You are on page 1of 264

Teradata Database

Resource Usage Macros and Tables
Release 13.10 B035-1099-109A October 2011

The product or products described in this book are licensed products of Teradata Corporation or its affiliates. Teradata, Active Enterprise Intelligence, Applications Within, Aprimo, Aprimo Marketing Studio, Aster, BYNET, Claraview, DecisionCast, Gridscale, Managing the Business of Marketing, MyCommerce, Raising Intelligence, Smarter. Faster. Wins., SQL-MapReduce, Teradata Decision Experts, Teradata Labs Logo, Teradata Raising Intelligence Logo, Teradata Source Experts, WebAnalyst, and Xkoto are trademarks or registered trademarks of Teradata Corporation or its affiliates in the United States and other countries. Adaptec and SCSISelect are trademarks or registered trademarks of Adaptec, Inc. AMD Opteron and Opteron are trademarks of Advanced Micro Devices, Inc. EMC, PowerPath, SRDF, and Symmetrix are registered trademarks of EMC Corporation. GoldenGate is a trademark of Oracle. Hewlett-Packard and HP are registered trademarks of Hewlett-Packard Company. Intel, Pentium, and XEON are registered trademarks of Intel Corporation. IBM, CICS, RACF, Tivoli, and z/OS are registered trademarks of International Business Machines Corporation. Linux is a registered trademark of Linus Torvalds. LSI is a registered trademark of LSI Corporation. Microsoft, Active Directory, Windows, Windows NT, and Windows Server are registered trademarks of Microsoft Corporation in the United States and other countries. NetVault is a trademark or registered trademark of Quest Software, Inc. in the United States and/or other countries. Novell and SUSE are registered trademarks of Novell, Inc., in the United States and other countries. Oracle, Java, and Solaris are registered trademarks of Oracle and/or its affiliates. QLogic and SANbox are trademarks or registered trademarks of QLogic Corporation. SAS and SAS/C are trademarks or registered trademarks of SAS Institute Inc. SPARC is a registered trademark of SPARC International, Inc. Symantec, NetBackup, and VERITAS are trademarks or registered trademarks of Symantec Corporation or its affiliates in the United States and other countries. Unicode is a registered trademark of Unicode, Inc. in the United States and other countries. UNIX is a registered trademark of The Open Group in the United States and other countries. Other product and company names mentioned herein may be the trademarks of their respective owners.

THE INFORMATION CONTAINED IN THIS DOCUMENT IS PROVIDED ON AN “AS-IS” BASIS, WITHOUT WARRANTY OF ANY KIND, EITHER EXPRESS OR IMPLIED, INCLUDING THE IMPLIED WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE, OR NON-INFRINGEMENT. SOME JURISDICTIONS DO NOT ALLOW THE EXCLUSION OF IMPLIED WARRANTIES, SO THE ABOVE EXCLUSION MAY NOT APPLY TO YOU. IN NO EVENT WILL TERADATA CORPORATION BE LIABLE FOR ANY INDIRECT, DIRECT, SPECIAL, INCIDENTAL, OR CONSEQUENTIAL DAMAGES, INCLUDING LOST PROFITS OR LOST SAVINGS, EVEN IF EXPRESSLY ADVISED OF THE POSSIBILITY OF SUCH DAMAGES.
The information contained in this document may contain references or cross-references to features, functions, products, or services that are not announced or available in your country. Such references do not imply that Teradata Corporation intends to announce such features, functions, products, or services in your country. Please consult your local Teradata Corporation representative for those features, functions, products, or services available in your country. Information contained in this document may contain technical inaccuracies or typographical errors. Information may be changed or updated without notice. Teradata Corporation may also make improvements or changes in the products or services described in this information at any time without notice. To maintain the quality of our products and services, we would like your comments on the accuracy, clarity, organization, and value of this document. Please email: teradata-books@lists.teradata.com Any comments or materials (collectively referred to as “Feedback”) sent to Teradata Corporation will be deemed non-confidential. Teradata Corporation will have no obligation of any kind with respect to Feedback and will be free to use, reproduce, disclose, exhibit, display, transform, create derivative works of, and distribute the Feedback and derivative works thereof without limitation on a royalty-free basis. Further, Teradata Corporation will be free to use any ideas, concepts, know-how, or techniques contained in such Feedback for any purpose whatsoever, including developing, manufacturing, or marketing products or services incorporating Feedback. Copyright © 2000 – 2011 by Teradata Corporation. All Rights Reserved.

Preface
Purpose
This book describes, and provides procedures for, Teradata Database resource usage data and macros.

Audience
This book is intended for system programmers, system administrators, and other database specialists responsible for administering or managing Teradata Database.

Supported Software Releases and Operating Systems
This book supports Teradata® Database 13.10. Teradata Database 13.10 supports: • • Microsoft Windows Server 2003 64-bit SUSE Linux Enterprise Server 10

Teradata Database client applications can support other operating systems.

Prerequisites
You should be familiar with basic computer technology, Teradata Database, and the system console environment. It will be helpful to review or reference the following books: • • • Introduction to Teradata Workload Management API: PM/API and Open API Performance Management

Resource Usage Macros and Tables

3

Preface Changes to This Book

Changes to This Book
Release Teradata Database 13.10 October 2011 Teradata Database 13.10 August 2010 Description Removed the note and some text from the “Channel Traffic Columns” section in the “ResUsageSpma Table” chapter. • Replaced instances of: • "VSS" with "TVS." • "collect period" and "collection period" with "gather period." • "32-bit" with "64-bit" to some of the extent driver I/O columns in "ResUsageSvpr Table." • "Windows and Linux" with "ALL." • "Teradata Manager" and "Teradata Dynamic Workload Manager" or "Teradata DWM" with appropriate references to Teradata Viewpoint. • Removed the following: • References to "MP-RAS," "xctl," "collection rate," and "Performance Monitor (PMON)." • The "Using the Active Row Filter Mode" topic from Chapter 2: “Planning Your Resource Usage Data.” • Appendix E: System Activity Reporter. • Combined Node Logging Rate and Vproc Logging Rate into a single Logging Rate in Chapter 3: “Resource Usage Procedures.” • Updated the following information: • The “About the "Invalid Platform" Column” section. • The SET RESOURCE syntax in Chapter 3: “Resource Usage Procedures.” • The description of the CollectIntervals column. • The descriptions of MailBoxDepth, MSGWORKTHREE and MSGWORKELEVEN in Chapter 7: “ResUsageSawt Table.”. • ResSpsView, ResSvprView, and ResGeneralInfoView views in Chapter 14: “Resource Usage Views.”. • The sample outputs for ResMemByGroup, ResMemMgmtOneNode, and ResMemMgmtByNode in Chapter 15: “Resource Usage Macros.”. • SpareCount columns 00-05 and 07-08 in Chapter 11: “ResUsageSps Table.”. • The NumSets description in Chapter 11: “ResUsageSps Table.”. • SpareCount columns 00-07 and SpareTmon columns 01-03 in Chapter 13: “ResUsageSvpr Table.”. • Changed the number of Performance Group (PGs) to 250. • Added views: ResSawtView and ResSpsView in Chapter 14: “Resource Usage Views.”.

4

Resource Usage Macros and Tables

Preface Changes to This Book

Release Teradata Database 13.10 August 2010 (continued)

Description • Marked FilePRowNDel, FileSRowNDel, FilePRowNUpd, FileSRowNUpd, FileAPtRowNUpd, FileAPtRowNDel, HostWriteFails, and HostReadFails valid on all platforms. • Marked ProcWorkType[i]Sum and ProcWorkType[i]Max invalid on all platforms. • The Vproc5 and VprocType5 fields are now valid on all platforms. • ResUsageSpdsk table is available for use and reports detailed usage of pdisks. • Updated descriptions of summary mode to clarify what is being reported. • Updated descriptions for file system fields in the ResUsageSpma table. • Fields that were once marked as invalid on all platforms that are now available for use on some or all platforms have been updated. • QWaitTimeMax, QLengthMax, and ServiceTimeMax have been changed to gather type "track" (not "count") which means they no longer need to be divided by the log intervals to obtain the average. • WorkTypeMax field in both the ResUsageSps and ResUsageAwt table now have the gather type "track." • Removed references to xschmon utility. It is no longer supported. • Added new fields to ResUsageSps and ResUsageSpdsk. • Added new flow control fields and Worktype fields to ResUsageSawt. • Added new fields to ResUsageSvpr to support Teradata Dynamic Workload Management software and DBQL. • Added new PDE, FSYS, and Teradata Virtual Storage fields to ResUsageSps. • Added new MI fields to ResUsageSawt and ResUsageSvpr. • Updated or added logical device fields such as input and output traffic columns as well as response time columns to ResUsageSldv. • Changed NodeType to CHAR(8) to accommodate new node types with longer names. • The field PGId has a new data type of SMALLINT. Also, summary mode of the ResUsageSps table uses the triplet of the PGId, VprType, and PPId fields. • Removed instances of gather type "countshft" because the only difference between countshft and count is how data collection is implemented. • Clarified CPU normalization in SPMA table • Marked all the Extent I/O Driver columns in the SVPR table as invalid.

Teradata Database 13.0 April 2009

Resource Usage Macros and Tables

5

enabling them to maximize their Teradata investment. • Order printed manuals: Under Print & CD Publications. 2 Enter your search criteria and click Search. view and join forums. • Download a documentation CD-ROM: 1 Under Online Publications.info. enter CD-ROM." The value must always be divided by the CollectIntervals value. www. Links include: • Executive reports. • ResUsageSps. and thought leadership • Technical information. and expert advice • Press releases.com/t/TEN/ Teradata Customer Education designs. and click Search. and download software patches. • Updated the definition of the ResGeneralInfo view and added the new ResSvprView view definition to Chapter 14.Preface Additional Information Release Teradata Database 13.teradataatyourservice.com/ 6 Resource Usage Macros and Tables .com/ Description Use the Teradata Information Products Publishing Library site to: • View or download a manual: 1 Under Online Publications.teradata. mentions and media resources www. Teradata Developer Exchange provides articles on using Teradata products. www. 2 In the Title or Keyword field. • ResUsageSps.WorkTypeMax values reports a maximum of sampled values and not the actual maximum of all inuse AWTs. technical alerts.com developer.WorkTypeInUse values now have a gather type "count. select General Search. develops and delivers education that builds skills and capabilities for our customers.com The Teradata home page provides links to numerous sources of information about Teradata. Use Teradata @ Your Service to access Orange Books. Additional Information URL www.teradata.teradata. case studies of customer experiences with Teradata. select How to Order. technical discussion forums. select General Search.0 April 2009 (continued) Description • The SpareTmon00 field in every table tracks the Capacity on Demand (COD) value. and knowledge repositories.teradata. and code downloads. solutions.

and value of this document. we would like your comments on the accuracy. Resource Usage Macros and Tables 7 .com. organization.teradata. clarity.Preface Additional Information To maintain the quality of our products and services. Please e-mail: teradatabooks@lists.

Preface Additional Information 8 Resource Usage Macros and Tables .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23 Resource Usage Macros and Tables 9 . . . . . . 19 Enabling Resource Usage Tables . . . . . . . . . . . . . . . . . 23 The Cost of Logging. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22 Using Active Row Filter Mode. . . . . . . . . . . . . . . . . . . . . . . . .3 Changes to This Book . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15 Setting Up and Maintaining Resource Usage Data . 15 Benefits of Using Resource Usage Data. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15 Accessing Resource Usage Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .3 Supported Software Releases and Operating Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21 Determining the Logging Value . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16 Data Gathering . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19 Resource Usage Tables. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .4 Additional Information . . . . . . . . . . . . . . . . . . . . . . .3 Audience . . . . . . . . 18 Chapter 2: Planning Your Resource Usage Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .3 Purpose . . . . . . . . . . 18 Application Programming Interfaces and Resource Usage Data . . . . . . . 20 Setting the Logging Rate. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Table of Contents Preface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17 Data Reporting . . . . . 23 Optimizing Resource Usage Logging . . . . . . . . . . . . . . . 21 Logging Rate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .6 Chapter 1: Introduction. . . . . . . . . . . . . . . 19 Tables Based on Needed Reports . . . . . . . . . . . . . . . . . . . . . . . . . . . 16 Gathering Resource Usage Data . . . . . . .3 Prerequisites . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17 Using Resource Usage Macros . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15 Overview of Resource Usage Data. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22 Using Summary Mode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . .31 Using One-Node Macros . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .42 Summary Mode in Resource Usage Tables. . . . . . . . . . . . . .29 Parameter Use for One-Node. . . . . . . . . . . . . . . . . . . . . .36 Purging Data . . . . .31 Saving and Analyzing Data . . . . . . . . . .Table of Contents Logging Cost Contributors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . and Group Macros . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .38 Inserting Rows into Resource Usage Tables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .36 Chapter 4: Resource Usage Tables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .24 Chapter 3: Resource Usage Procedures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .37 Physical Table Naming Conventions . .32 Using ENABLE and DISABLE LOGON Commands. . . . . . . . . . . . . . . . . . . . . . . . . .32 Function . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .27 Enabling RSS Logging. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .38 About the Invalid Platform Column. . . . . . . . . . . . . . . . . . . . . . . . . . . . .38 Occasional Event Data . . . . . . . . .31 Using ByGroup Macros . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .32 Executing Macros . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .42 Chapter 5: ResUsageScpu Table . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .48 10 Resource Usage Macros and Tables . . . . . . . . . . . . . . . . . . . . . . .48 Spare Columns . . . . . .41 About the Type of Data Column . . . . . . . . . All-Node. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .42 Column Names Ending In Sum . . . . . . . . . . . . . Multiple-Node. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .24 Operational Methods. . . . . . . . . . . . .27 Using ctl . . . . . . . . . . . . .28 General Macro Input Format. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .37 Relational Primary Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .45 Summary Mode. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .32 EXECUTE MACRO Syntax . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .38 Types of Resource Usage Table Columns . . . . . . . . . . . . .27 Using Database Window . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . 97 Chapter 11: ResUsageSps Table . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85 Summary Mode. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Table of Contents Chapter 6: ResUsageSpma Table . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 88 Spare Columns . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83 Spare Columns . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 116 Chapter 12: ResUsageSvdsk Table . . . . . . . . . . . 73 Summary Mode. 89 Chapter 10: ResUsageSpdsk Table . . . . . . 97 Spare Columns . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77 Spare Columns . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78 Chapter 8: ResUsageShst Table . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84 Chapter 9: ResUsageSldv Table . . 79 Summary Mode. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 124 Spare Columns . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 124 Resource Usage Macros and Tables 11 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91 Summary Mode. . . . . . . . . . . . . . . . . . . . . . . . . . 71 Chapter 7: ResUsageSawt Table . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51 Spare Columns . . . 99 Spare Columns . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 119 Summary Mode. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .184 ResAmpCpuByGroup Sample Output. . . . . . . . . . . . . . . . . . . . . . . . .188 ResPeCpuByGroup Sample Output . . . . . . . . . . . . . . . . . . . . . . . . .191 ResHostByLink Macros . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .192 12 Resource Usage Macros and Tables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .186 ResCPUByPE Sample Output.154 ResSawtView . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .175 Macro Output Format . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .143 Chapter 14: Resource Usage Views . . . . . . . . . . . . . .155 ResShstGroupView . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .187 ResCPUByPEOneNode Sample Output . . . . . . . . . . . . . . .190 ResCPUOneNode Sample Output . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .152 ResCPUUsageByPEView . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .183 ResCPUByAMPOneNode Sample Output . . . . . . . . . . . . . . . . . . . . . . . . . . . . .147 ResGeneralInfoView . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .180 ResAWTByNode Sample Output . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .158 ResSpsView . . . . .159 ResSvprView . . . . . . . . . . .125 Summary Mode. . . . . . . . .166 Chapter 15: Resource Usage Macros . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .157 ResSldvGroupView . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .190 ResCPUByGroup Sample Output .188 Normalized Viewing of CPU Usage by PE . . . . . . . . . . . . . . . .175 ResAWT Macros . . . . . . . . . . . . . .184 ResCPUByPE Macros . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .180 ResAWTByAMP Sample Output . . . . . . . . . . . . . . . . . . . . . .189 ResCPUByNode Sample Output . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .182 ResCPUByAMP Sample Output. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .148 ResCPUUsageByAMPView . . . . . . . . . . . . . . . . .188 ResCPUByNode Macros. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Table of Contents Chapter 13: ResUsageSvpr Table . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .184 Normalized Viewing of CPU Usage by AMP . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .142 Spare Columns . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .177 ResAWT Sample Output . . . . .181 ResCPUByAMP Macros . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . ResPdskByGroup Sample Output . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 195 ResLdvByNode Macros . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ResPdskByNode Macros: Pdisk Device Traffic . . . . . . . . . . . . . . . . . . . . 196 197 197 198 199 200 201 201 202 204 204 204 205 206 206 207 208 211 211 212 212 213 215 216 216 218 219 219 219 Appendix A: How to Read Syntax Diagrams . . . . . . . . . . . . . . . . . . . . . . . . ResMemByGroup Sample Output . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ResLdvOneNode Sample Output . . . . . . . . . . . . . . . . . . ResLdvByNode Sample Output . . . . . . . . . . . . . . ResPsByNode Macro Sample Output . ResNodeByNode Sample Output . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ResNetByGroup Sample Output . . . . . . . . . . . . ResVdskByGroup Sample Output . . . . . . . . . . . . . . . . . . . . . . . . ResMemMgmtByNode Sample Output. . . . . . . . . . . . . . . ResNode Macros . . . . . . . . . . ResNetByNode Sample Output . . 195 ResHostByGroup Sample Output . . . . . . . . . . . . . . ResVdskByNode Sample Output . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ResOneNode Sample Output . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ResPsByGroup Macro Sample Output . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ResMemMgmtOneNode Sample Output . . . . . . . . . . . . . . . . . . . . . . . . . . ResLdvByGroup Sample Output . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 221 Syntax Diagram Conventions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ResMemMgmtByNode Macros . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ResNodeByGroup Sample Output . . . . . . . . . ResPsByNodeWDJoin Macro Sample Output . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 194 ResHostOneNode Sample Output . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ResPdskOneNode Sample Output . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ResNetByNode Macros. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 221 Resource Usage Macros and Tables 13 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Table of Contents ResHostByLink Sample Output . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ResVdskByNode Macros . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ResNode Sample Output . . . . . . . . . . . . . . . . . . . . ResPs Macros. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ResNetOneNode Sample Output . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ResVdskOneNode Sample Output . . . . . . . ResPdskByNode Sample Output . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . .257 14 Resource Usage Macros and Tables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .249 Appendix D: Partition Assignments . . . . . . . . . . . . . . . . . . . . . . . . . . . .236 Appendix C: ResUsageIvpr Table . . . . . .248 Spare Columns . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .237 Summary Mode. . . . . . . . . . . . . . . .255 Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Table of Contents Appendix B: ResUsageIpma Table . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .252 Partition Assignment Listing . . . . . . . . . . . . . . . . . . . . . . . . . . . .252 Glossary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .251 Table Conventions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .227 Spare Columns . . . . . . . . . . . . . . .

This manual documents the resource usage data and settings for a variety of installation configurations and environments in Chapter 2: “Planning Your Resource Usage Data. and congestion Accessing Resource Usage Data Resource usage data is stored in system tables and views in the DBC database. Benefits of Using Resource Usage Data Resource usage data is useful for the following purposes: • • • • • • • Measuring system performance Measuring component performance Assisting with on-site job scheduling Identifying potential performance impacts Planning installation. down components. Macros installed with Teradata Database generate reports that display the data. Resource Usage Macros and Tables 15 .” To implement the settings you decide on. or ResUsage. Setting Up and Maintaining Resource Usage Data You need to decide what resource usage data you want to collect and the level of detail you want it to cover. and migration Analyzing performance degradation and improvement Identifying problems such as bottlenecks. is the collection and reporting of statistical information about the operation of your operating system and Teradata Database. See “Purging Data” on page 36. You can also write your own queries or macros on resource usage data. As with other database data.CHAPTER 1 Introduction Resource usage.” The only maintenance required is to purge old data regularly. upgrade. see Chapter 3: “Resource Usage Procedures. parallel inefficiencies. you can access resource usage data using SQL if you have the proper privileges.

AMP Worker Task (AWT) information memory management activity summary information Gathering Resource Usage Data Resource usage data gathering is a two-phase process as follows: • • Data gathering Data reporting 16 Resource Usage Macros and Tables . broadcast messaging. cylinder allocation. AWT statistics. data for each communication link. memory allocation. and time of session execution. Resource usage data covers … BYNET traffic on a node client-to-server traffic CPU utilization data tracking storage device traffic pdisk device traffic vdisk device traffic Priority Scheduler information Which includes … point-to-point messaging. see the following books: • • Workload Management API: PM/API and Open API Performance Management Overview of Resource Usage Data The following table lists topics covered by resource usage data. pdisk I/O. and migration statistics. user service.Chapter 1: Introduction Overview of Resource Usage Data Related Topics For additional information on performance analysis and system tuning. data by Performance Group (PG) from the Priority Scheduler and the ability to report resource usage data by Teradata Active System Management (ASM) workload definitions (WDs). overhead. and merge activities. positions of sessions in locking queues. the number of reads/writes and amount of data transferred as seen from the storage driver. all the cylinders allocated by an AMP (which can come from any pdisks in the clique). all data collected for a node or vproc.

Collect and Logging rates are independent. The resource usage data is written to the database using the data from the Log buffer. Data Reporting The reporting periods occur at the end of one or more gather intervals. and Teradata Database.Chapter 1: Introduction Gathering Resource Usage Data Data Gathering During the data gathering phase the RSS gathers information from the operating system. Collect and Logging rates. Each of the Teradata Dynamic Workload Management. The number of gather periods that occurred in any specific reporting period is indicated by the CollectInterval data field. Data Collection Macros and Routines Gather Buffer Collect Buffer Log Buffer Summary Log Teradata Dynamic Workload Management Buffer ResUsage Write Queue ResUsage Tables ResUsage Reports 1099A001 Resource Usage Macros and Tables 17 . Parallel Database Extensions (PDE). For the reporting period the respective reporting buffer is updated at the end of the respective reporting period and made accessible via the rssretrieve interface. Data gathering periods may not be uniformly spaced and are based on the Teradata Dynamic Workload Management.

and SQL Quick Reference for details about how to use the EXECUTE statement. see Workload Management API: PM/API and Open API. For more information on these APIs. Like other macros. resource usage macros consist of one or more Teradata SQL statements stored in Teradata Database and executed by a single EXECUTE statement.Chapter 1: Introduction Using Resource Usage Macros Using Resource Usage Macros Resource usage macros produce reports from data collected in the resource usage tables. In addition to the name of the macro. Application Programming Interfaces and Resource Usage Data The resource usage data are not used by just the resource usage macros. You can use the reports to analyze key operational statistics and evaluate the performance of your system. Resource usage data can also be used by the System Performance Monitor and Production Control Application Programming Interfaces (PMPC APIs). the EXECUTE statement for resource usage macros can include parameters to specify the following: • • • • A specific single-node A group of nodes Starting and ending dates and times Starting and ending nodes of a range of nodes Refer to Chapter 3: “Resource Usage Procedures” for more information on the resource usage macros. 18 Resource Usage Macros and Tables .

The following table lists more information. then you should enable them for logging. If you want to record detailed statistics covered by any of the resource usage tables.” then you need to enable the associated table.The default results in the ResUsageSpma (SPMA) table being logged every 10 minutes (600 seconds). the more overhead the system will use.CHAPTER 2 Planning Your Resource Usage Data This chapter describes how to: • • • • Enable resource usage tables Set the logging rate Use Summary Mode and Active Row Filter Mode Optimize resource usage logging Enabling Resource Usage Tables The default resource usage settings provide a good starting point for system monitoring. Tables Based on Needed Reports If you plan on using the report macros provided in Chapter 15: “Resource Usage Macros. along with specifying the largest logging period that will meet your needs. Naturally. You should not log data that you do not have a planned need for since this does incur additional database system overhead and uses up additional database space. The ResUsageSpma table provides a high level summary of how the system is operating and contains summarized or key elements from most of the other tables. the more tables you enable for logging and the shorter the logging period used. Resource Usage Macros and Tables 19 .

Chapter 2: Planning Your Resource Usage Data Enabling Resource Usage Tables For. you should enable when the system is first brought online to verify the following: • That all CPUs are functioning on all nodes • There is a good load balance among the CPUs ResUsageSpma System-wide node information provides a summary of overall system utilization incorporating the essential information from most of the other tables. saturation of a particular CPU on each node or on a particular node while others are idle could indicate a task always uses that CPU. ResUsageIpma ResUsageSawt System-wide node information. ResUsageShst Statistics on the host channels and LANs that communicate with Teradata Database. When you want to monitor the utilization of the AWT and determine if work is backing up because the AWTs are all being used. When You Should Enable When the performance analysis suggests that the overall performance is limited or to check if a program is spinning in an infinite loop on an individual processor. 20 Resource Usage Macros and Tables . For example. Also. this table is not used at customer sites. resulting in 100% transmitting and 100% receiving values simultaneously.” Resource Usage Tables The following table describes the tables and provides guidance about which ones to enable. “Enabling RSS Logging” on page 27. Chapter 15: “Resource Usage Macros. To determine details about the traffic over the IBM Host channels to determine if there is a bottleneck. Note: The BYNET can transmit and receive at the same time. Table Name ResUsageScpu Covers Statistics on the CPUs within the nodes.. Another method of determining BYNET utilization and traffic is to use the blmstat tool. intended primarily for Teradata engineers. To provide an overall history of the system operation. Generally. Data specific to the AWTs... instructions on setting resource usage tables instructions on using macros descriptions and examples of the macros See. “General Macro Input Format” on page 29 and “Executing Macros” on page 32. Use the columns in ResUsageSpma to view BYNET utilization..

Resource usage logging means the writing of resource data as rows to one or more of the resource usage database tables. the more frequently data is logged. RSS will automatically double the logging period which effectively summarizes the data by providing values for a time period twice that provided by the previous logging period. Data specific to each virtual processor and its file system. Spma. this table is not used at customer sites. When no longer busy. logical device statistics collected from the storage driver. the system resumes logging as before. see Chapter 3: “Resource Usage Procedures” for details on how to set the logging rate. Setting the Logging Rate The default for the Node Logging Rate is 600. Data by PG ID from the Priority Scheduler. this means that the database is busy.ResUsagexxxx. and the more disk space is used. and so forth) as listed in “Resource Usage Tables” on page 20. where xxxx is the name of the resource usage table (for example. ResUsageIvpr System-wide virtual processor information. To view details about the resources being used by each vproc on the system. The shorter the logging period. Resource Usage Macros and Tables 21 . Logging Rate Logging rate controls the frequency (number of seconds) at which resource usage data is logged to the resource usage tables. intended primarily for Teradata engineers. When You Should Enable To observe the balance of disk usage. Note: Events in the event logs related to this doubling of the logging period do not represent fatal errors but are informational to indicate that the automatic operations of the RSS are attempting to maintain data logging. Generally. To obtain detailed usage information about pdisks. ResUsageSpdsk ResUsageSps ResUsageSvpr Statistics collected from the pdisk device. When you need to track utilization by the query WD level. Ipma. When the system is so busy that the resource usage table logging gets backed up. This table is useful for looking for hot AMPS or PEs that may be CPU bound or throttled on other resources. When you have decided what rate to set. If you see the resource usage logging rates change without user intervention. Note: Use the ResUsageSvdsk table first to observe general system disk utilization unless specifically debugging at a low level. The storage device statistics are often difficult to interpret with disk arrays attached due to multi-path access to disks.Chapter 2: Planning Your Resource Usage Data Setting the Logging Rate Table Name ResUsageSldv Covers System-wide. The tables are named DBC.

22 Resource Usage Macros and Tables . For example. Rates and enabled tables may be changed at any time and the changes take effect immediately. then use summary mode to produce one row per vproc type instead of one row per vproc. You can select summary mode for each table individually. The ResUsageSpma table. 1 8 18 40 75 150 360 1200 2 9 20 45 80 180 400 1800 3 10 24 48 90 200 450 3600 4 12 25 50 100 225 600 5 15 30 60 120 240 720 6 16 36 72 144 300 900 A practical log interval minimum during production processing is 60 seconds. • • The white area of the table shows rates recommended only for short-term use for debugging a specific issue. See the description for each table for details on how summary mode affects that particular table. such as 120 seconds or 300 seconds can also be used. If the system becomes very busy. This helps minimize the cost of the data logging. The following table shows the valid logging rate. The default rate is 600 seconds. This effectively summarizes the data by providing values for a time period twice that of the previous logging period. if you want to log information provided in the ResUsageSvpr table but do not need data for each individual vproc. it will automatically double the logging period. Intermediate log intervals.Chapter 2: Planning Your Resource Usage Data Using Summary Mode Determining the Logging Value The system imposes the following rule on the logging rate: Intervals must evenly divide into 3600 (the number of seconds in an hour). The system automatically returns to logging back to the rate you set when it is no longer busy. Using Summary Mode You can use Summary Mode to reduce the system overhead from logging tables that produce multiple rows per logging period. in comparison. provides node level summary of key fields from most of the other ResUsage tables. The highlighted area of the table shows rates recommended for production processing. When more details are required than the ResUsageSpma table provides then the next level of information is provided by using summary mode logging for the table of interest. Summary Mode helps reduce overhead by combining data from multiple rows into one or more summary rows based on specific criteria for each table.

Instead the maximum value from the rows is used. Using Active Row Filter Mode Active Row Filter Mode reduces the overhead of logging for some of the resource usage table by limiting the data rows that are logged to the database. see “Summary Mode in Resource Usage Tables” on page 42. all the individual vproc rows of the same vproc type are combined into a single row. you need to divide the summary row data value by the number of rows that made up the summary mode row to get the average per vproc. it may appear that rows are missing when looking at the query results. For example. you can use the ResUsageSpma table or use the Vproc Manager utility. see “Enabling RSS Logging” on page 27. for the ResUsageSvpr table in summary mode. Resource Usage Macros and Tables 23 . If this space is never reclaimed. To determine if rows are not being logged to the database. For more information on Summary Mode. The rows written to the database take up space. PEs. and all other vproc types on your system. Optimizing Resource Usage Logging The Cost of Logging Logging resource usage data to database tables incurs costs: • • Writing to the database adds to the system I/O load. the ResUsageSvpr table MsgWorkQLenMax field in the summary mode row for the AMPs will contain the maximum value from all the AMP rows that would have been logged in non-summary mode. Note: Active Row Filtering should not be disabled for the ResUsageSps table. When active row filter is enabled. A similar computation needs to be done to derive the average value per PE from the summary row data value. This is because the index values of the inactive rows varies over time so that a row with one index may be logged one period but not in another. On a heavily loaded system. Since the data values are added together. divide the AMP summary row data value by the number of AMPs on that node to determine the average value per AMP. For example. Summary mode has either no effect on the values of the Housekeeping Columns or it is specifically detailed in the description of each affected field. this could affect the production workload throughput. (To determine the number of AMP.) Fields that represent a maximum statistic are not summed together. you should look in the event logs for messages indicating that rows have been lost. To enable Summary Mode.Chapter 2: Planning Your Resource Usage Data Using Active Row Filter Mode For example. it will eventually grow to consume all available space in user DBC. The fields that represent a minimum statistic are summarized by storing the minimum value from all the constituent rows.

Generally. Although it caches such data and will eventually catch up if given a chance. host driver. Restart resource usage logging by entering the appropriate SET RESOURCE command. when you issue the WRITE command. Avoid unnecessarily using or exhausting available disk space by doing the following: • Never enable logging on tables that you do not intend to use. it is possible that the RSS can fall behind in writing data to the database. 24 Resource Usage Macros and Tables . They depend on a number of interrelated factors: • • • • How busy is the system Which resource usage tables are enabled What resource usage logging rates are in effect The system configuration (vproc. ResUsageSps). The default rate is 600. Note: If resource usage logging terminates due to a lack of table space: a b Delete rows from the appropriate table or make more space for it in USER DBC. (For example.Chapter 2: Planning Your Resource Usage Data Optimizing Resource Usage Logging • In an extremely loaded system. with ctl. For example. the RSS will be forced to start discarding rows if the system load persists and its cache capacity has been exceeded. you should use a logging rate no smaller than 60. regardless of whether the database system is busy. logging only to the ResUsageSpma table provides a lot of useful information with a minimal operational load on the system. logical devices or device controllers) Operational Methods Use the following methods to optimize performance and reduce the cost of resource usage logging on your system: 1 Use Summary Mode to reduce the number of rows inserted into the resource usage tables if Summary Mode data provides sufficient information for your needs. Logging Cost Contributors Logging costs are difficult to quantify. New values take effect as soon as the adjustment command is issued.) 3 4 Purge old data from the ResUsage tables periodically. 2 For tables with a large number of rows (for example. • Use the largest rates that provide enough detail information for your purposes. These values can be adjusted any time. use Active Row Filter Mode to limit the number of rows written to the database each logging period and to minimize the amount of system resources used. CPU.

“Enabling RSS Logging” on page 27. Resource Usage Macros and Tables 25 ... and summarizing or filtering rows purging old data from resource usage tables See.Chapter 2: Planning Your Resource Usage Data Optimizing Resource Usage Logging Related Topics For instructions on. enabling resource usage tables... setting the logging rates. “Purging Data” on page 36.

Chapter 2: Planning Your Resource Usage Data Optimizing Resource Usage Logging 26 Resource Usage Macros and Tables .

Resource Usage Macros and Tables 27 . Windows or Linux Windows or Linux You can enable logging by running.. “Setting the Logging Rate” on page 21. “Using ctl” on page 27. see. set the logging rate. determine which tables and controlling rates apply to the resource usage macros you want to run. see the following topics: • • “Enabling Resource Usage Tables” on page 19. If you are running. “Using Database Window” on page 28.. Before you set the ResUsage tables. Using ctl The Control GDO Editor utility (ctl) is used to set various Teradata Database configuration settings. The RSS-related settings are presented on the RSS screen. For instructions.. For more information. see ctl in Utilities. Database Window (DBW)...CHAPTER 3 Resource Usage Procedures This chapter describes how to: • • • • Enable RSS logging Execute different types of macros Enable logons Purge old data Enabling RSS Logging By using one of the following interfaces you can enable tables. For detailed information on starting ctl and modifying the settings. the ctl utility from the Teradata Command Prompt.. and optionally summarize or filter rows.

tablename ALL ON OFF FE0CA030 SET LOGTABLE After the table is enabled for logging. You would enter the following: set logtable shst on set resource node log 600 28 Resource Usage Macros and Tables . tablename ON OFF 1095A010 SET SUMLOGTABLE Example The following example shows you how to enable table logging and set the Logging rate using the database commands in DBW. you can log rows in Summary Mode. 4 (Optional) Enable Summary Mode on the table specified using the command below. see "Database Window (xdbw)" in Utilities. Note: A rate of zero disables the logging function. For instructions on starting DBW. see “Using Summary Mode” on page 22. you must enable the table specified in both the RSS Table Logging Enable group and in the RSS Summary Mode Enable group. 3 Specify the table you want to enable logging to using the database command below. Note: To log rows in Summary Mode.Chapter 3: Resource Usage Procedures Enabling RSS Logging Using Database Window Use the database commands below to enable resource usage tables and set the logging rate from DBW on Windows or Linux. Suppose you want to enable the ResUsageShst table and set the logging rate for 10 minutes (600 seconds). SET RESOURCE NODE LOGGING LOG number 1099C002 where number is the number of seconds. To enable RSS logging from DBW 1 2 Open the Supvr window. Set the Node Logging Rate using the database command below. For more information.

Chapter 3: Resource Usage Procedures General Macro Input Format Related Topics For more information on.. all nodes.. Resource Usage Macros and Tables 29 . "Database Window (xdbw)" in Utilities. General Macro Input Format As shown in the table below. or group nodes as indicated. Utilities. there are four kinds of macros: • • • • Multiple-node One-node All-node ByGroup For any given line in the following table... one node. the macros on that line report the same statistics for either multiple nodes. ctl DBW See.

30 Description AWTs in use by node CPU usage by AMP Vprocs CPU usage by PE Vprocs CPU usage by nodes Host statistics Ldv disk statistics Memory management General network statistics General node-level statistics Priority Scheduler and Teradata ASM Workload statistics pdisk level I/O statistics AMP level I/O statistics ResLdvByNode ResMemMgmtByNode ResNetByNode ResNodeByNode ResPsByNode Multinode Macro ResAWTByAMP ResAWTByNode ResCPUByAMP ResCPUByPE ResCPUByNode ResCPUByAMPOneNode ResCPUByPEOneNode ResCPUOneNode ResHostOneNode ResLdvOneNode ResMemMgmtOneNode ResNetOneNode ResOneNode ResNode ResHostByLink One-Node Macro All-Node Macro ResAWT ResAmpCpuByGroup ResPeCpuByGroup ResCpuByGroup ResHostByGroup ResLdvByGroup ResMemByGroup ResNetByGroup ResNodeByGroup ResPsByGroup ByGroup Macro ResPdskByNode ResVdskByNode ResPdskOneNode ResVdskOneNode ResPdskByGroup ResVdskByGroup Resource Usage Macros and Tables .

ByGroup macros are similar to the other macros. Using One-Node Macros One-node macro versions are primarily used on single-node systems. The only difference is that they use the GroupId column of the views to report system usage for a specific set of nodes grouped by a GroupId. the nodes may become bottlenecks in the throughput of the system as a whole. In Teradata Database. you can use the corresponding multiple-node macro to report on just one node by supplying equal FromNode and ToNode parameters. The only differences are that the single-node version of each macro has both of the following: • • OneNode qualifier in the macro name. this macro reports statistics for all nodes in the group. Resource Usage Macros and Tables 31 . and group macros. For instructions on using these macros. A single node specification. multiple-node. this macro reports system-wide statistics. Because of the differences. All-Node. The input format of the ByGroup macros is the same as the other macros except ByGroup appears as the qualifier in the macro name. ToNode Node None. co-existing nodes are nodes of different model types in the same configurations. The default is ‘001-01’. Examples of redundant columns are the NodeId column and columns that focus on cross-node load balancing. Using ByGroup Macros ByGroup macro versions are used on systems with co-existing nodes. instead of the FromNode and ToNode parameters to specify a range of nodes. Note: The Database Administrator must identify the groupings of nodes when the system is first configured.Chapter 3: Resource Usage Procedures General Macro Input Format Parameter Use for One-Node. see “Executing Macros” on page 32. OneNode macros have the same general input format as the other macros. all-node. Multiple-Node. Alternatively. however. One-node versions are recommended. Macro Type Multiple node One node All node Group Number of Parameters Six (except ResHostByLink) Five Four Four Node Parameters Used FromNode. Therefore. ByGroup macros were developed to provide the system user with a summary of the performance data based on node groupings. None. because they eliminate redundant report columns on a single-node system. and Group Macros The following table explains parameter use for one-node.

. FromTime D . FromDate ( FromDate ( . for example. FromDate A . ToDate . ask your System Administrator to retain two sets of view and macro Data Definition Language (DDL) files in separate places. ToTime ToTime . Node FromTime ToTime GX02B001 where: 32 Resource Usage Macros and Tables .” EXECUTE MACRO Syntax The execution of each resource usage macro has the following form. You could. .” EXECUTE EXEC MacroNameMultiNode MacroNameAllNode MacroNameOneNode MacroNameByGroup ( ( . where xx represents the Teradata Database release number. . ToDate . ToDate ToDate . use ResNodeRxx. see Chapter 15: “Resource Usage Macros.Chapter 3: Resource Usage Procedures Executing Macros Saving and Analyzing Data If you expect an ongoing need to retain and analyze data from different Teradata Database releases. Executing Macros Function Macro execution is illustrated in the following diagram. For details about each macro and its resulting report. FromNode ToNode FromTime C . ). Rename the views and macros so that you can use either. as the name of the resource usage macro and use it when you want to analyze the data from that release. A . For information on interpreting the syntax diagrams. see Appendix A: “How to Read Syntax Diagrams. ToTime . B C D FromDate . FromTime B .

MacroNameByGroup Name of a ByGroup resource usage macro: • • • • • • FromDate ResAmpCpuByGroup ResCPUByGroup ResHostByGroup ResLdvByGroup ResMemByGroup ResNetByGroup • • • • • ResNodeByGroup ResPeCpuByGroup ResPdskByGroup ResPsByGroup ResVdskByGroup Start date to report resource usage data. Note: The character string date format has been changed from yymmdd to 'yyyy-mm-dd' to accommodate dates in the 21st century. The date may be entered either as a character string (for example. character format for May 31. 2007 would appear as '2007-05-31') or as a numeric value (for the same date in numeric format. End time to report resource usage data. See the FromDate syntax element column for a further explanation of date formats. The default is 000000. The default is 999999. The format is hhmmss. 1070531). The character string is the recommended format. The format is hhmmss. Resource Usage Macros and Tables 33 . The character string is the recommended format. See "String Date Validations" in SQL Data Manipulation Language for more detailed information on using numeric dates with macros. The default is the current system date.Chapter 3: Resource Usage Procedures Executing Macros Syntax element MacroNameMultiNode Description Name of a multinode resource usage macro: • • • • • • ResAwtByNode ResCPUByAMP ResCPUByPE ResCPUByNode ResLdvByNode ResMemMgmtByNode • • • • • ResNetByNode ResNodeByNode ResPdskByNode ResPsByNode ResVdskByNode MacroNameAllNode Name of an all-node resource usage macro: • ResNode • ResHostByLink The ResHostByLink and ResNode macros do not use the FromNode and ToNode parameters. ToDate End date to report resource usage data. FromTime ToTime Start time to report resource usage data.

type get config in the DBW Supvr window. The default is '001-01'. The format is 'nnn-nn'. producing a report for the period beginning 8:00 a. 2006. 34 Resource Usage Macros and Tables .Chapter 3: Resource Usage Procedures Executing Macros Syntax element FromNode Description Starting range of nodes to report resource usage data. 080000. and hyphen must be included in the forth character position. 240000. where: Statement Element ResCPUByAMP '2006-12-25' '2006-12-31' 080000 240000 '123-02' '125-04' Description Name of the resource usage macro Starting date of December 25. '123-02'. A hyphen must be included in the fourth character position.'2006-12-31'. 1-01 should be typed out as '001-01'. on December 25. Example 1: Executing the ResCPUByAMP Macro The following statement executes the ResCPUByAMP macro.m. 2006 and ending 12:00 midnight. Note: To identify the node ID numbers for your system.m. 2006 Starting time of 8:00 a. on December 31. EXECUTE ResCPUByAmp('2006-12-25'. ToNode Ending range of nodes to report resource usage data. Note: To identify the node ID numbers for your system.'125-04'). The format is 'nnnnn'. Ending time of 12:00 midnight Starting node of a range of nodes Ending node of a range of nodes See SQL Data Types and Literals for information on using numeric values for dates. 2006 Ending date of December 31. For example. type get config in the DBW Supervisor Window (Supvr). The format is 'nnn-nn'. Node Single-node ID to report resource usage data. The default is '000-00'. The default is '999-99'. It includes data for nodes 123-02 through 125-04. A hyphen must be included in the fourth character position.

m.m.Chapter 3: Resource Usage Procedures Executing Macros Example 2: Executing the ResCPUByAMPOneNode Macro The following statement executes the OneNode version of the ResCPUByAMP macro shown in Example 1.'123-02'). It uses the same starting and ending dates and times (using character string format). Resource Usage Macros and Tables 35 . 240000. Ending time of 12:00 midnight Node See SQL Data Types and Literals for information on using numeric values for dates.'2006-12-31'. Example 3: Executing the ResAMPCpuByGroup Macro The following statement executes the ByGroup version of the ResCPUByAmp macro shown in Example 1. where: Statement Element ResCPUByAMPByGroup '2006-12-25' '2006-12-31' 080000 240000 Description Name of the resource usage macro Starting date of December 25. 2006 Ending date of December 31.'2006-12-31'. 2006 Starting time of 8:00 a. 2006 Ending date of December 31. node 123-02. except the report is for a single-node. 240000). except the report is for a node grouping.080000. where: Statement Element ResCPUByAMPOneNode '2006-12-25' '2006-12-31' 080000 240000 '123-02' Description Name of the resource usage macro Starting date of December 25.080000. It uses the same starting and ending dates and times (using character string format). EXECUTE ResCpuByAmpOneNode ('2006-12-25'. Ending time of 12:00 midnight See SQL Data Types and Literals for information on using numeric values for dates. EXECUTE ResAMPCpuByGroup ('2006-12-25'. 2006 Starting time of 8:00 a.

See "Control GDO Editor (ctl)" in Utilities. You can directly remove old resource usage data by submitting SQL statements. use the following SQL statement to remove data more than five days old from the ResUsageSpma table: DELETE FROM ResUsageSpma WHERE TheDate < CURRENT_DATE . Purging Data The RSS does not automatically delete data from the resource usage tables.) To enable logons from: • • Database Window.7. Teradata command prompt. (DISABLE ALL LOGONS prevents all users. see "Changing Logon States and Restarting the System" in Database Administration. resource usage data stops logging to the tables even if there are still active sessions logged on. from logging on and also stops logging to the tables. For more information on enabling and disabling logons.Chapter 3: Resource Usage Procedures Using ENABLE and DISABLE LOGON Commands Using ENABLE and DISABLE LOGON Commands The DISABLE LOGONS command prevents new sessions from logging on. When logons are disabled. 36 Resource Usage Macros and Tables . For more information about the DELETE syntax. including user DBC. You need to purge data you no longer need on a regular basis. run ENABLE LOGONS or ENABLE ALL LOGONS. see "SQL Data Manipulation Language Statement Syntax" in SQL Data Manipulation Language. For example. use the Start With Logons field of the Screen Debug menu of ctl.

. AWT statistics WD resolution statistics Channel and LAN host information Resource Usage Macros and Tables 37 .CHAPTER 4 Resource Usage Tables This chapter describes: • • • How to name the physical table and insert rows into resource usage tables Types of resource usage table columns and data Summary Mode in resource usage tables Physical Table Naming Conventions Each physical table name follows this general naming convention: ResUsage Information_type Table_name where: Element Information_type Code S I Table_name Code pma vpr cpu ldv pdsk vdsk awt sps hst Description Node information vproc information CPU-specific information Logical device statistics pdisk device statistics vdisk device statistics Description System-wide information Internal Teradata Database information Is one of the following..

For each resource usage table column. this manual describes the: • • Column Name Type of Data 38 Resource Usage Macros and Tables . The primary index is nonunique because of duplicate rows that will appear with the same timestamp during daylight savings time. All columns described in the following chapters and appendixes are type FLOAT unless otherwise specified in the description of that column. therefore. Inserting Rows into Resource Usage Tables For information on how rows will be inserted into these tables based on the current resource usage control settings. • For more information on MULTISET tables. Note: The actual table definitions are obtainable by executing the SHOW TABLE statement. See SQL Data Definition Language for more information about SHOW TABLE. refer to “Using Summary Mode” on page 22. Occasional Event Data Occasional event data is considered outside the scope of resource usage and is. Types of Resource Usage Table Columns This manual describes what each of the resource usage table columns report (that is.Chapter 4: Resource Usage Tables Relational Primary Index Relational Primary Index All resource usage tables have the same nonunique primary index: • • The nonunique primary index consists of TheDate. and NodeID columns.” For information on the number of rows inserted in a resource usage table for each applicable log period. All nonexistent values are stored as NULL. TheTime. see Chapter 2: “Planning Your Resource Usage Data. Because the primary index is nonunique. logged in the ERRORLOG and the DBCINFO tables rather than in the resource usage tables.ResUsageXxxx.ColumnName reports) in a table format. Rows that have duplicate timestamps can be distinguished by the GmtTime column. all resource usage tables are created as MULTISET tables. This prevents the system from checking for duplicate rows. what each DBC. see "CREATE TABLE (Table Kind Clause)" in SQL Data Definition Language or "Duplicate Rows in Tables" in SQL Fundamentals.

The following table shows the types of statistics subdivided into their respective subcategories. gather elements and its general characteristics. Statistics columns which can be further categorized into subcategories. the WAL log. The WAL columns identify the log-based file system recovery scheme in which modifications to permanent data are written to a log file. current logging characteristics. memory or net columns. Other columns identify the purpose of operations being performed on disk segments such as cylinder migration or data updates. Identification of concurrency control activities is provided and subdivided into control done for user level processing. Category File System Subcategories • Cylinder Management • Cylinder Management Overhead Events • Data Block Prefetches • Data Segment Lock Requests • Segments Acquired • Segments Released • Synchronized Full File Scans • Write Ahead Logging (WAL) Description Some of the file system columns can be viewed as a subset of memory columns by expanding on the operations performed on disk memory segments. or identify the requests being made by database software on the file system. Statistic columns are further grouped by category and subcategory as shown below. system overhead processing. General Concurrency Control Database Locks Resource Usage Macros and Tables 39 . Column Name Type of Data Description Data Type Invalid Platform HOUSEKEEPING OR STATISTICS COLUMNS CATEGORYS Subcategory Each table has: • • Housekeeping columns which contain statistics on timestamp. and database locks. It does not include control specific to disk.Chapter 4: Resource Usage Tables Types of Resource Usage Table Columns • • • Description Data Type Invalid Platform The columns are grouped into either housekeeping columns or statistics columns. which are included in the disk. memory or net concurrency control. Categories and subcategories may vary from table to table. Operations counted are logical memory and physical disk reads and writes (including aging) and locking control activities.

