to have multiple If you plan to install and run the DB2 agent on one computer, they can use the
instances of the DB2 agent same network interface because they run as different processes.However, if you
running on the same system but want to have two DB2 agents on the same computer or want to run two
talking to different Tivoli instances of each agent, install two-network adapters. Each instance is configured
Enterprise Monitoring Server. for the host specific address so they can be recognized in the configuration
settings.
The agent is installed and This problem can occur when the root user account is used to install and start
running normally. After the agent. Verify whether you have used the root user account to install the
rebooting the computer, where agent.To change the user account from root to some other user account, see
Tivoli Enterprise Monitoring Installing as root.
Server was running, the agent is
not online.

Appendix C. Problem determination 193

Table 17. Agent problems and solutions (continued)
Problem Solution
Agent unable to connect: The This error message means that the agent is not able to connect to the computer
agent is started, but no reports where the Tivoli Enterprise Monitoring Server is running. The reason might be
are displayed on Tivoli any one of the following:
Enterprise Monitoring Server. Computer where the Tivoli Enterprise Monitoring Server is running is down
The log file includes the Ping the computer where the Tivoli Enterprise Monitoring Server is running
following error: and make sure that it is up and running.
Unable to find running CMS on Tivoli Enterprise Monitoring Server is not running
CMSLIST or Endpoint If the Tivoli Enterprise Monitoring Server is not running, recycle the Tivoli
unavailable Enterprise Monitoring Server and verify whether the agent is connecting.
Multiple NIC Cards on the computer where the Tivoli Enterprise Monitoring
Server is running.
If multiple NICs are installed on the computer where the Tivoli Enterprise
Monitoring Server is running, identify the Primary NIC and use the hostname
or IP address.
Verify that the Tivoli Enterprise Monitoring Server has been configured with
the Primary NIC’s IP address or hostname.
If you are using hostname, make sure that /etc/hosts has a valid entry for the
Primary NICs host name and its IP address.
On the Tivoli Enterprise Monitoring Server set the KDCB0_HOSTNAME
variable to the primary IP address of the computer. Use the same address to
configure the agent.
To connect to the Tivoli Enterprise Monitoring Server, configure the agent
with Primary NIC’s IP address or host name of the computer where the
Tivoli Enterprise Monitoring Server is running.
While configuring the agent, make sure that the port number that you are
connecting to is correct. If you are not using the default port number, make
sure that you are using the same port number used in Tivoli Enterprise
Monitoring Server. For more information, see the IBM Tivoli Monitoring
Installation and Setup Guide.
Agent is behind the Firewall
If you use a Firewall, identify whether you have any one of the following
scenarios:
v Hub monitoring server INSIDE, and agents OUTSIDE
v Hub and remote monitoring servers INSIDE, agents OUTSIDE
v Hub monitoring server INSIDE, remote monitoring server and agents
OUTSIDE
See Creating a firewall partition file for information about the
KDC_PARTITION file that enables communication across a firewall. For
additional information, see the IBM Tivoli Monitoring Installation and Setup
Guide.
Connecting to the monitoring server through a Virtual Private Network (VPN)
In some cases, the agent or a remote monitoring server needs to connect to
the hub monitoring server through a VPN. You must configure the
communication channel (pipe) to be ephemeral, as in the following example:
KDC_FAMILIES=ip.pipe port:port_number ephemeral:y ip use:n sna use:n

194 IBM Tivoli Monitoring for Databases: DB2 Agent: User’s Guide
Table 17. Agent problems and solutions (continued)
Problem Solution
When you restart the system that Manually start the monitoring agent.
hosts the Tivoli Enterprise
Monitoring Server, the DB2 agent
does not start automatically.
However, when you use
CandleAgent start, the agent
starts and continues running.

Tivoli Enterprise Portal problem determination

Table 18 lists problems that might occur with the Tivoli Enterprise Portal. This
appendix provides agent-specific problem determination information. See the IBM
Tivoli Monitoring Problem Determination Guide for general problem determination
information.
Table 18. Tivoli Enterprise Portal problems and solutions
Problem Solution
Historical data collection is The column, Sort By, Group By, and First/Last functions are not compatible with
unavailable because of the historical data collection feature. Use of these advanced functions will make a
incorrect queries in the Tivoli query ineligible for historical data collection.
Enterprise Portal.
Even if data collection has been started, you cannot use the time span feature if the
query for the chart or table includes any column functions or advanced query
options (Sort By, Group By, First / Last).

To ensure support of historical data collection, do not use the Sort By, Group By, or
First/Last functions in your queries.

See the IBM Tivoli Monitoring Administrator’s Guide the Tivoli Enterprise Portal
online Help for information on the Historical Data Collection function.
When you use a long process Truncation of process names in the portal display is the expected behavior. 64 bytes
name in the situation, the is the maximum name length.
process name is truncated.

Problem determination for remote deployment

Table 19 lists problems that might occur with remote deployment. This appendix
provides agent-specific problem determination information. See the IBM Tivoli
Monitoring Problem Determination Guide for general problem determination
information.

This section describes problems and solutions for remote deployment and removal
of agent software Agent Remote Deploy:
Table 19. Remote deployment problems and solutions
Problem Solution
While you are using the remote deployment feature Do not close or modify this window. It is part of the
to install Monitoring Agent for DB2, an empty installation process and will be dismissed automatically.
command window is displayed on the target
computer. This problem occurs when the target of
remote deployment is a Windows computer. (See the
IBM Tivoli Monitoring Installation and Setup Guide for
more information on the remote deployment
feature.)

Appendix C. Problem determination 195

Table 19. Remote deployment problems and solutions (continued)
Problem Solution
The removal of a monitoring agent fails when you This problem might happen when you attempt the remote
use the remote removal process in the Tivoli removal process immediately after you have restarted the
Enterprise Portal desktop or browser. Tivoli Enterprise Monitoring Server. You must allow time for
the monitoring agent to refresh its connection with the Tivoli
Enterprise Monitoring Server before you begin the remote
removal process.

