IBM Tivoli OMEGAMON XE on z/VM and Linux: User's … Srvr Data attributes 133 KVLTCPIPUsrData...

Tivoli® OMEGAMON XE on z/VM and LinuxVersion 4.20

User's Guide

SC27-2836-00

��

NoteBefore using this information and the product it supports, read the information in “Notices” on page 159.

March 2010

This edition applies to Version 4, Release 2, of IBM Tivoli OMEGAMON XE on z/VM and Linux (product number5698-A36) and to all subsequent releases and modifications until otherwise indicated in new editions.

© Copyright IBM Corporation 2006, 2010.US Government Users Restricted Rights – Use, duplication or disclosure restricted by GSA ADP Schedule Contractwith IBM Corp.

Contents

Figures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . vii

Tables. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ix

About this publication . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xiIntended audience . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xiWhat this guide contains . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xi

Part 1. Getting started. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1

Chapter 1. Introducing OMEGAMON XE on z/VM and Linux . . . . . . . . . . . . . . . . 3What's new in Version 4.2.0 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5

Replacement of the z/VM Systems: default workspace with the z/VM Systems: System Healthworkspace. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5

Overall system view . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5New product-provided situations . . . . . . . . . . . . . . . . . . . . . . . . . . 7

How OMEGAMON XE on z/VM and Linux works . . . . . . . . . . . . . . . . . . . . . 7Tivoli Management Services . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9

Tivoli Enterprise Portal and Tivoli Enterprise Portal Server . . . . . . . . . . . . . . . 10Tivoli Enterprise Monitoring Server . . . . . . . . . . . . . . . . . . . . . . . . 11Tivoli Data Warehouse and the warehouse proxy . . . . . . . . . . . . . . . . . . . 11

What you can do with OMEGAMON XE on z/VM and Linux . . . . . . . . . . . . . . . . . 11Monitor processor usage at the system level . . . . . . . . . . . . . . . . . . . . . 11Monitor paging and spooling activity . . . . . . . . . . . . . . . . . . . . . . . . 12Monitor resources and workloads across LPARs . . . . . . . . . . . . . . . . . . . . 12

Chapter 2. Monitoring with OMEGAMON XE on z/VM and Linux . . . . . . . . . . . . . . 13Workspaces . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13

Proactive monitoring using the Enterprise Status workspace. . . . . . . . . . . . . . . . 15Custom queries . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17Data filters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17Dynamic workspace linking . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18

Historical data. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20Tivoli Common Reporting tool . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21Viewing event data using the Tivoli Enterprise Console . . . . . . . . . . . . . . . . . . 21Data collection interval . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21

CP Monitor sample interval . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22Linux guest ApplData interval . . . . . . . . . . . . . . . . . . . . . . . . . . . 22

Attributes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22Using attributes in queries . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23

Situations and situation events . . . . . . . . . . . . . . . . . . . . . . . . . . . 23

Part 2. Resources provided by OMEGAMON XE on z/VM and Linux . . . . . . . 25

Chapter 3. Predefined workspaces . . . . . . . . . . . . . . . . . . . . . . . . . 27Accessing OMEGAMON XE on z/VM and Linux workspaces . . . . . . . . . . . . . . . . 28

Accessing the secondary workspaces . . . . . . . . . . . . . . . . . . . . . . . . 29From the Navigator . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29From the View menu . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29From a report . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29From a chart view . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29

Attribute groups used by predefined workspaces . . . . . . . . . . . . . . . . . . . . . 29

© Copyright IBM Corp. 2006, 2010 iii

z/VM Systems: System Health. . . . . . . . . . . . . . . . . . . . . . . . . . . . 30z/VM Linux Systems: default workspace . . . . . . . . . . . . . . . . . . . . . . . . 32CP Owned Devices workspace . . . . . . . . . . . . . . . . . . . . . . . . . . . 33Channel workspace. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35

FICON Channels workspace . . . . . . . . . . . . . . . . . . . . . . . . . . . 36DASD workspace . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37

Minidisk Cache workspace . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38CCW Translations workspace . . . . . . . . . . . . . . . . . . . . . . . . . . . 40Control Unit Cache workspace . . . . . . . . . . . . . . . . . . . . . . . . . . 40DASD Cache workspace . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42VDISK workspace . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43

LPAR workspace. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44Processor by LPAR Name workspace . . . . . . . . . . . . . . . . . . . . . . . . 46Processor workspace . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47

Network workspace. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48VSWITCH workspace . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50

Real Storage workspace . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51System workspace . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52

Spin Locks workspace. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53System Terminal workspace . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54

Take Action view . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55TCPIP workspace . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55TCPIP User workspace . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56Workload workspace . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57

Linux Workload workspace . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58ApplData workspace . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59

Enabling the collection of KVLUser ApplData . . . . . . . . . . . . . . . . . . . . 60Resource Constraint workspace . . . . . . . . . . . . . . . . . . . . . . . . . . 61

Chapter 4. Predefined situations . . . . . . . . . . . . . . . . . . . . . . . . . . 63Using the Situation Editor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63Investigating a situation event . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64Situation formulas . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65

Avoid using negative values . . . . . . . . . . . . . . . . . . . . . . . . . . . 65Predefined situations provided by OMEGAMON XE on z/VM and Linux . . . . . . . . . . . . 65

Summary of predefined situations . . . . . . . . . . . . . . . . . . . . . . . . . 66Predefined situations and formulas . . . . . . . . . . . . . . . . . . . . . . . . . 68

ZVM_Avail_Mean2G_Low . . . . . . . . . . . . . . . . . . . . . . . . . . . 68ZVM_Avail_Mean_Low . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68ZVM_CP_CPU_Critical . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69ZVM_CP_CPU_High . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69ZVM_CUCache_DataResp_High . . . . . . . . . . . . . . . . . . . . . . . . . 69ZVM_DASD_Cache_Hits_Low. . . . . . . . . . . . . . . . . . . . . . . . . . 70ZVM_DASD_Queue_Critical . . . . . . . . . . . . . . . . . . . . . . . . . . 71ZVM_DASD_Queue_High . . . . . . . . . . . . . . . . . . . . . . . . . . . 71ZVM_LPAR_Busy_Critical . . . . . . . . . . . . . . . . . . . . . . . . . . . 71ZVM_LPAR_Busy_High . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72ZVM_LPAR_Ovhd_Critical . . . . . . . . . . . . . . . . . . . . . . . . . . . 72ZVM_LPAR_Ovhd_High . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73ZVM_Page_Queue_High. . . . . . . . . . . . . . . . . . . . . . . . . . . . 73ZVM_Page_Used_Critical . . . . . . . . . . . . . . . . . . . . . . . . . . . 74ZVM_Page_Used_High . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74ZVM_PerfKit_Collector_Inactive . . . . . . . . . . . . . . . . . . . . . . . . . 75ZVM_Physical_CPU_Critical . . . . . . . . . . . . . . . . . . . . . . . . . . 76ZVM_Physical_CPU_High . . . . . . . . . . . . . . . . . . . . . . . . . . . 77ZVM_Spin_Exclusive_Pct_Critical . . . . . . . . . . . . . . . . . . . . . . . . 77

iv IBM Tivoli OMEGAMON XE on z/VM and Linux: User's Guide

ZVM_Spin_Exclusive_Pct_High . . . . . . . . . . . . . . . . . . . . . . . . . 77ZVM_Spool_Used_Critical . . . . . . . . . . . . . . . . . . . . . . . . . . . 78ZVM_Spool_Used_High . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78ZVM_Total_CPU_Critical . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79ZVM_Total_CPU_High. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79ZVM_Total_to_Virtual_High . . . . . . . . . . . . . . . . . . . . . . . . . . . 80ZVM_User_CPU_Critical . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80ZVM_User_CPU_High. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80ZVM_User_Scaled_CPU_Critical . . . . . . . . . . . . . . . . . . . . . . . . . 81ZVM_User_Scaled_CPU_High . . . . . . . . . . . . . . . . . . . . . . . . . 81ZVM_VDISK_Page_High_IORate_Low . . . . . . . . . . . . . . . . . . . . . . 82ZVM_Virtual_CPU_Critical . . . . . . . . . . . . . . . . . . . . . . . . . . . 82ZVM_Virtual_CPU_High . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83ZVM_Virtual_Scaled_CPU_Critical . . . . . . . . . . . . . . . . . . . . . . . . 83ZVM_Virtual_Scaled_CPU_High . . . . . . . . . . . . . . . . . . . . . . . . . 83ZVM_User_Wait_Page_Critical . . . . . . . . . . . . . . . . . . . . . . . . . 84ZVM_User_Wait_Page_High . . . . . . . . . . . . . . . . . . . . . . . . . . 84ZVM_User_Wait_CPU_Critical. . . . . . . . . . . . . . . . . . . . . . . . . . 85ZVM_User_Wait_CPU_High . . . . . . . . . . . . . . . . . . . . . . . . . . 85ZVM_Storage_Overcommit_Critical . . . . . . . . . . . . . . . . . . . . . . . . 86ZVM_Storage_Overcommit_High . . . . . . . . . . . . . . . . . . . . . . . . . 87ZVM_Eligible_List_High . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87

Chapter 5. Take Action commands . . . . . . . . . . . . . . . . . . . . . . . . . 89Issuing Take Action commands . . . . . . . . . . . . . . . . . . . . . . . . . . . 89

Part 3. Monitoring scenarios . . . . . . . . . . . . . . . . . . . . . . . . . 91

Chapter 6. Monitoring scenarios . . . . . . . . . . . . . . . . . . . . . . . . . . 93Solving a DASD problem. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93Solving a paging data set problem . . . . . . . . . . . . . . . . . . . . . . . . . . 94Monitoring Control Program (CP) CPU utilization . . . . . . . . . . . . . . . . . . . . . 95Monitoring LPAR CPU utilization . . . . . . . . . . . . . . . . . . . . . . . . . . . 96Monitoring system utilization by Linux on System z guests . . . . . . . . . . . . . . . . . 96Monitoring control unit cache activity . . . . . . . . . . . . . . . . . . . . . . . . . 97Monitoring virtual disk (VDISK) I/O activity . . . . . . . . . . . . . . . . . . . . . . . 98

Attributes reference . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101Attribute names. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101Attributes by attribute group . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101

KVLCP Device attributes . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101KVLChannel Data attributes . . . . . . . . . . . . . . . . . . . . . . . . . . . 102KVLControlUnit attributes . . . . . . . . . . . . . . . . . . . . . . . . . . . . 104KVLDASDCache attributes . . . . . . . . . . . . . . . . . . . . . . . . . . . 106KVLDevice attributes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 110KVLFChannel Data attributes . . . . . . . . . . . . . . . . . . . . . . . . . . . 111KVLHiperSocket attributes . . . . . . . . . . . . . . . . . . . . . . . . . . . . 112KVLLChannel Data attributes . . . . . . . . . . . . . . . . . . . . . . . . . . . 113KVLLPAR Info attributes . . . . . . . . . . . . . . . . . . . . . . . . . . . . 115KVLMinidisk Cache attributes. . . . . . . . . . . . . . . . . . . . . . . . . . . 118KVLProcessor Data attributes . . . . . . . . . . . . . . . . . . . . . . . . . . 121KVLPTKStat attributes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 123KVLSpinLock attributes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 124KVLSystem attributes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 125KVLSystem2 attributes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 129KVLUser Wait attributes . . . . . . . . . . . . . . . . . . . . . . . . . . . . 131

Contents v

KVLTCPIP Srvr Data attributes . . . . . . . . . . . . . . . . . . . . . . . . . . 133KVLTCPIPUsrData attributes . . . . . . . . . . . . . . . . . . . . . . . . . . . 135KVLUser ApplData attributes . . . . . . . . . . . . . . . . . . . . . . . . . . . 136KVLUser Workload attributes . . . . . . . . . . . . . . . . . . . . . . . . . . . 140KVLVdisk attributes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 143KVLVirtualSwitch attributes . . . . . . . . . . . . . . . . . . . . . . . . . . . 145

Support for problem solving . . . . . . . . . . . . . . . . . . . . . . . . . . . 149Using IBM Support Assistant . . . . . . . . . . . . . . . . . . . . . . . . . . . . 149Obtaining fixes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 149Receiving weekly support updates . . . . . . . . . . . . . . . . . . . . . . . . . . 150Contacting IBM Software Support . . . . . . . . . . . . . . . . . . . . . . . . . . 150

Determining the business impact . . . . . . . . . . . . . . . . . . . . . . . . . 151Describing problems and gathering information . . . . . . . . . . . . . . . . . . . . 151Submitting problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 152

Documentation library. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 153OMEGAMON XE on z/VM and Linux library . . . . . . . . . . . . . . . . . . . . . . 153

OMEGAMON XE on z/VM and Linux online help . . . . . . . . . . . . . . . . . . . 154IBM Tivoli Monitoring publications . . . . . . . . . . . . . . . . . . . . . . . . . . 154

Tivoli Enterprise Portal help system . . . . . . . . . . . . . . . . . . . . . . . . 155Related publications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 155Accessing terminology online. . . . . . . . . . . . . . . . . . . . . . . . . . . . 156Accessing publications online . . . . . . . . . . . . . . . . . . . . . . . . . . . 156Ordering publications. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 156Accessibility . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 157

Notices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 159Trademarks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 161

Glossary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 163

Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 167

vi IBM Tivoli OMEGAMON XE on z/VM and Linux: User's Guide

Figures

1. Architecture used by this monitoring agent . . . . . . . . . . . . . . . . . . . . . . 92. IBM Tivoli Monitoring components . . . . . . . . . . . . . . . . . . . . . . . . 103. Tivoli Enterprise Portal sample workspace for OMEGAMON XE on z/VM and Linux . . . . . . 144. Enterprise Status workspace . . . . . . . . . . . . . . . . . . . . . . . . . . 165. Tivoli Enterprise Portal Navigator with event indicators . . . . . . . . . . . . . . . . . 236. Link icon in a sample table view . . . . . . . . . . . . . . . . . . . . . . . . . 297. z/VM Systems System Health . . . . . . . . . . . . . . . . . . . . . . . . . . 328. z/VM Linux Systems default workspace . . . . . . . . . . . . . . . . . . . . . . 339. CP Owned Devices workspace . . . . . . . . . . . . . . . . . . . . . . . . . 34

10. Channel workspace . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3611. FICON Channels workspace . . . . . . . . . . . . . . . . . . . . . . . . . . 3712. DASD workspace . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3813. Minidisk Cache workspace . . . . . . . . . . . . . . . . . . . . . . . . . . . 3914. CCW Translations workspace . . . . . . . . . . . . . . . . . . . . . . . . . . 4015. Control Unit workspace . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4216. DASD Cache workspace . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4317. VDISK workspace . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4418. LPAR workspace . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4619. Processor by LPAR Name workspace . . . . . . . . . . . . . . . . . . . . . . . 4720. Processor workspace . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4821. Network workspace . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5022. VSWITCH workspace . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5123. Real Storage workspace . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5224. System workspace . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5325. Spin Locks workspace . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5426. System Terminal workspace. . . . . . . . . . . . . . . . . . . . . . . . . . . 5527. TCPIP workspace . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5628. TCPIP User workspace . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5729. Workload workspace . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5830. Linux Workload workspace . . . . . . . . . . . . . . . . . . . . . . . . . . . 5931. ApplData workspace . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6032. Resource Constraint workspace . . . . . . . . . . . . . . . . . . . . . . . . . 6133. Navigator with critical and warning event indicators . . . . . . . . . . . . . . . . . . 9334. ZVM_DASD_Queue_Critical situation formula . . . . . . . . . . . . . . . . . . . . 9335. Top 5 Device Busy bar chart . . . . . . . . . . . . . . . . . . . . . . . . . . 9336. Average Queued IO attribute . . . . . . . . . . . . . . . . . . . . . . . . . . 9437. Navigator with critical and warning event indicators . . . . . . . . . . . . . . . . . . 9438. Top 5 Page Extent Utilization bar chart . . . . . . . . . . . . . . . . . . . . . . . 9439. Paging and Spooling bar chart . . . . . . . . . . . . . . . . . . . . . . . . . . 9540. Navigator with critical and warning event indicators . . . . . . . . . . . . . . . . . . 9541. List of situation events - flyover pop-up. . . . . . . . . . . . . . . . . . . . . . . 9542. Top 5 CPU Users bar chart . . . . . . . . . . . . . . . . . . . . . . . . . . . 9543. LPAR Load bar chart . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9644. Top 5 CPU Linux Guest Systems bar chart . . . . . . . . . . . . . . . . . . . . . 9745. Navigator with warning event indicators . . . . . . . . . . . . . . . . . . . . . . 9746. Situation event flyover pop-up . . . . . . . . . . . . . . . . . . . . . . . . . . 9847. Top 5 Cache Control Unit Load Data Time bar chart . . . . . . . . . . . . . . . . . . 9848. Navigator with warning event indicators . . . . . . . . . . . . . . . . . . . . . . 9949. Top 5 Paging Rates per Second bar chart . . . . . . . . . . . . . . . . . . . . . 99

© Copyright IBM Corp. 2006, 2010 vii

viii IBM Tivoli OMEGAMON XE on z/VM and Linux: User's Guide

Tables

1. New predefined situations . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72. Links from OMEGAMON XE on z/VM and Linux to Tivoli Monitoring Agent for Linux OS . . . . 193. Attributes used by predefined workspaces . . . . . . . . . . . . . . . . . . . . . 294. Summary of predefined situations. . . . . . . . . . . . . . . . . . . . . . . . . 66

© Copyright IBM Corp. 2006, 2010 ix

x IBM Tivoli OMEGAMON XE on z/VM and Linux: User's Guide

About this publication

IBM® Tivoli® OMEGAMON® XE on z/VM® and Linux® version 4 release 2 (V4.2.0) is one of a suite of Tivolimonitoring products. The Tivoli monitoring products monitor your mainframe and distributed systems on avariety of platforms, and provide workstation-based reports you can use to track trends and troubleshootsystem problems.

This publication is intended to introduce you to the features of this product and to help you use thepredefined workspaces and monitoring tools it provides. In addition, it contains several typical monitoringscenarios that illustrate how the product can be used to monitor and manage your z/VM and Linuxsystems.

This publication is not intended as a guide to the features and functions of the Tivoli Enterprise Portal andthe services it provides. Where appropriate, this guide refers you to the appropriate documents.

Intended audienceThe primary audience for this publication is the z/VM system programmer or system administrator who isresponsible for monitoring the health of z/VM system resources. This publication is also addressed to theLinux on System z® programmer or system administrator who is responsible for monitoring the health ofsystems running in Linux guests.

The system programmer or system administrator responsibilities include:

v Planning for and overseeing product installation

v Monitoring and troubleshooting system and performance problems

v Analyzing performance data for problem determination

v Providing historical performance data for trend analysis

v Resolving system problems or forwarding them to someone who can resolve them

Users of this publication should be familiar with the following topics:

v Performance monitoring concepts

v IBM Tivoli Monitoring and the Tivoli Enterprise Portal interface

v The Linux on zSeries® operating system and its associated concepts

v The z/VM operating system and its associated concepts

v The Performance Toolkit for VM

v The Microsoft® Windows® operating system

What this guide containsThis guide contains the following sections:

Part 1 - Getting started

Part 1 contains the following chapters:

v Chapter 1, “Introducing OMEGAMON XE on z/VM and Linux,” on page 3 contains an overview of themonitoring agent.

v Chapter 2, “Monitoring with OMEGAMON XE on z/VM and Linux,” on page 13 provides overviews of thepredefined workspaces, situations, and attributes. This chapter also provides information on historicaldata collection, and on dynamic workspace linking between monitoring agents.

Part 2 - Resources provided by OMEGAMON XE on z/VM and Linux

© Copyright IBM Corp. 2006, 2010 xi

Part 2 contains the following chapters:

v Chapter 3, “Predefined workspaces,” on page 27 provided by this monitoring agent, and how you canuse them to help you meet your specific requirements.

v Chapter 4, “Predefined situations,” on page 63 describes the predefined situations available with thismonitoring agent.

v Chapter 5, “Take Action commands,” on page 89 describes the use of Take Action commands with thismonitoring agent. This chapter also provides instructions on how you can configure your environment toenable Take Action commands.

Part 3 - Monitoring scenarios

This chapter describes some common usage scenarios that might be similar to your experience in usingOMEGAMON XE on z/VM and Linux to monitor and to manage your environment.

Appendixes

The Appendixes section contains the following appendixes:

v “Attributes reference” on page 101 defines each of the attributes within the attribute groups and howattributes can be used.

v “Support for problem solving” on page 149 describes the options available for obtaining support for IBMsoftware products.

v “Documentation library” on page 153 contains information about the publications for Tivoli OMEGAMONXE on z/VM and Linux and for IBM Tivoli Monitoring and the commonly shared components of TivoliManagement Services. This appendix also includes a section on publications and Web sites that arerelevant to this monitoring agent.

v “Notices” on page 159 contains copyright and trademarks information.

xii IBM Tivoli OMEGAMON XE on z/VM and Linux: User's Guide

Part 1. Getting started

The following chapters introduce the Tivoli OMEGAMON XE on z/VM and Linux monitoring agent.

v Chapter 1, “Introducing OMEGAMON XE on z/VM and Linux,” on page 3 explains how OMEGAMON XEon z/VM and Linux works, describes the resources it provides, provides examples of how it can be usedto monitor, analyze, and manage operating systems, workloads, and system resources. This chapteralso describes the new features in this release.

v Chapter 2, “Monitoring with OMEGAMON XE on z/VM and Linux,” on page 13 provides informationabout using the predefined workspaces, configuring historical data collection for the product-providedmonitored attributes, and cross-product linking.

© Copyright IBM Corp. 2006, 2010 1

2 IBM Tivoli OMEGAMON XE on z/VM and Linux: User's Guide

Chapter 1. Introducing OMEGAMON XE on z/VM and Linux

IBM Tivoli OMEGAMON XE on z/VM and Linux, Version 4.2.0 gives you the ability to view data collectedfrom multiple systems on one easy-to-use, flexible interface called the Tivoli Enterprise Portal. WithOMEGAMON XE on z/VM and Linux, you can monitor z/VM data obtained from the Performance Toolkitfor VM (also called the Performance Toolkit). You can also display Linux on System z performance data.

Use this monitoring agent to take advantage of the flexibility of Linux with the reliability and performance ofmainframe servers. With z/VM virtualization technology, you can consolidate a significant number of Linuxservers, thus helping to minimize costs and to maximize the availability of your mission-criticalapplications. Another significant benefit of virtualization is the reduction of power usage, thus contributingto a green environment and lessening your energy costs.

Predefined workspaces enable you to start monitoring your enterprise as soon as the OMEGAMON XE onz/VM and Linux software is installed and configured. The user interface supports several formats forviewing data, such as graphs, bar charts, and tables.

The displayed information enables you to:

v Collect and analyze reliable, up-to-the-minute data that allow you to make faster, better informedoperating decisions.

v Identify, isolate and correct problems across z/VM and Linux instances quickly and easily.

v Manage your applications from a single point to identify problems at any time.

v Track performance across multiple platforms and systems.

v View and monitor workloads for virtual machines and workload groups, as well as analyze LPAR usage.

v View reports on z/VM usage of resources, such as CPU utilization, storage, I/O statistics, minidiskcache, and wait statistics on processor spin locks.

v View reports on Linux on System z usage of resources, such as paging data, CPU usage, and systemusage by Linux guests.

With OMEGAMON XE on z/VM and Linux, you can set threshold levels to alert you when systemconditions reach the defined thresholds. You can use advanced monitoring facilities that include:

v User-defined and predefined situations based on thresholds to raise different types of alerts. You cancustomize the predefined situations to meet the needs of your enterprise. See “Predefined situationsand formulas” on page 68.

v Navigation to greater levels of detailed performance data. For Linux guests, this monitoring agentprovides links to Tivoli Monitoring Agent for Linux OS workspaces directly from the Tivoli OMEGAMONXE on z/VM and Linux workspaces. See “Dynamic workspace linking” on page 18.

v Expert advice on how to identify and to resolve performance problems.

With this advanced monitoring approach on the Tivoli Enterprise Portal interface, you can manage multiplez/VM systems, as well as systems on other platforms, such as Windows, UNIX®, and z/OS®.

You can also connect to the z/VM host system by means of TCP/IP to access the Performance Toolkit.Once in the z/VM Performance Toolkit, you can navigate to specific screens to view more detailed dataand to investigate a problem. You use the “System Terminal workspace” on page 54 to connect to thez/VM host system.

Important: All Tivoli products that utilize the IBM Tivoli Monitoring V6 components are impacted byDaylight Savings Time (DST). Several countries have announced changes to their Daylight Savings Time(DST) policies and/or changes to their Time Zones. In order to assist customers with understanding theimpact of these changes to their Tivoli products, a Knowledge Collection with associated links has beencreated.


For more information, refer to Knowledge Collection: Non-US Daylight Savings Time (DST) and Time Zonechanges impact on Tivoli Products, found within the Technotes link at the following URL address:

http://www-306.ibm.com/software/sysmgmt/products/support/IBMTivoliMonitoringV6.html


http://www-306.ibm.com/software/sysmgmt/products/support/IBMTivoliMonitoringV6.html

What's new in Version 4.2.0The following changes have been made to this Tivoli OMEGAMON XE on z/VM and Linux V4.2.0 release:

v Replacement of the z/VM Systems default workspace with the z/VM Systems System Healthworkspace.

v Overall system view.

The sections that follow provide more detail about these changes.

Replacement of the z/VM Systems: default workspace with the z/VMSystems: System Health workspaceThe z/VM Systems branch is the top level of the Navigation tree for the z/VM operating system. Whenyou click z/VM Systems, the System Health workspace for this level of the tree is displayed. Thisworkspace provides data for all of the z/VM systems that are registered with the Tivoli EnterpriseMonitoring Server. Use this workspace to view the overall health of all your z/VM systems. See “z/VMSystems: System Health” on page 30 for details.

Overall system viewThe following attributes have been added to the “KVLSystem attributes” on page 125. This attribute groupprovides system-level data on CPU utilization by the z/VM Control Program and by the z/VM virtualmachines, on free-storage management, on paging rates, and on user activity. The following attributeswere added:

v Locked Dynamic Paging Area Storage in Kbytes

v Main Storage Available in Mbytes

v Main Storage Offline in Kbytes

v Pageable Dynamic Paging Area Storage in Mbytes

v Pageable Dynamic Paging Area Storage in Use Percent

v Reserved Real Storage in Mbytes

v Shared Segment Storage in Mbytes

v Standby Real Storage Size in Mbytes

v Subpool Storage in Kbytes

v Subpool Storage in Use Percent

v SYSGEN Storage Size in Mbytes

v Tasks Waiting for a Frame

v Tasks Waiting for a Page Rate per Second

v Trace Table Storage in Kbytes

v Virtual CPU Busy Percent

v XSTORE Size in Mbytes

v XSTORE Dedicated to Virtual Machines in Mbytes

v XSTORE Allocated to CP in Mbytes

v XSTORE In Use by CP Percent

v XSTORE Minimum Before Migrating in Kbytes

v XSTORE Page Allocation Rate per Second

v XSTORE Page Age in Seconds

v XSTORE Page Age at Migration in Seconds

Chapter 1. Introducing OMEGAMON XE on z/VM and Linux 5

The following attributes have been added to the new “KVLSystem2 attributes” on page 129. This attributegroup provides system-level data on CPU utilization by the z/VM Control Program and by the z/VM virtualmachines, on free-storage management, on paging rates, and on user activity. The following attributeswere added:

v Highest Normalized Individual Wait State Workload

v Highest Normalized Individual Wait State Percent

v Highest Individual Wait State

v Highest Individual Wait State Percent

v Highest Individual Wait State Workload

v Highest Normalized Individual Wait State

v LPAR Name

v Page Read Blocking Factor in pages

v Page Write Rate per Second

v Page Write Blocking Factor in pages

v Pages Moved Below 2G Line Rate per Second

v Pages Paged in via Fast Path Logic Rate per Second

v Pages Paged in via Long Path Logic Rate per Second

v Pages Paged out via Long Path Logic Rate per Second

v Paging I/O Rate from Storage to DASD Rate per second

v Spool Read Rate per second

v Spool Write Rate per second

v Storage Overcommit Minus VDISK Ratio

v Storage Overcommit Ratio

v System and Workload with Highest Normalized Wait State

v System and Workload with Highest Wait State

v System ID

v System Name

v Time

v Total Main Storage in Mbytes

v VDISK Main Storage Frames

v VDISK Maximum Blocks per User

v VDISK Maximum Storage Blocks

v VDISK Pages Migrated to DASD

v VDISK XSTORE Blocks

v XSTORE to DASD Blocking Factor in pages

The following attributes have been added to the“KVLLPAR Info attributes” on page 115. This attributegroup provides information about the overall utilization of system processor resources by the logicalpartitions defined for your system.

v Scaled Logical Load Percent

v Scaled LPAR Overhead Time Percent

v Scaled LPAR Suspend Time Percent

v Scaled VM Load Percent

The following attributes have been added to the“KVLMinidisk Cache attributes” on page 118. This attributegroup provides information about minidisk cache data, one line per interval. The information shown isbased on CP Monitor SYSTEM domain SAMPLE data.


v MDCACHE Read Hit Rate per Second

v MDCACHE Writes per Second

The following attributes have been added to the“KVLUser Workload attributes” on page 140. This attributegroup provides information about workload activity and storage utilization.

v Used DASD Page Slot Count

v Data Spaces Owned

v Locked Pages Above 2GB

v Locked Pages Below 2GB

v Main Storage Pages Moved to XSTOR per Second

v Page Steals per Second

v Pages Moved to Below 2GB per Second

v Reserved Pages

v Share Setting

v Share Setting Type

v XSTOR Pages Moved to Disk per Second

v XSTOR Pages Moved to Main Storage per Second

New product-provided situationsThe following table contains the new situations provided by this version of the monitoring agent:

Table 1. New predefined situations

Navigatoritem Name of situation

Column name and initial conditionalvalue State

Runs atstartup(Yes orNo)

z/VM LinuxSystems

ZVM_Storage_Overcommit_High KVLSystem2.Real_Storage_Overcommit >3.0 andKVLSystem2.Real_Storage_Overcommit <=4.0

High(Warning)

No

z/VM LinuxSystems

ZVM_Storage_Overcommit_Critical KVLSystem2.Real_Storage_Overcommit >4.0

Critical No

System ZVM_Eligible_List_High KVLSystem.Eligible_Users > 5 High(Warning)

No

Note: If you customize any product-provided situation, you may lose any changes you make when thesesituations are modified by future application of maintenance. To retain the changes you make to theexisting situations, make a copy of the situation using a unique name, and modify the copy instead of theoriginal. That way your situations will not be overwritten by maintenance.

See “Predefined situations and formulas” on page 68 for descriptions of the product-provided situations.

How OMEGAMON XE on z/VM and Linux worksThe OMEGAMON XE on z/VM and Linux product uses the Tivoli Management Services infrastructure(sometimes referred to as Tivoli Management Services). Tivoli Management Services components providesecurity, data transfer and storage, notification mechanisms, user interface presentation, andcommunication services for a number of products, including IBM Tivoli Monitoring and OMEGAMON XEmonitoring agents, in an agent-server-client architecture.


The OMEGAMON XE on z/VM and Linux monitoring agent runs as a process on a Linux on zSeries guest.The monitoring agent obtains z/VM data and Linux on System z data from the Performance Toolkit runningon z/VM. The acquired data displays in the Tivoli Enterprise Portal. This capability allows for quick problemresolution and enables support personnel to more easily work across platform boundaries. See the IBMTivoli OMEGAMON XE on z/VM and Linux Planning and Configuration Guide for information onconfiguring this monitoring agent. That guide also contains information on the supported versions of thesoftware used by this monitoring agent.

Figure 1 on page 9 depicts a sample OMEGAMON XE on z/VM and Linux environment. In this sampleenvironment, one z/VM image is running with three primary guests. The primary guests include two Linuxon zSeries guests and one guest running z/OS. The monitoring server, the Tivoli Enterprise Portal, and thedata warehouse are running on a separate host.

In this sample environment, both Linux systems are running the Tivoli Monitoring Agent for Linux OS. Oneof the Linux systems is running the OMEGAMON XE on z/VM and Linux monitoring agent. All of themonitoring agents are transferring requests and data using the shared technology components.

The environment includes a Conversational Monitor System (CMS) guest that is running the PerformanceToolkit. Data are collected from the z/VM operating system and the data are written to a discontiguoussaved segment (DCSS). The data are collected by the z/VM and Linux monitoring agent by reading thedata in the DCSS using the z/VM DCSS device driver support in Linux for zSeries.

The DCSS must be predefined for use by the Performance Toolkit and by the OMEGAMON XE on z/VMand Linux monitoring agent. The Performance Toolkit has read/write access to the DCSS. The monitoringagent has read-only access. A default DCSS called PERFOUT is provided with the Performance Toolkit forz/VM, v5.3 and above. You can customize the default PERFOUT DCSS to meet the requirements of yourenvironment.

Note: The default name PERFOUT is used in this document to refer to the formatted output DCSScreated by the FC MONCOLL SEGOUT command. If you choose to create your own with a differentname, substitute your name for PERFOUT.

Lastly, another CMS guest is used as a VM command server. The VM command server receivescommand requests from the Tivoli Enterprise Portal by means of the monitoring agent to execute VMand/or CMS commands to the z/VM operating system. You can use the optional Take Action featureavailable with this monitoring agent to emit commands to the z/VM operating system. See Chapter 5,“Take Action commands,” on page 89 for details.


Tivoli Management ServicesThe IBM Tivoli Monitoring client-server-agent implementation of Tivoli Management Services includes thefollowing components:

v Tivoli Enterprise Portal

v Tivoli Enterprise Portal Server

v Tivoli Enterprise Monitoring Server

v Tivoli Enterprise Monitoring Agents

v Tivoli Data Warehouse and the warehouse proxy

The major components of IBM Tivoli Monitoring are shown in Figure 2 on page 10.

Figure 1. Architecture used by this monitoring agent


For the latest information about the supported operating systems for the different IBM Tivoli Monitoringcomponents, see the following Web address:

http://www-306.ibm.com/software/sysmgmt/products/support/Tivoli_Supported_Platforms.html

Tivoli Enterprise Portal and Tivoli Enterprise Portal ServerThe Tivoli Enterprise Portal is the user interface for products that use Tivoli Management Services. TheTivoli Enterprise Portal is a thin Java™ application.

The Tivoli Enterprise Portal client has two modes of operation:

desktopThe application software is installed on your system.

browserAccess the Tivoli Enterprise Portal from a browser, using the Web address of the Tivoli EnterprisePortal Server. In browser mode, the software is downloaded to your system the first time you logon to Tivoli Enterprise Portal, and thereafter only when there are software updates.

The Tivoli Enterprise Portal has its own server, the Tivoli Enterprise Portal Server, that communicates witha central Tivoli Enterprise Monitoring Server to send requests to and retrieve data from monitoring agentson managed systems. The Tivoli Enterprise Portal Server is responsible for building and formatting theworkspaces viewed in the Tivoli Enterprise Portal clients.

You can find detailed instructions for using Tivoli Enterprise Portal in the Tivoli Enterprise Portal online helpand in the IBM Tivoli Monitoring publications (see “IBM Tivoli Monitoring publications” on page 154).

Figure 2. IBM Tivoli Monitoring components


http://www-306.ibm.com/software/sysmgmt/products/support/Tivoli_Supported_Platforms.html

Tivoli Enterprise Monitoring ServerThe Tivoli Enterprise Monitoring Server performs the following functions:

v Consolidates the data collected by monitoring agents and distributes the data to the Tivoli EnterprisePortal interface.

v Sends alerts to the Tivoli Enterprise Portal Server when specified conditions are met.

v Stores historical data and prototypes for configuration.

The monitoring server can run as a standalone server, or as a remote server in a hierarchy of servers thatreport to a master server called the hub monitoring server.

A hub Tivoli Enterprise Monitoring Server acts as the focal point for data collection and distribution. Itcommunicates with monitoring agents on the same or other systems, other Tivoli Enterprise MonitoringServers, and the Tivoli Enterprise Portal interface.

Remote monitoring servers communicate only with the agents that connect to them and with the hubmonitoring server.

Tivoli Data Warehouse and the warehouse proxyThe Tivoli Data Warehouse is an optional component of Tivoli Management Services. It is a long-term datastore for the performance and analysis data collected by the monitoring agents. The warehouse proxy is aprocess that periodically moves data from the hub Tivoli Enterprise Monitoring Server or each monitoringagent to the Tivoli Data Warehouse. The short-term history data for z/OS-based agents is maintained inthe persistent data store at either the Tivoli Enterprise Monitoring Server or the monitoring agent. Thisproxy agent retrieves the short-term history data and stores it into the warehouse.

For more information about the Tivoli Data Warehouse and the warehouse proxy, see the IBM TivoliMonitoring Installation and Setup Guide.

What you can do with OMEGAMON XE on z/VM and LinuxWhen a performance bottleneck is suspected or when job turnaround times become too long, the reasonhas to be determined so actions can be defined for improving performance. After you have decided (orbeen informed) that your system has a performance problem, there are three main areas that will need tobe checked:

v CPU

v Storage and paging

v General I/O

The data and resources provided by OMEGAMON XE on z/VM and Linux enable you to monitor and tomanage workload performance and resource usage in a variety of ways, especially as they pertain toperformance problems. The sections that follow describe some of the ways you can use this monitoringagent to manage and to improve the performance of your systems.

Monitor processor usage at the system levelThe “Workload workspace” on page 57 contains the All z/VM Workloads table that provides informationabout the total number of CPU seconds used by each workload, the percentage of total CPU used by thez/VM Control Program (CP) to manage the workload, and the total time each workload or eachaggregation of the group was logged on. This workspace also contains a bar chart that identifies the fivebiggest users of CPU. Additional bar charts on this workspace show page read/write data, page rate data,and working set size data.

Use the Workload workspace to monitor high CPU utilization by the Control Program as well to find outwhich z/VM workloads are consuming the most resources.


When you determine that a problem is not originating from z/VM, you can access the “Linux Workloadworkspace” on page 58 to examine the performance of the Linux guests. The Linux Workload workspaceprovides data about system usage by user ID for all Linux guest systems defined in your environment. Thepredefined workspace contains the Top 5 CPU Linux Guest Systems bar chart that displays the top fiveLinux guest users of CPU. For each Linux guest, it shows the percentage of CPU used by the system tomanage the workload, and the percentage of virtual CPU utilized.

The “ApplData workspace” on page 59, also accessed from the Workload workspace provides extensivemetrics about network utilization, CPU usage, and operating system data for each Linux guest system.

Monitor paging and spooling activityThe “CP Owned Devices workspace” on page 33 contains several views that display data on the activity ofthe paging and spooling devices on the overall z/VM system. You can use this workspace to determine theresource utilization by the paging and spooling devices to identify the busiest devices.

If you find that your paging disks have poor response times, you can check whether some of the I/O loadhas been caused by minidisks with a high activity on the same disk pack. Use the “Minidisk Cacheworkspace” on page 38, accessed from the “DASD workspace” on page 37, to monitor minidisk cacheusage.

You can also review the data on the “Real Storage workspace” on page 51. This workspace contains theSystem Page Rate linear gauge chart that shows the I/O paging rate for the whole system. Thisworkspace also contains the Page Wait Queue chart that displays the percentage of all virtual machines ina page wait state. A high percentage of users in a page wait state can indicate a paging bottleneck.

Monitor resources and workloads across LPARsOMEGAMON XE on z/VM and Linux provides information about the overall utilization of resources by thelogical partitions (LPARs) defined for your system. You can use the LPAR workspace to determine when tomove resources from logical partitions with less important work to logical partitions with more importantwork when that work is missing its goals.

Use the LPAR Utilization table in the “LPAR workspace” on page 44 to determine how well each logicalpartition is able to obtain CPU resources. This table provides data on overall system utilization by each ofthe defined logical partitions. The table shows data such as the total amount of time that all of the logicalprocessors for a logical partition were busy during the collection interval, the percentage of elapsed timethat a logical processor spent to manage LPARs, and the utilization of the system based on the number oflogical processors available.

The “Processor by LPAR Name workspace” on page 46, accessed from the LPAR Utilization table of the“LPAR workspace” on page 44, displays data on the logical partitions for which you elected to viewadditional information. Use this workspace to determine the resource usage by specific logical partitionsand the load for each of the defined processors.

Additionally, the “Processor workspace” on page 47 provides data on overall system utilization by each ofthe defined logical partitions. The table shows data such as the total amount of time that all of the logicalprocessors for a logical partition were busy during the collection interval, the percentage of elapsed timethat a logical processor spent to manage LPARs, and the utilization of the system based on the number oflogical processors available. This workspace also provides data on the types of processors being used.Use this workspace to determine how well each logical partition is able to obtain CPU resources.

For additional information on using the predefined workspaces and situations to analyze and to resolvesystem problems, see Chapter 6, “Monitoring scenarios,” on page 93.


Chapter 2. Monitoring with OMEGAMON XE on z/VM andLinux

This chapter is intended to familiarize you with the resources provided by OMEGAMON XE on z/VM andLinux. Predefined workspaces and situations enable you to start monitoring your enterprise as soon as theOMEGAMON XE on z/VM and Linux software is installed and configured.

The user interface supports several formats for viewing data, such as graphs, bar charts, and tables.Workspaces and situations can be customized to meet the needs of your enterprise.

This chapter provides overviews of the following topics:

v “Workspaces”

v “Custom queries” on page 17

v “Data filters” on page 17

v “Dynamic workspace linking” on page 18

v “Historical data” on page 20

v “Data collection interval” on page 21

v “Attributes” on page 22

v “Situations and situation events” on page 23

WorkspacesA workspace is the working area of the Tivoli Enterprise Portal window. You can start monitoring activityand system status immediately with the predefined workspaces. With just a few clicks of the mouse, youcan customize your own workspaces to give you summary overviews or to look at specific conditions.

OMEGAMON XE on z/VM and Linux provides a set of predefined workspaces. As you become morefamiliar with the product, you can modify the predefined workspaces or create new workspaces. For moreinformation about the predefined workspaces provided by the OMEGAMON XE on z/VM and Linuxproduct, see Chapter 3, “Predefined workspaces,” on page 27.

All of the workspaces that ship with OMEGAMON XE on z/VM and Linux are read-only. To make changesto these workspaces, save them under a different name. Descriptions of each workspace apply to thedefault settings of the original configuration.

Workspaces appear on the right of the application window. A workspace is divided into panes to showdifferent types of views. It can be divided into as many smaller frames, or panes, as you can reasonably fitinside the window. When you select an item in the Navigator, the default workspace opens with a set ofviews relevant to that level of the Navigator. Most workspaces contain at least one table. The columns in atable show the metrics that are displayed in the workspace. The rows in a table can contain links torelated workspaces that provide more detailed information. In addition to table views and graphs, aworkspace can contain other views, such as a notepad pane, a browser session, an event console, or aTake Action pane that gives you the ability to send commands to the operator console.

The source of data reported on in the Resource Constraint workspace is high-frequency sample data, thatis a set of counters and data that represent the state of the system at the moment they are sampled.High-frequency sampling is automatically activated whenever the MONITOR SAMPLE START command isissued, unless the RATE value has been previously set to STOP. Active high-frequency sampling in the CPMonitor is a prerequisite for Resource Constraint workspace views.

Whenever a table is refreshed with new values, those attributes that fall outside the threshold limits display

with a colored background: for critical; for warning; and for informational.


In addition, the column heading shows a , , or to indicate it has threshold cells, the scroll bararrows are tinted if any threshold cells are outside the viewable area, and the threshold expressionappears as flyover text when you move the mouse pointer over a threshold cell.

Every workspace has a set of properties associated with it: general properties that apply to the entireworkspace and properties for each view in the workspace. Use the Properties editor to customize theworkspace characteristics and to change the style and content of each view. You can also keep theoriginal workspace intact and create another workspace for the same item in the Navigator, customizing itfor the types of views you want and the information reported in charts and tables. Changes you make to aworkspace are available only to your user ID. System administrators can work in administration mode tocreate and edit workspaces that can be available to all users on the managed network.

Figure 3 shows a sample Tivoli Enterprise Portal application window for OMEGAMON XE on z/VM andLinux. Tivoli Enterprise Portal presents information in a single window comprising a navigation tree and aworkspace.

As you select items in the Navigator, each workspace presents views relevant to your selection. Everyworkspace has at least one view, and every view has a set of properties associated with it. You cancustomize the workspace by working in the Properties Editor to change the style and content of each view.You can also change, add, and delete views on a workspace.