Finally. These columns also track the number of times processing was switched to another process for multitasking purposes or to perform interrupt services. channel utilization. subdivided into the type of transmission. Memory management columns are also provided to identify events leading up to paging. Memory Memory related events. This helps identify the ability or inability of the system to effectively utilize resources. Some also give overhead and management information on the host channel and LAN. Time allotments are tracked by monitoring the time spent waiting for resources or processing code. logical memory and physical disk reads and writes (including paging and swapping). as well as physical utilization of the BYNET. subdivided into memory types. and Teradata net contention are identified as well. are collected for memory allocation and deallocation. Expanded detail is provided for work started but waiting on resources. Logical messages and direction are identified through subdivisions of the message class. with current characteristics and states. access. Controller overhead. The storage device statistics are calculated only on what can be derived from statistics collected by the operating system. a detailed snapshot of the memory is provided by tracking the current states per memory types. 40 Resource Usage Macros and Tables .Chapter 4: Resource Usage Tables Types of Resource Usage Table Columns Category Host Controller (SHST) Subcategories • • • • • • • • • • • • Channel Traffic Channel Management Controller Overhead User Commands User Command Arrival and Departure Memory Allocations Memory Availability Management Memory Pages Resident Memory Resident Paging Swapping Task Context Segment Usage Description These columns identify traffic on the host-to-node channels and LANs. These columns identify individual logical device activities for external storage components connected through the buses. since the disk array controllers do not provide us with any useful data for resource usage. Logical Device • • • • • Concurrent Operations Input and Output Traffic Outstanding Requests Response Time Seek Statistics Net • • • • • • • • • Broadcast Net Traffic Group Coordination Merge Services Net Controller Status and Miscellaneous Management Net Circuit Management Network Transport Per-Bynet Network Transport Data Point-to-point Net Traffic Work Mailbox Queue ChnSignal Status Tracking CPU Utilization Cylinder Read Process Allocation Process Block Counts Process Pending Snapshot Process Pending Wait Time Scheduled CPU Switching Process Scheduling • • • • • • • • These columns provide a CPU-level snapshot of work started. deaccess and memory control. Traffic over the BYNET is identified through the number and direction of messages. swapping and aging activities.

If your platform appears in that column for a field. These columns identify the secondary cache miss rate. It is not valid on Microsoft Windows. column is only valid on Linux. None. Resource Usage Macros and Tables 41 . The columns specific to the ResUsageSawt table also report the number of AWTs currently in use and the maximum number of AWTs for the current vproc on the node. The column is either obsolete or not valid on any of the platforms.Chapter 4: Resource Usage Tables About the Invalid Platform Column Category User Commands Subcategories • User command • User command Arrival and Departure None. • • • • • • • • AMP Worker Task In use and Max Array Data Priority Scheduler Worktype Descriptions Allocation I/O Migration Node Agent Description These columns describe the types of commands given to Teradata Database by the user and the progress of those commands. column is only valid on Windows. When the Invalid Platform column is blank. The following table explains the contents of the Invalid Platform column. Secondary Cache Misses Spare Teradata ASM Teradata Virtual Storage (VS) About the Invalid Platform Column The tables in this book that describe the resource usage tables contain an Invalid Platform column. column is valid on all platforms. Note: Teradata VS is available for purchase separately from Teradata Database. These columns collect and report statistics about the AWTs and Priority Scheduler. the column being described is valid on all platforms. then resource usage data for that particular field is either not collected by the system or is not valid and should not be used. see Teradata Virtual Storage. It is not valid on Linux. These columns identify individual pdisk and vdisk device activities. These columns are for future release or internal manipulation by Teradata developers. In the Invalid Platform column … ALL Linux Windows Means … do not use on any platform. For information about these columns.

Max fields have a Max suffix in the field name. are useful for calculating the average value for a gather period.Chapter 4: Resource Usage Tables About the Type of Data Column About the Type of Data Column There are four possible types of data reported in the Type of Data column: • • Count . Min . An example of a Max field in the ResUsageSvdsk table is the ReadRespMax field. Divide the resulting logged value by the value CollectIntervals to get the average value. • • Column Names Ending In Sum Column values ending in Sum.Min fields have a Min suffix in the field name. Summary Mode in Resource Usage Tables Summary mode combines data from the multiple data rows normally generated into one or more rows.Track fields gauge the current value of a countable item such as a queue length during a period of time. The CollectIntervals value is the number of gather periods per reporting period. An example of a Min field in the ResUsageSvpr table is the ReadResponseHotMin field. The track field reports the value at the end of the logging period. Max . This field reports the maximum value for the logging period. 42 Resource Usage Macros and Tables . Each sum column accumulates the values measured by the column at the end of every gather period. Track . This field reports the minimum value during each logging period.Count fields tallies the number of times an event occurred. the data is combined using the rules in the following table. which are the Count Type of Data. such as the number of disk reads or writes during a period of time. When multiple rows are condensed into a single row.

which will provide the sampled average number of AWTs in use over the reporting period.FlowCtlCnt field provides the total number of times that the system entered flow control state from a non-flow control state. then the value reported will be the FlowControlled state at the end of the last gather period. In Summary mode. the values from each of the non-summary mode rows are added together to produce the summary mode row value. the value reported is the sum of the current number of AWTs in use from each CollectInterval. On the other hand. If the row is being logged normally. Note: To obtain the average number of AWTs in use during the reporting period per AMP. Depending upon the usage of the field. For example.FlowControlled is a track field so that in summary mode. the value needs to be divided by the number of AMPs if the user wishes to determine the average number per AMP rather than the total. for the WorkTypeInuse00 field of the SAWT or SPS table. Summary Mode is applicable to all tables except ResUsageSpma and ResUsageIpma.ReadResponseHotMin is the minimum value from all the rows that are combined into a single summary row. • In summary mode. the field should always be divided by the number of CollectIntervals. the SummaryFlag value is set to ‘N’. the values are reported per AMP and no division by CollectIntervals is necessary since the total over the entire reporting period is desired. • In normal mode. ResUsageSvpr. the reported value for a Max field such as ResUsageSpdsk. The following table describes Summary Mode for the resource usage tables. the reported value for a Min field such as ResUsageSvpr.FlowControlled field. In summary mode. In summary mode. all the AMP vproc rows are combined into a single row and the FlowControlled field will report the summed value from each of the AMP vproc rows. the ResUsageSvpr. Max taking the maximum value from all the contributing rows. the SummaryFlag value is set to ‘S’. Track summing the values from all the contributing rows. For the ResUsageSvpr. the Track values from each row to be combined are summed together. the value may or may not need to be adjusted by the number of rows that were combined by summary mode or the number of data sampling periods as indicated by the CollectIntervals column. Count fields are also added together when there are multiple gather periods in the reporting period. this means that if there were 10 gather periods (the CollectIntervals column equals 10) in the reporting period. If summary mode is enabled. Min taking the minimum value from all the contributing rows. In this case. If the information for a row of a table is in Summary Mode. For example. This is the same as the value at the end of the reporting period. divide the summary mode reported value by the CollectIntervals value as well as the number of AMPs.ConcurrentWriteMax is the maximum value from all the rows that are combined into a single summary row.Chapter 4: Resource Usage Tables Summary Mode in Resource Usage Tables For the following Type of Data… Count Summary fields are combined by… summing the values from all the contributing rows. Note: The Track values are not combined across multiple gathering intervals (as represented by the CollectIntervals column). Resource Usage Macros and Tables 43 .

summarizing the vprocs of that type on that node. for each log interval. one row written to the database for each type of vproc on each node in the system. for each log interval. For details. one row for each pdisk type per node inserted each logging period. For example. node information specific to the CPUs within the nodes. For details. for large configurations. For details. For details. see Chapter 13: “ResUsageSvpr Table.” ResUsageSldv specific to each logical storage device.” ResUsageSawt specific to the AWTs. see Chapter 11: “ResUsageSps Table. summarizing the CPUs on that node. see Chapter 5: “ResUsageScpu Table. for each log interval. and PPid fields for each log interval.” ResUsageSvpr specific to each virtual processor and its file system. one row written to the database for each node in the system. for each log interval.Chapter 4: Resource Usage Tables Summary Mode in Resource Usage Tables The Table… ResUsageIpma ResUsageSpma ResUsageScpu contains ResUsage data… for available system-wide. For details. one row is written to the database for each node in the system. node information for available system-wide. see Chapter 10: “ResUsageSpdsk Table. for each log interval. summarizing all AMP vdisk data in each node. one row is written to the database for each type of host (network or channel-connected) on each node in the system. and the following information when Summary Mode is active… Summary Mode not applicable to this table.” Summary Mode not applicable to this table. summarizing all AWTs per node.” ResUsageShst specific to the host channels and LANs communicating with Teradata Database. see Chapter 9: “ResUsageSldv Table. summarizing the hosts of that type on that node.” ResUsageSps one row written to the database for each triplet of PGid.” 44 Resource Usage Macros and Tables .” ResUsageSpdsk specific to the pdisk device. For details. For details. see Chapter 12: “ResUsageSvdsk Table. Summary Mode not applicable to this table. enabling Summary Mode minimizes the amount of system resources used. the ResUsageSpdsk table may contain thousands of rows logged during each logging period. see Chapter 7: “ResUsageSawt Table. one row is written to the database for each node in the system. two rows written to the database: one summarizing the system logical devices and one summarizing the Teradata Database logical devices. see Chapter 8: “ResUsageShst Table. For details. VprType. ResUsageSvdsk specific to the vdisk device.

TheDate TheTime n/a n/a Date of the log entry. See the Secs and NominalSecs columns. The Invalid Platform column is a little counterintuitive.) The following table describes the ResUsageScpu table columns. by no more than one or two seconds).CHAPTER 5 ResUsageScpu Table This resource usage table contains resource usage information specific to the CPUs within the nodes. For example. Nominal time of the log entry. INTEGER DATE FLOAT Resource Usage Macros and Tables 45 . Note: Symmetric Multi-Processing (SMP) nodes have a chassis and cabinet number of 1. Type of Data Invalid Platform Column Name Description HOUSEKEEPING COLUMNS Data Type RELATIONAL PRIMARY INDEX COLUMNS These columns taken together form the nonunique primary index. NodeId n/a Identifies the Node. Note: Under conditions of heavy system load. If your platform appears in that column. The Node ID is formatted as CCC-MM. CPU information. entries may be logged late (typically. Table ResUsageScpu includes resource usage data for available system-wide. For more information see “Relational Primary Index” on page 38. then resource usage data for that particular column is either not collected or not valid and should not be used. the node ID of an SMP node is ‘001-01’. but this field will still contain the time value when the entry should have been logged. a node in chassis 9 of cabinet 3 has a node ID of ‘003-09’. Note: This table is created as a MULTISET table. where CCC denotes the three-digit cabinet number and MM denotes the two-digit chassis number of the node. For example. (For more information. see “About the Invalid Platform Column” on page 41.

Possible values are ‘N’ if the row is a non-summary row. the value is zero. FLOAT NodeType CPUId n/a n/a CHAR(8) SMALLINT Secs n/a Actual number of seconds in the log period represented by this row. Normally the same as NominalSecs. if Active Row Filter Mode is enabled. Identifies the summarization status of this row. INTEGER NominalSecs SummaryFlag n/a n/a SMALLINT CHAR Active count FLOAT 46 Resource Usage Macros and Tables . representing the per node system family type. and ‘S if the row is a summary row. to a per-second measurement. for example. Identifies the CPU within this node. but can be different in three cases: • The first interval after a log rate change • A sample logged late because of load on the system • System clock adjustments affect reported Secs Useful for normalizing the count statistics contained in this row. For example. For example. This field is useful when performing data calculations with small elapsed times where the difference between centisecond-based data and whole seconds results in a percentage error. then none of the data columns in the row have been updated during the logging period. In Summary Mode. A specified or nominal number of seconds in the logging period. Controls whether or not the rows will be logged to the ResUsage tables if Active Row Filter Mode is enabled. • a zero value. If Active is set to: • a non-zero value. then the row contains modified data columns. The values are 0 through NCPUs-1. Type of node. then the rows that have a zero Active field value will not be logged to the ResUsage tables. SMALLINT CentiSecs n/a Number of centiseconds in the logging period. 5600C or 5555H.Chapter 5: ResUsageScpu Table Column Name Type of Data Description MISCELLANEOUS HOUSEKEEPING COLUMNS Data Type Invalid Platform GmtTime n/a Greenwich Mean Time is not affected by the Daylight Savings Time adjustments that occur twice a year.

This number can vary from one period to the next. the CPU would be immediately dispatched for that process. they should total to 100 * Secs * number of CPUs on the node. privileged work performing system services on behalf of user execution processes which do not have root access. Time in centiseconds CPU is waiting for I/O completion. subdivided into the following columns: 1 CPUIdle . the values of these four columns. account for total CPU time on the node. FLOAT FLOAT FLOAT Windows CPUUExec count FLOAT Resource Usage Macros and Tables 47 .Idle time 2 CPUIoWait . since each CPU is always in exactly one of these four states. since the CPU is only recorded as being in this state if there are no processes eligible for execution. This is because if there were any such process. that is.Chapter 5: ResUsageScpu Table Column Name CollectIntervals Type of Data n/a Description The number of gather periods per reporting period.Idle and waiting for I/O completion 3 CPUUServ . Time in centiseconds CPU is busy executing user execution code. CPUIdle CPUIoWait count count Time in centiseconds the CPU is idle and not waiting for I/O. including activities performed for virtual processors. the value is the number of Gather operations that have been performed during the period. STATISTICS COLUMNS PROCESS SCHEDULING COLUMNS Data Type SMALLINT Invalid Platform CPU Utilization Columns Count all CPU activities. there is occasionally a very small plus or minus difference from this theoretical total.User execution These statistics are aggregates representing all CPUs on the node. In the Collect Buffer and Log Buffer. That is.User service 4 CPUUExec . CPUUServ count Time in centiseconds CPU is busy executing user service code. time spent in a user state on behalf of a process. Note: • • CPU idle time = CPUIdle + CPUIoWait CPU busy time = CPUUServ + CPUUExec Theoretically. In practice. for any given interval. that is. CPU utilization by user code is further subdivided by the vproc tables. Note: This represents another variety of Idle.

normally. so a displayed value of 500 represents a COD value of 50.0%. You can determine if a row is in Summary Mode by checking the SummaryFlag column for that row. summarizing all CPUs per node. Spare Columns The ResUsageScpu table has six spare columns (one of which is being used) as shown in the table below. Column Name SpareCount[00-01] SpareTrack[00-01] SpareTmon00 Type of Data count track n/a Description Spare counted statistic. The value represents the COD value in one tenths of a percent. IF the SummaryFlag column value is… ‘S’ ‘N’ THEN the data for that row is being logged… in Summary Mode. Number of CPUProcSwitches where a process replaced itself. one row is written to the database for each node. for each log interval. Spare tracked statistic. 48 Resource Usage Macros and Tables . that is. FLOAT FLOAT ALL ALL Summary Mode When Summary Mode is active for tables in this group. Note: This value is valid only on SUSE Linux Enterprise Server 10 systems and is a single value for the entire system. The SpareTmon00 field contains the Capacity on Demand (COD) value. CPUProcSwitches CPUProcSameSwitches count count Number of times the scheduler switched the CPUs currently active process to a new process.Chapter 5: ResUsageScpu Table Summary Mode Column Name Type of Data Description Data Type Invalid Platform Scheduled CPU Switching Columns Identify the number of times the CPU was switched by the scheduler from doing one type of work to another type of work. the new process was the same as the old process.

Chapter 5: ResUsageScpu Table Spare Columns Column Name SpareTmon01 Type of Data count Description Spare tmonitored statistic. The spare column fields expand to values 00 . and so on. Resource Usage Macros and Tables 49 . so that column names would be SpareCount00 or SpareTrack01.01.

Chapter 5: ResUsageScpu Table Spare Columns 50 Resource Usage Macros and Tables .

NodeId n/a Identifies the Node. entries may be logged late (typically. See the Secs and NominalSecs columns. For example. then resource usage data for that particular column is either not collected or not valid and should not be used. For more information see “Relational Primary Index” on page 38. Type of Data Invalid Platform Column Name Description HOUSEKEEPING COLUMNS Data Type RELATIONAL PRIMARY INDEX COLUMNS These columns taken together form the nonunique primary index. the node ID of an SMP node is ‘001-01’. always use the views provided in Chapter 14: “Resource Usage Views” to access the data rather than accessing the ResUsage table directly. This table is created as a MULTISET table. TheDate TheTime n/a n/a Date of the log entry. but this field will still contain the time when the entry should have been logged. The Invalid Platform column is somewhat counterintuitive. node information.” Note: Summary Mode is not applicable to this table. For information on this table. by no more than one or two seconds). The following table describes the ResUsageSpma table columns. For example. Note: Under conditions of heavy system load. where CCC denotes the three-digit cabinet number and MM denotes the two-digit chassis number of the node. Nominal time of the log entry. a node in chassis 9 of cabinet 3 has a node ID of ‘003-09’. The Node ID is formatted as CCC-MM. However. If your platform appears in that column. INTEGER DATE FLOAT Resource Usage Macros and Tables 51 . see Appendix B: “ResUsageIpma Table. The ResUsageSpma table is similar to the ResUsageIpma table.CHAPTER 6 ResUsageSpma Table The ResUsageSpma table includes resource usage data for available system-wide. Note: SMP nodes have a chassis and cabinet number of 1.

Current count of type 6 virtual processors running under the node. Under the current implementation. Vproc1 VprocType1 Vproc2 VprocType2 Vproc3 VprocType3 Vproc4 VprocType4 Vproc5 VprocType5 Vproc6 n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a Current count of type 1 (AMP) virtual processors running on the node. When the vproc is present on the node. Type 2 (PE). This is especially important in coexistence systems where the number of CPUs can vary across system nodes. CHAR(4) ALL SMALLINT CHAR(4) SMALLINT CHAR(4) SMALLINT CHAR(4) SMALLINT CHAR(4) SMALLINT CHAR(4) SMALLINT FLOAT NodeType NCPUs n/a n/a CHAR(8) SMALLINT 52 Resource Usage Macros and Tables . Type of node. This field is useful for normalizing the CPU utilization field values for the number of CPUs on the node. and Type 5 (VSS) vprocs exist. Type of virtual processor for Vproc2. Type 4 (RSG). the value is RSG. Type of virtual processor for Vproc1. Type of virtual processor for Vproc5. Type of virtual processor for Vproc3. the value is AMP.Chapter 6: ResUsageSpma Table Column Name Type of Data Description Data Type Invalid Platform MISCELLANEOUS HOUSEKEEPING COLUMNS These columns provide a generalized picture of the vprocs running on this node. only Type 1 (AMP). Number of CPUs on this node. Current count of type 5 (VSS) virtual processors running under the node. When the vproc is present on the node. When the vproc is present on the node. When the vproc is present on the node. the value is PE. Current count of type 4 (RSG) virtual processors running under the node. Current count of type 3 (GTW) virtual processors running under the node. the value is GTW. the value is VSS. VprocType6 n/a Type of virtual processor for Vproc6. 5600C or 5555H. shown as Type n virtual processors where n = 1 to 7. When the vproc is present on the node. For example. representing the per node system family type. GmtTime n/a Greenwich Mean Time is not affected by the Daylight Savings Time adjustments that occur twice a year. Type 3 (GTW). vproc types 6 through 7 are not currently used. Type of virtual processor for Vproc4. This column reports zeros and " " (blanks). Current count of type 2 (PE) virtual processors running under the node.

to a persecond measurement. Amount of memory on this node in megabytes. A per node normalization factor that is used to normalize the reported CPU values of the ResUsageSpma table. INTEGER NominalSecs CollectIntervals n/a n/a A specified or nominal number of seconds in the logging period. This column reports zeros and " " (blanks). The number of gather periods per reporting period.Chapter 6: ResUsageSpma Table Column Name Vproc7 Type of Data n/a Description Current count of type 7 virtual processors running under the node. For example. Normally the same as NominalSecs. SMALLINT CentiSecs n/a Actual number of centiseconds in the logging period. In the Collect Buffer and Log Buffer. CHAR(4) INTEGER ALL NodeNormFactor n/a INTEGER Secs n/a Actual number of seconds in the log period represented by this row. the value is the number of Gather operations that have been performed during the period. This number can vary from one period to the next.25. Useful for performing memory usage calculations. This field is useful when performing data calculations with small elapsed times where the difference between centisecond-based data and whole seconds results in a percentage error. but can be different in three cases: • The first interval after a log rate change • A sample logged late because of load on the system • System clock adjustments affect reported Secs Useful for normalizing the count statistics contained in this row. SMALLINT SMALLINT Resource Usage Macros and Tables 53 . for example. Data Type SMALLINT Invalid Platform VprocType7 MemSize n/a n/a Type of virtual processor for Vproc7. if the actual factor is 5. This value is scaled by a factor of 100. then the value of the NodeNormFactor will be 525. Note: This value is constant for the node and scaled up by a factor of 100 to preserve the two digit decimal resolution while using an integer field.

Process Allocation Columns These columns represent all currently allocated processes. The CollectIntervals value is the number of gather periods per reporting period. Sample count for sampled statistics for a Bynet. STATISTICS COLUMNS Data Type FLOAT FLOAT Invalid Platform Teradata VS Columns These columns identify pdisk I/O statistics that are reported by Teradata VS. blocked or suspended. divide this value by the CollectIntervals value. divide this value by the CollectIntervals value. For more information. ProcSuspendedSum count Total number of process suspended from execution. ProcReadySum count Number of runnable or ready tasks able to execute on CPUs when a CPU becomes available. FLOAT ALL FLOAT Windows FLOAT 54 Resource Usage Macros and Tables . Note: Teradata VS is available for purchase separately from Teradata Database. see Teradata Virtual Storage. For more information. For details about these columns. ready. see the CollectIntervals column. Note: NetSamples is used to normalize all net time monitored statistics to a percent-of-time basis. see the CollectIntervals column. For more information. Note: A task is a thread. see the CollectIntervals column. ProcBlockedSum count The total number of threads blocked waiting for I/O. awaiting another process to resume them (during a log interval). subdivided into the possible process states of running. The CollectIntervals value is the number of gather periods per reporting period. Note: To calculate the average number of processes suspended. The CollectIntervals value is the number of gather periods per reporting period. to calculate the average number of runnable or ready tasks. Note: To calculate the average number of processes blocked. For example. divide this value by the CollectIntervals value. Also. dividing (NetTxIdle/ NetSamples) yields the transmitter-idle time ratio for the net statistics.Chapter 6: ResUsageSpma Table Column Name Active NetSamples Type of Data n/a count Description Gets set to a non-zero value whenever one of the other data columns in the row is set.

The CollectIntervals value is the number of gather periods per reporting period. ProcPendNetThrottle count Number of processes blocked pending delivery of outstanding outgoing messages. FLOAT Process Pending Snapshot Columns Identify how many processes are blocked for each possible reason. Note: Always divide this value by CollectIntervals. Data Type FLOAT Invalid Platform ALL ProcReadyMax max Maximum number of runnable or ready tasks able to execute on CPUs when a CPU becomes available. a distinction should be made between the following two kinds of process blocks: • Block involves a process that is logically idle. In the Collect Buffer and Log Buffer. This number can vary from one period to the next. Note: In analyzing resource usage. divide the value by the CollectIntervals value. the second kind is represented by the remaining columns described here. FLOAT FLOAT FLOAT FLOAT Resource Usage Macros and Tables 55 . ProcPendMemAlloc count Number of processes blocked pending memory allocations. These columns total (minus ProcPendDBLock) approximately ProcBlockedSum. The first kind of block is represented by column ProcPendNetRead. Note: To calculate the average number of processes running. See "Note on Averages" above. • Block involves a process that has work to do but is being prevented from proceeding by some circumstance like a segment lock or flow control. Note: Always divide this value by CollectIntervals. Note: Always divide this value by CollectIntervals. The CollectIntervals value is the number of gather periods per reporting period. or for a timer to elapse. waiting to receive work on its primary mailbox. For more information. since we can only be blocked on one blocking type at a time.Chapter 6: ResUsageSpma Table Column Name ProcRunningSum Type of Data count Description Total number of processes running (executing) on CPUs during each log interval. This block does not affect throughput. This kind of block does affect throughput. ProcPendFsgWrite count Number of processes blocked pending an FSG write to disk. Note: Always divide this value by CollectIntervals. Note on Averages: To calculate the average number of processes pending. See "Note on Averages" above. ProcPendFsgRead count Number of processes blocked pending a File Segment (FSG) read from disk. See "Note on Averages" above. the value is the number of Gather operations that have been performed during the period. Note: A task is a thread. See "Note on Averages" above. divide this value by the CollectIntervals value. see the CollectIntervals column.

See "Note on Averages" above. FLOAT ProcPendQnl count Number of processes blocked pending a TSKQNL lock. See "Note on Averages" above. FLOAT ProcPendFsgLock count Number of processes blocked pending an FSG lock. FLOAT ProcPendDBLock count Number of processes blocked pending database locks. Data Type FLOAT Invalid Platform ProcPendMonitor count Number of processes blocked pending a user monitor. Note: Always divide this value by CollectIntervals. Note: Always divide this value by CollectIntervals. FLOAT ProcPendSegLock count Number of processes blocked pending a segment lock. Non-step work is not unanticipated work such as a new work request sent when a user initiates a request from the host. See "Note on Averages" above. See "Note on Averages" above. Note: Always divide this value by CollectIntervals. that is. Always divide this value by CollectIntervals. See "Note on Averages" above. FLOAT ProcPendMisc count Number of processes blocked pending miscellaneous events.Chapter 6: ResUsageSpma Table Column Name ProcPendNetRead Type of Data count Description Number of processes blocked pending non-step work. the number of processes blocked on any mailbox other than the work mailbox. See "Note on Averages" above. Note: Non-step work is anticipated work the process spawned off and is now waiting for some type of response from the spawned process or processes. See "Note on Averages" above. FLOAT 56 Resource Usage Macros and Tables . Note: Always divide this value by CollectIntervals. FLOAT ProcPendMonResume count Number of processes blocked pending a user monitor resume from a yield. Note: Always divide this value by CollectIntervals. Note: Always divide this value by CollectIntervals. Note: Always divide this value by CollectIntervals. See "Note on Averages" above.

Number of process blocks for a user monitor. Number of process blocks for an FSG lock. Number of process blocks for a disk or task context (scratch. Average time blocked can be approximated by dividing corresponding ProcWaitXxx by ProcBlksXxx. ProcWaitMemAlloc ProcWaitPageRead ProcWaitFsgRead count count count Total time processes were blocked pending memory allocations. The AWT can do other work while the lock is blocked. stack. and so on) segment lock. Number of process blocks for an FSG write to disk.Chapter 6: ResUsageSpma Table Column Name Type of Data Description Data Type Invalid Platform Process Block Counts Columns Identify how many times a process became blocked on which blocking type. Number of process blocks waiting only for timer expiration. Note: Since this time is only accounted for when a blocked process leaves the blocked state. it is possible for this statistic to be much larger than the amount of time available to all processes in a single log period. Number of process blocks for a user monitor resume from a yield. Number of process blocks for non-step work. Total time processes were blocked pending an FSG read from disk. FLOAT FLOAT FLOAT Resource Usage Macros and Tables 57 . FLOAT FLOAT FLOAT FLOAT FLOAT FLOAT FLOAT FLOAT FLOAT ProcBlksSegLock ProcBlksFsgLock ProcBlksTime ProcBlksMisc count count count count FLOAT FLOAT FLOAT FLOAT Process Pending Wait Time Columns Identify how much time in centiseconds processes were in the blocked state for each possible reason. Number of process blocks for miscellaneous events. ProcBlksMemAlloc ProcBlksQnl ProcBlksFsgRead ProcBlksFsgWrite ProcBlksNetThrottle ProcBlksMsgRead ProcBlksMonitor ProcBlksMonResume ProcBlksDBLock count count count count count count count count count Number of process blocks for memory allocations. Number of process blocks for delivery of outstanding outgoing messages. Number of process blocks for database locks. Number of process blocks for a TSKQNL lock. Number of process blocks for an FSG read from disk. Total time processes were blocked pending a page read from disk.

including activities performed for virtual processors. IOWait. Userv. Total time processes were blocked pending a user monitor.User service 4 CPUUExec . Total time processes were blocked pending nonstep work. Data Type FLOAT FLOAT FLOAT FLOAT FLOAT FLOAT FLOAT Invalid Platform ProcWaitFsgLock ProcWaitTime ProcWaitQnl ProcWaitMisc count count count count FLOAT FLOAT FLOAT FLOAT CPU Utilization Columns Count all CPU activities. Total time processes were blocked pending delivery of outstanding outgoing messages. Total time processes were blocked pending a user monitor resume from a yield. Uexec). CPU(Idle.Normalized user execution These columns represent the sum of all CPUs on the node.Normalized idle and waiting for I/O completion 7 CPUUServNorm . Total time processes were blocked pending an FSG lock. To obtain the average node CPU value for each column.Normalized idle time 6 CPUIOWaitNorm . and so on) segment lock. Total time processes were blocked pending a TSKQNL lock. Total time processes were blocked pending miscellaneous events.Chapter 6: ResUsageSpma Table Column Name ProcWaitFsgWrite ProcWaitNetThrottle ProcWaitMsgRead ProcWaitMonitor ProcWaitMonResume ProcWaitDBLock ProcWaitSegLock Type of Data count count count count count count count Description Total time processes were blocked pending an FSG write to disk. Total time processes were blocked pending some amount of elapsed time only. stack. Total time processes were blocked pending a disk or task context (scratch. Total time processes were blocked pending database locks.Idle and waiting for I/O completion 3 CPUUServ .User execution 5 CPUIdleNorm .Idle time 2 CPUIoWait . divide the column data by the number of CPUs per node (the value in the NCPUs column) and the number of centiseconds (CentiSecs column) in the logging interval. subdivided into the following columns: 1 CPUIdle . 58 Resource Usage Macros and Tables .Normalized user service 8 CPUUExecNorm .

You can calculate the CPU time by using the formula below. privileged work performing system services on behalf of user execution processes which do not have root access. This is related to the NodeType value reported in this resource usage table.Chapter 6: ResUsageSpma Table Column Name Note: • • • • • • Type of Data Description Data Type Invalid Platform CPU idle time = CPUIdle + CPUIoWait CPU busy time = CPUUServ + CPUUExec CPUIdleNorm = (CPUIdle * NodeNormFactor)/100 CPUIOWaitNorm = (CPUIoWait * NodeNormFactor)/100 CPUUServNorm = (CPUUServ * NodeNormFactor)/ 100 CPUUExecNorm = (CPUUExec * NodeNormFactor)/100 where the NodeNormFactor is the per node normalization factor. Time in centiseconds CPUs are busy executing user execution code. If there was a process available. IoWait. Time in centiseconds CPUs are idle and waiting for I/O completion. since a CPU is only in this state if there are no processes eligible for execution. FLOAT FLOAT Windows CPUUServ count Time in centiseconds CPUs are busy executing user service code. CPUIdle CPUIoWait count count Time in centiseconds CPUs are idle and not waiting for I/O. the CPU would be immediately dispatched for that process. Time in centiseconds CPUs are idle and not waiting on I/O. The normalization factor modifies the reported CPU times to the equivalent time of a specified virtual processor. Time in centiseconds CPUs are idle and waiting for I/O completion. the value is always 0. On Windows. FLOAT CPUUExec count FLOAT CPUIdleNorm CPUIOWaitNorm count count FLOAT FLOAT Windows Resource Usage Macros and Tables 59 . UExec) The CPU time returned in centiseconds is more accurate than those returned in seconds. that is. time spent in a user state on behalf of a process. On Windows. the value is always 0. use the following formula: 100 x Secs x NCPUs ≈ CentiSecs x NCPUs = CPU(Idle. that is. UServ. This does not add up to the reported CPU time. CPUIdleNorm + CPUIOWaitNorm + CPUUServNorm + CPUUExecNorm = CentiSecs * NCPUs * NodeNormFactor To calculate the non-normalized total CPU time. Note: This time represents another variety of Idle.

subdivided into (the only applicable) generic node memory type and a summarization of vproc memory types. of memory resident subdivided into memory types. time spent in a user state on behalf of a process. disk segments and task context types. It represents a delta from the previous reporting period. FLOAT ALL FLOAT ALL MemVprAllocs count FLOAT ALL MemVprAllocKB count FLOAT Windows MemTSysOhRes 60 Resource Usage Macros and Tables . that is. that is. Number of successful memory allocations and size-increasing memory alters for all vproc memory types. Memory Pages Resident Columns Identify the amount. it will report negative values as less memory is used. Time in centiseconds CPUs are busy executing user execution code. Note: The original meaning of this column was the total KBs attributed to allocations and sizeincreasing alters for vproc memory types. Each of these expands into six columns. MemTextAllocs count Number of successful memory allocations and size-increasing memory alters for non-system overhead text (code). Amount allocated can be derived by multiplying the number of allocations by the fixed page size. where [seg] is as follows: • • • • • • PDb = Permanent data block disk segments PCi =Permanent cylinder index disk segments SDb =Regular or restartable spool data block disk segments SCi = Regular or restartable spool cylinder index disk segments TJt = Transient journal table or WAL data block or WAL cylinder index APt = Append table or permanent journal table data block or cylinder index disk segments track Number of pages resident in memory for system overhead text. in number of pages or KBs. Thus. privileged work performing system services on behalf of user execution processes which do not have root access. that is. The value represents the change in memory.Chapter 6: ResUsageSpma Table Column Name CPUUServNorm Type of Data count Description Time in centiseconds CPUs are busy executing user service code. MEMORY COLUMNS Data Type FLOAT Invalid Platform CPUUExecNorm count FLOAT Memory Allocation Columns Identify the number and amount of memory allocations. Disk segment memory types are described by the single entries below. System Overhead Text is wired into memory upon startup and will not change.

The following formula is used by the RSS to calculate the MemFreeKB value. This value should be equal to the size of memory minus the total amount resident derived from adding all of the above memory resident columns and frozen disk segment resident column from ResUsageSvpr.Chapter 6: ResUsageSpma Table Column Name MemDSysOhRes Type of Data track Description Number of pages resident in memory for system overhead data. Number of times memory was aged. • active_slabs and pgsperslab come from /proc/ slabinfo. Buffers. Data Type FLOAT Invalid Platform ALL MemTextRes MemCtxtRes Mem[seg]KBRes MemFreeKB track track track track FLOAT FLOAT FLOAT FLOAT ALL ALL ALL Memory Availability Management Columns Identify overhead to managing memory when memory availability is a problem. MemFails count Number of failures performing memory allocations and size-increasing memory alters for vproc memory types as well as node memory types. MemFreeKB = MemFree + Buffers + Cached + SwapCached . The Linux operating system uses most free memory for buffers and caching to improve performance. and SwapCached come from /proc/meminfo. • fsgavailpgs come from the PDE FSG code. Cached. Number of pages resident in memory for text. the value reported is the approximate amount of memory that is available for use. KBs of free memory. but the operating system can reclaim that memory if it is needed by programs. System Overhead Data is wired into memory upon startup. Number of non-system overhead text pages dropped from memory to make more physical memory available. On Linux. Number of pages resident in memory for task context segments.fsgavailpgs*kbperpage (active_slabs*pgsperslab*kbperpage) where the values: • MemFree. FLOAT ALL MemAgings MemTextPageDrops count count FLOAT FLOAT ALL ALL Resource Usage Macros and Tables 61 . Current KBs resident in memory for (nonbackup) disk segments.

KBs dropped from memory by MemSwapDrops. Writes) and amount (ReadKB. NET COLUMNS FLOAT ALL MemSwapDropKB MemSwapReads count count FLOAT FLOAT ALL ALL MemSwapReadKB count FLOAT ALL Point-to-Point Net Traffic Columns Identify the number (Reads. This value is less than. the number of 4KB pages swapped out. and so on) pages that were paged out. KBs re-read from memory by MemSwapReads. stack. FLOAT ALL MemCtxtPageWrites count FLOAT MemCtxtPageReads count Number of task context (scratch. Number of disk segments that were re-read when they were previously dropped from memory because all its ancestor processes were swapped out. total processes allocated. On Linux. Number of task context (scratch. On Linux.Chapter 6: ResUsageSpma Table Column Name MemTextPageReads Type of Data count Description Number of non-system overhead text pages required to be read from disk when it was previously paged out. On Linux. and so on) pages that were paged in. Number of net point-to-point messages output from processes on the node via the message subsystem. FLOAT MsgPtPWrites count FLOAT MsgPtPReadKB count FLOAT 62 Resource Usage Macros and Tables . the number of 4KB pages paged minus the pages swapped in. It excludes TCP/IP traffic. WriteKB) of input and output messages passing through the Teradata Database nets through point-to-point (1:1) methods (PtP). MsgPtPReads count Number of net point-to-point messages input to processes on the node via the message subsystem. the number of 4KB pages swapped in. Total KBs of net point-to-point messages input to processes on the node via the message subsystem. Data Type FLOAT Invalid Platform MemProcSwapped track Current count of processes whose stack has been written to disk to make available more physical memory. stack. FLOAT MemSwapDrops count Number of disk segments that were dropped from memory because all its ancestor processes were swapped out. or equal to.

FLOAT MsgBrdWrites count FLOAT MsgBrdReadKB count FLOAT MsgBrdWriteKB count FLOAT Network Transport Data Columns Identify the number (Reads. FLOAT FLOAT FLOAT FLOAT FLOAT FLOAT FLOAT FLOAT Resource Usage Macros and Tables 63 . Total KBs of net broadcast messages output from processes on the node via the message subsystem. Amount of broadcast message data in KBs received by both Bynets. Amount of point-to-point message data in KBs received by both Bynets. Data Type FLOAT Invalid Platform Broadcast Net Traffic Columns Identify the number (Reads. Writes) and amount (ReadKB. Number of net broadcast messages output from processes on the node via the message subsystem. they do not take into consideration which Bynet performed the transport. WriteKB) of input and output messages passing through the Teradata Database nets through broadcast (1:many) methods (Brd). The number of broadcast messages received by both Bynets. Amount of broadcast message data in KBs transmitted by both Bynets. Note: If a single broadcast message is delivered to multiple processes in this node. MsgBrdReads count Number of net broadcast messages input to processes on the node via the message subsystem. that is. the NetBrdReads and NetBrdReadKB are only incremented once. The number of broadcast messages transmitted by both Bynets The number of point-to-point messages received by both Bynets. These statistics are nonspecific. The number of point-to-point messages transmitted by both Bynets. Total KBs of net broadcast messages input to processes on the node via the message subsystem.Chapter 6: ResUsageSpma Table Column Name MsgPtPWriteKB Type of Data count Description Total KBs of net point-to-point messages output from processes on the node via the message subsystem. Writes) and amount of input and output (PDE messages routed by the message subsystem) passing through the Teradata Database nets. NetMsgPtpWriteKB NetMsgBrdWriteKB NetMsgPtpReadKB NetMsgBrdReadKB NetMsgPtpWrites NetMsgBrdWrites NetMsgPtpReads NetMsgBrdReads count count count count count count count count Amount of point-to-point message data in KBs transmitted by both Bynets.

” Note: Circuit attempts for one or both Bynets can be computed as the sum of the applicable NetTxCircPtp and NetTxCircBrd columns. Total point-to-point KBs received over all Bynets.Chapter 6: ResUsageSpma Table Column Name Type of Data Description Data Type Invalid Platform Per-Bynet Network Transport Data Columns Identify the amount of input and output passing through the Teradata Database nets. FLOAT FLOAT FLOAT 64 Resource Usage Macros and Tables . netspecific statistics are not meaningful and are always zero. On a single-node system. that is. FLOAT FLOAT FLOAT FLOAT Net Controller Status and Miscellaneous Management Provide utilization and other status information about the Teradata Database net controllers. NetTxKBPtP NetRxKBPtP NetTxKBBrd NetRxKBBrd count count count count Total point-to-point KBs transmitted over all Bynets. Number of samples showing the receiver idle on a Bynet. net-specific statistics are not meaningful and are always zero. These statistics are net-specific. Additional detail is found in Appendix B: “ResUsageIpma Table. These statistics are not net-specific since all the Bynet statistics are reported in the net columns. On a single-node (vnet) system. All of these columns except for NetCircBackoffs are net-specific. Total broadcast KBs transmitted over all Bynets. Number of high priority point-to-point circuits transmitted on all Bynets. On a single-node (virtual network [vnet]) system. Number of samples showing the transmitter connected on a Bynet. Number of samples showing the receiver connected on a Bynet. NetTxRouting NetTxConnected NetRxConnected NetTxIdle NetRxIdle count count count count count Number of samples showing the transmitter routing on a Bynet. NetTxCircHPBrd NetRxCircPtp NetTxCircHPPtP count count count Number of high priority broadcast circuits transmitted on all Bynets. net-specific statistics are not meaningful and are always zero. Number of samples showing the transmitter idle on a Bynet. they relate to each specific Bynet. FLOAT FLOAT FLOAT FLOAT FLOAT Net Circuit Management Columns Identify the management of Teradata Database net circuits (Circ). Total number (both normal and high priority) of point-to-point circuits received on all Bynets. Total broadcast KBs received over all Bynets.

Maximum number of groups in use during each log interval. Data Type FLOAT FLOAT FLOAT Invalid Platform NetHWBackoffs NetTxCircPtp count count FLOAT FLOAT Group Coordination Messages Columns Identify messages that are communicated through the Teradata Database net for coordination of a process among a group of vprocs.Chapter 6: ResUsageSpma Table Column Name NetRxCircBrd NetTxCircBrd NetCircBackoffs Type of Data count count count Description Total number (both normal and high priority) of broadcast circuits received on all Bynets. NetMrgTxKB count Number of KBs transmitted. FLOAT Resource Usage Macros and Tables 65 . This is used in tracking down the slowest node or AMP in the system. Maximum number of semaphores in use during each log interval. Total number of channels in use during each log interval. or channels. Hardware backoffs reported by the BLM for all Bynets. without regard for which net was involved. Total number of groups in use during each log interval. Software backoffs. NetSemInUseSum NetSemInUseMax NetChanInUseSum NetChanInUseMax NetGroupInUseSum count max count max count Total number of semaphores in use during each log interval. MsgChnLastDone count Number of last done events that occurred on this node. FLOAT FLOAT FLOAT FLOAT FLOAT FLOAT NetGroupInUseMax max FLOAT Merge Services Columns Identify activity occurring through merge (many:1) methods (Mrg) on Teradata Database net. by merge transmission services for currently active merge operations. Maximum number of channels in use. Note: The last AMP to finish an operation may send a last done broadcast message indicating the work is done for this step. Total number (both normal and high priority) of broadcast circuits transmitted on all Bynets. Coordination is handled either through semaphores. without regard to which net. A node or AMP that has more last done messages than the others could be a bottleneck in the system performance. This number should be same across all nodes. groups. Total number (both normal and high priority) of point-to-point circuits transmitted on all Bynets. defined as BNS service blocked occurrences.

FLOAT FLOAT FLOAT FLOAT FLOAT FLOAT FLOAT FLOAT FLOAT 66 Resource Usage Macros and Tables . and KBs. without regard to which net. HOST CONTROLLER COLUMNS Data Type FLOAT Invalid Platform NetMrgTxRows count FLOAT NetMrgRxRows count FLOAT Channel Traffic Columns Identify the traffic between the host and the node in three levels of granularity: blocks. GENERAL CONCURRENCY CONTROL COLUMNS Database Locks Columns Identify database locking occurrences. DBLockBlocks DBLockDeadlocks count count Number of times a database lock was blocked. Number of data rows received. Number of messages written out to the host. Blocks are made up of some amount of variable sized messages. by merge transmission services for currently active merge operations. messages. without regard to which net. Number of messages read in from the host. FileAcqs count Total number of logical disk segments acquired. These columns identify the total disk memory segments acquired by the file system during the log period. KBs transferred in from the host. Acquires causing physical reads (AcqReads) and the amount read (AcqReadKB) are identified as a subset of logical acquires. without regard to which net. Number of times a database lock was deadlocked. Number of data rows transmitted. HostBlockReads HostBlockWrites HostMessageReads HostMessageWrites HostReadKB HostWriteKB count count count count count count Number of blocks read in from the host.Chapter 6: ResUsageSpma Table Column Name NetMrgRxKB Type of Data count Description Number of KBs received. For more detail. FILE SYSTEM COLUMNS Segments Acquired Columns Summarize logical and physical segments acquired by the file system. Number of blocks written out to the host. see “Segment Acquires Columns” on page 136 in the ResUsageSvpr Table chapter. ReadKB and WriteKB identify the KBs involved in the traffic. KBs transferred out to the host. by merge receive services for currently active merge operations. Logical acquires (Acqs) and the logical amount acquired (AcqKB) are identified. by merge receive services for currently active merge operations.

Total KBs physically read by FileAcqReads. there will be extra sectors read in on cylinder reads. For more information on cylinder read. the cylinder reads revert back to the original algorithm of a block-at-a-time read ahead. FilePres count Total number of times a logical data prefetch was performed (either as a cylinder read or individual block reads). FileWrites FileWriteKB count count Total number of disk segment immediate or delayed physical writes. The number of data blocks that are pre-read at a time is controlled by the DBS Control performance parameter ReadAhead Count. FLOAT FLOAT FLOAT FLOAT Data Block Prefetches Columns Summarize the effects of prefetching data blocks on the file system. FLOAT Resource Usage Macros and Tables 67 . see Performance Management. Note: Use the views provided in Chapter 14 instead of accessing the data for this field directly from this table. Either of these operations are generically called a prefetch. For more detail. FileRels FileRelKB count count Total number of logical disk segments released by tasks. For more detail. Total KBs logically released by FileRels. Total KBs physically written by FileWrites. Data Type FLOAT Invalid Platform FileAcqReads FileAcqReadKB count count Total number of disk segment acquires that caused a physical read. This also applies to the physical pre-reads. see “Segments Released Columns” on page 138 in the ResUsageSvpr Table chapter. Note: A prefetch is either a cylinder read operation or individual block reads operation. Note: Use the views provided in Chapter 14 instead of accessing the data for this field directly from this table. When all cylinder slots are in use. So the column FilePreKB is the sum of the size of data blocks logically read by either cylinder reads or data block pre-reads. FilePreReadKB includes both physical cylinder reads and single block pre-reads. see “Data Block Prefetches Columns” on page 137 in the ResUsageSvpr Table chapter. FLOAT FLOAT Segments Released Columns Summarize logical and physical segments released by the file system.Chapter 6: ResUsageSpma Table Column Name FileAcqKB Type of Data count Description Total KBs logically acquired by FileAcqs. The default is 1 block at a time pre-read. If you enable cylinder reads. An accurate calculation of the wasted kilobytes read by cylinder read is not possible since there are legitimate logical pre-reads that do not incur physical prereads.

The small Depot is typically used when the inplace writes are initiated by a foreground task. Note: Use the views provided in Chapter 14 instead of accessing the data for this field directly from this table. this field does not include the disk sectors in between the loaded data blocks. For cylinder reads. Data Type FLOAT Invalid Platform FilePreReads count Number of times a data prefetch was physically performed either as a cylinder read or individual blocks read. therefore. see “Data Segment Lock Requests Columns” on page 139 in the ResUsageSvpr Table chapter. FileLockBlocks FileLockDeadlocks FileLockEnters count count count Number of lock requests that were blocked.Chapter 6: ResUsageSpma Table Column Name FilePreKB Type of Data count Description Sum of the sizes of data blocks logically loaded with data prefetches (either cylinder reads or individual block reads). FLOAT FilePreReadKB count FLOAT Data Segment Lock Requests Columns Summarize the number of lock requests. For more detail. FileSmallDepotWrites count Number of small writes to the depot performed to protect in-place modifications. Each small Depot write protects a single in-place write of either a WAL data block or a database data block. Number of times a lock was requested. this field includes the disk sectors in between the loaded data blocks. and deadlocks on a disk segment. For cylinder reads. Number of deadlocks detected on lock requests. Small Depot writes are also counted against FileWrites. FLOAT FLOAT FLOAT Depot Columns Summarize the physical writes to the Depot used to protect in-place modifications. The size of the data prefetch (cylinder section or individual blocks being read) that is physically loaded from disk. FLOAT 68 Resource Usage Macros and Tables . FileWrites still indicates the total writes regardless of whether it was a Depot write or a database write. blocks.