Workspace problem determination

Table 20 shows problems that might occur with workspaces. This appendix
provides agent-specific problem determination information. See the IBM Tivoli
Monitoring Problem Determination Guide for general problem determination
information.
Table 20. Workspace problems and solutions
Problem Solution
The name of the attribute does not display in a When a chart or graph view that includes the attribute is scaled
bar chart or graph view. to a small size, a blank space is displayed instead of a truncated
name. To see the name of the attribute, expand the view of the
chart until there is sufficient space to display all characters of the
attribute’s name.
You see the following message: KFWITM083W You see this message because some links do not have default
Default link is disabled for the selected workspaces. Right-click the link to access a list of workspaces to
object; please verify link and link anchor select.
definitions.
At the bottom of each view, you see the Ensure that you configure all groups that supply data to the
following Historical workspace KFWITM220E view. In the Historical Configuration view, ensure that data
error: Request failed during execution. collection is started for all groups that supply data to the view.
You start collection of historical data but the data Managing options for historical data collection:
cannot be seen. v Basic historical data collection populates the Warehouse with
raw data. This type of data collection is turned off by default.
See Chapter 2, “Requirements and configuration for the
monitoring agent,” on page 3 for information on managing
this feature including how to set the interval at which data is
collected. By setting a more frequent interval for data
collection you reduce the load on the system incurred every
time data is uploaded.
v You use the Summarization and Pruning monitoring agent to
collect specific amounts and types of historical data. Be aware
that historical data is not displayed until the Summarization
and Pruning monitoring agent begins collecting the data. By
default, this agent begins collection at 2 AM daily. At that
point, data is visible in the workspace view. See the IBM Tivoli
Monitoring Administrator’s Guide to learn how to modify the
default collection settings.

Situation problem determination

This section provides information about both general situation problems and
problems with the configuration of situations. See the IBM Tivoli Monitoring
Problem Determination Guide for more information about problem determination for
situations.

196 IBM Tivoli Monitoring for Databases: DB2 Agent: User’s Guide
 General situation problems
Table 21 lists problems that might occur with specific situations.
Table 21. Specific situation problems and solutions
Problem Solution
You want to change the appearance of 1. Right-click an item in the Navigation tree.
situations when they are displayed in a
2. Select Situations in the pop-up menu. The Situation Editor window is
Workspace view.
displayed.
3. Select the situation that you want to modify.
4. Use the Status pull-down menu in the lower right of the window to
set the status and appearance of the Situation when it triggers.
Note: This status setting is not related to severity settings in IBM
Tivoli Enterprise Console.
Monitoring activity requires too much Check the RAS trace logging settings that are described in “Setting RAS
disk space. trace parameters” on page 184. For example, trace logs grow rapidly
when you apply the ALL logging option.
Monitoring activity requires too many Table 22 on page 198 describes the performance impact of specific
system resources. attribute groups. If possible, decrease your use of the attribute groups that
require greater system resources.
A formula that uses mathematical This formula is incorrect because situation predicates support only logical
operators appears to be incorrect. For operators. Your formulas cannot have mathematical operators.
example, if you were monitoring Linux, Note: The Situation Editor provides alternatives to math operators.
a formula that calculates when Free Regarding the example, you can select % Memory Free attribute and
Memory falls under 10 percent of Total avoid the need for math operators.
Memory does not work: LT
#’Linux_VM_Stats.Total_Memory’ / 10
If you are running a Version 350 Access the database detail. In the ″Release″ section change the version
Monitoring Agent for DB2 and you setting for the agent from 610 to 350. To enable Unicode and other
choose to alter the views to include a features, upgrade the monitoring agent to IBM Tivoli Monitoring, Version
Version 610 UNICODE attribute, be 6.1.0.
aware that data for this attribute is not
displayed and you see a blank column
in this view.
Situations that you create display the For a situation to have the correct severity in TEC for those situations
severity UNKNOWN in IBM Tivoli which are not mapped, you need to ensure that an entry exists in the
Enterprise Console. tecserver.txt file for the situation and that SEVERITY is specified.

See the “Configuring Tivoli Enterprise Console integration” chapter in the

IBM Tivoli Monitoring Administrator’s Guide for more information.
You see the 'Unable to get attribute Ensure that the agent attribute files are installed on the Tivoli Enterprise
name' error in the Tivoli Enterprise Monitoring Server.
Monitoring Server log after creating a
situation. The following example shows a typical log entry when you have this
problem:
(4320916A.0049-F60:kfaottev.c,1572,"Translate_ResultBuffer") \
Unable to get attribute name for tablename/column \
<UAG524400.UA4>. Ignored.

Consider performance impact of each attribute group: Table 22 on page 198 lists
the impact on performance (high, medium, or low) of each attribute group. The
multiple-instance attributes have been classified at the lowest level. That is, the
performance overhead will increase if you do not specify compare values for one
or more key values.

Appendix C. Problem determination 197

 When you want to prevent impact on performance by any of the attribute groups
listed in Table 22 you must avoid referencing that attribute group, as suggested in
this list:
v Disable the attribute group.
v Never select workspaces that reference the attribute group.
v Disable situations that reference the attribute group by using the ″Undistributed
situations″ option in the Situation Editor.
v Disable historical reporting that references the attribute group.
v Avoid using the ″Auto Refresh″ refresh feature in a Workspace because this
option causes a refresh of data for all attribute groups.
See the IBM Tivoli Monitoring User’s Guide for additional information on controlling
attribute group usage.
Table 22. Performance Impact by attribute group
Attribute group High Medium Low
Application (group 00, deprecated U
application attributes)
Application (group 00, Unicode application U
attributes)
Application (group 01) U
Buffer Pool Data U
Database (group 00) U
Database (group 01) U
Locking Conflict U
System Overview U
Tablespace U

Problems with configuration of situations

Table 23 lists problems that might occur with situations.