Tivoli Enterprise Portal can present data in the following types of graphical views:

v Table view

v Pie chart view

v Bar chart view

v Plot chart view

v Circular gauge view

Figure 3. Tivoli Enterprise Portal sample workspace for OMEGAMON XE on z/VM and Linux


v Linear gauge view

Additional function is provided in the following Tivoli Enterprise Portal views:

v Tivoli Enterprise Console® view that integrates alarms and events from the major network managementoperating systems and correlates them with other events. If your environment includes the TivoliEnterprise Console, you can add it to any workspace and enjoy the advantages of the combinedmonitoring features of the console and portal.

v Situation event console view that shows the list of alerts that have occurred over a certain period. Thisview lists each alert for situations associated with the current Navigator item and any other Navigatoritems on this branch. The console has a toolbar for filtering the view to show only the alerts you want tosee and a pop-up menu with items for managing alerts.

v Notepad view.

v Message log view that shows the status of the situations associated with the system.

v Take Action view that is used to send a command to the monitored system. This monitoring agent hasspecific requirements for enabling Take Action commands. See Chapter 5, “Take Action commands,” onpage 89. See also the IBM Tivoli OMEGAMON XE on z/VM and Linux Planning and ConfigurationGuide for details.

v Terminal view, from which you can start a 3270 or 5250 work session. From the Terminal view, you canlog on to the z/VM system and access the Performance Toolkit. Once in the Performance Toolkit, youcan navigate to specific screens to view more detailed data and to investigate a problem.

v Browser view, from which you can open a browser to see HTML pages and Web sites.

Proactive monitoring using the Enterprise Status workspaceWhen you log in to Tivoli Enterprise Portal, the Enterprise Status workspace displays.

The Enterprise Status workspace gives you a high level, enterprise-wide summary of recent message andsituation activity, and includes the following views:

v Situation Event Console view that displays a row for every situation whose status changes to any ofthe following:

– Open (including expired and reopened events)

– Closed

– Problem

– Acknowledged

You can also view the owner for each acknowledged event, and you can acknowledge an open eventfrom the pop-up menu. Link anchors in this view and in the My Acknowledged Events view open theEvent Details - Similar by Situation workspace or you can right-click a link anchor to see the other EventDetails workspaces available for this view and for My Acknowledged Events.

v My Acknowledged Events view that displays the events assigned to the current user. This view showsboth open and closed events. Closed events are displayed in this view to enable you to view the notesand actions taken against the closed events. For each event record listed, you can drill down into thedetails of the event by clicking the link icon to the left of the event.

To further refine the details displayed for an individual event, right-click the event, click Link to andchoose one of the following workspaces:

– Event Details - Similar Events by Situation Name: Includes the Similar Events by Situation Nameview on the Event Details workspace. This view displays those events that have a similar situationname as the event that you are viewing.

– Event Details - Similar Events by Source: Includes the Similar Events by Source view on the EventDetails workspace. This view displays events from the same source as the event you are viewing.

– Event Details - Similar Events by Resource: Includes the Similar Events by Resource view on theEvent Details workspace. This view displays those events that are generated by the same resourceas the event you are viewing.

Chapter 2. Monitoring with OMEGAMON XE on z/VM and Linux 15

v Message Log view that displays a row for every situation in the monitored network whose statuschanges to any of the following:

– Open (including expired and reopened events)

– Closed

– Stopped

– Problem

– Acknowledged

v Open Situation Counts - Last 24 Hours view that displays a bar for every situation in the monitorednetwork that has become true in the past 24 hours. The bar size indicates the number of times asituation event has been opened during this time. For additional information on any of these views, referto the online help for Tivoli Enterprise Portal.

More predefined workspaces for the Enterprise Navigator item are accessible through the Navigatorpop-up menu and through the View menu. These are the following:

v EIB Change Log

v Managed System Status

v Self-Monitoring Topology

For additional information on any of these views, refer to the online help for Tivoli Enterprise Portal.

Figure 4 shows the Enterprise Status workspace.

The Navigator displays in the top left corner of the screen. The Enterprise node is at the top level of thenavigation tree. The next level is the operating system level, such as z/VM Systems. Below the operatingsystem level, the z/VM Linux Systems node indicates that the OMEGAMON XE on z/VM and Linux agentis installed.

For this scenario, the z/VM Linux Systems node is being actively monitored; other management agentsmight be available offline. The message log is updated, and event indicator icons display in the Navigator

Figure 4. Enterprise Status workspace


whenever a situation is triggered. In this scenario, the critical message in the message log indicates thatthe ZVM_LPAR_Busy_High situation was triggered, impacting the LPAR workspace. As the message logshows, other situations have triggered during the last 24 hours. Moving the mouse pointer over the eventindicator in the Navigator opens the hover help listing of open events.

If you are working to resolve an issue related to an event, acknowledge the event. Acknowledgmentsenable operators responsible for handling events to communicate their ownership of the event and itsworking status. Acknowledging an event places a blue check mark next to the situation in the event list. Ifthe situation is still true when the acknowledgment expires or if you cancel the acknowledgement before itexpires, the indicator changes accordingly. Right-click an event to view the list of actions available andclick Acknowledge.

Click on the link icon to the left of the event text to view the situation values and expert advice.

In cases where you can issue a command to resolve the problem, use Take Action to enter a commandon the system where the problem occurred. A situation can include a Take Action command that executeswhen the situation becomes true. Also referred to as reflex automation, Take Action enables you toautomate a response to system conditions. See Chapter 5, “Take Action commands,” on page 89 to findout how Take Action commands are implemented for this monitoring agent.

Expert Advice is provided for each situation. The administrator can update the expert advice to reflectspecific conditions in your enterprise.

You can create, edit, delete or view a situation using the Situation Editor. “Predefined situations providedby OMEGAMON XE on z/VM and Linux” on page 65 describes the situations provided with OMEGAMONXE on z/VM and Linux along with a brief explanation of the Situation Editor. Refer to the IBM TivoliMonitoring Enterprise Portal User's Guide for more detailed information about creating and customizingsituations.

Custom queriesCustom queries pre-filter the data used in a table or chart to ensure fastest data retrieval and to ensurethat you see no extraneous data. In your predefined workspaces, queries to the Tivoli EnterpriseMonitoring Server (the host data management component for Tivoli Enterprise Portal) retrieve attributesfrom your managed systems. You can create your own queries with only the attributes you want to see inthe view.

Predefined queries are delivered as part of Tivoli Enterprise Portal. If you decide to use a predefined queryand modify it for your environment, first make a copy using Create Another and then edit the copy.

You can also retrieve data from any Open DataBase Connectivity (ODBC) database by writing aStructured Query Language (SQL) SELECT statement. Using the Tivoli Enterprise Portal Query editor, youcan write SQL queries to ODBC data sources for retrieval and display in table and chart views. After thedata have been retrieved by the query, you can further refine the view with a filter to display only certainattributes or only those attributes with certain values.

For detailed information on creating custom queries and on the predefined queries available with TivoliEnterprise Portal, see the Tivoli Enterprise Portal online help.

Data filtersOMEGAMON XE on z/VM and Linux applies source filters to queries used by table and graph views. Withsource filters, you can quickly identify likely problems without paging through multiple screens of data.Source filters improve response time by filtering data close to the source, at the monitoring agent and atthe monitoring server.


Note: Deleting source filters increases the amount of data displayed and can cause poor response time atthe Tivoli Enterprise Portal.

OMEGAMON XE on z/VM and Linux uses display filters to focus graphs on specific indicators of potentialproblems, such as virtual CPU equal to or greater than 80 percent. A graph can use a query with sourcefilters as well as with display filters. Display filters can be displayed or changed at the Tivoli EnterprisePortal using the Properties editor.

Source filters might be required on multiple attributes to limit tables and graphs to provide only the mostuseful data. Source filters can also be displayed or changed at the Tivoli Enterprise Portal in the Queryeditor. You can define filters to simplify performance management for operators. Operators can view orchange filters provided by OMEGAMON XE on z/VM and Linux or by a systems administrator.

Dynamic workspace linkingDynamic workspace linking allows you to easily navigate between workspaces that are provided bymultiple products. This feature aids problem determination and improves integration across the monitoringproducts, allowing you to quickly determine the root cause of a problem. Predefined cross-product linksprovided by the OMEGAMON XE products allow you to obtain additional information about systems,subsystems, resources, or network components that are being monitored by other monitoring agents.When you right-click on a link, the list of links displays. This list might contain links to workspaces providedby one or more monitoring products. The product you are linking to must be installed and configured.

The OMEGAMON XE on z/VM and Linux monitoring agent contains dynamic workspace links toworkspaces in the Tivoli Monitoring Agent for Linux OS. The predefined links are provided from the“ApplData workspace” on page 59 and from the “Linux Workload workspace” on page 58. Theseworkspaces are accessed from the “Workload workspace” on page 57 of this monitoring agent. Choose aworkspace from the list to navigate to that workspace. By linking to the target workspace in context, youreceive additional information that is related to the system, subsystem, or resource you are currentlyviewing. If you choose a workspace from the list and the target workspace is not available, you receivemessage KFWITM081E.

The following requirements must be met for dynamic workspace linking to be activated between these twomonitoring agents:

v The Tivoli Monitoring Agent for Linux OS must be installed and active in order to link from TivoliOMEGAMON XE on z/VM and Linux workspaces to Tivoli Monitoring Agent for Linux OS workspaces.

v You must install and enable the application support for the Linux OS agent on the appropriate sharedtechnology components. See the IBM Tivoli Monitoring Installation and Setup Guide for details oninstalling and configuring the Tivoli Monitoring Agent for Linux OS.

v If you have the Tivoli Monitoring Agent for Linux OS installed and configured, and you would like to linkto workspaces in that monitoring agent, you must configure the KLZ_SETLPARVMID=Y environmentvariable in the Tivoli Monitoring Agent for Linux OS configuration file (lz.ini). If you are running IBM TivoliMonitoring, V6.2.0 and above, you make an additional modification to the lz.ini file by commenting outthe CTIRA_HOSTNAME.

Important: If the Tivoli Monitoring Agent for Linux OS is started prior to adding the environment variableto the configuration file, a duplicate branch for the Linux OS agent may display in Tivoli EnterprisePortal. The original node will be dimmed in the Navigator. To clear the original node, right-click theNavigator item and click Clear offline entry. The original node will be deleted.

v You must enable data collection for the “KVLUser ApplData attributes” on page 136 at each Linux guestsystem that is to be monitored. Enabling data collection for this attribute group defines the workgroup asa Linux workgroup. This is important because the links are displayed on a table that queries thesystems in the Linux workgroup.


v You must enable the CP Monitor domain for the type of data you want to collect on the z/VM systemwhere the Performance Toolkit is running. For details on enabling the collection of Linux data and onenabling dynamic workspace linking, see the IBM Tivoli OMEGAMON XE on z/VM and Linux Planningand Configuration Guide.

v Your Tivoli Enterprise Portal user ID must be authorized to access the target product. Otherwise, links toworkspaces in the targeted product are not included in the list. See the online help system for TivoliEnterprise Portal for information on assigning access privileges.

See the IBM Tivoli OMEGAMON XE on z/VM and Linux Planning and Configuration Guide forconfiguration details. See the IBM Tivoli Monitoring: Agent for Linux OS User's Guide for descriptions ofthe workspaces available with the Linux OS monitoring agent.

Note: Dynamic workspace linking between the Tivoli OMEGAMON XE on z/VM and Linux monitoringagent and the Tivoli Monitoring Agent on Linux OS is not supported when the Linux system defined for theLinux OS agent is running as a guest under a second-level z/VM system.

The table that follows summarizes the links that are available:

Table 2. Links from OMEGAMON XE on z/VM and Linux to Tivoli Monitoring Agent for Linux OS

OMEGAMON XE on z/VMand Linux workspace

Target workspace in TivoliMonitoring Agent forLinux OS

Attributes used to locatetarget workspace

Attributes Used to FilterData in Target Workspace

Workload > Linux Workload- Linux Guest SystemWorkloads table

Linux Process Linux Guest ID None


Linux System Information Linux Guest ID None


Linux Virtual MemoryStatistics

Linux Guest ID None


Linux CPU Averages Linux Guest ID None


Linux Virtual MemoryUsage Trends

Linux Guest ID None

Workload > ApplData -Linux Guest Appl Data table





Linux Virtual Memory Linux Guest ID None


Linux Disk IO Rate Linux Guest ID None


Linux Network Linux Guest ID None


Linux Sockets Linux Guest ID None


Linux Capacity Usage Linux Guest ID None




Table 2. Links from OMEGAMON XE on z/VM and Linux to Tivoli Monitoring Agent for Linux OS (continued)

OMEGAMON XE on z/VMand Linux workspace

Target workspace in TivoliMonitoring Agent forLinux OS

Attributes used to locatetarget workspace

Attributes Used to FilterData in Target Workspace


Linux Virtual MemoryTrends

Linux Guest ID None

Workload > ApplData -Linux Guest Workload Datatable





Linux Virtual MemoryStatistics

Linux Guest ID None




Linux Virt. Memory UsageTrends

Linux Guest ID None

Historical dataIn addition to seeing real-time data in the table and chart views, you can store the data, using the TivoliData Warehouse, so that you can examine data over longer periods of time. The table view and the bar,pie, and plot charts have a tool for setting a time span, which causes previous data samples to bereported up to the time specified.

OMEGAMON XE on z/VM and Linux contains workspaces that provide additional views of historical data.

If you see the Time Span tool in a report, historical data is available for that report.

The data shown in historical table and chart views is retrieved from binary history files that were set upthrough the History Collection Configuration window. If you open a historical predefined workspace but seeonly real-time data, it means that the attribute group used in the workspace requires that historical datacollection be configured and started. Configuring historical data collection involves specifying the followingitems:

v The attribute groups to save data samplings from

v The collection interval

v The roll off interval, if any

v Where to store the collected data (either at the agent or at the monitoring server)

The binary files containing the historical data grow as new data are added at every sampling interval. Theirsize can increase quickly and take up a great deal of disk space. The larger a history file is, the longer ittakes to retrieve historical data into views. You must manually maintain these files unless you configure theTivoli Data Warehouse. You can move the data to the data warehouse database. You can also use thehistory file conversion programs provided by IBM that move data out of the historical files to delimited textfiles. After roll off, the history binary files are cleared of entries older than 24 hours to make room for newentries. You can then use the converted historical data with any reporting tool from a third-party vendor orload the converted data into a relational database to produce customized history reports. See the IBMTivoli Monitoring Administrator's Guide for information about converting history files to delimited flat files.


Note: You must start historical data collection for the monitoring agent for which you want historical databefore you can use the history feature. See the Tivoli Enterprise Portal online help for details onconfiguring historical data collection. See the IBM Tivoli Monitoring Installation and Setup Guide forinformation on installing and configuring the Tivoli Data Warehouse.

The Summarization and Pruning Agent is a mechanism for managing data in the Tivoli Data Warehouse.The data in the Tivoli Data Warehouse is a historical record of activity and conditions in your enterprise.Summarization of the data is the process of aggregating your historical data into time-based categories.For example, hourly, daily, weekly, and so on. Summarizing data allows you to perform historical analysisof the data over time. Pruning of the data keeps the database to manageable size and thus improvesperformance. Pruning of the database should be performed at regular intervals. See the IBM TivoliMonitoring Administrator's Guide for information on summarization and pruning of data.

Tivoli Common Reporting toolThe Tivoli Common Reporting tool is part of the product package for this monitoring agent. The TivoliCommon Reporting tool is a reporting feature available to users of Tivoli products. This tool provides aconsistent approach to viewing and administering reports.

The Tivoli Common Reporting zip files containing reports, documentation, graphics, and dynamic linklibraries for IBM Tivoli Monitoring and OMEGAMON monitoring agents are available from the IBM TivoliOpen Process Automation Library at http://www.ibm.com/software/tivoli/opal. Search on “Tivoli CommonReporting” to find report packages on OPAL.

A sample set of reports for the Tivoli OMEGAMON XE on z/VM and Linux monitoring agent are availableon OPAL for use with the Tivoli Common Reporting product. The Tivoli OMEGAMON XE on z/VM andLinux package consists of the following customizable reports:

v z/VM Linux System CPU Utilization

v z/VM VM System CPU Utilization

v z/VM VM System CP-Owned Device Utilization

v z/VM VM System Paging Utilization

v z/VM VM System TCP Server Statistics

Other sets of reports can be downloaded and installed using the Import facility. For information about thesoftware requirements for Tivoli Common Reporting, refer to the IBM Tivoli Common Reporting User’sGuide. See also the Tivoli Common Reporting: Reports for Tivoli OMEGAMON XE Agents and TivoliNetView® for z/OS guide.

Viewing event data using the Tivoli Enterprise ConsoleThe Tivoli Enterprise Monitoring Server generates Tivoli Enterprise Console events with classes that areunique to each monitoring agent. This monitoring agent provides a .baroc file with the Tivoli EnterpriseConsole classes that are generated by IBM Tivoli Monitoring. This is an optional feature and required onlyif you intend to view event data using the Tivoli Enterprise Console. To view event data in the TivoliEnterprise Console, install the monitoring agent .baroc file on the event server. Add the .baroc file afteryou have added application support for the agent to the monitoring server. For instructions on installing the.baroc file, and on adding application support, see the IBM Tivoli Monitoring Installation and Setup Guide.

Data collection intervalThe collection interval controls how often real-time data are collected. Each time you collect data, themonitoring agent will incur a certain amount of CPU usage. How much processing is incurred depends onthe number and types of resources you monitor.


Set the collection interval to meet your monitoring needs, ensuring that you are promptly alerted toproblems so they can be resolved quickly. But, also be aware that each data collection comes with a costin CPU usage.

CP Monitor sample intervalThe data for the OMEGAMON XE on z/VM and Linux monitoring agent is collected using the PerformanceToolkit. Data collection is linked to the sample interval set at the CP Monitor. The default sample interval is60 seconds. Setting a longer interval will reduce the data collection overhead for both CP MONITOR andthe Performance Toolkit.

To determine the CP Monitor interval set in your environment, issue a QUERY MONITOR INT commandon z/VM. This command displays the interval value currently in effect for sample monitoring on z/VM.

You control data collection for performance monitoring with the FCONTROL MONCOLL subcommand ofthe Performance Toolkit subcommand. A new option called SEGOUT has been added to FCONTROLMONCOLL that applies specifically to this monitoring agent. The use of this subcommand requiresprivilege class E.

You can access information on the Performance Toolkit and on the latest enhancements to thePerformance Toolkit as they pertain to this monitoring agent at the following Web address:

http://www.vm.ibm.com/related/perfkit/pksegout.html

See the z/VM: Performance Toolkit Reference for information on using the FCONTROL MONCOLLsubcommand.

Linux guest ApplData intervalThe interval set at the Linux guest for application monitor data collection determines how often new dataare provided to the CP Monitor. You may want to set the interval at the Linux guest at a rate different thanthe CP Monitor sample interval. For example, setting the interval at the Linux guest to once every thirtyminutes may be sufficient for your needs. It depends on your environment, and on how closely you want tomonitor the Linux systems. Setting each interval independently provides greater control of data collectionoverhead.

The default setting at the Linux guest is 10,000 milliseconds (once every ten seconds) of CPU time (notwall clock time). This setting should offer an adequate compromise between performance and currentdata.

As an example, the following command sets the interval at the Linux guest to 20 seconds:echo 20000 > /proc/sys/appldata/interval

The format of the command is as follows:echo <milliseconds> > /proc/sys/appldata/interval

where <milliseconds> is the number of milliseconds of CPU time (not wall clock time) that elapse beforenew data are passed to the CP Monitor.

For additional information, see “Enabling the collection of KVLUser ApplData” on page 60.

AttributesOMEGAMON XE on z/VM and Linux displays the z/VM data collected by the Performance Toolkit. It alsodisplays the Linux on zSeries data for the agents residing on the monitored Linux on zSeries systems. Itstores both types of data in system elements called attributes. You can use these attributes to monitorsystem performance, build custom workspaces, and create situations to alert you of pending problems.



Related attributes are grouped into attribute groups (also called attribute tables). Each table view containsinformation provided by a single attribute group.

For a complete description of the OMEGAMON XE on z/VM and Linux attributes, see “Attributesreference” on page 101 or the online help for this monitoring agent.

Using attributes in queriesGraph and table views use queries to specify which attribute values and monitored resources to requestfrom a monitoring agent. You can use the Query Editor to create a new query, modify an existing one, orapply filters and set styles to define the content and appearance of a view based on an existing query.

For instructions on using the Query Editor, see the Tivoli Enterprise Portal online help or the IBM TivoliMonitoring Enterprise Portal User's Guide.

Situations and situation eventsA situation describes a condition or a set of conditions that you set to determine whether a problem existsin one or more monitored systems and resources. A condition consists of an attribute, a value, and acomparison operator. The value of the attribute is compared with the value set for the condition todetermine whether the condition is met. For example, the ZVM_CP_CPU_Critical situation is true when thevalue of the CP Percent of CPU attribute is greater than or equal to 30 percent.

You can create complex situations that contain more than one condition, allowing you to compareattributes and values that represent characteristics of specific problems. When all the conditions of asituation are met (the situation is triggered), a situation event is registered. The operator is alerted tosituation events by indicator icons that are displayed in the Navigator. Operators can also be alerted bysound. As you move up the Navigator hierarchy, situation events are consolidated to show only theindicator with the highest severity level (Critical, followed by Warning, then Informational).

Figure 5 shows an example of the Tivoli Enterprise Portal Navigator with Critical, Warning, andInformational event indicators.

A situation can include a Take Action command that runs when the situation is triggered. You canautomate a response to a specific system condition. In addition, each situation can include text describingthe probable cause and expert advice so that you can address and resolve problems quickly.

Figure 5. Tivoli Enterprise Portal Navigator with event indicators


Note: If you create a situation and specify that an Action be run, be aware of the following:

v For this monitoring agent, select Run the action at the Managed System (Agent) for the Whereshould the action be run (Performed) option.

v When you select System Command on the Action tab, this field displays for you to type a command toissue at the system. Be sure to prefix all OMEGAMON XE on z/VM and Linux commands with VL:.

For information on enabling Take Action commands for this monitoring agent, see Chapter 5, “Take Actioncommands,” on page 89. See also the IBM Tivoli OMEGAMON XE on z/VM and Linux Planning andConfiguration Guide.

You can create or modify situations from the Tivoli Enterprise Portal user interface by using the SituationEditor. From the Situation Editor, you can specify situations to run at startup or you can start and stopsituations manually. OMEGAMON XE on z/VM and Linux provides a set of default situations to enable youto start monitoring your enterprise as soon as the product is installed and configured. Some of thepredefined situations are set to start automatically. Other situations must be started manually.

You can also use the Situation Editor to create new situations to meet the needs of your enterprise. Besure to save the original situation and use Create Another to create your own situation.

For information about the predefined situations provided with OMEGAMON XE on z/VM and Linux, see theproduct online help or Chapter 4, “Predefined situations,” on page 63. For instructions on using theSituation Editor, see the Tivoli Enterprise Portal online help or the IBM Tivoli Monitoring Enterprise PortalUser's Guide.


Part 2. Resources provided by OMEGAMON XE on z/VM andLinux

This section describes the predefined workspaces and situations provided by this monitoring agent, andhow you can use them to help you meet the specific needs of your environment. This section also explainsthe attributes associated with the workspaces and how to use Take Action commands.

Chapter 3, “Predefined workspaces,” on page 27 describes the predefined workspaces provided by thisagent

Chapter 4, “Predefined situations,” on page 63 describes the predefined situations available with thismonitoring agent.

Chapter 5, “Take Action commands,” on page 89 describes the use of Take Action commands with thismonitoring agent. This chapter also provides instructions on how you can configure your environment toenable Take Action commands.


Chapter 3. Predefined workspaces

The list of OMEGAMON XE on z/VM and Linux predefined workspaces appear in the Navigator tree viewin the Physical view of Tivoli Enterprise Portal. In addition to the workspaces listed in the tree view, youcan reach the secondary workspaces from the primary workspaces (that is, the workspaces listed in theNavigator tree view).

The hierarchy levels shown in the Navigator tree depend upon the customization of the Tivoli EnterprisePortal for your environment. IBM provides a set of predefined workspaces that do not require anycustomization by you. The list below shows the order and hierarchy of the predefined workspaces thatdisplay in OMEGAMON XE on z/VM and Linux. It is a representation of how the predefined workspacesare organized in the Navigator. The symbol indicates that the workspace is a secondary one. Somesecondary workspaces are accessible from more than one primary workspace or are primary workspacesas well. For more detailed information about a workspace, click its name in this list.

The following workspaces are provided by OMEGAMON XE on z/VM and Linux:

“z/VM Systems: System Health” on page 30

“z/VM Linux Systems: default workspace” on page 32

“CP Owned Devices workspace” on page 33

“Channel workspace” on page 35

“FICON Channels workspace” on page 36

“DASD workspace” on page 37

“Minidisk Cache workspace” on page 38

“CCW Translations workspace” on page 40

“Control Unit Cache workspace” on page 40

“DASD Cache workspace” on page 42

“VDISK workspace” on page 43

“LPAR workspace” on page 44

“Processor workspace” on page 47

“Processor by LPAR Name workspace” on page 46

“Network workspace” on page 48

“VSWITCH workspace” on page 50

“Real Storage workspace” on page 51

“System workspace” on page 52

“Spin Locks workspace” on page 53


“System Terminal workspace” on page 54

“TCPIP workspace” on page 55

“TCPIP User workspace” on page 56

“Workload workspace” on page 57

“Linux Workload workspace” on page 58

“ApplData workspace” on page 59

“Resource Constraint workspace” on page 61

All of the workspaces that ship with OMEGAMON XE on z/VM and Linux are read-only. To make changesto these workspaces, save them under a different name. See the Tivoli Enterprise Portal online help fordetails.

Note: Descriptions of each workspace apply to the default settings of the original configuration. Changesand additions that you make to a workspace are not described in the online help.

Additionally, data displays only when data collection for the CP Monitor domain associated with that typeof data is active. See the IBM Tivoli OMEGAMON XE on z/VM and Linux Planning and ConfigurationGuide for details on enabling data collection for this monitoring agent.

Accessing OMEGAMON XE on z/VM and Linux workspacesThe various levels of access to the OMEGAMON XE on z/VM and Linux workspaces are as follows:

v Enterprise is at the highest level. It encompasses all of the systems in your organization where themonitoring agents are installed.

v Operating platform is the operating system being monitored, such as Windows Systems, LinuxSystems, or z/VM Systems.

v System is the name of the computer where the monitoring agents are installed.

v Agent is the name of the monitoring agent installed on the system. If the agent name is dimmed,the agent is offline.

At this level of the Navigator, you can rename an item to add the z/VM system name to the agentname. To do so, right-click the item and click Properties. In the Properties window that opens, type thenew name in the Name field, add or change the description in the Description field, and click OK.

v Attribute is the category of attributes the agent is monitoring. Each entry is associated with one ormore workspaces that provide information on those monitored objects. For example, the DASDworkspace provides information about DASD activity.

To access the OMEGAMON XE on z/VM and Linux workspaces from the Navigator, perform the followingsteps:

1. Expand the Navigator tree by double-clicking the Enterprise item, if necessary.

2. Select z/VM Systems.

3. Under z/VM Systems, expand the entry that corresponds to the host name of the computer whereOMEGAMON XE on z/VM and Linux is installed.


4. Within that list of monitored subsystems, expand the entry for z/VM Linux Systems.

5. Select a workspace from the list of workspaces.

Accessing the secondary workspacesThe secondary workspaces for each primary workspace are accessible by one or more of the followingmethods:

From the Navigator1. Select the primary workspace.

2. Right-click the name of the selected workspace in the Navigator.

3. Select Workspaces from the context menu.

4. Select the desired secondary workspace.

From the View menu1. Select the primary workspace.

2. In the menu bar at the top of the Tivoli Enterprise Portal, select View > Workspaces.

3. Select the desired secondary workspace.

From a report1. Select the primary workspace.

2. If the workspace table (which by default displays at the bottom of the workspace) contains a link iconto the left of each row, you can click the icon to navigate to the default secondary workspacepertaining to the selected row. You can also right-click the icon and select a secondary workspace fromthe context menu.

Figure 6 shows a sample linked table view.

From a chart viewThe data displayed in some bar charts and plot charts are linked to secondary workspaces. To look for alink, right-click a bar or data point in the chart. If Link to > appears in the context menu, you can select asecondary workspace pertaining to the data in the chart.

Attribute groups used by predefined workspacesThe table that follows shows the relationships between the predefined workspaces and the attributegroups. In most cases, a workspace contains data or columns that have similar attributes in an attributegroup. Some of the listed workspaces are secondary workspaces.

Table 3. Attributes used by predefined workspaces

Workspace name Related attribute group

“z/VM Linux Systems: default workspace”on page 32

“KVLDevice attributes” on page 110, “KVLPTKStat attributes” on page123, “KVLSystem attributes” on page 125, and “KVLUser Workloadattributes” on page 140

Figure 6. Link icon in a sample table view

Chapter 3. Predefined workspaces 29

Table 3. Attributes used by predefined workspaces (continued)

Workspace name Related attribute group

“z/VM Systems: System Health” “KVLSystem attributes” on page 125 and “KVLSystem2 attributes” onpage 129

“ApplData workspace” on page 59 “KVLUser ApplData attributes” on page 136

“CP Owned Devices workspace” on page33

“KVLCP Device attributes” on page 101 and “KVLDevice attributes” onpage 110, and “KVLSystem attributes” on page 125

“Channel workspace” on page 35 “KVLChannel Data attributes” on page 102 and “KVLLChannel Dataattributes” on page 113

“Control Unit Cache workspace” on page40

“KVLControlUnit attributes” on page 104

“DASD workspace” on page 37 “KVLDevice attributes” on page 110

“DASD Cache workspace” on page 42 “KVLDASDCache attributes” on page 106

“CCW Translations workspace” on page 40 “KVLMinidisk Cache attributes” on page 118

“FICON Channels workspace” on page 36 “KVLFChannel Data attributes” on page 111

“LPAR workspace” on page 44 “KVLLPAR Info attributes” on page 115

“Linux Workload workspace” on page 58 “KVLUser Workload attributes” on page 140

“Minidisk Cache workspace” on page 38 “KVLMinidisk Cache attributes” on page 118

“Network workspace” on page 48 “KVLHiperSocket attributes” on page 112 and “KVLVirtualSwitchattributes” on page 145

“Processor workspace” on page 47 “KVLProcessor Data attributes” on page 121

“Processor by LPAR Name workspace” onpage 46

“KVLProcessor Data attributes” on page 121

“Real Storage workspace” on page 51 “KVLSystem attributes” on page 125

“Resource Constraint workspace” on page61

“KVLUser Wait attributes” on page 131

“Spin Locks workspace” on page 53 “KVLSpinLock attributes” on page 124

“System workspace” on page 52 “KVLSystem attributes” on page 125

“TCPIP workspace” on page 55 “KVLTCPIP Srvr Data attributes” on page 133

“TCPIP User workspace” on page 56 “KVLTCPIPUsrData attributes” on page 135

“VDISK workspace” on page 43 “KVLVdisk attributes” on page 143

“VSWITCH workspace” on page 50 “KVLVirtualSwitch attributes” on page 145

“Workload workspace” on page 57 “KVLUser Workload attributes” on page 140 and “KVLTCPIPUsrDataattributes” on page 135

z/VM Systems: System HealthThe z/VM Systems branch is the top level of the Navigation tree for the z/VM operating system. Whenyou click z/VM Systems, the System Health workspace displays. This workspace contains six views(excluding the navigator) and provides data for all of the z/VM systems that are registered with the TivoliEnterprise Monitoring Server. Each view contains one bar (and the tabular view contains one row of data)for every z/VM agent registered with this Tivoli Enterprise Monitoring Server.

Use this workspace to view the activity and health of your z/VM environment.

This workspace includes the following views:


v The System Health: CPU Utilization bar chart that displays the total CPU percent utilization of thesystem and its component parts: overhead and virtual CPU. If you are running multiple processors, thisvalue is the sum of the CPU utilization for all of the active processors and can be greater than 100%.

From this view, you can navigate to the “System workspace” on page 52 for a selected System ID. Togo to this workspace, right-click on the bar of the system for which you want additional data and click“System workspace” on page 52.

v The System Health: Real Memory Overcommit bar chart that displays the ratio of virtual memory toavailable real main storage.

From this view you can navigate to the “Real Storage workspace” on page 51 for a selected System ID.To go to this workspace, right-click on the bar of the system for which you want additional data and click“Real Storage workspace” on page 51.

v The System Health: DASD Page Space Utilization bar chart that displays the utilization percent of thedefined paging space on disk.

From this view you can navigate to the “CP Owned Devices workspace” on page 33 for a selectedSystem ID. To go to this workspace, right-click on the bar of the system for which you want additionaldata and click “CP Owned Devices workspace” on page 33.

v The System Health: Main Storage Paging Utilization bar chart that displays the overall systempaging rate for each system registered with the Tivoli Enterprise Monitoring Server.

From this view you can navigate to the “System workspace” on page 52 for a selected System ID. Togo to this workspace, right-click on the bar of the system for which you want additional data and click“System workspace” on page 52.

v The System Health: Highest Workload Wait Reason bar chart that displays the highest wait reasonfound for a single user on any of the systems registered with the Tivoli Enterprise Monitoring Server.

From this view you can navigate to the “Resource Constraint workspace” on page 61 for a selectedSystem ID. To go to this workspace, right-click on the bar of the system for which you want additionaldata and click “Resource Constraint workspace” on page 61.

v The z/VM Health: All Systems table displays the attributes collected for each z/VM agent registeredwith the Tivoli Enterprise Monitoring Server.

From this view you can navigate to any of the following workspaces for a selected System ID:

– “System workspace” on page 52

– “Real Storage workspace” on page 51

– “Resource Constraint workspace” on page 61

– “CP Owned Devices workspace” on page 33

To go to one of these workspaces, right-click on the table row of the system for which you wantadditional data and click the workspace that you want.

See the “KVLSystem attributes” on page 125 and “KVLSystem2 attributes” on page 129 for more detailson the data displayed in this view.

Figure 7 on page 32 shows a sample z/VM Systems System Health workspace.


z/VM Linux Systems: default workspaceWhen you click the z/VM Linux Systems node in the Navigator view, the z/VM Linux Systems defaultworkspace for this monitoring agent displays.

This workspace provides the following information:

v Gives you an instant view of whether or not the Performance Toolkit is active.

v Displays the type of data that are being collected during the current collection interval.

v Provides Top 5 system-wide data on DASD activity, on z/VM user CPU utilization, and on z/VM userworking set storage.

This workspace includes the following views:

v The z/VM PTK Collector Status table provides the status of the Performance Toolkit, and whether ornot it is active. If the Performance Toolkit is active, it is collecting data from z/VM and passing it on tothe monitoring agent. Under the column labelled Collector Name, this table also displays the types ofdata that are being collected. See the “KVLPTKStat attributes” on page 123 for more details on the datadisplayed in this view.

Important: If the Performance Toolkit stops updating the PERFOUT DCSS with data that it hascollected, the predefined situation ZVM_PerfKit_Collector_Inactive raises an alert, indicating aproblem. Investigate the cause immediately. For more information on the possible causes, see thedescription for ZVM_PerfKit_Collector_Inactive in “Predefined situations and formulas” on page 68. Seethe IBM Tivoli OMEGAMON XE on z/VM and Linux Planning and Configuration Guide for details onconfiguring your environment for this monitoring agent. See also the IBM Tivoli OMEGAMON XE onz/VM and Linux Troubleshooting Guide for troubleshooting information.

v The Top 5 Busy DASD Devices bar chart displays, for the five busiest devices, by volume serialnumber, the percentage of time each device was busy. See the “KVLDevice attributes” on page 110 formore details on the data displayed in this view.

Figure 7. z/VM Systems System Health


v The Top 5 Workloads by CPU Utilization bar chart displays, for the five busiest workloads, byworkload identifier, the following data:

– The percentage of total CPU (CP % of CPU) used by the Control Program to manage workloads.

– The total CPU (CPU Percent), which is the sum of the CP % of CPU and the virtual CPUpercentage, used on behalf of this workload.

– The percentage of virtual CPU (Virtual CPU %) being used for workloads.

Note: When multiple virtual processors are used, the percentages can exceed 100%.

See the “KVLSystem attributes” on page 125 for more details on the data displayed in this view.

v The Top 5 Workloads by Working Set bar chart shows, for the five busiest workloads, by workloadidentifier, the following data:

– The current number of pages (below the 2 GB line) physically in the main storage.

– The current number of pages (above the 2 GB line) physically in the main storage.

– The projected working set size, in pages.

See the “KVLUser Workload attributes” on page 140 for more details on the data displayed in this view.

Figure 8 shows a sample z/VM Linux Systems default workspace.

CP Owned Devices workspaceThe CP Owned Devices workspace provides several views that display data on the activity of the pagingand spooling devices on the overall z/VM system. You can use this workspace to determine, during thecollection interval, the resource utilization by the paging and spooling devices, to identify those deviceswith the top five service times, and to identify the five busiest devices.

Data Source:

The information shown is based on CP monitor STORAGE domain sample data, and on MONITORdomain configuration records.

Figure 8. z/VM Linux Systems default workspace


This predefined workspace contains the following views:

v The Paging and Spooling Space bar chart displays the top ten devices using the greatest amount ofpaging and spooling space. For each device managed by the z/VM Control Program, it shows thenumber of cylinders or blocks allocated, and the number of slots available on the device.

v The Top 5 Page Extent Utilization bar chart shows, by device address, the five devices using thegreatest amount of paging space. For each paging device, it shows the percentage of space in use onthe volume managed by the Control Program during the collection interval.

v The Top 5 Dump Extent Utilization bar chart shows, by device volume serial number, the five devicesusing the greatest percentage of the dump slots allocated to those devices. For each device, it showsthe percentage of dump slots in use on the volume managed by the Control Program during thecollection interval.

v The Top 5 Spool Extent Utilization bar chart shows, by device volume serial number, the five devicesusing the greatest amount of spooling space. For each spooling device, it shows the percentage ofspace in use on the volume managed by the Control Program during the collection interval.

v The CP Device Table (Paging and Spooling) displays data such as the number of available slots onthe device, the type of device, the number of cylinders or blocks allocated to a device, the LPARidentifier, and the percentage of space in use on the volume during the collection interval.

See the “KVLCP Device attributes” on page 101, the “KVLDevice attributes” on page 110, and the“KVLSystem attributes” on page 125 for more details on the data displayed in this view.

Figure 9 shows a sample CP Owned Devices workspace.

Figure 9. CP Owned Devices workspace


Channel workspaceThe Channel workspace displays channel load data for all active channels for which valid information wasfound. This workspace provides several views that display the activity of the busiest channels. You canuse this workspace to identify, during the collection interval, the channels utilizing the most resources.

Data Source:

Overall channel load analysis is based on data from the Extended Channel-Path Measurement Facility,where available; otherwise high-frequency sampling data from the CP monitor SYSTEM domain is used.

The predefined Channel workspace contains the following views:

v The Top 5 z/VM Channel Busy bar chart displays the five busiest channels on z/VM. For each installedchannel, the chart shows the percentage of busy conditions found, calculated for the entire period sincethe last RESET, or since the Performance Toolkit was last invoked. This bar chart also displays, foreach channel, the percentage of busy conditions found, calculated for the last interval. Channel busypercentages cannot be calculated for HiperSockets channels.

v The Top 5 LPAR Channel Busy - Owning Partition bar chart displays the five busiest channels forsystems running in the owning logical partition. The values shown pertain to the channel activitygenerated by the owning partition only, not the overall channel activity. For each installed channel, thechart shows the average percentage of busy conditions found, calculated for the entire period since thelast RESET, or since the Performance Toolkit was last invoked. This bar chart also displays, for eachchannel, the percentage of busy conditions found, calculated for the last interval. Note: Channel busypercentages cannot be calculated for HiperSockets channels.

v The Top 5 LPAR Channels by Interval Busy Percent - Owning Partition bar chart that shows thefive channels with the highest channel load distribution for systems running in the owning logicalpartition. The values shown pertain to the channel activity generated by the owning partition only. Foreach channel, the bar chart displays the relative frequency with which the channel busy percentage ofseparate measurement intervals occurred in any of the 10% ranges starting with 11-20 percent all theway through to 91-100 percent.

v The z/VM Channel Busy table displays, for each channel, data such as the hexadecimal channelmodel group, the channel path identifier (CHPID), the average percentage of busy conditions found,calculated for the entire period since the last RESET, or since the Performance Toolkit was last invoked,and the channel load distribution.

v The LPAR Channel Busy - Owning Partition table displays, for each channel within the owning logicalpartition, data such as the hexadecimal channel model group, the channel path identifier (CHPID), theaverage percentage of busy conditions found, calculated for the entire period since the last RESET, orsince the Performance Toolkit was last invoked, the channel load distribution. This table also indicateswhether the channel is shared with another logical partition or whether it is a dedicated channel.

The Channel workspace provides a link to a secondary OMEGAMON XE on z/VM and Linux workspace.From the Navigator, right-click the name of the Channel workspace, select Workspace, then select“FICON Channels workspace” on page 36.

See the “KVLChannel Data attributes” on page 102 for more details on the data displayed in this view.

Figure 10 on page 36 shows a sample Channel workspace.


FICON Channels workspaceThe FICON Channels workspace displays FICON® channel load data for all active FICON channels forwhich valid information was found.

This workspace provides several views that display the load data of the FICON channels. You can use thisworkspace to identify, during the collection interval, the FICON channels with the highest levels of activity.

Data Source:

FICON channel load analysis is based on data from the Extended Channel Path Measurement Facility,available from the CP monitor SYSTEM domain.

The FICON Channels predefined workspace contains the following views:

v The FICON Channel Utilization - System bar chart displays the FICON channel utilization on theoverall system. For the entire system, the bar chart shows percentages for the bus cycles utilization, thework units utilization, the data units write utilization, and the data units read utilization.

v The FICON Channel Utilization - Owning Partition bar chart displays the FICON channel utilizationfor the owning logical partition. This bar chart displays percentages for work units utilization, for dataunits write utilization, and for data units read utilization.

v The FICON Channel Utilization - Transfer Rates bar chart displays FICON channel transfer rates forthe entire system. The bar chart displays the total number of bytes written per second. The bar chartalso displays the total number of bytes read per second for the whole system.

v The FICON Channel Utilization table displays FICON channel utilization for the overall system. Thetable also displays FICON channel utilization for the owning logical partition. Data are provided such ashexadecimal channel path identifier, bus cycles utilization, work units utilization, data units write

Figure 10. Channel workspace


utilization, data units read utilization, and percentage of owning LPAR work units. This table alsoindicates whether the FICON channel is shared with another logical partition or whether it is a dedicatedFICON channel.

See the “KVLFChannel Data attributes” on page 111 for more details on the data displayed in thisworkspace.

Figure 11 shows a sample FICON Channels workspace.

DASD workspaceThe DASD workspace enables you to examine and to monitor the performance of the devices connectedto your system. This workspace provides several views that display the activity of the busiest I/O deviceson the overall z/VM system. You can use this workspace to determine, during the collection interval, thedevices with the worst connection time, the average number of I/O requests queued to a device, and theaverage service time for each device.

Data Source:

The information shown in this workspace is based on CP monitor MONITOR domain and I/O domainSAMPLE data.

The DASD predefined workspace contains the following views:

v The Top 5 Device Busy bar chart displays the five busiest devices. For each device, the chart showsthe percentage of time a device was busy.

v The Top 5 I/O Rate bar chart displays the five devices with the highest number of I/O requests started.

v The Top 5 Service Time for DASD I/O bar chart shows the five devices with the longest service timesduring this collection interval. For each device, metrics are collected for the average time a device wasconnected, the average time a device was disconnected, and the average time a device was in pendingstate, that is, a request was waiting in the I/O subsystem before being passed on to the device.

Figure 11. FICON Channels workspace


v The Top 5 I/O Queue Depth bar chart displays the average number of I/O requests queued on each ofthe five busiest devices.

v The DASD I/O Activity table provides data on system utilization by each of the assigned devices. Thetable shows data such as the average time a device was connected, the average number of I/Orequests queued for a device, and the average service time for a device, in milliseconds. Use this tableto determine how well each device is performing and where problems might be occurring.

The DASD workspace contains links to several secondary workspaces that provide additional detailed dataon DASD activity. To access these secondary workspaces from the Navigator, select the DASDworkspace, select Workspace, then select one of the following secondary workspaces:

v “Minidisk Cache workspace”

v “CCW Translations workspace” on page 40

v “Control Unit Cache workspace” on page 40

v “DASD Cache workspace” on page 42

v “VDISK workspace” on page 43

See the “KVLDevice attributes” on page 110 for more details on the data displayed in this view.

Figure 12 shows a sample DASD workspace.

Minidisk Cache workspaceThe Minidisk Cache workspace enables you to examine minidisk cache usage. Use this workspace todetermine minidisk cache storage utilization.

If you find that your devices have poor response times, check to see if some of the I/O load has beencaused by minidisks with a high activity on the same disk pack.