FileWrites still indicates the total writes regardless of whether it was a Depot write or a database write. Number of statements that departed normally. FLOAT FLOAT FLOAT FLOAT FLOAT FLOAT FLOAT User Command Arrival and Departure Columns Summarize the arrival and departure of user statements. replace. Number of statements that departed in error. Number of select commands. see Chapter 8: “ResUsageShst Table. The large Depot is typically used when blocks age out of memory in the background. For more detail. Number of statements that departed in failure or were aborted. grant or revoke commands. For more detail. FLOAT FLOAT FLOAT FLOAT ALL Resource Usage Macros and Tables 69 . create. therefore.” CmdDDLStmts CmdDeleteStmts CmdInsertStmts CmdSelectStmts CmdUpdateStmts CmdUtilityStmts CmdOtherStmts count count count count count count count Number of alter. the following calculation results in the average number of blocks protected by each large Depot write: FileLargeDepotBlocks / FileLargeDepotWrites USER COMMANDS COLUMNS Data Type FLOAT Invalid Platform FileLargeDepotBlocks count FLOAT User Command Columns Summarize the type of statements given to Teradata Database by the user. Number of utility commands. modify. Each large Depot write protects multiple in-place writes of either WAL data blocks or database data blocks. Total number of blocks (either WAL or database) that have been protected by large Depot writes.” CmdStmtsInProgCur CmdStmtSuccesses CmdStmtFailures CmdStmtErrors count count count count Current count of statements in progress. Large Depot writes are also counted against FileWrites. see Chapter 8: “ResUsageShst Table. Since a large Depot write protects multiple blocks. drop. Number of update commands. Number of delete commands. Number of insert commands.Chapter 6: ResUsageSpma Table Column Name FileLargeDepotWrites Type of Data count Description Number of large writes to the depot performed to protect in-place modifications. Number of other commands.

The sch peak value is the Max value since startup is never set and Max is the maximum allowed value. For more information. Note: This reported Max value is the maximum reached during each log period. AwtInuseMax max Peak number of AWTs (Max) on this node.Chapter 6: ResUsageSpma Table Column Name CmdStmtTime Type of Data count Description The sums of the resident time of each statement in progress during the log period. see Chapter 7: “ResUsageSawt Table. The AwtInuse value in a log period is the summation of the values of the gather periods comprising the log period. including the successes and failures. This is not the Peak or the Max value stored in the Priority Scheduler (sch) data structure and reported by the puma utility. This is why the AwtInuse value needs to be divided by the CollectIntervals value. 32. The AwtInuse value of a log period can be larger than the AwtInuseMax value of the log period if a log period consists of multiple gather periods. which would be 32. and the max value of each period is 10. if there are 4 gather periods in one log period. Divide the value for AwtInUse by the CollectIntervals value to obtain an average. For more information about the ResUsageSawt table and columns. TERADATA ASM COLUMNS Data Type FLOAT Invalid Platform ALL AMP Worker Task Columns Collect and report statistics about the AWTs. FLOAT FLOAT FLOAT AwtInuse max FLOAT 70 Resource Usage Macros and Tables . see the CollectIntervals column.” AwtFlowControlled AwtFlowCtlCnt count count Number of AMPs currently in flow control on the work input mailbox. For example. and 15. 7. The CollectIntervals value is the number of gather periods per reporting period. Number of times this log period that the node entered the flow control state from a non-flow controlled state. the max of the log period would be the max of the individual gather periods. Number of AWTs currently in use for this node.

divide ServiceTime by NumRequests. so a displayed value of 500 represents a COD value of 50.0%. so that column names would be SpareCount01 or SpareTrack02. Time in centiseconds that work requests waited on an input queue before being serviced. To get an approximate average ServiceTime per request during this period. Note: This value is valid only on SUSE Linux Enterprise Server 10 systems and is a single value for the entire system. For more information about the ResUsageSps table and columns. The spare column fields expand to values 00 . The value represents the COD value in one tenths of a percent. PSServiceTime count Time in centiseconds that work requests required for service. divide QWaitTime by NumRequests.” PSNumRequests PSQWaitTime count count Number of work requests received for all Performance Groups on this node. see Chapter 11: “ResUsageSps Table. SpareTmon[01-03] count Spare time monitored statistic. and so on. To get an approximate average QWaitTime per request during this period. Spare tracked statistic. The SpareTmon00 field contains the COD value. Resource Usage Macros and Tables 71 .03.Chapter 6: ResUsageSpma Table Spare Columns Column Name Type of Data Description Data Type Invalid Platform Priority Scheduler Columns Provides data specific to the Priority Scheduler. Column Name SpareCount[00-03] SpareTrack[00-03] SpareTmon00 Type of Data count track n/a Description Spare counted statistic. FLOAT FLOAT FLOAT Spare Columns The ResUsageSpma table has 12 spare columns (one of which is being used) as shown in the table below.

Chapter 6: ResUsageSpma Table Spare Columns 72 Resource Usage Macros and Tables .

5600C or 5555H. For example. Nominal time of the log entry. Note: SMP nodes have a chassis and cabinet number of 1. see “Summary Mode” on page 77. the node ID of an SMP node is ‘001-01’. For example. entries may be logged late (typically. For details. To consolidate and summarize the total number of rows written to the database. where CCC denotes the three-digit cabinet number and MM denotes the two-digit chassis number of the node. but this field will still contain the time value when the entry should have been logged. For example. a node in chassis 9 of cabinet 3 has a node ID of ‘003-09’. If table logging is enabled.CHAPTER 7 ResUsageSawt Table The ResUsageSawt table collects and reports statistics about the AWTs. then data is written to the database once for each log period. Type of node. The following table describes the ResUsageSawt table columns. MISCELLANEOUS HOUSEKEEPING COLUMNS GmtTime n/a Greenwich Mean Time is not affected by the Daylight Savings Time adjustments that occur twice a year. Note: Under conditions of heavy system load. The Node ID is formatted as CCC-MM. FLOAT INTEGER DATE FLOAT NodeType n/a CHAR(8) Resource Usage Macros and Tables 73 . Type of Data Invalid Platform Column Name Description HOUSEKEEPING COLUMNS Data Type RELATIONAL PRIMARY INDEX COLUMNS These columns taken together form the nonunique primary index. you can enable Summary Mode. TheDate TheTime n/a n/a Date of the log entry. representing the per node system family type. Note: This table is created as a MULTISET table. NodeId n/a Identifies the Node upon which the vproc resides. See the Secs and NominalSecs columns. by no more than one or two seconds).

then none of the data columns in the row have been updated during the logging period. Identifies the summarization status of this row. For details. if Active Row Filter Mode is enabled. This number can vary from one period to the next. Actual number of seconds in the log period represented by this row. All Vprocs in this table are AMPS so there is no VprType field provided. If the value is ‘N. If the value is ‘S. Controls whether or not the rows will be logged to the ResUsage tables if Active Row Filter Mode is enabled. In Summary Mode. to a per-second measurement. see “Summary Mode” on page 77. for example. For example. • a zero value. A specified or nominal number of seconds in the logging period. then the row contains modified data columns. this field is zero. Data Type INTEGER Invalid Platform Secs n/a SMALLINT CentiSecs n/a Number of centiseconds in the logging period.Chapter 7: ResUsageSawt Table Column Name VprId Type of Data n/a Description Identifies the vproc number.’ the row is a summary row. the value is the number of Gather operations that have been performed during the period. This field is useful when performing data calculations with small elapsed times where the difference between centisecond-based data and whole seconds results in a percentage error. but can be different in three cases: • The first interval after a log rate change • A sample logged late because of load on the system • System clock adjustments affect reported Secs Useful for normalizing the count statistics contained in this row. Normally the same as NominalSecs. SMALLINT 74 Resource Usage Macros and Tables . In the Collect Buffer and Log Buffer. then the rows that have a zero Active field value will not be logged to the ResUsage tables. If Active is set to: • a non-zero value.’ the row is a non-summary row. INTEGER NominalSecs SummaryFlag n/a n/a SMALLINT CHAR Active n/a FLOAT CollectIntervals n/a The number of gather periods per reporting period.

Current number of AWTs in use during the log period for each work type for the VprId vproc. Note: This value reports the SUM of the values reported in each gather period when there are multiple gather intervals in each log period. If the value is non-zero. It should be divided by the CollectIntervals column to get the average value. then the AMP is in flow control. Note: This value reports the SUM of the values reported in each gather period when there are multiple gather intervals in each log period. Use the tdntune utility to determine the settings for Flow Control.Chapter 7: ResUsageSawt Table Column Name Type of Data Description STATISTICS COLUMNS TERADATA ASM COLUMNS AMP Worker Task Columns Data Type Invalid Platform Collect and report statistics about the AWTs. WorkTypeMax00 WorkTypeMax15 max Maximum number of AWTs in use at one time during the log period for each work type for the VprId vproc. MailBoxDepth count The average depth of the AMP work mailbox. FlowControlled FlowCtlCnt track count Specifies if an AMP is in flow control. Number of times during the log period that the system entered the flow control state from a nonflow controlled state. and WorktypeInuse01 contains the values for MSGWORKONE. Maximum number of AWTs in use at any one time during the log period. WorktypeInuse00 contains the number of in use AWTs that are of Work Type MSGWORKNEW. this can be used to determine characteristics of the system during skew conditions or when there are AWT shortages. For information on Expedited Allocation Groups. The total time in milliseconds that an AMP is in flow control. This can be used to determine if the usage is close to the maximum values defined and what type of work they are doing. FLOAT FLOAT FLOAT FLOAT FlowCtlTime InuseMax WorkTypeInuse00 WorkTypeInuse15 count max count FLOAT FLOAT FLOAT Resource Usage Macros and Tables 75 . the WorkTypeMax field values are the Max of the values for all the AMPS. In Summary Mode. These columns allow the user to monitor the usage of the AWTs of each work type. WORK TYPE DESCRIPTIONS The WorkTypeInuse and WorkTypeMax array data columns above each contain 16 Work Type entries that are described here. It should be divided by the CollectIntervals column to get the average value. For example. Also. see "Priority Scheduler (schmon)" chapter of Utilities.

MultiLoad.Chapter 7: ResUsageSawt Table Column Name MSGWORKNEW Type of Data n/a Description Used for new work requests. If the utilities are not configured to use a separate pool of work types. Not used. n/a MSGWORKTWO MSGWORKTHREE MSGWORKFOUR MSGWORKFIVE n/a n/a n/a n/a n/a n/a n/a n/a MSGWORKSIX n/a n/a MSGWORKSEVEN n/a n/a MSGWORKEIGHT MSGWORKNINE MSGWORKTEN MSGWORKELEVEN n/a n/a n/a n/a n/a n/a n/a n/a 76 Resource Usage Macros and Tables . For example. This work type has the lowest number. New work for Expedited Allocation Groups. and FastExport. MSGWORKONE. Data Type n/a Invalid Platform MSGWORKONE n/a First level secondary work items. MultiLoad. Special types of database work. which means it is queued last. First level secondary work spawned work for utilities such as FastLoad. if utilities are configured to use a separate pool of work types. Second level secondary work for utilities such as FastLoad. MSGWORKONE. MultiLoad. However. and FastExport. and FastExport. This field is not normally used and MSGWORKNEW. and so on. and MSGWORKTWO. Start System Recover. MSGWORKONE. First level spawned work for Expedited Allocation Groups. Second level secondary work items. and MSGWORKTWO report work requests for utilities. A zero value is used for new work items. Second level spawned work for Expedited Allocation Groups. Each numbered work type is queued for delivery just before the one from which it is spawned. this field reports new work for utilities such as FastLoad. and MSGWORKTWO. Numbered work types are used for secondary work items. work type two (MSGWORKTWO) requests are spawned from work type one requests and queued for delivery before work type one requests. they use MSGWORKNEW. they use MSGWORKNEW. It also has the effect of honoring secondary requests needed to complete existing work items before any new ones are started. work type one (MSGWORKONE) is used for secondary work requests spawned by new work items. If the utilities are not configured to use a separate pool of work types.

The array number for MSGWORKABORT is 12. n/a MSGWORKCONTROL n/a Used for system control messages. You can determine if a row is in Summary Mode by checking the SummaryFlag column for that row. IF the SummaryFlag column value is… ‘S’ ‘N’ THEN the data for that row is being logged… in Summary Mode. normally. one row is written to the database for each node in the system for each log interval. This work type has a higher value than the numbered work types so that abort requests are honored before beginning any additional work item for the transactions being aborted.Chapter 7: ResUsageSawt Table Summary Mode Column Name MSGWORKABORT Type of Data n/a Description Used for transaction abort requests. This work type is delivered before any of the work items described above. Data Type n/a Invalid Platform MSGWORKSPAWN n/a Used for spawned abort requests and is delivered before normal aborts. These are delivered before any other kind of message. Resource Usage Macros and Tables 77 . The array number for MSGWORKSPAWN is 13. The AWT data will be combined for all the AMP vprocs on the node. The array number for MSGWORKNORMAL is 14. n/a MSGWORKNORMAL n/a Used for messages that do not fall within the standard work type hierarchy. The array number for MSGWORKCONTROL is 15. n/a Summary Mode When Summary Mode is active for the ResUsageSawt table.

Column Name SpareCount[00-01. if 12 of the normally 62 unreserved AWTs are removed from the pool by reserving them for expedited work. The SpareTmon00 field contains the COD value. and so on. 04-09] SpareCount[02-03] Type of Data count count Description Spare counted statistic. 78 Resource Usage Macros and Tables . • SpareCount02=Available. For example. If in this log period. than there would be only 40 available at the end of the log period. a value of 0 for SpareCount03 means there were no unreserved AWTs available in the pool at some point during the reporting period. SpareTrack[00-09] SpareTmon00 track n/a Spare tracked statistic. • SpareCount03=AvailableMin. SpareTmon08. Available is the number of unreserved AWTs from the pool that are not being used at the end of the interval.0. 10 unreserved AWTs are taken from the pool to service 10 queries that are still executing. SpareTrack03. The value represents the COD value in one tenths of a percent. The following SpareCount columns will be converted to the specified column names in Teradata Database 14. AvailableMin is the minimum number of unreserved AWTs available in the pool for each AMP for the logged period. Note: This value is valid only on SUSE Linux Enterprise Server 10 systems and is a single value for the entire system. For example. SpareTmon[01-09] count Spare time monitored statistic. so a displayed value of 500 represents a COD value of 50. The spare column fields expand to values 00–09.Chapter 7: ResUsageSawt Table Spare Columns Spare Columns The ResUsageSawt table has 30 spare columns (one of which is being used) as shown in the table below.0%. there could at most be 50 unreserved AWTs available. so that column names would be SpareCount00.

Note: Under conditions of heavy system load. entries may be logged late (typically. FLOAT INTEGER DATE FLOAT NodeType n/a CHAR(8) Resource Usage Macros and Tables 79 . 5600C or 5555H. MISCELLANEOUS HOUSEKEEPING COLUMNS GmtTime n/a Greenwich Mean Time is not affected by the Daylight Savings Time adjustments that occur twice a year. the node ID of an SMP node is ‘001-01’. For example. a node in chassis 9 of cabinet 3 has a node ID of ‘003-09’. The following table describes the ResUsageShst table columns. The Node ID is formatted as CCC-MM. representing the per node system family type. by no more than one or two seconds). Type of Data Invalid Platform Column Name Description HOUSEKEEPING COLUMNS Data Type RELATIONAL PRIMARY INDEX COLUMNS These columns taken together form the nonunique primary index.CHAPTER 8 ResUsageShst Table The ResUsageShst table: • • Contains resource usage information specific to the host channels and LANs communicating with Teradata Database. Note: This table is created as a MULTISET table. For example. host information. Note: SMP nodes have a chassis and cabinet number of 1. See the Secs and NominalSecs columns. but this field will still contain the time value when the entry should have been logged. NodeId n/a Identifies the Node upon which the vproc resides. Includes resource usage data for system-wide. Nominal time of the log entry. where CCC denotes the three-digit cabinet number and MM denotes the two-digit chassis number of the node. Type of node. For example. TheDate TheTime n/a n/a Date of the log entry.

HstId is always 0. HHH = three digit Host Group ID) with each field getting two or three decimal digits of the resulting 9 digit value. to a per-second measurement. for some reason. then VprId will be 65534. the rows are summarized into a single row. For LAN-connected hosts. VprId will be 65535. Data Type INTEGER Invalid Platform HstId n/a Identifies the host. Value is BBMMPPHHH (BB = Bus. INTEGER HstType n/a CHAR(8) Secs n/a SMALLINT CentiSecs n/a INTEGER NominalSecs SummaryFlag n/a n/a SMALLINT CHAR 80 Resource Usage Macros and Tables . In Summary Mode. If the value is ‘S. PP = Port. no PE on this node connects to this channel. If there are multiple PEs on this node connecting to this channel. The chassis number is always 0 for network-connected hosts. see “Summary Mode” on page 83. A specified or nominal number of seconds in the logging period.’ the row is a non-summary row. Possible values are “NETWORK” (LAN-connected host) and “IBMMUX” (channelconnected host). This field is useful when performing data calculations with small elapsed times where the difference between centisecond-based data and whole seconds results in a percentage error. for example. VprId is -1.’ the row is a summary row. VprId is the Gateway vproc ID. In Summary Mode. In Summary Mode. Actual number of seconds in the log period represented by this row. Number of centiseconds in the logging period. If.Chapter 8: ResUsageShst Table Column Name VprId Type of Data n/a Description Identifies the vproc number. If the value is ‘N. VprId should be the vproc id of the owning PE. For channel-connected hosts. Type of host. This value is useful for normalizing the statistics contained in this row. MM = Module Number (or chassis number). Identifies the summarization status of this row. For details.

FLOAT FLOAT FLOAT FLOAT FLOAT FLOAT Resource Usage Macros and Tables 81 . For example. Data Type FLOAT Invalid Platform CollectIntervals n/a The number of gather periods per reporting period. KBs transferred out to the host. messages and KBs. Number of messages written out to the host. In the Collect Buffer and Log Buffer. • a zero value.Chapter 8: ResUsageShst Table Column Name Active Type of Data count Description Controls whether or not the rows will be logged to the ResUsage tables if Active Row Filter Mode is enabled. then none of the data columns in the row have been updated during the logging period. then the rows that have a zero Active field value will not be logged to the ResUsage tables. the value is the number of Gather operations that have been performed during the period. KBs transferred in from the host. then the row contains modified data columns. if Active Row Filter Mode is enabled. Number of blocks written out to the host. Number of messages read in from the host. HostBlockReads HostBlockWrites HostMessageReads HostMessageWrites HostReadKB HostWriteKB count count count count count count Number of blocks read in from the host. STATISTICS COLUMNS HOST CONTROLLER COLUMNS SMALLINT Channel Traffic Columns Identify the traffic between the host and the node in three levels of granularity: blocks. This number can vary from one period to the next. Blocks are made up of some amount of variable sized messages. If Active is set to: • a non-zero value. ReadKB and WriteKB identify the KBs involved in the traffic.

Note: This is for TCHN only. Number of create or replace statement commands. Transactions consist of one or more requests. Statements are subdivided into the various statement types. HostQLenSum count Total number of messages queued for output to the host during each log interval. Number of alter. (Tpump commands cannot be distinguished. Number of FastLoad and MultiLoad utility commands. User Commands Columns Identify the type of commands given to Teradata Database by the user. Note: To calculate the average HostQLen divide the HostQLenSum by the CollectIntervals value to get the HostQLen average value during the logging period. Number of update commands. Number of failures transmitting from the host. Number of insert commands. archive and recovery). Requests consist of one or more statements.Chapter 8: ResUsageShst Table Column Name Type of Data Description Channel Management Columns Data Type Invalid Platform Identify overhead of channel management. Number of select commands. HostWriteFails count Number of failures transmitting to the host. UPDATE and DELETE statements). restore. CmdTransactions CmdRequests CmdAlterStmts CmdCreateStmts CmdDeleteStmts CmdGrantStmts CmdInsertStmts CmdSelectStmts CmdUpdateStmts CmdArchUtilityStmts CmdLoadUtilityStmts count count count count count count count count count count count Number of transaction commands. HostQLenMax HostReadFails max count Maximum number of messages queued in each log interval. and statement. Number of delete commands. dump. and are therefore counted by the INSERT. Number of archival utility commands (for example. Three levels of granularity are given: transaction. Note: This is for Teradata Channel software (TCHN) only. This average is an average of the values recorded at each of the gather periods that occur during the logging period. Number of grant or revoke commands. request. FLOAT FLOAT FLOAT FLOAT FLOAT FLOAT FLOAT FLOAT FLOAT FLOAT FLOAT FLOAT FLOAT FLOAT ALL FLOAT ALL 82 Resource Usage Macros and Tables . modify. Number of request commands. or drop statement commands.

for each log interval as follows: You can determine if a row is in Summary Mode by checking the SummaryFlag column for that row. CmdStmtSuccesses CmdStmtFailures CmdStmtErrors CmdStmtTime count count count count Number of statements that departed normally. Number of statements that departed in error. Note: To calculate the average number of statements in progress. The CollectIntervals value is the number of gather periods per reporting period. summarizing the hosts of that type on that node. one row is written to the database for each type of host (network or channel-connected) on each node in the system. Data Type FLOAT FLOAT Invalid Platform ALL User Command Arrival and Departure Columns Identify the arrival and departure times and status of user commands. CmdStmtsInProgMax CmdStmtsInProgSum max count Maximum number of statements in progress during each log interval. Number of statements that departed in failure or abortion. For more information.Chapter 8: ResUsageShst Table Summary Mode Column Name CmdMiscUtilityStmts CmdOtherStmts Type of Data count count Description Number of miscellaneous utility commands. FLOAT FLOAT FLOAT FLOAT ALL FLOAT FLOAT ALL ALL Summary Mode When Summary Mode is active for the ResUsageShst table. IF the SummaryFlag column value is… ‘S’ ‘N’ THEN the data for that row is being logged… in Summary Mode. The sums of the resident time of each statement in progress during the log period. Resource Usage Macros and Tables 83 . including the successes and failures. Total count of statements in progress during each log interval. Number of other commands. normally. divide this value by the CollectIntervals value. see the CollectIntervals column.

The value represents the COD value in one tenths of a percent. The spare column fields expand to values 00-09. SpareTmon01.0%. Column Name SpareCount[00-09] SpareTrack[00-09] SpareTmon00 Type of Data count track n/a Description Spare counted statistic. so that column names would be SpareCount00. Spare tracked statistic. and so on. so a displayed value of 500 represents a COD value of 50. Note: This value is valid only on SUSE Linux Enterprise Server 10 systems and is a single value for the entire system.Chapter 8: ResUsageShst Table Spare Columns Spare Columns The ResUsageShst table has 30 spare columns (one of which is being used) as shown in the table below. SpareTmon[01-09] count Spare time monitored statistic. The SpareTmon00 field contains the Capacity on Demand (COD) value. SpareTrack04. 84 Resource Usage Macros and Tables .

The Node ID is formatted as CCC-MM. Statistics from this table are collected from the storage devices. Note: Under conditions of heavy system load. Type of Data Invalid Platform Column Name Description HOUSEKEEPING COLUMNS Data Type RELATIONAL PRIMARY INDEX COLUMNS These columns taken together form the nonunique primary index. logical device information. Nominal time of the log entry. TheDate TheTime n/a n/a Date of the log entry. See the Secs and NominalSecs columns. For example. 5600C or 5555H. For example. Note: SMP nodes have a chassis and cabinet number of 1. NodeId n/a Identifies the Node upon which the vproc resides. a node in chassis 9 of cabinet 3 has a node ID of ‘003-09’. the node ID of an SMP node is ‘001-01’. where CCC denotes the three-digit cabinet number and MM denotes the two-digit chassis number of the node. entries may be logged late (typically. but this field will still contain the time value when the entry should have been logged. The following table describes the ResUsageSldv table columns. Note: This table is created as a MULTISET table. For more information see “Relational Primary Index” on page 38. representing the per node system family type. by no more than one or two seconds). Type of node. For example. FLOAT INTEGER DATE FLOAT NodeType n/a CHAR(8) Resource Usage Macros and Tables 85 .CHAPTER 9 ResUsageSldv Table The ResUsageSldv table contains resource usage information for system-wide. MISCELLANEOUS HOUSEKEEPING COLUMNS GmtTime n/a Greenwich Mean Time is not affected by the Daylight Savings Time adjustments that occur twice a year.

This field is useful when performing data calculations with small elapsed times where the difference between centisecond-based data and whole seconds results in a percentage error. the LdvId is derived from the Port number. the CtlId is set to 255. In summary mode. but can be different in three cases: • The first interval after a log rate change • A sample logged late because of load on the system • System clock adjustments affect reported Secs Useful for normalizing the count statistics contained in this row. Channel. target Id. Id. Note: For Linux. to a per-second measurement. The value is either DISK for database disk or SDSK for system disk. for example. The maximum controller ID is 255 decimal. Normally the same as NominalSecs. Data Type INTEGER INTEGER Invalid Platform LdvId n/a Represents the storage device in the Bus System where it resides. path Id. If the device address information is not available. The value is the decimal equivalent of the three digit controller ID in the LdvId. INTEGER NominalSecs n/a SMALLINT 86 Resource Usage Macros and Tables . For Windows. BYTE(4) LdvType Secs n/a n/a Type of logical device. and Lun information of the device. If the controller information is not available. and Lun information of the device. CHAR(4) SMALLINT CentiSecs n/a Number of centiseconds in the logging period. this field contains the device major and minor number. This allows the storage devices to be grouped by CtlId for controller based summarization. the LdvId is derived from the Host. Actual number of seconds in the log period represented by this row. The value in LdvId is -1. its value is set to 255.Chapter 9: ResUsageSldv Table Column Name VprId CtlId Type of Data n/a n/a Description Note: This column is obsolete in normal mode and the value is set to 65535. A specified or nominal number of seconds in the logging period. Represents the controller number.

then the rows that have a zero Active field value will not be logged to the ResUsage tables. STATISTICS COLUMNS LOGICAL DEVICE COLUMNS SMALLINT Input and Output Traffic Columns The following columns represent the number and amount. In the Collect Buffer and Log Buffer. Number of writes issued. then the row contains modified data columns. see “Summary Mode” on page 88. LdvReads LdvWrites LdvReadKB LdvWriteKB LdvReadRespMax count count count count max Number of reads issued. of data read and written to the logical device. The number of KBs (1024) read from the logical device.) Contains the maximum of the total write response times in centiseconds (Note that this is not the maximum of individual write I/O response times.’ the row is a summary row. If the value is ‘S.’ the row is a non-summary row. For details. FLOAT CollectIntervals n/a Number of gather periods per reporting period. the value is the number of Gather operations that have been performed during the period. In Summary Mode.Chapter 9: ResUsageSldv Table Column Name SummaryFlag Type of Data n/a Description Identifies the summarization status of this row. the rows are summarized into a single row. Data Type CHAR Invalid Platform Active count Controls whether or not the rows will be logged to the ResUsage tables if Active Row Filter Mode is enabled. For example. (Note that this is not the maximum of individual read I/O response times. then none of the data columns in the row have been updated during the logging period. This number can vary from one period to the next. If Active is set to: • A non-zero value. if Active Row Filter Mode is enabled. The number of KBs (1024) written to the logical device. in KBs. • A zero value. Contains the maximum of the total read response time in centiseconds.) FLOAT FLOAT FLOAT FLOAT FLOAT ALL LdvWriteRespMax max FLOAT ALL Resource Usage Macros and Tables 87 . If the value is ‘N.

Sum of the average of queued requests at each gather period. ReadActiveTotal WriteActiveTotal count count Total of read I/O active time in centiseconds. LdvWriteRespTot count Linux: Total of individual write response times in centiseconds. FLOAT FLOAT ALL FLOAT FLOAT FLOAT ALL ALL FLOAT ALL Summary Mode When Summary Mode is active for the ResUsageSldv table. Outstanding Requests Columns The following columns represent the number of outstanding operation requests and the amount of time with outstanding requests for the logical device. LdvConcurrentMax max Maximum number of concurrent requests during the log period. QReadLength QWriteLength LdvOutReqSum count count count Number of read operations in queue. Default value is always 0. Total time in centiseconds with (any) outstanding requests. To estimate an average value over the report period. divide LdvOutReqSum by the CollectIntervals column. FLOAT FLOAT ALL ALL FLOAT Windows FLOAT Windows Concurrent Operations Columns The following columns represent the number of concurrent operations performed on the logical device at a time. Total of write I/O active time in centiseconds. The values in this field should be less than or equal to the reported logging period. the following rows are written to the database for each node in the system for each log interval: 88 Resource Usage Macros and Tables . Windows: 0. Windows: 0. LdvOutReqAvg = LdvOutReqSum / CollectIntervals. Number of write operations in queue. LdvOutReqMax LdvOutReqTime max count Maximum value of the LdvOutReqSum field.Chapter 9: ResUsageSldv Table Summary Mode Column Name Type of Data Description Data Type Invalid Platform Response Time Columns The following columns represent the response time to requests given to the logical device. LdvReadRespTot count Linux: Total of individual read response times in centiseconds. Note: Do not use this field for any platform.

Chapter 9: ResUsageSldv Table Spare Columns

• •

One row summarizes the system logical devices One row summarizes Teradata Database logical devices

Also, you can determine if a row is in Summary Mode by checking the SummaryFlag column for that row.
IF the SummaryFlag column value is… ‘S’ ‘N’ THEN the data for that row is being logged… in Summary Mode. normally.

Spare Columns
The ResUsageSldv table has nine spare columns (one of which is being used) as shown in the table below.
Column Name SpareCount[00-02] SpareTrack[00-02] SpareTmon00 Type of Data count track n/a Description Spare counted statistic. Spare tracked statistic. The SpareTmon00 field contains the COD value. The value represents the COD value in one tenths of a percent, so a displayed value of 500 represents a COD value of 50.0%. Note: This value is valid only on SUSE Linux Enterprise Server 10 systems and is a single value for the entire system. SpareTmon[01-02] count Spare time monitored statistic.

The spare column fields expand to values 00-02, so that column names would be SpareCount00, SpareCount01, SpareCount02, SpareTrack00, and so on.

Resource Usage Macros and Tables

89

Chapter 9: ResUsageSldv Table Spare Columns

90

Resource Usage Macros and Tables

CHAPTER 10

ResUsageSpdsk Table

The ResUsageSpdsk table: • • Provides pdisk level statistics. Includes resource usage logs on cylinder I/O, allocation, and migration.

Note: This table is created as a MULTISET table. For more information see “Relational Primary Index” on page 38. The following table describes the ResUsageSpdsk table columns.
Type of Data Invalid Platform

Column Name

Description HOUSEKEEPING COLUMNS

Data Type

RELATIONAL PRIMARY INDEX COLUMNS These columns taken together form the nonunique primary index. TheDate TheTime n/a n/a Date of the log entry. Nominal time of the log entry. Note: Under conditions of heavy system load, entries may be logged late (typically, by no more than one or two seconds), but this field will still contain the time value when the entry should have been logged. See the Secs and NominalSecs columns. NodeId n/a Identifies the Node upon which the pdisk is connected. The Node ID is formatted as CCC-MM, where CCC denotes the threedigit cabinet number and MM denotes the two-digit chassis number of the node. For example, a node in chassis 9 of cabinet 3 has a node ID of ‘003-09’. Note: SMP nodes have a chassis and cabinet number of 1. For example, the node ID of an SMP node is ‘001-01’. MISCELLANEOUS HOUSEKEEPING COLUMNS GmtTime n/a Greenwich Mean Time is not affected by the Daylight Savings Time adjustments that occur twice a year. FLOAT INTEGER DATE FLOAT

Resource Usage Macros and Tables

91

Chapter 10: ResUsageSpdsk Table

Column Name PdiskGlobalId

Type of Data n/a

Description Identifies the pdisk in the system. Each pdisk in the system has a global ID which uniquely identifies the pdisk in the system. If a pdisk is connected to the nodes in a clique, all the nodes in that clique see the same pdisk global ID associated with that pdisk. In Summary Mode, the pdisk global ID is -1.

Data Type INTEGER

Invalid Platform

PdiskType

n/a

Type of pdisk. The pdisk can be one of the following: • DISK: This type of pdisk is a storage device. • FILE: This type of pdisk is a file.

CHAR(4)

PdiskDeviceId

n/a

Identifies the local pdisk device. For DISK pdisk, the pdisk device ID can be one of the following: • Linux: This is the pdisk major/minor number. The major number bit positions are 20-31 and the minor number is in bits 0-19. The format is similar to the one shown below. (MMMM MMMM MMMM mmmm mmmm mmmm mmmm mmmm) • Windows: This is the pdisk physical disk/ partition number. The physical disk number is in the lower 16 bits and the partition number is in the upper 16 bits. The format is similar to the one shown below. (0000 0000 0000 PPPP DDDD DDDD DDDD DDDD) For FILE pdisk, the pdisk device ID is -1. In Summary Mode, the pdisk device ID is -1.

BYTE(4)

NodeType

n/a

Type of node, representing the per node system family type. For example, 5600C or 5555H.

CHAR(8)

92

Resource Usage Macros and Tables

Chapter 10: ResUsageSpdsk Table

Column Name Secs

Type of Data n/a

Description Actual number of seconds in the log period represented by this row. Normally the same as NominalSecs, but can be different in three cases: • The first interval after a log rate change • A sample logged late because of load on the system • System clock adjustments affect reported Secs Useful for normalizing the count statistics contained in this row, for example, to a persecond measurement.

Data Type SMALLINT

Invalid Platform

CentiSecs

n/a

Number of centiseconds in the logging period. This field is useful when performing data calculations with small elapsed times where the difference between centisecondbased data and whole seconds results in a percentage error. A specified or nominal number of seconds in the logging period. Identifies the summarization status of this row. If the value is ‘N,’ the row is a nonsummary row. If the value is ‘S,’ the row is a summary row. In Summary Mode, the rows are summarized into a single row per pdisk type per node. For details, see “Summary Mode” on page 97.

INTEGER

NominalSecs SummaryFlag

n/a n/a

SMALLINT CHAR

Active

count

Controls whether or not the rows will be logged to the ResUsage tables if Active Row Filter Mode is enabled. If Active is set to: • a non-zero value, then the row contains modified data columns. • a zero value, then none of the data columns in the row have been updated during the logging period. For example, if Active Row Filter Mode is enabled, then the rows that have a zero Active field value will not be logged to the ResUsage tables.

FLOAT

Resource Usage Macros and Tables

93

Total of individual write response time in centiseconds. Total of squares of the individual read response time in centiseconds. ReadCnt WriteCnt ReadKB WriteKB ReadRespTot WriteRespTot ReadRespMax WriteRespMax ReadRespSq WriteRespSq ConcurrentReadMax ConcurrentWriteMax ConcurrentMax OutReqTime count count count count count count max max count count max max count count Number of logical device reads. the value is the number of Gather operations that have been performed during the period. Total of squares of the individual write response time in centiseconds. Total of individual read response time in centiseconds. Number of KBs (1024 bytes) written to the logical device. Number of logical device writes. Maximum number of concurrent read I/O requests. Time with outstanding requests (busy time). FLOAT FLOAT FLOAT FLOAT FLOAT FLOAT FLOAT FLOAT FLOAT FLOAT FLOAT FLOAT FLOAT FLOAT ALL 94 Resource Usage Macros and Tables . Maximum number of concurrent I/O requests. Maximum number of individual read response time in centiseconds. In the Collect Buffer and Log Buffer. This number can vary from one period to the next. STATISTICS COLUMNS Data Type SMALLINT Invalid Platform I/O Statistics Columns These columns identify the I/O statistics reported by the Extent Driver.Chapter 10: ResUsageSpdsk Table Column Name CollectIntervals Type of Data n/a Description The number of gather periods per reporting period. Maximum number of individual write response time in centiseconds. in centiseconds. Maximum number of concurrent write I/O requests. Number of KBs (1024 bytes) read from the logical device.

an option sold separately from Teradata Database. The Migrator estimates migration costs. Estimates the total cost (in centiseconds) incurred by migration I/Os completing during the log period. or saved during the log period.Chapter 10: ResUsageSpdsk Table Column Name MigrationBlockedIos Type of Data count Description Number of inputs and outputs that are blocked due to migration request. For more information. of all migration I/Os used. also known internally in the allocator as an extent. This count is for cylinders that were allocated on this device and migrated to a different location within the same device or migrated to a completely different device. Note: This field is for internal use only. Note: Each allocation is for a cylinder size worth of data. Migration write I/O response time. Data Type FLOAT Invalid Platform Allocation Columns These columns identify the allocation statistics reported by the Allocator process of the VSS vproc. Thus the column names begin with Ext for extent. ExtMigrateTotal count Total number of cylinders migrated to a different physical location. incurred. For detail s about these columns. The following formula calculates a ExtMigrateSlower value. Migration read I/O response time. in centiseconds. FLOAT FLOAT ExtMigrateReadRespTot ExtMigrateWriteRespTot ExtMigrateIOTimeCost count count count FLOAT FLOAT FLOAT Resource Usage Macros and Tables 95 . see the ExtMigrateFaster field. ExtMigrateFaster count Number of cylinders migrated to a faster location on a device. Note: These columns are populated and used by Teradata VS. see Teradata Virtual Storage. which is the number of cylinders migrated to slower locations: Migrate Slower = ExMigrateTotal ExMigrateFaster. Migration Columns The following columns identify the number of cylinders that migrated to a different location on a device as well as the time. where cost is the extra time waited by all non-migration I/Os as a result of the migration I/O. Do not use this field unless directed by Teradata Support Center.

response time changes in an un-quantified way. right before a particular log interval. The I/O time savings include the improvement in response time caused by the new data arrangement up to the time horizon. Its permanent improvement remains in effect as long as the workload does not change and newer migrations do not significantly alter the data arrangement.91ms. if. this percentage improvement is permanent. The average IO response time after the log interval should be (100%10%)*10ms = 9ms. The average I/O response time in the new log interval is (100%-1%)*9ms = 8. When the workload changes or new migrations affect data arrangement. ExtMigrateIOTimeImprove is useful because it predicts actual system performance at least for short periods of time and can be used to understand why the migration algorithm is doing what it is doing. Migration then logs an ExtMigrateIOTimeImprove of 1% in the next interval. ExtMigrateIOTimeBenefit does not include the cost of the migration I/Os and is a gross benefit. Note: This field is for internal use only. Data Type FLOAT Invalid Platform ExtMigrateIOTimeImprove count Estimates the percent improvement in average I/O response time due to migrations completing in the log interval. For example. Note: This field is for internal use only. Despite this. The Migrator estimates the migration benefit. the average I/O response time was 10 milliseconds (ms). ExtMigrateIOTimeImprove is only an estimate. not a net benefit. Do not use this field unless directed by Teradata Support Center. then the Migration logs an ExtMigrateIOTimeImprove value of 10% in this interval. Do not use this field unless directed by Teradata Support Center.Chapter 10: ResUsageSpdsk Table Column Name ExtMigrateIOTimeBenefit Type of Data count Description Estimates the total I/O time savings achieved by migrations completing in the log period. In theory. FLOAT 96 Resource Usage Macros and Tables .

Resource Usage Macros and Tables 97 . Column Name SpareCount[00-09] SpareTrack[00-09] SpareTmon00 Type of Data count track n/a Description Spare counted statistic. so that column names would be SpareCount00. rows are summarized into a single row for each pdisk type (for example.0%. SpareTmon[01-09] count Spare time monitored statistic. Spare tracked statistic. The spare column fields expand to values 00–09. IF the SummaryFlag column value is… ‘S’ ‘N’ THEN the data for that row is being logged… in Summary Mode. Note: This value is valid only on SUSE Linux Enterprise Server 10 systems and is a single value for the entire system. The value represents the COD value in one tenths of a percent. normally. so a displayed value of 500 represents a COD value of 50. DISK or FILE) for each node in the system per log interval. SpareTmon05 and so on. Spare Columns The ResUsageSpdsk table has 30 spare columns (one of which is being used) as shown in the table below. The SpareTmon00 field contains the COD value.Chapter 10: ResUsageSpdsk Table Summary Mode Summary Mode When Summary Mode is active for the ResUsageSpdsk table. You can determine if a row is in Summary Mode by checking the SummaryFlag column for that row. SpareTrack02.

Chapter 10: ResUsageSpdsk Table Spare Columns 98 Resource Usage Macros and Tables .

For a complete description of the Priority Scheduler and its components. If table logging is enabled on ResUsageSps. and other detail by Priority Scheduler Allocation Group. TheDate TheTime n/a n/a Date of the log entry. Information carried in the table is organized by: • • • • • Collection date/time Node Vproc Type Performance Group Performance Period/Allocation Group For those using Teradata ASM. PGId. see "Priority Scheduler (schmon)" chapter in Utilities. Type of Data Invalid Platform Column Name Description HOUSEKEEPING COLUMNS Data Type RELATIONAL PRIMARY INDEX COLUMNS These columns taken together form the nonunique primary index. See the Secs and NominalSecs columns. entries may be logged late (typically. and Performance Period ID (VprType. Note: Under conditions of heavy system load. The following table describes the ResUsageSps table columns.CHAPTER 11 ResUsageSps Table The ResUsageSps table contains data by Performance Group from the Priority Scheduler. but this column will still contain the time value when the entry should have been logged. each Workload Definition is the equivalent of one Performance Group in ResUsageSps. The ResUsageSps table carries information that is similar to what is displayed in Priority Scheduler monitor output. a row is written to the database once for every triplet of Vproc Type. by no more than one or two seconds). It allows you to see accumulated CPU. Performance Group ID. DATE FLOAT Resource Usage Macros and Tables 99 . Nominal time of the log entry. PPId) in the system for each log interval. Note: This table is created as a MULTISET table. number of active processes.

AMP. The VprId value is -1. Note: SMP nodes have a chassis and cabinet number of 1. The Node ID is formatted as CCCMM. and the PPId of 1 maps to the values 1 through 7. PPId n/a Identifies the performance period. The PPId is a mapping of the internal performance period value (ranges 0 to 7) to a RSS value (ranges 0 to 1). and MISC). A PPId of 0 maps to the value 0. representing the per node system family type. For example. and PGId set when a PGId uses more than one AGId during a logging period. The PPId column allows RSS to log two rows for each node. 5600C or 5555H. CHAR(4) BYTEINT FLOAT NodeType n/a CHAR(8) VprId n/a INTEGER 100 Resource Usage Macros and Tables .Chapter 11: ResUsageSps Table Column Name NodeId Type of Data n/a Description Identifies the Node upon which the vproc resides. the node ID of an SMP node is ‘001-01’. Data Type INTEGER Invalid Platform MISCELLANEOUS HOUSEKEEPING COLUMNS GmtTime n/a Greenwich Mean Time is not affected by the Daylight Savings Time adjustments that occur twice a year. PE. where CCC denotes the three-digit cabinet number and MM denotes the twodigit chassis number of the node. For example. Multiple Vprocs contribute to each Performance Group task. For example. VprType. See the AGId column for more information. VprType n/a Type of vproc (for example. Note: This column is obsolete. Identifies the vproc number. Type of node. a node in chassis 9 of cabinet 3 has a node ID of ‘003-09’.

Normally the same as NominalSecs. for example. while the value of a Workload Definition ID is not in a specific range (that is. The resulting join table outputs the WD name from WLcdefs. to a persecond measurement.Chapter 11: ResUsageSps Table Column Name PGId Type of Data n/a Description Identifies the Performance Group. INTEGER NominalSecs n/a SMALLINT Resource Usage Macros and Tables 101 .WlcName field. A specified or nominal number of seconds in the logging period.WlcPerfGroupMappings table. The mapping between Performance Group ID and Workload Definition ID can be determined by looking at the Teradata Viewpoint Workload Designer portlet or the TDWM. 0 indicates there is no WDId associated with the PG. but can be different in three cases: • The first interval after a log rate change • A sample logged late because of load on the system • System clock adjustments affect reported Secs Useful for normalizing the count statistics contained in this row.WlcId. The Performance Group ID value ranges from 0 to 250. Use this column to obtain the WD name from the WLcdefs table of the Teradata Dynamic Workload Management database by joining ResUsageSps. This column is useful when performing data calculations with small elapsed times where the difference between centisecond-based data and whole seconds results in a percentage error. Data Type SMALLINT Invalid Platform WDId track Workload Definition ID number. There is a one to one mapping between a Performance Group ID and a Workload Definition ID at any point in time.WDId with WLcdefs. the value is incremented and not reused). FLOAT Secs n/a Actual number of seconds in the log period represented by this row. SMALLINT CentiSecs n/a Number of centiseconds in the logging period.

’ the row is a nonsummary row. Data Type SMALLINT Invalid Platform SummaryFlag n/a Identifies the summarization status of this row. then none of the data columns in the Performance Group ID row have been updated during the logging period.’ the row is a summary row. • A zero value. If Active is set to: • A non-zero value. then the Performance Group ID row contains modified data columns. Controls whether or not the Performance Group ID rows will be logged to the ResUsage tables when Active Row Filter Mode is enabled.Chapter 11: ResUsageSps Table Column Name NCPUs Type of Data n/a Description Number of CPUs on this node. SMALLINT 102 Resource Usage Macros and Tables . In the Collect Buffer and Log Buffer. the value is the number of Gather operations that have been performed during the period. This column is useful for normalizing the CPU utilization column values for the number of CPUs on the node. CHAR Active count FLOAT CollectIntervals n/a The number of gather periods per reporting period. This number can vary from one period to the next. If the value is ‘N. If the value is ‘S. This is especially important in coexistence systems where the number of CPUs can vary across system nodes.

see the "Priority Scheduler (schmon)" chapter in Utilities. Number of logical data blocks read or written by Performance Group. or both.Chapter 11: ResUsageSps Table Column Name Type of Data Description STATISTICS COLUMNS TERADATA ASM COLUMNS Data Type Invalid Platform Priority Scheduler Columns The following columns provide a summary of the Priority Scheduler resource usage statistics. There is one Scheduling Set per session. Note: Allocation Groups with higher relative weights will have quicker access to system resources. RelWgt track Weight of the Allocation Group relative to the active Allocation Groups of the Resource Partition and the active Resource Partitions. Number of Scheduling Sets per PG/PPid combination. Note: A value of 200 is the system Allocation Group. This value cannot be assigned for user work. CPUTime count Milliseconds of CPU time consumed by all tasks that have the same VprType. For more information on the Allocation Group (AG). Number of AWT messages/requests that got assigned AWTs to them on the node. FLOAT FLOAT FLOAT IOBlks NumProcs count track FLOAT FLOAT NumSets track FLOAT NumRequests count FLOAT Resource Usage Macros and Tables 103 . PGId. This value can be any number from 0 to 200. see "Priority Scheduler (schmon)" chapter in Utilities. AGId track Identifies the current Allocation Group for the Performance Group ID that is being reported. For more information on allocation group weights. Number of tasks assigned to the Performance Group at the end of the gather period. and PPId values for a reporting period.

Chapter 11: ResUsageSps Table Column Name QWaitTime Type of Data count Description Total time for all messages delivered in the period (if not delivered then not counted). FLOAT 104 Resource Usage Macros and Tables . The Average number of work requests waiting on the input queue for service = QLength /(CollectIntervals * SpareTrack00). This is the amount of time the AWT was held through sleeps. and so on until it is released. The service time is the elapsed time from the time the message was received to the time the AWT was released. FLOAT ServiceTimeMax max Maximum time in milliseconds that work requests required for service. To calculate an approximate average ServiceTime for each request during this period. This value is derived each gather period from QWaitTime by dividing by the sample period and rounding the value. FLOAT QLength count FLOAT QLengthMax max Maximum number of work requests waiting on the input queue for service. I/O. This column is reported in DBC. divide ServiceTime by NumRequests. Sum of the average number of work requests waiting on the input queue for service. The values from each gather period are then summed together. always divide QLength by the AMPcount to get the desired average Qlength per AMP. To use this column. This value is derived by dividing the number of AMPs to display the maximum per AMP average number of work requests waiting on the input queue for service = QLengthMax / SpareTrack00.ResSpsView as QWaitTimeRequestAvg. CPU. FLOAT ServiceTime count Time in milliseconds that work requests required for service. Data Type FLOAT Invalid Platform QWaitTimeMax max Maximum time in milliseconds that work requests waited on an input queue before being serviced. Divide by NumRequests to obtain the average QwaitTime per Request.

Time in milliseconds CPUs are busy in the AWT executing user execution code. FILE SYSTEM COLUMNS Cylinder Read Columns The following columns represent file system resource usage statistics. Dispatcher. This is the system level time spent on a process. CPUUServAWT count Time in milliseconds CPUs are busy in the AWT executing user service code. This is the user level time spent on a process. This is the user level time spent on a process. Time in milliseconds CPUs are busy executing miscellaneous activities for user service code. This is the system level time spent on a process.Chapter 11: ResUsageSps Table Column Name Type of Data Description Data Type Invalid Platform PROCESS SCHEDULING COLUMNS CPU Utilization Columns The following columns represent CPU activities on the node associated with the AWT. Parser. This is the user level time spent on a process. This is the user level time spent on a process. Time in milliseconds CPUs are busy executing miscellaneous activities for user execution code. Time in milliseconds CPUs are busy in the Dispatcher or Parser executing user service code. or miscellaneous things. The Cylinder Read feature uses these statistics for tracking performance and utilization. This is the system level time spent on a process. FLOAT CPUUServDisp count FLOAT CPUUServPars count FLOAT ALL CPUUServMisc count FLOAT CPUUExecAWT count FLOAT CPUUExecDisp count FLOAT CPUUExecPars count FLOAT ALL CPUUExecMisc count FLOAT Resource Usage Macros and Tables 105 . Time in milliseconds CPUs are busy in the Dispatcher or Parser executing user execution code. Time in milliseconds CPUs are busy in the Parser executing user service code. Time in milliseconds CPUs are busy in the Parser executing user execution code. This is the system level time spent on a process.