This section provides information for problem determination for agents. Be sure to
consult the IBM Tivoli Monitoring Problem Determination Guide for more general
problem determination information.
Table 23. Problems with configuring situations that you solve in the Situation Editor
Problem Solution
Note: To get started with the solutions in this section, perform these steps:
1. Launch the Tivoli Enterprise Portal.
2. Click Edit > Situation Editor.
3. In the tree view, choose the agent whose situation you want to modify.
4. Choose the situation in the list. The Situation Editor view is displayed.
The situation for a specific agent is Open the Situation Editor. Access the All managed servers view. If the situation
not visible in the Tivoli Enterprise is absent, confirm that application support for Monitoring Agent for DB2 has
Portal. been added to the monitoring server. If not, add application support to the
server, as described in the IBM Tivoli Monitoring Installation and Setup Guide.
The monitoring interval is too Access the Situation Editor view for the situation that you want to modify.
long. Check the Sampling interval area in the Formula tab. Adjust the time interval
as needed.

198 IBM Tivoli Monitoring for Databases: DB2 Agent: User’s Guide
Table 23. Problems with configuring situations that you solve in the Situation Editor (continued)
Problem Solution
The situation did not activate at Manually recycle the situation as follows:
startup. 1. Right-click the situation and choose Stop Situation.
2. Right-click the situation and choose Start Situation.
Note: You can permanently avoid this problem by placing a check mark in the
Run at Startup option of the Situation Editor view for a specific situation.
The situation is not displayed. Click the Action tab and check whether the situation has an automated
corrective action. This action can occur directly or through a policy. The
situation might be resolving so quickly that you do not see the event or the
update in the graphical user interface.
An Alert event has not occurred Check the logs, reports, and workspaces.
even though the predicate has been
properly specified.
A situation fires on an unexpected Confirm that you have distributed and started the situation on the correct
managed object. managed system.
The product did not distribute the Click the Distribution tab and check the distribution settings for the situation.
situation to a managed system.
The situation does not fire. In the Formula tab, analyze predicates as follows:
1. Click the fx icon in the upper-right corner of the Formula area. The Show
Incorrect predicates are present in
formula window is displayed.
the formula that defines the
situation. For example, the a. Confirm the following details in the Formula area at the top of the
managed object shows a state that window:
normally triggers a monitoring v The attributes that you intend to monitor are specified in the formula.
event, but the situation is not true v The situations that you intend to monitor are specified in the formula.
because the wrong attribute is v The logical operators in the formula match your monitoring goal.
specified in the formula. v The numerical values in the formula match your monitoring goal.
b. (Optional) Click the Show detailed formula check box in the lower left
of the window to see the original names of attributes in the application
or operating system that you are monitoring.
c. Click OK to dismiss the Show formula window.
2. (Optional) In the Formula area of the Formula tab, temporarily assign
numerical values that will immediately trigger a monitoring event. The
triggering of the event confirms that other predicates in the formula are
valid.
Note: After you complete this test, you must restore the numerical values
to valid levels so that you do not generate excessive monitoring data based
on your temporary settings.

Table 24. Problems with configuration of situations that you solve in the Workspace area
Problem Solution
Situation events are not displayed Associate the situation with a workspace.
in the Events Console view of the Note: The situation does not need to be displayed in the workspace. It is
workspace. sufficient that the situation be associated with any workspace.
You do not have access to a Note: You must have administrator privileges to perform these steps.
situation. 1. Select Edit > Administer Users to access the Administer Users window.
2. In the Users area, select the user whose privileges you want to modify.
3. In the Permissions tab, Applications tab, and Navigator Views tab, select
the permissions or privileges that correspond to the user’s role.
4. Click OK.

Appendix C. Problem determination 199

Table 24. Problems with configuration of situations that you solve in the Workspace area (continued)
Problem Solution
A managed system seems to be 1. Select Physical View and highlight the Enterprise Level of the navigator
offline. tree.
2. Select View > Workspace > Managed System Status to see a list of
managed systems and their status.
3. If a system is offline, check network connectivity and status of the specific
system or application.

Table 25. Problems with configuration of situations that you solve in the Manage Tivoli Enterprise Monitoring Services
window
Problem Solution
After an attempt to restart the For UNIX, NetWare, or Windows, log on to the applicable system and perform
agents in the Tivoli Enterprise the appropriate queries.
Portal, the agents are still not
running.
The Tivoli Enterprise Monitoring Check the system status and check the appropriate IBM Tivoli Monitoring logs.
Server is not running.
The managed objects you created Check the managed system distribution on both the situation and the managed
are firing on incorrect managed object settings sheets.
systems.
For agents that can have multiple sub-nodes, such as database agents:
The icon is incorrect. Check the icon assignments in the template.
The situation is not assigned to a Check the situation assignments in the template of the associated managed
state in the template. object.
You assigned the situation to an Check the State settings sheet for the template.
incorrect state in the template.

Take Action command problem determination

Table 26 lists general problems that might occur with Take Action commands.
When each Take Action command runs it generates the log file listed in Table 11 on
page 182. This appendix provides agent-specific problem determination
information. See the IBM Tivoli Monitoring Problem Determination Guide for general
problem determination information.
Table 26. Take Action commands problems and solutions
Problem Solution
Some Take Action commands require several Allow several minutes. If you do not see a pop-up message
minutes to complete. advising you of completion, try to run the command manually.
Consult the log file for the Take Action command, name.log,
where name is the name of the Take Action command.

Support information
If you have a problem with your IBM software, you have the following options for
obtaining support for software products:
v “Searching knowledge bases” on page 201
v “Obtaining fixes” on page 201
v “Receiving weekly support updates” on page 202

200 IBM Tivoli Monitoring for Databases: DB2 Agent: User’s Guide
 v “Contacting IBM Software Support” on page 202

Searching knowledge bases

You can search the available knowledge bases to determine whether your problem
was already encountered and is already documented.

Searching the information center

Note: If you print PDF documents on other than letter-sized paper, set the option
in the File > Print window that allows Adobe Reader to print letter-sized
pages on your local paper.

The documentation CD contains the publications that are in the product library.
The format of the publications is PDF, HTML, or both.

IBM posts publications for this and all other Tivoli products, as they become
available and whenever they are updated, to the Tivoli software information center
Web site. Access the Tivoli software information center by first going to the Tivoli
software library at the following Web address:

http://www.ibm.com/software/tivoli/library

Scroll down and click the Product manuals link. In the Tivoli Technical Product
Documents Alphabetical Listing window, click M to access all of the IBM Tivoli
Monitoring product manuals.