Data Source:

Figure 12. DASD workspace


The information shown is based on the CP monitor SYSTEM domain SAMPLE data.

This Minidisk Cache predefined workspace contains the following views:

v The Cache Requests bar chart displays the number of read requests to the minidisk cache, persecond, where all the requested blocks were found in the cache, thereby avoiding an I/O operation, andit displays the number of requests per second to invalidate minidisk cache blocks, due to an I/O to avirtual device via a non-cachable I/O interface. This bar chart also shows the number of blocks persecond that were invalidated following an invalidation request.

v The Main Storage Frames bar chart displays the actual number of main storage page frames used forthe minidisk cache below the 2 GB line, the actual number of main storage page frames used for theminidisk cache above the 2 GB line, and the ideal number of main storage page frames in the minidiskcache (based on the average age of a page in the system's dynamic paging area, determined byarbiter).

v The Cache Age bar chart shows the estimated average age in seconds of a minidisk cache block,based on Little’s Law (cache size divided by pageout rate), and it displays the average age of pagingXSTORE blocks.

v The Cache Expanded Storage bar chart displays the ideal number of expanded storage blocks in theminidisk cache (based on average age of an XSTORE block, as determined by arbiter), and it displaysthe average age of paging XSTORE blocks.

v The Minidisk Cache Activity table provides such data as the size of the CP partition in expandedstorage, in blocks, the rate per second that blocks could not be moved into the minidisk cache becausetheir user’s fair share limit was exceeded, the maximum number of main storage page frames to beused for the minidisk cache, and the percentage of requests that were full hits, that is, where all therequested blocks were found in the cache.

See the “KVLMinidisk Cache attributes” on page 118 for more details on the data displayed in this view.

Figure 13 shows a sample Minidisk Cache workspace.

Figure 13. Minidisk Cache workspace


CCW Translations workspaceThe CCW Translations workspace displays the translated channel command word (CCW) activity for CCWchains directed at disks and network devices. Use this workspace to evaluate CCW translation activity.

Data Source:

The information shown is based on CP monitor SYSTEM domain SAMPLE data.

The CCW Translations predefined workspace contains the following views:

v The DASD CCW Translations bar chart displays, for DASD, the total number of CCWs handled persecond, the done rate for successfully translated CCWs per second, the number of CCWs per secondthat were found to be not eligible for translation, and the aborted translation attempts per second.

v The DASD and Network CCW Translations table displays data such as the done rate for successfullytranslated channel command words (CCWs) per second, for DASD, the total number of CCWs handledper second, the number of CCWs that were found to be not eligible for translation, per second fornetworks, and the done rate for networks.

See the “KVLMinidisk Cache attributes” on page 118 for more details on the data displayed in this view.

Figure 14 shows a sample CCW Translations workspace.

Control Unit Cache workspaceThe Control Unit Cache workspace provides data on the use and on the effectiveness of the cache. Usethis workspace to view load data for cached control units. Data displays for devices that have been activeduring the last measuring interval

Data Source:

The information shown is based on CP Monitor MONITOR and I/O domain SAMPLE data.

Figure 14. CCW Translations workspace



v The Top 5 Cache Control Unit Rate per Second bar chart displays, for the five busiest control units,by subsystem identifier of the control unit, the total I/O activity as it is recorded by the cache controlunit. That is, where multiple systems are connected to one control unit, the total I/O activity from allsystems is shown. This bar chart also displays the total I/O activity as determined from count fields inthe subchannel measurement blocks of the system that does the monitoring. That is, only the I/Orequests from that particular system are shown.

v The Top 5 Cache Control Unit Load Data Time bar chart displays, for the five busiest control units, bysubsystem identifier of the control unit, the device response time, in milliseconds. This is the deviceservice time, plus the time during which an I/O request was waiting to be started. It is calculated basedon the average I/O request queue length and the I/O rate.

v The Top 5 Cache Control Unit Load Busy Percent bar chart displays for the five busiest control units,by subsystem identifier of the control unit, the average busy percentage for all connected disks.

v The Top 5 Cache Control Unit Data Rate per Second bar chart displays, for the five busiest controlunits, by subsystem identifier of the control unit, the total nonsequential read rate per second, thesequential read rate per second, and the total fast-write rate per second.

v The Top 5 Available Cache Controller Memory bar chart displays, for the five busiest control units, bysubsystem identifier of the control unit, the size of configured cache storage, in megabytes, and theamount of cache storage that is available, in megabytes.

v The Cache Controller Utilization table displays data such as the LPAR identifier, the total I/O activityas it is recorded by the cache control unit, the subsystem identifier, the average disconnected time, inmilliseconds, and the sequential read rate per second.

From the Cache Controller Utilization table, you can link to the “DASD Cache workspace” on page 42.To access this workspace, click the link icon on the row of the control unit for which you would likeadditional data. The DASD Cache by Subsystem ID workspace will display with DASD cache datapertinent to the selected control unit.

See the “KVLControlUnit attributes” on page 104 for details on the data displayed in this view.

Figure 15 on page 42 shows a sample Control Unit workspace.


DASD Cache workspaceThe DASD Cache workspace provides data on device cache activity. Use this workspace to monitor theeffectiveness of the device cache and how it is being used.

Data Source:

The information shown is based on CP Monitor MONITOR and I/O domain SAMPLE data.


v The Top 5 Cache Rate per Second: Component Cache Rates bar chart displays, for the five busiestdevices, by volume serial number, the nonsequential read rate per second, the sequential read rate persecond, and the total fast-write rate per second.

v The Top 5 Percent Total Cache I/O for Staging/Destaging bar chart displays, for the five busiestdevices, by volume serial number, the percentage of nonsequential DASD to cache transfer operations,the percentage of sequential DASD to cache transfer operations, and the percentage of cache to DASDtransfer operations. Staging refers to the reading of data from the cache to the physical disk. Destagingrefers to the writing of data from the cache to the physical disk.

v The Top 5 Highest I/O Rate per Second bar chart displays for the five busiest devices, by volumeserial number, the number of normal I/O requests, per second, over this storage director, the number ofsequential I/O requests, per second, over this storage director, and the number of fast write I/O requestsper second, over this storage director.

v The Top 5 Cache Rate per Second bar chart displays, for the five busiest devices, by volume serialnumber, the total I/O rate per second, as indicated by subchannel measurement block data for thesystem that does the monitoring. That is, only the I/O activity for that system is included. This bar chartalso displays the total I/O rate for the disk as it is recorded by the cache control unit. That is, wheremultiple systems are connected to one control unit, the total I/O activity from all systems is shown.

Figure 15. Control Unit workspace


v The DASD Cache Activity table displays data such as the volume serial number of a device, thepercentage of DASD fast write hits, the percentage of DASD fast write operations that were forced tobypass the cache and access DASD directly, the percentage of read hits, the percentage of write hits,and the dual copy status.

From the DASD Cache Activity table, you can link to the “Control Unit Cache workspace” on page 40.To access this workspace, click the link icon on the row of the device for which you would like additionaldata. The Control Unit Cache by Controller ID workspace will display with control unit cache datapertinent to the selected device.

See the “KVLDASDCache attributes” on page 106 for details on the data displayed in this view.

Figure 16 shows a sample DASD Cache workspace.

VDISK workspaceThe VDISK workspace provides data on overall storage utilization by virtual disks.

Data Source:

The information shown is based on CP Monitor STORAGE domain SAMPLE data.


v The Top 5 Paging Rates per Second bar chart displays, for the five busiest virtual disks, by user ID ofthe owner of the VDISK and by device address, the number of pages read from DASD per second, thenumber of pages stolen per second, and the number of pages written to DASD per second.

v The Top 5 Expanded Storage Paging Rates per Second bar chart displays, for the five busiest virtualdisks, by user ID of the owner of the VDISK and by device address, the number of pages moved fromexpanded storage to central storage per second, the number of pages migrated from expanded storageto DASD per second, and the number of pages moved from central storage to expanded storage persecond.

Figure 16. DASD Cache workspace


v The Top 5 Pages in Use bar chart displays for the five busiest virtual disks, by user ID of the owner ofthe VDISK and by device address, the number of pages resident in central storage, the number oflocked data space pages, the number of slots occupied on auxiliary storage, and the number ofXSTORE blocks occupied by the data space.

v The Virtual Disk Activity table displays data such as the user ID of the owner of the VDISK, the nameof the logical partition, the virtual I/O rate, per second, to the VDISK, the size of the VDISK, the numberof pages written to DASD per second, and the number of pages read from DASD per second.

See the “KVLVdisk attributes” on page 143 for details on the data displayed in this view.

Figure 17 shows a sample VDISK workspace.

LPAR workspaceThe LPAR workspace provides information about the overall utilization of processor resources by thelogical partitions (LPARs) defined for your system. This workspace summarizes processor utilization bytype within each partition. You can use this workspace to determine when to move resources betweenlogical partitions.

Data Source:


The LPAR predefined workspace contains the following views:

v The LPAR Busy bar chart displays, for each assigned logical partition, the percentage of time that alllogical processors of a certain type were busy, divided by the number of processors of each type in thatpartition. Busy is defined as the percentage of elapsed time during which real processors were assignedto the logical processors. The LPAR busy utilization reported for a partition is the sum of the individualLPAR Busy Percent values for all of the logical processors of a given type that are defined to a logicalpartition, divided by the number of logical processors. LPAR Busy Percent will not exceed 100%.

Figure 17. VDISK workspace


This bar chart also displays the Physical CPU Busy value, which is the sum, as a percentage, of theload values for all active logical partitions with normal processors (non-IFL and non-ICF), including thegeneral LPAR management overhead, if available. Depending on the number of processors defined to alogical partition, the total percentage is calculated based on 100 percent multiplied by the number ofassigned processors.

v The LPAR Load bar chart displays data on the relative load of the partition on the whole processorcomplex, during the collection interval. This is expressed as the relation of the total time that logicalprocessors were dispatched for a partition, compared to the total amount of available processor time. Aload value displays only for those partitions with normal processors (non-IFL and non-ICF). Dependingon the number of processors defined to a logical partition, the total percentage is calculated based on100 percent multiplied by the number of assigned processors.

v The LPAR Weight bar chart provides the average weight of the logical processors defined for eachlogical partition. Weight values for dedicated processors range from 1 to 999. The string 'DED' displayswhen processors are dedicated.

v The Partition Suspension bar chart displays the average percentage of elapsed time that logicalprocessors were suspended. That is, they could not provide service to the guest system due to logicalpartition management time and contention for real processors. The "suspended" time is calculated asthe difference between the elapsed time and the sum of processor busy time and voluntary wait time forthe same processor, as seen by the z/VM system that is active in the partition.

v The LPAR Utilization table provides data on overall utilization of system processor resources for eachof the defined logical partitions. This table provides one row of data for each defined logical partition.The table shows data such as the total amount of time that all of the logical processors for a logicalpartition were busy during the collection interval, the percentage of elapsed time that a logical processorspent to manage LPARs, and the utilization of the system based on the number of logical processorsavailable. Use this table to determine how well each logical partition is able to obtain CPU resources.

The LPAR Utilization table contains one or more icons linking to the “Processor by LPAR Nameworkspace” on page 46. To access this workspace, click the link icon on the row of the logical partitionfor which you would like additional data and select Processor by LPAR Name.

The LPAR workspace provides a link to a secondary OMEGAMON XE on z/VM and Linux workspace.From the Navigator, select the LPAR workspace, select Workspace, then select “Processor workspace”on page 47.

See the “KVLLPAR Info attributes” on page 115 for more details on the data displayed in this view.

Figure 18 on page 46 shows a sample LPAR workspace.


Processor by LPAR Name workspaceThe Processor by LPAR Name workspace, accessed from the LPAR Utilization table of the “LPARworkspace” on page 44, displays data on the logical partitions for which you elected to view additionalinformation. Use this workspace to determine the resource usage by specific logical partitions and the loadfor each of the defined processors.

Data Source:


The Processor by LPAR Name predefined workspace contains the following views:


v The LPAR Load bar chart displays data on the relative load of the requesting partition. This isexpressed as the relation of the total time that logical processors were dispatched for this partition,compared to the total amount of available processor time.

v The LPAR Processor Busy bar chart displays, for the selected logical partition, the percentage of timethat the logical processor was busy, defined as the percentage of elapsed time during which a realprocessor was assigned to the logical processor. The value will also include wait time in the system thatis running in the logical partition if 'wait completion' has been enabled for the logical processor.

v The LPAR Processor Utilization table provides data on the system utilization by the selected logicalpartition. The table shows data such as the relative load of the partition on the whole processorcomplex. This is expressed as the relation of the total time that logical processors were dispatched forthis partition, compared to the total amount of available processor time.

This table also displays data such as the total logical CPU load calculated based on the sum of waittime plus processor busy time, and the average processor load as seen by the guest VM system,

Figure 18. LPAR workspace


calculated based on the sum of processor busy time plus elapsed time. Additionally, the table displaysidentifying information for the LPAR, such as its name, the CPU type of logical processor defined for theLPAR, and whether the LPAR is active.

See the “KVLProcessor Data attributes” on page 121 for more details on the data displayed in this view.

Figure 19 shows a sample Processor by LPAR Name workspace.

Processor workspaceThe Processor workspace, accessed from the “LPAR workspace” on page 44, displays data on systemutilization by the different types of processors assigned to the defined logical partitions. The default viewfor this workspace is to display data for processors with an LPAR Status of ACTIVE*.

Data Source:

The information shown is based on CP monitor SYSTEM and PROCESSOR domain SAMPLE data.



v The LPAR Load bar chart displays data on the relative load of the partition on the whole processorcomplex, during the collection interval. This is expressed as the relation of the total time that logicalprocessors were dispatched for a partition, compared to the total amount of available processor time. Aload value displays only for those partitions with normal processors (non-IFL and non-ICF). This barchart also displays the total logical CPU load calculated based on the sum of wait time plus processorbusy time, and it displays the average processor load as seen by the guest VM system, calculatedbased on the sum of processor busy time plus elapsed time. Additionally, it shows the status of theLPAR, that is, whether it is active.

Figure 19. Processor by LPAR Name workspace


v The LPAR Processor Busy bar chart displays, for each assigned logical partition, the percentage oftime that the logical processor was busy, defined as the percentage of elapsed time during which a realprocessor was assigned to the logical processor. The value will also include wait time in the system thatis running in the logical partition if 'wait completion' has been enabled for the logical processor. This barchart also provides data on the average percentage of elapsed time that the logical processor spent tomanage LPARs. This information is available only on systems with the LPAR management time facility.

v The LPAR Processor Utilization table provides information for each processor assigned to the definedlogical partitions. For each defined LPAR, there are one or more rows of processor data, depending onthe processors that have been configured for each LPAR. The table shows data such as the totalamount of time that all of the logical processors for a logical partition were busy during the collectioninterval, the percentage of elapsed time that a logical processor spent to manage LPARs, and theutilization of the system based on the number of logical processors available. Use this table todetermine how well each logical partition is able to obtain CPU resources.

See the “KVLProcessor Data attributes” on page 121 for more details on the data displayed in this view.

Figure 20 shows a sample Processor workspace.

Network workspaceThe Network workspace provides information about the utilization of the HiperSockets channels and thevirtual switch devices on the z/VM system. HiperSockets™ channels provide high-speed communicationbetween guest virtual machines. You can use this workspace to determine how well data are beingtransmitted across the network. This workspace also helps you monitor message traffic on logicalpartitions. If you define situations to monitor message traffic on LPARs, set the threshold for the messagecount to a value that makes sense for your environment and for the expected message traffic.

Data Source:

HiperSockets channel load analysis is based on data from the Extended Channel Path MeasurementFacility, available from the CP Monitor SYSTEM domain.

Figure 20. Processor workspace



v The Top 5 HiperSockets Message Rate per Second bar chart displays, for the five busiestHiperSockets channels, the rate of messages that were sent, per second, for the entire system, and therate of messages that were sent, per second, for specific logical partitions.

v The Top 5 HiperSockets Message Failure Rate per Sec bar chart displays, for the five busiestHiperSockets channels, data on the rate of sends that failed, per second, for the entire system, becausea receiver buffer was not available. This bar chart also shows the rate of sends that failed, per second,for an LPAR, because of the lack of a receiver buffer, or that failed for other problems that occurred.

v The Top 5 Virtual Switch Devices Receive Packets Rate per Second bar chart displays, for the fivebusiest virtual switches, data on the rate of virtual switch packets sent and received per second. Thischart also provides data on the rate of outbound packets discarded, per second, and data on the rate ofinbound packets discarded, per second.

v The HiperSockets Activity table provides data on overall HiperSockets activity, during the collectioninterval. The table shows data such as whether a channel is dedicated or whether it is being sharedwith other logical partitions, the rate of data units sent per second, the rate of messages sent persecond, and the channel path identifier. Use this table to determine how well each HiperSocketschannel is handling the flow of data.

v The Virtual Switch Activity table (located below the HiperSockets Activity table) contains data onvirtual switch utilization, during the collection interval. The table shows data such as the user ID of thevirtual machine to which a device is currently attached, the real device address, the timeout value forthe virtual switch in seconds, and the queue storage value. Use this table to determine how well thevirtual switch devices are performing.

See the “KVLHiperSocket attributes” on page 112 and the “KVLVirtualSwitch attributes” on page 145 formore details on the data displayed in this view.

The Network workspace provides a link to a secondary OMEGAMON XE on z/VM and Linux workspace.From the Navigator, select the Network workspace, select Workspace, then select the VSWITCHworkspace.

Figure 21 on page 50 shows a sample Network workspace.


VSWITCH workspaceThe VSWITCH workspace provides information about the activity of the virtual switch devices on the z/VMsystem, during the current collection interval. You can use this workspace to evaluate the performance ofyour network virtual switches.

Data Source:

The information shown for virtual switch activity is based on CP Monitor I/O domain SAMPLE data.


v The Top 5 Lock Requests Rate per Second bar chart displays, for the five busiest virtual switches, byreal device address, the number of lock requests made, per second, for the network lock.

v The Top 5 Lock Defers Rate per Second bar chart displays, for the five busiest virtual switches, byreal device address, data on the number of times per second that CP Monitor waited for the networklock. This bar chart also displays the number of times per second that CP Monitor waited for any lockwhen receiving data on the VSWITCH port, and the number of times per second that CP Monitor waitedfor any lock when sending data from the VSWITCH port.

v The Top 5 Transmit Packets Rate per Second bar chart displays data, for the five busiest virtualswitches, by real device address, on the rate of virtual switch packets sent per second. This chart alsoprovides data on the rate of outbound packets discarded, per second.

v The Top 5 Receive Packets Rate per Second bar chart displays data, for the five busiest virtualswitches, by real device address, on the rate of virtual switch packets received per second. This chartalso provides data on the rate of inbound packets discarded, per second.

v The Top 5 Signals Issued Rate per Second bar chart displays data, for the five busiest virtualswitches, by real device address, on the number of write signals per second. This chart also providesdata on the number of read signals issued per second, and on the number of sync signals issued persecond.

v The Virtual Switch Activity table (located below the middle bar charts) contains data on virtual switchutilization, during the collection interval. The table shows data such as the user ID of the virtual machineto which a device is currently attached, the real device address, the identifier for the virtual switch, thetimeout value for the virtual switch in seconds, and the queue storage value. Use this table to determinehow well the virtual switch devices are performing.

See “KVLVirtualSwitch attributes” on page 145 for more details on the data displayed in this view.

Figure 22 on page 51 shows a sample VSWITCH workspace.

Figure 21. Network workspace


Real Storage workspaceThe Real Storage workspace provides information about the factors affecting performance as it relates tothe usage of real storage on your z/VM system. You can use this workspace to determine, during thecollection interval, the use of the real storage, in terms of various types of frame counts and pagingmetrics.

Data Source:


The Real Storage predefined workspace contains the following views:

v The Storage Utilization bar chart displays the number of deferred pages waiting for a frame, thenumber of frames allocated to the dynamic paging area, as well as the number of frames used byfree-storage management, during the collection interval.

v The Available Frames Mean bar chart displays data on the average number of page frames availableon the below 2GB available list replenishment subsystem during the collection interval.

v The System Page Rate linear gauge chart shows the input/output paging rate for the entire system.

v The System Resource Utilization bar chart shows, for the collection interval, the average system-widepercentage of paging space in use, the percentage of spool space in use for the entire system, and thepercentage of temporary disk space in use for the entire system.

v The Page Wait Queue linear gauge chart displays the percentage of all virtual machines in a page waitstate.

v The z/VM Storage Utilization table displays storage utilization metrics for the z/VM system, such as thenumber of tasks that cannot be executed because they are waiting for a frame.

See the “KVLSystem attributes” on page 125 and the “KVLUser Workload attributes” on page 140 for moredetails on the data displayed in this view.

Figure 22. VSWITCH workspace


Figure 23 shows a sample Real Storage workspace.

System workspaceThe System workspace provides information about the factors affecting performance as it relates to theCPU usage of the virtual machine LPARs. CPU utilization is the percentage of time that all processorsavailable to a system were busy dispatching work.

If overall system CPU utilization is too high, you may want to review how competition for the CPU hasaffected the performance of a workload and you may want to review CPU use throughout the day.

Data Source:


The System predefined workspace contains the following views:

v The CPU Utilization panel contains two circular gauge charts. The CP Percent of CPU circular gaugechart displays the percentage of CPU used by the Control Program. The Percent of CPU circulargauge chart displays the total CPU utilization of the system by both the z/VM Control Program and thevirtual CPU. If you are running multiple processors, this value is the sum of the CPU utilization for allthe active processors and can be greater than 100%.

v The Current Users bar chart shows the average number of users who are logged on to the virtualmachine. It also displays the number of users who are dialed to the virtual machine, and it displays theaverage number of eligible users.

v The Total to Virtual Ratio circular gauge chart displays the ratio of total CPU time to virtual CPU time.

v The User Dispatching Queues bar chart shows the number of workloads on the eligible lists: E1, E2,and E3. This bar chart also provides data on the average number of users in the queue waiting to bedispatched.

Figure 23. Real Storage workspace


v The System Utilization table displays, during the collection interval, utilization metrics, such as theaverage number of users that are logged on, the percentage of CPU utilized by the z/VM ControlProgram, the number of users who are dialed to the z/VM system, and the percentage of all virtualmachines in an I/O wait state.

See the “KVLSystem attributes” on page 125 for more details on the data displayed in this view.

From the Navigator item for this workspace, you can connect to two secondary workspaces. To accessthese workspaces, from the Navigator, select the System workspace, select Workspace, then selecteither of the following secondary workspaces:

v “Spin Locks workspace”

v “System Terminal workspace” on page 54

Figure 24 shows a sample System workspace.

Spin Locks workspaceThe Spin Locks workspace provides detailed wait statistics on processor spin lock activity.

Data Source:

The information shown is based on CP Monitor SYSTEM domain SAMPLE data.

The predefined workspace contains the following views:

v The Top 5 Locks per Second bar chart displays, for the top five spin locks, by spin lock name, thetotal exclusive spin lock calls, per second, and the total shared spin lock calls, per second.

v The Top 5 Locks by Duration bar chart displays, for the top five spin locks, by spin lock name, theaverage total microseconds spent spinning on exclusive spin locks. The bar chart also displays the totalaverage number of microseconds spent spinning on shared spin locks.

v The Top 5 Percent of Elapsed Time in Spin Lock bar chart displays, for the top five spin locks, byspin lock name, the percentage of total elapsed time spent on exclusive spin locks, and the percentageof total elapsed time spent on shared spin locks.

v The Spin Lock Activity table displays wait statistics, for each CP Monitor spin lock, such as the totalexclusive and shared spin lock calls, per second, the percentage of total elapsed time spent on bothexclusive and shared spin locks, the total shared spin lock calls per second, the LPAR name, the z/VMsystem name, and the total exclusive spin lock calls per second.

See the “KVLSpinLock attributes” on page 124 for more details on the metrics displayed in this view.

Figure 25 on page 54 shows a sample Spin Locks workspace.

Figure 24. System workspace


System Terminal workspaceUse this workspace to log on to the z/VM system and to access the z/VM Performance Toolkit to furtherinvestigate a problem. Once in the z/VM Performance Toolkit, you can navigate to specific screens to viewmore detailed data and to investigate a problem.

You access the System Terminal workspace from the Navigator. Right-click the name of the Systemworkspace and select System Terminal.

The System Terminal workspace uses the terminal emulator adapter feature of Tivoli Enterprise Portal. Theterminal emulator adapter turns the view into a 3270, 5250, or Telnet interface so you can connect to anyTN3270, TN5250, or VT100 host system with TCP/IP.

For 3270 or 5250 terminal views, the terminal emulator adapter feature provides a scripting languageinterface that you can use to perform the following activities:

v You can record (capture) a host session. As you interact with a host session, the session is recorded asa set of script commands that can be saved under a name you specify and played back at a later time.This allows you to automate navigation to a specific set of screens.

v You can author complex scripts containing custom functions for manipulating host sessions. By default,the terminal connection remains active until you disconnect manually or until you end your TivoliEnterprise Portal work session. This means you can open other workspaces and return to this onewithout breaking the connection.

Click the Connection icon to provide information about the host system you are using, the port number,and the terminal type. This configuration information is used each time the workspace is opened andestablishes a connection to the specified host. You can save the terminal view in this workspace for futurework sessions. See the Tivoli Enterprise Portal online help for additional details about the terminalemulator feature.

Figure 25. Spin Locks workspace


Take Action viewThe System Terminal workspace includes a Take Action view that enables you to interact directly with yourapplications and operating system. You can use the Take Action view to send commands to the commandprocessor virtual machine, to select a predefined action, and to stop or to start a process at the system.The Take Action view can also be selected from the Tivoli Enterprise Portal toolbar. See Chapter 5, “TakeAction commands,” on page 89 for details on how the optional Take Action feature is implemented for thismonitoring agent. See the IBM Tivoli OMEGAMON XE on z/VM and Linux Planning and ConfigurationGuide for details on enabling the Take Action feature.

Figure 26 shows a sample System Terminal workspace.

TCPIP workspaceThe TCPIP workspace provides performance data about the Transmission Control Protocol/InternetProtocol (TCP/IP) server function that is part of the TCP/IP protocol suite running on the z/VM system.Use this workspace to monitor the general health and activity of the TCP/IP server.

Data Source:

The information shown is based on CP monitor APPLDATA domain SAMPLE data, provided by the TCP/IPserver machine.

The TCPIP predefined workspace contains the following views:

v The TCP/IP Server Bytes per Second bar chart displays, for each TCP/IP server, the number of bytesreceived per second, and the number of bytes transmitted per second, that are associated with inputand output requests.

v The TCP/IP Server Segments per Second bar chart displays, for each server, the rate of TransmissionControl Protocol (TCP) segments received per second, and the rate of TCP segments sent per second.

Figure 26. System Terminal workspace


It also provides data on the rate at which TCP segments were received that had errors, per second, aswell as the rate at which TCP segments were transmitted that included a reset, in seconds.

v The TCP/IP Server Connection Activity bar chart shows, for each server, the rate per second at whichTCP connection open requests were initiated, the rate per second at which TCP connection openrequests were accepted, the rate per second for TCP connection open failures, and the rate per secondfor TCP connections that were reset.

v The TCP/IP Server Activity table displays, for each server, during the collection interval, data such asthe number of bytes received per second, the rate per second at which TCP connection open requestswere initiated, and the rate per second at which TCP connection open requests were accepted.

From any row in the TCP/IP Server Activity table, you can link to the “TCPIP User workspace” thatprovides user data specific to the server associated with that row.

See the “KVLTCPIP Srvr Data attributes” on page 133 for more details on the data displayed in this view.

Figure 27 shows a sample TCPIP workspace.

TCPIP User workspaceThe TCPIP User workspace provides data about the main users of the TCP/IP function for the serverselected from one of the rows in the TCP/IP workspace. Use this workspace to monitor the activity of thevirtual machines that handle the local side of the connection.

Data Source:

The information shown is based on CP monitor APPLDATA domain EVENT data, provided in the form ofconnection OPEN/CLOSE records by the TCP/IP server machine.

The TCPIP User predefined workspace contains the following views:

Figure 27. TCPIP workspace


v The Byte Traffic bar chart displays data on the busiest virtual machines. For each virtual machine, dataare provided on the total number of bytes received per second during TCP sessions, and on the totalnumber of bytes transmitted per second during TCP sessions. It also provides data on the total numberof bytes received per second during User Datagram Protocol (UDP) sessions, and on the total numberof bytes transmitted per second during UDP sessions.

v The Total Segments per Second bar chart displays, for each virtual machine, the total number ofsegments received per second.

v The TCP/IP User Activity table that displays, for each server, during the collection interval, data suchas the total number of bytes received and sent during TCP sessions, the average elapsed time from thebeginning of a session to the end of a session, and the number of completed UDP sessions.

Note: Data does not display in this table for those TCP/IP users that have not completed any TCPsessions and that have not completed any UDP sessions during the collection interval.

See the “KVLTCPIPUsrData attributes” on page 135 for more details on the data displayed in this view.

Figure 28 shows a sample TCPIP User workspace.

Workload workspaceThe Workload workspace provides data about the system usage by user ID/workload for all users on thez/VM system.

Data Source:

The information shown is based on CP monitor USER domain records.

The Workload predefined workspace contains the following views:

v The Top 5 CPU Users bar chart displays the top five users of CPU. For each workload, it shows thepercentage of CPU used by the system to manage the workload, and the percentage of virtual CPUutilized.

v The Top 5 Page Rate bar chart displays the rate of page-ins and page-outs, per second, for the fiveworkloads with the greatest levels of activity.

v The Top 5 Paging Operations bar chart shows, for each workload, the number of page reads over thecollection interval, and the number of page writes over the collection interval.

v The Top 5 Working Set Size bar chart shows the top five workloads by number of pages in theirprojected working set. This value is calculated each time a workload drops from the queue. It is basedon the number of pages referenced during the last stay in the queue.

v The All z/VM Workloads table displays data such as the total number of CPU seconds used by eachworkload, the percentage of total CPU used by the z/VM Control Program to manage the workload, andthe total time each workload was logged on.

This workspace provides links to three other OMEGAMON XE on z/VM and Linux workspaces. From theNavigator, right-click the name of the Workload workspace, select Workspace, and select “LinuxWorkload workspace” on page 58, “ApplData workspace” on page 59, or “Resource Constraint workspace”on page 61.

See the “KVLTCPIPUsrData attributes” on page 135, “KVLUser Workload attributes” on page 140, and“KVLUser Wait attributes” on page 131 for more details on the data displayed in this view.

Figure 28. TCPIP User workspace


Figure 29 shows a sample Workload workspace.

Linux Workload workspaceThe Linux Workload workspace provides data about system usage by user ID for all Linux guest systemsdefined in your environment.

Data Source:

The information shown is based on CP monitor APPLDATA domain EVENT data, provided by the Linuxkernel.

The Linux Workload predefined workspace contains the following views:

v The Top 5 CPU Linux Guest Systems bar chart displays the top five Linux guest users of CPU. Foreach Linux guest, it shows the percentage of CPU used by the system to manage the workload, andthe percentage of virtual CPU utilized.

v The Top 5 Linux Guest System Page Rate bar chart displays the rate of page-ins and page-outs, persecond, for the five Linux guest systems with the greatest levels of activity.

v The Top 5 Linux Guest System Paging Operations bar chart shows, for each Linux guest system, thenumber of page reads over the collection interval, and the number of page writes over the collectioninterval.

v The Top 5 Linux Guest System Working Set Size bar chart shows the top five Linux guest systemsby number of pages in their projected working set. This value is calculated each time a workload dropsfrom the queue. It is based on the number of pages referenced during the last stay in the queue.

v The Linux Guest System Workloads table displays data such as the total number of CPU secondsused by each Linux guest system workload, the percentage of total CPU used by the z/VM ControlProgram to manage the Linux guest system workload, and the z/VM identifier for the Linux guestsystem.

The Linux Guest System Workloads table also contains dynamic workspace links to workspacesprovided by the Tivoli Monitoring Agent for Linux OS. Right-click the link icon from the Linux GuestSystem Workloads table to access these workspaces. To find out details about the linked workspaces,refer to the documentation for the Tivoli Monitoring Agent for Linux OS.

The following Tivoli Monitoring Agent for Linux OS workspaces are available from the Linux GuestSystem Workloads table:

– Linux Process

– Linux System Information

– Linux Virtual Memory Statistics

– Linux CPU Averages

– Linux Virtual Memory Usage Trends

Important: Predefined cross-product links to other product workspaces are not displayed if the otherproduct has not been installed and configured. If you have the Tivoli Monitoring Agent for Linux OSinstalled and you would like to link to workspaces in that monitoring agent, you must configure theKLZ_SETLPARVMID=Y environment variable in the Tivoli Monitoring Agent for Linux OS configurationfile (lz.ini). If you are running IBM Tivoli Monitoring, V6.2.0 and above, you make an additionalmodification to the lz.ini file by commenting out the CTIRA_HOSTNAME.

You also must install and enable the application support for this monitoring agent on the sharedtechnology components. See the IBM Tivoli Monitoring Installation and Setup Guide for details.

Figure 29. Workload workspace


Additionally, you must enable data collection for the KVLUser Appldata attribute group at each Linuxguest system that is to be monitored. Enabling data collection for this attribute group defines theworkgroup as a Linux workgroup. This is important because the links are displayed on a table thatqueries the systems in the Linux workgroup.

You must also enable the CP Monitor domain for this type of data on the z/VM system where thePerformance Toolkit is running. For details on enabling the collection of Linux data and on enablingdynamic workspace linking, see the IBM Tivoli OMEGAMON XE on z/VM and Linux Planning andConfiguration Guide. See also “Dynamic workspace linking” on page 18.

See the “KVLUser Workload attributes” on page 140 for more details on the data displayed in this view.

Figure 30 shows a sample Linux Workload workspace.

ApplData workspaceThe ApplData workspace provides, for each Linux guest system, extensive metrics about networkutilization, CPU usage, and operating system data.

Note: Data collection for the “KVLUser ApplData attributes” on page 136 from which some views of thisworkspace are built must be enabled at each Linux guest system that is to be monitored. In addition, youneed to enable data collection for the CP Monitor APPLDATA domain at the z/VM system where thePerformance Toolkit is running. For details on enabling data collection for this attribute group, see theplanning and configuration guide for this monitoring agent. For details on enabling the collection of Linuxdata, see the IBM Tivoli OMEGAMON XE on z/VM and Linux Planning and Configuration Guide.

Data Source:

The information shown is based on CP monitor APPLDATA domain EVENT data, provided by the Linuxkernel.

The ApplData predefined workspace contains the following views:

v The Linux Guest Workload Data table displays data such as the percentage of total CPU used by thez/VM Control Program to manage this workload, the average storage size for this workload, in kilobytes,the number of expanded storage blocks allocated to this workload by the Control Program for paging,and the projected working set size. See the “KVLUser Workload attributes” on page 140 for more detailson the data displayed in this table.

The Linux Guest Workload Data table also contains dynamic workspace links to workspaces providedby the Tivoli Monitoring Agent for Linux OS. Right-click the link icon from the Linux Guest WorkloadData table to access these workspaces. To find out details about the linked workspaces, refer to thedocumentation for the Tivoli Monitoring Agent for Linux OS.

The following Tivoli Monitoring Agent for Linux OS workspaces are available from the Linux GuestWorkload Data table:

– Linux Process


– Linux Virtual Memory Statistics


– Linux Virt. Memory Usage Trends

v The Linux Guest Appl Data table displays detailed data, such as the percentage of CPU used by thisLinux virtual machine running in kernel mode, the percentage of soft interrupts, the total size of the main

Figure 30. Linux Workload workspace


memory, and the size of the memory that is used for buffers. See the “KVLUser ApplData attributes” onpage 136 for more details on the data displayed in this view.

The Linux Guest Appl Data table contains dynamic workspace links to workspaces provided by theTivoli Monitoring Agent for Linux OS. Right-click the link icon from the Linux Guest Appl Data table toaccess these workspaces. To find out details about the linked workspaces, refer to the documentationfor the Tivoli Monitoring Agent for Linux OS.

The following Tivoli Monitoring Agent for Linux OS workspaces are available from the Linux Guest ApplData table:

– Linux Process


– Linux Virtual Memory

– Linux Disk IO Rate

– Linux Network

– Linux Sockets

– Linux Capacity Usage


– Linux Virtual Memory Trends

Important: Predefined cross-product links to other product workspaces are not displayed if the otherproduct has not been installed and configured. If you have the Tivoli Monitoring Agent for Linux OSinstalled and you would like to link to workspaces in that monitoring agent, you must configure theKLZ_SETLPARVMID=Y environment variable in the Tivoli Monitoring Agent for Linux OS configuration file(lz.ini). If you are running IBM Tivoli Monitoring, V6.2.0 and above, you make an additional modification tothe lz.ini file by commenting out the CTIRA_HOSTNAME.

You also must install and enable the application support for this monitoring agent on the sharedtechnology components. See the IBM Tivoli Monitoring Installation and Setup Guide for details.

Additionally, you must enable data collection for the “KVLUser ApplData attributes” on page 136 at eachLinux guest system that is to be monitored. Enabling data collection for this attribute group defines theworkgroup as a Linux workgroup. This is important because the links are displayed on a table that queriesthe systems in the Linux workgroup.

You must also enable the CP Monitor domain for the type of data you want to collect on the z/VM systemwhere the Performance Toolkit is running. For details on enabling the collection of Linux data and onenabling dynamic workspace linking, see the IBM Tivoli OMEGAMON XE on z/VM and Linux Planning andConfiguration Guide. See “Dynamic workspace linking” on page 18 for additional details.

Figure 31 shows a sample ApplData workspace.

Enabling the collection of KVLUser ApplDataYou must enable data collection for the “KVLUser ApplData attributes” on page 136 at each Linux guestsystem that is to be monitored. Instead of manually enabling the collection of data each time you use themonitoring agent, you might choose to collect the data automatically at startup time.

See the IBM Tivoli OMEGAMON XE on z/VM and Linux Planning and Configuration Guide for instructionson enabling the collection of this attribute group, either manually or automatically. See also the z/VM: CPCommands and Utilities for details on the MONITOR SAMPLE and on the MONITOR EVENT commandsand the operands associated with each command.

Figure 31. ApplData workspace


Resource Constraint workspaceThe Resource Constraint workspace provides information about what percentage of an interval a virtualmachine is prevented from running, due to various wait conditions. Wait percentage values never exceed100%. Often, all wait percentages for a majority of workloads are zero, including percent active. This waitpercentage indicates that the workload was idle during the interval.

Data Source:

The information shown is based on CP monitor USER domain records. You must have high frequencydata collection running.

The predefined workspace contains the following views:

v The Top 5 Workloads Waiting for Resources is sorted in descending order by the sum of the: CPUWait Percent, Loading Percent, Page Wait Percent, I/O Wait Percent, and Instruction Simulation WaitPercent. The top 5 workloads are shown.

v The Top 5 I/O Wait Percent bar chart is sorted in descending order by the percentage of activesamples where the user was in I/O wait state. The top 5 workloads are shown.

v The Top 5 CPU Wait Percent bar chart is sorted in descending order by the percentage of activesamples where the user was found waiting to run on the real CPU. The top 5 workloads are shown.

v The Top 5 Page Wait Percent bar chart is sorted in descending order by the percentage of activesamples where the user was found in page wait state. The top 5 workloads are shown.

v The All z/VM Workload Resource Constraint table displays user wait state data for the z/VM system,with user activity and wait state percentages shown.

See the “KVLUser Wait attributes” on page 131 for more details on the data displayed in this view.

Figure 32shows a sample Resource Constraint workspace.

Figure 32. Resource Constraint workspace


Chapter 4. Predefined situations

A situation is a logical expression involving one or more system conditions. Situations are used to monitorthe condition of systems in your network. You can manage situations from the Tivoli Enterprise Portal byusing the Situation editor.

OMEGAMON XE on z/VM and Linux provides a set of predefined situations that monitor the performanceand availability status of your z/VM and Linux on zSeries systems and resources. These situations checkfor specific conditions and can trigger Critical, Warning, or Informational situation event indicators (alsocalled alerts) in the Navigator. Some of the situations are set to start automatically. Other situations mustbe started manually. See “Summary of predefined situations” on page 66 for a list of the situations and tofind out whether or not a situation is autostarted.

When a situation triggers an alert, you can investigate the situation event by opening its workspace. If botha warning and a critical condition occur for the same workspace, the indicator always shows thehighest-level situation event (the critical one).

Using predefined situations can improve the speed with which you can begin using the OMEGAMON XEon z/VM and Linux. You can examine and, if necessary, change the conditions or values being monitoredby 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 thesesituations and save over them.

TipRather than editing a predefined situation, copy it with Create Another and then edit the copy. Thispractice prevents your edited situations from being overwritten during installation and configuration offuture versions of the product.

If you are working to resolve an issue related to an event, acknowledge the event. Acknowledgmentsenable operators responsible for handling events to communicate their ownership of the event and itsworking status. Acknowledging an event places a blue check mark next to the situation in the event list. Ifthe situation is still true when the acknowledgment expires or if you cancel the acknowledgement before itexpires, the indicator changes accordingly.

Right-click an event to view the list of actions and click Acknowledge.

Using the Situation EditorYou use the Situation Editor to examine the conditions or values being monitored and, if necessary,change them to ones better suited to your environment. You can also use the predefined situations asmodels for creating your own situations, using OMEGAMON XE on z/VM and Linux attributes. Before yournew situations can take effect, you must distribute them to the systems you want to monitor. Forinstructions, see the Tivoli Enterprise Portal online help or IBM Tivoli Monitoring Enterprise Portal User'sGuide, SC32-9409.

You can use the Situation Editor in the Tivoli Enterprise Portal to perform the following operations:

v Display a list of situations running on a specific managed system, the same types of managed systems,or on the enterprise.

v Create, edit, delete, or view a situation.

v Start and stop a situation.

v Associate a situation with the current Navigator item.


To open the Situation Editor, perform one of the following actions:

1. Click the Situation Editor button from the toolbar.

2. Right-click a Navigator entry and click Situations from the pop-up menu.

When you right-click a Navigator node and select Situations from the context menu, the Situation Editoropens with a list of the situations associated with the selected Navigator node. When you select theSituation Editor icon on the toolbar, the Situation Editor opens with a list of all OMEGAMON XE on z/VMand Linux situations.

The left frame of the Situation editor initially lists the situations associated with the Navigator item that youselected.

When you click a situation name or create a new situation, the right frame opens with the following tabs:

FormulaCondition being tested

DistributionList of managed systems (operating systems, subsystems, or applications) to which the situationcan be distributed.

Expert AdviceComments and instructions to be read in the event workspace

ActionCommand to be sent to the system when the situation becomes true

Note: For this monitoring agent, for the Where should the action be run (Performed) option,you must select Run the action at the Managed System (Agent). Additionally, when you selectSystem Command on this tab, this field displays for you to type a command to issue at thesystem. Be sure to prefix all OMEGAMON XE on z/VM and Linux commands with VL:.

Until Duration of the situation

The IBM Tivoli Monitoring Enterprise Portal User's Guide contains more information about predefined andcustom situations and how to use them to respond to alerts. For a list of the predefined situations for thismonitoring agent and a description of each situation, refer to “Predefined situations and formulas” on page68 and the information in that section for each individual situation.

See also the Tivoli Enterprise Portal online help for more detailed information about using the SituationEditor.

Investigating a situation eventWhen the conditions of a situation have been met, the situation evaluates to True, causing a situationevent indicator to be displayed in the Navigator. You can investigate the cause of a situation event byopening its workspace.

The situation event workspace shows the following two table views:

v One table contains the values of the attributes when the situation evaluated True.

v The other table contains the current values of the attributes.

The event workspace also displays any expert advice written by the author of the situation, as well as aTake Action view that enables you to send a command to the z/VM system.


TipYou can configure a Conversational Monitor System (CMS) guest system to receive commandrequests from the Tivoli Enterprise Portal through this monitoring agent. These Take Actioncommands can execute CP commands, CMS commands, and REXX executables on the z/VMoperating system. If you want to enable these types of Take Action commands, see the IBM TivoliOMEGAMON XE on z/VM and Linux Planning and Configuration Guide.

See the Tivoli Enterprise Portal online help or the IBM Tivoli Monitoring Enterprise Portal User's Guide fordetailed information on using situations.

Situation formulasSituations are expressions embedded in IF-TRUE statements of system conditions that you want tomonitor. This means that if the specified condition exists, then the situation is true and triggers an alert.

A condition consists of an attribute, a value, and a comparison operator. When a situation is activated, thevalue of the attribute is compared with the value set for the condition to determine whether the condition ismet. For example, the ZVM_PerfKit_Collector_Inactive situation is true when the value of theKVLPTKStat.Collector Status attribute equals Inactive for the Performance Toolkit component.