A number of calculations can be performed using this column: • Requests issued to FSG = FileFcrRequests . It records the number of attempts to use Cylinder Read independent of whether the request will be issued to FSG or not. The average number of data blocks in a successful Cylinder read can be calculated as: Average data blocks/ Cylinder Read = FileFcrBlocksRead / (FileFcrRequests FileFcrDeniedUser .Chapter 11: ResUsageSps Table Column Name FileFcrRequests Type of Data count Description Total number of requests for the File System to use Cylinder Read. This column is tracked and recorded by the FSG subsystem.FileFcrDeniedUser • Successful Cylinder Reads = FileFcrRequests .FileFcrDeniedKern) FLOAT ALL 106 Resource Usage Macros and Tables . A request can be denied due to insufficient data blocks or because there is insufficient space in the FSG cache. It records the number of requests for adaptive-style Cylinder Reads. This column is tracked and recorded by the File System. FLOAT ALL FileFcrBlocksRead count Number of data blocks read in using Cylinder Read. It records the total number of data blocks read in by successful Cylinder Read operations. Requests can also be denied at both the user and kernel level.FileFcrDeniedUser FileFcrDeniedKern Data Type FLOAT Invalid Platform ALL FileFcrRequestsAdaptive count Number of adaptive requests from File System. This column is tracked and recorded by the File System. Each of these items is counted in other FileFcr ResUsage columns.

Chapter 11: ResUsageSps Table

Column Name
FileFcrBlocksDeniedUser

Type of Data count

Description Number of data blocks in the Cylinder Read requests denied by the File System. This column is tracked and recorded by the File System. It records the number of Cylinder Read attempts that have been denied by the File System. A request can be denied by the File System due to insufficient number of data blocks being requested (for example, the FileFcrDeniedThreshUser column). For information, see the FileFcrDeniedThreshUser column.

Data Type FLOAT

Invalid Platform ALL

FileFcrBlocksDeniedKern

count

Number of data blocks in the Cylinder Read requests denied by the FSG subsystem. This column is tracked and recorded by the FSG subsystem. It records the number of Cylinder Read requests issued to the FSG subsystem which, for any reason, have been denied. A request can be denied due to insufficient data blocks (for example, the FileFcrDeniedThreshKern column) or because there is insufficient space in the FSG cache (for example, the FileFcrDeniedCache column). The FSG subsystem can reject a request containing insufficient data blocks that the File System thought had enough blocks because the FSG subsystem reduces the count by the number of data blocks that are already resident in the cache.

FLOAT

ALL

FileFcrBlocksDeniedCache

count

Number of data blocks in the Cylinder Read requests rejected by the FSG subsystem due to insufficient cache. This column is tracked and recorded by the FSG subsystem. It records the number of data blocks that were part of attempts to use Cylinder read that were denied by the FSG subsystem due to insufficient cache space; therefore, also incremented the FileFcrDeniedCache column.

FLOAT

ALL

Resource Usage Macros and Tables

107

Chapter 11: ResUsageSps Table

Column Name FileFcrBlocksDeniedThreshUser

Type of Data count

Description Number of data blocks in the Cylinder Read requests denied by the File System due to insufficient data blocks. This column is tracked and recorded by the File System. It records the number of Cylinder Read requests which have been denied due to the data block threshold criteria. There is a minimum threshold of data blocks for an individual Cylinder Read request. If the number of data blocks is below this threshold, the overhead of the Cylinder Read operation is considered too large and issuing individual data block reads is considered more efficient. Therefore, the Cylinder Read request is denied.

Data Type FLOAT

Invalid Platform ALL

FileFcrDeniedUser

count

Number of Cylinder Read requests denied by the File System. This column is tracked and recorded by the File System. It records the number of Cylinder Read attempts that have been denied by the File System. A request can be denied by the File System due to insufficient number of data blocks being requested (for example, the FileFcrDeniedThreshUser column). For information, see the FileFcrDeniedThreshUser column description.

FLOAT

ALL

FileFcrDeniedKern

count

Number of Cylinder Read requests denied by the FSG subsystem. This column is tracked and recorded by the FSG subsystem. It records the number of Cylinder Read requests issued to the FSG subsystem which, for any reason, have been denied. A request can be denied due to insufficient data blocks (for example, the FileFcrDeniedThreshKern column) or because there is insufficient space in the FSG cache (for example, the FileFcrDeniedCache column). The FSG subsystem can reject a request containing insufficient data blocks that the File System thought had enough blocks because the FSG subsystem reduces the count by the number of data blocks that are already resident in the cache.

FLOAT

ALL

108

Resource Usage Macros and Tables

Chapter 11: ResUsageSps Table

Column Name FileFcrDeniedCache

Type of Data count

Description Number of Cylinder Read requests denied by FSG due to insufficient cache. This column is tracked and recorded by FSG. It records the number of Cylinder Read requests which have been denied due to insufficient FSG cache space for a cylinders worth of data.

Data Type FLOAT

Invalid Platform ALL

FileFcrDeniedThreshUser

count

Number of Cylinder Read requests denied by the File System due to insufficient data blocks. This column is tracked and recorded by the File System. It records the number of Cylinder Read requests which have been denied due to the data block threshold criteria. There is a minimum threshold of data blocks for an individual Cylinder Read request. If the number of data blocks is below this threshold, the overhead of the Cylinder Read operation is considered too large and issuing individual data block reads is considered more efficient. Therefore, the Cylinder Read request is denied.

FLOAT

ALL

Segment Acquires Columns The following columns identify the total disk memory segments acquired by the file system during the log period. Logical acquires (Acqs) and the logical amount acquired (AcqKB) are described by single entries below, each of which expands into six actual columns, where [seg] is replaced as follows: PDb = Permanent data block disk segments PCi = Permanent cylinder index disk segments SDb = Regular or restartable spool data block disk segments SCi = Regular or restartable spool index disk segments Tjt = Transient journal table APt = Append table or permanent journal table data block or cylinder index disk segments File[seg]Acqs File[seg]AcqKB File[seg]AcqReads File[seg]AcqReadKB count count count count Total number of disk segments acquired. Total KBs acquired by File[seg]Acqs. Number of disk segment acquires that caused a physical read. KBs physically read by File[seg]AcqReads. FLOAT FLOAT FLOAT FLOAT

Resource Usage Macros and Tables

109

Chapter 11: ResUsageSps Table

Column Name

Type of Data

Description

Data Type

Invalid Platform

Data Block Prefetches Columns The following columns identify File Segment Prefetch activities. File segments prefetches are described by single entries below, each of which expands into six actual columns, where [seg] is replaced as follows: PDb = Permanent data block disk segments PCi = Permanent cylinder index disk segments SDb = Regular or restartable spool data block disk segments SCi = Regular or restartable spool index disk segments Tjt = Transient journal table APt = Append table or permanent journal table data block or cylinder index disk segments File[seg]Pres File[seg]PresKB File[seg]PreReads File[seg]PreReadKB count count count count Total number of disk segments prefetched. Total number of KBs prefetched by File[seg]Pres. Total number of disk segment prefetches that caused a physical read. Total number of KBs physically read by File[seg]PreReads. FLOAT FLOAT FLOAT FLOAT

Segments Released Columns The following columns identify the total disk memory segments released by the file system, as well as those segments that are dropped from memory during the log period. When a segment is release, the segment is either: Force out of memory (F) Remains resident in memory (R) Aged out of memory (A), from segments that remain resident Both the number of segments (Rels, Writes, Drps) and the size of the segments (RelKB, WriteKB, DrpKB) are counted. When a segment leaves memory, it must be written to disk only if the segment is dirty, that is, modified (Dy). Otherwise, the clean or unmodified (Cn) segment is simply dropped. (Most spool blocks are simply dropped from a task and put on the age queue. This may happen multiple times. Each of these will be counted as a resident release. If the system is low on memory and the age queue must be processed, this may also result in an age write or age drop. Forced writes are always also counted as either clean resident releases or forced drops, depending on whether age normal or age out now was specified.) Disk segment memory types are described by single entries below, each of which expands into six actual columns, where [seg] is replaced as follows: • • • • • • PDb = Permanent data block disk segments PCi = Permanent cylinder index disk segments SDb = Regular or restartable spool data block disk segments SCi = Regular or restartable spool cylinder index disk segments TJt = Transient journal table or WAL data block or WAL cylinder index APt = Append table or permanent journal table data block or cylinder index disk segments count Number of dirty disk segment resident releases. FLOAT

File[seg]DyRRels

110

Resource Usage Macros and Tables

Chapter 11: ResUsageSps Table

Column Name File[seg]DyRRelKB File[seg]FWrites

Type of Data count count

Description KBs released by File[seg]DyRRels. Number of disk segment forced releases or specific I/O requests causing an immediate physical write. Includes spool data that is aged out immediately and permanent data that is written immediately. KBs written by File[seg]FWrites.

Data Type FLOAT FLOAT

Invalid Platform

File[seg]FWriteKB

count

FLOAT

Cylinder Management Overhead Events Columns The following columns identify the number of times the file system software performed a cylinder management event. The table ResUsageIvpr further breaks down the I/Os associated with these events. See Appendix C: “ResUsageIvpr Table.” FileCylMigrs FileCylAllocs count count Number of cylinder migrations. Number of new cylinders allocated. Note: A new cylinder allocation event implies one logical cylinder index read and one logical cylinder index write. Synchronized Full File Scans Columns The following columns contain statistics relating to synchronized full-file scans. FileSyncScans FileSyncSubtables count track Number of attempts to synchronize a full file scan. Number of subtables scanned by one or more full file scanners who are willing to synchronize scans. Number of tasks involved in full file scans who are willing to synchronize with other scanners. Number of groups of scanners involved in full file scans. A group consists of scanners who are able to use the same read I/O to obtain data from disk. FLOAT FLOAT ALL ALL FLOAT FLOAT ALL ALL

FileSyncScanners

track

FLOAT

ALL

FileSyncGroups

track

FLOAT

ALL

ChnSignal Status Tracking columns The following columns track the chnsignal last done status (or track slowest vproc on the system for processing AMP steps). MsgChnLastDone count The number of last done events that occurred for this vproc. Note: The last AMP to finish an operation may send a last done broadcast message indicating the work is done for this step. This is used in tracking down the slowest AMP in the system. An AMP that has more last done messages than the others could be a bottleneck in the system performance. FLOAT ALL

Resource Usage Macros and Tables

111

Note: Only scratch pages are allocated. see Chapter 7: “ResUsageSawt Table. Current number of AWTs in use during the log period for each work type for the WD (PGid/VprType/PPid triplet). MemKBRes count The amount of memory resident that is specific to virtual processor activities.” Note: The system writes the data to the database once for every triplet of Vproc Type. The Max columns will not be able to be correlated to the ResUsageSawt table Max values in such a direct way since the ResUsageSps Max columns report the Max value of the ResUsageSps table InUse column for the WD and not the Max value of the ResUsageSawt table for all the WDs combined. The ResUsageSps table values should add up to the ResUsageSawt table for columns like WorkTypeInuse. and Performance Period ID (VprType. PPId). Total KBs attributed to SEG memory allocations. FLOAT ALL FlowCtlCnt WorkTypeInuse00 WorkTypeInuse15 count count FLOAT FLOAT ALL 112 Resource Usage Macros and Tables . Performance Group ID.Chapter 11: ResUsageSps Table Column Name Type of Data Description Data Type Invalid Platform Memory Allocation Columns The following columns represent the number and amount of memory allocations. FlowControlled count Number of times this log period that system entered the flow control state from a nonflow controlled state. FLOAT ALL FLOAT FLOAT FLOAT ALL Amp Worker Task Columns The following columns collect and report statistics about the AWTs. Number of AWTs currently in flow control on the work input mailbox. MemAllocs MemAllocKB MemCtxtAllocs count count count Number of successful SEG memory allocations. Number of successful SWAP memory allocations. subdivided into (the only applicable) generic node memory type. For more information about the ResUsageSawt table and columns. The data is reporting the contribution of the respective WD to the column and the values are not the same as the values reported in the ResUsageSawt table. PGId.

PPId). NetBrdReads NetBrdWrites count count Number of broadcast messages input to the vproc. Total KBs of point-to-point messages output to the vproc on behalf of the WD. The true maximum number of inuse AWTs of a WorkType may occur at a different time during the reporting period and not be seen at the end of the gather period and therefore not be reported. Number of broadcast messages output from the vproc. When multiple gather periods occur during the reporting period the value is the maximum of the sampled values. Note: The system writes the data to the database once for every triplet of Vproc Type. Number of point-to-point messages output from the vproc on behalf of the WD. and Performance Period ID (VprType. Maximum number of AWTs in use at one time during the log period for each work type for the WD (PGid/VprType/PPid triplet). Writes) and amount (ReadKB. FLOAT FLOAT FLOAT FLOAT Broadcast Net Traffic Columns The following columns identify the number (Reads. NetPtPReads NetPtPWrites NetPtPReadKB NetPtPWriteKB count count count count Number of point-to-point messages input to the vproc on behalf of the WD. PGId. the WorkTypeMax and WorkTypeInUse columns would report the same value. WriteKB) of input and output messages passing through the Teradata Database nets through broadcast (1:many) methods (Brd) per net. FLOAT FLOAT Resource Usage Macros and Tables 113 . If only a single gather period occurs during the reporting period. This is therefore a maximum sampled value. Writes) and amount (ReadKB. Total KBs of point-to-point messages input to the vproc on behalf of the WD. WriteKB) of input and output messages passing through either Teradata Database net through point-to-point (1:1) methods (PtP). Performance Group ID. NET COLUMNS Data Type FLOAT Invalid Platform Point-to-Point Net Traffic Columns The following columns identify the number (Reads.Chapter 11: ResUsageSps Table Column Name WorkTypeMax00 WorkTypeMax15 Type of Data max Description The value reported is the maximum of the WorkTypeInuse values seen at the end of each gather period during the reporting period.

Write Ahead Logging Columns The following columns identify the log-based file system recovery scheme in which modifications to permanent data are written to a log file. Number of map I/Os completed by the allocator. see Teradata Virtual Storage. FLOAT FLOAT FLOAT ALL ALL ALL I/O Columns These columns identify the I/O statistics reported from the extent driver. Number of migration requests completed by the Node Agent. Number of cylinder free requests received by the allocator. For details about these columns. NodeAgentMigrationsStarted NodeAgentMigrationsDone NodeAgentStatProcessed count count count Number of migration requests started by the Node Agent. the WAL log.Chapter 11: ResUsageSps Table Column Name Type of Data Description Data Type Invalid Platform Allocator Columns The following columns identify the number of requests or I/Os from the Allocator. an option sold separately from Teradata Database. Number of times the file system logically frees a cylinder. AllocatorExtentAllocReqs count count count count Number of cylinder allocation requests received by the allocator. FileWCylAllocs FileWCylFrees count count Number of new WAL cylinders allocated. Number of map I/Os initiated by the allocator. Note: These columns are populated and used by Teradata VS. Associated configuration settings appear in the ctl utility if you purchased Teradata Virtual Storage. FLOAT FLOAT ALL ALL 114 Resource Usage Macros and Tables . Number of statistics buffers processed by the Node Agent. FLOAT FLOAT FLOAT FLOAT ALL ALL ALL ALL AllocatorExtentFreeReqs AllocatorMapIOsStarted AllocatorMapIOsDone Node Agent Columns The following columns identify the migration and buffer processing statistics reported by the Node Agent.

Chapter 11: ResUsageSps Table Column Name Type of Data Description Data Type Invalid Platform Process Blocking and Waiting Columns The following columns count of blocks and wait time in milliseconds where [reason] is replaced with one of the following: • • • • • • SegNoVirtual FsgNIOs SegMDL MonResume NetThrottle Qnl • • • • • FsgRead FsgWrite DBLock Monitor SegLock • FsgLock • Time • FlowControla • CpuLimitb • Misc For example. ProcBlks[reason] count Number of times processes were blocked. Note: ProcBlksDBLock is invalid on all platforms. FLOAT FLOAT Resource Usage Macros and Tables 115 . Note: ProcWaitDBLock is invalid on all platforms. b. FlowControl refers to the delays caused by the flow control conditions. AG Limit. ProcBlksMisc. ProcWait[reason] count Total time processes were blocked pending. CpuLimit refers to delays due to the Priority Scheduler CPU Limits (System Limit. columns can appear as ProcBlksSegNoVirtual. ProcWaitSegNoVirtual. or RP Limit). and so on. The following column definition descriptions can be found in the ResUsageSpma table descriptions of the same names except where noted. a.

see “ResSpsView” on page 159.0: • SpareCount00 = WorkMsgSendDelay. When the task does receive an AWT in a later interval. • SpareCount01 = WorkMsgSendDelayMax. 116 Resource Usage Macros and Tables . the time attributed here will be counted again within QWaitTime of the interval where it was assigned an AWT. This column reports the maximum service time of a single task in a WD that is still running or has finished in the current reporting interval. This column reports service time consumed by a WD during the current reporting interval. This column reports the maximum delay time in milliseconds for messages that are still in the work box. This column is related to the QWaitTime column and represents a running total of delays attributed to the tasks that still have not been assigned an AWT within this interval. • SpareCount03 = WorkMsgReceiveDelayMax. This includes time used during previous intervals for that task. This column reports the longest time in milliseconds seen or still waiting at sample time (if the messages are not sent yet. This column reports the total time in milliseconds for all messages delivered in a period (if the messages are not sent yet. • SpareCount02 = WorkMsgReceiveDelay. If you use the spare columns above. Column Name SpareCount[00-05] Type of Data count Description The following SpareCount columns will be converted to the specified column names in Teradata Database 14. • SpareCount05 = WorkTimeInuseMax. It can be used to calculate the average usage of AWTs during the reporting period. Note: WorkTimeInUse is not the running sum of a WD that exists over multiple intervals. for example: Average AWTs used = WorkTimeInuse/ (Centisecs*10) This value is available in the ResSpsView as AwtUsedAvg. • SpareCount04 = WorkTimeInuse. This column reports the time for all messages not yet delivered at the end of each gather period. then they are not counted). then they are not counted).Chapter 11: ResUsageSps Table Spare Columns Spare Columns The ResUsageSps table has 30 spare columns (several of which are being used) as shown in the table below.

This column reports the number of messages that are delivered to the work box. This column reports the number of messages that are still waiting for AWTs at the end of each gather period. The following SpareCount columns will be converted to the specified column names in Teradata Database 14. SpareTrack[01-09] SpareTmon00 track n/a SpareTmon[01-09] count Spare time monitored statistic. so that column names would be SpareCount00. See “ResSpsView” on page 159 for an example of the view. AMPcount is the number of AMPs on the Node. SpareTrack09. AMPcount is used to divide columns that are reporting data from all the AMPs. so a displayed value of 500 represents a COD value of 50. SpareTrack00 track SpareCount00 will be converted to the column name AMPcount in Teradata Database 14. This allows the ResSpsView view to report the data columns on a per AMP basis. SpareTmon03. Note: This value is valid only on SUSE Linux Enterprise Server 10 systems and is a single value for the entire system. • SpareCount08 = WorkMsgReceiveDelayCnt.Chapter 11: ResUsageSps Table Spare Columns Column Name SpareCount[06 and 09] SpareCount[07-08] Type of Data count count Description Spare counted statistic. The value represents the COD value in one tenths of a percent.0. and so on. The spare column fields expand to values 00–09. The SpareTmon00 column contains the COD value. Spare tracked statistic.0: • SpareCount07 = WorkMsgSendDelayCnt.0%. Resource Usage Macros and Tables 117 .

Chapter 11: ResUsageSps Table Spare Columns 118 Resource Usage Macros and Tables .

TheDate TheTime n/a n/a Date of the log entry. Note: Under conditions of heavy system load. Invalid Platform Column Name Type of Data Description HOUSEKEEPING COLUMNS Data Type RELATIONAL PRIMARY INDEX COLUMNS These columns taken together form the nonunique primary index. To consolidate and summarize the total number of rows written to the database. Note: This table is created as a MULTISET table. See the Secs and NominalSecs columns. Nominal time of the log entry. a node in chassis 9 of cabinet 3 has a node ID of ‘003-09’. Note: SMP nodes have a chassis and cabinet number of 1. but this column will still contain the time value when the entry should have been logged. Includes resource usage logs on cylinder allocation. the node ID of an SMP node is ‘001-01’. For details. a row is written to the database once for every AMP vproc in the system for each log interval. entries may be logged late (typically. see “Summary Mode” on page 124. you can enable Summary Mode. INTEGER DATE FLOAT Resource Usage Macros and Tables 119 . If table logging is enabled on ResUsageSvdsk. The following table describes the ResUsageSvdsk table columns. NodeId n/a Identifies the Node upon which the vproc resides. by no more than one or two seconds). For example. and I/O statistics. For example.CHAPTER 12 ResUsageSvdsk Table The ResUsageSvdsk table: • • Provides AMP-level storage statistics. migration. The Node ID is formatted as CCCMM. where CCC denotes the three-digit cabinet number and MM denotes the two-digit chassis number of the node.

CentiSecs n/a Number of centiseconds in the logging period. If the value is ‘N. In Summary Mode. the rows are summarized into a single row. For details. NodeType Secs n/a n/a Type of node.’ the row is a non-summary row. The AMP vproc ID is numbered upward from 0. The maximum value is 8191. to a persecond measurement. A specified or nominal number of seconds in the logging period. representing the per node system family type. for example.’ the row is a summary row. Identifies the summarization status of this row.Chapter 12: ResUsageSvdsk Table Column Name Type of Data Description Data Type Invalid Platform MISCELLANEOUS HOUSEKEEPING COLUMNS GmtTime n/a Greenwich Mean Time is not affected by the Daylight Savings Time adjustments that occur twice a year. the value of the AMP vproc ID is -1. For example. see “Summary Mode” on page 124. 5600C or 5555H. but can be different in three cases: • The first interval after a log rate change • A sample logged late because of load on the system • System clock adjustments affect reported Secs Useful for normalizing the count statistics contained in this row. Normally the same as NominalSecs. Identifies the AMP vproc. This column is useful when performing data calculations with small elapsed times where the difference between centisecond-based data and whole seconds results in a percentage error. In Summary Mode. INTEGER CHAR(8) SMALLINT FLOAT VprId n/a INTEGER NominalSecs SummaryFlag n/a n/a SMALLINT CHAR 120 Resource Usage Macros and Tables . If the value is ‘S. Actual number of seconds in the log period represented by this row.

STATISTICS COLUMNS I/O Statistics Columns SMALLINT The following columns identify the I/O statistics reported by FSG for each AMP. Data Type FLOAT Invalid Platform CollectIntervals n/a The number of gather periods per reporting period. In the Collect Buffer and Log Buffer. Total of squares of the individual write response time in centiseconds. Number of KBs (1024 bytes) written to the logical device.Chapter 12: ResUsageSvdsk Table Column Name Active Type of Data count Description Controls whether or not the rows will be logged to the ResUsage tables if Active Row Filter Mode is enabled. then none of the data columns in the row have been updated during the logging period. This number can vary from one period to the next. Number of logical device writes. Total of squares of the individual read response time in centiseconds. then the rows that have a zero Active column value will not be logged to the ResUsage tables. Maximum number of individual write response time in centiseconds. if Active Row Filter Mode is enabled. the value is the number of Gather operations that have been performed during the period. Total of individual read response time in centiseconds. Maximum number of individual read response time in centiseconds. • a zero value. If Active is set to: • a non-zero value. ReadCnt WriteCnt ReadKB WriteKB ReadRespTot WriteRespTot ReadRespMax WriteRespMax ReadRespSq WriteRespSq count count count count count count max max count count Number of logical device reads. then the row contains modified data columns. Total of individual write response time in centiseconds. Number of KBs (1024 bytes) read from the logical device. FLOAT FLOAT FLOAT FLOAT FLOAT FLOAT FLOAT FLOAT FLOAT FLOAT Resource Usage Macros and Tables 121 . For example.

in centiseconds. ExtMigrateTotal count Total number of cylinders migrated to a different physical location. Estimates the total cost (in centiseconds) incurred by migration I/Os completing during the log period. where cost is the extra time waited by all non-migration I/Os as a result of the migration I/O. migrations whose gross benefits are positive) for the associated AMP. Note: Each allocation is for a cylinder size worth of data. Migration write I/O response time. which is the number of cylinders migrated to slower locations: Slower Migration = ExtMigrateTotal ExtMigrateFaster Cylinders are migrated to slower locations to make room for hotter cylinders to replace them. Maximum number of concurrent I/O requests. of all migration I/Os used. or saved during the log period. Time with outstanding requests (busy time). incurred. Allocation Columns Data Type FLOAT FLOAT FLOAT FLOAT Invalid Platform These columns identify the allocation statistics reported by the Allocator process of the VSS vproc. For details about these columns. Maximum number of concurrent write I/O requests. FLOAT FLOAT ExtMigrateReadRespTot ExtMigrateWriteRespTot ExtMigrateIOTimeCost count count count FLOAT FLOAT FLOAT 122 Resource Usage Macros and Tables .Chapter 12: ResUsageSvdsk Table Column Name ConcurrentReadMax ConcurrentWriteMax ConcurrentMax OutReqTime Type of Data max max max count Description Maximum number of concurrent read I/O requests. Do not use this column unless directed by Teradata Support. see Teradata Virtual Storage. ExtMigrateFaster count Number of cylinders migrated to faster locations (that is. see the ExtMigrateFaster column. Note: These columns are populated and used by Teradata VS. an option sold separately from Teradata Database. in centiseconds. also known internally in the allocator as an extent. Migration Columns The following columns identify the number of cylinders that migrated to a different location on a device as well as the time. For more information. Migration read I/O response time. The following formula calculates a Slower Migration value. Thus the column names begin with Ext for extent. Note: This column is for internal use only. The Migrator estimates migration costs.

if. then the Migration logs an ExtMigrateIOTimeImprove value of 10% in this interval. When the workload changes or new migrations affect data arrangement. For example. This value does not include the cost of the migration I/Os and is a gross benefit. Migration then logs an ExtMigrateIOTimeImprove of 1% in the next interval. Do not use this column unless directed by Teradata Support. not a net benefit. ExtMigrateIOTimeImprove is useful because it predicts actual system performance at least for short periods of time and can be used to understand why the migration algorithm is doing what it is doing. this percentage improvement is permanent. FLOAT Resource Usage Macros and Tables 123 . Do not use this column unless directed by Teradata Support. ExtMigrateIOTimeImprove is only an estimate. In theory. Despite this. The average I/O response time in the new log interval is (100%-1%)*9ms = 8. Note: This column is for internal use only. response time changes in an unquantifiable way. The average I/O response time after the log interval should be (100%10%)*10ms = 9ms. right before a particular log interval. Its permanent improvement remains in effect as long as the workload does not change and newer migrations do not significantly alter the data arrangement. The Migrator estimates the migration benefit. Data Type FLOAT Invalid Platform ExtMigrateIOTimeImprove count Estimates the percent improvement in average I/O response time due to migrations completing in the log interval. the average I/O response time was 10 milliseconds (ms).Chapter 12: ResUsageSvdsk Table Column Name ExtMigrateIOTimeBenefit Type of Data count Description Estimates the total I/O time savings achieved by migrations completing in the log period.91ms. Note: This column is for internal use only. The I/O time savings include the improvement in response time caused by the new data arrangement up to the time horizon.

so that column names would be SpareCount00. Column Name SpareCount[00-09] SpareTrack[00-09] SpareTmon00 Type of Data count track n/a Description Spare counted statistic. Spare tracked statistic. The spare column fields expand to values 00–09.Chapter 12: ResUsageSvdsk Table Summary Mode Summary Mode When Summary Mode is active for the ResUsageSvdsk table.0%. normally. SpareTrack02. The SpareTmon00 column contains the COD value. Note: This value is valid only on SUSE Linux Enterprise Server 10 systems and is a single value for the entire system. This row summarizes all AMP data in each node per log interval. You can determine if a row is in Summary Mode by checking the SummaryFlag column for that row. 124 Resource Usage Macros and Tables . and so on. Spare Columns The ResUsageSvdsk table has 30 spare columns (one of which is being used) as shown in the table below. so a displayed value of 500 represents a COD value of 50. one row is written to the database for each node in the system. SpareTmon[01-09] count Spare time monitored statistic. or SpareTmon06. IF the SummaryFlag column value is… ‘S’ ‘N’ THEN the data for that row is being logged… in Summary Mode. The value represents the COD value in one tenths of a percent.

TheDate TheTime n/a n/a Date of the log entry. Note: Under conditions of heavy system load. The Node ID is formatted as CCCMM. MISCELLANEOUS HOUSEKEEPING COLUMNS GmtTime n/a Greenwich Mean Time is not affected by the Daylight Savings Time adjustments that occur twice a year. entries might be logged late (typically. where CCC denotes the three-digit cabinet number and MM denotes the two-digit chassis number of the node.CHAPTER 13 ResUsageSvpr Table ResUsageSvpr logical table includes resource usage data for available system-wide. Note: SMP nodes have a chassis and cabinet number of 1. The following table describes the ResUsageSvpr table columns. by no more than one or two seconds). See the Secs and NominalSecs columns. NodeId n/a Identifies the Node upon which the vproc resides. However. Note: This table is created as a MULTISET table. For more information see “Relational Primary Index” on page 38. For example. but this column will still contain the time value when the entry should have been logged. always use the view “ResSvprView” on page 166 to access the data rather than accessing the ResUsageSvpr table directly. FLOAT INTEGER DATE FLOAT Resource Usage Macros and Tables 125 . a node in chassis 9 of cabinet 3 has a node ID of ‘003-09’. the node ID of an SMP node is ‘001-01’. Type of Data Invalid Platform Column Name Description HOUSEKEEPING COLUMNS Data Type RELATIONAL PRIMARY INDEX COLUMNS These columns taken together form the nonunique primary index. For example. virtual processor information. Nominal time of the log entry.

Chapter 13: ResUsageSvpr Table Column Name NodeType VprId Type of Data n/a n/a Description Type of node. 2 = PE. The vproc numbers within each type range are contiguous. For example. • NODE vprocs: numbered upward from 16384. AMP. to a persecond measurement. This column is useful when performing data calculations with small elapsed times where the difference between centisecond-based data and whole seconds results in a percentage error. 5600C or 5555H. INTEGER NominalSecs n/a SMALLINT 126 Resource Usage Macros and Tables . The VprId can be any of the following depending on the type: • AMP vprocs: numbered upward from 0. A specified or nominal number of seconds in the logging period. 0 = NODE. Normally the same as NominalSecs. Identifies the vproc number (non-Summary Mode) or the vproc type (Summary Mode. • RSG vprocs: numbered downward from 9215. RSG. or TVS (see Teradata Virtual Storage). The values can be NODE. 1 = AMP. but can be different in three cases: • The first interval after a log rate change • A sample logged late because of load on the system • System clock adjustments affect reported Secs Useful for normalizing the count statistics contained in this row. 4=RSG. • PE vprocs: numbered downward from 16383. • GTW vprocs: numbered upward from 8192. GTW. 3=GTW. Actual number of seconds in the log period represented by this row. 5=VSS). • VSS vprocs: numbered downward from 10238. representing the per node system family type. for example. CHAR(4) Secs n/a SMALLINT CentiSecs n/a Number of centiseconds in the logging period. Data Type CHAR(8) INTEGER Invalid Platform VprType n/a Type of vproc. PE. Each existing vproc type range should not overlap into the range of another existing vproc type on the system.

Possible values are ‘N’ if the row is a nonsummary row and ‘S’ if it is a summary row.47. then the row contains modified data columns. for example. For example. If Active is set to: • a non-zero value. In the Collect Buffer and Log Buffer. see "Process Scheduling Columns" in the Chapter 5: “ResUsageScpu Table. where [i] expands to the values 00 . the value is the number of Gather operations that have been performed during the period. then none of the data columns in the row have been updated during the logging period. Controls whether or not the rows will be logged to the ResUsage tables if Active Row Filter Mode is enabled. if Active Row Filter Mode is enabled. subdivided into 48 partitions. Data Type SMALLINT Invalid Platform SummaryFlag n/a Identifies the summarization status of this row.” CPUUServPart[i] count Time in centiseconds CPUs are busy in partition i doing user service. Partition 0 is reserved for use by PDE processes. then the rows that have a zero Active column value will not be logged to the ResUsage tables. This is the system level time spent on a process. FLOAT Resource Usage Macros and Tables 127 . This number can vary from one period to the next. CPUUservPart31. STATISTICS COLUMNS PROCESS SCHEDULING COLUMNS SMALLINT CPU Utilization Columns These columns represent CPU activities associated with this virtual processor. • a zero value.Chapter 13: ResUsageSvpr Table Column Name NCPUs Type of Data n/a Description Number of CPUs on this node. For definitions of user service and user execution. See Appendix D: “Partition Assignments” for more information on the other partitions. Each entry below represents 48 columns. This column is useful for normalizing the CPU utilization column values for the number of CPUs on the node. This is especially important in coexistence systems where the number of CPUs can vary across system nodes. CHAR Active count FLOAT CollectIntervals n/a The number of gather periods per reporting period.

This column is tracked and recorded by the File System.FileFcrDeniedKern) FLOAT FLOAT ALL FLOAT 128 Resource Usage Macros and Tables . It records the total number of data blocks read in by successful Cylinder Read operations. FileFcrRequests count Total number of requests for the File System to use Cylinder Read. This is the user level time spent on a process. Data Type FLOAT Invalid Platform Cylinder Read Columns These columns represent file system resource usage statistics. This column is tracked and recorded by the File System. A request can be denied due to insufficient data blocks or because there is insufficient space in the FSG cache. It records the number of requests for adaptive-style Cylinder Reads. Note: This column is not currently used.Chapter 13: ResUsageSvpr Table Column Name CPUUExecPart[i] Type of Data count Description Time in centiseconds CPUs are busy in partition i doing user execution. FileFcrBlocksRead count Number of data blocks read in using Cylinder Read.FileFcrDeniedUser • Successful Cylinder Reads = FileFcrRequests . It records the number of attempts to use Cylinder Read independent of whether the request will be issued to FSG or not. Each of these items is counted in other FileFcr ResUsage columns.FileFcrDeniedUser FileFcrDeniedKern FileFcrRequestsAdaptive count Number of adaptive requests from File System. A number of calculations can be performed using this column: • Requests issued to FSG = FileFcrRequests . The average number of data blocks in a successful Cylinder read can be calculated as: Average data blocks/ Cylinder Read = FileFcrBlocksRead / (FileFcrRequests FileFcrDeniedUser . The Cylinder Read feature uses these statistics for tracking performance and utilization. This column is tracked and recorded by the FSG subsystem. Requests can also be denied at both the user and kernel level.

This column is tracked and recorded by the FSG subsystem. FLOAT Resource Usage Macros and Tables 129 . the FileFcrDeniedCache column). have been denied.Chapter 13: ResUsageSvpr Table Column Name FileFcrDeniedUser Type of Data count Description Number of Cylinder Read requests denied by the File System. This column is tracked and recorded by the File System. the FileFcrDeniedThreshUser column). It records the number of Cylinder Read requests issued to the FSG subsystem which. The FSG subsystem can reject a request containing insufficient data blocks that the File System thought had enough blocks because the FSG subsystem reduces the count by the number of data blocks that are already resident in the cache. A request can be denied due to insufficient data blocks (for example. It records the number of Cylinder Read attempts that have been denied by the File System. FLOAT FileFcrDeniedKern count Number of Cylinder Read requests denied by the FSG subsystem. therefore. FLOAT FileFcrBlocksDeniedKern count Number of data blocks contained in the rejected requests for Cylinder Read. A request can be denied by the File System due to insufficient number of data blocks being requested (for example. It records the number of data blocks that were part of attempts to use Cylinder Read that were denied by the File System. therefore. Data Type FLOAT Invalid Platform FileFcrBlocksDeniedUser count Number of data blocks contained in the rejected requests for Cylinder Read. This column is tracked and recorded by the File System. for any reason. This column is tracked and recorded by the FSG subsystem. also incremented the FileFcrDeniedUser column. see the FileFcrDeniedThreshUser column. also incremented the FileFcrDeniedKern column. For information. the FileFcrDeniedThreshKern column) or because there is insufficient space in the FSG cache (for example. It records the number of data blocks that were part of attempts to use Cylinder Read that were denied by the FSG subsystem.

This column is tracked and recorded by the FSG subsystem. the overhead of the Cylinder Read operation is considered too large and issuing individual data block reads is considered more efficient. also incremented the FileFcrDeniedThreshUser column. also incremented the FileFcrDeniedCache column. the Cylinder Read request is denied. FLOAT FileFcrDeniedThreshUser count Number of Cylinder Read requests denied by the File System due to insufficient data blocks. It records the number of Cylinder Read requests which have been denied due to the data block threshold criteria.Chapter 13: ResUsageSvpr Table Column Name FileFcrDeniedCache Type of Data count Description Number of Cylinder Read requests denied by FSG due to insufficient cache. It records the number of data blocks that were part of attempts to use Cylinder Read that were denied by the File System due to the number of blocks being below the threshold. therefore. therefore. It records the number of Cylinder Read requests which have been denied due to insufficient FSG cache space for a cylinders worth of data. This column is tracked and recorded by FSG. FLOAT FileFcrBlocksDeniedThreshUser count Number of data blocks contained in Cylinder Read requests rejected for threshold by the File System. There is a minimum threshold of data blocks for an individual Cylinder Read request. It records the number of data blocks that were part of attempts to use Cylinder read that were denied by the FSG subsystem due to insufficient cache space. Data Type FLOAT Invalid Platform FileFcrBlocksDeniedCache count Number of data blocks contained in Cylinder Read requests rejected by the FSG subsystem due to insufficient cache. This column is tracked and recorded by the File System. If the number of data blocks is below this threshold. FLOAT 130 Resource Usage Macros and Tables . Therefore. This column is tracked and recorded by the File System.

This is used in tracking down the slowest AMP in the system. also incremented the FileFcrDeniedThreshKern column. Note: The last AMP to finish an operation may send a last done broadcast message indicating the work is done for this step. Therefore. FSG must reevaluate the threshold for a request that the File System considered valid since FSG eliminates any data blocks from the request list that are already resident in the cache. It records the number of Cylinder Read requests which have been denied due to the data block threshold criteria. It records the number of data blocks that were part of attempts to use Cylinder read that were denied by the FSG subsystem due to the number of blocks being below the threshold. This column is tracked and recorded by the FSG subsystem. An AMP that has more last done messages than the others could be a bottleneck in the system performance. This column is tracked and recorded by the FSG subsystem. the Cylinder Read request is denied. FLOAT Resource Usage Macros and Tables 131 . MsgChnLastDone count The number of last done events that occurred for this vproc.Chapter 13: ResUsageSvpr Table Column Name FileFcrDeniedThreshKern Type of Data count Description Number of Cylinder Read requests denied by the FSG subsystem due to insufficient data blocks. Data Type FLOAT Invalid Platform FileFcrBlocksDeniedThreshKern count Number of data blocks contained in Cylinder Read requests rejected for threshold by the FSG subsystem. This could reduce the count that the File System thought was above the threshold to one that is now below. There is a minimum threshold of data blocks for an individual Cylinder Read request. FLOAT ChnSignal Status Tracking Columns These columns track the chnsignal last done status (or track slowest vproc on the system for processing AMP steps). therefore. If the number of data blocks is below this threshold. the overhead of the Cylinder Read operation is considered too large and issuing individual data block reads is considered more efficient.

Chapter 13: ResUsageSvpr Table Column Name Type of Data Description Data Type Invalid Platform Process Pending Counts and Wait Time Columns These columns identify the number of processes blocked on database locks. Note: Only scratch pages will get allocated. each of which expands into six actual columns. Number of times processes blocked for database locks. Total KBs attributed to allocations and sizeincreasing memory alters for disk segments. Teradata Database context amounts are not included since they can be calculated by multiplying the fixed page size by the number of page allocations. MemPCiAlloc. Total time processes were blocked pending database locks. subdivided into segment types. and how long they were blocked. Disk segment memory types are described by single entries below. FLOAT ALL FLOAT FLOAT FLOAT ALL Mem[seg]AllocKB MemCtxtAllocs count count FLOAT FLOAT ALL 132 Resource Usage Macros and Tables . where [seg] is replaced as follows: • PDb = Permanent data block disk segments • PCi = Permanent cylinder index disk segments • SDb = Regular or restartable spool data block disk segments • SCi = Regular or restartable spool cylinder index disk segments • TJt = Transient journal table or WAL data block or WAL cylinder index • APt = Append table or permanent journal table data block or cylinder index disk segments For example MemPDbAlloc. Mem[seg]Allocs count Number of successful memory allocations and size-increasing memory alters on disk segments. MEMORY COLUMNS Memory Allocations Columns These columns represent the number and amount of memory allocations specific to virtual processor activities. and so on. All other task context pages will appear resident at some point soon after component restart. The columns do not include any memory allocations specific to the node the vproc is running under. MemSDb. Number of successful memory allocations and size-increasing memory alters on task context pages. ProcPendDBLock ProcBlksDBLock ProcWaitDBLock count count count Number of processes blocked pending database locks.

FLOAT ALL FLOAT FLOAT FLOAT ALL ALL ALL MemCtxtRes MemPKBResFrz Mem[seg]KBResCU Mem[seg]KBResCA track FLOAT ALL Mem[seg]KBResDA track FLOAT ALL Resource Usage Macros and Tables 133 . ‘Regular’ disk segment memory types are described by single entries below. Note: MemBaseKBResCU tracks of the sum of the data block sizes in the FSG cache (for both general purpose preloads and cylinder read preloads). subdivided into memory types. Disk memory segments can be in one of four states: • Clean (unmodified) and Unaccessed by any process (CU) • Dirty (modified) and Unaccessed (DU) • Clean and Accessed (CA) • Dirty and Accessed (DA). Current KBs resident in memory for regular disk segments that are currently dirty and accessed. where [seg] is replaced as follows: • • • • • • PDb = Permanent data block disk segments PCi = Permanent cylinder index disk segments SDb = Regular or restartable spool data block disk segments SCi = Regular or restartable spool cylinder index disk segments TJt =Transient journal table or WAL data block or WAL cylinder index APt = Append table or permanent journal table data block or cylinder index disk segments track track track Current pages resident in memory for task context segments. Current KBs resident in memory for regular disk segments that are currently clean and accessed. Current KBs resident in memory for frozen segments. Permanent segments for an entire table can be user-locked-in to memory. Mem[seg]KBResDU track Current KBs resident in memory for regular disk segments that are currently dirty and unaccessed. These are called frozen segments (Frz). The columns do not include any memory allocations specific to the node the vproc is running under. each of which expands into six actual columns. Current KBs resident in memory for regular disk segments that are currently clean and not accessed. and no state subdivision is necessary because they cannot be aged or forced out of memory.Chapter 13: ResUsageSvpr Table Column Name Type of Data Description Data Type Invalid Platform Memory Resident Columns These columns represent the amount of memory resident specific to virtual processor activities.

MemCtxtAccesses MemCtxtAccessKB MemCtxtDeaccesses count count count Number of segments accessed. FLOAT FLOAT ALL ALL Swapping Columns These columns identify the effects on disk segments when all processes accessing them get swapped out. MemSwapDrops count Number of disk segments that were dropped from memory because all accessor processes were swapped out.) KBs of segments deaccessed. MemCtxtPageReads MemCtxtPageWrites count count Number of task context pages that were paged in. Number of disk segments that were re-read when they were previously dropped from memory because all accessor processes were swapped out. Number of segments destroyed. (Deaccessed segments remain in memory until paged out through aging. FLOAT FLOAT FLOAT FLOAT MemCtxtDeaccessKB MemCtxtDestroys count count FLOAT FLOAT MemCtxtDestroyKB count FLOAT 134 Resource Usage Macros and Tables . NET COLUMNS Point-to-Point Net Traffic Columns These columns identify the number (Reads. Writes) and amount (ReadKB. Number of segments deaccessed.) KBs of segments destroyed. KBs dropped from memory by MemSwapDrops. Number of task context pages that were paged out. (Destroyed segments are immediately dropped from memory. FLOAT ALL MemSwapDropKB MemSwapReads count count FLOAT FLOAT ALL ALL MemSwapReadKB count FLOAT ALL Task Context Segment Usage Columns These columns identify the usage of task context segments and how they leave memory. WriteKB) of input and output messages passing through either Teradata Database net through point-to-point (1:1) methods (PtP) NetPtPReads count Number of point-to-point messages input to the vproc.Chapter 13: ResUsageSvpr Table Column Name Type of Data Description Data Type Invalid Platform Paging Columns These columns identify paging activities on pages containing Teradata Database context pages only. KBs of segments accessed. KBs re-read from memory by MemSwapReads.

MsgWorkQLenSum count Total number of work requests waiting during each log interval. Total KBs of point-to-point messages input to the vproc. Total KBs of broadcast messages output from the vproc. unlike the Sum count. divide this value by the CollectIntervals value. For more information. Total KBs of point-to-point messages output to the vproc. Therefore the Sum count can be 0 even though the Max count can be non-zero over the log period. Data Type FLOAT FLOAT FLOAT Invalid Platform Broadcast Net Traffic Columns These columns identify the number (Reads. FLOAT FLOAT FLOAT FLOAT Work Mailbox Queue Columns These columns identify the virtual processor work mailbox queue length where requested work awaits the allocation of a process to perform the work.Chapter 13: ResUsageSvpr Table Column Name NetPtPWrites NetPtPReadKB NetPtPWriteKB Type of Data count count count Description Number of point-to-point messages output from the vproc. MsgWorkQLenMax max Maximum number of work requests waiting during each log interval. NetBrdReads NetBrdWrites NetBrdReadKB NetBrdWriteKB count count count count Number of broadcast messages input to the vproc. Total KBs of broadcast messages input to the vproc. Writes) and amount (ReadKB. Number of broadcast messages output from the vproc. The Max count. The CollectIntervals value is the number of gather periods per reporting period. The Sum count tracks the current count at the end of each gather period and sums the counts over the log period. see the CollectIntervals column. FLOAT FLOAT Resource Usage Macros and Tables 135 . Note: To calculate the average work requests waiting. WriteKB) of input and output messages passing through the Teradata Database nets through broadcast (1:many) methods (Brd) per net. tracks the maximum count over the log period.

Total number of database locks held during each log interval. divide this value by the CollectIntervals value. Disk segment memory types are described by single entries below. see the CollectIntervals column. To calculate the average number of requests blocked. To calculate the average number of database locks held. Total number of requests blocked on database locks during each log interval. where [seg] is replaced as follows: • • • • • • PDb = Permanent data block disk segments PCi = Permanent cylinder index disk segments SDb = Regular or restartable spool data block disk segments SCi = Regular or restartable spool cylinder index disk segments TJt = Transient journal table or WAL data block or WAL cylinder index APt = Append table or permanent journal table data block or cylinder index disk segments FLOAT ALL FLOAT FLOAT ALL ALL FLOAT FLOAT FLOAT FLOAT ALL 136 Resource Usage Macros and Tables . divide this value by the CollectIntervals value. Number of times a database lock was blocked. DBLockEnters DBLockBlocks DBLockDeadlocks DBLockBlocksSum count count count count Number of times a database lock was entered for holding. each of which expands into six actual columns. DBLockBlocksMax DBLocksHeldSum max count Maximum number of requests blocked on database locks during each log interval. For more information. Acquires causing physical reads (AcqReads) and the amount read (AcqReadKB) are identified as a subset of logical acquires. DBLocksHeldMax max Maximum number of database locks held during each log interval. The CollectIntervals value is the number of gather periods per reporting period. For more information.Chapter 13: ResUsageSvpr Table Column Name Type of Data Description Data Type Invalid Platform GENERAL CONCURRENCY CONTROL COLUMNS Database Locks Columns These columns identify database locking activities. The CollectIntervals value is the number of gather periods per reporting period. Logical acquires (Acqs) and the logical amount acquired (AcqKB) are identified. FILE SYSTEM COLUMNS Segment Acquires Columns These columns identify the total disk memory segments acquired by the file system during the log period. Number of times a database lock was deadlocked. see the CollectIntervals column.

this column does not include the disk sectors in between the loaded data blocks. Either of these operations are generically called a prefetch. Total number of KBs logically prefetched (either a cylinder read or an individual block read) by File[seg]Pres. this column includes the disk sectors in between the loaded data blocks.Chapter 13: ResUsageSvpr Table Column Name File[seg]Acqs File[seg]AcqKB File[seg]AcqReads File[seg]AcqReadKB Type of Data count count count count Description Total number of disk segments acquired. Data Type FLOAT FLOAT FLOAT FLOAT Invalid Platform Data Block Prefetches Columns These columns identify File Segment Prefetch activities. File segment prefetches are described by single entries below. each of which expands into six actual columns. FLOAT FLOAT File[seg]PresKB count FLOAT File[seg]PreReadKB count FLOAT Resource Usage Macros and Tables 137 . File[seg]PreReads count Total number of disk segment prefetches (either a cylinder read or an individual block read) that caused a logical read. File[seg]Pres count Total number of times a logical data prefetch was performed (either a cylinder read or an individual block read). where [seg] is replaced as follows: • PDb = Permanent data block disk segments • PCi = Permanent cylinder index disk segments • SDb = Regular or restartable spool data block disk segments • SCi = Regular or restartable spool cylinder index disk segments • TJt = Transient journal table or WAL data block or WAL cylinder index • APt = Append table or permanent journal table data block or cylinder index disk segments Note: A prefetch is either a cylinder read operation or an individual block read operation. For cylinder reads. Total KBs acquired by File[seg]Acqs. KBs physically read by File[seg]AcqReads. Total number of KBs physically read by File[seg]PreReads. Number of disk segment acquires that caused a physical read. For cylinder reads.