Searching the Internet

If you cannot find an answer to your question in the information center, search the
Internet for the latest, most complete information that might help you resolve your
problem.

The IBM Software Support Web site provides the latest information about known
product limitations and workarounds in the form of technotes for your product.
You can view this information at the following Web site:

http://www.ibm.com/software/support

To search for information on IBM products through the Internet (for example, on
Google), be sure to consider the following types of documentation:
v IBM technotes
v IBM downloads
v IBM Redbooks
v IBM developerWorks
v Forums and newsgroups

Obtaining fixes
A product fix might be available to resolve your problem. To determine what fixes
are available for your IBM software product, follow these steps:
1. Go to the Software support Web site at
http://www.ibm.com/software/support.
2. Click the Download tab.
3. Select the operating system in the Operating system menu.
4. Type search terms in the Enter search terms field.

Appendix C. Problem determination 201

 5. As appropriate, use other search options to further define your search.
6. Click Search.
7. From the list of downloads returned by your search, click the name of a fix to
read the description of the fix and to optionally download the fix.

For more information about the types of fixes that are available, see the IBM
Software Support Handbook at
http://techsupport.services.ibm.com/guides/handbook.html.

Receiving weekly support updates

To receive weekly e-mail notifications about fixes and other software support news,
follow these steps:
1. Go to the IBM Software Support Web site at
http://www.ibm.com/software/support.
2. Click My account in the upper right corner of the page.
3. Click Subscribe to IBM e-news. (If you have already subscribed and want to
modify your subscription preferences, click Modify subscriptions and follow
the instructions on screen.)
4. Follow the instructions on screen to provide the following data:
v Your personal contact information.
v Your areas of interest.
v The types of subscriptions and regional versions that you want to receive.
5. Review the subscription confirmation to confirm your settings.

Contacting IBM Software Support

IBM Software Support provides assistance with product defects.

Before contacting IBM Software Support, your company must have an active IBM
software maintenance contract, and you must be authorized to submit problems to
IBM. The type of software maintenance contract that you need depends on the
type of product you have:
v For IBM distributed software products (including, but not limited to, Tivoli,
Lotus, and Rational products, as well as DB2 and WebSphere products that run
on Windows, or UNIX operating systems), enroll in Passport Advantage in one
of the following ways:
Online
Go to the Passport Advantage Web site at
http://www.lotus.com/services/passport.nsf/
WebDocs/Passport_Advantage_Home and click How to Enroll.
By phone
For the phone number to call in your country, go to the IBM Software
Support Web site at
http://techsupport.services.ibm.com/guides/contacts.html and click the
name of your geographic region.
v For customers with Subscription and Support (S & S) contracts, go to the
Software Service Request Web site at
https://techsupport.services.ibm.com/ssr/login.
v For customers with IBMLink, CATIA, Linux, OS/390, iSeries, pSeries, z/Series,
and other support agreements, go to the IBM Support Line Web site at
http://www.ibm.com/services/us/index.wss/so/its/a1000030/dt006.

202 IBM Tivoli Monitoring for Databases: DB2 Agent: User’s Guide
v For IBM eServer software products (including, but not limited to, DB2 and
WebSphere products that run in z/Series, pSeries, and iSeries environments),
you can purchase a software maintenance agreement by working directly with
an IBM sales representative or an IBM Business Partner. For more information
about support for eServer software products, go to the IBM Technical Support
Advantage Web site http://www.ibm.com/servers/eserver/techsupport.html.

If you are not sure what type of software maintenance contract you need, call
1-800-IBMSERV (1-800-426-7378) in the United States. From other countries, go to
the contacts page of the IBM Software Support Handbook on the Web at
http://techsupport.services.ibm.com/guides/contacts.html and click the name of
your geographic region for phone numbers of people who provide support for
your location.

To contact IBM Software support, follow these steps:

1. “Determining the business impact”
2. “Describing problems and gathering information”
3. “Submitting problems” on page 204

Determining the business impact

When you report a problem to IBM, you are asked to supply a severity level.
Therefore, you need to understand and assess the business impact of the problem
that you are reporting. Use the following criteria:
Severity 1
The problem has a critical business impact. You are unable to use the
program, resulting in a critical impact on operations. This condition
requires an immediate solution.
Severity 2
The problem has a significant business impact. The program is usable, but
it is severely limited.
Severity 3
The problem has some business impact. The program is usable, but less
significant features (not critical to operations) are unavailable.
Severity 4
The problem has minimal business impact. The problem causes little impact
on operations, or a reasonable circumvention to the problem was
implemented.

Describing problems and gathering information

When describing a problem to IBM, be as specific as possible. Include all relevant
background information so that IBM Software Support specialists can help you
solve the problem efficiently. To save time, know the answers to these questions:
v What software versions were you running when the problem occurred?
v Do you have logs, traces, and messages that are related to the problem
symptoms? IBM Software Support is likely to ask for this information.
v Can you re-create the problem? If so, what steps were performed to re-create the
problem?
v Did you make any changes to the system? For example, did you make changes
to the hardware, operating system, networking software, and so on.
v Are you currently using a workaround for the problem? If so, be prepared to
explain the workaround when you report the problem.

Appendix C. Problem determination 203

 See “Gathering product information for IBM Software Support” on page 179 for
further tips for gathering information for IBM Software Support.

Submitting problems
You can submit your problem to IBM Software Support in one of two ways:
Online
Click Submit and track problems on the IBM Software Support site at
http://www.ibm.com/software/support/probsub.html. Type your
information into the appropriate problem submission form.
By phone
For the phone number to call in your country, go to the contacts page of
the IBM Software Support Handbook at
http://techsupport.services.ibm.com/guides/contacts.html and click the
name of your geographic region.

If the problem you submit is for a software defect or for missing or inaccurate
documentation, IBM Software Support creates an Authorized Program Analysis
Report (APAR). The APAR describes the problem in detail. Whenever possible,
IBM Software Support provides a workaround that you can implement until the
APAR is resolved and a fix is delivered. IBM publishes resolved APARs on the
Software Support Web site daily, so that other users who experience the same
problem can benefit from the same resolution.