The following shows the formula for the ZVM_PerfKit_Collector_Inactive situation:IfKVLPTKStat.Collector Name equals Performance Toolkit Collector andKVLPTKStat.Collector Status equals Inactive thenthe situation ZVM_Perfkit_Collector_Inactive is true.

For information about the attributes you can use in situations, see “Attributes reference” on page 101.

Avoid using negative valuesIf you define situations that use a counter or a range of numbers, always provide positive values. Forexample, use a greater-than-or-equal-to-zero expression, as shown in some of the predefined situationsdescribed below. This practice prevents a situation from falsely tripping when the monitoring agentencounters an undefined attribute value. Undefined attribute values are interpreted as negative numbersand might erroneously raise alerts for a situation that specifies a negative number.

Predefined situations provided by OMEGAMON XE on z/VM and LinuxThe predefined situations packaged with the OMEGAMON XE on z/VM and Linux product are distributedwhen the Tivoli Enterprise Monitoring Server is seeded (that is, initialized with application data).

Note: Some of the predefined situations are started automatically. This means that they are set to run onstartup of the Tivoli Enterprise Monitoring Server. Some of the situations are not set to start automatically.You must start these situations manually. The table that follows indicates the situations that are set to runautomatically at startup and those that must be manually set to run at startup.

The rest of this chapter describes the predefined situations and provides their formulas.

Chapter 4. Predefined situations 65

Summary of predefined situationsThe following table contains a summary of the situations provided by this monitoring agent:

Table 4. Summary of predefined situations

NavigatorItem Name of Situation

Column Name and InitialConditional Value State

Runs atstartup(Yes or No)

Defaultworkspace

“ZVM_PerfKit_Collector_Inactive” on page 75 Collector Name =Performance ToolkitCollector AND CollectorStatus = Inactive

Critical Yes

DASD “ZVM_CUCache_DataResp_High” on page69

Average Response Timemsec > 50

High(Warning)

No

DASD “ZVM_DASD_Cache_Hits_Low” on page 70 Cache_Total_Percent > 0AND Cache_Total_Percent< 30 AND Caching Status =ACTIVATED

Low(Warning)

No

DASD “ZVM_DASD_Queue_Critical” on page 71 Average Queued IO >=25.00

Critical No

DASD “ZVM_DASD_Queue_High” on page 71 Average Queued IO >=10.00 AND Average QueuedIO < 25.00

High(Warning)

No

LPAR “ZVM_LPAR_Busy_Critical” on page 71 LPAR Busy Percent >= 90AND LPAR Weight < 65535

Critical No

LPAR “ZVM_LPAR_Busy_High” on page 72 LPAR Busy Percent >= 80AND LPAR Busy Percent <90 and LPAR Weight <65535

High(Warning)

No

LPAR “ZVM_LPAR_Ovhd_Critical” on page 72 LPAR Overhead Percent >=40

Critical No

LPAR “ZVM_LPAR_Ovhd_High” on page 73 LPAR Overhead Percent >=15 AND LPAR OverheadPercent < 40

High(Warning)

No

LPAR “ZVM_Physical_CPU_Critical” on page 76 Physical CPU Busy >=90.00 AND LPAR Status =Active

Critical No

LPAR “ZVM_Physical_CPU_High” on page 77 Physical CPU Busy >=80.00 AND Physical CPUBusy < 90.00 and LPARStatus = Active

High(Warning)

No

RealStorage

“ZVM_Page_Used_Critical” on page 74 Pct Page Space In Use >=50

Critical No

RealStorage

“ZVM_Page_Used_High” on page 74 Pct Page Space In Use >=40 AND Pct Page Space InUse < 50

High(Warning)

No

System “ZVM_Avail_Mean2G_Low” on page 68 Available Frames Mean >2GB <= Avail Low Thresh2G

High(Warning)

No

System “ZVM_Avail_Mean_Low” on page 68 Available Frames Mean <=Avail Low Thresh

High(Warning)

No

System “ZVM_CP_CPU_Critical” on page 69 CP Percent of CPU >= 30 Critical No


Table 4. Summary of predefined situations (continued)




System “ZVM_CP_CPU_High” on page 69 CP Percent of CPU >= 20AND CP Percent of CPU <30

High(Warning)

No

System “ZVM_Page_Queue_High” on page 73 Page Wait Queue > 25 High(Warning)

No

System “ZVM_Spin_Exclusive_Pct_Critical” on page77

Exclusive Time Spinning onLocks Percent > 30

Critical No

System “ZVM_Spin_Exclusive_Pct_High” on page 77 Exclusive Time Spinning onLocks Percent > 25 ANDExclusive Time Spinning onLocks Percent <= 30

High(Warning)

No

System “ZVM_Spool_Used_Critical” on page 78 Pct Spool Space In Use >=95

Critical No

System “ZVM_Spool_Used_High” on page 78 Pct Spool Space In Use >=80 AND Pct Spool Space InUse < 95

High(Warning)

No

System “ZVM_Total_CPU_Critical” on page 79 Percent of CPU >= 90 Critical No

System “ZVM_Total_CPU_High” on page 79 Percent of CPU >= 80 ANDPercent of CPU < 90

High(Warning)

No

System “ZVM_Total_to_Virtual_High” on page 80 Total to Virtual Ratio >=2.00

High(Warning)

No

VDISK “ZVM_VDISK_Page_High_IORate_Low” onpage 82

SIT(“ZVM_Page_Queue_High”on page 73) = TRUE ANDVirtual I/Os per Second <10.0

Low(Warning)

No

Workload “ZVM_User_CPU_Critical” on page 80 CPU Percent >= 90.00 Critical No

Workload “ZVM_User_CPU_High” on page 80 CPU Percent >= 80.00 andCPU Percent < 90.00

High(Warning)

No

Workload “ZVM_User_Scaled_CPU_Critical” on page81

CPU Scaled Percent >=90.00

Critical No

Workload “ZVM_User_Scaled_CPU_High” on page 81 CPU Scaled Percent >=80.00 and CPU ScaledPercent < 90.00

High(Warning)

No

Workload “ZVM_Virtual_CPU_Critical” on page 82 Virtual CPU % >= 90.00 Critical No

Workload “ZVM_Virtual_CPU_High” on page 83 Virtual CPU % >= 80.00and Virtual CPU % < 90.00

High(Warning)

No

Workload “ZVM_Virtual_Scaled_CPU_Critical” on page83

Virtual CPU Scaled Percent>= 90.00

Critical No

Workload “ZVM_Virtual_Scaled_CPU_High” on page83

Virtual CPU Scaled Percent>= 80.00 and Virtual CPUScaled Percent < 90.00

High(Warning)

No

ResourceConstraint

“ZVM_User_Wait_Page_Critical” on page 84 Page_Wait_Percent > 40 Critical No

ResourceConstraint

“ZVM_User_Wait_Page_High” on page 84 Page_Wait_Percent > 25AND Page_Wait_Percent<= 40

High(Warning)

No


Table 4. Summary of predefined situations (continued)




ResourceConstraint

“ZVM_User_Wait_CPU_Critical” on page 85 CPU_Wait_Percent > 40 Critical No

ResourceConstraint

“ZVM_User_Wait_CPU_High” on page 85 CPU_Wait_Percent > 25AND CPU_Wait_Percent <=40

High(Warning)

No

ResourceConstraint

“ZVM_User_Wait_CPU_High” on page 85 CPU_Wait_Percent > 25AND CPU_Wait_Percent <=40

High(Warning)

No

z/VM LinuxSystems

“ZVM_Storage_Overcommit_Critical” onpage 86

KVLSystem2.Real_Storage_Overcommit > 4.0

Critical No

z/VM LinuxSystems

“ZVM_Storage_Overcommit_High” on page87

KVLSystem2.Real_Storage_Overcommit > 3.0 andKVLSystem2.Real_Storage_Overcommit <= 4.0

High(Warning)

No

z/VM LinuxSystems

“ZVM_Eligible_List_High” on page 87 KVLSystem.Eligible_Users> 5

High(Warning)

No

Predefined situations and formulasThis section provides the names, descriptions, logic, and threshold values for the predefined situations ofthis monitoring agent.

ZVM_Avail_Mean2G_LowRaises an alert when the average number of page frames on the above 2GB available list is less than orequal to the low threshold for frames on the above 2GB available list replenishment subsystem. When thisoccurs, CP begins a replenishment scan to add pages to the available list. The replenishment scan maytake pages from active workloads to replenish the available list. This situation must be started manually.

The triggering of this situation indicates that there are not enough free page frames to satisfy pagerequests for active workloads. This could result in greater page wait and possible thrashing of the pagingsubsystem. Consider adding real memory to the system. You can also reduce the paging demand byremoving workload where possible.

The values used to trigger this situation can be found on the “Real Storage workspace” on page 51available with this monitoring agent. See also the Available List Log (AVAILLOG) screen of thePerformance Toolkit. The AVAILLOG screen shows additional information about the health of the pagingsubsystem.

The following shows the formula for the ZVM_Avail_Mean2G_Low situation:

*IF *VALUE KVLSystem.Available Frames Mean > 2GB *LE KVLSystem.Avail Low Thresh 2G

ZVM_Avail_Mean_LowRaises an alert when the average number of page frames on the below 2GB available list replenishmentsubsystem is less than or equal to the low threshold for page frames on the below 2GB available listreplenishment subsystem. When this occurs, CP begins a replenishment scan to add pages to theavailable list. The replenishment scan may take pages from active workloads to replenish the available list.This situation must be started manually.


The triggering of this situation indicates that there are not enough free page frames to satisfy pagerequests for active workloads. This could result in greater page wait and possible thrashing of the pagingsubsystem. Consider adding real memory to the system. You can also reduce the paging demand byremoving workload where possible.

The values used to trigger this situation can be found on the “Real Storage workspace” on page 51available with this monitoring agent. See also the Available List Log (AVAILLOG) screen of thePerformance Toolkit. The AVAILLOG screen shows additional information about the health of the pagingsubsystem.

The following shows the formula for the ZVM_Avail_Mean_Low situation:

*IF *VALUE KVLSystem.Available Frames Mean *LE KVLSystem.Avail Low Thresh

ZVM_CP_CPU_CriticalRaises an alert when the percentage of CPU used by the z/VM Control Program is greater than or equalto 30 percent. The percentage includes both supervisor time spent for users and supervisor time spent forsystem services. The value displayed is the average of all the %CPU values reported for all processorsassigned to the logical partition. This situation must be started manually.

Use the “Workload workspace” on page 57 to identify those users that are using functions causing CP toprocess supervisor calls. You also might want to check the load on the paging subsystem. CP may bebusy processing a high number of page faults caused by guest operating systems. Consider modifying thethreshold for this situation if high CP CPU utilization is normal in your environment. Additional informationcan be found on the General CPU screen of the Performance Toolkit. The General CPU screen reportsthe %CPU utilization by individual processor.

The following shows the formula for the ZVM_CP_CPU_Critical situation:

*IF *VALUE KVLSystem.CP Percent of CPU *GE 30

ZVM_CP_CPU_HighRaises an alert when the percentage of CPU used by the z/VM Control Program is greater than or equalto 20 percent and less than 30 percent. This includes both supervisor time spent for users and supervisortime spent for system services. The value displayed is the average of all the %CPU values reported for allprocessors assigned to the logical partition. This situation must be started manually.

Use the “Workload workspace” on page 57 to identify those users that are using functions causing CP toprocess, for example, supervisor calls. You also might want to check the load on the paging subsystem.CP may be busy processing a high number of page faults caused by guest operating systems. Considermodifying the threshold for this situation if high CP CPU utilization is normal in your environment.Additional information can be found on the General CPU screen of the Performance Toolkit. The GeneralCPU screen reports the %CPU utilization by individual processor.

The following shows the formula for the ZVM_CP_CPU_High situation:

*IF *VALUE KVLSystem.CP Percent of CPU *GE 20 *AND *VALUE KVLSystem.CP Percent of CPU *LT 30

ZVM_CUCache_DataResp_HighRaises an alert when the response time for completing an I/O operation to a device on a cached controlunit exceeds the specified number of milliseconds (ms). The response time is the sum of function pending,of disconnected time, and of connected time (average device service time), plus the time during which anI/O request was waiting to be started. The value is calculated based on the average I/O request queuelength and on the I/O rate. This situation must be started manually.


A long response time indicates that a device is experiencing prolonged delays due to high activity.Consider rebalancing the I/O load across lightly loaded devices, if possible. Long response times may alsobe caused by excessive non-zero seeking.

Sharing DASD devices between separate systems can also cause device contention and queuing. Whensearching for a cause with control units shared across multiple systems, combined utilization must beconsidered. Disk performance can often be improved by minimizing I/O rates.

z/VM offers various components to help avoid high I/O, such as the following:

v Minidisk cache (MDC)

v Virtual disk in storage (VDISK)

v Discontiguous saved segment (DCSS)

MDC is a write-through cache that may benefit read I/O. MDC is supported by real and expanded storageon the processor. VDISK can be useful in Linux environments as a swap device. Both MDC and VDISKrequire sufficient memory to be used. A DCSS is a method of sharing memory across guests. In a Linuxenvironment, the block device driver can be configured to read file systems from a DCSS.

If long response times are encountered on devices that are used for paging, system performance may besignificantly degraded. Consider making additional memory available to z/VM Control Program (CP) toreduce paging. Spread CP-owned allocations across as many devices as possible and use dedicatedDASD devices.

See the “CP Owned Devices workspace” on page 33 for additional information on the location andutilization of CP paging space. See also the CP Owned Device (DEVICE CPOWNED) and theCPOWNED Device Log (CPOWNLOG devno) of the Performance Toolkit for additional information onpaging space activity. Refer to the Performance Toolkit Reference for details.

The following shows the formula for the ZVM_CUCache_DataResp_High situation:

*IF *VALUE KVLCTLUnit.Average Response Time msec *GT 50.0

ZVM_DASD_Cache_Hits_LowRaises an alert when the total percentage of cache utilization for a device, as measured by the CacheTotal Percent attribute, is within the range of the values specified. To reduce false alerts from devices thatare not active, only devices with a caching status of ACTIVATED are checked. The total percentage ofcache utilization is a measurement of the efficiency of the cache. The triggering of this situation indicatesthat the hardware caching capability available on the device is not being used to its full capacity. A lowvalue for Cache Total Percent indicates that records from I/O operations are not found in the cache (alsoknown as a cache miss), and require a disk I/O to complete the operation. Cache Total Percent valuesdisplay in the “DASD Cache workspace” on page 42 that is accessed from the “DASD workspace” onpage 37. The value displayed for each device is a percentage in the range of 0 - 100. This situation mustbe started manually.

In certain cases, it may be normal for the Cache Total Percent value to be very low. When I/O to a cachedevice is mostly sequential, such as during a disk backup, a restore, or a copy operation, the resultingvalue can be low. In these instances, the cache provides little benefit, as its contents are not being used tosatisfy most I/O requests. See the Performance Toolkit Reference for more information on the CACHEXTcommand and on the DEVICE command.

The following shows the formula for the ZVM_DASD_Cache_Hits_Low situation:

*IF *VALUE KVLDASDCache.Cache Total Percent *GT 0 AND *VALUE KVLDASDCache.Cache Total Percent*LT 30 AND *VALUE KVLDASDCache.Caching_Status *EQ ACTIVATED


ZVM_DASD_Queue_CriticalRaises an alert when the average number of I/O requests queued on a device is greater than or equal to25 requests, during the collection interval. The I/O queue length is not meaningful for devices that arededicated (attached) to guest systems. Guest system queuing cannot be shown by host performancemonitors. Refer to the guest systems performance monitor for queuing information on dedicated devices. Asimilar restriction applies to CP-owned devices with page or spool areas. Multiple pending I/O requests forpage or SPOOL are queued in a different place and are not included in this attribute. See the DASD I/OActivity Table in the “DASD workspace” on page 37 to view device data. See also the Queue Lngth fieldof the CP Owned Device screen in the Performance Toolkit. This situation must be started manually.

If a device has excessive I/O queuing, consider rebalancing the I/O load across lightly loaded devices, ifpossible. Device queuing may also be caused by excessive non-zero seeking. Sharing DASD devicesbetween separate systems can also cause device contention and queuing. Refer to the DASD SeeksDistances (SEEKDIST) screen and to the DASD Seeks Locations (SEEKLOC) screen in thePerformance Toolkit for help with analyzing seek operations by DASD device.

The following shows the formula for the ZVM_DASD_Queue_Critical situation:

*IF *VALUE KVLDevice.Average Queued IO *GE 25.00

ZVM_DASD_Queue_HighRaises an alert when the average number of I/O requests queued on a device is greater than or equal to10 requests and less than 25 requests, during the collection interval. Note that the I/O queue length is notmeaningful for devices that are dedicated (attached) to guest systems. Guest system queuing cannot beshown by host performance monitors. Refer to the guest systems performance monitor for queuinginformation on dedicated devices. A similar restriction applies to CP-owned devices with page or SPOOLareas. Multiple pending I/O requests for page or SPOOL are queued in a different place and are notincluded in this attribute. See the DASD I/O Activity Table in the “DASD workspace” on page 37 to viewdevice data. See also the Queue Lngth field of the CP Owned Device screen in the Performance Toolkit.This situation must be started manually.

If a device has excessive I/O queuing, consider rebalancing the I/O load across lightly loaded devices, ifpossible. Device queuing may also be caused by excessive non-zero seeking. Sharing DASD devicesbetween separate systems can also cause device contention and queuing. Refer to the SEEKDIST andSEEKLOC screens in the Performance Toolkit for help with analyzing seek operations by DASD device.

The following shows the formula for the ZVM_DASD_Queue_High situation:

*IF *VALUE KVLDevice.Average Queued IO *GE 10.00 *AND *VALUE KVLDevice.Average Queued IO *LT25.00

ZVM_LPAR_Busy_CriticalRaises an alert when the CPU utilization for a logical partition exceeds the specified percentage. Thisindicates that all of the logical processors of a specific type were found to be busy for the reportedpercentage of the elapsed time interval. Busy is defined as the portion of elapsed time during which realprocessors were assigned to logical processors. The LPAR busy utilization reported for a partition is thesum of the individual LPAR Busy Percent percentage values for all of the logical processors of a giventype that are defined to a logical partition, divided by the number of logical processors. The LPAR BusyPercent value will not exceed 100%. Customize this situation to match your processor configuration. Thissituation is not set to start automatically at startup of the monitoring agent.

Use the “LPAR workspace” on page 44 to find out about the overall utilization of resources by the logicalpartitions defined for your system. In the LPAR workspace tabular views, the LPAR Weight columncontains the string "DED" for partitions with dedicated processors. When you create a situation predicateor a threshold that tests the LPAR Weight attribute, use the numeric value 65535 instead of the string


"DED". The LPAR Weight value of 65535 tests true for partitions with dedicated processors. From the“LPAR workspace” on page 44, you can navigate to the “Processor by LPAR Name workspace” on page46 to view utilization data by individual processor.

The triggering of this situation indicates that there may be a problem with a workload in a partitionconsuming excessive processor resources. This problem could be caused by a looping virtual machine.This problem could also be caused by the normal contention for processor resources due to increasedworkload. In the latter case, consider modifying the situation threshold value to align with yourenvironment.

Use the “Workload workspace” on page 57 to identify the workloads that are using a large percentage ofCPU. If the source is found to be a guest operating system, use the performance tools for that system todetermine the cause.

The following shows the formula for the ZVM_LPAR_Busy_Critical situation:

*IF *VALUE KVLLPAR_Info.LPAR Busy Percent *GE 90.00 *AND *VALUE KVLLPAR_Info.LPAR_Weight *LT65535

ZVM_LPAR_Busy_HighRaises an alert when the CPU utilization for a logical partition is within the range specified in the situationformula. This indicates that all of the logical processors of a specific type were found to be busy for thereported percentage of the elapsed time interval. Busy is defined as the portion of elapsed time duringwhich real processors were assigned to logical processors. The LPAR busy utilization reported for apartition is the sum of the individual LPAR Busy Percent values for all of the logical processors of a giventype that are defined to a logical partition, divided by the number of logical processors. The LPAR BusyPercent value will not exceed 100%. Customize this situation to match your processor configuration. Thissituation must be started manually.

Use the “LPAR workspace” on page 44 to find out about the overall utilization of resources by the logicalpartitions defined for your system. In the LPAR workspace tabular views, the LPAR Weight columncontains the string "DED" for partitions with dedicated processors. When you create a situation predicateor a threshold that tests the LPAR Weight attribute, use the numeric value 65535 instead of the string"DED". The LPAR Weight value of 65535 tests true for partitions with dedicated processors. From the“LPAR workspace” on page 44, you can navigate to the “Processor by LPAR Name workspace” on page46 to view utilization data by individual processor.

The triggering of this situation indicates that there may be a problem with a workload in this partitionconsuming excessive processor resources. This problem could be caused by a looping virtual machine.This problem could also be caused by the normal contention for processor resources due to increasedworkload. In the latter case, consider modifying the situation threshold value to align with yourenvironment.

Use the “Workload workspace” on page 57 to identify the workloads that are using a large percentage ofCPU. If the source is found to be a guest operating system, use the performance tools for that system todetermine the cause.

The following shows the formula for the ZVM_LPAR_Busy_High situation:

*IF *VALUE KVLLPAR_Info.LPAR Busy Percent *GE 80.00 *AND *VALUE KVLLPAR_Info.LPAR Busy Percent*LT 90.00 *AND *VALUE KVLLPAR_Info.LPAR_Weight *LT 65535

ZVM_LPAR_Ovhd_CriticalRaises an alert when the percentage of elapsed time that a logical processor spent to manage a logicalpartition exceeds the specified percentage. This indicates that the LPAR scheduler consumed the specifiedpercentage of processor resources in support of a logical partition. The displayed value is the sum of the


overhead percentage values reported for all the logical processors of a given type that are defined to thelogical partition. This situation must be started manually.

Use the “LPAR workspace” on page 44 to find out about the overall utilization of resources by the logicalpartitions defined for your system. From the “LPAR workspace” on page 44, you can navigate to the“Processor by LPAR Name workspace” on page 46 to view utilization data by individual processor. Seealso the %ovhd values on the LPAR Load screen of the Performance Toolkit for additional data. Refer tothe Performance Toolkit Reference for details on the LPAR Load screen.

Important: During the time that a real processor is performing LPAR management, it cannot service theguest system.

The following shows the formula for the ZVM_LPAR_Ovhd_Critical situation:

*IF *VALUE KVLLPAR_Info.LPAR Overhead Percent *GE 40.00

ZVM_LPAR_Ovhd_HighRaises an alert when the percentage of elapsed time that a logical processor spent to manage a logicalpartition is within the range specified in the situation formula. This indicates that the LPAR schedulerconsumed the specified percentage of processor resources in support of a logical partition. The reportedvalue is the sum of the overhead percentage values reported for all the logical processors of a given typethat are defined to the logical partition. This situation must be started manually.

Use the “LPAR workspace” on page 44 to find out about the overall utilization of resources by the logicalpartitions defined for your system. From the “LPAR workspace” on page 44, you can navigate to the“Processor by LPAR Name workspace” on page 46 to view utilization data by individual processor. Seealso the %ovhd values on the LPAR Load screen of the Performance Toolkit for additional data. Refer tothe Performance Toolkit Reference for details on the LPAR Load screen.

Important: During the time that a real processor is performing LPAR management, it cannot service theguest system.

The following shows the formula for the ZVM_LPAR_Ovhd_High situation:

*IF *VALUE KVLLPAR_Info.LPAR Overhead Percent *GE 15.00 *AND *VALUE KVLLPAR_Info.LPAR OverheadPercent *LT 40.00

ZVM_Page_Queue_HighRaises an alert when the percentage of in-queue users in the page wait state exceeds the specified value.

In-queue virtual machines are those found, at the end of the sample interval, in the scheduler dispatch list.They can also be virtual machines that are waiting to be added to the scheduler dispatch list. The virtualmachine where the Performance Toolkit is collecting data is not included in this count. A virtual machinethat is waiting for a page cannot perform work until the page request is satisfied. If numerous virtualmachines are waiting for pages, system performance may be adversely affected.

Examine storage utilization by checking the working set sizes of the logged-on virtual machines. Refer tothe “Workload workspace” on page 57 for metrics on resident pages in use and on the working set sizeallocated to each virtual machine. Overcommitted virtual memory may be causing excessively high pagingrates. Consider adding real memory, or reducing the working set size of logged-on virtual machines. Youcan reduce the working set size by establishing lower virtual storage limits in the z/VM Control Program(CP) directory. Log off any unnecessary virtual machines with large working sets to reclaim the memorythat they are using.

If you used the SET RESERVED command of CP to reserve pages for preferred virtual machines,consider reducing the amount reserved. SET RESERVED can be useful in a paging environment to


ensure that virtual machines on which others are dependent do not wait for paging. For example, if avirtual machine acting as a file server for many other virtual machines must wait for paging, then all thedependent virtual machines must also wait. By minimizing page wait by the file server, one shortens thetime the other virtual machines need to be active and to have their pages active. Be aware that abusingthe SET RESERVED command can create memory constraints and excessive paging.

Virtual disks in storage (VDISKs) reside in the dynamic paging area. See the “VDISK workspace” on page43 to determine which virtual machines are using the VDISK feature. Look for VDISKs that are not beingused to full capacity, and that can be removed from the system. For Linux on System z virtual machines,explore sharing memory between machines by using a named saved system (NSS) or a discontiguoussaved segment (DCSS). See the CP Planning and Administration guide for information on using a namedsaved system. See the CP Command and Utilities Reference for information on using a DCSS.

Note: Extensive use of VDISKs can result in increased paging subsystem load. Therefore, do not useVDISKs in memory-constrained or high paging rate environments.

Refer to the Paging Log (PAGELOG), the System Counters (SYSTRANS), the CPU Load andTransactions (CPU), the Storage Utilization Log (STORLOG), and the User Page Data (UPAGE)screens of the Performance Toolkit for additional information. Refer to the Performance Toolkit Referencefor details on the Performance Toolkit screens.

The following shows the formula for the ZVM_Page_Queue_High situation:

*IF *VALUE KVLSystem.Page Wait Queue *GT 25

ZVM_Page_Used_CriticalRaises an alert when the percentage of any one of the allocations listed below is greater than or equal to50 percent:

v Allocated paging space

v Allocated spooling space

v Allocated dump space

If the situation triggers because of a spool allocation, spool files or overflow paging may be consuming alarge percentage of the available space. For detailed information, refer to the SPOOL File Summary(SPOOL) screen of the Performance Toolkit. The SPOOL screen shows how many disk blocks areconsumed by each file. Use this as a guide when determining which spool files would yield the most freespace if deleted. You might want to consider adding more space if spool utilization is generally high. Youcould also offload spool using the SPXTAPE command with the PURGE option if the files need to beretained. This situation is not set to run automatically at startup of the monitoring agent.

If the situation raises for a paging allocation, available paging space may be dangerously low. Consideradding additional paging space or reducing the active workload on the system by logging off guestsystems that are not critical. Refer to the attribute Working Set Size in the “Workload workspace” on page57 to determine which workloads are consuming the most real storage. See also the User Page Data(UPAGE) screen of the Performance Toolkit. This screen provides working set size data and the count ofpages that are locked or reserved for workloads. To avoid raising this exception unnecessarily, considermodifying the threshold for this situation if it is not appropriate for your environment.

The following shows the formula for the ZVM_Page_Used_Critical situation:

*IF *VALUE KVLSystem.Pct Page Space In Use *GE 50

ZVM_Page_Used_HighRaises an alert when the percentage of any one of the allocations listed below is greater than or equal to40 percent and less than 50 percent:





If the situation is triggered because of a SPOOL allocation, spool files or overflow paging may beconsuming a large percentage of the available space. Refer to the SPOOL File Summary (SPOOL)screen of the Performance Toolkit. This screen shows how many disk blocks are consumed by each file.Use this as a guide for determining which spool files would yield the most free space if deleted. You mightwant to consider adding more space if spool utilization is typically high. You could also offload spool usingthe SPXTAPE command with the PURGE option if the files need to be retained. This situation must bestarted manually.

If the situation raises for a paging allocation, available paging space may be dangerously low. Consideradding additional paging space or reducing the active workload on the system by logging off guestsystems that are not critical.

Refer to the Working Set Size attribute in the “Workload workspace” on page 57 to determine whichworkloads are consuming the most real storage. See also the User Page Data (UPAGE) screen of thePerformance Toolkit. This screen provides working set size data and the count of pages that are locked orreserved for workloads.

If the situation raises for a dump allocation, consider processing the dumps using the DUMPLOADcommand and deleting them from the dump area. To avoid raising this exception unnecessarily, considermodifying the threshold for this situation if it is not appropriate for your environment.

The following shows the formula for the ZVM_Page_Used_High situation:

*IF *VALUE KVLSystem.Pct Page Space In Use *GE 40 *AND *VALUE KVLSystem.Pct Page Space In Use*LT 50

ZVM_PerfKit_Collector_InactiveRaises an alert when the Performance Toolkit has stopped updating the PERFOUT DCSS with data that ithas collected. This situation is set to run automatically at startup of the monitoring agent. TheOMEGAMON XE on z/VM and Linux monitoring agent allows three sample intervals to expire without anupdate to the PERFOUT DCSS before setting the Collector Status attribute to Inactive. This minimizesfalse triggering of the situation by allowing the Performance Toolkit some time to recover from a transienterror condition or processing delay. When this situation raises, investigate the cause immediately.

Review the following list to ensure that your configuration is complete:

v Check the monitoring agent log for information that could help you to pinpoint the problem. See the IBMTivoli OMEGAMON XE on z/VM and Linux Planning and Configuration Guide for the locations of the logfiles.

v Ensure that the Performance Toolkit is running.

v Check the status of this monitoring agent to see if it is running. You do this by issuing the followingcommand:./cinfo -R

When you use the -R option of the cinfo command, the processes currently running display. If the agentis active, the product code vl for this monitoring agent, displays, along with its running status. See theIBM Tivoli Monitoring Installation and Setup Guide for additional information on this command.

v The z/VM CP Monitor must be running for the Performance Toolkit to obtain data. If the CP Monitor isnot running, issue the MONITOR START command. Also verify that the necessary Monitor Sample andEvent domains are enabled. See the z/VM: CP Commands and Utilities Reference for furtherinformation. See also the IBM Tivoli OMEGAMON XE on z/VM and Linux Planning and ConfigurationGuide.


v From the Performance Toolkit screen, issue the FC MONCOLL SEGOUT QUERY command todetermine the status of data collection in the PERFOUT DCSS.

v Verify that this monitoring agent is reading data from the same PERFOUT DCSS to which thePerformance Toolkit is writing data. Use the CP command QUERY NSS USERS <dcss_name>, wheredcss_name is the name of the DCSS reported by the FC MONCOLL SEGOUT QUERY command. TheQUERY NSS command should report only two users sharing the PERFOUT DCSS:

– The virtual machine where this monitoring agent is running.

– The virtual machine where the Performance Toolkit is running.

If the response to the FC MONCOLL SEGOUT QUERY command indicates that SEGOUT datacollection is OFF, issue the FC MONCOLL SEGOUT ON command to start collecting data in thePERFOUT DCSS. See the z/VM: Performance Toolkit Reference for instructions on using the FCMONCOLL subcommand.

v The Performance Toolkit also stops collecting data when the size of the DCSS has been exceeded.Increase the size of the DCSS so that it is large enough to allow the update. After increasing the size ofthe DCSS, restart the Performance Toolkit. See the IBM Tivoli OMEGAMON XE on z/VM and LinuxPlanning and Configuration Guide for information on estimating the size of the DCSS.

Note: The FC MONCOLL SEGOUT ON command can be placed in the FCONX $PROFILE on the A-diskof the Performance Toolkit virtual machine. This automatically starts data collection in the PERFOUTDCSS during initialization of the Performance Toolkit.

You can access information on the Performance Toolkit and on the latest enhancements to thePerformance Toolkit as they relate to this monitoring agent at the following Web address:


The following shows the formula for the ZVM_PerfKit_Collector_Inactive situation:

*IF *VALUE KVLPTKStat.Collector Name *EQ 'Performance Toolkit Collector' *AND *VALUEKVLPTKStat.Collector Status *EQ INACTIVE

ZVM_Physical_CPU_CriticalRaises an alert when the combined percentage of CPU utilization for all physical processors across alllogical partitions in a complex is greater than or equal to 90 percent. This indicates that all the processorswere busy for the specified percentage of the elapsed time interval. Utilization applies to those logicalpartitions with non-IFL and non-ICF processors only, and includes both LPAR management time and timespent servicing the guest system. The Physical CPU Busy field on the “LPAR workspace” on page 44 isa processor complex-wide metric. Physical CPU Busy is the sum of all the %Load values, plus thegeneral LPAR management overhead for the partitions.

Use the “LPAR workspace” on page 44 to determine which partitions in the complex are using the mostprocessor resources. See also the LPAR Load screen of the Performance Toolkit for additionalinformation. This situation must be started manually.

To prevent this situation from triggering for all defined partitions, a check is included for whether the datawas retrieved by this partition. A high Physical CPU Busy percentage may indicate a looping process, oran abnormally high system load. To determine the cause of the problem, it helps to understand theprocessor load of all of the partitions in the complex. Then use the performance monitoring tools for thoseenvironments to identify the cause. The thresholds used in this situation may not be appropriate for yourinstallation if your system typically has high processor utilization.

The following shows the formula for the ZVM_Physical_CPU_Critical situation:

*IF *VALUE KVLLPAR_Info.Physical CPU Busy *GE 90.00 *AND *VALUE KVLLPAR_Info.LPAR Status *EQACTIVE*



ZVM_Physical_CPU_HighRaises an alert when the combined percentage of CPU utilization for all physical processors across alllogical partitions in a complex is greater than or equal to 80 percent and less than 90 percent. Thisindicates that all the processors were busy for the specified percentage of the elapsed time interval.Utilization applies to those logical partitions with non-IFL and non-ICF processors only, and includes bothLPAR management time and time spent servicing the guest system. The Physical CPU Busy field on the“LPAR workspace” on page 44 is a processor complex-wide metric. Physical CPU Busy is the sum of allthe percentage load values, plus the general LPAR management overhead for the partitions. This situationmust be started manually.

Use the “LPAR workspace” on page 44 to determine those partitions in the complex that are using themost processor resources. See also the LPAR Load screen of the Performance Toolkit.

To prevent this situation from triggering for all defined partitions, a check is included for whether the datawas retrieved by this partition. A high Physical CPU Busy percentage may indicate a looping process, oran abnormally high system load. To determine the cause of the problem, it helps to understand theprocessor load of all of the partitions in the complex. Then use the performance monitoring tools for thoseenvironments to identify the cause. The thresholds used in this situation may not be appropriate for yourinstallation if your system typically has high processor utilization.

The following shows the formula for the ZVM_Physical_CPU_High situation:

*IF *VALUE KVLLPAR_Info.Physical CPU Busy *GE 80.00 *AND *VALUE KVLLPAR_Info.Physical CPU Busy*LT 90.00 *AND *VALUE KVLLPAR_Info.LPAR Status *EQ ACTIVE*

ZVM_Spin_Exclusive_Pct_CriticalRaises an alert when the percentage of time spent spinning on a monitored exclusive lock exceeds thespecified value. Spinning on an exclusive lock means that work on a processor is suspended until therequired lock becomes available. The z/VM Control Program (CP) scheduler is prevented from completingprocessing due to the need to serialize on a critical internal data structure shared by multiple processors.This situation must be started manually.

Frequent spinning on an exclusive lock can significantly impact the performance of a system. It may not bepossible to reduce the spin locks initiated by CP. Observing an unusually high percentage of time spentspinning on an exclusive lock may indicate a problem that needs further investigation by z/VM support.Increasing the number of processors available in a partition increases locking, due to more processorsaccessing shared internal structures. Higher spin times may occur when the logical partition hosting z/VMdoes not get sufficient processor time.

Note: This metric does not include any spinning on locks performed by guest operating systems for theirown internal locks.

Use the “Spin Locks workspace” on page 53 to obtain detailed wait statistics on processor spin lockactivity. See also the Spin Lock Log (LOCKLOG) screen of the Performance Toolkit for a recent history ofCP lock activity. Refer to the Performance Toolkit Reference for additional information.

The following shows the formula for the ZVM_Spin_Exclusive_Pct_Critical situation:

*IF *VALUE KVLSpinlock.Exclusive Time Spinning on Locks Percent *GT 30.00

ZVM_Spin_Exclusive_Pct_HighRaises an alert when the percentage of time spent spinning on a monitored exclusive lock is within therange specified in the situation formula. Spinning on an exclusive lock indicates that work on a processoris suspended until the required lock becomes available. The z/VM Control Program (CP) scheduler isprevented from completing processing due to the need to serialize on a critical internal data structure


shared by multiple processors. Frequent spinning on an exclusive lock can significantly impact theperformance of a system. This situation must be started manually.

It may not be possible to reduce the spin locks initiated by CP. Observing an unusually high percentage oftime spent spinning on an exclusive lock may indicate a problem that needs further investigation by z/VMsupport. Increasing the number of processors available in a partition increases locking, due to moreprocessors accessing shared internal structures. Higher spin times may occur when the logical partitionhosting z/VM does not get sufficient processor time.

Note: This metric does not include any spinning on locks performed by guest operating systems for theirown internal locks.

Use the “Spin Locks workspace” on page 53 to obtain detailed wait statistics on processor spin lockactivity. See also the Spin Lock Log (LOCKLOG) screen of the Performance Toolkit for a recent history ofCP lock activity. Refer to the Performance Toolkit Reference for additional information.

The following shows the formula for the ZVM_Spin_Exclusive_Pct_High situation:

*IF *VALUE KVLSpinlock.Exclusive Time Spinning on Locks Percent *GT 25.00 AND *VALUEKVLSpinlock.Exclusive Time Spinning on Locks Percent *LE 30.00

ZVM_Spool_Used_CriticalRaises an alert when the percentage of allocated spool area used on all defined spool volumes is greaterthan or equal to 95 percent. Spool files or overflow paging may be consuming a large percentage of theavailable space. For detailed information, refer to the SPOOL File Summary (SPOOL) screen of thePerformance Toolkit. This screen shows how many disk blocks are consumed by each file. Use this as aguide for determining which spool files would yield the most free space if removed from the spool area.This situation must be started manually.

If the spool area is in danger of being completely full, you can perform any one of these actions or all ofthese actions:

v Delete spool files that are not needed.

v If files must be retained, offload spool files using the SPXTAPE command.

v Increase the size of the spool area.

To avoid raising this exception unnecessarily, consider modifying the threshold specified for this situation ifit is not appropriate for your environment and needs to be adjusted.

The following shows the formula for the ZVM_Spool_Used_Critical situation:

*IF *VALUE KVLSystem.Pct Spool Space In Use *GE 95

ZVM_Spool_Used_HighRaises an alert when the percentage of allocated spool area used on all defined spool volumes is greaterthan or equal to 80 percent and less than 95 percent. spool files or overflow paging may be consuming alarge percentage of the available space. You can navigate to the SPOOL File Summary (SPOOL) screenof the Performance Toolkit. This screen shows how many disk blocks are consumed by each file. Use thisas a guide for determining which spool files would yield the most free space if they were removed from thespool area. This situation must be started manually.

If the spool area is in danger of being completely full, you can perform any one of these actions or all ofthese actions:

v Delete spool files that are not needed.

v If files must be retained, offload spool files using the SPXTAPE command.

v Increase the size of the spool area.


To avoid raising this exception unnecessarily, consider modifying the threshold specified for this situation ifit is not appropriate for your environment and needs to be adjusted.

The following shows the formula for the ZVM_Spool_Used_High situation:

*IF *VALUE KVLSystem.Pct Spool Space In Use *GE 80 *AND *VALUE KVLSystem.Pct Spool Space In Use*LT 95

ZVM_Total_CPU_CriticalRaises an alert when the percentage of CPU utilized by both the Control Program (overhead) and theguest virtual machine workload is greater than or equal to 90 percent. Overhead includes both supervisortime spent for users and supervisor time spent for system services. The value reported is the sum of allthe percent CPU values reported for all processors assigned to the logical partition. This situation must bestarted manually.

Use the “Workload workspace” on page 57 to identify any virtual machines that are consuming a largeamount of CPU. See also the General CPU screen on the Performance Toolkit. This screen displays thepercentage of CPU utilization by individual processor.

Since overhead is included in the total CPU percent, the Control Program may be contributing to highCPU. You might want to check the load on the paging subsystem. The Control Program may be busyprocessing a high number of page faults caused by guest operating systems.

For virtual machines running guest operating systems, a monitoring product for that platform can provideadditional information. It is possible that the high CPU utilization is caused by a user application or aprocess stuck in a loop. The application in question should be examined for inefficient or defective code. Itis also possible that the total CPU is high simply because there is legitimate processing work to be donethat is consuming all available CPU. A processor upgrade may be needed. Consider modifying thethreshold for this situation if high total CPU utilization is normal in your environment.

The following shows the formula for the ZVM_Total_CPU_Critical situation:

*IF *VALUE KVLSystem.Percent of CPU *GE 90

ZVM_Total_CPU_HighRaises an alert when the percentage of CPU utilized by both the Control Program (overhead) and theguest virtual machine workload is greater than or equal to 80 percent and less than 90 percent. Overheadincludes both supervisor time spent for users and supervisor time spent for system services. The valuereported is the sum of all the percent CPU values reported for all processors assigned to the logicalpartition. This situation must be started manually.

Use the “Workload workspace” on page 57 to identify any virtual machines that are consuming a largeamount of CPU. See also the General CPU screen of the Performance Toolkit. This screen displays thepercentage of CPU utilization by individual processor.

Since overhead is included in the total CPU percent, the Control Program may be contributing to highCPU. You might want to check the load on the paging subsystem. The Control Program may be busyprocessing a high number of page faults caused by guest operating systems.

For virtual machines running guest operating systems, a monitoring product for that platform can provideadditional information. It is possible that the high CPU utilization is caused by a user application or aprocess stuck in a loop. The application in question should be examined for inefficient or defective code. Itis also possible that the total CPU is high simply because there is legitimate processing work to be donethat is consuming all available CPU. A processor upgrade may be needed. Consider modifying thethreshold for this situation if high total CPU utilization is normal in your environment.


The following shows the formula for the ZVM_Total_CPU_High situation:

*IF *VALUE KVLSystem.Percent of CPU *GE 80 *AND *VALUE KVLSystem.Percent of CPU *LT 90

ZVM_Total_to_Virtual_HighRaises an alert when the ratio of total CPU time used to virtual CPU time used is greater than or equal to2. This ratio is the mean total to virtual ratio of all virtual machines on the system and indicates how muchprocessor overhead is incurred for virtual machine processing. A lower total to virtual ratio means lessprocessor time has been used by the Control Program and more time has been used by virtual machineprocessing. This situation must be started manually.

Review the Total to Virtual Ratio linear gauge chart in the “System workspace” on page 52. See also theUSER screen of the Performance Toolkit. This screen displays the mean T/V Ratio.

Consider changing the thresholds used in this situation as they may not be appropriate for yourinstallation.

The following shows the formula for the ZVM_Total_to_Virtual_High situation:

*IF *VALUE KVLSystem.Total to Virtual Ratio *GE 2.00

ZVM_User_CPU_CriticalRaises an alert when the percentage of total CPU used by the system to manage the workload is greaterthan or equal to 90 percent. The value reported is the percentage of total CPU used by the system tomanage a workload. The value is averaged over the number of virtual processors in a multiple processorconfiguration. When multiple virtual processors are used, the percentage can sometimes exceed 100%.This situation must be started manually.

Review the data displayed in the All z/VM Workloads table in the “Workload workspace” on page 57. Seealso the %CPU value in the User Resource Usage (USER) screen of the Performance Toolkit. Thatscreen displays data for all the virtual processors defined to the virtual machine. See the PerformanceToolkit Reference for information on the User Resource Usage (USER) screen.

A monitoring product for the guest platform can provide additional information for virtual machines that arerunning guest operating systems. It is possible that the high CPU utilization is caused by a looping userapplication or process. Examine the application in question for inefficient or defective code. It is alsopossible that the total CPU is high simply because there is legitimate processing work to be done that iscausing high CPU utilization. A processor upgrade may be needed. Consider modifying the threshold forthis situation if high CPU utilization is normal for this workload.

The following shows the formula for the ZVM_User_CPU_Critical situation:

*IF *VALUE KVLUser_Workload.CPU Percent *GE 90.00

ZVM_User_CPU_HighRaises an alert when the percentage of total CPU used by the system to manage the workload is greaterthan or equal to 80 percent and less than 90 percent. The value reported is the percentage of total CPUused by the system to manage a workload. The value is averaged over the number of virtual processors ina multiple processor configuration. When multiple virtual processors are used, the percentage cansometimes exceed 100%. This situation must be started manually.