Since the system also drops the block from memory as soon as the write is complete. KBs written by File[seg]FWrites. it must be written to disk only if the segment is dirty (Dy). there is no point keeping them in the cache. it would be reacquired (whenever the system creates spool data. When a segment leaves memory. both DyAWrites and CnADrps are incremented. where [seg] is replaced as follows: • • • • • • PDb = Permanent data block disk segments PCi = Permanent cylinder index disk segments SDb = Regular or restartable spool data block disk segments SCi = Regular or restartable spool cylinder index disk segments TJt = Transient journal table or WAL data block or WAL cylinder index APt = Append table or permanent journal table data block or cylinder index disk segments count count count Number of dirty disk segment resident releases. KBs released by File[seg]DyRRels. Since these operations do not access a block multiple times. the segment is either: • Forced (F) • Remains resident in memory (R) • Aged out of memory (A). If a block that was examined did not contain any rows that qualify for the modification. When it is reacquired it will no longer be modified. If the segment is modified. Number of disk segment forced releases or specific I/O requests causing an immediate physical write. FLOAT FLOAT FLOAT File[seg]DyRRels File[seg]DyRRelKB File[seg]FWrites File[seg]FWriteKB File[seg]DyAWrites count count FLOAT FLOAT 138 Resource Usage Macros and Tables . However if the block was modified. If the segment is unmodified. Drps) and the size of the segments (RelKB. it must be written out first. from segments that are currently resident Both the number of segments (Rels. Writes. when it is released the system issues the write as part of the release so it is counted as a forced write (FWrite fields). each of which expands into six actual columns. a subsequent step will read it) and released again. as well as those segments that are dropped from memory during the log period. Otherwise. Includes spool data that is aged out immediately and permanent data that is written immediately. Number of dirty disk segments aged out of memory causing a delayed physical write. the segment might be removed from memory. The release will still be counted as a dirty resident release. Each of these will be counted as a dirty resident release (DyRRel fields). if there is contention for room in the FSG cache. If a block survives in the cache. when the block is released it will be dropped from memory immediately (FDrp fields). On the other hand.Chapter 13: ResUsageSvpr Table Column Name Type of Data Description Data Type Invalid Platform Segments Released Columns These columns identify the total disk memory segments released by the file system. When a segment is released. so the subsequent release will be counted as a clean resident release (CnRRel fields). only CnADrps is incremented. unmodified segment is simply dropped. Disk segment memory types are described by single entries below. the clean (Cn). this release is also counted as a forced drop (FDrp fields). This is counted as a dirty age write (DyAWrite fields). modified. Most spool blocks for a small table remain resident when they are created and age there. since the block survived in a modified state. Full table modification operations make one pass on the table and modify each block only once.WriteKB. that is. Because it is a modified segment. that is. DrpKB) are counted.

If the segment is unmodified. If the segment is modified. only CnADrpKB is incremented. Note: CnADrps counts includes the DyAWrites counts.locks blocked = locks with immediate grants. subtract the DyAWriteKB value from the CnADrpKB value. KBs released by File[seg]CnRRels. FLOAT FLOAT FileLockDeadlocks count FLOAT Resource Usage Macros and Tables 139 . only CnADrpKB is incremented. only CnADrps is incremented. Number of disk segment forced releases causing an immediate memory drop. (Total locks . FLOAT FLOAT FLOAT File[seg]FDrpKB File[seg]CnADrps count count FLOAT FLOAT File[seg]CnADrpKB count KBs dropped by File[seg]CnADrps. subtract the DyAWrites value from the CnADrps value. Note: CnADrpKB counts includes the DyAWriteKB counts. Data Type FLOAT Invalid Platform File[seg]CnRRels File[seg]CnRRelKB File[seg]FDrps count count count Number of clean disk segment resident releases. both DyAWriteKB and CnADrpKB are incremented. If the segment is modified. If the segment is unmodified. If the segment is modified. including those implied for segment acquires.Chapter 13: ResUsageSvpr Table Column Name File[seg]DyAWriteKB Type of Data count Description KBs written by File[seg]DyAWrites. Number of clean disk segments aged out of memory. both DyAWrites and CnADrps are incremented. Number of lock requests that were blocked.) Number of deadlocks detected on lock requests. FileLockEnters FileLockBlocks count count Number of lock requests on disk segments. FLOAT Data Segment Lock Requests Columns These columns identify the number of lock requests. To calculate the KBs of clean segments that aged out of memory. both DyAWriteKB and CnADrpKB are incremented. KBs dropped by File[seg]FDrps. blocks. Segments that are never to be part of the memory cache (the age queue) are counted as forced drops. If the segment is unmodified. To calculate the clean segments that aged out of memory. and deadlocks on a disk segment.

Number of cylinder defragments performed. Number of map I/Os initiated by the allocator. FileSyncScans FileSyncSubtables count track Number of attempts to synchronize a full table scan. The table ResUsageIvpr further breaks down the I/Os associated with these events. See Appendix C: “ResUsageIvpr Table. Number of cylinder free requests received by the allocator. FLOAT FLOAT FLOAT 140 Resource Usage Macros and Tables . FLOAT FLOAT Allocation Columns These columns identify the allocation statistics reported by the Allocator. Number of subtables scanned by one or more full table scanners who are willing to synchronize scans.” FileCylAllocs FileCylFrees FileCylMigrs FileMCylPacks FileCylDefrags count count count count count Number of new cylinders allocated. Number of logical or physical cylinders freed. Note: The following columns have been moved from ResUsageIvpr to ResUsageSvpr to avoid costly joins. A group consists of scanners who are able to use the same read I/O to obtain data from disk. FLOAT FLOAT FileSyncScanners track FLOAT FileSyncGroups track FLOAT Write Ahead Logging Columns These columns identify the number of times the file system software performed a cylinder management event associated with the WAL log. Number of cylinder migrations. AllocatorExtentAllocReqs AllocatorExtentFreeReqs AllocatorMapIOsStarted count count count Number of cylinder allocation requests received by the allocator. Number of times the file system logically frees a cylinder. Number of groups of scanners involved in full table scans. FileWCylAllocs FileWCylFrees count count Number of new WAL cylinders allocated.Chapter 13: ResUsageSvpr Table Column Name Type of Data Description Data Type Invalid Platform Cylinder Management Overhead Events Columns These columns identify the number of times the file system software performed a cylinder management event. Number of tasks involved in full table scans who are willing to synchronize with other scanners. Number of MiniCylPacks performed. FLOAT FLOAT FLOAT FLOAT FLOAT Synchronized Full Table Scans Columns These columns contain statistics relating to synchronized full table scans.

Data Type FLOAT Invalid Platform Node Agent Columns These columns identify the migration and buffer processing statistics reported by the Node Agent. see Teradata Virtual Storage. Number of writes while holding a MI no modification (nomod) write lock. FSG I/O Columns These columns identify the I/O statistics reported from the FSG. FLOAT FLOAT FLOAT FLOAT FLOAT FLOAT FLOAT ALL MIWriteTimeMax max FLOAT MISleeps count FLOAT Resource Usage Macros and Tables 141 . MI write lock hold time in milliseconds. Total write time while holding MI no modification (nomod) write lock in milliseconds. Number of times spent waiting to get a MI lock. Master Index Columns MIWriteLocks MIWriteLockTime MIWriteLockTimeMax MIWrites MIWriteTime count count max count count Number of write locks acquired on a MI. For details about these columns. NodeAgentMigrationsStarted NodeAgentMigrationsDone NodeAgentStatProcessed count count count Number of migration requests started by the Node Agent. Maximum write time while holding a MI no modification (nomod) write lock in milliseconds. Maximum time gap between I/O completions in milliseconds on an AMP. FLOAT FLOAT FLOAT Extent Driver I/O Columns These columns identify the I/O statistics reported from the extent driver. Maximum MI write lock hold time in milliseconds. IoRespMax IoGapMax max max Maximum I/O response time in milliseconds on an AMP. Number of migration requests completed by the Node Agent.Chapter 13: ResUsageSvpr Table Column Name AllocatorMapIOsDone Type of Data count Description Number of map I/Os completed by the allocator. Associated configuration settings appear in the ctl utility if you have purchased Teradata VS. Note: These columns are populated and used by Teradata VS. Number of statistics buffers processed by the Node Agent. an option sold separately from Teradata Database.

Chapter 13: ResUsageSvpr Table Summary Mode

Column Name MISleepTime MISleepTimeMax

Type of Data count max

Description Total amount of time spent waiting to get a MI lock in milliseconds. Maximum amount of time spent waiting to get a MI lock.

Data Type FLOAT FLOAT

Invalid Platform

Summary Mode
When Summary Mode is active for the ResUsageSvpr table, one row is written to the database for each type of vproc on each node in the system, summarizing the vprocs of that type on that node, for each log interval. You can determine if a row is in Summary Mode by checking the SummaryFlag column for that row.
IF the SummaryFlag column value is… ‘S’ ‘N’ THEN the data for that row is being logged… in Summary Mode. normally.

142

Resource Usage Macros and Tables

Chapter 13: ResUsageSvpr Table Spare Columns

Spare Columns
The ResUsageSvpr table has 30 spare columns as shown in the table below.
Column Name SpareCount[00-07] Type of Data count Description The following SpareCount columns will be converted to the specified column names in Teradata Database 14.0: • SpareCount00 = DBMergeTried. The number of times the data block being modified was attempted to be merged with some number of adjacent data blocks as part of the modification. • SpareCount01 = DBMergeDone. The number of times the data block being modified has successfully merged with some number of adjacent data blocks as part of the modification. Subtracting DBMergeDone from DBMergeTried will result in the number of data block merge operations that were tried and failed. • SpareCount02 = DBMergeElim. The number of data blocks eliminated due to data block merges. If n data blocks are merged into 1 large block (where n is the number of data blocks), this number is incremented by n-1. • SpareCount03 = DBMergeExtrIO. The number of additional physical I/Os performed in the data block merge process over and above those that would have been done if no data block merges were attempted. This includes any extra physical I/Os that were performed regardless of whether a particular merge attempt succeeded or not. • SpareCount04 = FileACPCylsSkipped. The number of cylinders AutoCylPack scanned at the MI level and decided nothing needed to be done. • SpareCount05 = FileACPCylsMigr. The number of successful migrations performed by AutoCylPack.

Resource Usage Macros and Tables

143

Chapter 13: ResUsageSvpr Table Spare Columns

Column Name SpareCount[00-07] (continued)

Type of Data count

Description • SpareCount06 = FileACPCylsUnFSEOnly. The number of cylinders the background task, Automatic Cylinder Packing (ACP), selected for performing a migration, but could not because of a CI that is marked down, a locking conflict, or a recently modified CI. Instead (except for the down CI case), ACP cleaned-upa the unfree sector entries (UNFSEs) on these cylinders. For more information on ACP, see Utilities or Performance Management. • SpareCount07 = FileACPCylsPostponed. The number of cylinders AutoCylPack selected for performing a migration, but could not be performed at the current time. This can happen due to conflicts with foreground tasks modifying the cylinder at around the same time as AutoCylPack. AutoCylPack, therefore, postpones the work until the next time it scans the MI from the beginning. When AutoCylPack sees this cylinder again, if the cylinder still qualifies, it is selected again for processing. Spare counted statistic. Spare tracked statistic. The SpareTmon00 column contains the COD value. The value represents the COD value in one tenths of a percent, so a displayed value of 500 represents a COD value of 50.0%. Note: This value is valid only on SUSE Linux Enterprise Server 10 systems and is a single value for the entire system.

SpareCount[08-09] SpareTrack[00-09] SpareTmon00

count track n/a

144

Resource Usage Macros and Tables

Chapter 13: ResUsageSvpr Table Spare Columns

Column Name SpareTmon[01-03]

Type of Data count

Description The following SpareCount columns will be converted to the specified column names in Teradata Database 14.0: • SpareTmon01= FSGCacheWaits. The number of times the file system waits for memory to become available in the file segment cache when trying to read data from disk. • SpareTmon02 = FSGCacheWaitTime. The total amount of time the file system waits for memory to become available in the file segment cache when trying to read data from disk. • SpareTmon03 = MAX. The maximum amount of time the file system waits for memory to become available in the file system cache when trying to read data from disk.

SpareTmon[04-09]

count

Spare time monitored statistic.

a. Unfree sectors are those that have no current data block, but cannot yet be used for a new data block. The file system cleans up the unfree sectors, which entails deleting the entries in the CI that say the sectors are unfree and creating new entries that say these sectors are free.

The spare column fields expand to values 00-09, so that column names would be SpareCount08, SpareTrack04, SpareTmon01, and so on.

Resource Usage Macros and Tables

145

Chapter 13: ResUsageSvpr Table Spare Columns

146

Resource Usage Macros and Tables

For the easiest setup. Note: Always use the views to access the data when it is available through the views rather than accessing the ResUsage table directly. For a coexisting system. however. where viewname is the name of the view whose most recent SQL create text is to be reported. If the columns are modified. For details on using the SHOW VIEW statement. the values need to be set up when the system is installed or reconfigured so that each type of node is assigned to a specific group ID. execute SHOW VIEW viewname. the resource usage macros that use these views may not work properly. let the group with the fewest nodes be assigned under the WHEN clause and the group with most nodes be assigned via the ELSE clause in the case statement.CHAPTER 14 Resource Usage Views This chapter provides the definitions of the resource usage views. GroupId. Caution: Do not change or delete columns in these views. To see the view definitions. safely add columns. however. The following views report the table column. Resource Usage Macros and Tables 147 . see SQL Data Definition Language. A homogenous system requires no changes to use this macro because all the nodes will be assigned to group A. You can.

( CPUIoWait* (NodeNormFactor / 100) / NULLIFZERO(NCPUs) ) AS CPUWaitIONorm. Vproc5. ( FileAcqReadKB + FilePreReadKB + 148 Resource Usage Macros and Tables . VprocType4. Vproc1. VprocType7. NodeNormFactor. NCPUs. CollectIntervals. ( CPUUExec * (NodeNormFactor / 100) / NULLIFZERO(NCPUs) ) AS CPUUserNorm. ( FileAcqReads + FilePreReads + FileWrites ) AS DiskSegmentIO. NominalSecs. '001-06') THEN 'A' ELSE 'A' END AS GroupId. ( CPUUServ / NULLIFZERO(NCPUs) ) AS CPUOpSys. Vproc7. Vproc2.ResGeneralInfoView ResGeneralInfoView provides an overview of system operation. Secs. NodeId (FORMAT '999-99') AS NodeId. Note: The data columns in this view will change as the columns in the ResUsageSpma table change. ( FileWrites + MemCtxtPageWrites ) AS LogicalDeviceWrites. VprocType5. '001-05') THEN 'A' WHEN NodeId IN ('001-02'. CASE WHEN NodeId IN ('001-01'. Vproc6. ( (CPUUServ + CPUUExec) * (NodeNormFactor / 100)/ NULLIFZERO(NCPUs) ) AS CPUBusyNorm. TheTime. ( FileAcqReads + FilePreReads + MemTextPageReads + MemCtxtPageReads + MemSwapReads ) AS LogicalDeviceReads. Vproc3. REPLACE VIEW DBC. MemSize. '001-03'. ( CPUUServ * (NodeNormFactor / 100) / NULLIFZERO(NCPUs) ) AS CPUOpSysNorm. VprocType2. ( CPUUExec / NULLIFZERO(NCPUs) ) AS CPUUser. GmtTime. '001-04'. Vproc4. CentiSecs. Reserved. ( CPUIoWait / NULLIFZERO(NCPUs) ) AS CPUWaitIO. /* transformed fields */ ( (CPUUServ + CPUUExec) / NULLIFZERO(NCPUs) ) AS CPUBusy. VprocType1. NodeType.ResGeneralInfoView AS SELECT /* housekeeping fields */ TheDate. ( FileAcqReads + FilePreReads + FileWrites + MemTextPageReads + MemCtxtPageReads + MemCtxtPageWrites + MemSwapReads ) AS LogicalDeviceIO. VprocType6. VprocType3.

ProcBlockedSum/NULLIFZERO(CollectIntervals) (FORMAT '------9') AS ProcBlockedAvg. ( CmdDDLStmts + CmdDeleteStmts + CmdInsertStmts + CmdSelectStmts + CmdUpdateStmts + CmdUtilityStmts + CmdOtherStmts ) AS UserStmtsArriving. VssWriteCnt AS TvsWriteCnt. ( NetMsgPtPReads + NetMsgPtPWrites ) AS NetPtoPIO. NetGroupInUseSum. ProcPendFsgRead. ProcPendFsgWrite. SpareCount01 AS TvsWriteMax. ProcRunningSum. ProcPendNetRead. ProcReadyMax. SpareTmon00 AS COD. ( ProcBlksDBLock + ProcBlksMemAlloc + ProcBlksMisc + ProcBlksMonitor + ProcBlksMonResume + ProcBlksNetThrottle + ProcBlksSegLock + ProcBlksFsgLock + ProcBlksFsgRead + ProcBlksFsgWrite ) (FORMAT '------9') AS ProcBlocks. ( NetMsgBrdWriteKB + NetMsgPtPWriteKB ) AS NetWriteKB. ( NetMsgBrdReadKB + NetMsgPtPReadKB ) AS NetReadKB. ( ProcWaitDBLock + ProcWaitMemAlloc + ProcWaitMisc + ProcWaitMonitor + ProcWaitMonResume + ProcWaitNetThrottle + ProcWaitPageRead + ProcWaitSegLock + ProcWaitFsgLock + ProcWaitFsgRead + ProcWaitFsgWrite ) (FORMAT '------9') AS ProcWaits. ProcReadySum. CmdStmtsInProgCur AS UserStmtsInProgress. 0 AS NetChannelSR. ( ProcBlockedSum + ProcReadySum + ProcRunningSum )/NULLIFZERO(CollectIntervals) AS ProcActiveAvg. ProcBlockedSum. NetChanInUseSum.Resource Usage Macros and Tables 149 /* paging or swapping count times pagesize (= 4K) */ (MemTextPageReads + MemCtxtPageReads + MemSwapReads) * 4 ) AS LogicalDeviceReadKB. ProcPendMemAlloc. VssReadRespTot AS TvsReadRespTot. . /* 14. NetSemInUseSum. ( NetMsgBrdWrites + NetMsgPtPWrites ) AS NetWrites. ( NetMsgBrdReads + NetMsgPtPReads ) AS NetReads. ProcSuspendedSum. ( NetTxCircPtP + NetTxCircBrd ) (FORMAT '------9') AS NetAttempts. /* Spare Field Usage */ SpareCount00 AS TvsReadMax. ProcPendDBLock. ProcPendMonResume. ( FileWriteKB + /* paging or swapping count times pagesize (= 4K) */ MemCtxtPageWrites * 4 ) AS LogicalDeviceWriteKB.0 hdr field Capacity On Demand factor */ /* SPMA table fields */ Active. ProcPendNetThrottle. ( NetCircBackoffs ) (FORMAT '------9') AS NetBackoffs. VssWriteRespTot AS TvsWriteRespTot. ProcPendMonitor. ( MemTextPageReads + MemCtxtPageReads + MemCtxtPageWrites + MemSwapReads ) AS PageOrSwapIO. ( NetMsgBrdReads + NetMsgBrdWrites ) AS NetMultiIO. /* TVS Teradata Virtual Storage fields will be renamed in the table in the 14.0 release */ VssReadCnt AS TvsReadCnt.

ProcWaitMonResume. MemVprAllocs. MemAPtKBRes. ProcWaitTime. 150 Resource Usage Macros and Tables . CPUIoWaitNorm. ProcBlksMsgRead. ProcWaitFsgWrite. ProcBlksFsgWrite. MemSwapDropKB. ProcBlksMonResume. MemSwapDrops. MemCtxtRes. CPUIoWait. ProcWaitMisc. MemTextPageDrops. MemPCiKBRes. ProcBlksNetThrottle. ProcBlksTime. MemCtxtPageWrites. MemProcSwapped. ProcPendQnl. MemFails. ProcWaitMsgRead. ProcWaitPageRead. MemAgings. ProcPendFsgLock. ProcWaitSegLock. MemFreeKB. MemTJtKBRes. CPUUServ. ProcWaitMemAlloc. ProcBlksMisc. CPUIdleNorm. CPUUExec. ProcWaitNetThrottle. ProcBlksSegLock. MemSCiKBRes. MemTextAllocs. MemSDbKBRes. ProcBlksQnl. MemTextRes. ProcBlksMonitor. MemTextPageReads. MemDSysOhRes. MemVprAllocKB.ProcPendSegLock. ProcPendMisc. CPUIdle. ProcBlksFsgLock. CPUUServNorm. CPUUExecNorm. ProcWaitFsgRead. ProcBlksDBLock. ProcWaitQnl. MemTSysOhRes. ProcBlksFsgRead. MemCtxtPageReads. MemPDbKBRes. ProcWaitDBLock. MemSwapReads. ProcBlksMemAlloc. ProcWaitMonitor. ProcWaitFsgLock.

NetTxCircPtP. NetTxConnected. FilePreReads. NetTxKBBrd. MsgBrdWrites. NetHWBackoffs. NetMsgBrdReadKB.Resource Usage Macros and Tables 151 MemSwapReadKB. NetMsgBrdReads. NetTxRouting. NetTxKBPtP. MsgBrdReadKB. HostMessageReads. MsgBrdWriteKB. NetMrgRxRows. NetRxCircPtP. NetMrgTxRows. NetTxCircHPPtP. FilePreReadKB. HostMessageWrites. NetMsgPtPReadKB. . NetRxConnected. FilePreKB. MsgPtPWriteKB. NetGroupInUseMax. NetTxIdle. NetMsgBrdWriteKB. NetRxIdle. MsgPtPReads. FileWriteKB. FilePres. DBLockDeadlocks. HostBlockWrites. FileWrites. MsgPtPReadKB. NetMsgBrdWrites. FileAcqKB. NetTxCircHPBrd. NetMsgPtPReads. FileAcqs. NetRxKBBrd. NetRxCircBrd. NetMrgRxKB. HostReadKB. NetSemInUseMax. HostBlockReads. FileAcqReads. NetTxCircBrd. FileRels. MsgBrdReads. FileAcqReadKB. MsgPtPWrites. FileRelKB. NetMsgPtPWrites. HostWriteKB. DBLockBlocks. NetMrgTxKB. NetMsgPtPWriteKB. NetRxKBPtP. NetSamples. NetCircBackoffs. NetChanInUseMax. NetCircAttempts.

SpareTmon01. CmdDDLStmts. AMPWorkTaskServ. PSQWaitTime.ResCPUUsageByAMPView ( TheDate. NCPUs. COMMENT ON VIEW DBC. SpareTrack03. AMPWorkTaskExec. SpareCount02. SpareCount03. FileLargeDepotBlocks. SpareTmon00. NodeID. NCPUs. SpareTmon02. FileLargeDepotWrites. AMPMiscUserExec. AMPTotalUserExec. Vproc. CmdUpdateStmts. CmdStmtSuccesses. REPLACE VIEW DBC. FileLockDeadlocks. SpareTmon03 FROM ResUsageSpma WITH CHECK OPTION. PSServiceTime. GroupId. SpareCount01. CmdInsertStmts. CmdSelectStmts. NodeId. CmdStmtErrors. SpareTrack02. MsgChnLastDone. AwtFlowCtlCnt. CmdOtherStmts. PSNumRequests. AwtInuseMax. SpareCount00. VprId. AMPTotalUserServ ) AS SELECT TheDate. TheTime. Secs. SpareTrack01. /* GroupId */ . CmdStmtTime. FileSmallDepotWrites. Secs. AMPMiscUserServ. AwtInuse. TheTime. FileLockEnters. CmdStmtsInProgCur. AwtFlowControlled. CmdStmtFailures.FileLockBlocks. 152 Resource Usage Macros and Tables ResCPUUsageByAMPView ResCPUUsageByAMPView describes CPU usage per AMP. SpareTrack00. CmdUtilityStmts. CmdDeleteStmts.ResGeneralInfoView AS 'View of general system information'.

/* AMPMiscUserServ */ ( CPUUServPart01+CPUUServPart02+CPUUServPart03 + CPUUServPart04+CPUUServPart05+CPUUServPart06+CPUUServPart07 + CPUUServPart08+CPUUServPart09+CPUUServPart10 + CPUUServPart12+CPUUServPart13+CPUUServPart14+CPUUServPart15 + CPUUServPart16+CPUUServPart17+CPUUServPart18+CPUUServPart19 + CPUUServPart20+CPUUServPart21+CPUUServPart22+CPUUServPart23 + CPUUServPart24+CPUUServPart25+CPUUServPart26+CPUUServPart27 + CPUUServPart28+CPUUServPart29+CPUUServPart30+CPUUServPart31 + CPUUServPart32+CPUUServPart33+CPUUServPart34+CPUUServPart35 + CPUUServPart36+CPUUServPart37+CPUUServPart38+CPUUServPart39 + CPUUServPart40+CPUUServPart41+CPUUServPart42+CPUUServPart43 + CPUUServPart44+CPUUServPart45+CPUUServPart46+CPUUServPart47). /* AMPTotalUserServ */ (CPUUServPart00+CPUUServPart01+CPUUServPart02+CPUUServPart03 + CPUUServPart04+CPUUServPart05+CPUUServPart06+CPUUServPart07 + CPUUServPart08+CPUUServPart09+CPUUServPart10+CPUUServPart11 + CPUUServPart12+CPUUServPart13+CPUUServPart14+CPUUServPart15 + CPUUServPart16+CPUUServPart17+CPUUServPart18+CPUUServPart19 + CPUUServPart20+CPUUServPart21+CPUUServPart22+CPUUServPart23 + CPUUServPart24+CPUUServPart25+CPUUServPart26+CPUUServPart27 + CPUUServPart28+CPUUServPart29+CPUUServPart30+CPUUServPart31 + CPUUServPart32+CPUUServPart33+CPUUServPart34+CPUUServPart35 + CPUUServPart36+CPUUServPart37+CPUUServPart38+CPUUServPart39 + CPUUServPart40+CPUUServPart41+CPUUServPart42+CPUUServPart43 + CPUUServPart44+CPUUServPart45+CPUUServPart46+CPUUServPart47) FROM ResUsageSvpr WHERE VprType like 'AMP%' WITH CHECK OPTION. .Resource Usage Macros and Tables 153 CASE WHEN NodeId IN ('001-01'. /* AMPWorkTaskServ */ CPUUServPart11. '001-03'. /* AMPWorkTaskExec */ CPUUExecPart11. /* AMPMiscUserExec*/ ( CPUUExecPart01+CPUUExecPart02+CPUUExecPart03 + CPUUExecPart04+CPUUExecPart05+CPUUExecPart06+CPUUExecPart07 + CPUUExecPart08+CPUUExecPart09+CPUUExecPart10 + CPUUExecPart12+CPUUExecPart13+CPUUExecPart14+CPUUExecPart15 + CPUUExecPart16+CPUUExecPart17+CPUUExecPart18+CPUUExecPart19 + CPUUExecPart20+CPUUExecPart21+CPUUExecPart22+CPUUExecPart23 + CPUUExecPart24+CPUUExecPart25+CPUUExecPart26+CPUUExecPart27 + CPUUExecPart28+CPUUExecPart29+CPUUExecPart30+CPUUExecPart31 + CPUUExecPart32+CPUUExecPart33+CPUUExecPart34+CPUUExecPart35 + CPUUExecPart36+CPUUExecPart37+CPUUExecPart38+CPUUExecPart39 + CPUUExecPart40+CPUUExecPart41+CPUUExecPart42+CPUUExecPart43 + CPUUExecPart44+CPUUExecPart45+CPUUExecPart46+CPUUExecPart47). /* AMPTotalUserExec */ (CPUUExecPart00+CPUUExecPart01+CPUUExecPart02+CPUUExecPart03 + CPUUExecPart04+CPUUExecPart05+CPUUExecPart06+CPUUExecPart07 + CPUUExecPart08+CPUUExecPart09+CPUUExecPart10+CPUUExecPart11 + CPUUExecPart12+CPUUExecPart13+CPUUExecPart14+CPUUExecPart15 + CPUUExecPart16+CPUUExecPart17+CPUUExecPart18+CPUUExecPart19 + CPUUExecPart20+CPUUExecPart21+CPUUExecPart22+CPUUExecPart23 + CPUUExecPart24+CPUUExecPart25+CPUUExecPart26+CPUUExecPart27 + CPUUExecPart28+CPUUExecPart29+CPUUExecPart30+CPUUExecPart31 + CPUUExecPart32+CPUUExecPart33+CPUUExecPart34+CPUUExecPart35 + CPUUExecPart36+CPUUExecPart37+CPUUExecPart38+CPUUExecPart39 + CPUUExecPart40+CPUUExecPart41+CPUUExecPart42+CPUUExecPart43 + CPUUExecPart44+CPUUExecPart45+CPUUExecPart46+CPUUExecPart47). '001-04'. '001-05') THEN 'A ' WHEN NodeId IN ('001-02'. '001-06') THEN 'A' ELSE 'A' END.

/* PEMiscUserServ */ ( CPUUServPart01+CPUUServPart02+CPUUServPart03 + CPUUServPart04+CPUUServPart05+CPUUServPart06+CPUUServPart07 + CPUUServPart08+CPUUServPart09+CPUUServPart10+CPUUServPart11 + CPUUServPart15 + CPUUServPart16+CPUUServPart17+CPUUServPart18+CPUUServPart19 + CPUUServPart20+CPUUServPart21+CPUUServPart22+CPUUServPart23 + CPUUServPart24+CPUUServPart25+CPUUServPart26+CPUUServPart27 + CPUUServPart28+CPUUServPart29+CPUUServPart30+CPUUServPart31 + CPUUServPart32+CPUUServPart33+CPUUServPart34+CPUUServPart35 + CPUUServPart36+CPUUServPart37+CPUUServPart38+CPUUServPart39 + CPUUServPart40+CPUUServPart41+CPUUServPart42+CPUUServPart43 + CPUUServPart44+CPUUServPart45+CPUUServPart46+CPUUServPart47). NCPUs. '001-05') THEN 'A ' WHEN NodeId IN ('001-02'. /* PEMiscUserExec */ ( CPUUExecPart01+CPUUExecPart02+CPUUExecPart03 + CPUUExecPart04+CPUUExecPart05+CPUUExecPart06+CPUUExecPart07 + CPUUExecPart08+CPUUExecPart09+CPUUExecPart10+CPUUExecPart11 + CPUUExecPart15 + CPUUExecPart16+CPUUExecPart17+CPUUExecPart18+CPUUExecPart19 + CPUUExecPart20+CPUUExecPart21+CPUUExecPart22+CPUUExecPart23 + CPUUExecPart24+CPUUExecPart25+CPUUExecPart26+CPUUExecPart27 + CPUUExecPart28+CPUUExecPart29+CPUUExecPart30+CPUUExecPart31 + CPUUExecPart32+CPUUExecPart33+CPUUExecPart34+CPUUExecPart35 + CPUUExecPart36+CPUUExecPart37+CPUUExecPart38+CPUUExecPart39 + CPUUExecPart40+CPUUExecPart41+CPUUExecPart42+CPUUExecPart43 + CPUUExecPart44+CPUUExecPart45+CPUUExecPart46+CPUUExecPart47). /* PEDispServ */ CPUUServPart13. /* PEDispExec */ CPUUExecPart13. Vproc. '001-06') THEN 'A' ELSE 'A' END. /* PEParsExec */ CPUUExecPart14. /* PESessExec */ CPUUExecPart12. PETotalUserServ ) AS SELECT TheDate. REPLACE VIEW DBC. PEDispExec. NodeId. NodeID(FORMAT'999-99'). '001-03'. TheTime. PEMiscUserServ. PEMiscUserExec. PEDispServ. /* PEParsServ */ CPUUServPart14. VprId. Secs. PESessServ. PESessExec. Secs. PEParsExec. NCPUs. /* PESessServ */ CPUUServPart12.ResCPUUsageByPEView ( TheDate. /* GroupId */ CASE WHEN NodeId IN ('001-01'. '001-04'. GroupId. TheTime. PEParsServ. /* PETotalUserExec */ (CPUUExecPart00+CPUUExecPart01+CPUUExecPart02+CPUUExecPart03 + CPUUExecPart04+CPUUExecPart05+CPUUExecPart06+CPUUExecPart07 + CPUUExecPart08+CPUUExecPart09+CPUUExecPart10+CPUUExecPart11 + CPUUExecPart12+CPUUExecPart13+CPUUExecPart14+CPUUExecPart15 + 154 Resource Usage Macros and Tables . PETotalUserExec.ResCPUUsageByPEView ResCPUUsageByPEView describes CPU usage by each PE.

REPLACE VIEW DBC. NodeId (FORMAT '999-99') AS NodeId. TheTime. /* transformed fields */ MailBoxDepth/CollectIntervals . Reserved.Resource Usage Macros and Tables 155 CPUUExecPart16+CPUUExecPart17+CPUUExecPart18+CPUUExecPart19 + CPUUExecPart20+CPUUExecPart21+CPUUExecPart22+CPUUExecPart23 + CPUUExecPart24+CPUUExecPart25+CPUUExecPart26+CPUUExecPart27 + CPUUExecPart28+CPUUExecPart29+CPUUExecPart30+CPUUExecPart31 + CPUUExecPart32+CPUUExecPart33+CPUUExecPart34+CPUUExecPart35 + CPUUExecPart36+CPUUExecPart37+CPUUExecPart38+CPUUExecPart39 + CPUUExecPart40+CPUUExecPart41+CPUUExecPart42+CPUUExecPart43 + CPUUExecPart44+CPUUExecPart45+CPUUExecPart46+CPUUExecPart47). NominalSecs. Secs. '001-04'. '001-05') THEN 'A' WHEN NodeId IN ('001-02'. /* PETotalUserServ */ (CPUUServPart00+CPUUServPart01+CPUUServPart02+CPUUServPart03 + CPUUServPart04+CPUUServPart05+CPUUServPart06+CPUUServPart07 + CPUUServPart08+CPUUServPart09+CPUUServPart10+CPUUServPart11 + CPUUServPart12+CPUUServPart13+CPUUServPart14+CPUUServPart15 + CPUUServPart16+CPUUServPart17+CPUUServPart18+CPUUServPart19 + CPUUServPart20+CPUUServPart21+CPUUServPart22+CPUUServPart23 + CPUUServPart24+CPUUServPart25+CPUUServPart26+CPUUServPart27 + CPUUServPart28+CPUUServPart29+CPUUServPart30+CPUUServPart31 + CPUUServPart32+CPUUServPart33+CPUUServPart34+CPUUServPart35 + CPUUServPart36+CPUUServPart37+CPUUServPart38+CPUUServPart39 + CPUUServPart40+CPUUServPart41+CPUUServPart42+CPUUServPart43 + CPUUServPart44+CPUUServPart45+CPUUServPart46+CPUUServPart47) FROM ResUsageSvpr WHERE VprType like 'PE%' WITH CHECK OPTION. /* Capacity On Demand */ AS MailBoxDepthAvg. SummaryFlag. ResSawtView ResSawtView is based on the ResUsageSawt table. '001-03'. CentiSecs. NodeType. AvailableMin. GmtTime. CASE /* Coexistence reporting support */ WHEN NodeId IN ('001-01'. CollectIntervals. /* Spare Field usage */ SpareCount02 AS SpareCount03 AS SpareTmon00 AS Available. '001-06') THEN 'A' ELSE 'A' END AS GroupId.ResSawtView AS SELECT /* housekeeping fields */ TheDate. VprId. COD.

WorkTypeInuse04. WorkTypeInuse08. WorkTypeInuse11. MailBoxDepth. WorkTypeMax09. FlowCtlTime. WorkTypeMax02. WorkTypeInuse11Avg. WorkTypeMax00. WorkTypeInuse05Avg. WorkTypeMax12. WorkTypeMax13. WorkTypeInuse09Avg. FlowControlled. WorkTypeMax05. WorkTypeInuse02Avg. WorkTypeInuse05. WorkTypeInuse07Avg. WorkTypeInuse13Avg. WorkTypeMax03. WorkTypeInuse06Avg. WorkTypeMax10. WorkTypeInuse06. WorkTypeInuse03. WorkTypeInuse00Avg. WorkTypeInuse15.156 Resource Usage Macros and Tables ( WorkTypeInuse00 WorkTypeInuse04 WorkTypeInuse08 WorkTypeInuse12 + + + + WorkTypeInuse01 WorkTypeInuse05 WorkTypeInuse09 WorkTypeInuse13 AS AS AS AS AS AS AS AS AS AS AS AS AS AS AS AS AS + WorkTypeInuse02 + WorkTypeInuse06 + WorkTypeInuse10 + WorkTypeInuse14 WorkTypeInuseAvg. InuseMax. WorkTypeMax01. WorkTypeInuse10. WorkTypeInuse09. FlowCtlCnt. WorkTypeMax15. WorkTypeInuse14. WorkTypeInuse08Avg. WorkTypeInuse12. WorkTypeInuse04Avg. WorkTypeMax11. WorkTypeInuse14Avg. WorkTypeInuse01. WorkTypeInuse15Avg. WorkTypeMax07. WorkTypeMax08. WorkTypeMax04. WorkTypeInuse03Avg. + + + + WorkTypeInuse03 WorkTypeInuse07 WorkTypeInuse11 WorkTypeInuse15 + + + ) / CollectIntervals WorkTypeInuse00/CollectIntervals WorkTypeInuse01/CollectIntervals WorkTypeInuse02/CollectIntervals WorkTypeInuse03/CollectIntervals WorkTypeInuse04/CollectIntervals WorkTypeInuse05/CollectIntervals WorkTypeInuse06/CollectIntervals WorkTypeInuse07/CollectIntervals WorkTypeInuse08/CollectIntervals WorkTypeInuse09/CollectIntervals WorkTypeInuse10/CollectIntervals WorkTypeInuse11/CollectIntervals WorkTypeInuse12/CollectIntervals WorkTypeInuse13/CollectIntervals WorkTypeInuse14/CollectIntervals WorkTypeInuse15/CollectIntervals /* Remaining table fields */ Active. . WorkTypeMax06. WorkTypeInuse00. WorkTypeInuse02. WorkTypeInuse07. WorkTypeInuse13. WorkTypeInuse10Avg. WorkTypeInuse12Avg. WorkTypeMax14. WorkTypeInuse01Avg.

HostWriteKB. ResShstGroupView ResShstGroupView is based on the ResUsageShst table. SpareTrack09. '001-04'.ResSawtView AS 'View of Sawt table data'. SpareTrack06. TheTime. SpareTrack00. HstId. SpareTmon05. Secs. SpareCount05. SpareTrack04. NodeId. NominalSecs. HostMessageWrites. SpareCount07. HostWriteFails ) AS SELECT TheDate. SpareCount03. . HostReadKB. COMMENT ON VIEW DBC.Resource Usage Macros and Tables 157 SpareCount00. GroupId. SpareTrack07. SpareTrack08. HostBlockWrites. CollectIntervals. SpareTrack02. SpareCount08. SpareTmon02. SpareCount09. HstType. /* GroupId */ CASE WHEN NodeId IN ('001-01'. SpareTrack05. HostQLenSum. SpareCount04. HstId. NominalSecs. NodeId. HostBlockReads. '001-05') THEN 'A ' WHEN NodeId IN ('001-02'. '001-06') THEN 'A' ELSE 'A' END. SpareTmon08. SpareCount06. SpareTmon09 FROM ResUsageSawt WITH CHECK OPTION. SpareCount01. SpareTrack03. HstType. SpareTmon03. SpareTrack01. SpareTmon04. HostQLenMax. SpareTmon01. VprId. SpareTmon07. SpareCount02. HostReadFails. TheTime. Secs. '001-03'. HostMessageReads. VprId. SpareTmon00. REPLACE VIEW DBC. SpareTmon06.ResShstGroupView ( TheDate.

Secs. LdvOutReqSum. LdvWriteRespMax. HostQLenSum. NominalSecs. HostWriteKB.ResSldvGroupView ( TheDate. NodeId. LdvOutReqMax. HostWriteFails FROM ResUsageShst WITH CHECK OPTION. LdvReadRespMax. LdvOutReqTime ) AS SELECT TheDate. TheTime. LdvWriteKB. LdvWriteRespMax. /* GroupId */ CASE WHEN NodeId IN ('001-01'. TheTime. NodeId. HostMessageReads. LdvReadRespTot. 158 Resource Usage Macros and Tables ResSldvGroupView ResSldvGroupView is based on the ResUsageSldv table. LdvType. LdvWriteKB. LdvReadKB. LdvId. '001-06') THEN 'A' ELSE 'A' END. LdvReadRespTot. LdvWrites. LdvOutReqMax. HostMessageWrites. LdvReads. LdvOutReqTime FROM ResUsageSldv WITH CHECK OPTION. '001-03'. . LdvReadKB. NominalSecs. GroupId. Secs. VprId. HostReadFails. HostBlockWrites. HostBlockReads. LdvWriteRespTot.CollectIntervals. CollectIntervals. LdvConcurrentMax. LdvId. LdvReadRespMax. '001-05') THEN 'A' WHEN NodeId IN ('001-02'. HostReadKB. HostQLenMax. REPLACE VIEW DBC. LdvOutReqSum. VprId. LdvWrites. LdvType. LdvReads. CollectIntervals. '001-04'. LdvWriteRespTot. LdvConcurrentMax.

'001-06') THEN 'A' ELSE 'A' END AS GroupId. SpareCount05 AS WorkTimeInuseMax. CentiSecs. '001-05') THEN 'A' WHEN NodeId IN ('001-02'. NominalSecs. * AwtReleases reports the number of AWTs released (completed requests) * while NumRequests reports number that arrived. */ /* Total delay time of messages successfully sent to the workbox */ SpareCount00 AS WorkMsgSendDelay. SpareCount03 AS WorkMsgReceiveDelayMax. PGid. /* Average number of AWTs used based on WorkTimeInuse */ SpareCount04/(Centisecs*10) AS AwtUsedAvg. REPLACE VIEW DBC. GmtTime. CASE /* Coexistence reporting support */ WHEN NodeId IN ('001-01'. TheTime. CollectIntervals. SpareCount04 AS WorkTimeInuse. Secs. /* Longest delay of messages successfully sent to the workbox */ SpareCount01 AS WorkMsgSendDelayMax. PPid. VprType. /* Spare Field usage */ /* * WorkTimeInuse reports the service time consummed by a WD during the current * reporting interval. '001-04'. This includes time used during previous intervals for * that task. NCPUs. This is not the running sum of a WD that * exists over multiple intervals. VprId.ResSpsView AS SELECT /* housekeeping fields */ TheDate.*/ /* So that would be SpareCount02 */ SpareCount02 AS WorkMsgReceiveDelay. show view ResSpsView. * ServiceTimeMax reports the maximum time it took a single task in a WD * to be serviced that has completed in this reporting interval. * Average AWTs used = WorkTimeInuse/(Centisecs*10) * WorkTimeInuseMax reports the maximum service time of a single task in a WD * that is still running or has finished in the current reporting * interval. '001-03'. Resource Usage Macros and Tables 159 . SummaryFlag. NodeType. This value may be much larger than the reporting period.ResSpsView ResSpsView is based on the ResUsageSps table. * ServiceTime reports the time it took a WD to be serviced and is * reported when the AWT is released (task is done). /* be the total delay time of messages still waiting for AWTs. NodeId (FORMAT '999-99') AS NodeId.

(FilePDbAcqReads + FilePDbPreReads) AS PhysicalReadPerm. (NetPtPWrites + NetBrdWrites) AS NetWrites. (CPUUExecAWT + CPUUExecDisp + CPUUExecMisc) AS CPUUExec. FilePDbFWrites AS PhysicalWritePerm. averaged per AMP */ (WorkTypeInuse00 + WorkTypeInuse01 + WorkTypeInuse02 + WorkTypeInuse03 + WorkTypeInuse04 + WorkTypeInuse05 + WorkTypeInuse06 + WorkTypeInuse07 + WorkTypeInuse08 + WorkTypeInuse09 + WorkTypeInuse10 + WorkTypeInuse11 + WorkTypeInuse12 + WorkTypeInuse13 + WorkTypeInuse14 + WorkTypeInuse15) / ( SpareTrack00 * CollectIntervals ) AS WorkTypeInuseAmp. /* Avg delay for msgs delivered in period per Request */ WorkMsgSendDelay / NULLIFZERO(WorkMsgSendDelayCnt) AS WorkMsgSendDelayRequestAvg.0 */ SpareTmon00 AS COD. MAX(SUM(AWT[AMP])) */ (WorkTypeMax00 + WorkTypeMax01 + WorkTypeMax02 + WorkTypeMax03 + WorkTypeMax04 + WorkTypeMax05 + WorkTypeMax06 + WorkTypeMax07 + WorkTypeMax08 + WorkTypeMax09 + WorkTypeMax10 + WorkTypeMax11 + WorkTypeMax12 + WorkTypeMax13 + WorkTypeMax14 + WorkTypeMax15) AS WorkTypeInuseMax. /* transformed fields */ (CpuTime * 10)/(CentiSecs * NCPUs) AS CpuPct. /* Max of (WorkTypesInuse for all AMPs (NOT per AMP) */ /* Can NOT divide by AMPs. (NetPtPReads + NetBrdReads) AS NetReads. QLength / ( SpareTrack00 * CollectIntervals ) AS QLengthAmpAvg. housekeeping field in 14. /* report average service time per request */ ServiceTime / NULLIFZERO(AwtReleases) AS ServiceTimeRequestAvg. 160 Resource Usage Macros and Tables /* report average wait time per request for messages */ /* already delivered to AWTs */ QWaitTime / NULLIFZERO(NumRequests) AS QWaitTimeRequestAvg. (FilePCiAcqReads + FileSDbAcqReads + FileSCiAcqReads + FileTJtAcqReads + FileAPtAcqReads + FilePCiPreReads + FileSDbPreReads + FileSCiPreReads + FileTJtPreReads + FileAPtPreReads) AS PhysicalReadOther. /* number of messages accumulating the send side delay time */ SpareCount07 AS WorkMsgSendDelayCnt.1 AwtReleases */ SpareCount06 AS AwtReleases. FilePDbDyRRels AS LogicalWritePerm./* 13. (FilePCiAcqs + FileSDbAcqs + FileSCiAcqs + FileTJtAcqs + FileAPtAcqs + FilePCiPres + FileSDbPres + FileSCiPres + FileTJtPres + FileAPtPres) AS LogicalReadOther. (FilePCiDyRRels + FileSDbDyRRels + FileSCiDyRRels + FileTJtDyRRels + FileAPtDyRRels) . /* requested by Anita Richards to match WorkMsgSendDelayRequestAvg */ /* avg delay per request on receive side for messages not */ /* yet delivered to AWTs */ WorkMsgReceiveDelay / NULLIFZERO(WorkMsgReceiveDelayCnt) AS WorkMsgReceiveDelayRequestAvg. FilePDbFWriteKB AS PhysicalWritePermKB. /* housekeeping field in 14. SpareTrack00 AS AMPcount. (FilePDbAcqs + FilePDbPres) AS LogicalReadPerm. (ProcWaitFsgRead + ProcWaitFsgWrite + ProcWaitFsgNIOs) AS IODelayTime. (ProcBlksFsgRead + ProcBlksFsgWrite + ProcBlksFsgNIOs) AS IODelay. /* All WorkTypes in use. (CPUUServAWT + CPUUServDisp + CPUUServMisc) AS CPUUServ. /* number of messages accumulating the receive side delay time */ SpareCount08 AS WorkMsgReceiveDelayCnt.0 */ /* Capacity On Demand factor.