204 IBM Tivoli Monitoring for Databases: DB2 Agent: User’s Guide
Appendix D. Accessibility
Accessibility features help users with physical disabilities, such as restricted
mobility or limited vision, to use software products successfully. The major
accessibility features in this product enable users to do the following:
v Use assistive technologies, such as screen-reader software and digital speech
synthesizer, to hear what is displayed on the screen. Consult the product
documentation of the assistive technology for details on using those technologies
with this product.
v Operate specific or equivalent features using only the keyboard.
v Magnify what is displayed on the screen.

In addition, the product documentation was modified to include the following

features to aid accessibility:
v All documentation is available in both HTML and convertible PDF formats to
give the maximum opportunity for users to apply screen-reader software.
v All images in the documentation are provided with alternative text so that users
with vision impairments can understand the contents of the images.

Navigating the interface using the keyboard

Standard shortcut and accelerator keys are used by the product and are
documented by the operating system. Refer to the documentation provided by
your operating system for more information.

Magnifying what is displayed on the screen

You can enlarge information on the product windows using facilities provided by
the operating systems on which the product is run. For example, in a Microsoft
Windows environment, you can lower the resolution of the screen to enlarge the
font sizes of the text on the screen. Refer to the documentation provided by your
operating system for more information.

© Copyright IBM Corp. 2005 205

206 IBM Tivoli Monitoring for Databases: DB2 Agent: User’s Guide
Appendix E. Notices
This information was developed for products and services offered in the U.S.A.
IBM not offer the products, services, or features discussed in this document in
other countries. Consult your local IBM representative for information on the
products and services currently available in your area. Any reference to an IBM
product, program, or service is not intended to state or imply that only that IBM
product, program, or service be used. Any functionally equivalent product,
program, or service that does not infringe any IBM intellectual property right be
used instead. However, it is the user’s responsibility to evaluate and verify the
operation of any non-IBM product, program, or service.

IBM have patents or pending patent applications covering subject matter described
in this document. The furnishing of this document does not give you any license to
these patents. You can send license inquiries, in writing, to:

IBM Director of Licensing

IBM Corporation
North Castle Drive
Armonk, NY 10504-1785 U.S.A.

For license inquiries regarding double-byte (DBCS) information, contact the IBM
Intellectual Property Department in your country or send inquiries, in writing, to:

IBM World Trade Asia Corporation

Licensing
2-31 Roppongi 3-chome, Minato-ku
Tokyo 106, Japan

The following paragraph does not apply to the United Kingdom or any other
country where such provisions are inconsistent with local law:

INTERNATIONAL BUSINESS MACHINES CORPORATION PROVIDES THIS

PUBLICATION ″AS IS″ WITHOUT WARRANTY OF ANY KIND, EITHER
EXPRESS OR IMPLIED, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED
WARRANTIES OF NON-INFRINGEMENT, MERCHANTABILITY OR FITNESS
FOR A PARTICULAR PURPOSE.

Some states do not allow disclaimer of express or implied warranties in certain

transactions, therefore, this statement might not apply to you.

This information could include technical inaccuracies or typographical errors.

Changes are periodically made to the information herein; these changes will be
incorporated in new editions of the publication. IBM make improvements and/or
changes in the product(s) and/or the program(s) described in this publication at
any time without notice.

Any references in this information to non-IBM Web sites are provided for
convenience only and do not in any manner serve as an endorsement of those Web
sites. The materials at those Web sites are not part of the materials for this IBM
product and use of those Web sites is at your own risk.

© Copyright IBM Corp. 2005 207

 IBM use or distribute any of the information you supply in any way it believes
appropriate without incurring any obligation to you.

Licensees of this program who wish to have information about it for the purpose
of enabling: (i) the exchange of information between independently created
programs and other programs (including this one) and (ii) the mutual use of the
information which has been exchanged, should contact:

IBM Corporation
2Z4A/101
11400 Burnet Road
Austin, TX 78758 U.S.A.

Such information be available, subject to appropriate terms and conditions,

including in some cases payment of a fee.

The licensed program described in this document and all licensed material
available for it are provided by IBM under terms of the IBM Customer Agreement,
IBM International Program License Agreement or any equivalent agreement
between us.

Any performance data contained herein was determined in a controlled

environment. Therefore, the results obtained in other operating environments vary
significantly. Some measurements have been made on development-level systems
and there is no guarantee that these measurements will be the same on generally
available systems. Furthermore, some measurement have been estimated through
extrapolation. Actual results vary. Users of this document should verify the
applicable data for their specific environment.

Information concerning non-IBM products was obtained from the suppliers of

those products, their published announcements or other publicly available sources.
IBM has not tested those products and cannot confirm the accuracy of
performance, compatibility or any other claims related to non-IBM products.
Questions on the capabilities of non-IBM products should be addressed to the
suppliers of those products.

All statements regarding IBM’s future direction or intent are subject to change or
withdrawal without notice, and represent goals and objectives only.

This information is for planning purposes only. The information herein is subject to
change before the products described become available.

This information contains examples of data and reports used in daily business
operations. To illustrate them as completely as possible, the examples include the
names of individuals, companies, brands, and products. All of these names are
fictitious and any similarity to the names and addresses used by an actual business
enterprise is entirely coincidental.

COPYRIGHT LICENSE:

This information contains sample application programs in source language, which

illustrate programming techniques on various operating systems. You copy, modify,
and distribute these sample programs in any form without payment to IBM, for
the purposes of developing, using, marketing or distributing application programs
conforming to the application programming interface for the operating system for
which the sample programs are written. These examples have not been thoroughly

208 IBM Tivoli Monitoring for Databases: DB2 Agent: User’s Guide
 tested under all conditions. IBM, therefore, cannot guarantee or imply reliability,
serviceability, or function of these programs. You copy, modify, and distribute these
sample programs in any form without payment to IBM for the purposes of
developing, using, marketing, or distributing application programs conforming to
IBM’s application programming interfaces.

If you are viewing this information in softcopy form, the photographs and color
illustrations might not appear.