Review the data displayed in the All z/VM Workloads table in the “Workload workspace” on page 57. Seealso the %CPU value in the User Resource Usage (USER) screen of the Performance Toolkit. Thatscreen displays data for all the virtual processors defined to the virtual machine.


A monitoring product for the guest platform can provide additional information for virtual machines that arerunning guest operating systems. It is possible that the high CPU utilization is caused by a looping userapplication or process. Examine the application in question for inefficient or defective code. It is alsopossible that the total CPU is high simply because there is legitimate processing work to be done that iscausing high CPU utilization. A processor upgrade may be needed. Consider modifying the threshold forthis situation if high CPU utilization is normal for this workload.

The following shows the formula for the ZVM_User_CPU_High situation:

*IF *VALUE KVLUser_Workload.CPU Percent *GE 80.00 *AND *VALUE KVLUser_Workload.CPU Percent *LT90.00

ZVM_User_Scaled_CPU_CriticalRaises an alert when the percentage of total CPU used by the system to manage and run the workload isgreater than or equal to 90 percent. The value reported is the percentage of total CPU used by the systemto manage and run a workload. The value is scaled over the number of virtual processors in amultiprocessor configuration and does not exceed 100%.

Review the data displayed in the All z/VM Workloads table in the “Workload workspace” on page 57 ofthis monitoring agent. See also the Total CPU or %CPU (if multiple virtual processors are defined) valuein the User Resource Details (USER userid) screen of the Performance Toolkit. This window shows datafor all the virtual processors defined to the virtual machine.

It is possible that the high CPU utilization is caused by a looping user application or process. Examine theapplication in question for inefficient or defective code. It is also possible that the total CPU is highbecause there is legitimate processing work to be done that is causing high CPU utilization. A processorupgrade might be needed. Consider modifying the threshold for this situation if high CPU utilization isnormal for this workload.

The following shows the formula for the ZVM_User_Scaled_CPU_Critical situation:

*IF *VALUE KVLUser_Workload.CPU Scaled Percent *GE 90.00

ZVM_User_Scaled_CPU_HighRaises an alert when the percentage of total CPU used by the system to manage and run the workload isgreater than or equal to 80 percent and less than 90 percent. The value reported is the percentage of totalCPU used by the system to manage and run a workload. The value is scaled over the number of virtualprocessors in a multi-processor configuration and does not exceed 100%

Review the data displayed in the All z/VM Workloads table in the “Workload workspace” on page 57 ofthis monitoring agent. See also the Total CPU or %CPU (if multiple virtual processors are defined) valuein the User Resource Details (USER user id) screen of the Performance Toolkit. This window showsdata for all the virtual processors defined to the virtual machine.

It is possible that the high CPU utilization is caused by a looping user application or process. Examine theapplication in question for inefficient or defective code. It is also possible that the total CPU is high simplybecause there is legitimate processing work to be done that is causing high CPU utilization. A processorupgrade might be needed. Consider modifying the threshold for this situation if high CPU utilization isnormal for this workload.

The following shows the formula for the ZVM_User_Scaled_CPU_High situation:

*IF *VALUE KVLUser_Workload.CPU Scaled Percent *GE 80.00 *AND *VALUE KVLUser_Workload.CPU ScaledPercent *LT 90.00


ZVM_VDISK_Page_High_IORate_LowA virtual disk in storage (VDISK) appears as a Fixed Block Architecture (FBA) disk device (virtual devicetype 9336). The Control Program emulates this device by moving pages between the primary addressspace of the virtual machine and the VDISK address space. I/O requests to VDISK devices are low incost. However, there is a trade-off between real I/O requests avoided by using an emulated device, andthe system paging space consumed by the allocation. This situation must be started manually.

Note: This situation embeds situation “ZVM_Page_Queue_High” on page 73. You must start situationZVM_Page_Queue_High first, prior to starting this situation. Additionally, this situation is defined to beraised only when it evaluates to TRUE for 10 consecutive sampling intervals.

This situation triggers under the following conditions:

v The system paging rate appears to be high.

v A defined VDISK is not being used sufficiently.

If this situation triggers often, consider detaching the VDISK (if possible) to release the page space itconsumes back to CP for use in general paging. It is possible that the low level of VDISK I/O and the highpaging rate are transitory.

In instances where you are using a VDISK for a Linux swap device, there may be benefit in creatingmultiple smaller VDISKS for swapping, instead of one large VDISK. The pages associated with virtual diskblocks are not created until they are actually referenced.

Use the “VDISK workspace” on page 43, accessed from the “DASD workspace” on page 37, for details onoverall storage utilization by virtual disks.

Note: Extensive use of VDISKs can result in increased paging subsystem load. Therefore, VDISK usagein memory-constrained, high paging rate, or highly overcommitted environments must be carefullyevaluated.

The following shows the formula for the ZVM_VDISK_Page_High_IORate_Low situation:

*IF *SIT (ZVM Page Queue High) *EQ TRUE *AND *VALUE KVLVdisk.Virtual I/O's per Second *LT 10.0

ZVM_Virtual_CPU_CriticalRaises an alert when the percentage of CPU used by the guest virtual machine workload is greater thanor equal to 90 percent. When multiple virtual processors are used, the percentage can sometimes exceed100%. This situation must be started manually.

See the Top 5 CPU Users bar chart in the “Workload workspace” on page 57 to find out which workloadsare using the most CPU. Virtual CPU does not include the Control Program overhead. Refer to the UserResource Usage (USER) screen of the Performance Toolkit for additional details. For virtual machinesrunning guest operating systems, a monitoring product for that platform can provide more information.

It is possible that the high virtual CPU utilization is caused by a looping user application or process. Theapplication in question should be examined for inefficient or defective code. It is also possible that theVirtual CPU is high simply because there is legitimate processing work to be done that is using all theavailable resources. A processor upgrade may be needed. Consider modifying the threshold for thissituation if high virtual CPU utilization is normal in your environment.

The following shows the formula for the ZVM_Virtual_CPU_Critical situation:

*IF *VALUE KVLUser_Workload.Virtual CPU % *GE 90.00


ZVM_Virtual_CPU_HighRaises an alert when the percentage of CPU used by the guest virtual machine workload is greater thanor equal to 80 and less than 90 percent. This situation must be started manually.

See the Top 5 CPU Users bar chart in the “Workload workspace” on page 57 to find out which workloadsare using the most CPU. Virtual CPU does not include the Control Program overhead. When multiplevirtual processors are used, the percentage can sometimes exceed 100%. Refer to the User ResourceUsage (USER) screen of the Performance Toolkit for additional details. For virtual machines running guestoperating systems, a monitoring product for that platform can provide more information.

It is possible that the high virtual CPU utilization is caused by a looping user application or process. Theapplication in question should be examined for inefficient or defective code. It is also possible that thevirtual CPU is high simply because there is legitimate processing work to be done that is using all theavailable resources. A processor upgrade may be needed. Consider modifying the threshold for thissituation if high virtual CPU utilization is normal in your environment.

The following shows the formula for the ZVM_Virtual_CPU_High situation:

*IF *VALUE KVLUser_Workload.Virtual CPU % *GE 80.00 *AND *VALUE KVLUser_Workload.Virtual CPU %*LT 90.00

ZVM_Virtual_Scaled_CPU_CriticalRaises an alert when the percentage of total CPU used by the system to run the workload is greater thanor equal to 90 percent. The value reported is the percentage of total CPU used by the system to run aworkload. The value is scaled over the number of virtual processors in a multiprocessor configuration anddoes not exceed 100%.

Review the data that is shown in the All z/VM Workloads table in the “Workload workspace” on page 57.See also the Emulat. CPU or %EM (if multiple virtual processors are defined) value in the User ResourceDetails (USER userid) window of the Performance Toolkit. This window displays data for all the virtualprocessors defined to the virtual machine. It is possible that the high CPU utilization is caused by alooping user application or process. Examine the application in question for inefficient or defective code. Itis also possible that the total CPU is high simply because there is legitimate processing work to be donethat is causing high CPU utilization. A processor upgrade might be needed. Consider modifying thethreshold for this situation if high CPU utilization is normal for this workload.

The following shows the formula for the ZVM_Virtual_Scaled_CPU_Critical situation:

*IF *VALUE KVLUser_Workload.Virtual CPU Scaled Pct *GE 90.00

ZVM_Virtual_Scaled_CPU_HighRaises an alert when the percentage of total CPU used by the system to run the workload is greater thanor equal to 80 percent and less than 90 percent. The value reported is the percentage of total CPU usedby the system to run a workload. The value is scaled over the number of virtual processors in amultiprocessor configuration and does not exceed 100%.

Review the data that is shown in the All z/VM Workloads table in the “Workload workspace” on page 57.See also the Emulat. CPU or %EM (if multiple virtual processors are defined) value in the User ResourceDetails (USER userid) window of the Performance Toolkit. This window displays data for all the virtualprocessors defined to the virtual machine. It is possible that the high CPU utilization is caused by alooping user application or process. Examine the application in question for inefficient or defective code. Itis also possible that the total CPU is high simply because there is legitimate processing work to be donethat is causing high CPU utilization. A processor upgrade might be needed. Consider modifying thethreshold for this situation if high CPU utilization is normal for this workload.

The following shows the formula for the ZVM_Virtual_Scaled_CPU_High situation:


*IF *VALUE KVLUser_Workload.Virtual CPU Scaled Pct *GE 80.00 *AND *VALUEKVLUser_Workload.Virtual CPU Scaled Pct *LT 90.00

ZVM_User_Wait_Page_CriticalRaises an alert when the percentage of samples during which the workload was found to be active andwaiting for the completion of a page read operation is greater than 40 percent. This information is from thehigh frequency user state sampling facility of the CP Monitor facility. A workload enters page wait statewhen it references a page that is not present in host storage and must be brought in from auxiliarystorage. This state reflects time waiting for page reads from DASD. Page writes and paging fromexpanded storage do not affect this state.

The triggering of this situation indicates that there are not enough free page frames to satisfy pagerequests for active workloads. This can result in greater page wait and possible thrashing of the pagingsubsystem. Consider adding real memory to the system. You can also reduce the paging demand byremoving workload where possible.

If you used the SET RESERVED CP command to reserve pages for preferred virtual machines, considerreducing the amount reserved. The SET RESERVED can be useful in a paging environment to ensure thatvirtual machines on which others are dependent do not wait for paging. For example, if a virtual machineacting as a file server for many other virtual machines must wait for paging, then all the dependent virtualmachines must also wait. By minimizing page wait by the file server, you shorten the time the other virtualmachines need to be active and to have their pages active. Be aware that abusing the SET RESERVEDcommand can create memory constraints and excessive paging.

Virtual disks in storage (VDISKs) are located in the dynamic paging area. See the VDISK workspace todetermine which virtual machines are using the VDISK feature. Look for VDISKs that are not being used tofull capacity, and that can be removed from the system. For Linux on System z virtual machines, exploresharing memory between machines by using a named saved system (NSS) or a discontiguous savedsegment (DCSS).


See the Paging Log (PAGELOG), the System Counters (SYSTRANS), the CPU Load andTransactions (CPU), the Storage Utilization Log (STORLOG), and the User Page Data (UPAGE)screens of the Performance Toolkit for additional information. See the Performance Toolkit Reference forinformation on the Performance Toolkit screens.

The following shows the formula for the ZVM_User_Wait_Page_Critical situation:

*IF *VALUE KVLUser_Wait.Page_Wait_Percent > 40

ZVM_User_Wait_Page_HighRaises an alert when the percentage of samples during which the workload was found to be active andwaiting for the completion of a page read operation is greater than 25 percent and less than or equal to 40percent. This information is from the high frequency user state sampling facility of the CP Monitor facility. Aworkload enters page wait state when it references a page that is not present in host storage and must bebrought in from auxiliary storage. This state reflects time waiting for page reads from DASD. Page writesand paging from expanded storage do not affect this state.

The triggering of this situation indicates that there are not enough free page frames to satisfy pagerequests for active workloads. This can result in greater page wait and possible thrashing of the pagingsubsystem. Consider adding real memory to the system. You can also reduce the paging demand byremoving workload where possible.


If you used the SET RESERVED CP command to reserve pages for preferred virtual machines, considerreducing the amount reserved. SET RESERVED can be useful in a paging environment to ensure thatvirtual machines on which others are dependent do not wait for paging. For example, if a virtual machineacting as a file server for many other virtual machines must wait for paging, then all the dependent virtualmachines must also wait. By minimizing page wait by the file server, you shorten the time the other virtualmachines need to be active and to have their pages active. Be aware that abusing the SET RESERVEDcommand can create memory constraints and excessive paging.

Virtual disks in storage (VDISKs) are located in the dynamic paging area. See the VDISK workspace todetermine which virtual machines are using the VDISK feature. Look for VDISKs that are not being used tofull capacity, and that can be removed from the system. For Linux on System z virtual machines, exploresharing memory between machines by using a named saved system (NSS) or a discontiguous savedsegment (DCSS).


See the Paging Log (PAGELOG), the System Counters (SYSTRANS), the CPU Load andTransactions (CPU), the Storage Utilization Log (STORLOG), and the User Page Data (UPAGE)screens of the Performance Toolkit for additional information. See the Performance Toolkit Reference forinformation about the Performance Toolkit screens.

The following shows the formula for the ZVM_User_Wait_Page_High situation:

*IF *VALUE KVLUser_Wait.Page_Wait_Percent > 25 and KVLUser_Wait.Page_Wait_Percent <= 40

ZVM_User_Wait_CPU_CriticalRaises an alert when the percentage of samples during which the workload was found to be active andwaiting for the CPU is greater than 40 percent. This information is from the high frequency user statesampling facility of the CP Monitor facility. A workload is in this state when it is found waiting to run on areal processor. High CPU wait time indicates a bottleneck in processor resources. The value does notexceed 100% even in a multi-processor configuration.

Review the data displayed in the All z/VM Workloads table in the Workload workspace of this monitoringagent. See also the %CPU value in the User Resource Usage (USER) screen of the Performance Toolkit.The User Resource Usage (USER) screen displays data for all of the virtual processors defined to thevirtual machine. See the Performance Toolkit Reference for information on the User Resource Usage(USER) screen.

A monitoring product for the guest platform can provide additional information for virtual machines that arerunning guest operating systems. The high CPU utilization might be caused by a looping user applicationor process. The application in question might be examined for inefficient or defective code. The total CPUmight be high because there is legitimate processing work to be performed that is causing high CPUutilization. A processor upgrade might be needed. You might modify the threshold for this situation if highCPU utilization is normal for this workload.

The following shows the formula for the ZVM_User_Wait_CPU_Critical situation:

*IF *VALUE KVLUser_Wait.CPU_Wait_Percent > 40

ZVM_User_Wait_CPU_HighRaises an alert when the percentage of samples during which the workload was found to be active andwaiting for the CPU is greater than 25 percent and less than or equal to 40 percent. This information isfrom the high frequency user state sampling facility of the CP Monitor facility. A workload is in this state


when it is found waiting to run on a real processor. High CPU wait time indicates a bottleneck in processorresources. The value does not exceed 100% even in a multi-processor configuration.

Review the data displayed in the All z/VM Workloads table in the Workload workspace of this monitoringagent. See also the %CPU value in the User Resource Usage (USER) screen of the Performance Toolkit.The User Resource Usage (USER) screen displays data for all of the virtual processors defined to thevirtual machine. See the Performance Toolkit Reference for information on the User Resource Usage(USER) screen.

A monitoring product for the guest platform can provide additional information for virtual machines that arerunning guest operating systems. The high CPU utilization might be caused by a looping user applicationor process. The application in question might be examined for inefficient or defective code. The total CPUmight be high because there is legitimate processing work to be performed that is causing high CPUutilization. A processor upgrade might be needed. You might modify the threshold for this situation if highCPU utilization is normal for this workload.

The following shows the formula for the ZVM_User_Wait_CPU_High situation:

*IF *VALUE KVLUser_Wait.CPU_Wait_Percent > 25 and KVLUser_Wait.CPU_Wait_Percent <= 40

ZVM_Storage_Overcommit_CriticalRaises an alert when the ratio of virtual storage to real main storage is greater than 4 to 1. Virtual storageis the sum of the virtual machine sizes (Average Storage Size) from the OMEGAMON XE on z/VMWorkload group for all logged-on virtual machines. The sum is divided by the amount of storageconfigured for z/VM to use. The storage-over-commit ratio is an indication of how many virtual machinepages are allocated for each real storage page. A system with a storage ratio that is too large mightexperience excessive paging rates, or other performance problems, because the real memory isovercommitted. A system with a small ratio (less than or near 1 to 1) that has sufficient CPU capacityavailable should be able to handle more workload.

Overcommitted virtual memory might be causing excessively high paging rates. Consider adding realmemory, or reducing the virtual storage size of logged-on virtual machines. You can reduce the virtualstorage size by establishing lower limits in the z/VM Control Program (CP) directory. Log off anyunnecessary virtual machines with large virtual storage sizes to reclaim the memory they are using. Seethe All z/VM Workloads view in the Workload workspace of this monitoring agent. The virtual storage sizefor all logged-on users is displayed. See also the Resource Constraint workspace to determine ifworkloads are waiting for pages. A high page wait percent indicates that the storage and pagingsubsystem require evaluation.

Virtual disks in storage (VDISKs) are located in the dynamic paging area. See the VDISK workspace ofthis monitoring agent to determine which virtual machines are using the VDISK feature. Look for VDISKsthat are not being used to full capacity that can be removed from the system. For Linux on System zvirtual machines, explore sharing memory between machines by using a named saved system (NSS) or adiscontiguous saved segment (DCSS).

Note: Extensive use of VDISKs can result in increased paging subsystem load. Therefore, VDISK usagein memory-constrained, high paging rate, or highly over committed environments must be carefullyevaluated.

See the Paging Log (PAGELOG), the System Counters (SYSTRANS), the CPU Load and Transactions(CPU), the Storage Utilization Log (STORLOG), and the User Page Data (UPAGE) screens of thePerformance Toolkit for additional information. See the z/VM: Performance Toolkit Reference forinformation about the Performance Toolkit screens.

The following shows the formula for the ZVM_Storage_Overcommit_Critical situation:


*IF *VALUE KVLSystem2.Real_Storage_Overcommit > 4.0

ZVM_Storage_Overcommit_HighRaises an alert when the ratio of virtual storage to real main storage is greater than 3 to 1, and less thanor equal to 4 to 1. Virtual storage is the sum of the virtual machine sizes (Average Storage Size) from theOMEGAMON XE on z/VM Workload group for all logged-on virtual machines. The sum is divided by theamount of main storage configured for z/VM to use. The storage overcommit ratio is an indication of howmany virtual machine pages are allocated for each real storage page. A system with a storage ratio that istoo large might experience excessive paging rates, or other performance problems, because the realmemory is overcommitted. A system with a small ratio (less than or near 1 to 1) that has sufficient CPUcapacity available should be able to handle additional workload.

Overcommitted virtual memory might be causing excessively high paging rates. Consider adding realmemory, or reducing the virtual storage size of logged-on virtual machines. You can reduce the virtualstorage size by establishing lower limits in the z/VM Control Program (CP) directory. Log off anyunnecessary virtual machines with large virtual storage sizes to reclaim the memory they are using. Seethe All z/VM Workloads view in the Workload workspace of this monitoring agent. The virtual storage sizefor all logged-on users is displayed. See also the Resource Constraint workspace to determine ifworkloads are waiting for pages. A high page wait percent indicates that the storage and pagingsubsystem require evaluation.

Virtual disks in storage (VDISKs) are located in the dynamic paging area. See the VDISK workspace ofthis monitoring agent to determine which virtual machines are using the VDISK feature. Look for VDISKsthat are not being used to full capacity that can be removed from the system. For Linux on System zvirtual machines, explore sharing memory between machines by using a named saved system (NSS) or adiscontiguous saved segment (DCSS).

Note: Extensive use of VDISKs can result in increased paging subsystem load. Therefore, VDISK usagein memory-constrained, high paging rate, or highly over committed environments must be carefullyevaluated.

See the Paging Log (PAGELOG), the System Counters (SYSTRANS), the CPU Load and Transactions(CPU), the Storage Utilization Log (STORLOG), and the User Page Data (UPAGE) screens of thePerformance Toolkit for additional information. See the z/VM: Performance Toolkit Reference forinformation about the Performance Toolkit screens.

The following shows the formula for the ZVM_Storage_Overcommit_High situation:

*IF *VALUE KVLSystem2.Real_Storage_Overcommit > 3.0 and KVLSystem2.Real_Storage_Overcommit <=4.0

ZVM_Eligible_List_HighRaises an alert when the average number of users waiting in the z/VM scheduler eligible list is greaterthan 5. The eligible list contains a list of virtual machines that are unable to run because they are waitingfor resources. When a virtual machine is logged on, it is placed in the dormant list. It is moved to theeligible list by the scheduler only when it has work to do. As resources become available, the schedulermoves virtual machines from the eligible list to the dispatch list. The dispatch list consists of the logged-onvirtual machines that are currently in the set of virtual machines being considered for dispatching. That is,the virtual machines in the dispatch list are competing for use of the real processors and all otherscheduled system resources. The eligible list forms when there is more demand by virtual machines forreal system resources than is available. Therefore, the virtual machines in the eligible list are ready to runand are waiting to enter the dispatch list. A virtual machine might appear in the eligible list multiple timesas each virtual processor defined to a virtual machine is a dispatchable entity.


Review the Resource Constraint workspace to determine the wait states causing the eligible list to form.High CPU wait time indicates a bottleneck in processor resources. A user ID is in this state when it isfound waiting to run on a real processor. You might need to add real processing capacity, or deferworkload.

High I/O wait time indicates that you should check the effectiveness of minidisk cache, virtual disk instorage, and the I/O hardware configuration. This state indicates that the guest is waiting for completion ofan I/O and thus is prevented from running.

High simulation wait time can indicate performance problems in connectivity functions, loss of hardwareassists, or other simulation-related bottlenecks. When CP is simulating a hardware function such asinstructions, interrupts, timer updates, or unique z/VM functions such as IUCV, the virtual processor isprevented from running and is in this state.

A virtual machine enters page wait state when it references a page that is not present in host storage andmust be brought in from auxiliary storage. This state reflects time waiting for page reads from DASD. Pagewrites and paging from expanded storage do not affect this state. High page wait time indicates thestorage and paging subsystem need evaluation. If key server virtual machines are in page wait, you mightwant to consider reserving pages for these machines.

Certain functions in CP are serialized by a state called console function mode. While in this wait state, thevirtual processor is not permitted to run. This state might include such things as certain CP commands.High console function wait can indicate problems in the network, excessive CP command usage, andpossibly master processor contention.

CP might place a user who has gone idle, or is waiting for what is expected to be a short-term wait, intoTest Idle wait. While in test idle, the user is still in the dispatch list. This state can last up to approximately300 milliseconds. If no new work arrives for the virtual machine during the test idle period, the virtualmachine is dropped from the dispatch list and added to the dormant list. Using this approach, much of theoverhead of dropping and adding users to the dispatch list can be avoided. A high amount of time in testidle is not necessarily bad. There are actually two types of test idle states; one where the user has anoutstanding communication with another virtual machine (SVM wait) and one where there is not an SVMwait.

A virtual processor is in the I/O active state if there is an asynchronous I/O outstanding causing the user tobe in the dispatch list, but not prevented from running. It is normal for some virtual machines to have highI/O active wait times in cases where they use continuous channel programs. This situation can occur withsome network-related machines. Virtual machines with increased I/O active wait time can indicateperformance problems in networks or other I/O done asynchronously.

Active Page Wait is similar to active I/O wait, except instead of an outstanding asynchronous I/O, there isan outstanding page fault that the virtual machine was prepared to handle asynchronously. If a user isready to run but has exceeded the maximum share setting, the user might be placed on the limit list for aperiod of time. During that time, the user is in the limit list wait state.

The following shows the formula for the ZVM_Eligible_List_High situation:

*IF *VALUE KVLSystem.Eligible_Users > 5


Chapter 5. Take Action commands

The Take Action feature available with this monitoring agent allows you to execute CP commands, CMScommands, and REXX executables on the z/VM operating system. The CMS guest defined for Take Actionoperates as a dedicated, unattended, disconnected virtual machine that waits to process messagescontaining Take Action commands sent from the monitoring agent.

You can use the Take Action view to perform the following actions:

v Send commands to the Command Processor or to another operations automation tool, such asOperations Manager.

v Select a predefined action.

v Start a process, such as a user-defined CMS executable, a REXX executable, or a program.

v Issue a CP FORCE command or issue a CP SMSG command, where supported.

The predefined actions available depend upon where the window was invoked. The Take Action view canalso be selected from the Tivoli Enterprise Portal toolbar. See the Tivoli Enterprise Portal online help foradditional details about this feature.

To enable Take Action commands, the following requirements must be met:

Note: The following requirements are optional. They are required only if you intend to use the Take Actionfeature to issue CP commands, CMS commands, and REXX executables from the Tivoli Enterprise Portalto z/VM.

v The Take Action Command Processor REXX executable (KVLCMD EXEC) that ships with OMEGAMONXE on z/VM and Linux must be installed and started on the service guest virtual machine beforesubmitting Take Action commands at the Tivoli Enterprise Portal.

v A Conversational Monitor System (CMS) guest system must be configured to receive commands fromthe Tivoli Enterprise Portal. This guest system must be disconnected and must reside on the samemanaged system as the monitoring agent. The guest system receives command requests from the TivoliEnterprise Portal through the z/VM monitoring agent.

v sudo, the superuser do utility for Linux-based systems, is required for the monitoring agent to run withtemporary root authority to send Take Action commands to the Command Processor virtual machine.This allows a non-root Linux guest user to run vmcp without having root authority.

v The file called KVL CONFIG that ships with this monitoring agent must be modified to include theallowed user IDs for the Linux guests where the monitoring agent is running. This file also contains a listof commands that are to be excluded from running, and it lets you specify the size of the log file.

v When you issue a Take Action command, your user ID must be authorized on the relevant system forthe requested command.

See the Program Directory for installation information. For configuration information, see the IBM TivoliOMEGAMON XE on z/VM and Linux Planning and Configuration Guide. See also the z/VM: CMSCommands and Utilities Reference and the z/VM: CP Commands and Utilities Reference guides.

Issuing Take Action commandsTo issue a Take Action command from a workspace, right-click anywhere in the workspace, select TakeAction, then Select.

The Take Action window enables you to select a predefined action or enter an action command yourself.The actions available depend on where the window was invoked.


Name of the action command. In the Name list, click the action of your choice or leave it Select action ifyou are entering the command yourself. For this monitoring agent, select Issue z/VM Commands from theName drop-down list.

Command is the command to send to the application or operating system on which it is running, and islimited to 512 bytes. Commands for this monitoring agent are prefixed with VL:.

Destination System A list of possible target systems is displayed here after you enter a command in theCommand field. Click an application host in the list. If you want to send the command to multiple systems,use Ctrl+click to select more or Shift+click to select all systems between.

For information on enabling Take Action commands, see the IBM Tivoli OMEGAMON XE on z/VM andLinux Planning and Configuration Guide.


Part 3. Monitoring scenarios

This chapter describes some common usage scenarios that might be similar to your experience in usingOMEGAMON XE on z/VM and Linux to monitor and to manage your environment.


Chapter 6. Monitoring scenarios

OMEGAMON XE on z/VM and Linux provides predefined workspaces and situations to help you startmonitoring your z/VM and Linux on zSeries systems and resources very quickly. As you become morefamiliar with the product, you can modify the workspaces and situations to meet the specific needs of yourenterprise. This chapter describes some common usage scenarios, using product-provided situations, thatmight be helpful to you in using OMEGAMON XE on z/VM and Linux to monitor and to manage yourenvironment. You can modify the predefined situation to meet the specific conditions in your environment.Be sure to save the original situation and use Create Another Situation. See the Tivoli Enterprise Portalonline help for details on creating situations.

The monitoring scenarios are addressed to the z/VM system programmer or system administrator who isresponsible for ensuring the availability of z/VM system resources. The scenarios are also addressed tothe Linux on System z programmer or system administrator who is responsible for ensuring the availabilityof systems running on Linux guests.

Solving a DASD problemIn this scenario, you identify the devices with the highest I/O.

1. When you log on to the Tivoli Enterprise Portal, you see that the Enterprise and z/VM Systems nodesof the Navigator are overlaid by a critical situation event indicator.

2. Move your cursor over the critical situation event indicator to see a list of situation events on yoursystems. The predefined situation “ZVM_DASD_Queue_Critical” on page 71 is one of the situationslisted.

3. To see the conditions tested by the predefined situation, open the Situation Editor by clicking its iconin the toolbar. From the list that displays, select ZVM_Linux_Systems.

4. Double-click ZVM_DASD_Queue_Critical. This situation tests only one condition. This situation raisesan alert when the average number of I/O requests queued on a device is greater than or equal to 25requests, during the collection interval.

5. To view the formula for this situation, click the Show Formula icon. The formula for this situationappears as:

6. To examine and to monitor the performance of the devices connected to your system, expand thez/VM Linux Systems node and select the “DASD workspace” on page 37. This workspace providesseveral views that display the activity of the busiest input and output (I/O) devices on the overall z/VMsystem. You can use this workspace to determine, during the collection interval, the devices with theworst connection times, the average number of I/O requests queued to a device, and the averageservice time for each device.

7. To view the status of the five busiest devices in your enterprise, look at the data in the Top 5 DeviceBusy bar chart.

Figure 33. Navigator with critical and warning event indicators

Figure 34. ZVM_DASD_Queue_Critical situation formula

Figure 35. Top 5 Device Busy bar chart


8. Use this bar chart to identify the five busiest devices. For each device, the chart shows the percentageof time a device was busy.

9. Additionally, the DASD workspace contains the DASD I/O Activity table. Scroll through the pages ofthis table until you find the metrics for the device that you are monitoring. In this example, the AverageQueued IO attribute displays the metrics in red, indicating that a critical threshold was reached orexceeded.

If a device has excessive I/O queuing, consider rebalancing the I/O load across lightly loaded devices,if possible. Device queuing may also be caused by excessive non-zero seeking.

Refer to the DASD Seeks Distances (SEEKDIST) screen and to the (DASD Seeks Locations(SEEKLOC) screen in the Performance Toolkit for help with analyzing seek operations by DASDdevice.

Sharing DASD devices between separate systems can also cause device contention and queuing.Refer to the guest systems performance monitor for queuing information on dedicated devices.

Solving a paging data set problemIn this scenario, you identify a paging data set utilization problem.

1. When you log on to the Tivoli Enterprise Portal, you see that the Enterprise and the z/VM Systemsnodes of the Navigator are overlaid by a critical situation event indicator.

2. Move your cursor over the critical situation event indicator to see a list of situation events on yoursystems. The predefined situation “ZVM_Page_Used_Critical” on page 74 is one of the situationslisted. This situation raises an alert when the percentage of any one of the allocations listed below isgreater than or equal to 95 percent:




3. To see the status of paging space utilization in your enterprise, expand the z/VM Linux Systems nodeand select the “CP Owned Devices workspace” on page 33.

This workspace contains the Top 5 Page Extent Utilization bar chart that shows, by device address,the five devices using the greatest amount of paging space. For each paging device, it shows thepercentage of space in use on the volume managed by the Control Program during the collectioninterval.

This workspace also contains the Paging and Spooling Space bar chart that displays the top tendevices using the greatest amount of paging and spooling space. For each device managed by thez/VM Control Program, it shows the number of cylinders or blocks allocated, and the number of slotsavailable on the device.

Figure 36. Average Queued IO attribute


Figure 38. Top 5 Page Extent Utilization bar chart


4. Review these bar charts to identify those systems that are using the most paging space.

If the situation triggers because of a spool allocation, spool files or overflow paging may be consuming alarge percentage of the available space. For detailed information, refer to the SPOOL File Summary(SPOOL) screen of the Performance Toolkit. The SPOOL screen shows how many disk blocks areconsumed by each file. Use this as a guide when determining which spool files would yield the most freespace if deleted. You might want to consider adding more space if spool utilization is generally high. Youcould also offload spool using the SPXTAPE command with the PURGE option if the files need to beretained.

If the situation raises for a paging allocation, available paging space may be dangerously low. Consideradding additional paging space or reducing the active workload on the system by logging off guestsystems that are not critical.

Refer to the attribute Working Set Size in the “Workload workspace” on page 57 to determine whichworkloads are consuming the most real storage. See also the User Page Data (UPAGE) screen of thePerformance Toolkit. This screen provides working set size data and the count of pages that are locked orreserved for workloads. To avoid raising this exception unnecessarily, consider modifying the threshold forthis situation if it is not appropriate for your environment.

Monitoring Control Program (CP) CPU utilizationIn this scenario, you identify the five biggest users of functions that are causing CP to process supervisorcalls, and thus using up a high amount of CPU.


2. Move your cursor over the critical situation event indicator to see a list of situation events on yoursystems. The predefined situation “ZVM_CP_CPU_Critical” on page 69 is one of the situations listed.

This situation raises an alert when the percentage of CPU used by the z/VM Control Program (CP) isgreater than or equal to 30 percent. The percentage includes both supervisor time spent for users andsupervisor time spent for system services. The value reported is the average of all the %CPU valuesreported for all processors assigned to the logical partition that are displayed on the CPU screen of thePerformance Toolkit.

3. To see the status of CPU utilization by CP in your enterprise, expand the z/VM Linux Systems nodeand select the “Workload workspace” on page 57. The Workload workspace contains a table called Allz/VM Workloads that provides data such as the total number of CPU seconds used by each workload,the percentage of total CPU used by the z/VM Control Program to manage the workload, and the totaltime each workload or each aggregation of the group was logged on. It also contains a bar chart calledTop 5 CPU Users.

Figure 39. Paging and Spooling bar chart


Figure 41. List of situation events - flyover pop-up

Figure 42. Top 5 CPU Users bar chart

Chapter 6. Monitoring scenarios 95

4. Use this bar chart to identify the five biggest users of functions that are causing CP to processsupervisor calls.

You might want to check the load on the paging subsystem. CP may be busy processing a high number ofpage faults caused by guest operating systems. Consider modifying the threshold for this situation if highCP CPU utilization is normal in your environment. You can also review the General CPU screen of thePerformance Toolkit.

Monitoring LPAR CPU utilizationIn this monitoring scenario, you identify those logical partitions that are using a large amount of CPU.


2. Move your cursor over the critical situation event indicator to see a list of situation events on yoursystems. The predefined situation “ZVM_LPAR_Busy_Critical” on page 71 is one of the situationslisted. This situation raises an alert when the CPU utilization for a logical partition is greater than orequal to 90 percent. The triggering of this situation indicates that there may be a problem with aworkload in this partition consuming excessive processor resources. This could be caused by a loopingvirtual machine, or by the normal contention for processor resources due to workload. In the lattercase, consider modifying the sample threshold value to match your environment.

3. To see the status of CPU utilization by the logical partitions defined in your enterprise, expand thez/VM Linux Systems node and select the “LPAR workspace” on page 44. The LPAR workspacecontains a table called LPAR Utilization that shows data such as the total amount of time that all ofthe logical processors for a logical partition were busy during the collection interval, the percentage ofelapsed time that a logical processor spent to manage LPARs, and the utilization of the system basedon the number of logical processors available. Use this table to determine how well each logicalpartition is able to obtain CPU resources. It also contains a bar chart called LPAR Load. This bar chartdisplays data on the relative load of a logical partition on the whole processor complex. Depending onthe number of processors defined to a logical partition, the total percentage is calculated based on 100percent multiplied by the number of assigned processors.

4. Use this bar chart to identify the logical partitions that are using a large amount of CPU. You can alsouse the “Workload workspace” on page 57 to identify the workloads that are using a large percentageof CPU. If the source is found to be a guest operating system, use the performance tools for thatsystem to determine the cause.

Monitoring system utilization by Linux on System z guestsIn this scenario, you focus in on the system utilization of Linux guests in your enterprise.


2. Move your cursor over the critical situation event indicator to see a list of situation events on yoursystems. The predefined situation “ZVM_Total_CPU_Critical” on page 79 is one of the situations listed.

This situation raises an alert when the percentage of CPU utilized by both the Control Program(overhead) and the guest virtual machine workload is greater than or equal to 90 percent. The valuereported is the sum of all the percent CPU values reported for all processors assigned to the logicalpartition.

3. To see the system utilization of Linux guests in your enterprise, expand the z/VM Linux Systems nodeand select the “Workload workspace” on page 57. The Workload workspace contains links to two other

Figure 43. LPAR Load bar chart


OMEGAMON XE on z/VM and Linux workspaces. From the Navigator, right-click the name of theWorkload workspace, select Workspace, and select either the “Linux Workload workspace” on page58 or the “ApplData workspace” on page 59.

4. The “Linux Workload workspace” on page 58 provides data about system usage by user ID for allLinux guest systems defined in your environment. Among other panels, this workspace contains theTop 5 CPU Linux Guest Systems bar chart.

5. Use this bar chart to identify the top five Linux guest users of CPU. For each Linux guest, it shows thepercentage of CPU used by the system to manage the workload, and the percentage of virtual CPUutilized. Other bar charts display the guest system paging rates and working set sizes.

6. The “ApplData workspace” on page 59 contains tables that display extensive metrics about systemutilization by the guest virtual machines. Review these tables to obtain data such as the percentage ofCPU used by this Linux virtual machine, running in kernel mode, the percentage of soft interrupts, thetotal size of the main memory, in megabytes, and the size of the memory that is used for buffers, inmegabytes.

Note: Data collection for the attribute group from which this workspace is built must be enabled ateach Linux guest system that is to be monitored. In addition, you will need to enable data collection forthe associated CP Monitor domains at the z/VM system where the Performance Toolkit is running. See“ApplData workspace” on page 59 for details. See also the IBM Tivoli OMEGAMON XE on z/VM andLinux Planning and Configuration Guide.

Both the “Linux Workload workspace” on page 58 and the “ApplData workspace” on page 59 contain linksto workspaces provided by the Tivoli Monitoring Agent for Linux OS. You can access these workspaces ifyou have the Tivoli Monitoring Agent for Linux OS agent installed and configured. To access theworkspaces, right-click the link icon from the Linux Guest System Workloads table of the Linux Workloadworkspace, or right-click the link icon from the Linux Guest Workload Data table of the ApplDataworkspace. To find out details about the linked workspaces that provide additional Linux on zSeries data,refer to the documentation for the Tivoli Monitoring Agent for Linux OS.

See also the General CPU screen on the Performance Toolkit. This screen displays the percentage ofCPU utilization by individual processor. Since overhead is included in the total CPU percent, the ControlProgram may be contributing to high CPU. You might want to check the load on the paging subsystem.The Control Program may be busy processing a high number of page faults caused by guest operatingsystems. It is possible that the high CPU utilization is caused by a user application or a process stuck in aloop. The application in question should be examined for inefficient or defective code. It is also possiblethat the total CPU is high simply because there is legitimate processing work to be done that is consumingall available CPU. A processor upgrade may be needed. Consider modifying the threshold for this situationif high total CPU utilization is normal in your environment.

Monitoring control unit cache activityIn this scenario, you identify the cached control unit devices with the longest response times.

1. When you log on to the Tivoli Enterprise Portal, you see that the Enterprise and z/VM Systems nodesof the Navigator are overlaid by a warning situation event indicator.

2. Move your cursor over the warning situation event indicator to see a list of situation events on yoursystems. The predefined situation “ZVM_CUCache_DataResp_High” on page 69 is listed, indicating

Figure 44. Top 5 CPU Linux Guest Systems bar chart

Figure 45. Navigator with warning event indicators


the name of the affected managed system, and the date and time the situation was triggered.

3. To see the conditions tested by the predefined situation, you open the Situation Editor by clicking itsicon in the toolbar. From the list that displays, select ZVM_Linux_Systems.

4. Double-click ZVM_CUCache_DataResp_High. This situation tests only one condition. This situationraises an alert when the average response time for completing an I/O operation to a device on acached control unit exceeds 50 milliseconds, during the collection interval.

5. To view the formula for this situation, click the Show Formula icon. You can modify the predefinedsituation to meet the specific conditions in your environment. Be sure to save the original situation anduse Create Another Situation. See the Tivoli Enterprise Portal online help for details on creatingsituations.

6. To examine and to monitor the performance of the cached control units, expand the z/VM LinuxSystems node and select the “DASD workspace” on page 37. This workspace contains links toseveral secondary workspaces that provide additional data on DASD activity.

From the Navigator, select the DASD workspace, select Workspace, then select the “Control UnitCache workspace” on page 40.

Among other views, the Control Unit Cache workspace contains the Top 5 Cache Control Unit LoadData Time bar chart. Use this bar chart to view, for the five busiest control units, by subsystemidentifier of the control unit, the device response time, in milliseconds. This is the device service time,plus the time during which an I/O request was waiting to be started. This metric is calculated based onthe average I/O request queue length and the I/O rate.

7. Additionally, the “Control Unit Cache workspace” on page 40 contains the Top 5 Cache Control UnitLoad Busy Percent table. Scroll through the pages of this table until you find the metrics for thecontrol unit that you are monitoring. You may want to look at the metrics for the size of configuredcache storage, the amount of cache storage that is available, and the average busy percentage for allconnected disks.

A long response time indicates that a device is experiencing prolonged delays due to high activity.Consider rebalancing the I/O load across lightly loaded devices, if possible.

Sharing DASD devices between separate systems can also cause device contention and queuing. Whensearching for a cause with control units shared across multiple systems, combined utilization must beconsidered. Disk performance can be improved by minimizing I/O rates.

If long response times are encountered on devices that are used for paging, system performance may besignificantly degraded. Consider making additional memory available to CP to reduce paging. SpreadCP-owned allocations across as many devices as possible and use dedicated DASD devices. See the “CPOwned Devices workspace” on page 33 for additional information on the location and utilization of CPpaging space. See also the CP Owned Device (DEVICE CPOWNED) and the CPOWNED Device Log(CPOWNLOG devno) of the Performance Toolkit for additional information on paging space activity. Referto the Performance Toolkit Reference for details.

Monitoring virtual disk (VDISK) I/O activityIn this scenario, you identify the cached control unit devices with the longest response times.

Figure 46. Situation event flyover pop-up

Figure 47. Top 5 Cache Control Unit Load Data Time bar chart


1. When you log on to the Tivoli Enterprise Portal, you see that the Enterprise and z/VM Systems nodesof the Navigator are overlaid by a warning situation event indicator.

2. Move your cursor over the warning situation event indicator to see a list of situation events on yoursystems. The predefined situation “ZVM_VDISK_Page_High_IORate_Low” on page 82 is listed,indicating the name of the affected managed system, and the date and time the situation wastriggered.

3. To see the conditions tested by the predefined situation, you open the Situation Editor by clicking itsicon in the toolbar. From the list that displays, select ZVM_Linux_Systems.

4. Double-click ZVM_VDISK_Page_IORate_Low. This situation triggers when the“ZVM_Page_Queue_High” on page 73 situation has evaluated to TRUE for 10 consecutive samples.ZVM_Page_Queue_High raises an alert when the percentage of in-queue users in the page wait stateexceeds the specified value. A virtual machine that is waiting for a page cannot perform work until thepage request is satisfied. If numerous virtual machines are waiting for pages, system performance maybe adversely affected.

5. To view the formula for this situation, click the Show Formula icon.

6. To examine and to monitor the performance of the virtual disks, expand the z/VM Linux Systemsnode and select the “DASD workspace” on page 37. This workspace contains links to severalsecondary workspaces that provide additional data on DASD activity.

From the Navigator, select the DASD workspace, select Workspace, then select the “VDISKworkspace” on page 43.

Among other views, the VDISK workspace contains the Top 5 Paging Rates per Second bar chart.This bar chart displays, for the five busiest virtual disks, by user ID of the owner of the VDISK and bydevice address, the number of pages read from DASD per second, the number of pages stolen persecond, and the number of pages written to DASD per second.

If this situation triggers often, consider detaching the VDISK (if possible) to release the page space itconsumes back to CP for use in general paging. It is possible that the low level of VDISK I/O and the highpaging rate are transitory.

In instances where you are using a VDISK for a Linux swap device, there may be benefit in creatingmultiple smaller VDISKS for swapping, instead of one large VDISK. The pages associated with virtual diskblocks are not created until they are actually referenced.


Figure 48. Navigator with warning event indicators

Figure 49. Top 5 Paging Rates per Second bar chart


Attributes reference

Attributes are characteristics or properties of the objects monitored by OMEGAMON XE on z/VM andLinux. Related attributes are organized into attribute groups (also called attribute tables). The attributes areused to define the queries that collect the information displayed in tables and charts in the OMEGAMONXE on z/VM and Linux workspaces and to create situations that trigger alerts in response to specifiedconditions.

The table view in each workspace displays data collected for a single attribute group. See “Attributegroups used by predefined workspaces” on page 29 for a tabular representation of the relationshipsbetween the predefined workspaces and the attribute groups.