FileSyncSubtables. FileFcrDeniedCache. (ProcBlksSegNoVirtual + ProcBlksSegMDL + ProcBlksSegLock) AS ProcBlksSeg. FileCylAllocs. FilePDbDyRRelKB AS LogicalWritePermKB. FileWCylAllocs. FileSyncGroups. MsgChnLastDone. FileWCylFrees. CPUUExecPars. ( FileSDbDyRRelKB + FilePCiDyRRelKB + FileSCiDyRRelKB + FileTJtDyRRelKB + FileAPtDyRRelKB) AS LogicalWriteOtherKB. (ProcBlksMisc + ProcBlksNetThrottle + ProcBlksQnl +ProcBlksTime + ProcBlksFlowControl) AS ProcBlksMisc. FileFcrBlocksDeniedKern. FileFcrBlocksDeniedUser. + FileSCiFWrites + FileTJtFWrites + FileAPtFWrites) AS PhysicalWriteOther. FileFcrBlocksRead. FileSyncScanners. FileSyncScans. AllocatorMapIOsDone. + + FileTJtAcqKB + FileAPtAcqKB + + FileTJtPresKB + FileAPtPresKB) AS LogicalReadOtherKB. FileFcrBlocksDeniedCache. /* ProcBlksMonResume = covered by DBLocks */ (ProcWaitMisc + ProcWaitMonResume + ProcWaitNetThrottle +ProcWaitQnl + ProcWaitTime + ProcWaitFlowControl) AS ProcWaitMisc. AllocatorMapIOsStarted. obsolete Sps table fields removed in 14. + + + + FilePDbPresKB FileSDbPresKB FileSCiAcqKB FileSCiPresKB 161 . FileFcrDeniedKern. FileFcrRequests. FileFcrRequestsAdaptive.0 CPUUServPars. FileFcrDeniedUser. AllocatorExtentAllocReqs. NodeAgentStatProcessed. NodeAgentMigrationsDone. ( FilePDbPreReadKB + FilePDbAcqReadKB ) AS PhysicalReadPermKB. MemCtxtAllocs. (ProcWaitSegNoVirtual + ProcWaitSegMDL + ProcWaitSegLock) AS ProcWaitSeg. MemKBRes. FileFcrBlocksDeniedThreshUser. (FilePCiFWriteKB + FileSDbFWriteKB + FileSCiFWriteKB + FileTJtFWriteKB + FileAPtFWriteKB) AS PhysicalWriteOtherKB. NodeAgentMigrationsStarted. FileCylMigrs. (FilePCiFWrites + FileSDbFWrites ( FilePDbAcqKB ( FileSDbAcqKB FilePCiAcqKB FilePCiPresKB ) AS LogicalReadPermKB. AllocatorExtentFreeReqs. FlowCtlCnt. FileFcrDeniedThreshUser. ( FilePCiAcqReadKB + FileSDbAcqReadKB + FileSCiAcqReadKB + FileTJtAcqReadKB + FileAPtAcqReadKB + FilePCiPreReadKB + FileSDbPreReadKB + FileSCiPreReadKB + FileTJtPreReadKB + FileAPtPreReadKB) AS PhysicalReadOtherKB.Resource Usage Macros and Tables /* * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * */ AS LogicalWriteOther. FlowControlled.

0:hdr COD: CodFactor */ . SpareTrack05. SpareTmon02.1 13. SpareCount09. MemAllocKB. CPUUServDisp. CPUUServMisc. SpareTmon07.1 WorkMsgSendDelay WorkMsgSendDelayMax WorkMsgReceiveDelay WorkMsgReceiveDelayMax WorkTimeInuse WorkTimeInuseMax AwtReleases WorkMsgSendDelayCnt WorkMsgReceiveDelayCnt */ */ */ */ */ */ */ */ */ */ /* 13. NumRequests. SpareTmon04. SpareTrack04.1 13. IOBlks. CPUUServAWT. SpareTmon08. RelWgt. SpareTrack02. SpareTmon05. SpareCount05. SpareCount02. SpareTrack01.1 13.1 13. NumProcs. SpareCount07. AGid.1 13. SpareTmon06. SpareCount03. SpareTmon03. SpareTmon01. QLengthMax. SpareCount00. SpareCount04. CpuTime. SpareCount06. SpareTmon09. WDid. ServiceTime. CPUUExecAWT. CPUUExecDisp. ServiceTimeMax. */ /* WorkMsgReceiveDelay in SpareCount02 is wait time of /* messages still waiting for AWTs.1 13. SpareCount01.162 Resource Usage Macros and Tables /* Remaining table fields */ Active. SpareTrack08. SpareTrack07.1 AMPcount /* 14. QLength. CPUUExecMisc.1 13.1 13. QWaitTimeMax. NumSets. SpareCount08. */ */ /* /* /* /* /* /* /* /* /* 13. SpareTrack09. SpareTrack06. MemAllocs. QWaitTime. /* QWaitTime = wait time of messages delivered. SpareTmon00. SpareTrack03. SpareTrack00.

FileTJtAcqReads. FilePDbAcqReads. FileAPtAcqs. WorkTypeMax09. WorkTypeMax15. WorkTypeMax12. WorkTypeMax13. FilePDbAcqKB. WorkTypeInuse04. WorkTypeMax08. WorkTypeMax04. WorkTypeInuse10. WorkTypeMax00. WorkTypeMax10. WorkTypeMax01. WorkTypeInuse13. FileSDbAcqReads. FileSCiAcqReads. WorkTypeMax02. FileAPtAcqReads. WorkTypeInuse09. WorkTypeInuse15. FileSDbAcqs. FileSCiAcqKB. WorkTypeMax03. WorkTypeInuse11. FilePDbAcqs. WorkTypeInuse08. WorkTypeMax06. FilePCiAcqReadKB. FileAPtAcqReadKB. FileAPtAcqKB. WorkTypeMax05.Resource Usage Macros and Tables 163 /* AWT fields */ WorkTypeInuse00. FileTJtAcqs. FileTJtAcqKB. WorkTypeInuse12. FileSDbAcqReadKB. WorkTypeInuse14. FileSCiAcqs. WorkTypeInuse02. WorkTypeInuse03. FileSCiAcqReadKB. WorkTypeInuse07. . FilePCiAcqKB. FileSDbAcqKB. FilePCiAcqs. FileTJtAcqReadKB. FilePDbAcqReadKB. WorkTypeInuse06. WorkTypeInuse01. WorkTypeMax14. WorkTypeInuse05. WorkTypeMax07. WorkTypeMax11. FilePCiAcqReads.

FileSCiPresKB. FilePDbFWrites. FileSCiPres. NetPtPReads. FileAPtPreReads. FileSDbDyRRels. FileSCiPreReads. FileTJtFWriteKB. FileAPtPresKB. FileTJtDyRRels. FileSDbPreReads. FilePCiPresKB. FileSDbPreReadKB. FilePCiDyRRels. NetBrdReads. NetPtPReadKB. FileSDbDyRRelKB. FilePDbDyRRels. FilePDbFWriteKB. FilePDbPreReadKB. FileSDbPresKB. FileAPtDyRRelKB. FileTJtPresKB. NetBrdWrites. FileTJtPreReadKB. FileSCiDyRRelKB. FilePCiDyRRelKB. FileSDbFWriteKB. FileAPtPres. FilePCiFWrites. NetPtPWriteKB. FileSDbPres. FileAPtFWriteKB. FilePDbDyRRelKB. FileTJtDyRRelKB.164 Resource Usage Macros and Tables FilePDbPres. FileTJtPres. FileAPtPreReadKB. FileSDbFWrites. FilePCiPreReadKB. FileAPtDyRRels. FilePDbPresKB. FilePCiPres. FileSCiFWriteKB. FileTJtPreReads. FilePCiFWriteKB. FileAPtFWrites. FileSCiDyRRels. FileSCiPreReadKB. FileSCiFWrites. NetPtPWrites. . FileTJtFWrites. FilePDbPreReads. FilePCiPreReads.

/* Process blocking reasons: ProcWait time in milliseconds */ ProcWaitSegNoVirtual. ProcWaitFsgWrite. ProcBlksSegMDL. ProcBlksTime. ProcBlksDBLock. ProcBlksMonResume. ProcBlksFsgLock. ProcBlksFsgNIOs. ProcBlksNetThrottle. ReadsWarm. /* Process blocking reasons: ProcBlks count of blocks */ ProcBlksSegNoVirtual. ProcBlksCpuLimit. ProcWaitNetThrottle. ProcBlksFlowControl. ProcWaitFsgNIOs. ProcWaitSegMDL. WritesHot. ProcWaitMonResume. ProcWaitTime. ProcWaitSegLock. ProcBlksQnl. ProcWaitMonitor. ReadsCold. ProcWaitDBLock. ProcWaitFsgLock. WritesWarm. ProcBlksSegLock.Resource Usage Macros and Tables 165 /* TVSA fields */ ReadsHot. ProcBlksMonitor. ProcWaitFsgRead. ProcWaitCpuLimit FROM ResUsageSps WITH CHECK OPTION. ProcWaitQnl. ProcBlksFsgRead. ProcWaitFlowControl. . WritesCold. ProcBlksFsgWrite.

thetime. REPLACE VIEW DBC. vprid. Note: The data columns in this view will change as the columns in the ResUsageSvpr table change. ProcBlksDBLock. Reserved. CollectIntervals. SpareTmon00 SpareTmon01 SpareTmon02 SpareTmon03 AS AS AS AS COD. SpareCount03 AS DBMergeExtrIO. 166 Resource Usage Macros and Tables /* SVpr table fields (remaining) */ Active. GmtTime. MemPCiAllocs.ResSvprView AS SELECT /* housekeeping fields */ NodeID. /* transformed fields */ /* user defined co-existence system node groupings */ CASE WHEN NodeId IN ('001-01'. Secs. MemSDbAllocs. NodeType. FSGCacheWaitTime. SummaryFlag. NCPUs. thedate. FSGCacheWaitTimeMax. VprType.ResSvprView Use the ResSvprView to access the ResUsageSvpr table data. SpareCount02 AS DBMergeElim. This view allows data to be properly presented and reports all the columns available from the ResUsageSvpr table. SpareCount04 AS FileACPCylsSkipped. SpareCount07 AS FileACPCylsPostponed. MemPDbAllocs. '001-03'. SpareCount01 AS DBMergeDone. MemSCiAllocs. . SpareCount06 AS FileACPCylsUnFSEOnly. FSGCacheWaits. NominalSecs. ProcWaitDBLock. '001-05') THEN 'A' WHEN NodeId IN ('001-02'. CentiSecs. /* Spare field usage */ SpareCount00 AS DBMergeTried. ProcPendDBLock. '001-06') THEN 'A' ELSE 'A' END AS GroupId. '001-04'. SpareCount05 AS FileACPCylsMigr.

MemAPtAllocKB. MemPCiKBResCU. NetPtPWrites. MemTJtKBResCA. NetBrdWriteKB. MsgWorkQLenMax. DBLocksHeldMax. MemSwapReadKB. MemSDbAllocKB. NetPtPReadKB. MemSDbKBResCA. MemSDbKBResDA. MemSCiKBResDU. MemPDbKBResCA. MemPCiKBResDA. MemPDbKBResCU. MemPCiKBResCA. MemCtxtDestroys. NetBrdWrites. MemAPtKBResDU. MemPKBResFrz. MemCtxtDeaccesses. MemAPtKBResCU. NetPtPReads. MemCtxtRes. NetBrdReads. MemCtxtAccessKB. MemSwapReads. MemCtxtDestroyKB. FilePDbAcqs. MemSCiKBResDA. MemCtxtAllocs. MemTJtAllocKB. DBLockEnters. . MemAPtKBResCA. MemCtxtDeaccessKB. MemSDbKBResCU. MemPDbKBResDU. NetPtPWriteKB. MemSCiKBResCA.Resource Usage Macros and Tables 167 MemTJtAllocs. MemCtxtPageWrites. MemCtxtPageReads. MemSDbKBResDU. MemPDbKBResDA. DBLockDeadlocks. MemSCiAllocKB. MemSwapDropKB. MemPCiKBResDU. MemSCiKBResCU. MemAPtAllocs. MemPDbAllocKB. MemTJtKBResDU. NetBrdReadKB. MemTJtKBResDA. MemTJtKBResCU. MemPCiAllocKB. MemCtxtAccesses. MemSwapDrops. MemAPtKBResDA. DBLockBlocks. DBLockBlocksMax.

FileSCiAcqReadKB. FileAPtPreReadKB. FilePDbAcqKB. FileAPtPresKB. FileAPtAcqKB. FilePDbAcqReadKB. FilePDbDyRRels. FilePCiFWrites. FileTJtPreReads. FileSCiDyRRelKB. 168 Resource Usage Macros and Tables . FileSDbAcqKB. FileTJtDyRRelKB. FileSDbPreReadKB. FilePDbFWrites. FileSCiPreReadKB. FileAPtPreReads. FileSCiAcqKB. FilePCiDyRRels. FilePCiPres. FileAPtPres. FileSDbAcqs. FilePCiPreReadKB. FilePDbDyRRelKB. FileAPtAcqReadKB. FilePCiDyRRelKB. FileTJtPreReadKB. FileSCiAcqReads. FilePDbPreReadKB. FilePDbPres. FileSCiDyRRels. FileSDbDyRRels. FileSDbFWrites. FileTJtAcqReads. FileAPtAcqReads. FileTJtDyRRels. FileSDbPresKB. FileSCiPresKB. FilePDbPreReads. FileAPtAcqs. FileTJtAcqKB. FileSCiAcqs.FilePCiAcqs. FileTJtPres. FileSDbAcqReads. FileSDbPreReads. FileTJtAcqs. FilePDbAcqReads. FileSCiFWrites. FileSCiPres. FileSDbDyRRelKB. FileSDbPres. FilePCiPresKB. FileAPtDyRRels. FileAPtDyRRelKB. FilePCiPreReads. FileTJtPresKB. FileSDbAcqReadKB. FileSCiPreReads. FileTJtAcqReadKB. FilePDbPresKB. FilePCiAcqReads. FilePCiAcqReadKB. FilePCiAcqKB.

FileSDbCnADrpKB. FileCylFrees. FileAPtCnADrps. FilePCiDyAWriteKB. FilePCiDyAWrites. FileSDbFDrpKB. FileTJtDyAWriteKB. FileSCiFDrpKB. FileMCylPacks. FilePCiCnADrps. FileAPtCnRRels. FilePCiCnRRels. FileAPtCnRRelKB. FileTJtCnADrps. FileSCiDyAWriteKB. FileSDbCnRRels. FileSCiCnADrpKB. FileLockDeadlocks. FileTJtCnRRelKB. FileTJtDyAWrites. FilePCiCnADrpKB. FileSDbFWriteKB. FileTJtCnRRels. FilePDbDyAWrites. FileSCiCnRRels. FileSCiCnRRelKB. FilePDbDyAWriteKB. FileSCiFWriteKB. FilePCiFWriteKB. FileLockBlocks. FileAPtDyAWriteKB. FileSDbDyAWrites. FileSDbCnADrps. FileAPtCnADrpKB. FileCylMigrs. FilePDbFDrpKB. FileTJtFWriteKB. FileAPtFWriteKB.Resource Usage Macros and Tables 169 FileTJtFWrites. FileSCiDyAWrites. FilePDbFWriteKB. FilePDbCnRRels. . FileSDbFDrps. FilePCiFDrps. FileSDbDyAWriteKB. FileLockEnters. FileAPtFDrpKB. FileTJtCnADrpKB. FileTJtFDrpKB. FilePCiFDrpKB. FileAPtFWrites. FileCylAllocs. FileAPtDyAWrites. FileSCiFDrps. FilePDbCnADrpKB. FileSCiCnADrps. FileTJtFDrps. FileSDbCnRRelKB. FilePDbCnADrps. FilePDbFDrps. FilePDbCnRRelKB. FileAPtFDrps. FilePCiCnRRelKB.

CPUUServPart15. CPUUServPart00. 170 Resource Usage Macros and Tables . CPUUServPart33. FileFcrDeniedUser. CPUUServPart31. CPUUServPart19. CPUUServPart14. CPUUServPart37.FileCylDefrags. FileSyncScanners. CPUUServPart12. CPUUServPart24. CPUUServPart20. CPUUServPart36. CPUUServPart26. CPUUServPart09. FileFcrBlocksDeniedCache. CPUUServPart32. FileFcrDeniedCache. CPUUServPart17. FileFcrBlocksDeniedKern. CPUUServPart05. FileSyncSubtables. DBLockBlocksSum. FileFcrRequests. DBLocksHeldSum. FileFcrRequestsAdaptive. FileWCylFrees. FileFcrDeniedKern. FileSyncGroups. FileFcrBlocksDeniedThreshKern. MsgChnLastDone. CPUUServPart23. FileFcrBlocksDeniedThreshUser. CPUUServPart22. CPUUServPart25. FileWCylAllocs. CPUUServPart08. FileFcrDeniedThreshUser. CPUUServPart03. CPUUServPart10. FileFcrBlocksDeniedUser. CPUUServPart38. CPUUServPart27. FileFcrDeniedThreshKern. FileSyncScans. CPUUServPart29. CPUUServPart35. CPUUServPart11. CPUUServPart16. CPUUServPart04. CPUUServPart02. CPUUServPart13. CPUUServPart18. CPUUServPart34. CPUUServPart06. CPUUServPart21. MsgWorkQLenSum. CPUUServPart01. CPUUServPart30. CPUUServPart07. CPUUServPart28. FileFcrBlocksRead.

CPUUExecPart21. CPUUExecPart16. CPUUServPart40. CPUUExecPart13. CPUUExecPart18. CPUUExecPart08. CPUUExecPart30. CPUUExecPart33. CPUUServPart42. CPUUExecPart25.Resource Usage Macros and Tables 171 CPUUServPart39. CPUUExecPart05. AllocatorExtentAllocReqs. CPUUExecPart26. CPUUServPart45. CPUUServPart46. CPUUExecPart42. CPUUExecPart19. CPUUExecPart36. CPUUExecPart32. CPUUServPart47. CPUUExecPart37. CPUUExecPart23. CPUUExecPart38. CPUUServPart44. NodeAgentMigrationsStarted. CPUUExecPart11. CPUUExecPart27. . CPUUExecPart04. CPUUServPart43. AllocatorExtentFreeReqs. CPUUExecPart34. CPUUExecPart28. CPUUExecPart01. CPUUExecPart10. CPUUExecPart03. CPUUExecPart24. CPUUExecPart12. CPUUExecPart43. CPUUExecPart00. CPUUExecPart41. CPUUExecPart14. CPUUExecPart29. NodeAgentMigrationsDone. CPUUExecPart02. CPUUExecPart39. CPUUExecPart15. CPUUExecPart44. CPUUExecPart47. CPUUExecPart17. CPUUExecPart46. CPUUExecPart35. AllocatorMapIOsStarted. CPUUExecPart07. CPUUExecPart22. CPUUExecPart09. CPUUExecPart06. AllocatorMapIOsDone. CPUUServPart41. CPUUExecPart20. CPUUExecPart45. CPUUExecPart31. CPUUExecPart40.

SpareCount03. SpareTmon00. SpareCount01. SpareCount00. SpareCount07.NodeAgentStatProcessed. WriteResponseHotMin. SpareTmon01. MISleepTime. SpareTrack04. SpareCount04. SpareTrack06. SpareTrack02. ReadsCold. IoRespMax. WriteResponseHotTotal. WritesCold. ReadResponseWarmMax. SpareTmon06. ReadsWarm. SpareTmon04. WriteResponseWarmMax. WriteResponseWarmTotal. ReadResponseColdMin. SpareTrack07. MIWriteLocks. SpareCount02. SpareCount09. 172 Resource Usage Macros and Tables . SpareTrack08. SpareTrack09. IoGapMax. WritesWarm. SpareTrack03. ReadResponseColdMax. ReadResponseHotTotal. WriteResponseColdMax. SpareTmon02. SpareCount08. MIWriteTime. SpareTmon05. MISleeps. SpareTrack00. ReadResponseWarmTotal. ReadResponseHotMax. ReadsHot. MIWrites. WriteResponseWarmMin. WriteResponseColdMin. SpareTrack01. WriteResponseColdTotal. SpareCount05. ReadResponseHotMin. SpareCount06. ReadResponseColdTotal. WritesHot. SpareTrack05. ReadResponseWarmMin. WriteResponseHotMax. SpareTmon03. MIWriteLockTime. MISleepTimeMax. MIWriteLockTimeMax. MIWriteTimeMax.

COMMENT ON VIEW DBC. SpareTmon09 FROM ResUsageSvpr WITH CHECK OPTION.Resource Usage Macros and Tables 173 SpareTmon07. . SpareTmon08.ResSvprView AS 'View of ResUsageSvpr (per Vproc) table data'.

Chapter 14: Resource Usage Views ResSvprView 174 Resource Usage Macros and Tables .

.9%'. host type.99% .99% .99% 999-99 999-99 999.99') unless the number represents a percentage or sum of percentages.. A virtual processor type..-----.99% 999. and so on Description The date at the end of a log interval.---. Page <num> 2nd 3rd Stat Stat .-------. or a special type for certain reports.99 99999.99 99999...99 99999. Numbers are generally displayed with the appropriate fixed format (for example.-------99999.99 99999.. AAAA 999-99 999-99 999.. 'zz9.. and so on NodeID VprocID HostID GroupID The appropriate statistics.99'). Macro Output Format Resource usage macros provide output in the following general format. -------. Statistics on each line cover the time period ending at the indicated time. The time at the end of a log interval... It is one or more of the following: • • • • 1st Stat.99 where: Column Date Time Type 1st ID... 'zzzz9..------99/99/99 99:99:99 AAAA 999-99 999-99 999.-----.. The appropriate identifier.99 99999.99 99999... <Report Date> <Title of Report> --------------1st 2nd 1st Date Time Type Id Id Stat -------...99 99999..99% 99:99:99 AAAA 999-99 999-99 999.99 99999. depending on the macro. which varies. Details are given with the descriptions of each macro in this chapter.. 'zz9' or 'zz9. Percentages are displayed with the appropriate format (for example. Resource Usage Macros and Tables 175 ..99 99999. logical device type. 2nd ID.Chapter 15: Resource Usage Macros Macro Output Format CHAPTER 15 Resource Usage Macros This chapter describes the output format of the resource usage macros and each macro. 2nd Stat..

Chapter 15: Resource Usage Macros Macro Output Format Unless otherwise specified. logging must be enabled on the ResUsage table used by the view or macro. Repeated date. time. A question mark does not mean there is an error. It represents data that is not available. All reports are ordered by date. and so on. such as per second. Usage Notes To get current data. 176 Resource Usage Macros and Tables . time. by dividing KBs by number of blocks read. for example. Instead. 1st ID. Question Marks Question marks used as values in the output examples are generated when a division by zero is made. they are converted to a normalized value. When there are no blocks read. all statistical numbers are expressed as either: • • • Percentage Milliseconds (ms) Kilobytes (KBs) Columns whose values depend on the logging rate are never reported as raw data. type. 2nd ID. and ID column values are suppressed until a new row presents a different value. KB is divided by zero. The numbers in the columns are calculated. but indicates that there is no information to report for this time period. type.

. ToNode and Node parameters. ResAWT ResAWTByAMP ResAWTByNode Collects and reports the average AWT.. AMP ID. Name the node log rate.ToNode). Input Format Examples The input forms of these three macros are described below.ToDate.ToDate. in use for all AMPs in the system. after the Date and Time columns.FromTime. Resource Usage Macros and Tables 177 . See “Executing Macros” on page 32 for a description of the FromDate. EXEC ResAWTByAMP (FromDate.FromTime. ToDate. in the ResAWTByAMP output.. on all AMPs in each node. and ResAWTByNode macros. the 22 statistics columns.ToDate. Statistics columns 1 2 3 through 24 Description Node ID. The following table describes the 24 statistics columns.ToTime.ToTime). in use for each AMP.ToTime). FromTime.FromNode.FromTime. after the Date and Time columns. ResAWTByAMP. In the ResAWT output report. FromNode. provide a summary of the AWTs resource usage.. EXEC ResAWTByNode (FromDate. EXEC ResAWT (FromDate. Output Examples The reports in the following sections are sample output reports from the ResAWT.Chapter 15: Resource Usage Macros ResAWT Macros ResAWT Macros Function Macro. Summary of AWTs resource usage. ToTime. Usage Notes For any of these macros the following usage notes apply: • • Logging must be enabled on ResUsageSawt.

current number of AWTs in use during the log period for each new first-level secondary work type for the VprId vproc. current number of AWTs in use during the log period for each message that does not fall within the standard work type hierarchy for the VprId vproc. in the ResAWTByNode output report.. current number of AWTs in use during the log period for each transaction abort request for the VprId vproc... Work Ctl AWTs 178 Resource Usage Macros and Tables .Chapter 15: Resource Usage Macros ResAWT Macros The following table describes the 23 statistics columns. Mail box Depth In Flow Ctl? Flow Ctls Per Sec Work Type In Use Work New AWTs Work One AWTs New + One AWTs Work Two AWTs Work 3 AWTs Work Abrt AWTs Work Spwn AWTs Work Norm AWTs Reports the. current number of AWTs in use during the log period for each second-level secondary work type for the VprId vproc. Summary of AWTs resource usage. AMP that is or is not in flow control. Statistics columns 1 2 through 23 Description Node ID. and ResAWTByAMP. current number of AWTs in use during the log period for each new work type for the VprId vproc.. current number of AWTs in use during the log period for each first-level secondary work type for the VprId vproc. The following table describes the columns in all output reports (with the exception of ResAWTByNode. Column. current depth of the AMP work mailbox. which has the NodeId column. number of times during the log period that the system entered the flow control state from a non-flow controlled state. summary of the previous two columns: Work New AWTs and Work One AWTs. which has the Node ID and AMP ID columns). after the Date and Time columns. Note: This column is not currently used. current number of AWTs in use during the log period for each spawned abort request for the VprId vproc. current number of AWTs in use during the log period for each third-level secondary work type for the VprId vproc.

current number of AWTs in use during the log period for each expedited Allocation Groups for the VprId vproc. see Chapter 7: “ResUsageSawt Table. the maximum number of AWTs in use at one time during the log period for each first-level secondary work type for the VprId vproc. the maximum number of AWTs in use at one time during the log period for each spawned abort request for the VprId vproc. See the "Priority Scheduler (schmon)" chapter in Utilities for details.. the maximum number of AWTs in use at one time during the log period for each transaction abort request for the VprId vproc. the maximum number of AWTs in use at one time during the log period for each expedited Allocation Groups for the VprId vproc. Note: This column is not currently used.) Max Work New AWTs Max Work One AWTs Max Work Two AWTs Max Work 3 AWTs Max Work Abrt AWTs Max Work Spwn AWTs Max Work Norm AWTs Max Work Ctl AWTs Max Work Exp AWTs For a complete description of the columns above.” Resource Usage Macros and Tables 179 . the maximum number of AWTs in use at one time during the log period for each message that does not fall within the standard work type hierarchy for the VprId vproc.. (Expedited Allocation Groups exist when using the reserved AWT pool. Work Exp AWTs Reports the. See the "Priority Scheduler (schmon) chapter in Utilities for details.. the maximum number of AWTs in use at one time during the log period for each second-level secondary work type for the VprId vproc.Chapter 15: Resource Usage Macros ResAWT Macros Column. (Expedited Allocation Groups exist when using the reserved AWT pool. the maximum number of AWTs in use at one time during the log period for each third-level secondary work type for the VprId vproc.) the maximum number of AWTs in use at one time during the log period for each new work type for the VprId vproc..

ResAWT Sample Output
07/08/17 AMP Worker Task Summary Average Usage per AMP Across System Max Work Abrt AWTs ---1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 Max Work Spwn AWTs ---0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 Max Max Max Work Work Work Norm Ctl Exp AWTs AWTs AWTs ---- ---- --0 0 3 0 0 2 0 0 2 0 0 3 0 0 3 0 0 4 0 0 3 0 0 3 0 0 3 0 0 4 0 0 3 0 0 3 0 0 3 0 0 4 0 0 3 0 0 3 0 0 4 0 0 4 0 0 4 0 0 3 0 0 4 0 0 3 0 0 3 Page 1 Max Max Max Max Mail In Flow Work Work New+ Work Work Work Work Work Work Work Work Work Work Work Box Flow Ctls New One One Two 3 Abrt Spwn Norm Ctl Exp New One Two 3 Date Time Depth Ctl? PerSec AWTs AWTs AWTs AWTs AWTs AWTs AWTs AWTs AWTs AWTs AWTs AWTs AWTs AWTs -------- -------- ------ ----- ------ ---- ---- ---- ---- ---- ---- ---- ---- ---- ---- ---- ---- ---- ---07/08/12 23:00:00 8 1.00 0.01 31 22 54 0 0 0 0 0 0 0 35 25 2 1 23:01:00 9 0.00 0.02 28 25 54 0 0 0 0 0 0 0 33 27 1 1 23:02:00 6 0.00 0.00 31 22 54 0 0 0 0 0 0 0 35 26 1 1 23:03:00 3 0.00 0.00 31 22 53 0 0 0 0 0 0 0 35 26 1 1 23:04:00 2 0.00 0.00 31 22 52 0 0 0 0 0 0 0 36 27 1 1 23:05:00 5 0.00 0.00 31 22 53 0 0 0 0 0 0 0 36 26 1 1 23:06:00 3 0.00 0.00 27 24 51 0 0 0 0 0 0 0 34 27 1 1 23:07:00 1 0.00 0.00 21 19 40 0 0 0 0 0 0 1 35 29 1 1 23:08:00 1 0.00 0.00 26 20 46 0 0 0 0 0 0 0 35 29 1 1 23:09:00 2 0.00 0.00 30 20 50 0 0 0 0 0 0 0 37 26 1 1 23:10:00 1 0.00 0.00 29 16 46 0 0 0 0 0 0 0 38 25 2 1 23:11:00 2 0.00 0.00 30 19 49 0 0 0 0 0 0 0 38 27 1 1 23:12:00 2 0.00 0.00 31 21 52 0 0 0 0 0 0 0 37 26 1 1 23:13:00 1 0.00 0.00 29 19 49 0 0 0 0 0 0 0 36 26 1 1 23:14:00 1 0.00 0.00 29 18 47 0 0 0 0 0 0 0 36 25 1 1 23:15:00 2 0.00 0.00 29 19 48 0 0 0 0 0 0 0 37 25 2 1 23:16:00 3 0.00 0.00 34 18 52 0 0 0 0 0 0 0 37 25 1 1 23:17:00 6 0.00 0.00 35 19 54 0 0 0 0 0 0 0 40 22 1 1 23:18:00 8 0.00 0.00 30 23 53 0 0 0 0 0 0 0 37 24 1 1 23:19:00 1 0.00 0.01 25 24 49 0 0 0 0 0 0 0 36 30 1 1 23:20:00 1 0.00 0.00 28 18 46 0 0 0 0 0 0 0 36 30 2 2 23:21:00 7 0.00 0.01 34 20 54 0 0 0 0 0 0 0 36 26 1 1 23:22:00 2 0.00 0.01 30 22 53 0 0 0 0 0 0 0 36 27 1 1

180 Resource Usage Macros and Tables

ResAWTByAMP Sample Output
07/08/17 AMP Worker Task Summary Usage per AMP Max Work New AWTs ---33 34 34 34 34 35 35 34 34 33 34 34 34 33 34 34 34 34 33 34 32 33 33 33 Max Work One AWTs ---24 24 24 24 25 23 24 23 23 25 24 24 24 24 24 24 24 24 25 24 26 26 25 26 Max Work Two AWTs ---2 1 1 1 1 1 1 1 2 0 1 1 0 0 1 1 1 1 1 0 0 0 0 0 Max Max Max Max Work Work Work Work 3 Abrt Spwn Norm AWTs AWTs AWTs AWTs ---- ---- ---- ---1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 Max Work Ctl AWTs ---0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 Max Work Exp AWTs ---3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 2 1 1 1 Page 1

Mail In Flow Work AMP Box Flow Ctls New Date Time ID Depth Ctl? PerSec AWTs -------- -------- ------ ------ ----- -----07/08/12 23:00:00 0 3 0.00 0.00 31 1 1 0.00 0.00 32 2 12 0.00 0.00 31 3 14 0.00 0.00 32 4 4 0.00 0.00 32 5 11 0.00 0.00 32 6 12 0.00 0.00 30 7 17 0.00 0.00 32 8 5 0.00 0.00 31 9 4 0.00 0.00 32 10 1 0.00 0.00 30 15 0.00 0.00 32 12 7 0.00 0.00 32 13 1 0.00 0.00 29 14 1 0.00 0.00 31 15 4 0.00 0.00 32 16 27 1.00 0.22 31 17 5 0.00 0.00 32 18 1 0.00 0.00 29 19 11 0.00 0.02 32 23:01:00 0 1 2 3 14 11 19 12 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 28 29 29 28

Work One AWTs ---22 22 23 22 22 22 24 22 23 22 23 22 22 24 23 22 23 22 23 22 25 25 25 26

New+ One AWTs ---53 54 54 54 54 54 54 54 54 54 53 54 54 53 54 54 54 54 52 54 53 54 54 54

Work Work Work Work Work Two 3 Abrt Spwn Norm AWTs AWTs AWTs AWTs AWTs ---- ---- ---- ---- ---0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0

Work Ctl AWTs ---0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

Work Exp AWTs ---0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

Resource Usage Macros and Tables
07/08/17

4 5 6 7 8 9 10 11 12 13 14 15 16 23:02:00 0 1 2

3 13 19 8 12 3 4 13 3 1 1 11 10 9 6 1

0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00

0.00 0.00 0.13 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.17 0.00 0.00 0.00

29 29 28 29 29 28 29 29 30 26 26 28 29 31 31 31

25 25 26 25 25 26 25 25 24 26 25 26 25 22 23 23

54 54 54 54 54 54 54 54 54 52 51 54 54 53 54 54

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

33 32 32 32 33 33 33 32 32 31 33 32 32 33 32 34

26 25 26 26 26 26 26 25 25 26 26 26 27 25 26 25

0 0 1 0 0 0 0 0 0 0 0 0 0 1 0 0

0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0

0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

1 1 1 1 1 2 1 2 2 2 1 1 2 2 2 2

ResAWTByNode Sample Output
AMP Worker Task Summary Average Usage per AMP By Node Page Max Work New AWTs ---35 33 35 35 36 36 34 35 35 37 38 38 37 36 36 37 37 40 37 36 36 Max Work One AWTs ---25 27 26 26 27 26 27 29 29 26 25 27 26 26 25 25 25 22 24 30 30 Max Work Two AWTs ---2 1 1 1 1 1 1 1 1 1 2 1 1 1 1 2 1 1 1 1 2 1 Max Work Ctl AWTs ---0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 Max Work Exp AWTs ---3 2 2 3 3 4 3 3 3 4 3 3 3 4 3 3 4 4 4 3 4 Max Max Max Max Work Work Work Work 3 Abrt Spwn Norm AWTs AWTs AWTs AWTs ---- ---- ---- ---1 1 0 0 1 1 0 0 1 1 0 0 1 1 0 0 1 1 0 0 1 1 0 0 1 1 0 0 1 1 0 0 1 1 0 0 1 1 0 0 1 1 0 0 1 1 0 0 1 1 0 0 1 1 0 0 1 1 0 0 1 1 0 0 1 1 0 0 1 1 0 0 1 1 0 0 1 1 0 0 2 1 0 0

Mail In Node Box Flow Date Time ID Depth Ctl? -------- -------- ------ ------ ----07/08/12 23:00:00 1-04 8 1.00 23:01:00 1-04 9 0.00 23:02:00 1-04 6 0.00 23:03:00 1-04 3 0.00 23:04:00 1-04 2 0.00 23:05:00 1-04 5 0.00 23:06:00 1-04 3 0.00 23:07:00 1-04 1 0.00 23:08:00 1-04 1 0.00 23:09:00 1-04 2 0.00 23:10:00 1-04 1 0.00 23:11:00 1-04 2 0.00 23:12:00 1-04 2 0.00 23:13:00 1-04 1 0.00 23:14:00 1-04 1 0.00 23:15:00 1-04 2 0.00 23:16:00 1-04 3 0.00 23:17:00 1-04 6 0.00 23:18:00 1-04 8 0.00 23:19:00 1-04 1 0.00 23:20:00 1-04 1 0.00

Flow Ctls PerSec -----0.01 0.02 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.01 0.00

Work New AWTs ---31 28 31 31 31 31 27 21 26 30 29 30 31 29 29 29 34 35 30 25 28

Work One AWTs ---22 25 22 22 22 22 24 19 20 20 16 19 21 19 18 19 18 19 23 24 18

New+ One AWTs ---54 54 54 53 52 53 51 40 46 50 46 49 52 49 47 48 52 54 53 49 46

Work Work Work Work Work Two 3 Abrt Spwn Norm AWTs AWTs AWTs AWTs AWTs ---- ---- ---- ---- ---0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

Work Ctl AWTs ---0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

Work Exp AWTs ---0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0

181

Chapter 15: Resource Usage Macros ResCPUByAMP Macros

ResCPUByAMP Macros
Function
Macro... ResCPUByAMP ResCPUByAMPOneNode ResAmpCpuByGroup Reports the following... how each AMP on each node utilizes the CPUs. how each AMP on a specific node utilizes the CPUs. the summary of AMP CPU usage by node grouping.

Input Format Examples
The input forms of these three macros are described below.
EXECUTE ResCPUByAMP (FromDate,ToDate,FromTime,ToTime,FromNode,ToNode); EXECUTE ResCPUByAMPOneNode (FromDate,ToDate,FromTime,ToTime,Node); EXECUTE ResAmpCpuByGroup (FromDate,ToDate,FromTime,ToTime);

See “Executing Macros” on page 32 for a description of the FromDate, ToDate, FromTime, ToTime, FromNode, ToNode and Node parameters.

Usage Notes
For any of these macros the following usage notes apply: • Logging must have been enabled on ResUsageSvpr at some time before macro execution. See Chapter 2: “Planning Your Resource Usage Data” for an explanation of how to enable/ disable logging. Name the node log rate. Note: It is not necessary that logging for the table and the rate be enabled at the moment the macro is executed. For a description of partitions and partition assignments in Teradata Database, see Appendix D: “Partition Assignments.”

Output Examples
The reports in the following sections are sample output reports from the ResCPUByAMP, the ResCPUByAMPOneNode, and the ResAmpCpuByGroup macros, respectively, where:
Column... Awt User Serv% Reports the percent of time AMPs were busy doing user... service for the AMP Worker Task (Awt) partition.

182

Resource Usage Macros and Tables

Chapter 15: Resource Usage Macros ResCPUByAMP Macros

Column... Misc User Serv% Awt User Exec% Misc User Exec% Total User Serv% Total User Exec% Total Busy%

Reports the percent of time AMPs were busy doing user... service for miscellaneous (all other except Partition 0) AMP partitions. execution within the AMP Worker Task (Awt) partition. execution within miscellaneous (all other except Partition 0) AMP partitions. servicea work. This is the sum of the Awt User Serv%, the Misc User Serv%, and AMP Partition 0 user service%.a executionb work. This is the sum of the Awt User Exec%, Misc User Exec%, and AMP Partition 0 user execution.b service and execution work. This is the sum of the Total User Serv% and the Total User Exec% columns.

a. Service is the time that a CPU is busy executing user service code, which is privileged work performing system-level services on behalf of user execution processes that do not have root privileges. b. Execution is the time a CPU is busy executing user execution code, which is the time spent in a user state on behalf of a process.

Note: The above CPU statistics represent the aggregate of all time spent in the indicated way by all CPUs on the node. Because there are multiple CPUs, the Total Busy % should be compared to a theoretical maximum of 100% times the number of CPUs on the node. The Node CPU column in the following sample outputs reports the number of CPUs (NCPUs). For more information on how to monitor busy AMP Worker Tasks (AWTs), see "AWT Monitor (awtmon)" in Utilities.

ResCPUByAMP Sample Output
01/07/12 Vproc Id ----0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 Node Id -----001-01 001-01 001-01 001-01 001-01 001-01 001-01 001-01 001-01 001-01 001-01 001-01 001-01 001-01 001-01 001-01 001-01 001-01 Node CPUs -------4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 CPU USAGE BY AMP Awt User Serv% ------0.36% 0.26% 0.41% 0.34% 0.25% 0.19% 0.38% 0.31% 0.31% 0.36% 0.39% 0.32% 0.29% 0.21% 0.30% 0.30% 0.40% 0.26% Misc User Serv% ------0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% Awt User Exec% ------0.05% 0.12% 0.12% 0.05% 0.18% 0.06% 0.08% 0.09% 0.08% 0.09% 0.11% 0.12% 0.11% 0.09% 0.06% 0.19% 0.09% 0.08% Misc User Exec% ------0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% Page Total User Serv% ------0.36% 0.30% 0.45% 0.38% 0.28% 0.29% 0.45% 0.34% 0.34% 0.40% 0.41% 0.36% 0.30% 0.22% 0.31% 0.32% 0.46% 0.38% 1 Total User Exec% ------0.05% 0.12% 0.12% 0.05% 0.18% 0.06% 0.08% 0.09% 0.08% 0.09% 0.11% 0.12% 0.11% 0.09% 0.06% 0.19% 0.09% 0.08% Total Busy % ------0.41% 0.42% 0.58% 0.42% 0.45% 0.35% 0.52% 0.42% 0.41% 0.49% 0.52% 0.49% 0.41% 0.31% 0.38% 0.51% 0.55% 0.45%

Date -------01/07/12

Time -------09:57:00 09:57:20 09:57:40 09:58:00 09:58:20 09:58:40 09:59:00 09:59:20 09:59:40

Resource Usage Macros and Tables

183

Chapter 15: Resource Usage Macros ResCPUByAMP Macros

10:00:00

0 1

001-01 001-01

4 4

0.32% 0.28%

0.00% 0.00%

0.08% 0.09%

0.00% 0.00%

0.34% 0.31%

0.08% 0.09%

0.41% 0.40%

Note: The NodeID column only appears in the ResCPUByAMP output report.

ResCPUByAMPOneNode Sample Output
01/07/12 CPU Usage by AMP for Node 001-01 (4 CPUs) Vproc Id ----0 1 0 1 0 1 0 1 0 1 0 1 0 1 Awt User Serv% ------0.36% 0.26% 0.41% 0.34% 0.25% 0.19% 0.38% 0.31% 0.31% 0.36% 0.39% 0.32% 0.29% 0.21% Misc User Serv% ------0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% Awt User Exec% ------0.05% 0.12% 0.12% 0.05% 0.18% 0.06% 0.08% 0.09% 0.08% 0.09% 0.11% 0.12% 0.11% 0.09% Misc User Exec% ------0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% Total User Serv% ------0.36% 0.30% 0.45% 0.38% 0.28% 0.29% 0.45% 0.34% 0.34% 0.40% 0.41% 0.36% 0.30% 0.22% Page Total User Exec% ------0.05% 0.12% 0.12% 0.05% 0.18% 0.06% 0.08% 0.09% 0.08% 0.09% 0.11% 0.12% 0.11% 0.09% 68 Total Busy % ------0.41% 0.42% 0.58% 0.42% 0.45% 0.35% 0.52% 0.42% 0.41% 0.49% 0.52% 0.49% 0.41% 0.31%

Date -------01/07/12

Time -------09:57:00 09:57:20 09:57:40 09:58:00 09:58:20 09:58:40 09:59:00

NCPUs ----4 4 4 4 4 4 4 4 4 4 4 4 4 4

ResAmpCpuByGroup Sample Output
01/07/12 Group Id ----A A A A A A A A A A A A A A Node CPUs -------4 4 4 4 4 4 4 4 4 4 4 4 4 4 AMP CPU USAGE BY GROUP Awt User Serv% ------0.32% 0.33% 0.35% 0.36% 0.27% 0.29% 0.36% 0.35% 0.34% 0.41% 0.28% 0.35% 0.34% 0.26% Misc User Serv% ------0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% Awt User Exec% ------0.07% 0.08% 0.07% 0.09% 0.09% 0.06% 0.06% 0.11% 0.07% 0.04% 0.09% 0.09% 0.06% 0.08% Misc User Exec% ------0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% Total User Serv% ------0.36% 0.36% 0.37% 0.39% 0.28% 0.34% 0.40% 0.38% 0.36% 0.43% 0.28% 0.43% 0.42% 0.29% Page Total User Exec% ------0.07% 0.08% 0.07% 0.09% 0.09% 0.06% 0.06% 0.11% 0.07% 0.04% 0.09% 0.09% 0.06% 0.08% 45 Total Busy % ------0.43% 0.44% 0.44% 0.48% 0.37% 0.40% 0.46% 0.49% 0.43% 0.47% 0.37% 0.53% 0.48% 0.37%

Date -------01/07/12

Time -------09:51:40 09:52:00 09:52:20 09:52:40 09:53:00 09:53:20 09:53:40 09:54:00 09:54:20 09:54:40 09:55:00 09:55:20 09:55:40 09:56:00

Note: The GroupID column only appears in the ResAmpCpuByGroup output report.

Normalized Viewing of CPU Usage by AMP
Some users may prefer to view CPU usage by AMP in a normalized fashion. Conceptually, this restates each of the above statistics in terms of percentage of total CPU capacity of the node.

184

Resource Usage Macros and Tables

Vproc.2. SEL TheDate.3.Chapter 15: Resource Usage Macros ResCPUByAMP Macros The following SQL example shows how to perform this normalization for the Total Busy % statistic. Resource Usage Macros and Tables 185 .TITLE ‘Total// Busy// %’) FROM ResCpuUsageByAMPView WHERE TheDate = CURRENT_DATE AND TheTime>080000 ORDER BY 1. NodeId. (AmpTotalUserExec+AmpTotalUserServ) /Secs/NCPUs (FORMAT ‘zz9%’. TheTime.

. execution bwithin the Parser partition of the PE. the PE CPU utilization summarized by a node grouping.Chapter 15: Resource Usage Macros ResCPUByPE Macros ResCPUByPE Macros Function Macro.ToTime). Pars User Serv% Disp User Serv% Ses User Serv% Misc User Serv% Pars User Exec% Disp User Exec% Reports the percent of time PEs are busy doing user.b execution within the Dispatcher partition of the PE.. Usage Notes For any of these macros the following usage notes apply: • • Logging must be enabled on ResUsageSvpr.Node). ResCPUByPE ResCPUByPEOneNode ResPeCpuByGroup Reports. Output Examples The reports in the following sections are sample output reports from the ResCPUByPE. EXEC ResCPUByPE (FromDate. 186 Resource Usage Macros and Tables . except Partition 0) PE partitions. EXEC ResCPUByPEOneNode (FromDate.. FromTime. See “Executing Macros” on page 32 for a description of the FromDate..ToNode).. how each PE on each node is utilizing the CPUs.FromTime.ToTime..ToDate.FromTime.a service for the Dispatcher partition of the PE. servicea for the Parser partition of the PE..ToDate. where: Column.FromTime. ResCPUByPEOneNode.FromNode. service for miscellaneous (all other. FromNode. how each PE on a specific node is utilizing the CPUs. Input Format Examples The input forms of the these three macros are described below. Name the node log rate.ToTime.. ToNode and Node parameters. respectively. service for the Session Control partition of the PE. EXEC ResPeCpuByGroup (FromDate. ToDate. and ResPeCPUByGroup macros.ToDate. ToTime.

00% Note: The NodeId column only appears in the ResCPUByPE output report.00% 0.00% 0. Ses User Exec% Misc User Exec% Total User Serv% Total User Exec% Total Busy% Reports the percent of time PEs are busy doing user.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0. This is the sum of the four user service columns above plus PE Partition 0 user service. which is privileged work performing system-level services on behalf of user execution processes that do not have root privileges.00% 0.00% 0.-----.------.00% 0.------.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0. This is the sum of the four user execution columns above plus PE Partition 0 user execution.00% 0.00% 0.00% 0.00% 0.00% 0. the Total Busy % should be compared to a theoretical maximum of 100% times the number of CPUs on the node.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0. Note: The above CPU statistics represent the aggregate of all time spent in the indicated way by all CPUs on the node.00% 0. Execution is the time a CPU is busy executing user execution code.00% 0.00% 0.00% 0.00% 0.00% 0.00% 16383 001-01 4 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0. a.------.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.. service and execution work.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.------01/07/12 09:57:00 16382 001-01 4 0. The Node CPU column in the following sample outputs reports the number of CPUs (NCPUs).00% 0.00% 0.00% 0.00% 0..00% 0.00% 0.00% 0.00% 0.00% 0.00% 0. ResCPUByPE Sample Output 01/07/12 CPU USAGE BY PE Misc User Exec% ------0.00% 0.00% 0. execution within the Session Control partition of the PE.00% 0.00% 0..------.-----.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.Chapter 15: Resource Usage Macros ResCPUByPE Macros Column.------0.00% 0.------.00% 0.00% 0.00% 0.----.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.-------.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0. execution work.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% Page 1 Total Total Total User User Busy Serv% Exec% % -----.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 09:57:20 16382 001-01 16383 001-01 09:57:40 16382 001-01 16383 001-01 09:58:00 16382 001-01 16383 001-01 09:58:20 16382 001-01 16383 001-01 09:58:40 16382 001-01 16383 001-01 09:59:00 16382 001-01 16383 001-01 09:59:20 16382 001-01 16383 001-01 09:59:40 16382 001-01 16383 001-01 10:00:00 16382 001-01 16383 001-01 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 0.00% 0.00% 0.00% 0.00% 0. b.00% 0.00% 0.00% 0.00% 0.00% 0. execution within miscellaneous (all other.00% 0.. Because there are multiple CPUs.------.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% Pars Disp Ses Misc Pars Disp Ses Vproc Node Node User User User User User User User Date Time Id Id CPUs Serv% Serv% Serv% Serv% Exec% Exec% Exec% -------. Resource Usage Macros and Tables 187 .00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0. Service is the time that a CPU is busy executing user service code.---.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0. This is the sum of the Total User Serv% and the Total User Exec% columns.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0. service work. which is the time spent in a user state on behalf of a process.00% 0. except Partition 0) PE partitions.00% 0.00% 0.