Trademarks
IBM, the IBM logo, AIX, Candle, CandleNet Portal, DB2, DB2 Connect, DB2
Universal Database, developerWorks®, Distributed Relational Database
Architecture, DRDA, eServer™, iSeries™, Lotus®, MVS, OMEGAMON, OS/2,
OS/390®, OS/400, Passport Advantage®, pSeries®, Rational®, Redbooks™, Tivoli,
the Tivoli logo, Tivoli Enterprise, Tivoli Enterprise Console, VTAM®, WebSphere®,
and zSeries® are trademarks or registered trademarks of International Business
Machines Corporation in the United States, other countries, or both.

Microsoft, Windows, and Windows NT are registered trademarks of Microsoft

Corporation in the United States, other countries, or both.

UNIX is a registered trademark of The Open Group in the United States and other
countries.

Other company, product, and service names be trademarks or service marks of

others.

Appendix E. Notices 209

210 IBM Tivoli Monitoring for Databases: DB2 Agent: User’s Guide
Index
A Buffer Pool (continued)
workspaces (continued)
accessibility x, 205 list 15
actions Buffer Pool Data attribute group 68
See Take Action commands Buffer Pool Detail workspace 20
agent Buffer Pool workspace 156
problem determination 192 built-in problem determination features 179
trace logs 181
agent deployment 4
agent installation problems 187
Application C
workspaces calculate historical data disk space 128
list 15 capacity planning for historical data 128
Application attribute group (group 00) 28 collecting data 13
Application attribute group (group 01) 65 commands, Take Action 143
Application attribute group (Unicode group 00) 46 components 2
Application I/O Activity workspace 16 configuration 3
Application Identification workspace 16 Connection workspace 24
Application Lock Activity workspace 17 contacting support 202
Application Overview workspace 17 conventions
Application Package and Catalog Cache Activity operating system xi
workspace 17 typeface x
Application Sort and Hash Join Activity workspace 18 customer support
Application SQL Activity workspace 18 See support
Application SQL Statement Text workspace 18 customizing
Application Summary workspace 19 monitoring environment 11
Application Time Information workspace 19 situations 12
Application Top Ten Summary
workspaces
descriptions 19 D
Application Top Ten Summary workspace 155 data
application version requirements 3 collecting 13
attribute groups trace logs 180
Application (group 00) 28 viewing 13
Application (group 01) 65 data provider
Application (Unicode group 00) 46 See agent
Buffer Pool Data 68 Database
Database Information (group 00) 75 workspaces
Database Information (group 01) 92 list 16
list of all 28 database agent installation problems 187
Locking Conflict 105 Database Bottom Ten Summary by BP Hit Ratio
more information 27 workspace 21
overview 27 Database I/O Activity workspace 22
performance impact 197 Database Identification / Status / Logging Information
System Overview 110 workspace 21
Tablespace 120 Database Information attribute group (group 00) 75
attributes Database Information attribute group (group 01) 92
more information 27 Database Lock Activity workspace 22
overview 27 Database Overview workspace 156
Database Pkg/Cat Cache workspace 22
Database Sort/Hash Join Activity workspace 23
B Database SQL Activity workspace 23
Backup Database action 144 Database Summary by BP Hit Ratio workspace 23
books Database workspace 21
feedback viii DB2 properties 4
online viii deployment, agent 4
ordering viii detecting problems, modifying situation values 12
see publications ix directory names, notation xi
Buffer Pool disk capacity planning for historical data 128
workspaces disk space requirements 3
descriptions 20

© Copyright IBM Corp. 2005 211

E K
education knowledge bases for support 201
see Tivoli technical training x
environment
customizing 11
features 1
L
legal notices 207
monitoring real-time 9
Locking Conflict
real-time monitoring 9
workspaces
environment variables, notation xi
descriptions 24
event
list 16
mapping 159
Locking Conflict attribute group 105
events
logging
investigating 10
agent trace logs 181, 182
workspaces 10
built-in features 179
installation log files 181
location and configuration of logs 180
F trace log files 180
features, Monitoring Agent for DB2 1
files
agent trace 181
installation trace 181
M
manuals
other trace log 182
feedback viii
trace logs 180
online viii
fixes, obtaining 201
ordering viii
see publications ix
memory requirements 3
G messages
gathering support information 179 built-in features 179
General Information workspace 24 modifying situation values to detect problems 12
Monitoring Agent for DB2
components 2
H features 1
purposes 9
historical data
using 9
calculate disk space 128
monitoring, viewing the real-time environment 9
disk capacity planning 128
historical data, collecting and viewing 13

N
I notation
environment variables xi
IBM Software Support
path names xi
See support
typeface xi
IBM Tivoli Enterprise Console
event mapping 159
IBM Tivoli Monitoring for Databases: DB2 Agent
performance considerations 197 O
information centers for support 201 ODBC source 6
information, additional online publications
attributes 27 accessing ix
policies 153 for support 201
procedural 9 operating systems 3
situations 131 operation of resource, recovering 10
Take Action commands 143 ordering publications ix
workspaces 15 other requirements 4
installation 3
log file 181
problems 187
interface, user 2
P
path names, for trace logs 180
problem determination for Tivoli Enterprise Portal 195
path names, notation xi
Internet
performance considerations 197
for product support 201
performance impact
investigating an event 10
attribute groups 197
policies
list of all 153
more information 153