Attribute namesEvery attribute is identified by a unique name composed of the attribute group name followed by a periodand the name of the attribute item. For example, the attribute CP % of CPU a member of theKVLUser_Workload group, stores a value representing the percentage of total CPU used by the z/VMControl Program to manage this workload at the time of the data collection.KVLUser_Workload.CP % of CPU

Attributes by attribute groupThe rest of this chapter lists the attributes (in alphabetical order by attribute group) for the OMEGAMONXE on z/VM and Linux product. You can use these attributes to monitor system performance, to buildcustom workspaces, and to create situations that alert you of pending problems.

KVLCP Device attributesThe KVLCP Device attribute group provides information about the devices managed by the z/VM ControlProgram. Use the KVLCP Device attributes to create situations that monitor these devices.

The KVLCP Device attribute group contains the following attributes:

Allocation The number of cylinders or blocks allocated for the extent managed by the z/VM ControlProgram. The value format is an integer.

Available Slots The number of slots available on the device managed by the z/VM Control Program at thetime the sample was taken. The value format is an integer.

Note: If the monitoring agent cannot determine the type of device being used, a zero value displays inthis column. This most likely means that monitoring has not been turned on for the device being managedby the z/VM Control Program.

Device Address The real address of the storage device managed by the z/VM Control Program. Thevalue format is an alphanumeric text string with a maximum of 4 characters.

Device End Extent The ending cylinder or block number allocated on the device managed by the z/VMControl Program. The value format is an integer.

Device Percent Full The percentage of space on the volume in use at the time the sample was taken.The value format is an integer.

Device Slots Used The number of used slots on the device managed by the z/VM Control Program at thetime the sample was taken. The value format is an integer.


Note: If the monitoring agent cannot determine the type of device being used, a zero value displays inthis column. This most likely means that monitoring has not been turned on for the device being managedby the z/VM Control Program.

Device Start Extent The starting cylinder or block number allocated on the device managed by the z/VMControl Program. The value format is an integer.

Device Type The type of device managed by the z/VM Control Program. The value format is analphanumeric text string with a maximum of 8 characters.

Note: If the monitoring agent cannot determine the type of device being used, UNKNOWN displays in thiscolumn. This most likely means that monitoring has not been turned on for the device being managed bythe z/VM Control Program.

On z/VM, to turn on monitoring for that device, issue the following Control Program command:CP MONITOR SAMPLE ENABLE I/O DEVICE rdev

where rdev is the device number that you want to enable. See the z/VM: CP Commands and UtilitiesReference for information on using CP commands.

Device VOLSER The volume serial number for the device managed by the z/VM Control Program. Thevalue format is an alphanumeric text string with a maximum of 8 characters.

LPAR Name The name assigned to the logical partition. The value format is an alphanumeric text stringwith a maximum of 8 characters.

PAGING SPOOLING The primary purpose of the allocation. The following options are possible:

v PAGING. The area on the disk reserved for CP page storage.

v SPOOLING. The area on the disk reserved for CP spool files.

v DUMP. The area on the disk reserved for CP dumps.

v DIRECT. The area on the disk reserved for the CP directory.

v T-DISK. The area on the disk reserved for temporary disk allocation.

v MDISK. The area on the disk reserved for minidisk cache.

v UNKNOWN. Device allocation usage is unknown.

v NOT-FOUND. Device allocation usage is not recognized.

System ID The name that uniquely identifies the active z/VM system. The value format is an alphanumerictext string with a maximum of 8 characters.

System Name The host name of the system on which the agent is running. The value format is analphanumeric text string no longer than 64 characters.

Time The date and time at the end of the CP Monitor sample interval. This attribute is intended for loggingand reporting data collection times rather than for creating situations. See the Tivoli Enterprise Portal helpfor instructions on specifying timestamp attributes.

See also:

“Attribute groups used by predefined workspaces” on page 29

KVLChannel Data attributesThe KVLChannel Data attribute group provides information about channel data, one line per channel.Overall channel load analysis is based on data from the Extended Channel-Path Measurement Facility,


where available; otherwise high-frequency sampling data from the CP monitor SYSTEM domain is used.The workspace associated with this attribute group displays data for all channels for which validinformation was found. Use the KVLChannel Data attributes to create situations that monitor channelactivity.

The KVLChannel Data attribute group contains the following attributes:

Average Busy Conditions Percent The average percentage of busy conditions found, calculated for theentire period since the last RESET, or since the Performance Toolkit was last invoked. Channel busypercentages cannot be calculated for HiperSockets channels. The value format is in the range of 0.00 -100.00. For example, the value 12.00 represents 12.00%.

CHPID In a channel subsystem, a value assigned to each installed channel path of the system thatuniquely identifies that path to the system. The value format is an alphanumeric text string no longer than4 characters.

Channel Busy Distribution 0-10 Percent The channel load distribution for each of the active channels. Itshows the relative frequency with which the channel busy percentage of separate measurement intervalswas found in the range 0 - 10. The value format is in the range of 0.00 - 100.00. For example, the value12.00 represents 12.00%.

Channel Busy Distribution 11-20 Percent The channel load distribution for each of the active channels.It shows the relative frequency with which the channel busy percentage of separate measurement intervalswas found in the range 11 - 20. The value format is in the range of 0.00 - 100.00. For example, the value12.00 represents 12.00%.







Attributes reference 103



Group Qualifier The hexadecimal channel model group qualifier. Note that a valid qualifier is availableonly on systems with the extended channel measurement facility. The value format is a hexadecimalidentifier no longer than 4 characters.

Interval Busy Conditions Percent The percentage of ’busy’ conditions found, calculated for the lastinterval. Channel ’busy’ percentages cannot be calculated for HiperSocket channels. The value format is inthe range of 0.00 - 100.00. For example, the value 12.00 represents 12.00%.


Model Group Description of the channel model group. The value format is an alphanumeric text string nolonger than 8 characters.




See also:


KVLControlUnit attributesThe KVLControlUnit attribute group provides overall load information on cached control units. This groupprovides data on the use and on the effectiveness of the cache. Data displays for devices that have beenactive during the last measuring interval. Use the KVLControlUnit attributes to create situations thatmonitor control unit cache activity.

Average Connect Time msec The average connected time, in milliseconds. This is the average time thatthe device was connected to the channel while executing I/O requests. That is, while the device wasreceiving commands or transferring data. This metric includes the SEARCH time. The value format is adecimal number, with one digit to the right of the decimal point. An example is 14.6.

Average Disconnect Time msec The average disconnected time, in milliseconds. This is the averagetime that a device remained disconnected from the channel while executing I/O requests. For DASD, thisincludes the time spent for executing the SEEK and SET SECTOR orders and also any re-connectmisses. High values may indicate overloaded paths. High values may also indicate many long SEEKs to


data at opposite ends of the device. The value also includes the ’device-active-only’ time where thatinformation is available. The value format is a decimal number, with one digit to the right of the decimalpoint. An example is 14.6.

Average Pending Time msec The average function pending time, in milliseconds. This is the averagetime an I/O request remained pending in the I/O subsystem due to path busy conditions. The value formatis a decimal number, with one digit to the right of the decimal point. An example is 14.6.

Average Read Hits Percent The average percentage of read hits. The value format is an integer.

Average Response Time msec The average response time, in milliseconds, for the device. This is thedevice service time, plus the time during which an I/O request was waiting to be started. It is calculatedbased on the average I/O request queue length and the I/O rate. The value format is a decimal number,with one digit to the right of the decimal point. An example is 14.6.

Average Service Time msec The average service time, in milliseconds. That is, the sum of functionpending, disconnected, and connected time. The value format is a decimal number, with one digit to theright of the decimal point. An example is 14.6.

Average Write Hits Percent The average percentage of write hits, for DASD and cache fast writeoperations. The value format is an integer.

Busy Percent The average busy percentage for all connected disks. The value format is an integer.

Cache I/O Rate The total I/O activity as it is recorded by the cache control unit. That is, where multiplesystems are connected to one control unit, the total I/O activity from all systems is shown. The valueformat is a decimal number, with one digit to the right of the decimal point. An example is 14.6.

Cache Fast Write Percent The average percentage of cache fast-write operations, based on the sum ofall write operations. The value format is an integer.

Cache Fast Write Hit Percent The average percentage of cache fast write hits. The value format is aninteger.

Control Unit Type The type and model of the control unit. The value format is an alphanumeric text stringwith a maximum of 8 characters.

DASD Fast Write Hit Percent The average percentage of DASD fast write hits. The value format is aninteger.

DASD Fast Write Percent The average percentage of DASD fast-write operations, based on the sum ofall write operations. The value format is an integer.

Fast Write Rate per Second The total fast-write rate per second. The value format is a decimal number,with one digit to the right of the decimal point. An example is 0.5.


Nonsequential Read Rate per Second The total nonsequential read rate per second. The value format isa decimal number in the range of 0.0 through 100.0, with one digit to the right of the decimal point. Anexample is 10.2.

Nonvolatile Storage Available Mbytes The amount of nonvolatile storage that is available, in megabytes.The value format is an integer.


Nonvolatile Storage Configured Mbytes The size of configured nonvolatile storage, in megabytes. Thevalue format is an integer.

Read Rate Percent The average percentage on read operations, based on the sum of read and write.The value format is a decimal number, with one digit to the right of the decimal point. An example is 14.6.

Sequential Read Rate per Second The total sequential read rate per second. The value format is adecimal number, with one digit to the right of the decimal point. An example is 14.6.

Subchannel Measurement Block I/O Rate The total I/O activity as determined from count fields in thesubchannel measurement blocks of the system that does the monitoring. That is, only the I/O requestsfrom that particular system are shown. The value format is a decimal number, with one digit to the right ofthe decimal point. An example is 14.6.

Subsystem ID The subsystem identifier for the control unit. The value format is an alphanumeric textstring with a maximum of 4 characters.




Volatile Storage Available Mbytes The amount of cache storage that is available, in megabytes. Thevalue format is an integer.

Volatile Storage Configured Mbytes The size of configured cache storage, in megabytes. The valueformat is an integer.

See also:


KVLDASDCache attributesThe KVLDASDCache attribute group provides performance data for disks connected to a cached 3880-13control unit, to a cached 3880-23 control unit, or to an IBM DASD subsystem. Use the KVLDASDCacheattributes to create situations that monitor the effectiveness of the cache and how it is being used.

Base PAV Address The base address for the Parallel Access Volume (PAV). The value format is analphanumeric text string with a maximum of 4 characters. This attribute is blank when the PAV address isnot available.

Bypass and Inhibit Cache Load Percent The percentage of I/O operations that bypassed cachingvoluntarily. The value format is in the range of 0.00 - 100.00. For example, the value 12.00 represents12.00%.

Bypass Cache Requests per Second The number of bypass cache requests per second. The valueformat is a decimal number, with one digit to the right of the decimal point. An example is 14.6.

Cache Fast Write Percent The percentage of cache fast write operations, based on total write activity.The value format is an integer.


Cache Fast Write Hits Percent The percentage of cache fast write hits, based on the sum of all cachefast-write operations. The value format is an integer.

Cache to DASD Transfers Percent The percentage of cache to DASD transfer operations. The valueformat is an integer.

Cache Read Percent The percentage of cachable read operations, based on total I/O activity. The valueformat is an integer.

Cache Read Hits Percent The percentage of read hits, based on the sum of all cachable read operations.The value format is an integer.

Cache Total Percent The percentage of total hits (read + DASD fast write + cache fast write), based onthe sum of all cachable read and write operations. The value format is an integer.

Cache Write Hits Percent The percentage of write hits (DASD fast write + cache fast write), based on thesum of all DASD and cache fast write operations. The value format is an integer.

Caching Status The caching status. Possible options are:

v ACTIVATED. Caching is activated.

v DEACTIVATED. Caching is deactivated.

v DEACTIVATE PENDING. Deactivate is pending.

v UNKNOWN. Not able to determine the caching status.

Controller ID The storage controller subsystem identifier. The value format is an alphanumeric string witha maximum of four characters.

DASD Fast Write Bypass Percent The percentage of DASD fast write operations that were forced tobypass the cache and access DASD directly due to nonvolatile storage constraints. The value format is aninteger.

DASD Fast Write Hits Percent The percentage of DASD fast write hits, based on the sum of all DASDfast-write operations. The value format is an integer.

DASD Fast Write Percent The percentage of DASD fast write operations, based on total write activity.The value format is an integer.

DASD Fast Write Status The status of DASD fast write. Possible options are:

v ACTIVATED. DASD fast write is activated.

v DEACTIVATED. DASD fast write is deactivated.

v DEACTIVATE PENDING. DASD fast write deactivate is pending.

v UNKNOWN. Status of DASD fast write is not known.


Device Type The type and model of the device managed by the z/VM Control Program. If the device is aParallel Access Volume (PAV) device, PAV displays. The value format is an alphanumeric text string with amaximum of 8 characters.

Note: If the monitoring agent cannot determine the type of device being used, UNKNOWN displays in thiscolumn. This most likely means that monitoring has not been turned on for the device being managed bythe z/VM Control Program.


On z/VM, to turn on monitoring for that device, issue the following CP command:CP MONITOR SAMPLE ENABLE I/O DEVICE rdev

where rdev is the device number that you want to enable. See the z/VM: CP Commands and UtilitiesReference for information on using CP commands.

Dual Copy Indicator The dual copy indicator. Possible options are:

v NO DUPLEX. Dual copy has been suspended.

v PRIMARY. Disk is the primary device of a dual copy pair.

v SECONDARY. Disk is the secondary device of a dual copy pair.

v UNKNOWN. Status is not known.

Dual Copy Status The dual copy status. Possible options are:

v DUPLEX PAIR AVAILABLE. Duplex pair is available.

v DUPLEX PAIR PENDING. Duplex pair is pending.

v FAILED DUPLEX. Failed duplex.

v SUSPENDED DUPLEX. Duplex has been suspended.

v DUPLEX NOT ACTIVE. Duplex is not active.

v UNKNOWN. Status is not known.

Fast Write I/O Rate per Second The number of fast write I/O requests per second over this storagedirector. The value format is a decimal number, with one digit to the right of the decimal point. An exampleis 14.6.

Fast Write Rate per Second The total fast-write rate per second (cache fast write + DASD fast write).The value format is a decimal number, with one digit to the right of the decimal point. An example is 14.6.

Fast Write Read Percent The percentage of fast-write read requests. The value format is an integer.

Fast Write Read Hit Percent The percentage of fast-write read requests that did not need DASD access.The value format is an integer.

I/O Requests The sum of normal, sequential, and fast write (for IBM DASD subsystems only) I/Orequests, per second (read + write). The value format is a decimal number, with one digit to the right ofthe decimal point. An example is 14.6.

Inhibit Cache Load Requests per Second The number of inhibit cache load requests per second. Thevalue format is a decimal number, with one digit to the right of the decimal point. An example is 14.6.


Nonsequential DASD to Cache Transfers Percent The percentage of nonsequential DASD to cachetransfer operations. The value format is an integer.

Nonsequential Read Rate per Second The nonsequential read rate (read normal + read cache fastwrite), per second. The value format is a decimal number, with one digit to the right of the decimal point.An example is 14.6.

Normal I/O Rate per Second The number of normal I/O requests, per second, over this storage director.The value format is a decimal number, with one digit to the right of the decimal point. An example is 14.6.

Normal Read Percent The percentage of normal read requests. The value format is an integer.


Normal Read Hit Percent The normal read percentage. That is, the percentage of read requests that didnot need DASD access. The value format is an integer.

Read Percent The overall percentage of read requests (normal, sequential, and fast write) to the total I/Oactivity of the device. The value format is an integer.

Read Hit Percent The overall percentage of read hits. That is, the percentage of read requests that didnot need DASD access (normal, sequential, and fast write). The value format is an integer.

Sequential DASD to Cache Transfers Percent The percentage of sequential DASD to cache transferoperations. The value format is an integer.

Sequential I/O Rate per Second The number of sequential I/O requests, per second, over this storagedirector. The value format is a decimal number, with one digit to the right of the decimal point. An exampleis 14.6.

Sequential Read Percent The percentage of sequential read requests. That is, the percentage of readrequests that did not require DASD access. The value format is an integer.

Sequential Read Hit Percent The percentage of sequential read hits. The value format is an integer.

Sequential Read Rate per Second The sequential read rate (read sequential) per second. The valueformat is a decimal number, with one digit to the right of the decimal point. An example is 14.6.

Storage Director ID 3880-13/23 storage director ID. For IBM DASD subsystems, the last two hexadecimaldigits of the service set identifier (SSID) will be inserted. The value format is an alphanumeric text stringwith a maximum of 4 characters.




Timeouts The number of timeouts that occurred while waiting for data from the control unit. A zerodisplays when no timeouts are found. The value format is an integer.

Total Cache Rate per Second The total I/O rate, per second, for the disk as it is recorded by the cachecontrol unit. That is, where multiple systems are connected to one control unit, the total I/O activity from allsystems is shown. The value format is a decimal number, with one digit to the right of the decimal point.An example is 14.6.

Total Subchannel Measurement Block Rate per Second The I/O rate per second, as indicated bysubchannel measurement block data for the system that does the monitoring. That is, only the I/O activityfor that system is included. The value format is a decimal number, with one digit to the right of the decimalpoint. An example is 14.6.

VOLSER The volume serial number of the device managed by the z/VM Control Program. The valueformat is an alphanumeric text string with a maximum of 8 characters.


Write Hit Percent The overall percentage of write hits to write requests. On 3880 control units, this refersto "write normal" hits. On IBM DASD subsystems, this refers to DASD fast write hits. The value format isan integer.

See also:


KVLDevice attributesThe KVLDevice attribute group provides information about the DASD managed by the z/VM ControlProgram. Use the KVLDevice attributes to create situations that monitor DASD.

The KVLDevice attribute group contains the following attributes:

Average Disconnect Time The average time, in milliseconds, that this device was disconnected duringthe collection interval. The value format is a decimal number, with two digits to the right of the decimalpoint. An example is 0.70.

Average Pending Time The average time, in milliseconds, that a request was waiting in the I/Osubsystem before it was passed to the device. The value format is a decimal number, with two digits to theright of the decimal point. An example is 0.70.

Average Queued IO The average number of input and output (I/O) requests queued on this device. Thevalue format is a decimal number, with two digits to the right of the decimal point. An example is 0.70.

Average Service Time The average time, in milliseconds, that this device was in service during thecollection interval. The value format is a decimal number, with two digits to the right of the decimal point.An example is 0.70.

Connection Time The average time, in milliseconds, this device was connected during the collectioninterval. The value format is a decimal number, with two digits to the right of the decimal point. An exampleis 0.70.


Device Type The type of device managed by the z/VM Control Program. The value format is analphanumeric text string with a maximum of 8 characters.


Number IO per Second The average rate per second of I/O data going to this device. The value format isan integer.

Number of Start IOs This attribute is not supported in any version of this monitoring agent.

Percent Busy The percentage of time the device was busy. The value format is an integer.





Volume Serial Number The volume serial number of the device managed by the z/VM Control Program.The value format is an alphanumeric text string with a maximum of 8 characters.

See also:


KVLFChannel Data attributesThe KVLFChannel Data attribute group provides Fiber Connectivity (FICON) channel load metrics. FICONchannel load analysis is based on data from the Extended Channel Path Measurement Facility, availablefrom the CP monitor SYSTEM domain. The workspace associated with this attribute group displays datafor all channels for which valid information was found. Use the KVLFChannel Data attributes to createsituations that monitor FICON channel load activity.

The KVLFChannel Data attribute group contains the following attributes:

Bus Cycles Percent The bus cycles utilization for the whole system. The value format is in the range of0.00 - 100.00. For example, the value 12.00 represents 12.00%.

CHPID In a channel subsystem, a value assigned to each installed channel path of the system thatuniquely identifies that path to the system. The value format is a hexadecimal identifier no longer than 4characters.

Channel Shared Indicator The 'shared' indicator for this channel.

The following options are possible:

v YES. Channels that are shared with other logical partitions.

v NO. Channels that are dedicated to this logical partition.

The value format is an alphanumeric text string no longer than 3 characters.

Data Read Percent The data units read utilization for the whole system. The value format is in the rangeof 0.00 - 100.00. For example, the value 12.00 represents 12.00%.

Data Write Percent The data units write utilization for the whole system. The value format is in the rangeof 0.00 - 100.00. For example, the value 12.00 represents 12.00%.


Owning LPAR Data Read Percent The data units read utilization for the owning logical partition. Thevalue format is in the range of 0.00 - 100.00. For example, the value 12.00 represents 12.00%.

Owning LPAR Data Write Percent The data units write utilization for the owning logical partition. Thevalue format is in the range of 0.00 - 100.00. For example, the value 12.00 represents 12.00%.

Owning LPAR Work Units Percent The work units utilization for the owning logical partition. The valueformat is in the range of 0.00 - 100.00. For example, the value 12.00 represents 12.00%.




System Reads per Second The total number of bytes read per second for the entire system. The valueformat is a decimal number, with two digits to the right of the decimal point. An example is 0.70.

System Writes per Second The total number of bytes written per second for the entire system. The valueformat is a decimal number, with two digits to the right of the decimal point. An example is 0.70.


Work Units Percent The work units utilization for the entire system. The value format is in the range of0.00 - 100.00. For example, the value 12.00 represents 12.00%.

See also:


KVLHiperSocket attributesThe KVLHiperSocket attribute group provides data about HiperSockets network utilization during thecurrent collection interval. Use the KVLHiperSocket attributes to create situations that track HiperSocketsutilization to ensure that settings do not reach or exceed predefined thresholds.

The KVLHiperSocket attribute group contains the following attributes:

Channel Path The hexadecimal identifier for the channel path. The value format is an alphanumeric textstring with a maximum of 4 characters. When specifying a hexadecimal value, prefix the number with zeroand the letter X. For example, 0X123b for hex value 123b.

Failed LPAR No Buffer The number of sends that failed, per second, for this logical partition, due to thelack of a receiver buffer. The value format is a decimal number, with two digits to the right of the decimalpoint. An example is 0.70.

Failed LPAR Other The number of sends that failed, per second, for this logical partition, due to otherproblems. The value format is a decimal number, with two digits to the right of the decimal point. Anexample is 0.70.

Failed Total No Buffer The number of sends that failed, per second, for the whole system, due to the lackof a receiver buffer. The value format is a decimal number, with two digits to the right of the decimal point.An example is 0.70.


Sharing Indicator The sharing indicator for the channel. Possible options are:

v YES. Channel is shared with other logical partitions.

v NO. Channel is not shared with other logical partitions.





Transferred LPAR Data Units The rate per second that data units are sent for this logical partition. Thevalue format is a decimal number, with two digits to the right of the decimal point. An example is 0.70.

Transferred LPAR Messages The rate per second that messages are sent for this logical partition. Thevalue format is a decimal number, with two digits to the right of the decimal point. An example is 0.70.

Transferred Total Data Units The rate per second that data units are sent for the whole system. Thevalue format is a decimal number, with two digits to the right of the decimal point. An example is 0.70.

Transferred Total Messages The rate per second that messages are sent for the entire system. Thevalue format is a decimal number, with two digits to the right of the decimal point. An example is 0.70.

See also:


KVLLChannel Data attributesThe KVLLChannel Data attribute group provides information about LPAR channel load. LPAR channel loadanalysis is based on CP monitor SYSTEM domain data from the Channel-Path Measurement Facilitywhere available. Otherwise data from the Extended Channel-Path Measurement Facility is used. Forsystems running in a logical partition, the values shown pertain to the channel activity generated by thispartition only, not the overall channel activity. The display associated with this attribute group shows datafor all channels for which valid information was found. Use the KVLLChannel Data attributes to createsituations that monitor channel activity.

The KVLLChannel Data attribute group contains the following attributes:

Average Busy Conditions Percent The percentage of ’busy’ conditions found, calculated for the entireperiod since the last RESET, or since the Performance Toolkit was last invoked. Channel ’busy’percentages cannot be calculated for HiperSocket channels. The value format is in the range of 0.00 -100.00. For example, the value 12.00 represents 12.00%.

CHPID In a channel subsystem, a value assigned to each installed channel path of the system thatuniquely identifies that path to the system. The value format is a hexadecimal identifier no longer than 4characters.

The value format is an alphanumeric text string no longer than 3 characters.

Channel Busy Distribution 0-10 Percent The channel load distribution for each of the active channels, inthe range 0 - 10. It shows the relative frequency with which the channel busy percentage of separatemeasurement intervals was found. The value format is in the range of 0.00 - 100.00. For example, thevalue 12.00 represents 12.00%.

Channel Busy Distribution 11-20 Percent The channel load distribution for each of the active channels,in the range 11 - 20. It shows the relative frequency with which the channel busy percentage of separatemeasurement intervals was found. The value format is in the range of 0.00 - 100.00. For example, thevalue 12.00 represents 12.00%.










Channel Shared Indicator The 'shared' indicator for this channel.

The following options are possible:

v Yes. Channels that are shared with other logical partitions.

v No. Channels that are dedicated to this logical partition.

Group Qualifier The hexadecimal channel model group qualifier. Note that a valid qualifier is availableonly on systems with the extended channel measurement facility. The value format is a hexadecimalidentifier no longer than 4 characters.

Interval Busy Conditions Percent The percentage of ’busy’ conditions found, calculated for the wholeperiod since the last RESET, or since the Performance Toolkit was last invoked. Channel ’busy’percentages cannot be calculated for HiperSockets channels. The value format is in the range of 0.00 -100.00. For example, the value 12.00 represents 12.00%.



Model Group The hexadecimal channel model group. The value format is an alphanumeric text string nolonger than 8 characters.




See also:


KVLLPAR Info attributesThe KVLLPAR Info attribute group provides information about the overall utilization of system processorresources by the logical partitions defined for your system. Use these attributes to create the situationsthat monitor the activity of the logical partitions to ensure that system resources are being utilizedefficiently.

The KVLLPAR Info attribute group contains the following attributes:

LPAR Busy Percent The percentage of time that all of the processors of a specific type were found to bebusy for the reported percentage of the elapsed time interval. Busy is defined as the portion of elapsedtime during which real processors were assigned to logical processors. The LPAR busy utilization reportedfor a partition is the sum of the individual LPAR Busy Percent values for all of the logical processors of agiven type that are defined to a logical partition, divided by the number of logical processors. LPAR BusyPercent will not exceed 100%.

LPAR CPU The number of logical processors defined for this logical partition. The value format is aninteger.

LPAR Capped Indicates whether or not capping is set for the logical partition. If capping is set, CPUcycles are allocated up to the specified maximum. Possible options are:

v YES. Capping has been set for the logical partition.

v NO. Capping is not active.

v UNKNOWN. Data are not available for those systems. This can occur when the logical partition isinactive. That is, no processors are assigned to the logical partition.

LPAR Load The relative load of the partition on the whole processor complex. This is expressed as therelation of the total time that logical processors were dispatched for this partition, compared to the totalamount of available processor time. A load value will display only for those partitions with normalprocessors (non-IFL and non-ICF). Depending on the number of processors defined to a logical partition,the total percentage is calculated based on 100 percent multiplied by the number of assigned processors.The value format is a decimal number, with two digits to the right of the decimal point. An example is 0.70.



LPAR Number The assigned logical partition number. The value format is an integer.

LPAR Overhead Percent The average percentage of elapsed time that the logical processor spent tomanage LPARs. This information is available only on systems with the LPAR management time facility.The value format is in the range of 0.00 - 100.00. For example, the value 12.00 represents 12.00%.

LPAR Overhead Time The general LPAR management time, expressed as a percentage of the availableCPU time (elapsed time multiplied by the number of physical processors). This does not include LPARmanagement overhead for specific logical processors. The information is available only on systems withthe LPAR management time facility. The value format is in the range of 0.00 - 100.00. For example, thevalue 12.00 represents 12.00%.

LPAR Partition ID The user partition ID number for the logical partition. The value format is analphanumeric text string with a maximum of 4 characters.

LPAR Status Indicates the status of the logical partition during the reporting interval. The logical partitionthat was used to collect the LPAR data will display with an asterisk, for example, ACTIVE*. Possibleoptions are:

v ACTIVE

v ACTIVE* - this partition requested data

v INACTIVE

LPAR Suspend Time The percentage of elapsed time that a logical processor was suspended, that is, itcould not provide service to the guest system due to LPAR management time and contention for realprocessors. The "suspended" time is calculated as the difference between the elapsed time and the sumof processor busy time and voluntary wait time for the same processor, as seen by the z/VM system thatis active in the partition. Depending on the number of processors defined to a logical partition, the totalpercentage is calculated based on 100 percent multiplied by the number of assigned processors.

LPAR Wait The wait completion definition for the logical processor. The following are possible values:

v YES. Displays if the wait completion flag is turned on. The logical processor continues to run on a realprocessor until its time slice is up, even though it is waiting.

v NO. Displays if the wait completion flag is turned off.

LPAR Weight The 'weight' defined for the logical processor. Weight values for dedicated processors rangefrom 1 to 999. The string 'DED' displays when processors are dedicated. The value format is a decimalnumber, with two digits to the right of the decimal point. An example is 0.70.

Logical CPU Load The total logical CPU load calculated based on the sum of wait time plus processorbusy time. This value is calculated only for VM systems running in an LPAR. The resulting value dependsnot only on the actual activity of the guest system (the processor time used), but also on the activity ofsystems in other logical partitions on the same physical system. The 'suspended' time will increase if thereis more contention for the real processors, thus causing less voluntary wait time (see the LPAR SuspendTime value). Less voluntary wait time in turn will result in a higher 'Logical CPU load' being calculated,even if the actual CPU time used by a processor did not change. Depending on the number of processorsdefined to a logical partition, the total percentage is calculated based on 100 percent multiplied by thenumber of assigned processors. This attribute provides a good indicator of CPU bottlenecks. The valueformat is a decimal number, with two digits to the right of the decimal point. An example is 0.70.

Physical CPU Busy The sum of the %Load values for all active logical partitions with normal processors(non-IFL and non-ICF), plus the general LPAR management overhead, if available. That is, data displaysfor the overall load of normal processors in the processor complex. This does not include the load ofpartitions with IFL or ICF processors. The value format is a decimal number, with two digits to the right ofthe decimal point. An example is 0.70.


Processor Type The CPU type of the logical processors defined for the partition. Possible values are:

v CP. Central Processor.

v ICF. Internal Coupling Facility Processor.

v IFL. Integrated Facility for Linux Processor.

v ZIIP. IBM System z Integrated Information Processor.

v ZAAP. IBM zSeries Application Assist Processor.

v Special. Other type of processor.

v Unknown. No processor is defined for this logical partition.

Scaled Logical Load Percent The total logical CPU load calculated based on the sum of wait time plusprocessor busy time. This value is calculated only for VM systems running in an LPAR. The resulting valuedepends not only on the actual activity of the guest system (the processor time used), but also on theactivity of systems in other logical partitions on the same physical system. The 'suspended' time willincrease if there is more contention for the real processors, thus causing less voluntary wait time (see theLPAR Suspend Time value). Less voluntary wait time in turn will result in a higher 'Logical CPU load' beingcalculated, even if the actual CPU time used by a processor did not change. Depending on the number ofprocessors of the same type defined to a logical partition, the percentage is calculated based on a total of100 percent. This attribute provides a good indicator of CPU bottlenecks. The value format is a decimalnumber, with two digits to the right of the decimal point. An example is 0.70.

Scaled LPAR Overhead Time Percent The general LPAR management time, expressed as a percentageof the available CPU time (elapsed time). This does not include LPAR management overhead for specificlogical processors. The information is available only on systems with the LPAR management time facility.The value format is in the range of 0.00 - 100.00. For example, the value 12.00 represents 12.00%. Thisvalue is LPAR Overhead Time percent divided by the number of logical processors of the same typedefined to the partition.

Scaled LPAR Suspend Time Percent The percentage of elapsed time that a logical processor wassuspended, that is, it could not provide service to the guest system due to LPAR management time andcontention for real processors. The 'suspended' time is calculated as the difference between the elapsedtime and the sum of processor busy time and voluntary wait time for the same processor, as seen by thez/VM system that is active in the partition. The total percentage is calculated based on 100 percent for allassigned processors of the same type available to the partition.

Scaled VM Load Percent The average processor load as seen by the guest VM system, based on thesum of processor busy time plus elapsed time. This value is calculated only for VM systems running in anLPAR. The value is calculated based on the number of processors defined for a logical partition. The totalpercentage is calculated based on 100 percent for all assigned processors of the same type available tothe partition. The value format is a decimal number, with two digits to the right of the decimal point. Anexample is 0.70.

System ID The name that uniquely identifies the active system. The value format is an alphanumeric textstring with a maximum of 8 characters.



Total LPAR Busy Percent The percentage of time that all logical processors were busy, defined as thepercentage of elapsed time during which real processors were assigned to the logical processors. Thevalue includes wait time in the system that is running in the logical partition if 'wait completion' has been


enabled for the logical processor. Depending on the number of processors defined to a logical partition,the total percentage is calculated based on 100 percent multiplied by the number of assigned processors.

VM CPU Load The average processor load as seen by the guest VM system, based on the sum ofprocessor busy time plus elapsed time. This value is calculated only for VM systems running in an LPAR.Depending on the number of processors defined to a logical partition, the total percentage is calculatedbased on 100 percent multiplied by the number of assigned processors. The value format is a decimalnumber, with two digits to the right of the decimal point. An example is 0.70.

See also:


KVLMinidisk Cache attributesThe KVLMinidisk Cache attribute group provides information about minidisk cache storage. Theinformation shown is based on CP monitor SYSTEM domain SAMPLE data. Use the KVLMinidisk Cacheattributes to create situations that monitor minidisk cache storage utilization.

The KVLMinidisk Cache attribute group contains the following attributes:

Actual Cache Size in Blocks The number of expanded storage blocks used for minidisk caching at theend of the interval. The value format is an integer.

Actual Frames Below 2G The actual number of main storage page frames used for the minidisk cachebelow the 2 GB line. The value format is an integer.

Actual Frames Above 2G The actual number of main storage page frames used for the minidisk cacheabove the 2 GB line. These data are available only for z/VM systems running in 64-bit mode. The valueformat is an integer.

Actual XSTORE in Blocks The actual number of expanded storage blocks used for the minidisk cache.The value format is a decimal number, with two digits to the right of the decimal point. An example is 0.70.

Avg XSTORE Age in Seconds The average age of paging XSTORE blocks. The value format is adecimal number, with two digits to the right of the decimal point. An example is 0.70.

Bias The bias for minidisk cache usage of real storage. A number less than 1 is a bias against minidiskcache; a number greater than 1 is a bias for minidisk cache. The value format is a decimal number, withtwo digits to the right of the decimal point. An example is 0.70.

Block Invalidates per Second The number of blocks per second that were invalidated following aninvalidation request. The value format is a decimal number, with two digits to the right of the decimal point.An example is 0.70.

Block Life in Seconds The estimated average age in seconds of a minidisk cache block, based on Little’sLaw (cache size divided by pageout rate). The value format is a decimal number, with two digits to theright of the decimal point. An example is 0.70.

CCWs Translated per Second The done rate for successfully translated channel command words(CCWs) per second, for DASD. The value format is a decimal number, with two digits to the right of thedecimal point. An example is 0.70.

CCWs Not Eligible per Second The number of CCWs that were found to be not eligible for translation,per second for DASD. The value format is a decimal number, with two digits to the right of the decimalpoint. An example is 0.70.


Fair Share Exceed per Second The rate per second that blocks could not be moved into the minidiskcache because their user’s fair share limit was exceeded. The value format is a decimal number, with twodigits to the right of the decimal point. An example is 0.70.

Fair Share Size in Blocks The fair share limit for the minidisk cache, that is, the maximum number ofblocks any one user is allowed to insert into the minidisk cache per fair share interval. The value format isan integer.

Full Read Hit Percent The percentage of requests that were full hits, that is, where all the requestedblocks were found in the cache. The value format is in the range of 0.00 - 100.00. For example, the value12.00 represents 12.00%.

Ideal Frames The ideal number of main storage page frames in the minidisk cache (based on the averageage of a page in the system's dynamic paging area, determined by arbiter). The value format is an integer.

Ideal XSTORE in Blocks The ideal number of expanded storage blocks in the minidisk cache (based onaverage age of an XSTORE block, as determined by arbiter). The value format is an integer.

InTransit Insertions per Second The insertions per second into the in-transit waiting queue that keepstrack of users requesting blocks that are already on their way to the minidisk cache. The value format is adecimal number, with two digits to the right of the decimal point. An example is 0.70.

Invalidates Request per Second The number of requests per second to invalidate minidisk cache blocks,due to an I/O to a virtual device via a non-cacheable I/O interface. The value format is a decimal number,with two digits to the right of the decimal point. An example is 0.70.


MDC Bias The MDC bias for main storage. The value format is a decimal number, with two digits to theright of the decimal point. An example is 0.70.

MDC Bias XSTORE The bias for minidisk cache use of expanded storage. A number less than 1 is a biasagainst minidisk cache; a number greater than 1 is a bias for minidisk cache. The value format is adecimal number, with two digits to the right of the decimal point. An example is 0.70.

MDCACHE Read Hit Rate per Second The number of read operations per second to cached minidiskswhere the block to be read was already in the minidisk cache. The value format is an integer.

MDCACHE Writes per Second The number of write operations per second to the minidisk cache. Thevalue format is an integer.

Max Cache Size in Blocks The maximum number of blocks that can be used for the minidisk cache. Thevalue format is an integer.

Maximum Storage Frames The maximum number of main storage page frames to be used for theminidisk cache. The value format is an integer.

Maximum XSTORE in Blocks The maximum number of expanded storage blocks to be used for theminidisk cache. The value format is an integer.

Minidisk Cache Reads per Second The number of read requests to the minidisk cache, per second,where all the requested blocks were found in the cache, thereby avoiding an I/O operation. The valueformat is a decimal number, with two digits to the right of the decimal point. An example is 0.70.


Minimum Storage Frames The minimum number of main storage page frames to be used for theminidisk cache. The value format is an integer.

Minimum XSTORE in Blocks The minimum number of expanded storage blocks to be used for theminidisk cache. The value format is an integer.

Network CCWs Aborted per Second The number of attempts at translation of CCWs, per second, fornetwork devices. The value format is an integer.

Network CCWs Translated per Second The done rate for networks. The value format is an integer.

Network CCWs Not Eligible per Second The number of CCWs that were found to be not eligible fortranslation, per second for networks. The value format is an integer.

Pages Deleted per Second The number of main storage pages deleted from cache per second. Thevalue format is a decimal number, with two digits to the right of the decimal point. An example is 0.70.

Page In per Second The rate at which minidisk cache blocks are moved, per second, from expandedstorage to central storage. The value format is a decimal number, with two digits to the right of the decimalpoint. An example is 0.70.

Page Out per Second The rate at which minidisk cache blocks are moved per second from centralstorage to expanded storage. The value format is a decimal number, with two digits to the right of thedecimal point. An example is 0.70.

Partition Size in Blocks The size of the CP partition in expanded storage, in blocks. The value format isan integer.

Steal Invoked per Second The rate at which the steal function was invoked to steal expanded storagepages from the minidisk cache per second. The value format is a decimal number, with two digits to theright of the decimal point. An example is 0.70.




Total CCWs per Second The total number of CCWs handled per second for DASD. The value format is adecimal number, with two digits to the right of the decimal point. An example is 0.70.

Total Network CCWs Translated per Second The total number of CCWs handled per second fornetworks. The value format is a decimal number, with two digits to the right of the decimal point. Anexample is 0.70.

Translations Aborted per Second The aborted translation attempts per second for DASD. The valueformat is a decimal number, with two digits to the right of the decimal point. An example is 0.70.

XSTORE Pages Deleted per Second The number of XSTORE pages deleted from cache per second.The value format is a decimal number, with two digits to the right of the decimal point. An example is 0.70.


XSTORE Steal Invoked per Second The number of times the steal function was invoked to stealexpanded storage pages from cache per second. The value format is a decimal number, with two digits tothe right of the decimal point. An example is 0.70.

See also:


KVLProcessor Data attributesThe KVLProcessor Data attribute group provides information about the CPU utilization of each processorassigned to each defined LPAR. Use the KVLProcessor Data attributes to create situations that monitorCPU utilization by processors.

LPAR Busy The percentage of time that the logical processor was busy, defined as the percentage ofelapsed time during which a real processor was assigned to the logical processor. The value will alsoinclude wait time in the system that is running in the logical partition if 'wait completion' has been enabledfor the logical processor. The value format is in the range of 0.00 - 100.00. For example, the value 12.00represents 12.00%.

LPAR CPU The number of logical processors defined for this logical partition. The value format is aninteger.

LPAR Capped Indicates whether or not capping has been defined and has been turned on for thepartition. If capping is turned on, CPU cycles are allocated up to the specified maximum. Possible optionsare:

v YES. Capping has been set for the logical partition.

v NO. Capping is not active.

v UNKNOWN. Data are not available for those systems. This can occur when the logical partition isinactive. That is, no processors are assigned to the logical partition.

LPAR Load The relative load of the partition on the whole processor complex. This is expressed as therelation of the total time that logical processors were dispatched for this partition, compared to the totalamount of available processor time. A load value will display only for those partitions with normalprocessors (non-IFL and non-ICF). The value format is a decimal number, with two digits to the right of thedecimal point. An example is 0.70.


LPAR Name Processor Number This attribute is for display purposes only. It is a combination of thefollowing two attributes:

v LPAR Name

v Processor Number

Note: Do not use this attribute to create a query or to create a situation. Use the LPAR Name attributeand the Processor Number attribute instead.

LPAR Number The assigned logical partition number. The value format is an integer.

LPAR Overhead Percent The average percentage of elapsed time that the logical processor spent tomanage LPARs. This information is available only on systems with the LPAR management time facility.The value format is in the range of 0.00 - 100.00. For example, the value 12.00 represents 12.00%.


LPAR Overhead Time The general LPAR management time, expressed as a percentage of the availableCPU time (elapsed time multiplied by the number of physical processors). This does not include LPARmanagement overhead for specific logical processors. The information is available only on systems withthe LPAR management time facility. The value format is a decimal number, with two digits to the right ofthe decimal point. An example is 0.70.

LPAR Partition ID The user partition ID number for the logical partition. The value format is analphanumeric text string with a maximum of 4 characters.

LPAR Status Indicates whether or not this logical partition retrieved the performance data for the LPAR.The logical partition that was used to collect the LPAR data will display with an asterisk, for example,ACTIVE*. Possible options are:

v INACTIVE

v ACTIVE

v ACTIVE* - this specific partition requested data

LPAR Suspend Time The percentage of elapsed time that a logical processor was suspended, that is, itcould not provide service to the guest system due to LPAR management time and contention for realprocessors. The "suspended" time is calculated as the difference between the elapsed time and the sumof processor busy time and voluntary wait time for the same processor, as seen by the z/VM system thatis active in the partition. The value format is in the range of 0.00 - 100.00. For example, the value 12.00represents 12.00%.

LPAR Wait The wait completion definition for the logical processor. It can contain one of the followingvalues:

v YES. Displays if the wait completion flag is turned on. The logical processor continues to run on a realprocessor until its time slice is up, even though it is waiting.

v NO. Displays if the wait completion flag is turned off.

LPAR Weight The 'weight' defined for this type of logical processor. Weight values for dedicatedprocessors range from 1 to 999. The string 'DED' is inserted when processors are dedicated. The valueformat is a decimal number, with two digits to the right of the decimal point. An example is 0.70.

Logical CPU LoadThe logical CPU load calculated based on the sum of processor wait time plusprocessor busy time. Note that the resulting value depends not only on the actual activity of the guestsystem (the processor time used), but also on the activity of systems in other logical partitions on thesame physical system. The 'suspended' time (see LPAR Suspend Time) will increase if there is morecontention for the real processors, thus causing less voluntary wait time. Less voluntary wait time results ina higher logical CPU load being calculated, even if the actual CPU time used by a processor is the same.The value format is a decimal number, with two digits to the right of the decimal point. An example is 0.70.

Physical CPU Busy The sum of the %Load values for all active logical partitions with normal processors(non-IFL and non-ICF), plus the general LPAR management overhead, if available. That is, data displaysfor the overall load of normal processors in the processor complex. This does not include the load ofpartitions with IFL or ICF processors. The value format is a decimal number, with two digits to the right ofthe decimal point. An example is 0.70.

Processor Number The CPU identification number allocated to each of the logical processors that weredefined. The value format is an integer.

Processor Type The CPU type of the logical processors defined for the partition. Possible values are:

v CP. Central Processor.

v ICF. Internal Coupling Facility Processor.

v IFL. Integrated Facility for Linux Processor.


v ZIIP. IBM System z Integrated Information Processor.

v ZAAP. IBM zSeries Application Assist Processor.

v Special. Other type of processor.

v Unknown. No processor is defined for this logical partition.