00% 0. SEL TheDate.2% 0.00% 0. 188 Resource Usage Macros and Tables .00% 0.00 Ses User Exec% -----0.00% 0.00% 0.Chapter 15: Resource Usage Macros ResCPUByPE Macros ResCPUByPEOneNode Sample Output 01/09/13 Vproc Node Time Id CPUs ------.00% 0.00% 0.18% 0.05% CPU Usage by PE for Node 001-01 (4 CPUs) Disp Ses Misc Pars User User User User Serv% Serv% Serv% Exec% -----.00% 0.00% Misc User Serv% -----0.00% 0.00% 0.00% 0.00% 0.56% 0.02% 0.13% Total User Exec% -----0.00 0.00% 0.00% 0.00% 0.05% 0.00% Total User Exec% ----0.01% 0.00% 0.01% 0.00% Pars User Exec% -----0.02% 0.00% 0.---15:41:00 2 4 15:42:00 15:43:00 15:44:00 15:45:00 15:46:00 01/08/27 16:21:00 2 2 2 2 2 2 4 4 4 4 4 4 Pars User Serv% -----0.00% 0.00% 0.00% 0.08% 0.00% 0.0% 0.Vproc.----0.00 0.7% 0.00% 0.00% 0.00 0.3.0% 0.00% 0.00% 0.00% 0.00% 0. Normalized Viewing of CPU Usage by PE Some users may prefer to view CPU usage by PEs in a normalized fashion.0% 0.00% 0.00 0.00% Page 8 Total Busy % ----0.00% 0.00% 0.00% 0.0% 0.------01/07/12 04:55:40 A 4 0. this restates each of the above statistics in terms of percentage of total CPU capacity of the node.70% Page 4 Total Busy % -----0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.03% 0.00% 0.00% 0.00 0.00% 0.20% 0. Conceptually.00% 0.55% 0.----.00% 0.58% 0.00% 0.00% 0.0% 0.18% 0.TITLE ‘Total// Busy// %’) FROM ResCpuUsageByPEView WHERE TheDate = CURRENT_DATE AND TheTime>080000 ORDER BY 1.00% 0.00% 04:56:00 04:56:20 04:56:40 04:57:00 04:57:20 04:57:40 04:58:00 04:58:20 A A A A A A A A 4 4 4 4 4 4 4 4 0.04% 0.00% Misc User Exec% -----0.00% Misc User Exec% ------0.00% 0.00% 0.2% 0.00% Total User Serv% -----0.00% 0.58% 0.00% 0.00% 0.00% 0.00% 0.00% PE CPU USAGE BY GROUP Ses User Serv% ------0.18% 0.00 0.56% 0.00% 0.00% 0.6% 0.00% 0.00% 0.00% 0. TheTime.00% 0.0% 0.00% 0.00% 0.00% 0.00% Disp User Serv% -----0.----.08% 0.69% Disp User Exec% -----0.58% 0.01% 0.00% 0.00% 0.8% Date -----01/08/21 ResPeCpuByGroup Sample Output 01/07/12 Pars Group Node User Date Time Id CPUs Serv% ----------.00% Ses User Exec% -----0.20% 0.00% 0.05% 0.00% 0.02% 0.00% 0.00% 0.00% 0.00% 0.00% 0. The following SQL example shows how to perform this normalization for the Total Busy % statistic.01% 0.00% 0.00% Note: The GroupID column only appears in the ResPeCpuByGroup output report.---.00% Total User Serv% -----0.00% 0.00% 0.00% 0.0% 0.00% 0. (PETotalUserExec+PETotalUserServ) /Secs/NCPUs (FORMAT ‘zz9%’.02% 0.00% 0.02% 0.00% 0.02% 0.6% 0.00% 0.00% 0.02% 0.00% 0.00% 0.00 0.NodeId.-----.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.2% 0.00% 0.00% 0.00% 0.00% 0.05% 0.00% 0.00% 0.0% Disp User Exec% -----0.00% 0.00 0.03% 0.00% 0.2.0% 0.02% 0.-----.00% 0.18% 0.00% 0.00% 0.

ToTime).ToDate. a specific node is utilizing its CPUs... ToTime. EXEC ResCPUOneNode (FromDate. ToNode and Node parameters.ToTime.FromTime. EXEC ResCPUByNode (FromDate.FromTime. ResCPUByNode ResCPUOneNode ResCPUByGroup Reports how. Resource Usage Macros and Tables 189 . and the ResCPUByGroup.FromNode. FromNode. FromTime.FromTime.ToDate.Chapter 15: Resource Usage Macros ResCPUByNode Macros ResCPUByNode Macros Function Macro.Node). The following columns are the averages for all CPUs on the node. ToDate. Usage Notes For any of these macros the following usage notes apply: • • Logging must be enabled on ResUsageSpma. the ResCPUOneNode macro. each individual node is utilizing its CPUs.ToNode). EXEC ResCPUByGroup (FromDate.ToTime. See “Executing Macros” on page 32 for a description of the FromDate. a specified Node Group is utilizing the system CPUs. Name the node log rate. Input Format Examples The input forms of these three macros are described below... Output Examples The reports in the following sections are sample output reports from the ResCPUByNode.ToDate.

6% 1..1% 1.5% 1.-----. This is the sum of Total User Serv % and the Total User Exec % columns. which is the time spent in a user state on behalf of a process.2% 15.5% 1.6% 0.5% Total User Exec% -----0.2% 1.7% Page 45 Date -------01/07/12 Time -------09:51:40 09:52:00 09:52:20 09:52:40 09:53:00 09:53:20 Note: The NodeId column only appears in the ResCPUByNode output report.2% 1.3% 0.1% 15.1% 0. I/O Wait % Total User Serv % Total User Exec % Total Busy % Lists percentage of time spent .9% 1. ResCPUByNode Sample Output 01/07/12 CPU USAGE BY NODE Node Id -----001-01 001-01 001-01 001-01 001-01 001-01 I/O Wait % -----16. busy doing user service and execution work.2% 0.5% 16.. busy doing user service work.2% 1. where: This variable… User service User execution Describes the time a CPU is busy executing… user service code.1% 0.3% 1.2% 0.7% 0.2% Total Busy % -----1. user execution code. which is privileged work performing system-level services on behalf of user execution process that do not have root privileges.5% 1.3% 1.4% 16. idle and waiting for I/O completion.5% Total User Serv% -----1.2% Total Total Total User User Busy Serv% Exec% % -----.0% 17.6% 16.2% 1.9% 1.2% 0.3% 0.2% 1.1% 0.7% 1.2% 16.5% 1.6% 0. busy doing user execution work..5% 17.0% 1.8% 15..9% 16.2% 1.-----1.2% 1.4% 1.2% 17.8% Page 01 Date -------01/07/12 Time -------09:44:20 09:44:40 09:45:00 09:45:20 09:45:40 09:46:00 09:46:20 190 Resource Usage Macros and Tables .8% 1.2% 0.Chapter 15: Resource Usage Macros ResCPUByNode Macros This column .2% 1.5% 1. ResCPUOneNode Sample Output 01/07/12 CPU Usage for Node 001-01 I/O Wait % -----16.

0% 0.0% 0.2% 2.0% 0.0% 0.3% 1.0% 0.0% 0.0% 0.2% 0.0% 0.0% 0.8% 1.3% 0.9% 1.4% 2.6% 0.0% A B A B A B A B A B A B A B A B A B A B A B A B A B A B 0.1% 1.9% 1.9% 0.6% 0.0% 0.0% 0.0% 0.1% 0. Resource Usage Macros and Tables 191 .0% 0.0% 0.7% 1.9% 0.6% 0.9% 0.8% 0.6% 0.6% 1.6% 0.0% 0.0% 0.2% Page 2 Date -------00/10/16 Time -------11:25:00 B 11:30:00 11:35:00 11:40:00 11:45:00 11:50:00 11:55:00 12:00:00 12:05:00 12:10:00 12:15:00 12:20:00 12:25:00 12:30:00 12:35:00 Note: The GroupID column only appears in the ResCpuByGroup output report.8% 1.0% 0.2% 2.0% 0.3% 1.9% 1.7% 2.6% 1.6% 1.0% 0.9% 1.3% 1.0% 0.-----.0% 0.7% 0.6% 0.0% 0.-----A 0.0% 1.1% 0.0% 0.0% 0.5% 1.6% 0.0% 0.2% 1.0% 0.1% 1.8% 1.2% 0.8% 1.8% 0.6% 0.1% 0.7% 2.0% 0.0% 0.2% 1.6% 0.5% 1.1% 0.4% 1.2% 1.9% 0.8% 0.-----.4% 1.0% 0.9% Total Busy % -----0.0% 0.0% 0.0% 0.0% 0.6% 1.5% 0.0% 0.4% 0.3% 2.5% 0.2% 1.8% 1.0% 0.0% 0.0% 0.0% 0.6% 1.0% 0.0% 0.0% 0.9% 0.Chapter 15: Resource Usage Macros ResCPUByNode Macros ResCPUByGroup Sample Output 00/10/16 CPU USAGE BY Group I/O Total Total Group Wait User User Id % Serv% Exec% ----.0% 1.6% 2.1% 0.3% 1.5% 1.3% 0.6% 2.0% 0.6% 0.9% 0.0% 0.6% 1.7% 0.

when doing an export.FromTime.FromTime. the ResUsageSpma table may show 30 million rows/log period..ToDate. ToTime. See “Executing Macros” on page 32 for a description of the FromDate. For any of these macros the following usage notes apply: • • Logging must be enabled on ResUsageShst.. and Node parameters. how high are they? If not high. EXEC ResHostByLink (FromDate. if there is still a traffic problem.. EXEC ResHostOneNode (FromDate. every communication link in the system. Usage Notes The ResHostByLink macros help you answer the following questions: • • Is my set up correct? Am I making good use of the channels? If not. consider studying the LAN traffic. for example.Chapter 15: Resource Usage Macros ResHostByLink Macros ResHostByLink Macros Function Macro. Study the incoming traffic. Note: The ResHostByLink macro syntax does not include the FromNode and ToNode parameters to specify a range of nodes.FromTime. the communication links of a specific node. EXEC ResHostByGroup (FromDate. then there may not be enough host resources. Problems with incoming traffic may be simply caused by an incorrect configuration.. ToDate.ToTime. 192 Resource Usage Macros and Tables .ToDate. Name the node log rate. FromTime. Once configured correctly. the communication links of a node grouping. Input Format Examples The input forms of these three macros are described below. ResHostByLink ResHostOneNode ResHostByGroup Reports the host traffic for.ToTime).ToTime).Node).ToDate.

number of KBs written per second. number of successful blocks written per second. for channel-connected hosts KBs Read/ Sec KBs Write/ Sec Blks Read/ Sec Blks Write/ Sec Blk Read Fail % Blk Write Fail % KBs/Blk Read KBs/Blk Write Msgs/Blk Read Msgs/Blk Write Avg ReqQ Len Max ReqQ Len number of KBs read per second. average number of messages per block written. maximum number of messages queued for output to the host.. and the ResHostByGroup. for LAN-connected hosts • IBMMUX. number of successful blocks read per second.. percentage of block read attempts that failed. average number of messages per block read. average number of KBs per block read. Host Type Reports the. average number of messages queued for output to the host. where: Column.Chapter 15: Resource Usage Macros ResHostByLink Macros Output Examples The reports in the following sections are sample output reports from the ResHostByLink. average number of KBs per block written. percentage of block write attempts that failed.. the ResHostOneNode macros. type of host connection: • NETWORK.. respectively. Resource Usage Macros and Tables 193 .

0 23.0 0.4 0.1 0.0 ? 0.9 1.9 1.0 0.0 0.7 5317.0% 0.0 250605.1 Blk Read Fail% ----0.0 0.0 0.4 0.0% KBs KBs Msgs Msgs /Blk /Blk /Blk /Blk Read Write Read Write -----.-----00/10/16 11:07:00 105-04 65535 NETWORK 0 24.0% ? 0.0 0.ResHostByLink Sample Output 00/10/16 HOST COMMUNICATIONS BY COMMUNICATION LINK Blks Blks Read Write /Sec /Sec ----.9 ? ? 1.0 0.4 0.8 0.6 408.0 0.0 0.3 97.0 0.0 31.2 5276.0 13.0 0.0 0.5 ? ? 0.8 246270.0 23.4 1720.----.1 0.5 ? ? ? ? ? 200.0 0.0% ? ? ? ? ? 0.3 34.0% 0.0 0.0% 0.3 22.0% ? 0.6 ? 206.0 0.0 0.----.1 0.0 0.5 ? ? 1037.0 0.0 0.9 ? ? ? 0.1 11:22:42 105-04 65535 NETWORK 0 81105.0 ? ? ? ? ? 385.9 ? ? 1.0 0.0 0.0 0.0% ? ? ? ? 0.3 0.2 ? 391.0 Note: The NodeId column only appears in the ResHostByLink output report.0 81.8 ? ? ? 198.0 0.1 22.0% ? ? 0.1 0.0 0.0 0.1 0.0 ? ? ? ? ? 0.0 0.9 8.0% 0.0% 0.0 0.0 0.0 0.0 0.5 0.1 47.0 0.0 0.0 0.0 0.----.0 0.0 0.0 0.0 0.0 0.0 11.0 0.5 186.0 0.0% 0.5 97.0 ? 0.2 0.0 ? ? ? ? ? 0.8 205.-----.----350.1 0.0% ? ? ? ? 0.0% ? ? ? ? ? 0.6 0.6 46.0 0.0 0. .0 0.0 1 194 Resource Usage Macros and Tables KBs KBs Node Vproc Host Host Read Write Date Time Id Id Type Id /Sec /Sec ------.0 0.0 0.2 176635.8 ? ? 85.0 0.0 0.1 11:42:42 105-04 65535 NETWORK 0 59002.0 0.9 1.0% 0.0 0.5 ? ? 407.6 0.0 0.0 0.0 0.0 0.0 22.----.0 0.0 0.8 0.0% 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 391.0 0.0% ? 0.5 ? 84.0 0.0 Page Max ReqQ Len ---0.0 0.3 5148.0 0.0 0.3 ? ? 0.3 1722.7 ? ? 215.0 0.046.4 86.0 105-05 65535 NETWORK 0 IBMMUX 202 IBMMUX 304 106-04 65535 NETWORK 106-05 65535 NETWORK 0 0 44.0 0.0 0.0% Blk Write Fail% ----0.0 0.0 0.9 1.0 0.0 11.0 0.0 1.2 0.0 0.0 0.0 0.0 0.0% ? ? 0.4 ? 86.0 0.1 0.0% ? ? 0.0 0.1 0.1 0.0 1721.0 0.0 0.-----.0% ? 0.0 0.0 0.7 0.------.5 917.1 0.1 IBMMUX 101 0.8 47.9 0.0 46.0 0.1 11:32:42 105-04 65535 NETWORK 0 IBMMUX 101 105-05 65535 NETWORK 0 IBMMUX 202 IBMMUX 304 106-04 65535 NETWORK 106-05 65535 NETWORK 0 0 80303.----0.0 0.0% 0.0 1.3 34.0 ? 0.9 Avg ReqQ Len ----0.8 ? ? ? ? ? ? ? 398.3 106-05 65535 NETWORK 0 7.0% 0.1 0.0% ? ? 0.9 1.0 0.8 1097.0 105-05 65535 NETWORK 0 IBMMUX 202 IBMMUX 304 106-04 65535 NETWORK 106-05 65535 NETWORK 0 0 0.0 0.0 0.0 0.0 0.1 0.3 IBMMUX 101 0.9 0.9 ? ? ? 0.0 0.0 0.0 0.0% 0.9 1.0 ? 0.0 0.0% 0.3 IBMMUX 101 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 105-05 65535 NETWORK 0 IBMMUX 202 IBMMUX 304 106-04 65535 NETWORK 0 0.9 11.------.0 0.1 0.7 0.0 0.

0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 44.0 0.0 0.0 0.0 0.0 0.0 0.1 0.0 0.0 Blks Write /Sec ----0.0 0.9 ? ? ? ? ? 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.----.0 0.0 0.0 0.0 Blk Read Fail% ----? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? Blk Write Fail% ----? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? KBs /Blk Read ------? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? KBs /Blk Write ------? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? Msgs /Blk Read ----? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? Msgs /Blk Write ----? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? Avg ReqQ Len ----0.0 0.0% ? ? 0.0 0.0 0.0 0.0 0.0 0.0% ? ? ? ? ? 0.0 0.0 0.0 0.0 0.0 0.0 0.1 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.-------0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.----.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.3 0.0 0.0 0.0% ? ? ? ? ? 0.0 0.0 0.0 0.0 0.0 0.0 0.----.0 0.0 0.-------.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.ResHostOneNode Sample Output 00/10/16 Vproc Host Host Date Time Id Type Id -------.----00/10/16 11:07:00 65535 NETWORK 0 IBMMUX 202 IBMMUX 304 11:22:42 65535 NETWORK IBMMUX IBMMUX 11:32:42 65535 NETWORK IBMMUX IBMMUX 11:42:42 65535 NETWORK IBMMUX IBMMUX 0 202 304 0 202 304 0 202 304 Host Communications for Node 105-05 KBs KBs Read Write /Sec /Sec -------.0 Max ReqQ Len ----0.0 0.0 0.0 0.0 0.0 0.0 0.-------.0 0.0 0.0 ? ? ? 0.4 0.0 Blks Blks Blk Blk KBs Read Write Read Write /Blk /Sec /Sec Fail% Fail% Read -----.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.8 ? ? ? ? ? 86.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 22.0 0.9 ? ? ? ? ? 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.-----.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0. .0 0.0 0.0 0.0 0.0 0.0 ? ? ? 0.0 0.0 0.0 0.0 0.0 0.0 Note: The GroupID column only appears in the ResHostByGroup output report.0 0.0 0.0 0.0 0.0 0.0 0.5 ? ? Page Msgs /Blk Write ----? ? ? 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 ? ? ? 0.0 0.0 0.0 0.0 0.0 0.0 Resource Usage Macros and Tables 195 ResHostByGroup Sample Output Date -------00/10/16 00/10/16 00/10/16 00/10/16 00/10/16 00/10/16 00/10/16 00/10/16 00/10/16 00/10/16 00/10/16 00/10/16 00/10/16 00/10/16 00/10/16 00/10/16 00/10/16 00/10/16 00/10/16 00/10/16 00/10/16 00/10/16 00/10/16 00/10/16 00/10/16 Time -------11:30:00 11:30:00 11:35:00 11:35:00 11:40:00 11:40:00 11:45:00 11:45:00 11:50:00 11:50:00 11:55:00 11:55:00 12:00:00 12:00:00 12:05:00 12:05:00 12:10:00 12:10:00 12:15:00 12:15:00 12:20:00 12:20:00 12:25:00 12:25:00 12:30:00 Group Id -----A B A B A B A B A B A B A B A B A B A B A B A B A Host Type -------NETWORK NETWORK NETWORK NETWORK NETWORK NETWORK NETWORK NETWORK NETWORK NETWORK NETWORK NETWORK NETWORK NETWORK NETWORK NETWORK NETWORK NETWORK NETWORK NETWORK NETWORK NETWORK NETWORK NETWORK NETWORK KBs Read /Sec -------0.1 0.0 0.----0.0 0.0 0.0 0.0 KBs Write /Sec -------0.0 0.0 0.5 ? ? KBs /Blk Write ------? ? ? 206.0% ? ? 391.0 Max ReqQ Len ---0.5 ? ? ? ? ? 84.0 0.0 0.0 0.0 0.0 0.6 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.3 ? ? 1 Avg ReqQ Len ---0.0 0.0 0.0 0.0 0.0 0.7 ? ? Msgs /Blk Read ----? ? ? 0.0 Blks Read /Sec ----0.

EXEC ResLdvByNode (FromDate.ToDate. ResLdvByNode ResLdvOneNode ResLdvByGroup Reports the logical device traffic channeled through.ToTime. average number of logical device reads per second. average response time for a logical device read or write in seconds. FromNode. 196 Resource Usage Macros and Tables . Name the node log rate.. a specific node by totaling all its controller links into one summarized node output line. EXEC ResLdvByGroup (FromDate.FromNode. where: Column. ToDate. ToNode and Node parameters.FromTime.Node). Reads / Sec Writes / Sec Rd KB / I/O Wrt KB / I/O Avg I/O Resp Reports the. ToTime..Chapter 15: Resource Usage Macros ResLdvByNode Macros ResLdvByNode Macros Function Macro. and the ResLdvByGroup macros. respectively. FromTime. the ResLdvOneNode..ToTime. average number of KBs per logical device read.FromTime. average number of KBs per logical device write.. Output Examples The reports in the following sections are sample output reports from the ResLdvByNode.FromTime. a node grouping.. Input Format Examples The input forms of these three macros are described below....ToNode).ToDate. See “Executing Macros” on page 32 for a description of the FromDate.ToTime). average number of logical device writes per second.ToDate. EXEC ResLdvOneNode (FromDate. Usage Notes For any of these macros the following usage notes apply: • • Logging must be enabled on ResUsageSldv. each node by totaling its controller links into one summarized node output line.

13 0.6% 1.57 10:12:30 001-01 0.67 31.0 0.48 0.0 0.1% Date -------06/09/26 Ldv Type ---DISK Time -------10:09:45 10:10:00 10:10:15 10:10:30 10:10:45 10:11:00 10:11:15 10:11:30 10:11:45 10:12:00 10:12:15 10:12:30 10:12:45 10:09:45 10:10:00 10:10:15 10:10:30 10:10:45 10:11:00 10:11:15 Reads / Sec -------0.57 107.9% 3.0 0.07 0.0% 1.00 0.53 3.000 0.5% 1.00 0.000 0.30 5.51 Writes / Sec -------2.87 1.8% 2.27 16.28 55.1% 11.000 0.000 0.42 109.88 10:10:30 001-01 0.000 0.00 1.000 0.000 0.33 5.0 0.73 9.72 0.3 1.000 0.0 0.0 0.3 1.000 0.00 2.87 8.0 1.000 0.85 10:11:15 001-01 0.000 0.3% 0.000 0.---.21 8.20 1.1% 9.0 4.-------.48 0.44 0.000 0.3% 0.00 1.97 0.64 100.00 7.0 0.000 0.13 0.000 0.53 1.3% 0.7% 1.53 111.0 1 Rqst Time % -----1.7% 1.30 10:11:30 001-01 0.0 Out Rqst Time % -----1.50 0.000 0.2 1.2% 9.2% 9.1% 11.44 10:12:45 001-01 0.0 0. percent of time there are outstanding requests.29 55.3 0.1% 9.51 0.89 5.1% 9.6% 1.0% 1.45 8.0 0.000 0.00 4.0 0.2 1.20 5.6% Note: The NodeId column only appears in the ResLdvByNode output report.1 0.27 5.85 4.3% 0.000 0.000 0.000 0.0% 1.0 0.64 100.0 0.53 111.20 5.2% 8.00 5..32 98.24 ****** ****** LOGICAL DEVICE TRAFFIC BY NODE Out I/O Resp ------0.32 98.1 1.00 40.45 8.22 3.000 0.87 14.8% 2. average number of outstanding requests.01 8.0% 100.0 1.00 1.000 Avg Out Rqsts ----0.98 9.00 1..000 0.0 1.000 0.33 1.40 7.00 0.000 0.00 0.0 0.00 4.65 8.40 1.0% 1.000 0.0 0.000 0.00 3.2% 0.67 3.9% 3.0 0.73 9.00 0.3% 0.00 0.96 10:11:00 001-01 0.71 10:12:00 001-01 0.40 5.0 0.20 6.01 8.00 2.0 0.33 5.0 0.0 0.00 0.2% 9.98 9.00 1.00 1.00 0.1 1.00 0..000 0.87 40.00 0. maximum number of concurrent requests during the log period.000 0.00 7.0 0.4% 5.3% 0.57 107.000 0.20 3.Chapter 15: Resource Usage Macros ResLdvByNode Macros Column.1 1.000 0.57 7.00 3.00 0.00 1.88 6.1% 9.00 0.0% 100.-------.4% 5.1 1.33 2.07 0.27 10:12:15 001-01 0.000 0. ResLdvByNode Sample Output 06/09/26 Avg Avg Ldv Node Reads Writes KB Date Type Time Id / Sec / Sec / I/O -------.-----06/09/26 DISK 10:09:45 001-01 0.000 0.000 0.000 0.42 109.89 10:10:15 001-01 0.1% 9. Max Concur Rqsts Avg Out Rqsts Out Rqst Time % Reports the.00 2.00 0.000 Page Out Rqsts ----0.000 0.44 14.22 10:10:45 001-01 0.0 0.5% 1.1% 9.000 0.65 10:11:45 001-01 0.000 0.0 0.20 10:10:00 001-01 0.85 2.96 5.0 0.-------.71 31.88 0. ResLdvOneNode Sample Output 06/09/26 LOGICAL DEVICE TRAFFIC FOR NODE 001-01 Avg I/O Resp ------0..27 2.02 100.000 0.34 0.1% 9.83 SDSK Resource Usage Macros and Tables 197 .0 0.00 0.1% 9.29 SDSK 10:09:45 001-01 10:10:00 001-01 10:10:15 001-01 10:10:30 001-01 10:10:45 001-01 10:11:00 001-01 10:11:15 001-01 10:11:30 001-01 10:11:45 001-01 10:12:00 001-01 0.0 1.83 410.2% 0.-----.0 4.02 100.98 1.00 0.21 8.0 0.000 0.50 0.40 16.000 0.85 KB / I/O -----4.2% 9.00 1.44 0.000 0.

000 0.000 0.3% 0.45 8. 198 Resource Usage Macros and Tables .0 0.85 2.000 0.40 1.000 0.14 ? 46.000 0.0 0.44 0.34 0.000 0.30 ? 5.00 0.2% 9.50 0.76 105.2% 9.00 0.48 0.1% 9.28 3.87 1.28 111.15 0.01 8.14 4.3% 0.33 2.40 5.94 0.21 8.000 0.00 1.0 0.65 ? 8.000 0.000 0.00 ? 113.0 0.00 0.1% 9.0 0.67 3.0 0.0% 1.000 0.4% 5.0 0.0 0.00 1.90 39.29 121.00 2.--.33 12.08 0.00 0.---06/09/26 A DISK Time --------10:09:45 10:10:00 10:10:15 10:10:30 10:10:45 10:11:00 10:11:15 10:11:30 10:11:45 10:12:00 10:12:15 10:12:30 10:12:45 10:09:45 10:10:00 10:10:15 10:10:30 10:10:45 10:11:00 10:11:15 10:11:30 10:11:45 10:12:00 10:12:15 10:12:30 10:12:45 Reads / Sec -------0.6% 9.08 ****** ****** ****** ****** 295.1% 9.00 0.00 Writes / Sec -------2.000 0.94 Avg I/O Resp ------0.00 0.93 94.000 0.44 ? 14.2% 8.3% 0.2% 0.98 0.57 107.0 0.000 0.73 9.29 32.88 ? 6.1% 9.00 0.000 0.53 1.0% 1.00 7.85 11.89 ? 5.00 0.0 0.2% 1.1 0.000 0.64 109.1 0.0 0.000 0.000 SDSK Note: The GroupID column only appears in the ResLdvByGroup output report.000 0.27 139.000 0.1 0.00 0.87 40.0 0.88 0.97 0.12 12.43 50.72 0.20 3.000 0.000 0.6% 1.8% 2.1% 9.27 2.20 1.72 0.000 0.0 0.0 0.98 9.9% 3.01 12.00 0.00 0.0 0.34 0.000 0.0 0.1% 9.000 0.97 0.6% 9.000 0.000 0.85 ? 4.98 0.0% ResLdvByGroup Sample Output 06/09/26 LOGICAL DEVICE TRAFFIC BY GROUP Page 1 Max Concur Rqsts ----0.0 0.13 9.0% 100.2% 1.000 0.0 0.51 0.000 0.1% 11.96 ? 5.39 410.0 9.0 Out Rqst Time % -----1.00 0.20 ? 5.15 0.000 0.0 0.59 0.0 0.33 1.72 105.3 0.71 ? 31.000 0.00 0.78 16.88 0.0 0.1% 9.0 0.02 102.24 ****** ****** 285.0 0.07 0.1% 9.2% 8.-----? 4.5% 1.22 ? 3.Chapter 15: Resource Usage Macros ResLdvByNode Macros 10:11:30 10:11:45 10:12:00 10:12:15 10:12:30 10:12:45 0.00 0.08 0.39 Rd KB Wrt KB / I/O / I/O ------.13 0.27 ? 16.14 4.57 ? 7.0 0.000 0.0 0.7% 1.28 3.0% Grp Ldv Date Id Type -------.99 32.0 0.000 0.59 0.

Name the node log rate. ReadCnt/Sec WriteCnt/Sec Rd KB/ I/O Wrt KB/ I/O Avg I/O Resp Reports the. ToDate. EXEC ResPdskByNode (FromDate. by a physical node.. Usage Notes For any of these macros the following usage notes apply: • • Logging must be enabled on ResUsageSpdsk. ResPdskByNode ResPdskOneNode ResPdskByGroup Reports the device traffic. FromTime.ToDate..ToDate..ToTime. EXEC ResPdskOneNode (FromDate. which reports by NodeType column). ToTime. average number of KBs per device read.ToNode). EXEC ResPdskByGroup (FromDate.FromTime. FromNode.. See “Executing Macros” on page 32 for a description of the FromDate. Column. average number of device writes per second...ToTime.FromNode. Resource Usage Macros and Tables 199 ..Node). which reports by NodeId columns. Input Format Examples The input forms of these three macros are described below.FromTime. ToNode and Node parameters.FromTime.ToTime).Chapter 15: Resource Usage Macros ResPdskByNode Macros: Pdisk Device Traffic ResPdskByNode Macros: Pdisk Device Traffic Function Macro. for a specified node.. Output Examples The following table describes the statistics columns in all output reports (with the exception of ResPdiskByNode. average number of KBs per device write. and ResPdiskByGroup. average response time for a device read or write in seconds. node grouping. average number of device reads per second.ToDate.

77 4637.18 0.0% 0.001 0.05 5262.12 0.14 0.0% 0.07 0.11 0.0% 0.67 3856.0% 0.13 0.000 0.34 0.19 0.41 0. Max Concur Rqsts Out Rqst Time % Reports the.001 0.0% 0.001 0.0% 0.59 4532.0% 0.11 0.71 5730.18 0.06 0.0% 0.05 0.40 Note: The NodeId column only appears in the ResPdskByNode output report.45 0.43 6192.17 0.27 3602.0% 0.0% 0.60 6731.0% 0.002 0..15 3623.46 3990.05 0..51 0.04 6260.42 0.28 0. ResPdskByNode Sample Output 07/11/28 PDISK TRAFFIC BY NODE Avg I/O Resp ------0.0% 0.07 0.0% 0.46 5218.19 0.001 0.001 0.71 2560.00 2523.29 4176.37 0.0% 0.002 0.11 0.53 0.63 4943.12 0.002 0.67 5616.06 0.59 5139.002 Page Out Rqst Time % -----0.20 0.002 0.0% 0.0% 0.36 5990.001 0. maximum number of concurrent requests during the log period.002 0.12 4785.002 0..23 4975.64 5274.0% 1 Date -------07/11/28 Pdisk Type -----DISK Time -------13:20:00 13:21:00 13:22:00 13:24:00 13:25:00 13:26:00 13:28:00 13:29:00 13:30:00 13:32:00 13:33:00 13:34:00 13:36:00 13:37:00 13:38:00 13:39:00 13:41:00 13:42:00 13:43:00 13:45:00 13:46:00 13:47:00 Node Id -----001-01 001-01 001-01 001-01 001-01 001-01 001-01 001-01 001-01 001-01 001-01 001-01 001-01 001-01 001-01 001-01 001-01 001-01 001-01 001-01 001-01 001-01 Reads / Sec -------0.55 4758.00 Wrt KB / I/O ------******* 4202.0% 0.13 Writes / Sec -------0.33 4846.36 5546.0% 0.002 0.001 0. percent of time there are outstanding requests.14 3990.003 0.17 0.69 5017.46 3990.85 3396.35 0.Chapter 15: Resource Usage Macros ResPdskByNode Macros: Pdisk Device Traffic Column.11 0.26 4485.23 0.51 5125.33 4874. 200 Resource Usage Macros and Tables .30 5677.17 0.002 0.29 5961.0% 0.14 0.38 Rd KB / I/O ------******* ******* ******* 4717.07 0.40 0.001 0.10 0.001 0.34 5745..05 0.42 0.004 0.0% 0.38 0.49 0.00 6537.59 5218.14 0.41 0.0% 0.59 4660.001 0.

0 0.001 0.0% 0.26 4485.002 0.0% 0.002 Out Rqst Time % -----0.002 0.0% 0.45 0.002 0.38 0.0% Note: The GroupID column only appears in the ResPdskByGroup output report.0 0.001 0.18 0.17 0.0 0.37 0.59 4532.0% 0.00 ? ? 0.07 0.05 0.15 3623.002 0.0% 0.0% 0.00 ? ? 0.36 5990.00 ? ? 0.41 0.002 0.0% Page 1 Date -------07/11/28 Pdisk Type -----DISK Time -------13:20:00 13:21:00 13:22:00 13:24:00 13:25:00 13:26:00 13:28:00 13:29:00 13:30:00 13:32:00 13:33:00 13:34:00 13:36:00 13:37:00 13:38:00 13:39:00 13:41:00 13:42:00 13:43:00 13:45:00 13:46:00 13:47:00 13:49:00 Reads / Sec -------0.0% 0.14 0.00 6537.18 0.00 0.12 0.0% 0.40 5878.05 Writes / Sec -------0.00 ? ? 0.00 3693.001 0.07 0.00 0.55 4758.52 Note: The NodeId column only appears in the ResPdskOneNode output report.06 0.00 0.0 0.0% 0.64 5274.11 0.-----0.0% 0.0% 0.00 ? ? 0.28 0.001 0.00 ? ? 0.002 0.003 0.14 0.0% 0.29 5961.17 0.21 Rd KB / I/O ------******* ******* ******* 4717.001 0.36 5546.000 0.0 0.17 0.41 0.30 5677.00 0.05 5262.00 ? ? 0.00 0.Chapter 15: Resource Usage Macros ResPdskByNode Macros: Pdisk Device Traffic ResPdskOneNode Sample Output 07/11/28 PDISK Traffic for Node 001-01 Avg I/O Resp ------0.10 0.33 4874. ResPdskByGroup Sample Output 06/09/26 PDISK TRAFFIC BY GROUP Avg I/O Resp ------? ? ? ? ? ? ? ? ? ? ? ? ? Page 1 Node Pdisk Date Type Type -------.0% 0.0% 0.00 0.0% 0.38 0.00 0.42 0.00 0.42 0.11 0.0% 0.00 0.85 3396.0% 0.23 0.06 0.0% 0.27 3602.0% 0.0% 0.0% 0.0 0.05 0.49 0. Resource Usage Macros and Tables 201 .05 0.0 0.43 6192.00 0.71 Wrt KB / I/O ------******* 4202.53 0.0% 0.69 5017.19 0.46 3990.001 0.001 0.001 0.00 ? ? 0.13 0.0 0.13 0.0% 0.001 0.11 0.34 0.0 0.0% 0.0% 0.002 0.0 0.00 ? ? 0.0% 0.14 3990.0% 0.0% 0.04 6260.77 4637.46 3990.40 0.14 0.0% 0.00 ? ? 0.59 5218.00 ? ? 0.00 0.-----06/09/26 4400 DISK Time -------10:09:45 10:10:00 10:10:15 10:10:30 10:10:45 10:11:00 10:11:15 10:11:30 10:11:45 10:12:00 10:12:15 10:12:30 10:12:45 ReadCnt WriteCnt Rd KB Wrt KB / Sec / Sec / I/O / I/O -------.------.71 2560.51 0.23 4975.001 0.20 0.34 5745.0% 0.60 6731.------0.71 5730.63 4943.11 0.29 4176.00 2523.00 0.0 0.0% 0.0% 0.19 0.002 0.59 4660.33 4846.---.59 5139.-------.004 0.12 0.00 ? ? 0.002 0.35 0.51 5125.00 ? ? Max Out Concur Rqst Rqsts Time % -----.00 0.12 4785.46 5218.67 3856.67 5616.0% 0.0 0.07 0.0% 0.0 0.

ToDate. where: Column. ToTime. each individual node. average number of text page allocations per second.FromTime.ToDate..FromTime. FromNode. a node grouping.ToDate.ToTime.FromNode.Node).ToTime.ToNode). VPR Alocs/ Sec average number of vproc-specific page and segment allocations per second on a node. Input Format Examples The input forms of these three macros are described below.. and the ResMemByGroup. % Mem Free Text Alocs/ Sec Reports the. EXEC ResMemMgmtOneNode (FromDate.FromTime.. FromTime... EXEC ResMemByGroup (FromDate.Chapter 15: Resource Usage Macros ResMemMgmtByNode Macros ResMemMgmtByNode Macros Function Macro.. ToNode and Node parameters. Usage Notes For any of these macros the following usage notes apply: • • Logging must be enabled on ResUsageSpma.ToTime). See “Executing Macros” on page 32 for a description of the FromDate.. 202 Resource Usage Macros and Tables . EXEC ResMemMgmtByNode (FromDate. Name the node log rate. the ResMemMgmtOneNode macros. Output Examples The reports in the following sections are sample output reports from the ResMemMgmtByNode. ResMemMgmtByNode ResMemMgmtOneNode ResMemByGroup Reports memory management activity for. current snapshot of the percent of memory that is unused. ToDate. a specific node.. text pages are allocations of memory for code that is not associated with system-level overhead tasks. respectively.

Page drops are text pages that are dropped from memory to increase the amount of available memory. average number of paged. Page writes include only task context pages. and the disk segments (Swp Drps/ Sec). Includes both the memory text pages and task context pages (Pg Rds/ Sec). such as scratch. that were dropped. reread back into memory after being swapped out. average size. of reread disk segments that had been previously dropped from memory because their ancestor processes were swapped out. swapped page. after being swapped. or segment reads per second. and the disk segments (Swp Rds/ Sec)... or segment writes per second. or segment drops per second. P+S Writes/ Sec P+S IO % Resource Usage Macros and Tables 203 . average number of text pages dropped from memory per second. out per second. average total number of paged. average number of times memory pages were aged out per second. current number of processes that are swapped out. percent of memory allocation attempts that failed. Swap reads include all reread disk segments that had been previously dropped from memory because their ancestor processes were swapped out. Swap Reads/ Sec average number of disk segments reread back into memory. swapped page. Swap Drops/ Sec average number of disk segments dropped from memory per second. swapped page. Page Reads/ Sec average number of memory pages read from disk per second. Swap drops include all disk segments dropped from memory because their ancestor processes were swapped out. This statistic includes both the memory text pages (Pg Drps/ Sec).. of disk segments dropped from memory because their ancestor processes were swapped out. in KBs. in KBs. Page reads include both memory text pages and task context pages. percent of total logical device inputs and outputs that are paging or swapping inputs and outputs. Page Writes/ Sec average number of memory pages written to disk per second. stack. P+S Drops/ Sec P+S Reads/ Sec average number of paged..Chapter 15: Resource Usage Macros ResMemMgmtByNode Macros Column. KB/Swp Drp KB/Swp Rd average size. KB/ VPR Aloc Aloc Fail % Ages/ Sec # Proc Swp Page Drops/ Sec Reports the. and so on. average KBs per vproc-specific page and segment allocation on a node.

---.0 0.0 0.0 2 0.----.8 0.0 1 0.0 0.0 ? ? 0.---.0 5 2.0 0.---.-----.----.0 0.0 0.0 ? ? 0.-----.0 ? ? 0.0 89.0 0.---.----.0 ? ? 0.0 90 46.0 0.----.8 97% 001-01 15% 0.-------08/09/29 11:35:00 11:36:00 11:37:00 11:38:00 Note: The GroupID column only appears in the ResMemByGroup output report.3 3.---.0 0.----.0 0.0 ? ? 0. .-----.---.0 ? ? 0 5 2 73% Date Time -------.---15% 0.---.---A 15% 0.---.0 0.---.0 0 0.0 0 0.0 0.0 0.6 1.----.0 ? ? 0.0 53 1.----.-----.0 0. ResMemMgmtOneNode Sample Output 08/09/29 Memory Management Activity for Node 001-01 Page 1 % Text VPR KB Aloc # Page Page Page Swap Swap KB KB P+S P+S P+S Mem Alocs Alocs /VPR Fail Ages Proc Drops Reads Wrts Drops Reads /Swp /Swp Drops Reads Writes P+S Free /Sec /Sec Aloc % /Sec Swap /Sec /Sec /Sec /Sec /Sec Drp Rd /Sec /Sec /Sec IO % ---.-----.4 0.4 0.---.0 50% A 16% 0.0 90 46.0 ? ? 0 2 0 50% 16% 0.----.3 3.0 0.-----.0 2.---.0 0 0.---001-01 15% 0.0 ? ? 0 53 2 97% 15% 0.0 0.0 0.------.8 0.0 ? ? 0.0 52.-----.4 0.0 2 0.0 ? ? 0.0 0.---.0 0.0 2.------.0 ? ? 0.0 0.------.8 46.0 0.0 10.-----.0 ? ? 0.0 10.5 92% Date Time -------.4 0.0 ? ? 0.0 0 0.---.0 ? ? 0.1 0.0 0.-------08/09/29 11:35:00 11:36:00 11:37:00 Note: The NodeId column only appears in the ResMemMgmtByNode output report.0 90 46.6 1.----.0 53 1.0 89.0 ? ? 0 1 0 45% 16% 0.0 ? ? 0 90 46 97% 16% 0.---.---.-----.0 0 0.-----.0 ? ? 0.0 ? ? 0.8 97% A 15% 0.0 53 1.0 0 0.0 0.0 10 3.---.0 0.0 ? ? 0.---.0 0 0.0 0.0 0 0.0 0 0.0 ? ? 0.---.----.0 0.8 0.0 52.0 ? ? 0.0 ? ? 0.0 0.-----.------.0 0 0.0 0.0 2 0.----.0 0 0.---.0 50% 001-01 16% 0.0 ? ? 0.-----.1 0.0 0.0 ? ? 0.0 0.5 0.0 0 0.---.4 97% A 16% 0.8 46.0 0 0.-----.ResMemMgmtByNode Sample Output 08/09/29 MEMORY MANAGEMENT ACTIVITY BY NODE Page 1 % Text VPR KB Aloc # Page Page Page Swap Swap KB KB P+S P+S P+S Node Mem Alocs Alocs /VPR Fail Ages Proc Drops Reads Wrts Drops Reads /Swp /Swp Drops Reads Writes P+S Id Free /Sec /Sec Aloc % /Sec Swap /Sec /Sec /Sec /Sec /Sec Drp Rd /Sec /Sec /Sec IO % -----.-------08/09/29 11:35:00 11:36:00 11:37:00 11:38:00 11:40:00 11:41:00 ResMemByGroup Sample Output 08/09/29 MEMORY MGMT ACTIVITY BY GROUP Page 1 % Text VPR KB Aloc # Page Page Page Swap Swap KB KB P+S P+S P+S Group Mem Alocs Alocs /VPR Fail Ages Proc Drops Reads Wrts Drops Reads /Swp /Swp Drops Reads Writes P+S Id Free /Sec /Sec Aloc % /Sec Swap /Sec /Sec /Sec /Sec /Sec Drp Rd /Sec /Sec /Sec IO % ----.----.----.4 97% 204 Resource Usage Macros and Tables Date Time -------.0 ? ? 0 10 3 92% 16% 0.0 0.

where: Column. the routing path was busy. Total Reads/ Sec average number of net reads per second. A value over 100% does not imply a problem. See “Executing Macros” on page 32 for a description of the FromDate. Resource Usage Macros and Tables 205 .ToTime.ToTime).FromTime. percent of total net circuit attempts that caused software backoffs (BNS service-blocked occurrences). a specific node. the ResNetOneNode.. Output Examples The reports in the following sections are sample output reports from the ResNetByNode. ToDate.ToDate.. Name the node log rate.. ToTime. EXEC ResNetByGroup (FromDate. each node.. Usage Notes For any of these macros the following usage notes apply: • • Logging must be enabled on ResUsageSpma. ResNetByNode ResNetOneNode ResNetByGroup Reports net traffic for. but shows that there were multiple attempts to send new messages while the Bynet path was busy. Input Format Examples The input forms of these three macros are described below.ToDate. % Retries Reports the..ToDate.ToTime.FromNode. That implies that the end point was busy or.Node). FromNode.FromTime.. EXEC ResNetByNode (FromDate. ToNode and Node parameters. and the ResNetByGroup macros. respectively.FromTime. nodes summarized by node groups. this can be a normal level of activity.. in switch node terms.ToNode). Note: This value reflects how many times the hardware backed off a connection because the switch nodes could not route to the end point. EXEC ResNetOneNode (FromDate. FromTime. On a busy system..Chapter 15: Resource Usage Macros ResNetByNode Macros ResNetByNode Macros Function Macro.

51 0.07 35.18 20.39 0.02 0.71 12.6% 15.4 0.5 0.44 0.48 0.-----.08 2.40 0.52 5.47 11.------.92 22.38 0.0% 0.0% 0.85 5.Chapter 15: Resource Usage Macros ResNetByNode Macros Column.51 0.97 0.4 0.02 33.-----001-03 0.------.-------2000/10/16 11:20:00 11:25:00 11:30:00 11:35:00 11:40:00 11:45:00 11:50:00 % PtP --92 93 90 90 91 92 73 74 93 93 76 83 77 75 % Brd --8 7 10 10 9 8 27 26 7 7 24 17 23 25 ResNetOneNode Sample Output 00/10/16 Net Activity for Node 001-03 % Retries -----0.0% 0.9 Page 2 Date Time ---------.18 24.4 001-04 0.90 0.0% 0.63 14.. vnet) systems.87 3. Note: In the following examples.54 Total Writes /Sec ------0.4 0.87 10.17 0.07 2.4 13. average KBs per net read or write.8 8.95 1.52 0.47 0.85 1.5% 2.47 12.4 0.78 2.3 1.0% 0.0% 0.0% 0. the values in the Total Reads/ Sec and Total Writes/ Sec are expected to be equal on SMP (single-node.1% 0.7% 20.43 1.85 0.0% 0.42 0.97 0.4 001-03 001-04 001-03 001-04 001-03 001-04 001-03 001-04 001-03 001-04 001-03 001-04 0.6 0.16 13.07 2.3% 4.0% 0.4 Page 1 % PtP --8 96 80 98 98 84 93 93 92 93 92 94 79 93 % Brd --92 4 20 2 2 16 7 7 8 7 8 6 21 7 Date -------00/10/16 Time -------10:19:00 10:20:00 10:21:00 10:22:00 10:23:00 10:30:00 10:35:00 10:40:00 10:45:00 10:50:00 10:55:00 11:00:00 11:05:00 11:10:00 206 Resource Usage Macros and Tables .56 38.01 KB /IO -----1. average number of net reads and writes per second.58 0.47 1.03 24.39 0.37 0.9% 28.0% Total Reads /Sec ------0.81 1.44 0.5 0. The GroupID column only appears in the ResNetByGroup output report.71 0..33 0. For all the examples.74 0.5 0.00 76.5% 24.77 0.79 18.42 33.4 1.53 0.39 0.13 2..07 2.4 0.4 0.55 0.15 4.18 0.5 0.0% 0.53 15.44 37.55 20.45 0.47 Total IOs /Sec ------0.0% 0.65 2.8 1.5 0.23 0.4 0.4 0.0% 0. ResNetByNode Sample Output 00/10/16 NET ACTIVITY BY NODE Total Total Total Node % ReReads Writes IOs KB Id tries /Sec /Sec /Sec /IO -----. Total Writes/ Sec Total IOs/ Sec KB/ IO % PtP % Brd Reports the.1% 3.39 0.46 0..72 0.0% 0.------.46 0.3 12.95 0. percent of total net reads and writes that are broadcast reads and writes.11 43.0% 0.32 0.8 1. average number of net writes per second.5 0.00 0.09 0.57 0.1 0.0% 0.25 0.0% 2.8 17.0% 0.84 23. percent of total net reads and writes that are point-to-point reads and writes.51 35.37 38.58 71. the NodeId column appears only in the ResNetByNode output report.0% 0.20 4.0% 0.