212 IBM Tivoli Monitoring for Databases: DB2 Agent: User’s Guide
policies (continued)
overview 153
S
predefined 153 situations
problem determination 179, 186 general problem determination 198
agents 192 list of all 132
built-in features 179 more information 131
describing problems 203 overview 131
determining business impact 203 predefined 132
information centers for 201 specific problem determination 197
installation 187 UDB_Agent_WaitToken_High 133
installation logs 181 UDB_Agents_Stolen_High 133
knowledge bases for 201 UDB_Appl_BP_Hit_Ratio_Low 133
remote deployment 195 UDB_Appl_CatCache_Hit_Low 134
situations 196, 198 UDB_Appl_HJoinOflws_High 134
submitting problems 204 UDB_Appl_HJoinSmOflw_High 134
Take Action commands 200 UDB_Appl_Lock_Warning 134
Tivoli Enterprise Portal 195 UDB_Appl_PkgCache_Hit_Low 134
uninstallation 187 UDB_Appl_SQL_Fail_High 134
uninstallation logs 181 UDB_Appl_Wait_Lock 134
workspaces 196 UDB_BP_DrtyPg_Steal_Clns 134
problems UDB_BP_DrtyPg_thresh_Clns 135
detecting 12 UDB_BP_Hit_Ratio_Low 135
problems and workarounds 186 UDB_Buff_Max_Used_Pct_Crit 135
procedures 9 UDB_Buff_Max_Used_Pct_Warn 135
properties, DB2 4 UDB_Buff_Used_Pct_Crit 135
publications UDB_Buff_Used_Pct_Warn 135
accessing online ix UDB_Ce_Max_Used_Pct_Crit 135
feedback viii UDB_Ce_Max_Used_Pct_Warn 135
for support 201 UDB_Ce_Used_Pct_Crit 135
online viii UDB_Ce_Used_Pct_Warn 135
ordering viii, ix UDB_Database_Lock_Warning 136
purposes UDB_DB_BP_Hit_Ratio_Low 136
collecting data 13 UDB_DB_Cat_Cache_Hit_Ratio_Crit 136
customizing monitoring environment 11 UDB_DB_Cat_Cache_Hit_Ratio_Warn 136
investigating events 10 UDB_DB_Cur_Cons_Pct_Crit 136
monitoring with custom situations 12 UDB_DB_Cur_Cons_Pct_Warn 136
problem determination 179 UDB_DB_Dlk_Rb_Pct_For_Int_Crit 136
recovering resource operation 10 UDB_DB_Dlk_Rb_Pct_For_Int_Warn 137
viewing data 13 UDB_DB_File_Closed_High 137
viewing real-time monitoring environment 9 UDB_DB_Int_Deadlock_Rb_Pct_Crit 137
UDB_DB_Int_Deadlock_Rb_Pct_Warn 137
UDB_DB_Invalid_Pkgs_Crit 137
UDB_DB_Invalid_Pkgs_Warn 137
Q UDB_DB_Invalid_Sys_Pkgs_Crit 137
queries, using attributes 27 UDB_DB_Invalid_Sys_Pkgs_Warn 137
UDB_DB_Invalid_Triggers_Crit 137
UDB_DB_Invalid_Triggers_Warn 137
R UDB_DB_Lock_Waits_Pct_Crit 138
real-time data, viewing 9 UDB_DB_Lock_Waits_Pct_Warn 138
Rebind All Packages action 145 UDB_DB_Pool_Hit_Idx_Pct_Crit 138
Rebind Package action 146 UDB_DB_Pool_Hit_Idx_Pct_Warn 138
recovering the operation of a resource 10 UDB_DB_Pool_Hit_Ratio_Pct_Crit 138
remote deployment UDB_DB_Pool_Hit_Ratio_Pct_Warn 138
problem determination 195 UDB_DB_Pri_Log_Used_Pct_Crit 138
Reorg Table action 146 UDB_DB_Pri_Log_Used_Pct_Warn 138
requirements UDB_DB_Sec_Log_Used_Pct_Crit 138
application versions 3 UDB_DB_Sec_Log_Used_Pct_Warn 138
disk space 3 UDB_DB_Sort_Overflow_High 139
memory 3 UDB_DB_SQL_Fail_High 139
operating system 3 UDB_DB_SQL_Stmts_Fail_Pct_Crit 139
other 4 UDB_DB_SQL_Stmts_Fail_Pct_Warn 139
resource, recovering operation 10 UDB_Ma_Max_Used_Pct_Crit 139
Run Statistics action 147 UDB_Ma_Max_Used_Pct_Warn 139
UDB_Max_Agent_Overflows_High 139
UDB_Pip_Sort_Hit_Ratio_Pct_Crit 139
UDB_Pip_Sort_Hit_Ratio_Pct_Warn 139
UDB_Piped_Sorts_Rej_Pct_Crit 139
UDB_Piped_Sorts_Rej_Pct_Warn 140