VM CPU Load The CPU load for the z/VM guest system, calculated based on the sum of processor busytime plus elapsed time. The value format is a decimal number, with two digits to the right of the decimalpoint. An example is 0.70.

See also:


KVLPTKStat attributesThe KVLPTKStat attribute group provides data on how the Performance Toolkit is being used during thecurrent collection interval. It also lets you know whether or not the Performance Toolkit component isactive. Use the KVLPTKStat attributes to learn about the type of data being collected by the PerformanceToolkit. The product-provided situation ZVM_PerfKit_Collector_Inactive triggers when the PerformanceToolkit is no longer active and collecting data for the monitoring agent.

The KVLPTKStat attribute group contains the following attributes:

Collector Name The name of the Performance Toolkit component. This attribute also displays the type ofz/VM data being collected by the Performance Toolkit and displayed on the Tivoli Enterprise Portalinterface. For example, if you are collecting data on the devices being managed by the Control Program,the row called CP-Owned will contain data on these devices. The value format is an alphanumeric textstring no longer than 32 characters.

Collector Status Identifies whether or not the Performance Toolkit component is active. This attribute alsodisplays whether or not a particular type of data are currently being collected. The possible values are:

v Active

v Inactive

Important: If the Performance Toolkit stops updating the PERFOUT DCSS with data that it has collected,the predefined situation ZVM_PerfKit_Collector_Inactive raises an alert, indicating a problem. If this occurs,investigate the cause immediately. For more information on the possible causes, see“ZVM_PerfKit_Collector_Inactive” on page 75. See the IBM Tivoli OMEGAMON XE on z/VM and LinuxPlanning and Configuration Guide for details on configuring your environment for this monitoring agent.


Number of rows this interval The number of rows collected for this component during the last collectioninterval. The value format is an integer.



System Name The host name of the system on which the monitoring agent is running. The value format isan alphanumeric text string no longer than 64 characters.


z/VM Collection Interval The current collection interval, in seconds, as received from the PerformanceToolkit. This setting is defined on the Performance Toolkit Collector. The value format is an integer.

See also:


KVLSpinLock attributesThe KVLSpinLock attribute group provides wait statistics on processor spin locks. Use this attribute groupto create situations that monitor spin lock data to ensure that settings do not reach or exceed predefinedthresholds.

Exclusive Spin Lock Rate per Second The total exclusive spin lock calls, per second. The value formatis a decimal number, with one digit to the right of the decimal point. An example is 1.4.

Exclusive Time Spinning on Locks in Microseconds The average total microseconds spent spinning onexclusive spin locks. The value format is a decimal number, with two digits to the right of the decimalpoint. An example is 0.70.

Exclusive Time Spinning on Locks Percent The percentage of total elapsed time spent on exclusivespin locks. The value format is in the range of 0.00 - 100.00. For example, the value 12.00 represents12.00%.


Lock Name The name of the CP monitor spin lock. The value format is an alphanumeric text string with amaximum of 8 characters.

Shared Spin Lock Rate per Second The total shared spin lock calls, per second. The value format is adecimal number, with one digit to the right of the decimal point. An example is 1.4.

Shared Time Spinning on Locks in Microseconds The total average number of microseconds spentspinning on shared spin locks. The value format is a decimal number, with two digits to the right of thedecimal point. An example is 0.70.

Shared Time Spinning on Locks Percent The percentage of total elapsed time spent on shared spinlocks. The value format is in the range of 0.00 - 100.00. For example, the value 12.00 represents 12.00%.

Time Spinning on Locks in Microseconds The total average number of microseconds spent spinning onboth exclusive and shared spin locks. The value format is a decimal number, with two digits to the right ofthe decimal point. An example is 0.70.


Time Spinning on Locks Percent The percentage of total elapsed time spent on both exclusive andshared spin locks. The value format is in the range of 0.00 - 100.00. For example, the value 12.00represents 12.00%.

Total Spin Lock Calls per Second The total exclusive and shared spin lock calls, per second. The valueformat is a decimal number, with one digit to the right of the decimal point. An example is 1.4.




See also:


KVLSystem attributesThe KVLSystem attribute group provides system-level data on CPU utilization by the z/VM ControlProgram and by the z/VM virtual machines, on free-storage management, on paging rates, and on useractivity. Use the System attributes to create situations that monitor the performance of your z/VM system.

The KVLSystem attribute group contains the following attributes:

Active Users The number of users who had any activity since the last sampling interval. The value formatis an integer.

Available Frames High Thresh The high threshold for frames on the below 2GB the available listreplenishment subsystem. The value format is an integer.

Available Frames High Thresh > 2 GB The high threshold for frames on the above 2GB available listreplenishment subsystem. The value format is an integer.

Available Frames Mean The average number of page frames available on the below 2GB available listreplenishment subsystem. The value format is an integer.

Available Frames Mean > 2 GB The average number of page frames on the above 2GB available listreplenishment subsystem. The value format is an integer.

Available Pages Low Thresh The number of page frames on the below 2GB available list replenishmentsubsystem. The value format is an integer.

Available Pages Low Thresh > 2 GB The low threshold on the above 2GB available list replenishmentsubsystem. The value format is an integer.

Average Number of Eligible Users The average number of users waiting on the eligible list. An eligiblelist contains a list of virtual machines that are waiting for resources. The value format is a decimal number,with two digits to the right of the decimal point. An example is 0.70.

Average Number of Users The average number of users logged on. The value format is an integer.


CP Percent of CPU The percentage of CPU utilized by the z/VM Control Program, and if you are runningmultiple processors, this value is the average of the CPU utilization for all the active processors. The valueformat is an integer.

DPA PCT of Real Stor The percentage of real storage available to the dynamic paging area. The valueformat is in the range of 0.00 - 100.00. For example, the value 12.00 represents 12.00%.

Deferred Pages The number of deferred pages waiting for a frame. The value format is an integer.

Demand Scan Fails The number of times the demand scan was invoked and could not replenish theavailable list to its threshold. The value format is an integer.

Eligible List E1 The number of short running (interactive) users in the eligible list for the E1 queue.

Eligible List E2 The number of medium-running users in the eligible list for the E2 queue.

Eligible List E3 The number of long-running users in the eligible list for the E3 queue.

Free Stor Releases The rate per second of free-storage releases during the collection interval. The valueformat is an integer.

Free Stor Requests The rate per second of free-storage requests during the collection interval. The valueformat is an integer.

Free Stor Used The number of page frames used by free-storage management during the collectioninterval. The value format is an integer.

I/O Operation Wait Queue The percentage of all virtual machines in the queue that are in an I/O waitstate. The value format is an integer.

In Queue Users The average number of users in the queue waiting to be dispatched. The value format isan integer.

Locked Dynamic Paging Area Storage in Kbytes The amount of DPA storage consisting of lockedpages (including CP FREE storage space).

LPAR Busy Percent The percentage of time this logical processor was busy, based on the number ofavailable logical processors. The value format is in the range of 0.00 - 100.00. For example, the value12.00 represents 12.00%.


Main Storage Available in Mbytes Total main storage size. Note that this value reflects the real storagesize, including HSA (Hardware Storage Area) frames and that less storage might actually be used by yoursystem if less storage has been defined during system generation.

Main Storage Offline in Kbytes Total main storage that is offline.

Non-Trivial Trans Rate The rate per second that non-trivial transactions are processed during thecollection interval. The value format is a decimal number, with two digits to the right of the decimal point.An example is 0.70.

Number Active CPU's The number of active CPUs. The value format is an integer.


Number of Dialed Users The number of users who are dialed to the z/VM system. The value format is aninteger.

Number of Dynamic Frames The number of page frames allocated to the dynamic paging area. Thevalue format is an integer.

Number of Frames The number of page frames below the 2 GB line currently on the available list. Thevalue format is an integer.

Number of Frames > 2 GB The number of page frames above the 2 GB line currently on the availablelist. The value format is an integer.

Page Index The loading user capacity of the system. This value is equal to the number of pagingexposures. That is, the number of paging I/O requests that are executing concurrently. You can use thisvalue to determine the extent to which the paging resources are creating a bottleneck. The value format isan integer.

Page Rate The rate per second that pages are being read in by the system. The value format is a decimalnumber, with two digits to the right of the decimal point. An example is 0.70.

Page Wait Queue The percentage of all virtual machines in the queue that are in a page wait state. Thevalue format is an integer.

Pageable Dynamic Paging Area Storage in Mbytes The size of the pageable part of the DynamicPaging Area (DPA), that is, total DPA size minus locked pages, minus trace table.

Pageable Dynamic Paging Area Storage in Use Percent The percentage of pageable storage used upby the working sets of currently active users. Due to rounding, this value can exceed 100%.

Pct Page Space In Use The average system-wide percentage of paging space in use. The value format isan integer.

Pct Spool Space In Use The percentage of spool space in use for the entire system. The value format isan integer.

Pct Tdisk Space In Use The percentage of temporary disk space in use for the entire system. The valueformat is an integer.

Percent of CPU The total CPU utilization of the system by both the z/VM Control Program and the virtualCPU, and if you are running multiple processors, this value is the sum of the CPU utilization for all theactive processors and can be greater than 100%. The value format is an integer.

Reserved Real Storage in Mbytes Real machine reserved storage size.

Shared Segment Storage in Mbytes The amount of main storage currently used by shared segments.

Single Read Page Faults The rate per second of page faults for single-page reads during the collectioninterval. The value format is a decimal number, with two digits to the right of the decimal point. An exampleis 0.70.

Standby Real Storage Size in Mbytes Real machine standby storage size. Standby storage is calculatedbased on the amount of installed storage that is not currently claimed by active logical partitions and theRESERVED storage specified for this partition.

Subpool Storage in Kbytes Storage taken up by subpools.


Subpool Storage in Use Percent The percentage of subpool free storage that is currently in use.

SYSGEN Storage Size in Mbytes Storage size defined for the z/VM system.



System Paging Rate The system-wide page read rate, per second. The value format is an integer.

System Transaction Rate The rate per second of total CPU time (ratio) that transactions are executed onthe system. The value format is a decimal number, with two digits to the right of the decimal point. Anexample is 0.70.

Tasks in Wait The number of tasks that cannot be executed because they are waiting for a frame. Thevalue format is an integer.

Tasks Waiting for a Frame The number of CP tasks waiting for a page frame.

Tasks Waiting for a Page Rate per Second The number of CP tasks that were found to be waiting for apage, per second.


Total to Virtual Ratio The ratio of total CPU time to virtual CPU time. The value format is a decimalnumber, with two digits to the right of the decimal point. An example is 0.70.

Trace Table Storage in Kbytes Storage occupied by CP trace table.

Trivial Transaction Rate The rate per second that trivial transactions are processed during the collectioninterval. The value format is a decimal number, with two digits to the right of the decimal point. An exampleis 0.70.

Users Waiting for Resources The percentage of users in the queue waiting for a constrained resource,during the collection interval. The value format is an integer.

Virtual CPU Busy Percent The difference between the total CPU utilization (CPU busy) and the amountof CPU used for overhead (CP Busy). This leaves the amount of CPU that is used by virtual machines.

XSTORE Allocated to CP in Mbytes Expanded storage size allocated for CP use.

XSTORE Dedicated to Virtual Machines in Mbytes Size of XSTORE dedicated to specific virtualmachines.

XSTORE In Use by CP Percent Percentage of CP XSTORE partition currently in use.

XSTORE Minimum Before Migrating in Kbytes The minimum XSTORE size that is to be kept free forallocation of new blocks. Used XSTORE blocks are migrated to DASD if less free XSTORE space isavailable.


XSTORE Page Age at Migration in Seconds The average age of XSTORE pages at migration. If theaverage age at migration is low and near to the average age of all XSTORE blocks in storage, the size ofthe CP partition is probably too small.

XSTORE Page Age in Seconds The average age of XSTORE pages in the CP partition.

XSTORE Page Allocation Rate per Second The rate at which XSTORE pages were allocated.

XSTORE Size in Mbytes Total size of expanded storage.

See also:


KVLSystem2 attributesThe KVLSystem2 attribute group provides system-level data on CPU utilization by the z/VM ControlProgram and by the z/VM virtual machines, on free-storage management, on paging rates, and on useractivity. Use the KVLSystem2 attributes to create situations that monitor the performance of your z/VMsystem.

The KVLSystem2 attribute group contains the following attributes:

Highest Normalized Individual Wait State Workload After normalizing for the percent of the interval thatthe workload was active, this is the workload which has the highest individual wait state. Wait statesconsidered are limited to CPU, Loading/Page, I/O, Instruction Simulation, Console Function, and PageActive Wait. If no workload is found waiting for one of these reasons this attribute is blank. See alsoHighest Individual Wait State Workload.

Highest Normalized Individual Wait State Percent After normalizing for the percent of the interval thatthe workload was active, this shows the value of the specific wait state of the workload on the z/VMsystem that has the highest wait state. Wait states considered are limited to CPU, Loading/Page, I/O,Instruction Simulation, Console Function, and Page Active Wait. If no workload is found waiting for one ofthese reasons this attribute is zero. See also Highest Individual Wait State Percent.

Highest Individual Wait State This shows the specific wait state of the workload on the z/VM system thathas the highest wait state. Wait states considered are limited to CPU, Loading/Page, I/O, InstructionSimulation, Console Function, and Page Active Wait. If no workload is found waiting for one of thesereasons this attribute contains: 'Unknown'. See also: Highest Normalized Individual Wait State.

Highest Individual Wait State Percent This shows the value of the specific wait state of the workload onthe z/VM system that has the highest wait state. Wait states considered are limited to CPU, Loading/Page,I/O, Instruction Simulation, Console Function, and Page Active Wait. If no workload is found waiting forone of these reasons this attribute is zero. See also: Highest Normalized Individual Wait State Percent.

Highest Individual Wait State Workload The workload which has the highest individual wait state. Waitstates considered are limited to CPU, Loading/Page, I/O, Instruction Simulation, Console Function, andPage Active Wait. If no workload is found waiting for one of these reasons this attribute is blank. See also:Highest Normalized Individual Wait State Workload.

Highest Normalized Individual Wait State After normalizing for the percent of the interval that theworkload was active, this shows the specific wait state of the workload on the z/VM system that has thehighest wait state. Wait states considered are limited to CPU, Loading/Page, I/O, Instruction Simulation,Console Function, and Page Active Wait. If no workload is found waiting for one of these reasons thisattribute contains: 'Unknown'. See also Highest Individual Wait State.



Page Read Blocking Factor in pages Average blocking factor for page read operations, that is, averagenumber of pages read with a single read operation. Both the page read and page write blocking factorsare indicators of the efficiency of block paging (the higher the number, the greater the efficiency). Due tothe page-in logic employed, the page read blocking factor is usually approximately half the page writeblocking factor.

Page Write Rate per Second The number of pages per second being written by the system.

Page Write Blocking Factor in pages Average blocking factor for page write operations.

Pages Moved Below 2G Line Rate per Second The rate at which pages were moved below the 2-GBline for translation. The field is available only for z/VM systems running in 64-bit mode.

Pages Paged in via Fast Path Logic Rate per Second The rate at which pages were paged in by usingfast path logic.

Pages Paged in via Long Path Logic Rate per Second The rate at which pages were paged in by usinglong path logic.

Pages Paged out via Long Path Logic Rate per Second The rate at which pages were paged out byusing long path logic.

Paging I/O from Storage to DASD Rate per Second Number of I/O operations per second for movingpages between main storage and paging devices. Due to the block paging logic, this number is usuallyconsiderably lower than the number of pages transferred.

Spool Read Rate per Second The rate at which pages are read from SPOOL devices.

Spool Write Rate per Second The rate at which pages are written out to SPOOL devices.

Storage Overcommit Minus VDISK Ratio The ratio of the sum of the working set sizes of all users, tothe total real storage of the z/VM system minus the storage used for VDISKS.

Storage Overcommit Ratio The ratio of the sum of the working set sizes of all users to the total realstorage of the z/VM system.

System and Workload with Highest Normalized Wait State This attribute is only for display purposes.After normalizing for the percent of the interval that the workload was active, it is a combination of theworkload on the z/VM system that has the highest wait state, along with the system name. Wait statesconsidered are limited to the following:

v CPU

v Loading/Page

v I/O

v Instruction Simulation

v Console Function

v Page Active Wait

If no workload is found waiting for one of these reasons this attribute is blank.

System and Workload with Highest Wait State This attribute is only for display purposes. It is acombination of the workload on the z/VM system that has the highest wait state, along with the systemname. Wait states considered are limited to the following:


v CPU

v Loading/Page

v I/O

v Instruction Simulation

v Console Function

v Page Active Wait

If no workload is found waiting for one of these reasons this attribute is blank.




Total Main Storage in Mbytes Total main storage size, including any offline frames.

VDISK Main Storage Frames The number of main storage page frames used by virtual disks.

VDISK Maximum Storage Blocks The system limit set for maximum storage utilization by virtual disks inblocks.

VDISK Pages Migrated to DASD The number of virtual disk pages that have been migrated to DASD.

VDISK XSTORE Blocks The number of expanded storage blocks used by virtual disks.

XSTORE to DASD Blocking Factor in pages Average blocking factor for migrating pages out fromXSTORE to DASD. Both the page read and page write blocking factors are indicators of the efficiency ofblock paging (the higher the number, the greater the efficiency). Due to the page-in logic employed, thepage read blocking factor is usually approximately half the page write blocking factor.

See also:


KVLUser Wait attributesThe KVLUser Wait attribute group provides information about resource constraint analysis, which isdisplayed as one line per user.

The KVLUser Wait attribute group contains the following attributes:

Active Percent Percentage of samples where the user was active, either consuming CPU or waiting forsome service.

Console Function Wait Percent Percentage of active samples where the user was in console functionwait, that is, waiting for the central processor to execute one of the console function commands. By usingthese commands, a user can alter the state and configuration of the virtual machine (for example LINK,ATTACH, and DETACH).

Eligible List Percent Percentage of active measurements where the user was in the eligible list (anyclass).


I/O Wait Percent Percentage of active samples where the user was in I/O wait state.

Instruction Simulation Wait Percent Percentage of active samples where the user was in instructionsimulation wait, that is, waiting for the central processor to complete simulation of a privileged instruction.

Limit List Wait Percent Percentage of active measurements where the user was in the limit list, that is inthe dispatch list waiting for CPU, but prevented from running by a maximum share setting.

Loading Percent Percentage of active samples where the user was in loading state, that is with a highpage read rate.


Other State Percent Percentage of active measurements where the user was found in other states.

Page Active Wait Percent Percentage of page active wait, that is the virtual machine has loaded a waitPSW while waiting for the completion of a page read operation. This state is to be expected for guestsystems running with a PAGEX ON setting.

Page Wait Percent Percentage of active samples where the user was found in page wait state.

Q0 List Percent Percentage of active measurements where the user was in Q0 (in the dispatch list asclass 0 user).




Running Percent Percentage of active samples where the user was running on a real processor.

Sum of Top 5 Wait States The sum of the I/O wait, instruction simulation wait, page wait, CPU wait, andloading wait percent values. This value has a range of 0 to 500.

SVM and Dormant Wait List Percent Percentage of active samples where the user was in both SVMwait and in the dormant list. This status is counted as active since the machine is actually waiting forsomething, and not idle.

SVM and Eligible List Wait Percent Percentage of active samples where the user was in both SVM waitand in the eligible list (resource wait).

SVM and Test Idle Wait Percent Percentage of active samples where the user was in both SVM wait andtest idle wait. SVM wait state indicates that the user was waiting for the completion of a communicationwith a service virtual machine.


Test Idle Wait Percent Percentage of active samples where the user was in test idle wait, that is waitingto be dropped from the dispatch list.



Wait While I/O Active Percent Percentage of active measurements where the user was found in a stateother than the preceding states, while an I/O operation started by this machine had not yet completed.

Waiting Percent Percentage of active samples where the user was waiting to run on the real CPU.

Waiting User ID User ID.

See also:


KVLTCPIP Srvr Data attributesThe KVLTCPIP Srvr Data attribute group provides information on TCP/IP server utilization during thecurrent collection interval. Use the KVLTCPIP Srvr Data attributes to keep track of network traffic to ensurethat settings do not reach or exceed predefined thresholds.

The KVLTCPIP Srvr Data attribute group contains the following attributes:

ARP Received The rate per second of Address Resolution Protocol (ARP) requests received. The valueformat is a decimal number, with two digits to the right of the decimal point. An example is 0.70.

ARP Reply Xmit The rate per second at which ARP replies were transmitted. The value format is adecimal number, with two digits to the right of the decimal point. An example is 0.70.

ARP Request Xmit The rate per second at which ARP requests were transmitted. The value format is adecimal number, with two digits to the right of the decimal point. An example is 0.70.

Activity Control Block The activity control block pool level. The value format is an integer.

BSD Socket Pool The BSD-type socket control block pool level. The value format is an integer.

Client Control Block The client control block pool level. The value format is an integer.

Conn Open Accept The rate, per second, at which the TCP connection open requests were accepted.The value format is a decimal number, with two digits to the right of the decimal point. An example is 0.70.

Conn Open Failed The rate, per second, for TCP connection open failures. The value format is a decimalnumber, with two digits to the right of the decimal point. An example is 0.70.

Conn Open Init The rate, per second, at which the TCP connection open requests were initiated. Thevalue format is a decimal number, with two digits to the right of the decimal point. An example is 0.70.

Connections Reset The reset rate, per second, for TCP connections. The value format is a decimalnumber, with two digits to the right of the decimal point. An example is 0.70.

Data Buffer Regular The regular data buffer pool level. The value format is an integer.

Data Buffer Small The small data buffer pool level. The value format is an integer.

Data Buffer Tiny The tiny data buffer pool level. The value format is an integer.


Envelope Large Pool The large envelope pool level. The value format is an integer.

Envelope Regular Pool The regular envelope pool level. The value format is an integer.

Fixed Page Pool The fixed page storage pool level. The value format is an integer.

IO Bytes Received The rate per second of bytes received. The value format is a decimal number, withtwo digits to the right of the decimal point. An example is 0.70.

IO Bytes Sent The rate per second of bytes sent. The value format is a decimal number, with two digits tothe right of the decimal point. An example is 0.70.

IO Requests Read The rate per second of read requests. The value format is a decimal number, with twodigits to the right of the decimal point. An example is 0.70.

IO Requests Write The rate per second of write requests. The value format is a decimal number, with twodigits to the right of the decimal point. An example is 0.70.


Raw IP Pool The raw IP control block pool level. The value format is an integer.

Segment Acknowledge Pool The segment acknowledgement control block pool level. The value format isan integer.

Segment Received in Error The rate per second during which TCP segments were received that haderrors. The value format is a decimal number, with two digits to the right of the decimal point. An exampleis 0.70.

Segments Received The rate per second during which TCP segments were received. The value format isa decimal number, with two digits to the right of the decimal point. An example is 0.70.

Segments Retransmitted The rate per second during which segments were retransmitted. The valueformat is a decimal number, with two digits to the right of the decimal point. An example is 0.70.

Segments Sent The rate per second during which TCP segments were transmitted. The value format is adecimal number, with two digits to the right of the decimal point. An example is 0.70.

Segments Sent Reset The rate per second during which TCP segments were transmitted that included areset. The value format is a decimal number, with two digits to the right of the decimal point. An exampleis 0.70.

Socket Pool The socket control block pool level. The value format is an integer.



TCP Pool The TCP control block pool level. The value format is an integer.

TCPIP Server The name of the TCP/IP Server. The value format is an alphanumeric text string with amaximum of 8 characters.



UDP Pool The User Datagram Protocol (UDP) control block pool level. The value format is an integer.

See also:


KVLTCPIPUsrData attributesThe KVLTCPIPUsrData attribute group provides data on TCP/IP usage by the virtual machines during thecurrent collection interval. Use the KVLTCPIPUsrData attributes to ensure that settings to do not reach orexceed predefined thresholds.

Note: Data does not display in the TCP/IP User Activity table of the “TCPIP User workspace” on page56 for those TCP/IP users that have not completed any TCP sessions and any UDP sessions during thecollection interval.

The TCPIPUsrData attribute group contains the following attributes:

Average Session Duration The average elapsed time, in seconds, from the beginning of a session to theend of a session.


Max Buffer Q Size Input The maximum input buffer queue size. The value format is an integer.

Max Buffer Q Size Output The maximum output buffer queue size. The value format is an integer.

Smoothed Round Trip The smoothed round trip time, in seconds. The value format is an integer.



TCP Completed Sessions The number of completed Transmission Control Protocol (TCP) sessions. Thevalue format is an integer.

TCPIP Server The name that identifies the TCP/IP server to which the user is connected. The valueformat is an alphanumeric text string with a maximum of 8 characters.


Total Segments The total number of segments. The value format is an integer.

Total TCP KBytes Received The number of kilobytes received during TCP sessions. The value format isan integer.


Total TCP KBytes Sent The number of kilobytes sent during TCP sessions. The value format is aninteger.

Total UDP KBytes Received The number of kilobytes received during User Datagram Protocol (UDP)sessions. The value format is an integer.

Total UDP KBytes Sent The number of kilobytes sent during UDP sessions. The value format is aninteger.

UDP Completed Sessions The number of completed UDP sessions. The value format is an integer.

Unacknowledged Segments The maximum number of segments that have not been acknowledged. Thevalue format is an integer.

User ID The user ID of the virtual machine that handled the local side of the connection. The value formatis an alphanumeric text string with a maximum of 8 characters.

Variance Round Trip The round trip variance time, in seconds. The value format is an integer.

See also:


KVLUser ApplData attributesThe KVLUser ApplData attribute group provides information about Linux workload activity and storageutilization. Use this attribute group to create situations that monitor the activity and performance of theLinux virtual machines to ensure that settings do not reach or exceed predefined thresholds.

Note: Data collection for this attribute group must be enabled at each Linux guest system that is to bemonitored. In addition, you must enable the CP Monitor APPLDATA domain on the z/VM system where thePerformance Toolkit is running.

See the IBM Tivoli OMEGAMON XE on z/VM and Linux Planning and Configuration Guide for instructionson enabling data collection at the Linux guest for this group of attributes.

The KVLUser ApplData attribute group contains the following attributes:

Appldata Flag field The APPLDATA flag is an internal attribute that is used to determine whether thevalue for an attribute was normalized. Normalization is applied to APPLDATA metrics that are created by aLinux system that is at a level prior to SUSE Linux Enterprise Server 10 or prior to Red Hat EnterpriseLinux 5. That is, the CPU percent utilization metrics are adjusted based on the virtual CPU percentreported for the Linux guest. If the percentage of time stolen is available in the Linux APPLDATA, andLinux systems are at a SUSE Linux Enterprise Server 10 level or at a Red Hat Enterprise Linux 5 level,processor utilization metrics are already normalized.

Possible values are the following:

v x'01' - Normalization has occurred.

v x'00' - Normalization was not needed.

Normalization of data is applied to the following KVLUser ApplData attributes:

v Kernel CPU - The percentage of CPU used by the Linux virtual machine, running in kernel mode.

v Nice CPU - The percentage of CPU time in ’nice’ mode. A process runs in ’nice’ mode if the schedulingpriority is lower than normal. If the superuser has increased the scheduling priority of some processesto values higher than normal, this is not considered ’nice’ mode.


v Total CPU - The percentage of CPU total active time.

v User CPU - The percentage of CPU time in user mode, that is, executing ordinary user code. If a userprogram calls a system routine like open(), the execution time needed for the open() system routine iscounted as kernel mode, while the normal processing of the user program is counted as user mode.

Avg Processes Last Minute The average number of processes found running during the last minute. Thevalue format is a decimal number, with two digits to the right of the decimal point. An example is 0.70.

Avg Processes Last 5 Minutes The average number of processes found running during the last fiveminutes. The value format is a decimal number, with two digits to the right of the decimal point. Anexample is 0.70.

Avg Processes Last 15 Minutes The average number of processes found running during the last fifteenminutes. The value format is a decimal number, with two digits to the right of the decimal point. Anexample is 0.70.

Bad Packets The number of bad packets received, per second. The value format is a decimal number,with two digits to the right of the decimal point. An example is 0.70.

Block I/O Read Rate The block I/O data read rate, in kilobytes per second. The value format is a decimalnumber, with two digits to the right of the decimal point. An example is 0.70.

Block I/O Write Rate The block I/O data write rate, in kilobytes per second. The value format is a decimalnumber, with two digits to the right of the decimal point. An example is 0.70.

Buffers and Free Cache The size of the memory that is reserved for buffers and for free cache, inmegabytes. The value format is a decimal number, with two digits to the right of the decimal point. Anexample is 0.70.

Buffer Cache Used The size of the memory that is used for buffers, in megabytes. The Linux buffer cacheis a disk cache that is intended to relieve processes from having to wait for relatively slow disks to retrieveor to store data. Linux automatically takes unused memory for disk buffers. When free memory becomesscarce, buffer frames are automatically released. The value format is a decimal number, with two digits tothe right of the decimal point. An example is 0.70.

Bytes Received The rate per second of bytes received. The value format is a decimal number, with twodigits to the right of the decimal point. An example is 0.70.

Bytes Sent The rate per second of bytes transmitted. The value format is a decimal number, with twodigits to the right of the decimal point. An example is 0.70.

Kernel CPU The percentage of CPU used by this Linux virtual machine, running in kernel mode.

Note: If the APPLDATA values generated for this attribute are created by a Linux guest that is at a levelprior to SUSE Linux Enterprise Server 10 or prior to Red Hat Enterprise Linux 5, normalization of data isapplied. That is, the CPU percent utilization metrics are adjusted based on the virtual CPU percentreported for the Linux guest. The metrics for processor utilization are based within the context of Linuxguests running in the z/VM environment. When steal time is available in the Linux APPLDATA, and Linuxsystems are at a SUSE Linux Enterprise Server 10 level or at a Red Hat Enterprise Linux 5 level,processor utilization metrics are already normalized.The value format is in the range of 0.00 - 100.00. For example, the value 12.00 represents 12.00%.



Linux Guest ID The z/VM identifier for the Linux guest system. This value is blank for all other guestsystems. The value format is an alphanumeric text string with a maximum of 8 characters.

Linux Kernel Product ID The identifier for the Linux kernel product. The value format is an alphanumerictext string with a maximum of 16 characters. The following is an example of the type of information thatdisplays:LINUXKRNL 260100

where:

v LINUXKRNL - Linux kernel.

v 26 - Version 2.6 of the Linux kernel.

v 01 - Release 01.

v 00 - APPLDATA record modification level.

Nice CPU The percentage of CPU time in ’nice’ mode. A process runs in ’nice’ mode if the schedulingpriority is lower than normal. If the superuser has increased the scheduling priority of some processes tovalues higher than normal, this is not considered ’nice’ mode.


No Space Buffer Errors The rate per second, of no space found in Linux buffers. The value format is adecimal number, with two digits to the right of the decimal point. An example is 0.70.

No Space Linux Errors The rate per second, of no space available in Linux. The value format is adecimal number, with two digits to the right of the decimal point. An example is 0.70.

Number of Interfaces The number of networking interfaces defined. The value format is a decimalnumber, with two digits to the right of the decimal point. An example is 0.70.

Packets Received The rate per second of packets received. The value format is a decimal number, withtwo digits to the right of the decimal point. An example is 0.70.

Packets Sent The rate per second of packets transmitted. The value format is a decimal number, with twodigits to the right of the decimal point. An example is 0.70.

Page Allocation Rate The rate of page allocations (the number of pages obtained from the available list),in 4 kilobyte pages per second. The value format is a decimal number, with two digits to the right of thedecimal point. An example is 0.70.

Page Faults Major The rate per second of major page faults for the process. A minor page fault occurs if -from the Linux operating system’s perspective - disk access is involved in handling the page fault. Thevalue format is a decimal number, with two digits to the right of the decimal point. An example is 0.70.

Page Faults Minor The rate per second of minor page faults for the process. A major page fault occurs if- from the Linux operating system’s perspective - disk access is not involved in handling the page fault.The value format is a decimal number, with two digits to the right of the decimal point. An example is 0.70.


Percent CPU Idle The percentage of time spent by the virtual machine in a CPU idle state. The valueformat is in the range of 0.00 - 100.00. For example, the value 12.00 represents 12.00%.

Percent High Memory Utilization The percentage of high memory used. The value format is in the rangeof 0.00 - 100.00. For example, the value 12.00 represents 12.00%.

Percent I/O Wait The percentage of time spent by the virtual machine in an I/O wait state. The valueformat is in the range of 0.00 - 100.00. For example, the value 12.00 represents 12.00%.

Percent IRQ The percentage of interrupts (IRQs). The value format is in the range of 0.00 - 100.00. Forexample, the value 12.00 represents 12.00%.

Percent Main Memory Utilization The percentage of main memory used. The value format is in therange of 0.00 - 100.00. For example, the value 12.00 represents 12.00%.

Percent Stolen The percentage of time that the hypervisor is in control of the processor and its resources.

Percent Swap Space Utilization The percentage of swap space used. The value format is in the range of0.00 - 100.00. For example, the value 12.00 represents 12.00%.

Processes Waiting for I/O The number of processes waiting for I/O. The value format is an integer.

Runnable Processes The number of runnable processes at sampling time. The value format is an integer.

Shared Memory The size of the memory that is usable by more than one process, in megabytes. If anypart of memory can be used by more than one process, it is counted as shared memory. The value formatis a decimal number, with two digits to the right of the decimal point. An example is 0.70.

Percent Soft IRQs The percentage of soft interrupts (IRQs). The value format is in the range of 0.00 -100.00. For example, the value 12.00 represents 12.00%.

Swap In Rate The number of pages swapped in, at the rate of 4-kilobyte pages per second. Note that theLinux kernel uses swapping only if there is not enough memory available. Swapping is normally bad forperformance and application responsiveness. Exploiting VM virtual disks in storage as swapping devicescan help reduce the negative impact to performance or swapping. The value format is a decimal number,with two digits to the right of the decimal point. An example is 0.70.

Swap Out Rate The number of pages swapped out, at the rate of 4-kilobyte pages per second. Note thatthe Linux kernel uses swapping only if there is not enough memory available. Swapping is normally badfor performance and application responsiveness. Exploiting VM virtual disks in storage as swappingdevices can help reduce the negative impact to performance or swapping. The value format is a decimalnumber, with two digits to the right of the decimal point. An example is 0.70.



Time The date and time that summary Linux performance information (APPLDATA) was written by theLinux guest. This attribute is intended for logging and reporting data collection times rather than forcreating situations. See the Tivoli Enterprise Portal help for instructions on specifying timestamp attributes.

Total CPU The percentage of CPU total active time.



Total High Memory The total size of the high memory, in megabytes. The value format is a decimalnumber, with two digits to the right of the decimal point. An example is 0.70.

Total Main Memory The total size of the main memory, in megabytes. This value does not include spacereserved by the kernel. The value format is a decimal number, with two digits to the right of the decimalpoint. An example is 0.70.

Total Processes The total number of processes at sampling time. The value format is an integer.

Total Swap Space The total amount of swap space, both used and available, in megabytes. The valueformat is an integer.

Transmit Collisions The rate per second of collisions while transmitting. The value format is a decimalnumber, with two digits to the right of the decimal point. An example is 0.70.

Transmit Errors The rate per second of packet transmit problems. The value format is a decimal number,with two digits to the right of the decimal point. An example is 0.70.

User CPU The percentage of CPU time in user mode, that is, executing ordinary user code. If a userprogram calls a system routine like open(), the execution time needed for the open() system routine iscounted as kernel mode, while the normal processing of the user program is counted as user mode.


User ID The user identification or group name of the Linux guest. The value format is an alphanumerictext string with a maximum of 8 characters.

Virtual CPUs The number of virtual CPUs defined for the Linux guest system. The value format is aninteger.

See also:


KVLUser Workload attributesThe KVLUser Workload attribute group provides information about workload activity and storage utilization.Use this attribute group to create situations that monitor workload activity to ensure that settings do notreach or exceed predefined thresholds.


The KVLUser Workload attribute group contains the following attributes:

Average Storage Size The average storage size for this workload, in kilobytes. The value format is aninteger.

CP % of CPU The percentage of available CPU used by the system to manage this workload per virtualprocessor. The value format is a percentage value with two decimal places. For example, the value 12.00represents 12.00%.

CPU Percent The percentage of CPU utilized by this workload per virtual processor. The value format is apercentage value with two decimal places. For example, the value 12.00 represents 12.00%.

CP Seconds The total number of z/VM Control Program seconds used by this workload, calculated to thenearest second, during the collection interval. The value format is an integer.

CPU Seconds The total number of CPU seconds used by this workload, calculated to the nearest second,during the collection interval. The value format is an integer.

Used DASD Page Slot Count The number of page slots used on system paging areas on DASD.

Data Spaces Owned The number of non-base address spaces owned by the virtual machine.

Expanded Storage Size in Mbytes The number of pages allocated to this workload that are currentlyresident in expanded storage (without dedicated expanded storage). The value format is an integer.

Expanded Storage Pages Moved Count The rate per second during which page frames were movedfrom expanded storage to main storage, plus the rate at which page frames were migrated out from mainstorage to expanded storage, plus the rate at which page frames were migrated out from expandedstorage to DASD. The value format is an integer.

Expanded Storage Paging Block Count The number of expanded storage blocks allocated to thisworkload by the Control Program for paging. The value format is an integer.

Locked Pages Above 2GB The number of private pages that are locked above the 2-GB line.

Locked Pages Below 2GB The number of pages locked in the user’s private address spaces below the2-GB line for z/VM V5.2 or later. For older levels of z/VM, the number represents the total of the user’slocked private pages.


Linux Guest ID The z/VM identifier for the Linux guest system. This value is blank for all other guestsystems. The value format is an alphanumeric text string with a maximum of 20 characters.

Main Storage Pages Moved to XSTOR per Second The rate at which page frames were migrated outfrom main storage to XSTORE.

Page Rate The rate of page-ins and page-outs for this workload, in pages per second. The value format isa decimal number, with two digits to the right of the decimal point. An example is 0.70.

Page Reads Pages per Second The rate per second at which pages were read in during a pagingoperation, over the specified period of time. The value format is a decimal number, with two digits to theright of the decimal point. An example is 0.70.


Page Steals per Second The rate at which pages were stolen from the virtual machine’s address spaceswhile they were private.

Pages Moved to Below 2GB per Second The rate at which page frames were moved below the 2-GBline for translation.

Page Writes Pages per Second The rate per second at which pages were written out during a pagingoperation, over the specified period of time. The value format is a decimal number, with two digits to theright of the decimal point. An example is 0.70.

Reserved Pages The number of reserved pages for the user.

Resident Pages The current number of pages (below the 2GB line) physically in the main storage for thisworkload. The value format is an integer.

Resident Pages > 2GB The current number of pages (above the 2GB line) physically in the main storagefor this workload. The value format is an integer.

Session Time The total amount of time, to the nearest minute, that this workload was logged on or thetotal amount of time, to the nearest minute, that this aggregation of the group was logged on, during thecollection interval. The value format is an integer.

Share Setting Absolute or relative share value set for the user. A virtual machine’s share is dynamicallydistributed across its started virtual processors rather than its defined virtual processors. See: Type ofShare Setting attribute for whether the setting is relative or absolute.




Total CPU Percent The percentage of total CPU used by the system to manage this workload. The valueformat is a percentage value with two decimal places. For example, the value 12.00 represents 12.00%.The percentage can exceed 100% when multiple virtual processors are used.

Total CP % of CPU The percentage of total CPU used by the z/VM Control Program to manage thisworkload. The value format is in the range of 0.00 - 100.00. For example, the value 12.00 represents12.00%.

Total Virtual CPU % The percentage of total virtual CPU utilization for the workload specified. The valueformat is a percentage value with two decimal places. For example, the value 12.00 represents 12.00%.The percentage can exceed 100% when multiple virtual processors are used.

Share Setting Type Indicates via the text a Relative or Absolute share value, that is the type of share setfor the workload and reported in the Share Setting attribute. The following are the types of share setting:

v No limit: this is the default share setting. It allows z/VM to adjust the users share value according tosystem load with no restrictions.

v Maximum: a maximum share value has been set for the user. The Share value shown in the ShareSetting attribute is the actual value, not the maximum.

v Limit hard: the share value for the user can never exceed the maximum value.


v Absolute: the user will receive a targeted minimum of the system resources.

v Dedicated: the user is running on a dedicated processor, and the Share setting has no meaning.

User ID The identifier of the user or the group name of the workload. The value format is an alphanumerictext string with a maximum of 8 characters.

Virtual CPU % The percentage of virtual CPU utilization for the workload specified per virtual processor.The value format is a percentage value with two decimal places. For example, the value 12.00 represents12.00%.

Virtual CPUs The number of virtual processors in the virtual machine configuration. The value format is aninteger.

Virtual Seconds The total number of virtual CPU seconds used by this workload, calculated to thenearest second, during the collection interval. The value format is an integer.

Working Set Size A user's projected working set size, in pages. This value is calculated each time a userdrops from the queue. The value is based on the number of pages referenced during the last stay in thequeue. The value format is an integer.

Workload Group The name of the group to which this workload belongs. This attribute is primarily used todetermine which virtual machines are Linux guest hosts. This value will either be blank or will displayLINUX if this is a Linux guest host. The value format is an alphanumeric text string with a maximum of 8characters.

XSTOR Pages Moved to Disk per Second The rate at which page frames were migrated out fromXSTORE to DASD. The value format is a decimal number, with one digit to the right of the decimal point.An example is 0.7.

XSTOR Pages Moved to Main Storage per Second The rate at which page frames were moved fromXSTORE to main storage. The value format is a decimal number, with one digit to the right of the decimalpoint. An example is 0.7.

See also:


KVLVdisk attributesThe KVLVdisk attribute group provides information about virtual disk (VDISK) activity. Use this attributegroup to create situations that monitor VDISK activity to ensure that settings do not reach or exceedpredefined thresholds.

Data Space Name The name of the data space. The value format is an alphanumeric text string with amaximum of 24 characters.

Device Number The virtual device number of the VDISK. The value format is an alphanumeric text stringwith a maximum of 8 characters.


Locked Pages The number of locked data space pages. The value format is an integer.

Number of Links The number of links to the virtual disk. The value format is an integer.


Occupied Slots The number of slots occupied on auxiliary storage. The value format is an integer.

Pages from Central Storage per Second The number of data space pages moved from central storageto expanded storage, per second. The value format is a decimal number, with two digits to the right of thedecimal point. An example is 0.70.

Pages Read from DASD per Second The number of data space pages read from DASD, per second.The value format is a decimal number, with two digits to the right of the decimal point. An example is 0.70.

Pages Stolen per Second The number of data space pages stolen per second. The value format is adecimal number, with two digits to the right of the decimal point. An example is 0.70.

Pages to Central Storage per Second The number of data space pages moved from expanded storageto central storage, per second. The value format is a decimal number, with two digits to the right of thedecimal point. An example is 0.70.

Pages to DASD per Second The number of data space pages migrated from expanded storage to DASD,per second. The value format is a decimal number, with two digits to the right of the decimal point. Anexample is 0.70.

Pages Written to DASD per Second The number of data space pages written to DASD, per second. Thevalue format is a decimal number, with two digits to the right of the decimal point. An example is 0.70.

Resident Pages The number of data space pages resident in central storage. The value format is aninteger.




VDISK Owner The user ID of the owner of the VDISK. The value format is an alphanumeric text stringwith a maximum of 8 characters.

VDISK Owner Device Number For display purposes only. This attribute is a combination of the followingtwo attributes:

v VDISK Owner - the user ID of the owner of the VDISK.

v Device Number - the virtual device number of the VDISK.

Note: Do not use this attribute to create a query or to create a situation. Use the VDISK Ownerattribute and the Device Number attributes instead.

VDISK Size The size of the VDISK, expressed as the number of 512-byte blocks for which it was defined.The value format is an integer.

Virtual I/Os per Second The virtual I/O rate, per second, to the VDISK. The value format is a decimalnumber, with two digits to the right of the decimal point. An example is 0.70.

XSTORE Pages The number of XSTORE blocks occupied by the data space. The value format is aninteger.


See also:


KVLVirtualSwitch attributesThe KVLVirtualSwitch attribute group provides data on virtual switch activity during the current collectioninterval. Use the KVLVirtualSwitch attributes to ensure that situation settings do not reach or exceedpredefined thresholds.

Group Name The name of the link aggregation port group in use for this virtual switch. The value formatis an alphanumeric text string with a maximum of 8 characters.

Input Buffers Processed per Second The number of input buffers processed per second. The valueformat is a decimal number, with one digit to the right of the decimal point. An example is 0.5.

Input Queue Overflows per Second The number of input queue overflows per second. The value formatis a decimal number, with one digit to the right of the decimal point. An example is 0.5.

Interrupts Processed per Second The number of interrupts, such as Peripheral Component Interconnect(PCI) interrupts, that resulted in input queue processing, per second. The value format is a decimalnumber, with one digit to the right of the decimal point. An example is 0.5.

Interrupts Received per Second The number of interrupts, such as PCI interrupts, received per second.The value format is a decimal number, with one digit to the right of the decimal point. An example is 0.5.

IP Address The LAN Management IP address. This address is used by clients in switch management. Forexample, a Network Management System can request Bridge MIB information from an SNMP agent usingthe IP address. If an SNMP agent has not been assigned, the IP address will not display. The value formatis an alphanumeric text string with a maximum of 45 characters.

LAN Lock Defers per Second The number of times per second that the CP monitor waited for thenetwork lock. The value format is a decimal number, with one digit to the right of the decimal point. Anexample is 0.5.

LAN Management Userid The LAN Management userid. This is the user ID of the z/VM TCP/IP stackthat has assigned an SNMP agent to the virtual switch in its HOME statement. If an SNMP agent has notbeen assigned, no data displays for this attribute. The value format is an alphanumeric text string with amaximum of 8 characters.


Load Balance Interval in Seconds The load balancing interval. This is the number of seconds betweenbalancing operations. The value format is an integer.

Lock Requests Rate per Second The number of lock requests made for the network lock, per second.The value format is a decimal number, with one digit to the right of the decimal point. An example is 0.5.

MAC Address The LAN Management Media Access Control (MAC) address. The value format is analphanumeric text string with a maximum of 17 characters.

Microcode Level The Open Systems Adapter (OSA) device microcode level. The value format is analphanumeric text string with a maximum of 8 characters.


Output Buffers Processed per Second The number of output buffers processed per second. The valueformat is a decimal number, with one digit to the right of the decimal point. An example is 0.5.

Output Queue Overflows per Second The number of output queue overflows per second. The valueformat is a decimal number, with one digit to the right of the decimal point. An example is 0.5.

QSV The queue storage value. Possible values are 1 through 8. The value format is an alphanumeric textstring with a maximum of 4 characters.

Read Signals per Second The number of read signals issued per second. The value format is a decimalnumber, with one digit to the right of the decimal point. An example is 0.5.

Real Device Address The real device address. The value format is an alphanumeric text string no longerthan 4 characters.

Receive Bytes The rate per second of bytes received. The value format is a decimal number, with twodigits to the right of the decimal point. An example is 0.70.

Receive Lock Defers per Second The number of times per second that CP monitor waited for any lockwhen receiving data on this VSWITCH port. The value format is a decimal number, with one digit to theright of the decimal point. An example is 0.5.

Receive Packets Discarded Rate per Second The rate per second of inbound packets discarded. Thevalue format is a decimal number, with two digits to the right of the decimal point. An example is 0.70.

Receive Packets Rate per Second The rate per second of packets received. The value format is adecimal number, with two digits to the right of the decimal point. An example is 0.70.

Send Lock Defers per Second The number of times per second that CP monitor waited for any lockwhen sending data from this VSWITCH port. The value format is a decimal number, with one digit to theright of the decimal point. An example is 0.5.

Session Layer The session layer, either 2 or 3 of the Open Systems Interconnection (OSI) seven-layermodel. The value format is an integer.

Sync Signals per Second The number of sync signals issued per second. The value format is a decimalnumber, with one digit to the right of the decimal point. An example is 0.5.




Total LACP PDUs Received The number of Link Aggregation Control Protocol (LACP) Packet Data Units(PDUs) received on this port. This is the total value for the monitor interval. The value format is an integer.

Total LACP PDUs Sent The number of LACP PDUs sent on this port. This is the total value for themonitor interval. The value format is an integer.

Total PDU Responses Received The number of marker PDUs received. This is the total value for themonitor interval. The value format is an integer.


Total PDU Responses Sent The number of marker PDUs sent to this port in response to receiving amarker PDU from the partner port. This is the total value for the monitor interval. The value format is aninteger.

Total PDUs Sent The number of marker PDUs sent to this port. This is the total value for the monitorinterval. The value format is an integer.

Total Timeouts The total number of times the virtual switch timed out while waiting for a marker responsePDU for a marker request sent by CP monitor to a partner port. The value format is an integer.

Transmit Bytes The rate per second of bytes transmitted. The value format is a decimal number, with twodigits to the right of the decimal point. An example is 0.70.

Transmit Packets Rate per Second The rate per second of packets transmitted. The value format is adecimal number, with two digits to the right of the decimal point. An example is 0.70.

Transmit Packets Discarded Rate per Second The rate per second of outbound packets discarded. Thevalue format is a decimal number, with two digits to the right of the decimal point. An example is 0.70.

User ID The user ID of the virtual machine to which the device is currently attached. The value format isan alphanumeric text string no longer than 8 characters.

VSWITCH Name The LAN identifier for the virtual switch. When a virtual switch is created, it is assignedan 8-byte identifier, or LAN ID.

VSWITCH Timeout The timeout value, in seconds, for the virtual switch. The value format is an integer.

Write Signals per Second The number of write signals issued per second. The value format is a decimalnumber, with one digit to the right of the decimal point. An example is 0.5.

See also:



Support for problem solving

If you have a problem with your IBM software, you want to resolve it quickly. This section describes thefollowing options for obtaining support for IBM software products:

v “Using IBM Support Assistant”

v “Obtaining fixes”

v “Receiving weekly support updates” on page 150

v “Contacting IBM Software Support” on page 150

Using IBM Support AssistantThe IBM Support Assistant is a free, stand-alone application that you can install on any workstation. Youcan then enhance the application by installing product-specific plug-in modules for the IBM products youuse.

The IBM Support Assistant saves you the time it takes to search the product, support, and educationalresources. The IBM Support Assistant helps you gather support information when you need to open aproblem management record (PMR), which you can then use to track the problem.

The product-specific plug-in modules provide you with the following resources:

v Support links

v Education links

v Ability to submit problem management reports

For more information, and to download the IBM Support Assistant, see http://www.ibm.com/software/support/isa. After you download and install the IBM Support Assistant, follow these steps to install theplug-in for your Tivoli product:

1. Start the IBM Support Assistant application.

2. Select Updater on the Welcome page.

3. Select New Properties and Tools or select the New Plug-ins tab (depending on the version of IBMSupport Assistant installed).

4. Under Tivoli, select your product, and then click Install. Be sure to read the license and description.

If your product is not included on the list under Tivoli, no plug-in is available yet for the product.

5. Read the license and description, and click I agree.

6. Restart the IBM Support Assistant.

Obtaining fixesA product fix might be available to resolve your problem. To determine which fixes are available for yourTivoli software product, follow these steps:

1. Go to the IBM Software Support Web site at http://www.ibm.com/software/support.

2. Under Select a brand and/or product, select Tivoli.

If you click Go, the Search within all of Tivoli support section is displayed. If you don't click Go, yousee the Select a product section.

3. Select your product and click Go.

4. Under Download, click the name of a fix to read its description and, optionally, to download it.

If there is no Download heading for your product, supply a search term, error code, or APAR numberin the field provided under Search Support (this product), and click Search.


http://www.ibm.com/software/support/isa

http://www.ibm.com/software/support/isa

http://www.ibm.com/software/support

You can find information specific to the Tivoli OMEGAMON XE on z/VM and Linux monitoring agent at thefollowing Web address:

http://www-306.ibm.com/software/sysmgmt/products/support/IBMTivoliOMEGAMONonzVMLinux.html.

The support page for this monitoring agent displays.

For more information about the types of fixes that are available, see the IBM Software Support Handbookat http://techsupport.services.ibm.com/guides/handbook.html.

Receiving weekly support updatesTo 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 support in the far upper-right corner of the page under Personalized support.

3. If you have already registered for My support, sign in and skip to the next step. If you have notregistered, click register now. Complete the registration form using your e-mail address as your IBMID and click Submit.

4. The Edit profile tab is displayed.

5. In the first list under Products, select Software. In the second list, select a product category (forexample, Systems and Asset Management). In the third list, select a product sub-category (forexample, Application Performance & Availability or Systems Performance). A list of applicableproducts is displayed.

6. Select the products for which you want to receive updates. For example, IBM Tivoli OMEGAMON XEon z/VM and Linux.

7. Click Add products.

8. After selecting all products that are of interest to you, click Subscribe to email on the Edit profiletab.

9. In the Documents list, select Software.

10. Select Please send these documents by weekly email.

11. Update your e-mail address as needed.

12. Select the types of documents you want to receive.

13. Click Update.

If you experience problems with the My support feature, you can obtain help in one of the following ways:

OnlineSend an e-mail message to [email protected], describing your problem.

By phoneCall 1-800-IBM-4You (1-800-426-4968).

Contacting IBM Software SupportIBM Software Support provides assistance with product defects. The easiest way to obtain that assistanceis to open a PMR or ETR directly from the IBM Support Assistant (see “Using IBM Support Assistant” onpage 149).

Before contacting IBM Software Support, your company must have an active IBM software maintenancecontract, and you must be authorized to submit problems to IBM. The type of software maintenancecontract that you need depends on the type of product you have:


http://www-306.ibm.com/software/sysmgmt/products/support/IBMTivoliOMEGAMONonzVMLinux.html

http://techsupport.services.ibm.com/guides/handbook.html

http://www.ibm.com/software/support

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:

OnlineGo to the Passport Advantage Web site at http://www-306.ibm.com/software/howtobuy/passportadvantage/pao_customers.htm .

By phoneFor the phone number to call in your country, go to the IBM Software Support Web site athttp://techsupport.services.ibm.com/guides/contacts.html and click the name of your geographicregion.

v For customers with Subscription and Support (S & S) contracts, go to the Software Service RequestWeb site at https://techsupport.services.ibm.com/ssr/login.

v For customers with IBMLink, CATIA, Linux, OS/390®, iSeries®, pSeries®, zSeries, and other supportagreements, go to the IBM Support Line Web site at http://www.ibm.com/services/us/index.wss/so/its/a1000030/dt006.

v For IBM eServer™ software products (including, but not limited to, DB2 and WebSphere products thatrun in zSeries, pSeries, and iSeries environments), you can purchase a software maintenanceagreement by working directly with an IBM sales representative or an IBM Business Partner. For moreinformation about support for eServer software products, go to the IBM Technical Support AdvantageWeb site at 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 SoftwareSupport Handbook on the Web at http://techsupport.services.ibm.com/guides/contacts.html and click thename 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 152

Determining the business impactWhen you report a problem to IBM, you are asked to supply a severity level. Use the following criteria tounderstand and assess the business impact of the problem that you are reporting:

Severity 1The problem has a critical business impact. You are unable to use the program, resulting in acritical impact on operations. This condition requires an immediate solution.

Severity 2The problem has a significant business impact. The program is usable, but it is severely limited.

Severity 3The problem has some business impact. The program is usable, but less significant features (notcritical to operations) are unavailable.

Severity 4The problem has minimal business impact. The problem causes little impact on operations, or areasonable circumvention to the problem was implemented.

Describing problems and gathering informationWhen describing a problem to IBM, be as specific as possible. Include all relevant background informationso that IBM Software Support specialists can help you solve the problem efficiently. To save time, knowthe answers to these questions:

Support for problem solving 151

http://www.lotus.com/services/passport.nsf/ WebDocs/Passport_Advantage_Home

http://www.lotus.com/services/passport.nsf/ WebDocs/Passport_Advantage_Home

http://techsupport.services.ibm.com/guides/contacts.html

https://techsupport.services.ibm.com/ssr/login

http://www.ibm.com/services/us/index.wss/so/its/a1000030/dt006

http://www.ibm.com/services/us/index.wss/so/its/a1000030/dt006

http://www.ibm.com/servers/eserver/techsupport.html


v Which 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 SoftwareSupport 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 workaroundwhen you report the problem.

Submitting problemsYou can submit your problem to IBM Software Support in one of two ways:

OnlineClick 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 submissionform.

By phoneFor the phone number to call in your country, go to the contacts page of the IBM Software SupportHandbook at http://techsupport.services.ibm.com/guides/contacts.html and click the name of yourgeographic region.

If the problem you submit is for a software defect or for missing or inaccurate documentation, IBMSoftware Support creates an Authorized Program Analysis Report (APAR). The APAR describes theproblem in detail. Whenever possible, IBM Software Support provides a workaround that you canimplement until the APAR is resolved and a fix is delivered. IBM publishes resolved APARs on theSoftware Support Web site daily, so that other users who experience the same problem can benefit fromthe same resolution.


http://www.ibm.com/software/support/probsub.html

http://www.ibm.com/software/support/probsub.html


Documentation library

This appendix contains information about the publications related to OMEGAMON XE on z/VM and Linuxand to IBM Tivoli Monitoring and the commonly shared components of Tivoli Management Services. Thesepublications are listed in the following categories:

v “OMEGAMON XE on z/VM and Linux library”

v “IBM Tivoli Monitoring publications” on page 154

v “Related publications” on page 155

See IBM Tivoli Monitoring and OMEGAMON XE Products: Documentation Guide, SC23-8816, forinformation about accessing and using the publications. You can find the Documentation Guide, underMonitoring, in the IBM Tivoli Monitoring and OMEGAMON XE Information Center at http://publib.boulder.ibm.com/infocenter/tivihelp/v15r1/.

To find a list of new and changed publications, click What's new on the Welcome page of the IBM TivoliMonitoring and OMEGAMON XE Information Center. To find publications from the previous version of aproduct, click Previous information centers on the Welcome page for the product.

OMEGAMON XE on z/VM and Linux libraryThe following publications comprise the OMEGAMON XE on z/VM and Linux library:

v IBM Tivoli OMEGAMON XE on z/VM and Linux Program Directory, GI11-4135

Provides hardware and software prerequisites for the OMEGAMON XE on z/VM and Linux installationand instructions for the VMSES/E part of the installation.

v IBM Tivoli OMEGAMON XE on z/VM and Linux Planning and Configuration Guide, SC27-2837

Provides information for installing and configuring OMEGAMON XE on z/VM and Linux.

v IBM Tivoli OMEGAMON XE on z/VM and Linux User's Guide, SC27-2836

Introduces the features and describes the workspaces, attributes, and predefined situations for theOMEGAMON XE on z/VM and Linux product. Supplements the online help provided with this product byincluding product-specific monitoring scenarios.

v IBM Tivoli OMEGAMON XE on z/VM and Linux Troubleshooting Guide, GC27-2838

Provides information and messages to assist you in troubleshooting problems with the OMEGAMON XEon z/VM and Linux monitoring agent.

v IBM Tivoli OMEGAMON XE on z/VM and Linux Readme , GI11-9442

Contains late-breaking information about the OMEGAMON XE on z/VM and Linux product limitationsand workarounds.

v IBM Tivoli OMEGAMON XE on z/VM and Linux Quick Start Guide , GI11-9441

Provides overview information to get you started installing and configuring OMEGAMON XE on z/VMand Linux.

You can access the OMEGAMON XE on z/VM and Linux library at the following Web address:

http://publib.boulder.ibm.com/infocenter/tivihelp/v15r1/topic/com.ibm.omegamon_xezvm.doc_4.2/welcome.htm

You can find additional information at the following Web address: http://www.ibm.com/software/sysmgmt/products/support

Select IBM Tivoli OMEGAMON XE on z/VM and Linux from the Support for specific Tivoli productsdrop-down list. You will be taken to the support page for this monitoring agent.


http://publib.boulder.ibm.com/infocenter/tivihelp/v15r1/




http://www.ibm.com/software/sysmgmt/products/support

http://www.ibm.com/software/sysmgmt/products/support

OMEGAMON XE on z/VM and Linux online helpThe online help aids operators in understanding and using the provided data, attributes, and situations tomonitor performance and availability in the context of this product.

Note: This set of information is nested within the Tivoli Enterprise Portal help, which describes globalfeatures that are useful in helping you to use OMEGAMON XE on z/VM and Linux. For more information,see “Tivoli Enterprise Portal help system” on page 155.

IBM Tivoli Monitoring publicationsTo use the information in this guide effectively, you must have some prerequisite knowledge about IBMTivoli Monitoring (also called IBM Tivoli Management Services) and the Tivoli Enterprise Portal interface,which you can obtain from the following guides. The publications are available in the IBM Tivoli MonitoringInformation Center at the following Web address:


v IBM Tivoli Monitoring Installation and Setup Guide, GC32-9407

Provides information on installing and setting up the Tivoli Enterprise Monitoring Server and the TivoliEnterprise Portal Server and client. It also describes how to install the distributed monitoring agents.

v Configuring IBM Tivoli Enterprise Monitoring Server on z/OS, SC27-2313

Describes how to configure and customize the Tivoli Enterprise Monitoring Server on z/OS. Thepublication also contains planning information and information about setting up security on your TivoliEnterprise Monitoring Server.

v IBM Tivoli Monitoring Administrator's Guide, SC32-9408

Describes the support tasks and functions required for the Tivoli Enterprise Portal Server and clients,including Tivoli Enterprise Portal user administration.

v IBM Tivoli Monitoring Enterprise Portal User's Guide, SC32-9409

Describes how to use the Tivoli Enterprise Portal interface. It includes a tutorial about monitoring thatcovers workspaces, navigation, views, and responding to alerts. Different types of views and situationsfor event-based monitoring are also included, as well as information on automation policies.

v Exploring IBM Tivoli Monitoring, SC32-1803

Provides a series of exercises that help users explore ITM Tivoli Monitoring.

v IBM Tivoli Universal Agent API and Command Programming Reference Guide, SC32-9461

Contains the procedures for implementing the IBM Tivoli Universal Agent application programminginterfaces (APIs) and provides descriptions, syntax, and return status codes for the API calls andcommand-line interface commands.

v IBM Tivoli Monitoring Troubleshooting Guide, GC32-9458

Provides information and messages to assist users with troubleshooting problems with the software.

v IBM Tivoli Monitoring: Upgrading from Tivoli Distributed Monitoring, GC32-9459

Provides information on how to upgrade from Tivoli Distributed Monitoring.

v IBM Tivoli Universal Agent User's Guide, SC32-9459

Introduces you to the IBM Tivoli Universal Agent, an agent of IBM Tivoli Monitoring. The IBM TivoliUniversal Agent enables you to use the monitoring and automation capabilities of IBM Tivoli Monitoringto monitor any type of data you collect.

v Introducing IBM Tivoli Monitoring, GI11-4071

Gives a basic introduction to the features of IBM Tivoli Monitoring.

v IBM Tivoli OMEGAMON XE and Tivoli Management Services on z/OS Upgrade Guide, GC32-1980



Provides instructions for performing a staged upgrade for OMEGAMON XE V4.1.0 monitoring agents,discusses the basic upgrade requirements and the effects of running a mixed environment forglobalization and data presentation. This guide also lists the required sequence of steps for performinga staged upgrade in two scenarios.

Additionally, the following user's guide provides information about the IBM Tivoli Monitoring for Linux OSagent. The OMEGAMON XE on z/VM and Linux monitoring agent contains dynamic links to workspaces inthe IBM Tivoli Monitoring for Linux OS agent. See “Dynamic workspace linking” on page 18 for details.

v IBM Tivoli Monitoring: Agent for Linux OS User's Guide, SC32-9447

Provides information about using the IBM Tivoli Monitoring for Linux OS agent. This includesdescriptions of the workspaces, the situations, and the attributes provided by this agent.

Tivoli Enterprise Portal help systemThe Tivoli Enterprise Portal help system provides context-sensitive reference information regarding IBMTivoli Monitoring. The help system also documents features and customization options, such as the LinkWizard that helps you to modify the default linkages provided in the pop-up menus of rows in aworkspace.

Related publicationsYou can access the entire library for the IBM z/VM Operating System at the following Web address:

http://www.vm.ibm.com/library/

The following z/VM publications are relevant to this monitoring agent:

v z/VM: Performance Toolkit Guide, SC24-6156

This publication provides, in concert with theSC24-6157, contains all information required for using thePerformance Toolkit for VM (previously known as the VM/ESA® Full Screen Operator Console andGraphical Real Time Monitor, FCON/ESA, or just FCON).

v z/VM: Performance Toolkit Reference, SC24-6157

This publication, in concert with thez/VM: Performance Toolkit Guide, contains all information requiredfor using the Performance Toolkit for VM (previously known as the VM/ESA Full Screen OperatorConsole and Graphical Real Time Monitor, FCON/ESA, or just FCON).

You can also access information on the Performance Toolkit and on the latest enhancements to thePerformance Toolkit as they pertain to this monitoring agent at the following Web address:


v z/VM: CP Commands and Utilities Reference, SC24-6081

Lists and describes IBM z/VM Control Program (CP) commands and utilities for users of every privilegeclass.

v z/VM: CMS Commands and Utilities Reference, SC24-6073

Provides reference information about Conversational Monitor System (CMS) commands and utilities forIBM z/VM.

v Device Drivers, Features and Commands, SC33-8281

Provides information about the device drivers available to Linux for the control of zSeries® and S/390®

devices and attachments with the kernel 2.6 (April 2004 stream). It also provides information oncommands and parameters relevant to configuring Linux for zSeries and S/390®.

v Saved Segments Planning and Administration, SC24-6116

Provides general information on Discontinuous Saved Segments (DCSS).

v Directory Maintenance VM/ESA V1R5.0 Command Reference, SC20-1839

Provides reference information for the z/VM Directory Maintenance Facility (DirMaint™) Function Level510, for use on IBM z/VM Version 5.

Documentation library 155

http://www.vm.ibm.com/library/


You may also want to review the IBM Redbooks® for z/VM, as well as those listed below, at the followingWeb address:

http://www.vm.ibm.com/pubs/redbooks/

v Running Linux on IBM System z9® and zSeries under z/VM, SG24-6311

v z/VM and Linux on IBM System z: The Virtualization Cookbook for SLES9, SG24-6695

v Linux on IBM eServer zSeries and S/390: Performance Measurement and Tuning,

You can find links to hints and tips about z/VM performance at the following Web site:

http://vm.ibm.com/perf

Additionally, the following Web site contains links to a collection of observations from the LinuxPerformance Team for Linux on System z:

http://www.vm.ibm.com/developerworks/linux/linux390/perf/index.html

Accessing terminology onlineThe 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 convenientlocation. You can access the Terminology Web site at the following Web address:

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

Accessing publications onlineThe documentation CD contains the publications that are in the product library. The format of thepublications is PDF, HTML, or both. Refer to the readme file on the CD for instructions on how to accessthe documentation.

IBM posts publications for this and all other Tivoli products, as they become available and whenever theyare updated, to the Tivoli Information Center Web site at at the following Web address:

http://publib.boulder.ibm.com/tividd/td/link/tdprodlist.html.

In the Tivoli Software Information Center window, click the letter that matches the first letter of your productname to access your product library. For example, click M to access the IBM Tivoli Monitoring library orclick O to access the IBM Tivoli OMEGAMON library.

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

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

http://www.elink.ibmlink.ibm.com/publications/servlet/pbi.wss

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

v In the United States: 800-879-2755


http://www.vm.ibm.com/pubs/redbooks/

http://www.vm.ibm.com/perf

http://www.ibm.com/developerworks/linux/linux390/perf/index.html

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

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

http://publib.boulder.ibm.com/tividd/td/link/tdprodlist.html

http://www.elink.ibmlink.ibm.com/publications/servlet/pbi.wss

v In Canada: 800-426-4968

In other countries, contact your software account representative to order Tivoli publications. To locate thetelephone number of your local representative, perform the following steps:

1. Go to http://www.elink.ibmlink.ibm.com/publications/servlet/pbi.wss.

2. Select your country from the list and click Go.

3. Click About this site in the main panel to see an information page that includes the telephone numberof your local representative.

AccessibilityAccessibility features help users with a physical disability, such as restricted mobility or limited vision, touse software products successfully. With this product, you can use assistive technologies to hear andnavigate the interface. You can also use the keyboard instead of the mouse to operate all features of thegraphical user interface.

Documentation library 157

Notices

This information was developed for products and services offered in the U.S.A. IBM may not offer theproducts, services, or features discussed in this document in other countries. Consult your local IBMrepresentative for information on the products and services currently available in your area. Any referenceto an IBM product, program, or service is not intended to state or imply that only that IBM product,program, or service may be used. Any functionally equivalent product, program, or service that does notinfringe any IBM intellectual property right may be used instead. However, it is the user's responsibility toevaluate and verify the operation of any non-IBM product, program, or service.

IBM may 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 licenseinquiries, in writing, to:

IBM Director of LicensingIBM CorporationNorth Castle DriveArmonk, NY 10504-1785 U.S.A.

For license inquiries regarding double-byte (DBCS) information, contact the IBM Intellectual PropertyDepartment in your country or send inquiries, in writing, to:

IBM World Trade Asia CorporationLicensing2-31 Roppongi 3-chome, Minato-kuTokyo 106, Japan

The following paragraph does not apply to the United Kingdom or any other country where suchprovisions 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 NOTLIMITED TO, THE IMPLIED WARRANTIES OF NON-INFRINGEMENT, MERCHANTABILITY OR FITNESSFOR A PARTICULAR PURPOSE.

Some states do not allow disclaimer of express or implied warranties in certain transactions, therefore, thisstatement might not apply to you.

This information could include technical inaccuracies or typographical errors. Changes are periodicallymade to the information herein; these changes will be incorporated in new editions of the publication. IBMmay make improvements and/or changes in the product(s) and/or the program(s) described in thispublication at any time without notice.

Any references in this information to non-IBM Web sites are provided for convenience only and do not inany manner serve as an endorsement of those Web sites. The materials at those Web sites are not part ofthe materials for this IBM product and use of those Web sites is at your own risk.

IBM may use or distribute any of the information you supply in any way it believes appropriate withoutincurring any obligation to you.

Licensees of this program who wish to have information about it for the purpose of enabling: (i) theexchange 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 Corporation2Z4A/10111400 Burnet RoadAustin, TX 78758 U.S.A.

Such information may be available, subject to appropriate terms and conditions, including in some casespayment of a fee.

The licensed program described in this document and all licensed material available for it are provided byIBM under terms of the IBM Customer Agreement, IBM International Program License Agreement or anyequivalent agreement between us.

Any performance data contained herein was determined in a controlled environment. Therefore, the resultsobtained in other operating environments may vary significantly. Some measurements may have beenmade on development-level systems and there is no guarantee that these measurements will be the sameon generally available systems. Furthermore, some measurement may have been estimated throughextrapolation. Actual results may vary. Users of this document should verify the applicable data for theirspecific environment.

Information concerning non-IBM products was obtained from the suppliers of those products, theirpublished announcements or other publicly available sources. IBM has not tested those products andcannot 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.

All IBM prices shown are IBM's suggested retail prices, are current and are subject to change withoutnotice. Dealer prices may vary.

This information is for planning purposes only. The information herein is subject to change before theproducts described become available.

This information contains examples of data and reports used in daily business operations. To illustratethem as completely as possible, the examples include the names of individuals, companies, brands, andproducts. All of these names are fictitious and any similarity to the names and addresses used by anactual business enterprise is entirely coincidental.

COPYRIGHT LICENSE:

This information contains sample application programs in source language, which illustrate programmingtechniques on various operating platforms. You may copy, modify, and distribute these sample programs inany form without payment to IBM, for the purposes of developing, using, marketing or distributingapplication programs conforming to the application programming interface for the operating platform forwhich the sample programs are written. These examples have not been thoroughly tested under allconditions. IBM, therefore, cannot guarantee or imply reliability, serviceability, or function of theseprograms. You may copy, modify, and distribute these sample programs in any form without payment toIBM for the purposes of developing, using, marketing, or distributing application programs conforming toIBM‘s application programming interfaces.

Each copy or any portion of these sample programs or any derivative work, must include a copyrightnotice as follows:

© (your company name) (year). Portions of this code are derived from IBM Corp. Sample Programs. ©Copyright IBM Corp. _enter the year or years_. All rights reserved.


If you are viewing this information in softcopy form, the photographs and color illustrations might notdisplay.

TrademarksIBM, the IBM logo, ibm.com®, Performance Toolkit for VM™, and z/VM® are trademarks or registeredtrademarks of International Business Machines Corporation in the United States, other countries, or both. Ifthese and other IBM trademarked terms are marked on their first occurrence in this information with atrademark symbol (® or ™), these symbols indicate U.S. registered or common law trademarks owned byIBM at the time this information was published. Such trademarks may also be registered or common lawtrademarks in other countries. A current list of IBM trademarks is available on the Web at "Copyright andtrademark information" at http://www.ibm.com/legal/copytrade.shtml.

Adobe®, Acrobat, Portable Document Format (PDF), and PostScript® are either registered trademarks ortrademarks of Adobe Systems Incorporated in the United States, other countries, or both.

Java™ and all Java-based trademarks and logos are trademarks or registered trademarks of SunMicrosystems, Inc. in the United States, other countries, or both.

Linux is a registered trademark of Linus Torvalds in the United States, other countries, or both.

Microsoft and Windows are trademarks of Microsoft Corporation in the United States, other countries, orboth.

UNIX is a registered trademark of The Open Group in the United States and other countries.

Other company, product, or service names may be trademarks or service marks of others.

Notices 161

http://www.ibm.com/legal/copytrade.shtml

Glossary

Aadding application support. Before you can use amonitoring agent, the Tivoli Enterprise Monitoring Serverto which it reports must be initialized with applicationdata. This step adds product-provided situations,templates, and other sample data to the TivoliEnterprise Monitoring Server Enterprise InformationBase (EIB) tables.

application data. The product-provided situations,templates, and other sample data to the EnterpriseInformation Base (EIB) tables of the Tivoli EnterpriseMonitoring Server.

attribute. A system or application element beingmonitored by the monitoring agent, such as Disk Nameand Disk Read/Writes Per Second. An attribute can alsobe a field in an ODBC-compliant database.

attribute table. A set of related attributes that can becombined in a data view or a situation. When you openthe view or start the situation, Tivoli Enterprise Portalretrieves data samples of the selected attributes. Eachtype of agent has a set of attribute groups.

authorized virtual machine. In z/VM, a GCS virtualmachine associated with an authorized user ID.

CCCW. A doubleword at the location in main storagespecified by the channel address word. One or moreCCWs make up the channel program that directs datachannel operations.

channel (CHN). (1) A path along which signals can besent, for example, an input/output channel. (2) Thesystem element that controls one channel path, whosemode of operation depends on the type of hardware towhich it is attached.

channel address word (CAW). An area in storagethat specifies the location in main storage at which achannel program begins.

channel path (CHP). A single interface between acentral processor and one or more control units alongwhich signals and data can be sent to perform I/Orequests.

channel path identifier (CHPID). In a channelsubsystem, a value assigned to each installed channelpath of the system that uniquely identifies that path tothe system.

channel program. One or more channel commandwords (CCWs) that direct the operation of a datachannel.

client/server. In communications, the model ofinteraction in distributed data processing in which aprogram at one site sends a request to a program atanother site and awaits a response. The requestingprogram is called a client; the answering program iscalled a server.

command. A request from a user to the z/VM systemto perform a particular operation. A CMS command canalso be the name of a CMS file with a file type of EXECor MODULE, in which a sequence of operations aredefined. See also Conversational Monitor System(CMS).

Configuration Tool. Tool used to install the previousCandle products and to configure some of the TivoliMonitoring Agent zSeries products (which are nowinstalled using the System ModificationProgram/Extended (SMP/E) tool).

Control Program (CP). A component of the z/VMsystem that manages the resources of a singlecomputer so that multiple computing systems seem toexist. Each apparent system, or virtual machine, is thefunctional equivalent of the real computer, and CPsimulates the real machine architecture in the virtualmachine.

Conversational Monitor System (CMS). Acomponent of z/VM that runs in a virtual machine andprovides both the interactive z/VM end-user interfaceand the general z/VM application programminginterface. CMS runs only under the control of the z/VMControl Program (CP).

Ddevice driver. In z/VM: (1) A file that contains thecode needed to use an attached device. (2) A programthat enables a computer to communicate with a specificperipheral device, such as a printer or a CD drive. (3) Acollection of subroutines that control the interfacebetween I/O device adapters and the processor. (4) InCMS Pipelines, a stage that reads data from or writesdata to I/O and storage devices, host environments(including CP, CMS, and XEDIT), and REXX and EXEC2 variables.

direct access storage device (DASD). In z/VM, amass storage medium in which the data access time iseffectively independent of the data location.

discontiguous saved segment (DCSS). In z/VM, asaved segment that begins and ends on a megabyte


boundary and is not a segment space or a member of asegment space. A DCSS can contain logical savedsegments. See also segment space.

Eexpert advice. A description within the Situation Editorof each situation provided with a monitoring agent tohelp you quickly gather and interpret data.

Gguest operating system. An operating system, suchas Linux or z/OS, running in a virtual machine managedby the z/VM Control Program (CP).

guest virtual machine. In z/VM, a virtual machine inwhich an operating system is running.

guest virtual storage. In z/VM, the storage thatappears to the operating system running in a virtualmachine.

HHiperSockets. (1) A hardware channel that provideshigh-speed TCP/IP communication between logicalpartitions (LPARs) on the same IBM zSeries server. Ituses an adaptation of the queued direct I/O (QDIO)architecture. (2) The virtualization of the HiperSocketschannel in z/VM, which provides high-speedcommunication between guest virtual machines.

historical data management. The procedures appliedto short-term binary history files that perform roll off toeither a data warehouse or to delimited text files anddelete entries in the short-term history files over 24hours old to make room for new entries.

host. The z/VM Control Program (CP) in its capacityas manager of a virtual machine in which anotheroperating system is running.

hub Tivoli Enterprise Monitoring Server. The TivoliEnterprise Monitoring Server that has been elected toact as the focal point to which all Tivoli Enterprise PortalServers connect.

IIBM Tivoli Monitoring. A client-server implementationcomprising a Tivoli Enterprise Monitoring Server, TivoliEnterprise Portal Server, Tivoli Enterprise Portal client,and monitoring agents that collect and distribute data tothe Tivoli Enterprise Monitoring Server.

inter-user communication vehicle (IUCV). In z/VM, aCP interface for passing data between virtual machinesor between CP and a virtual machine.

Mmember saved segment. In z/VM, a saved segmentthat begins and ends on a page boundary and belongsto up to 64 segment spaces. A member saved segmentis accessed by its own name or by the name of asegment space to which it belongs. A member savedsegment may contain logical saved segments. Seediscontiguous saved segment (DCSS) and segmentspace.

monitoring agent. The agent process probes amanaged system for data and sends information back toTivoli Enterprise Portal formatted into table and chartviews.

monitoring interval. A specified time, scalable toseconds, minutes, hours, or days, for how often theTivoli Enterprise Monitoring Server checks to see if asituation has become true. The minimum monitoringinterval is 30 seconds; the default is 15 minutes.

Llogical partition (LPAR). In z/VM, a subset of theprocessor hardware that is defined to support theoperation of a system control program (SCP).

NNavigator. The left pane of the Tivoli Enterprise Portalwindow. The Navigator Physical view shows yournetwork enterprise as a physical hierarchy of systemsgrouped by platform. OMEGAMON DE users can alsocreate other views to create logical hierarchies groupedas you specify, such as by department or function.

OOMEGAMON Web Services. An openstandards-based interface to IBM Tivoli Monitoring usingSOAP requests. Any Tivoli Monitoring Agent can bedynamically queried, so performance and availabilitydata can be processed by other applications.

Ppaging. In z/VM, transferring pages between realstorage and external page storage.

Performance Toolkit for VM. An optional feature ofz/VM that gathers, analyzes, and displays VMperformance data. It can also process Linuxperformance data obtained from the ResourceManagement Facility.

privilege class. In z/VM, the authorization required touse a subset of the CP commands and DIAGNOSEcodes and certain CP system functions. The privilege


class corresponds to a type of user. Each CP commandand DIAGNOSE code belongs to one or more privilegeclasses. A user is assigned one or more privilegeclasses in the z/VM directory entry for that virtualmachine.

Rremote Tivoli Enterprise Monitoring Server. TheTivoli Enterprise Monitoring Server that passes itscollected data to the hub Tivoli Enterprise MonitoringServer to be made available to clients, creating anenterprise-wide view.

Sseeding. See adding application support.

shared segment. See discontiguous savedsegment (DCSS).

situation. A set of conditions that, when met, createsan event. A condition consists of an attribute, anoperator such as greater than or equal to, and a value.It can be read as, “If - system condition - compared to -value - is true.” An example of a situation is: IF - CPUusage - GT - 90% - TRUE. The expression “CPU usageGT 90%” is the situation condition.

TTake Action. A command window on Tivoli EnterprisePortal from which you can enter your command orchoose from a list of predefined commands. It also hasa list of systems on which to effect the command. Thismonitoring agent has specific requirements for enablingTake Action commands. See the IBM TivoliOMEGAMON XE on z/VM and Linux Planning andConfiguration Guide for details.

target libraries. SMP/E-controlled libraries that containthe data from the distribution media

threshold. A level set in the system at which amessage is sent or an error-handling program is called.For example, in a user auxiliary storage pool, the usercan set the threshold level in the system values, andthe system notifies the system operator when that levelis reached.

Tivoli Enterprise Monitoring Server. The IBM TivoliMonitoring component that retrieves data from themonitoring agents and delivers data to the TivoliEnterprise Portal Server, sends alerts to the TivoliEnterprise Portal Server when conditions specified insituations are met, receives commands from the TivoliEnterprise Portal and passes them to the appropriatemonitoring agents, and (optionally) provides a repository

for short-term historical data. This component can beinstalled on z/OS, Windows, and some UNIX operatingsystems.

Tivoli Enterprise Portal Server. The IBM TivoliMonitoring server you log on to. The Tivoli EnterprisePortal Server connects to the hub Tivoli EnterpriseMonitoring Server. It enables retrieval, manipulation andanalysis of data from IBM Tivoli Monitoring managedsystems.

Vview. A windowpane, or frame, in a workspace. Itmight contain data from an agent in a chart or table, orit might contain a terminal session or browser, forexample. A view can be split into two separate,autonomous views.

Wworking set. The estimated number of pages of realstorage that a virtual machine needs to run.

workspace. The viewing area of the Tivoli EnterprisePortal window, excluding the Navigator. Eachworkspace comprises one or more views. EveryNavigator item has its own default workspace and mighthave multiple workspaces.

Glossary 165

Index

Aaccessibility 157agents, monitoring 9alerts 63ApplData

workspace 59ApplData workspace

enabling data collection for KVLUser ApplDataattributes 60

attribute groupsOMEGAMON XE on z/VM and Linux product 101used by predefined workspaces 29

attributesattribute groups 101defined 22KVLChannel Data 102KVLControlUnit 104KVLCP Device 101KVLDASDCache 106KVLDevice 110KVLFChannel Data 111KVLHiperSocket 112KVLLChannel Data 113KVLLPAR Info 115KVLMinidisk Cache 118KVLProcessor Data 121KVLPTKStat 123KVLSpinLock 124KVLSystem 125, 129KVLTCPIP Srvr Data 133KVLTCPIPUsrData 135KVLUser ApplData 136KVLUser Wait 131KVLUser Workload 140KVLVirtualSwitch 145names 101OMEGAMON XE on z/VM and Linux product 101using in queries 23

audienceexpertise xiresponsibilities xi

Bbrowser client 10

Cclient 10collection interval

defining 21components

IBM Tivoli Monitoring 9conditions, situation 65CP CPU utilization

monitoring 95

CP Monitor sample intervalsetting 22

customer supportSee Software Support

Ddesktop client 10

EEnterprise Status

Message Log 16My Acknowledged Events 15Open Situation Counts - Last 24 Hours 16Situation Event Console 15

Ffixes, obtaining 149

Hhistorical data

collecting 20

IIBM Redbooks 149IBM Support Assistant 149IBM Tivoli Monitoring publications 154introduction, product 3

LLinux guest ApplData interval

setting 22

Mmanuals

see documentation library 153see publications 156

monitor processor usage at the system level 11monitor resources and workloads across LPARs 12monitoring agents 9monitoring scenarios

monitoring control unit cache activity 97, 98monitoring CP CPU utilization 95monitoring LPAR CPU utilization 96monitoring system utilization by Linux on System z

guests 96solving a DASD problem 93solving a paging data set problem 94

monitoring server 11


NNavigator

nodes 27nodes

Navigator 27

OOMEGAMON XE on z/VM and Linux

architecture overview 8how this monitoring agent works 7

OMEGAMON XE on z/VM and Linux productattributes 101introduction 3publications 153situations 63workspaces 27

online publicationsaccessing 156

ordering publications 156overview, product 3

Ppaging data sets

scenario 94performance considerations

data collection interval 21Performance Toolkit

guides and Web site 155predefined situations

descriptions and formulas 68ZVM_Avail_Mean_Low 68ZVM_Avail_Mean2G_Low 68ZVM_CP_CPU_Critical 69ZVM_CP_CPU_High 69ZVM_CUCache_DataResp_High 69ZVM_DASQ_Queue_Critical 71ZVM_DASQ_Queue_High 71ZVM_DevCache_Hits_Low 70ZVM_Eligible_List_High 87ZVM_LPAR_Busy_Critical 71ZVM_LPAR_Busy_High 72ZVM_LPAR_Ovhd_Critical 72ZVM_LPAR_Ovhd_High 73ZVM_Page_Queue_High 73ZVM_Page_Used_Critical 74ZVM_Page_Used_High 74ZVM_PerfKit_Collector_Inactive 75ZVM_Physical_CPU_Critical 76ZVM_Physical_CPU_High 77ZVM_Spin_Exclusive_Pct Critical 77ZVM_Spin_Exclusive_Pct_High 77ZVM_Spool_Used_Critical 78ZVM_Spool_Used_High 78ZVM_Storage_Overcommit_Critical 86ZVM_Storage_Overcommit_High 87ZVM_Total_CPU_Critical 79ZVM_Total_CPU_High 79ZVM_Total_to_Virtual_High 80

predefined situations (continued)ZVM_User_CPU_Critical 80ZVM_User_CPU_High 80ZVM_User_Scaled_CPU_Critical 81ZVM_User_Scaled_CPU_High 81ZVM_User_Wait_CPU_Critical 85ZVM_User_Wait_CPU_High 85ZVM_User_Wait_Page_Critical 84ZVM_User_Wait_Page_High 84ZVM_VDISK_Page_High_IORate_Low 82ZVM_Virtual_CPU_Critical 82ZVM_Virtual_CPU_High 83ZVM_Virtual_Scaled_CPU_Critical 83ZVM_Virtual_Scaled_CPU_High 83

problem determinationdescribing problems 151determining business impact 151submitting problems 152

problem resolution 149publications 153

accessing online 156IBM Tivoli Monitoring 154OMEGAMON XE on z/VM and Linux product 153ordering 156related 155see documentation library 153see publications 156

Qqueries 23

RRedbooks 149

Ssample interval

CP Monitor interval 22Linux guest ApplData interval 22

scenariospaging data set problem 94ZVM_Page_Used_Critical situation 94

Situation Editordefined 24opening 64tabs 64using 63

situation events 23, 63, 64situations

conditions 65defined 23formulas 65investigating an event 64negative values 65OMEGAMON XE on z/VM and Linux product 63predefined 65product-provided 65summary of predefined 66


Software Supportcontacting 150describing problems 151determining business impact 151overview 149receiving weekly updates 150submitting problems 152

support assistant 149System Terminal workspace

Take Action view 55

TTake Action

issuing commands 89prerequisites 89requirements 89types of commands that can be issued 89using 89

Take Action commanddefined 23

Tivoli Data Warehouse and the warehouse proxy 11Tivoli Enterprise Monitoring Agents 9Tivoli Enterprise Monitoring Server 9Tivoli Enterprise Portal

browser client 10defined 10desktop client 10views 14workspaces 13

Tivoli Management Services components 9Tivoli software information center 156

Uuser's guide

chapters xi

Wwhat you can do with OMEGAMON XE on z/VM and

Linux 11monitor paging and spooling activity 12monitor processor usage at the system level 11monitor resources and workloads across LPARs 12

workspacesaccessing from a chart view 29accessing from a report 29accessing from the Navigator 29accessing from the View menu 29accessing primary 28accessing the secondary 29ApplData 59attribute groups used by workspaces 29CCW Translations workspace 40Channel workspace 35CP Owned Devices 33creating custom queries 17DASD 37data filters 17defined 13

workspaces (continued)Enterprise Status 15FICON Channels workspace 36linking to another monitoring agent 18Linux Workload 58LPAR 44Minidisk Cache workspace 38predefined 27proactive monitoring 15Processor by LPAR Name workspace 46Processor workspace 47product-provided 27Real Storage 51Resource Constraint 61System 52System Terminal 54TCPIP 55TCPIP User 56Workload 57z/VM Linux Systems default workspace 32

ZZVM_Page_Used_Critical situation 94

Index 169

��

Printed in USA

SC27-2836-00

IBM Tivoli OMEGAMON XE on z/VM and Linux: User's … Srvr Data attributes 133 KVLTCPIPUsrData...

Documents

Transcript of IBM Tivoli OMEGAMON XE on z/VM and Linux: User's … Srvr Data attributes 133 KVLTCPIPUsrData...