Index 213
situations (continued) typeface conventions x
UDB_Post_Threshold_Sorts_High 140
UDB_Rb_Max_Used_Pct_Crit 140
UDB_Rb_Max_Used_Pct_Warn 140
UDB_Rb_Used_Pct_Crit 140
U
UDB_Agent_WaitToken_High situation 133
UDB_Rb_Used_Pct_Warn 140
UDB_Agents_Stolen_High situation 133
UDB_Status_Warning 140
UDB_Appl_BP_Hit_Ratio_Low situation 133
UDB_TS_Sp_Used_DMS_Tab_Pct_Crit 140
UDB_Appl_CatCache_Hit_Low situation 134
UDB_TS_Sp_Used_DMS_Tab_Pct_Warn 140
UDB_Appl_HJoinOflws_High situation 134
UDB_TS_Status_Warn 140
UDB_Appl_HJoinSmOflw_High situation 134
values, modifying 12
UDB_Appl_Lock_Warning 134
situations, using attributes 27
UDB_Appl_PkgCache_Hit_Low 134
Start DB2 action 149
UDB_Appl_SQL_Fail_High 134
starting monitoring agent on UNIX 5
UDB_Appl_Wait_Lock 134
Stop DB2 action 149
UDB_BP_DrtyPg_Steal_Clns 134
stopping monitoring agent on UNIX 5
UDB_BP_DrtyPg_thresh_Clns 135
support
UDB_BP_Hit_Ratio_Low situation 135
about 200
UDB_Buff_Max_Used_Pct_Crit 135
contacting 202
UDB_Buff_Max_Used_Pct_Warn 135
describing problems 203
UDB_Buff_Used_Pct_Crit 135
determining business impact of problems 203
UDB_Buff_Used_Pct_Warn 135
gathering information for 179
UDB_Ce_Max_Used_Pct_Crit 135
information centers for 201
UDB_Ce_Max_Used_Pct_Warn 135
knowledge bases for 201
UDB_Ce_Used_Pct_Crit 135
obtaining fixes 201
UDB_Ce_Used_Pct_Warn 135
on Internet 201
UDB_Database_Lock_Warning 136
submitting problems 204
UDB_DB_BP_Hit_Ratio_Low 136
weekly update option 202
UDB_DB_Cat_Cache_Hit_Ratio_Crit 136
System Overview
UDB_DB_Cat_Cache_Hit_Ratio_Warn 136
workspaces
UDB_DB_Cur_Cons_Pct_Crit 136
list 16
UDB_DB_Cur_Cons_Pct_Warn 136
System Overview attribute group 110
UDB_DB_Dlk_Rb_Pct_For_Int_Crit 136
System Overview workspace 25
UDB_DB_Dlk_Rb_Pct_For_Int_Warn 137
UDB_DB_File_Closed_High 137
UDB_DB_Int_Deadlock_Rb_Pct_Crit 137
T UDB_DB_Int_Deadlock_Rb_Pct_Warn 137
Table Space UDB_DB_Invalid_Pkgs_Crit 137
workspaces UDB_DB_Invalid_Pkgs_Warn 137
list 16 UDB_DB_Invalid_Sys_Pkgs_Crit 137
Table Space workspace 25 UDB_DB_Invalid_Sys_Pkgs_Warn 137
Tablespace attribute group 120 UDB_DB_Invalid_Triggers_Crit 137
tacmd addSystem command 5 UDB_DB_Invalid_Triggers_Warn 137
Take Action commands 10 UDB_DB_Lock_Waits_Pct_Crit 138
Backup Database 144 UDB_DB_Lock_Waits_Pct_Warn 138
list of all 143 UDB_DB_Pool_Hit_Idx_Pct_Crit 138
more information 143 UDB_DB_Pool_Hit_Idx_Pct_Warn 138
overview 143 UDB_DB_Pool_Hit_Ratio_Pct_Crit 138
predefined 143 UDB_DB_Pool_Hit_Ratio_Pct_Warn 138
problem determination 200 UDB_DB_Pri_Log_Used_Pct_Crit 138
Rebind All Packages 145 UDB_DB_Pri_Log_Used_Pct_Warn 138
Rebind Package 146 UDB_DB_Sec_Log_Used_Pct_Crit 138
Reorg Table 146 UDB_DB_Sec_Log_Used_Pct_Warn 138
Run Statistics 147 UDB_DB_Sort_Overflow_High 139
Start DB2 149 UDB_DB_SQL_Fail_High 139
Stop DB2 149 UDB_DB_SQL_Stmts_Fail_Pct_Crit 139
Update Database Configuration 150 UDB_DB_SQL_Stmts_Fail_Pct_Warn 139
Update DB Manager Configuration 150 UDB_Ma_Max_Used_Pct_Crit 139
tasks for using 9 UDB_Ma_Max_Used_Pct_Warn 139
Tivoli Enterprise Portal UDB_Max_Agent_Overflows_High 139
problem determination 195 UDB_Pip_Sort_Hit_Ratio_Pct_Crit 139
Tivoli software information center ix UDB_Pip_Sort_Hit_Ratio_Pct_Warn 139
Tivoli technical training x UDB_Piped_Sorts_Rej_Pct_Crit 139
trace logs 180 UDB_Piped_Sorts_Rej_Pct_Warn 140
directories 180 UDB_Post_Threshold_Sorts_High 140
trademarks 209 UDB_Rb_Max_Used_Pct_Crit 140
training, Tivoli technical x UDB_Rb_Max_Used_Pct_Warn 140
troubleshooting 179 UDB_Rb_Used_Pct_Crit 140

214 IBM Tivoli Monitoring for Databases: DB2 Agent: User’s Guide
UDB_Rb_Used_Pct_Warn 140 workspaces (continued)
UDB_Status_Warning 140 Application Top Ten Summary 19, 155
UDB_TS_Sp_Used_DMS_Tab_Pct_Crit 140 Buffer Pool 20, 156
UDB_TS_Sp_Used_DMS_Tab_Pct_Warn 140 Buffer Pool Detail 20
UDB_TS_Status_Warn 140 Connection 24
uninstallation Database Bottom Ten Summary by BP Hit Ratio 21
log file 181 Database I/O Activity 22
problems 187 Database Identification / Status / Logging Information 21
Update Database Configuration action 150 Database Lock Activity 22
Update DB Manager Configuration action 150 Database Overview 156
user interfaces options 2 Database Pkg/Cat Cache 22
Database Sort/Hash Join Activity 23
Database SQL Activity 23
V Database Summary by BP Hit Ratio 23
Database workspace 21
values, modifying situations 12
event 10
variables, notation for xi
General Information 24
viewing data 13
list of all 15
viewing real-time monitoring environment 9
Locking Conflict 24
views
more information 15
Application I/O Activity workspace 16
overview 15
Application Identification workspace 16
predefined 15
Application Lock Activity workspace 17
problem determination 196
Application Overview workspace 17
System Overview 25
Application Package and Catalog Cache Activity 17
Table Space 25
Application Sort and Hash Join Activity workspace 18
Application SQL Activity workspace 18
Application SQL Statement Text 18, 19
Application Time Information 19
Buffer Pool Detail workspace 20
Connection workspace 24
Database Bottom Ten Summary by BP Hit Ratio
workspace 21
Database I/O Activity workspace 22
Database Identification / Status / Logging information
workspace 21
Database Lock Activity workspace 22
Database Package/Catalog Cache workspace 22
Database Sort/Hash Join Activity workspace 23
Database SQL Activity workspace 23
Database Summary by BP Hit Ratio workspace 23
Database workspace 21
General Information workspace 24
System Overview workspace 25
Table Space workspace 25

W
weekly update support option 202
Windows agent installation problems 187
workarounds 186
agents 192
remote deployment 195
situations 196
Take Action commands 200
Tivoli Enterprise Portal 195
workspaces 196
workspaces
Application I/O Activity 16
Application Identification 16
Application Lock Activity 17
Application Overview 17
Application Package and Catalog Cache Activity 17
Application Sort and Hash Join Activity 18
Application SQL Activity 18
Application SQL Statement Text 18, 19
Application Time Information 19

Index 215
216 IBM Tivoli Monitoring for Databases: DB2 Agent: User’s Guide



Printed in USA

SC32-9449-00