4 0.8 17.8 1.4 13.94 31.0 1.35 19.3 12.65 33.4% 3.00 76.84 23.8% 57.0% 0.3 17.7% 34.9% Reads Writes /Sec /Sec ------.01 44.4 93 7 0.47 12.6 5.16 0.4 1.42 33.11 22.10 14.55 20.63 14.70 IOs /Sec ------1.0% 0.18 24.3 1.13 33.37 13.1% 3.16 61.83 70.46 0.87 10.46 88.4 92 92 8 8 ResNetByGroup Sample Output 00/10/16 NET ACTIVITY BY Group Group % ReTime Id tries -------.2 23.09 26.02 0.9% 28.40 0.56 38.97 33.7% 20.77 40.0% 11:25:00 11:30:00 11:35:00 11:40:00 11:45:00 11:50:00 11:55:00 12:00:00 12:05:00 12:10:00 12:15:00 12:20:00 A B A B A B A B A B A B A B A B A B A B A B A B 0.7 14.6% 73.33 0.9 22.7 2.0% 0.97 17.77 37.3% 4.74 73.1% 55.8 18.8% 24.2% 10.4 0.34 54.--.0 1.47 0.37 0.47 11.70 34.45 44.------0.Chapter 15: Resource Usage Macros ResNetByNode Macros 11:15:00 11:20:00 0.5 0.07 35.0 13.0% 2.5 23.16 13.2 90 90 91 92 73 74 93 93 76 83 77 75 95 96 73 70 90 91 91 93 73 73 95 93 10 10 9 8 27 26 7 7 24 17 23 25 5 4 27 30 10 9 9 7 27 27 5 7 Date ---------2000/10/16 Resource Usage Macros and Tables 207 .72 0.39 0.18 20.85 0.83 71.03 28.16 11.39 0.8% 41.44 0.27 33.51 35.4 0.5% 2.02 22.1% 5.0 2.57 25.81 1.37 38.--0.6% 15.01 26.33 28.3% 48.55 0.13 36.86 Page 2 KB % % /IO PtP Brd -----.71 0.67 52.92 22.86 0.1% 5.-----11:20:00 B 0.65 16.02 33.71 12.53 15.9% 6.0% 0.-----.11 43.79 18.58 71.44 37.12 21.18 38.97 14.39 0.8 8.8 1.12 36.03 24.0% 0.46 0.5% 24.72 32.32 0.

The following table describes the 19 statistics columns. respectively. 208 Resource Usage Macros and Tables .ToNode). node by node.ToTime)..ToDate. EXEC ResOneNode (FromDate. the ResOneNode.FromTime...Chapter 15: Resource Usage Macros ResNode Macros ResNode Macros Function Macro.FromTime.ToDate. and the ResNodeByGroup macros. Note: The ResNode macro syntax does not include the FromNode and ToNode parameters to specify a range of nodes.Node). after the Date and Time columns. See “Executing Macros” on page 32 for a description of the FromDate.ToDate.ToTime. ResNode ResOneNode ResNodeByNode ResNodeByGroup Provides a summary of resource usage.ToTime.FromTime. Output Examples The reports in the following sections are sample output reports from the ResNode.ToDate. FromTime. ToTime.FromNode. ToNode. Input Format Examples The input forms of these four macros are described below. EXEC ResNodeByNode (FromDate.. Name the node log rate. averaged across all nodes. FromNode. ToDate. for a node grouping. and Node parameters.FromTime.ToTime). EXEC ResNodeByGroup (FromDate. Usage Notes For any of these macros the following usage notes apply: • • Logging must be enabled on ResUsageSpma. in the ResNode output report. for a specific node. the ResNodebyNode. EXEC ResNode (FromDate.

Resource Usage Macros and Tables 209 . Net interface. Logical device interface. Logical device interface. after the Date and Time columns. Statistics columns 1 and 2 3 through 6 7 through 12 13 through 15 16 and 17 Description CPU usage. Memory interface. The following table describes the 17 statistics columns. Memory interface. Statistics columns 1 and 2 3 through 6 7 through 11 12 through 14 15 and 16 Description CPU usage.Chapter 15: Resource Usage Macros ResNode Macros Statistics columns 1 through 3 4 through 8 9 through 14 15 through 17 18 and 19 Description CPU usage. Net interface. Statistics columns 1 2 and 3 Description GroupId as defined in the associated view as a grouping of one or more nodes. General node process scheduling. Net interface. Memory interface. The following table describes the 17 statistics columns. after the Date and Time columns. The following table describes the 16 statistics columns. Logical device interface. in the ResOneNode output report. General node process scheduling. CPU usage. General node process scheduling. after the Date and Time columns. in the ResNodeByGroup output report. in the ResNodebyNode output report.

Net interface. percent of memory that is unused. Column… CPU Bsy % CPU Eff % (ResNode report) Reports the… percent of time the CPUs are busy. which has a GroupId column). It is the average for all nodes of total busy divided by the total busy time of the busiest node. Average number of times memory pages were aged out per second. percent of logical device reads and writes that are disk segment reads and writes. and ResNodeByGroup. WIO % Ldv IOs /Sec Ldv Eff % (ResNode report) P+S % of IOs Read % of IOs Ldv KB / IO Fre Mem % Mem Aloc / Sec Mem Fai % Mem Age /Sc A+R % of IOs TPtP IOs /Sec TMlt IOs /Sec percent of time the CPUs are idle and waiting for completion of an I/O operation. per node. average number of memory allocations per second. per node. total point-to-point net reads and writes per second. percent of logical device reads and writes that are for paging or swapping purposes. average size of a logical device read or write. percent of logical device reads and writes that are reads. It is the average number of I/Os per node divided by the number of I/Os performed by the node with the most I/Os. per node. total multicast (broadcast or merge) net reads and writes per second. Parallel efficiency is the total percent of time nodes are busy. based on average CPU usage per node. The following table describes the statistics columns in all output reports (with the exception of ResNodeByNode. parallel efficiency of node CPU usage. General node process scheduling. 210 Resource Usage Macros and Tables . percent of memory allocation attempts that failed. parallel efficiency of the logical device (disk) I/Os. average number of logical device reads and writes per second for each node. Memory interface.Chapter 15: Resource Usage Macros ResNode Macros Statistics columns 4 through 7 8 through 12 13 through 15 16 and 17 Description Logical device interface. which has a NodeId column. per node.

spent waiting for a blocked process other than message and timer waits. per node.--.--.Chapter 15: Resource Usage Macros ResNode Macros Column… Net Rtry % Prc Blks / Sec ms /Blk Net Rx Bsy % Net Tx Bsy % Reports the… percent of transmission attempts that resulted in retries.----.--.----21 1 1 20 1 1 21 1 1 22 1 1 19 1 1 23 1 1 22 1 1 20 1 1 24 1 1 21 1 1 20 1 1 23 1 1 21 1 1 24 1 1 22 1 1 Mem Fai % --0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 Mem A+R TPtP TMlt Net Prc Age %of IOs IOs Rtry Blks /Sc IOs /Sec /Sec % /Sec --.------.--.--17 0 0 16 0 0 16 0 0 16 0 0 16 0 0 16 0 0 17 0 0 16 0 0 17 0 0 16 0 0 16 0 0 16 0 0 16 0 0 17 0 0 16 0 0 Ldv Fre Mem KB Mem Aloc /IO % /Sec --.----.--. average time.--.--.--0 100 0 0 0 41 0 100 0 0 0 41 0 100 0 0 0 39 0 100 0 0 0 40 0 100 0 0 0 40 0 100 0 0 0 40 0 100 0 0 0 40 0 100 0 0 0 40 0 100 0 0 0 40 0 100 0 0 0 40 0 100 0 0 0 40 0 100 0 0 0 40 0 100 0 0 0 40 0 100 0 0 0 41 0 100 0 0 0 39 ms / Blk ---477 371 2810 540 362 371 555 390 366 520 439 433 499 1142 378 Page 01 Net Net Rx Tx Bsy% Bsy% --. percent of time the network was busy either receiving.--.--.--.----.--1/07/12 09:44:20 2 16 09:44:40 1 17 09:45:00 1 16 09:45:20 2 17 09:45:40 1 17 09:46:00 2 17 09:46:20 2 16 09:46:40 1 16 09:47:00 2 16 09:47:20 2 16 09:47:40 2 17 09:48:00 1 15 09:48:20 2 16 09:48:40 2 16 09:49:00 1 16 General Resource Usage Summary for Node 001-01 Ldv P+S Rd IOs %of %of /Sec IOs IOs ----.---? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? Resource Usage Macros and Tables 211 .----15 16 100 0 0 22 1 1 0 0 100 0 0 0 40 16 16 100 0 0 23 1 1 0 0 100 0 0 0 39 16 16 100 0 0 19 1 1 0 0 100 0 0 0 39 17 16 100 0 0 24 1 1 0 0 100 0 0 0 40 16 16 100 0 0 22 1 1 0 0 100 0 0 0 39 16 16 100 0 0 22 1 1 0 0 100 0 0 0 40 17 17 100 0 0 22 1 1 0 0 100 0 0 0 41 15 16 100 0 0 24 1 1 0 0 100 0 0 0 39 14 16 100 0 0 24 1 1 0 0 100 0 0 0 40 16 17 100 0 0 23 1 1 0 0 100 0 0 0 41 16 16 100 0 0 22 1 1 0 0 100 0 0 0 39 17 17 100 0 0 23 1 1 0 0 100 0 0 0 41 16 17 100 0 0 22 1 1 0 0 100 0 0 0 41 16 16 100 0 0 23 1 3 0 0 100 0 0 0 42 15 16 100 0 0 18 1 1 0 0 100 0 0 0 39 16 16 100 0 0 20 1 1 0 0 100 0 0 0 40 16 16 100 0 0 21 1 1 0 0 100 0 0 0 39 16 16 100 0 0 19 1 1 0 0 100 0 0 0 40 16 17 100 0 0 23 1 1 0 0 100 0 0 0 42 15 16 100 0 0 19 1 1 0 0 100 0 0 0 39 16 16 100 0 0 22 1 1 0 0 100 0 0 0 40 15 16 100 0 0 19 1 1 0 0 100 0 0 0 40 15 16 100 0 0 18 1 1 0 0 100 0 0 0 39 15 16 100 0 0 20 1 1 0 0 100 0 0 0 40 16 16 100 0 0 22 1 1 0 0 100 0 0 0 40 16 16 100 0 0 22 1 1 0 0 100 0 0 0 40 16 17 100 0 0 19 1 1 0 0 100 0 0 0 40 ms Net Net / Rx Tx Blk Bsy% Bsy% ----.---.------.---386 ? ? 561 ? ? 383 ? ? 437 ? ? 532 ? ? 383 ? ? 417 ? ? 446 ? ? 426 ? ? 1173 ? ? 456 ? ? 416 ? ? 395 ? ? 2481 ? ? 463 ? ? 353 ? ? 662 ? ? 505 ? ? 341 ? ? 537 ? ? 480 ? ? 395 ? ? 491 ? ? 461 ? ? 1093 ? ? 442 ? ? 450 ? ? ResOneNode Sample Output 01/07/12 CPU Bsy WIO Date Time % % ------.--.---.-------. number of times per second. in milliseconds. ResNode Sample Output 01/07/12 CPU CPU Bsy Eff Date Time % % ------.--01/07/12 04:45:40 2 100 04:46:00 2 100 04:46:20 1 100 04:46:40 2 100 04:47:00 2 100 04:47:20 1 100 04:47:40 1 100 04:48:00 2 100 04:48:20 1 100 04:48:40 2 100 04:49:00 2 100 04:49:20 2 100 04:49:40 2 100 04:50:00 2 100 04:50:20 1 100 04:50:40 1 100 04:51:00 2 100 04:51:20 1 100 04:51:40 2 100 04:52:00 2 100 04:52:20 1 100 04:52:40 1 100 04:53:00 2 100 04:53:20 1 100 04:53:40 2 100 04:54:00 2 100 04:54:20 1 100 GENERAL RESUSAGE SUMMARY Page 4 Average across all nodes Ldv Ldv P+S Rd Ldv Fre Mem Mem Mem A+R TPtP TMlt Net Prc WIO IOs Eff %of %of KB Mem Aloc Fai Age %of IOs IOs Rty Blks % /Sec % IOs IOs /IO % /Sec % /Sec IOs /Sec /Sec % /Sec --.--.----.--. percent of time the network was busy transmitting. that processes other than message and timer waits are blocked.--.

--.----.Chapter 15: Resource Usage Macros ResNode Macros 09:49:20 09:49:40 09:50:00 09:50:20 2 2 1 2 16 17 16 17 17 16 16 16 0 0 0 0 0 0 0 0 20 20 19 20 1 1 1 1 1 1 3 1 0 0 0 0 0 0 0 0 100 100 100 100 0 0 0 0 0 0 0 0 0 0 0 0 41 41 40 40 543 363 2598 506 ? ? ? ? ? ? ? ? ResNodeByNode Sample Output 01/07/12 Node by Node General Resource Usage Summary Page 8 Mem A+R TPtP TMlt Net Prc ms Net Net Age %of IOs IOs Rty Blks / Rx Tx /Sc IOs /Sec /Sec % /Sec Blk Bsy% Bsy% ---.--.--.---0 100 0 0 0 42 442 ? ? 0 100 0 0 0 41 4887 ? ? 0 100 0 0 0 41 3844 ? ? 0 100 0 0 0 43 402 ? ? 0 100 0 0 0 40 521 ? ? 0 100 0 0 0 40 529 ? ? 0 100 0 0 0 40 387 ? ? 0 100 0 0 0 39 452 ? ? 0 100 0 0 0 40 381 ? ? 0 100 0 0 0 40 1281 ? ? 0 100 0 0 0 39 452 ? ? CPU Ldv P+S Rd Ldv Fre Mem Mem Bsy WIO IOs %of %of KB Mem Aloc Fai Date Time NodeId % % /Sec IOs IOs /IO % /Sec % -----.----.----.--.--.----.---.---.-------.--.--.---.--.----.--.---.---.-------.--.--01/07/12 04:55:40 001-01 2 17 17 0 0 26 1 1 0 04:56:00 001-01 2 15 16 0 0 22 1 3 0 04:56:20 001-01 2 14 16 0 0 22 1 7 0 04:56:40 001-01 2 16 17 0 0 23 1 6 0 04:57:00 001-01 2 16 16 0 0 23 1 1 0 04:57:20 001-01 1 16 16 0 0 24 1 1 0 04:57:40 001-01 1 15 16 0 0 22 1 1 0 04:58:00 001-01 1 16 16 0 0 22 1 1 0 04:58:20 001-01 1 15 16 0 0 22 1 1 0 04:58:40 001-01 1 16 16 0 0 22 1 1 0 04:59:00 001-01 2 16 16 0 0 21 1 1 0 ResNodeByGroup Sample Output 01/07/12 GENERAL RESOURCE USAGE SUMMARY BY GROUP Page 8 ms Net Net / Rx Tx Blk Bsy Bsy% -----.----.----.--.---442 ? ? 4887 ? ? 3844 ? ? 402 ? ? 521 ? ? 529 ? ? 387 ? ? 452 ? ? 381 ? ? 1281 ? ? 452 ? ? CPU Ldv P+S Rd Ldv Fre Mem Mem A+R TPtP TMlt Net Prc Group Bsy WIO IOs %of % of KB Mem Aloc Fai %of IOs IOs Rty Blks Date Time Id % % /Sec IOs IOs /IO % /Sec % IOs /Sec /Sec % /Sec -------.--.----.-----.--.--.----.----01/07/12 04:55:40 A 2 17 17 0 0 26 1 1 0 100 0 0 0 42 04:56:00 A 2 15 16 0 0 22 1 3 0 100 0 0 0 41 04:56:20 A 2 14 16 0 0 22 1 7 0 100 0 0 0 41 04:56:40 A 2 16 17 0 0 23 1 6 0 100 0 0 0 43 04:57:00 A 2 16 16 0 0 23 1 1 0 100 0 0 0 40 04:57:20 A 1 16 16 0 0 24 1 1 0 100 0 0 0 40 04:57:40 A 1 15 16 0 0 22 1 1 0 100 0 0 0 40 04:58:00 A 1 16 16 0 0 22 1 1 0 100 0 0 0 39 04:58:20 A 1 15 16 0 0 22 1 1 0 100 0 0 0 40 04:58:40 A 1 16 16 0 0 22 1 1 0 100 0 0 0 40 04:59:00 A 2 16 16 0 0 21 1 1 0 100 0 0 0 39 212 Resource Usage Macros and Tables .-----.----.

Each defined workload is internally associated with a priority scheduler Performance Group. For information on working with Teradata ASM rules.FromTime. see the Teradata Viewpoint Workload Designer portlet. you must have Teradata ASM Category 3 rule (Workloads) enabled and the workloads defined. Name the node log rate. ResPsByNode ResPsByGroup Provides a summary of the Priority Scheduler resource usage. FromNode. ToDate. Usage Notes For any of these macros the following usage notes apply: • • Logging must be enabled on ResUsageSps. ToNode.ToDate. EXEC ResPsByNodeWDJoin (FromDate.FromTime. Input Format Examples The input forms of the macros are described below..FromTime. These macros display the critical workloads in the context of their Allocation Group relationships. In order for the ResPsWDJoin and ResPsByNodeWDJoin macros to function. ToTime.Chapter 15: Resource Usage Macros ResPs Macros ResPs Macros Function Macro. FromTime...ToNode). EXEC ResPsByGroup (FromDate.FromNode.ToDate.ToTime). Therefore.ToTime). See “Executing Macros” on page 32 for a description of the FromDate. Resource Usage Macros and Tables 213 . by node.FromNode. EXEC ResPsByNode (FromDate.ToTime. the GroupId column is not needed and the ResUsageSps table view is not provided. Note: Coexistence support can be accomplished using the NodeType column to do a group by in SQL directly. for each logging period.ToTime. produces one row of data for each Performance Group ID.ToDate. by coexistence group. produces one row of data for each node type in the system per logging period. and Node parameters. which in turn is associated with Allocation Groups.ToDate.FromTime.. Output Examples The reports in the following sections are sample output reports from the ResPsByNode and ResPsByGroup macros. EXEC ResPsWDJoin (FromDate.ToNode).

in the ResPsWDJoin output report. Workload name. Column… CPU Bsy % Reports the… percent of CPU time consumed by a task associated or running under the Performance Group. Statistics columns 1 2 3 through 12 Description Node ID. The following table describes the summary statistics columns in all output reports (with the exception of the ResPsWDJoin and ResPsByNodeWDJoin macros which have the CPU ms column). Allocation Group ID. Relative weight. in the ResPsByNode output report. Performance Group ID. after the Date and Time columns. Summary of the Priority Scheduler and Teradata ASM workload statistics. in the ResPsByGroup output report. 214 Resource Usage Macros and Tables . in the ResPsByNodeWDJoin output report. after the Date and Time columns. Statistics columns 1 2 through 11 Description Node type.Chapter 15: Resource Usage Macros ResPs Macros The following table describes the 12 statistics columns. after the Date and Time columns. The following table describes the 14 statistics columns. Summary of the Priority Scheduler statistics. The following table describes the 15 statistics columns. Performance period ID. Statistics columns 1 2 3 4 5 6 through 15 Description Node ID. after the Date and Time columns. The following table describes the 11 statistics columns. Summary of the Priority Scheduler statistics.

average QWaitTime for each request during a specified period. see Chapter 11: “ResUsageSps Table.--------07/05/07 14:40:00 15:00:00 ------1-05 1-05 1-04 1-04 1-05 1-05 1-04 1-04 1-05 1-05 1-04 1-04 1-05 1-05 1-04 1-04 1-04 1-05 1-05 1-04 1-04 1-05 1-05 ---1 40 0 40 0 40 0 40 0 40 0 40 0 40 0 3 40 0 40 0 40 0 40 15:20:00 15:40:00 16:00:00 17:00:00 Resource Usage Macros and Tables 215 . number of work requests waiting on the input queue for service.---? 0 0 0 ? 0 ? 2 ? 0 ? 0 ? ? 0 ? ? ? 0 ? 0 ? 0 ? 3 0 0 10 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 ----0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 ---0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 Page Avg Svc Time ----? 62 ? 215 ? 162 ? 97 ? 93 ? 50 ? 102 ? ? 130 ? 71 ? 112 ? 62 6 Max Svc Time ----0 100 0 220 0 270 0 180 0 160 0 80 0 150 0 0 160 0 260 0 140 0 90 -------. maximum time in milliseconds that work requests waited on an input queue before being serviced. For a complete description of the above columns. number of requests of the AWT. maximum number of work requests waiting on the input queue for service.” ResPsByNode Macro Sample Output 07/05/07 Date Time NodeID PGid PS by Node Usage Summary Page CPU Bsy % --0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 0 1 0 1 IO Blks/ Sec ----0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 Num Num Avg Max Q Q Procs Requests QWait QWait Length Len Time Time Max ---0 50 0 54 0 50 0 54 0 50 0 54 0 50 0 0 54 0 50 0 54 0 0 ------0 5 0 2 0 3 0 3 0 3 0 4 0 3 0 0 3 0 5 0 3 0 50 ----. average ServiceTime for each request during a specified period. number of processes assigned to the Performance Group at the end of the gather period.Chapter 15: Resource Usage Macros ResPs Macros Column… IO Blks / Sec Num Procs Num Requests Avg QWait Time Max QWait Time Q Length Q Length Max Avg Svc Time Max Svc Time Reports the… average number of logical data blocks read and (or) written by Performance Group per second. maximum time in milliseconds that work requests required for service.

-------07/05/04 13:00:00 13:10:00 13:20:00 14:00:00 Q Avg Max Q Len Svc Svc Length MaxTime Time Time ------.------0 0 30 200 0 0 102 210 0 0 30 160 0 0 52 240 ResPsByNodeWDJoin Macro Sample Output 07/08/09 Workload Usage Summary (Average Usage per AMP By Node) IO Blks /Sec ----0 0 0 0 7 0 36 0 0 0 0 2 0 0 10 9 0 0 0 3 0 2 4 10 0 0 0 0 7 Avg QWait Time -----? ? ? 4 3 ? 3 ? 2 ? ? 4 ? ? 5 ? ? ? 13 4 ? 22 9 2 ? ? ? 3 3 Max QWait Time ------0 0 0 70 220 0 400 0 30 0 0 210 0 0 340 0 0 0 130 30 0 220 240 50 0 0 0 160 290 Q Len Max -----0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 Avg Svc Time ------? ? ? 2 18 ? 35 ? 14 ? ? 1548 ? ? 1303 ? ? ? 309 7072 ? 3875 333 24 ? ? ? 12 45 Page 1 Max Svc Time ------0 0 0 100 1160 0 4120 0 280 0 0 41680 0 0 47370 1063680 0 0 3720 67660 0 500 4740 1980 0 0 0 830 2730 Date -------07/08/06 Time -------17:57:00 Node ID -----1-04 AG ID --1 2 3 4 5 7 Rel Wgt --1 2 5 10 48 11 Workload (WD) Name -----------------------------? ? ? ? All_Tactical TDWM Continious Load LobLoader Teradata Manger WD-ConsoleH WD-ConsoleR ADW_Strategic DWD_OLAP Java Stored Procedures Mixedsql Multiuser Simulation PEstress WD-ConsoleM WD-Default Penalty_box WD-ConsoleL All_Tactical qmiles ? ? ? ? ? All_Tactical PP ID -0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 CPU ms -----0 0 0 36 1084 0 4156 0 19 0 0 2506 0 0 2848 995 0 0 10 2613 0 911 3769 1048 0 0 0 64 957 Num Procs -----0 0 0 0 0 0 0 0 0 0 0 13 0 0 21 7 0 0 0 4 0 0 0 6 0 0 0 0 0 Num Requests -------0 0 0 48 364 0 2679 0 18 0 0 110 0 0 124 0 0 0 12 43 0 1 87 270 0 0 0 64 340 Q Length -----0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 8 5 9 10 200 1-05 1 2 3 4 5 2 12 100 1 2 5 10 48 .----.ResPsByGroup Macro Sample Output 07/05/04 PS by Group Usage Summary CPU IO Avg Node Bsy Blks Num Num QWait Type % /Sec Procs Requests Time ---.---.------.-------.----.-----UNKN 0 0 3 5 1 UNKN 0 0 5 0 0 UNKN 0 0 3 3 0 UNKN 0 0 3 3 0 Max QWait Time ------10 0 10 10 Page 1 216 Resource Usage Macros and Tables Date Time ------.-----.

if the question mark is used as a value in any other column. the it indicates there is no information to report for this time period (see “Question Marks” on page 176 for details). However. .Resource Usage Macros and Tables 217 TDWM 7 11 Continious Load LobLoader Teradata Manger WD-ConsoleH WD-ConsoleR ADW_Strategic DWD_OLAP Java Stored Procedures Mixedsql Multiuser Simulation PEstress WD-ConsoleM WD-Default Penalty_box WD-ConsoleL All_Tactical qmiles ? 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 8223 0 19 0 0 1802 0 0 1573 2724 0 0 9 1313 0 670 2916 1062 0 85 0 0 0 0 4 0 0 6 4 0 0 0 2 0 2 3 10 0 1 0 0 0 0 3 0 0 5 4 0 0 0 4 0 0 0 6 0 3401 0 17 0 0 66 0 0 427 0 0 0 12 10 0 3 69 226 ? 2 ? 4 ? ? 6 ? ? 4 ? ? ? 4 5 ? 84 12 2 0 550 0 80 0 0 50 0 0 170 0 0 0 20 30 0 510 530 360 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 ? 25 ? 23 ? ? 5564 ? ? 1529 ? ? ? 134 13160 ? 1886 345 32 0 4220 0 100 0 0 82900 0 0 82110 0 0 0 960 98350 0 1890 4710 3570 8 5 9 10 200 2 12 100 Note: Question marks used as values in the Workload (WD) Name column in the output above mean there is no associated workload to the PG ID/PP ID.

See “Executing Macros” on page 32 for a description of the FromDate. EXEC ResVdskByGroup (FromDate. a physical node. which has the NodeId column. EXEC ResVdskByNode (FromDate.Node).ToTime. ToTime. average number of logical device writes per second. FromTime. FromNode.Chapter 15: Resource Usage Macros ResVdskByNode Macros ResVdskByNode Macros Function Macro... EXEC ResVdskOneNode (FromDate. average number of logical device reads per second..FromTime.FromNode.. average number of KBs per logical device read. ToDate.ToTime).ToDate. ToNode and Node parameters.ToDate.. Input Format Examples The input forms of these three macros are described below. Usage Notes For any of these macros the following usage notes apply: • • Logging must be enabled on ResUsageSvdsk. Column... 218 Resource Usage Macros and Tables . Read Cnt / Sec Write Cnt / Sec Rd KB / I/O Wrt KB / I/O Avg I/O Resp Reports the. and ResVdskByGroup. a node grouping.ToNode)..FromTime. Name the node log rate. which has the NodeType column). average response time for a logical device read or write in seconds. Output Examples The following table describes the statistics columns in all output reports (with the exception of ResVdiskByNode.FromTime. average number of KBs per logical device write.ToDate.ToTime. for a specified node. ResVdskByNode ResVdskOneNode ResVdskByGroup Reports the logical device traffic by.

2% 45.4% 97.118 13.118 0.90 12.01 Resource Usage Macros and Tables 219 .17 111.76 0.30 1785.7% 85.-----.70 1759.77 1785.15 0.40 ? 45.00 2.7% 0.77 2.93 112.94 ResVdskByGroup Sample Output 06/09/26 VDISK TRAFFIC BY GROUP Page Avg Max Out I/O Concur Rqst Resp Rqsts Time % ------.0% 97.77 12.67 43.54 109.7% 98.8% 98.143 20.69 116.119 0.40 ? 2.00 47.17 127..081 12.25 1.20 9.17 127.3% 54.76 18.126 109.023 4.8% 98.87 WriteCnt Rd KB / Sec / I/O -------.00 46.3% 0.6% 3.-------06/09/26 4400 10:09:45 10:10:00 10:10:15 10:10:30 10:10:45 10:11:00 10:11:15 10:11:30 ReadCnt / Sec -------0.3% 98.67 2.56 0.0 97.79 45.50 0.8% 38.023 0.13 1374.73 5.00 12.-----0.90 12.0% 36.5 54.70 1759.039 0.00 43.0% 0.00 0.56 16.34 0. Out Rqst Time % Max Concur Rqsts Reports the.065 71.47 0.34 118.37 5.43 3..79 2.10 12.3% 0.2% 45.064 0.Chapter 15: Resource Usage Macros ResVdskByNode Macros Column.00 WriteCnt Rd KB / Sec / I/O -------.4% 0.---.107 14.50 12.60 275.27 41.43 0.50 47.118 112.5 98.34 118.25 5.40 0.07 112.038 0.17 ? Note: The NodeId column only appears in the ResVdskByNode output report.93 112.9% 41.13 1374.60 275.83 4.065 0.107 0.01 38.32 0.152 20.119 14.00 43. ResVdskByNode Sample Output 06/09/26 VDISK TRAFFIC BY NODE Avg Wrt KB I/O I/O Resp ------.039 18.5 9.40 ? 45.081 38.47 2..00 2.00 4.7% 85.07 0.------5.43 45.038 32.5 97.20 121.17 111.69 0.00 46.00 2.0 98.77 2.01 0.1% 1 Date -------06/09/26 Time -------10:09:45 10:10:00 10:10:15 10:10:30 10:10:45 10:11:00 10:11:15 10:11:30 10:11:45 10:12:00 10:12:15 10:12:30 10:12:45 Node Id -----001-01 001-01 001-01 001-01 001-01 001-01 001-01 001-01 001-01 001-01 001-01 001-01 001-01 Read Cnt Write Cnt Rd KB / Sec / Sec I/O / -------.7% 98.143 0.73 0.15 71.00 14.50 0.17 127.98 0.143 113.79 45.27 0. ResVdskOneNode Sample Output 06/09/26 VDISK Traffic for Node 001-01 Avg I/O Resp ------0.84 0.67 2.30 1785.017 Page Out Rqst Time % -----9.0% 97.20 0.10 12.119 112.------5.00 9.00 2.17 111.25 Wrt KB / I/O ------46.32 100.10 12.54 0.6% 3.3% 98.081 0.5 85.20 121.3% 54.83 4.37 1.------0.152 0. maximum number of concurrent requests during the log period.------46.9% 41.27 41.84 32.00 5.4% 97.023 116.-------.54 109.77 1785.43 2. percent of time there are outstanding requests.69 116.77 12.152 118..77 12.98 113.77 45.17 ? Wrt KB / I/O ------46.94 0.73 0.00 14.7% 0.00 0.5 98.064 16.87 14.8% 38.43 2.50 47.87 5.40 0.32 100.93 0.98 113.107 100.20 121.0% 36.90 12.00 2.40 41.1% 1 Date -------06/09/26 Time -------10:09:45 10:10:00 10:10:15 10:10:30 10:10:45 10:11:00 10:11:15 10:11:30 10:11:45 10:12:00 10:12:15 10:12:30 10:12:45 ReadCnt / Sec --------0.07 112.9% 1 Node Date Type Time -------.83 46.126 0.017 Page Out Rqst Time % -----9.8% 0.126 17.13 1374.00 3.

0% 36.60 2.00 1759.77 0.017 3.5 2.37 5. 220 Resource Usage Macros and Tables .1% Note: The NodeType column only appears in the ResVdskByGroup output report.30 9.56 16.20 0.0 2.50 0.065 0.15 71.8% 38.064 0.17 1785.00 1785.47 ? 38.Chapter 15: Resource Usage Macros ResVdskByNode Macros 10:11:45 10:12:00 10:12:15 10:12:30 10:12:45 5.70 12.0 3.0 2.76 18.94 0.43 3.6% 3.038 0.0 41.2% 45.00 275.00 1.039 0.84 32.

• lowercase letters represent a keyword that you must type in lowercase. and proceeds. • lowercase bold letters represent an excerpt from the diagram. Syntax diagrams show all keywords in uppercase. unless operating system restrictions require them to be in lowercase. which marks the end of the diagram. Do not use commas when typing a number with more than 3 digits. • lowercase italic letters represent a variable such as a column or table name. This applies to both uppercase and lowercase words. to the vertical bar. Type all punctuation exactly as it appears in the diagram.APPENDIX A How to Read Syntax Diagrams This appendix describes the conventions that apply to reading the syntax diagrams used in this book. The excerpt is defined immediately following the diagram that contains it. Spaces Punctuation Use one space between items such as keywords or variables. • UPPERCASE LETTERS represent a keyword. A digit ranging from 0 through 9. Word Keywords and variables. Resource Usage Macros and Tables 221 . Syntax Diagram Conventions Notation Conventions Item Letter Number Definition / Comments An uppercase or lowercase alphabetic character ranging from A through Z. left to right. The only part of a path that reads from right to left is a loop. Substitute the variable with a proper value. such as a Linux command. Paths The main path along the syntax diagram begins at the left with a keyword. • UNDERLINED LETTERS represent the default value. Paths that do not have an arrow or a vertical bar only show portions of the syntax.

Required Entries Required entries appear on the main path: SHOW FE0CA003 If you can choose from more than one entry. in a stack. Optional entries appear below the main path: SHOW CONTROLS FE0CA004 222 Resource Usage Macros and Tables .Appendix A: How to Read Syntax Diagrams Syntax Diagram Conventions Continuation Links Paths that are too long for one line use continuation links. The first entry appears on the main path: SHOW CONTROLS VERSIONS FE0CA005 Optional Entries You may choose to include or disregard optional entries. the choices appear vertically. Continuation links are circled letters indicating the beginning and end of a link: A A FE0CA002 When you see a circled letter in a syntax diagram. go to the corresponding circled letter and continue reading.

over the item or items that you can repeat: .Appendix A: How to Read Syntax Diagrams Syntax Diagram Conventions If you can optionally choose from more than one entry. the unabbreviated form always appears on the main path. Syntax diagrams show loops as a return path above the main path. the following formats are valid: • • SHOW CONTROLS SHOW CONTROL Loops A loop is an entry or a group of entries that you can repeat one or more times. . all the choices appear below the main path: READ SHARE ACCESS JC01A010 Some commands and statements treat one of the optional choices as a default value. SHOW CONTROLS CONTROL FE0CA042 In the above syntax. This value is UNDERLINED. ( cname 4 3 ) JC01B012 Resource Usage Macros and Tables 223 . Strings String literals appear in apostrophes: 'msgtext ' JC01A004 Abbreviations If a keyword or a reserved word has a valid abbreviation. It is presumed to be selected if you type the command or statement without specifying one of the options. The shortest valid abbreviation appears beneath.

a separator character is required between entries Excerpts Sometimes a piece of a syntax phrase is too large to fit into the diagram. col_pos JC01A014 excerpt A 224 Resource Usage Macros and Tables . The boldface excerpt name and the excerpted phrase appears immediately after the main diagram. there is a maximum number of entries allowed there is a minimum number of entries required THEN. cname . a delimiter character is required around entries the beginning and end characters appear outside the return path. Such a phrase is indicated by a break in the path. In the example. marked by (|) terminators on each side of the break. the separator character is a comma. If the diagram does not show a separator character. the character appears on the return path. Generally.. you must type at least three groups of column names. In the example. the number appears in a circle on the return path.Appendix A: How to Read Syntax Diagrams Syntax Diagram Conventions Read loops from right to left. the number appears in a square on the return path... a space is not needed between delimiter characters and entries. The following conventions apply to loops: IF. you may type cname a maximum of 4 times. In the example. The name for the excerpted piece appears between the terminators in boldface type. use one blank space. The excerpted phrase starts and ends with a plain horizontal line: LOCKING A HAVING con excerpt where_cond .. the delimiter characters are the left and right parentheses. In the example.

Appendix A: How to Read Syntax Diagrams Syntax Diagram Conventions Multiple Legitimate Phrases In a syntax diagram. it is possible for any number of phrases to be legitimate: dbname DATABASE tname TABLE vname VIEW JC01A016 In this example. CREATE VIEW CV viewname cname AS LOCKING LOCK A A DATABASE dbname FOR ACCESS SHARE READ WRITE EXCLUSIVE EXCL MODE IN B tname TABLE vname VIEW . B SEL . . any of the following phrases are legitimate: • • • • • • dbname DATABASE dbname tname TABLE tname vname VIEW vname Sample Syntax Diagram . cname . col_pos JC01A018 Resource Usage Macros and Tables 225 . FROM expr tname .aname qual_cond C C HAVING cond qual_cond WHERE cond GROUP BY .

The text never refers to this string. 226 Resource Usage Macros and Tables .Appendix A: How to Read Syntax Diagrams Syntax Diagram Conventions Diagram Identifier The alphanumeric string that appears in the lower right corner of every diagram is an internal identifier used to catalog the diagram.

For example. The Node ID is formatted as CCC-MM. vproc types 6 through 7 are not currently used. Type 2 (PE). Type 3 (GTW). Nominal time of the log entry. Under the current implementation. The following table describes the ResUsageIpma table columns. node information. by no more than one or two seconds). shown as Type n virtual processors where n = 1 to 7. This table is created as a MULTISET table. only Type 1 (AMP). a node in chassis 9 of cabinet 3 has a node ID of ‘003-09’. For more information see “Relational Primary Index” on page 38. MISCELLANEOUS HOUSEKEEPING COLUMNS These columns provide a generalized picture of the vprocs running on this node. For example. Type of Data Invalid Platform Column Name Description HOUSEKEEPING COLUMNS Data Type RELATIONAL PRIMARY KEY COLUMNS These columns taken together form the nonunique primary index.APPENDIX B ResUsageIpma Table The ResUsageIpma table includes resource usage data for system-wide. Note: Under conditions of heavy system load. entries may be logged late (typically. and Type 5 (VSS) vprocs exist. FLOAT INTEGER DATE FLOAT Resource Usage Macros and Tables 227 . See the Secs and NominalSecs columns. NodeId n/a Identifies the Node. GmtTime n/a Greenwich Mean Time is not affected by the Daylight Savings Time adjustments that occur twice a year. Type 4 (RSG). but this field will still contain the time value when the entry should have been logged. Note: SMP nodes have a chassis and cabinet number of 1. the node ID of an SMP node is ‘001-01’. where CCC denotes the threedigit cabinet number and MM denotes the two-digit chassis number of the node. TheDate TheTime n/a n/a Date of the log entry. Note: Summary Mode is not applicable to this table.

Data Type CHAR(8) Invalid Platform NCPUs n/a SMALLINT Vproc1 VprocType1 Vproc2 VprocType2 Vproc3 VprocType3 Vproc4 VprocType4 Vproc5 VprocType5 n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a Current count of type 1 (AMP) virtual processors running under the node. CHAR(4) SMALLINT ALL VprocType7 n/a Type of virtual processor for Vproc7. Current count of type 2 (PE) virtual processors running under the node. Number of CPUs on this node. Type of virtual processor for Vproc3. Current count of type 7 virtual processors running under the node. Value is always AMP. SMALLINT CHAR(4) SMALLINT CHAR(4) SMALLINT CHAR(4) SMALLINT CHAR(4) SMALLINT CHAR(4) Vproc6 n/a SMALLINT VprocType6 Vproc7 n/a n/a Type of virtual processor for Vproc6. Value is always RSG.Appendix B: ResUsageIpma Table Column Name NodeType Type of Data n/a Description Type of node. Value is always PE. The value is always TVS (see Teradata Virtual Storage). This column reports zeros and " " (blanks). This column reports zeros and " " (blanks). Value is always GTW. 5600C or 5555H. Type of virtual processor for Vproc1. Current count of type 4 (RSG) virtual processors running under the node. Type of virtual processor for Vproc2. Current count of type 3 (GTW) virtual processors running under the node. Current count of type 6 virtual processors running under the node. Current count of type 5 (VSS) virtual processors running under the node. Type of virtual processor for Vproc4. This field is useful for normalizing the CPU utilization field values for the number of CPUs on the node. representing the per node system family type. CHAR(4) ALL 228 Resource Usage Macros and Tables . For example. This is especially important in coexistence systems where the number of CPUs can vary across system nodes. Type of virtual processor for Vproc5.

INTEGER NominalSecs n/a SMALLINT Resource Usage Macros and Tables 229 . but can be different in three cases: • The first interval after a log rate change • A sample logged late because of load on the system • System clock adjustments affect reported Secs Useful for normalizing the count statistics contained in this row. Useful for performing memory usage calculations. then the value of the NodeNormFactor will be 525.” Data Type INTEGER Invalid Platform NodeNormFactor n/a INTEGER Secs n/a Actual number of seconds in the log period represented by this row. INTEGER CentiSecs n/a Number of centiseconds in the logging period. Normally the same as NominalSecs.Appendix B: ResUsageIpma Table Column Name MemSize Type of Data n/a Description Amount of memory on this node in megabytes. to a persecond measurement. Note: The normalization factor is related to the NodeType value reported in the ResUsageSpma table. see Chapter 6: “ResUsageSpma Table. for example. A specified or nominal number of seconds in the logging period. For example. This value is scaled by a factor of 100. For information on this value. This field is useful when performing data calculations with small elapsed times where the difference between centisecondbased data and whole seconds results in a percentage error. A per node normalization factor that is used to normalize the reported CPU values of the ResUsageSpma table.25. if the actual factor is 5.

For example. that is. FLOAT Windows CPUProcSameSwitches count FLOAT ALL Interrupted CPU Switching Columns Identify the number of times an interrupt was issued for the node and/or its CPUs. In the Collect Buffer and Log Buffer. CPUProcSwitches count Number of times the scheduler switched a CPU’s currently active process to a new process. This field is a subset of CPUProcSwitches. STATISTICS COLUMNS PROCESS SCHEDULING COLUMNS SMALLINT Scheduled CPU Switching Columns Identify the number of times CPUs were switched by the scheduler from one type of work to another type of work. then none of the data columns in the row have been updated during the logging period. • a zero value. the value is the number of Gather operations that have been performed during the period. Number of CPUProcSwitches where a process replaced itself. FLOAT FLOAT FLOAT FLOAT ALL ALL ALL ALL 230 Resource Usage Macros and Tables . Number of times a CPU was interrupted by the IOP. Number of times the node was interrupted to handle a disk request. ProcNetInts IOPtoCPUInts ProcDiskInts ProcHostInts count count count count Number of times the node was interrupted for Teradata a net request. then the row contains modified data columns. If Active is set to: • a non-zero value. the new process was the same as the old process. if Active Row Filter Mode is enabled. Data Type FLOAT Invalid Platform CollectIntervals n/a The number of gather periods per reporting period. then the rows that have a zero Active field value will not be logged to the ResUsage tables. Number of times the node was interrupted to handle a host request.Appendix B: ResUsageIpma Table Column Name Active Type of Data count Description Controls whether or not the rows will be logged to the ResUsage tables if Active Row Filter Mode is enabled. This number can vary from one period to the next.

FLOAT FLOAT FLOAT FLOAT FLOAT FLOAT FLOAT ALL Resource Usage Macros and Tables 231 . The number of timer interrupts. • These columns do not include any memory deallocated specific to a vproc running under the node. Number of processor messages written by the node. MsgHashReads MsgHashWrites MsgProcReads MsgProcWrites MsgGroupReads MsgGroupWrites count count count count count count Number of hash messages read by this node. Number of group messages read by this node. MEMORY COLUMNS Data Type FLOAT FLOAT FLOAT FLOAT Invalid Platform ALL Windows ALL ALL Memory Page Deallocation Columns Represent the number of memory page deallocations specific to generic node activities. Number of group messages written by this node. Number of hash messages written by this node. • The amount deallocated can be derived by multiplying the number of deallocations by the fixed page size. MemTextDestroys count Number of memory deallocations and sizedecreasing memory alters for non-system overhead text (code).Appendix B: ResUsageIpma Table Column Name ProcLanInts ProcGenClockInts ProcCPUClockInts ProcInterCPUInts Type of Data count count count count Description Number of times the node was interrupted to handle a LAN request. and so on • Mailbox messages (Mbox) are data sent between vprocs through mailbox IDs for similar purposes as channel messages. NET COLUMNS Message Type Columns Subdivide all messages sent and received into the type of message. subdivided into memory types. since Hash + Proc + Group + Local messages = Chan + MBox messages. Number of processor messages read by this node. Number of times a CPU was interrupted to service a CPU specific clock event. Number of times the a CPU was interrupted to service an inter-CPU request. A duplicated accounting is done with two different perspectives. where: • • • • • Hash messages (Hash) are data sent to a destination through its primary or fallback hash value Processor messages (Proc) are data sent to a destination through a vproc ID Group messages (Group) are broadcasted messages to be received by members of a group Local messages (Local) are messages communicated locally within the node Channel messages (Chan) are data sent between vprocs through channel IDs for purposes of a private conversation to perform functions such as row redistribution.

Total amount of time read hash messages took for process delivery. On a single-node system. Number of mailbox messages written by this node. Data Type FLOAT FLOAT FLOAT FLOAT FLOAT FLOAT Invalid Platform Message Delivery Time Columns Identify the time it took for hash. FLOAT FLOAT FLOAT FLOAT FLOAT FLOAT FLOAT FLOAT Net Circuit Management Columns Identify the management of Teradata net circuits (Circ) and raw data traffic on the network (hardware) on all networks.Appendix B: ResUsageIpma Table Column Name MsgLocalReads MsgLocalWrites MsgChanReads MsgChanWrites MsgMboxReads MsgMboxWrites Type of Data count count count count count count Description Number of local messages read by this node. Total amount of time read group messages took for mailbox delivery. Total amount of time read group messages took for process delivery. Total amount of time read processor messages took for mailbox delivery. Note: All of these columns except for NetBackoffs are net-specific. group and local messages to reach their destination. Number of local messages written by this node. FLOAT 232 Resource Usage Macros and Tables . Total amount of time read processor messages took for process delivery. Total amount of time read local messages took for process delivery. Number of channel messages read by this node. Two times are provided: • Message transmission to mailbox delivery (MDelivery) • Mailbox delivery to process delivery (PDelivery) MsgHashMDelivery MsgProcMDelivery MsgGroupMDelivery MsgLocalMDelivery MsgHashPDelivery MsgProcPDelivery MsgGroupPDelivery MsgLocalPDelivery count count count count count count count count Total amount of time read hash messages took for mailbox delivery. net-specific statistics are not meaningful and are always zero. Number of channel messages written by this node. Total amount of time read local messages took for mailbox delivery. NetBackoffs count Software backoffs. processor. Number of mailbox messages read by this node. defined as BNS service blocked occurrences without regard for which net was involved.

Appendix B: ResUsageIpma Table Column Name NetTxCircPtp Type of Data count Description Total number (both normal and high priority) of point-to-point circuits transmitted on all Bynets. Total number (both normal and high priority) of broadcast circuits received on all Bynets. that is. Number of high priority broadcast circuits transmitted on all Bynets. it relates to each specific Bynet. On a single-node system. FLOAT FLOAT FLOAT FLOAT Net Miscellaneous Contention Management Columns Identify some additional contention management not addressed in the other contention management areas. Number of times the net software was blocked because the channel was not in RxReady state on the receiver. Note: NetBrdWindowOverrun is net-specific. net-specific statistics are not meaningful and are always zero. NetMsgFCSleep NetMsgFCBlock NetMsgResourceBlock count count count Number of times a transmitter process was put to sleep because it was flow controlled. Total broadcast data KBs transmitted over all Bynets Total point-to-point data KBs received over all Bynets. Total broadcast data KBs received over all Bynets. Total number (both normal and high priority) of point-to-point circuits received on all Bynets. Number of times the net software was blocked because the receiver could not get the necessary resources. Bynet Network Transport Data Columns Data Type FLOAT Invalid Platform NetTxCircBrd count FLOAT NetTxCircHPPtP NetTxCircHPBrd NetRxCircPtp count count count FLOAT FLOAT FLOAT NetRxCircBrd count FLOAT NetTxKBPtP NetTxKBBrd NetRxKBPtP NetRxKBBrd count count count count Total point-to-point data KBs transmitted over all Bynets. FLOAT FLOAT FLOAT NetMsgChannelBlock count FLOAT Resource Usage Macros and Tables 233 . Number of times the net software was blocked because the receiver was flow controlled. Total number (both normal and high priority) of broadcast circuits transmitted on all Bynets. Number of high priority point-to-point circuits transmitted on all Bynets.

FLOAT FLOAT FLOAT 234 Resource Usage Macros and Tables . The number of concurrent active merges on all Bynets. NetPtPQueue NetPtPQueueMax NetBrdQueue count max count Aggregated sample point-to-point normal priority queue length on all Bynets. Number of broadcast window errors. Data Type FLOAT Invalid Platform NetMsgRxBlock count FLOAT NetMrgBlock count FLOAT NetBrdWindowOverrun NetActiveMrg NetMrgBufWaits NetBackoffExhausted count count count count FLOAT FLOAT FLOAT FLOAT ALL NetBrdWindowError NetConfigurations NetProtocolFilter NetTxSoftBackoffs NetRxSoftBackoffs count count count count count FLOAT FLOAT FLOAT FLOAT FLOAT ALL ALL ALL ALL ALL Net Queues Columns Identify lengths of the various internal queues used by the network controllers. Aggregated sample broadcast normal priority queue length on all Bynets. that is. they relate to each specific Bynet. Number of soft backoffs transmitted on all Bynets. net-specific statistics are not meaningful and are always zero. Number of protocol filters executed. On a single-node system. Broadcast window overruns on all Bynets. Number of times a merge message was blocked until delivery of outstanding outgoing messages. Number of soft backoffs received on all Bynets. • Example: Dividing (NetPtPQueue/NetSamples) yields the average point-to-point queue length over all samples on all networks taken during this log interval. Number of network configurations and reconfigurations. The maximum value of NetPtPQueue over all gather intervals in this reporting interval. Number of transmit circuits that were backedoff too many times and had to be converted to blocking circuits.Appendix B: ResUsageIpma Table Column Name NetMsgGroupBlock Type of Data count Description Number of times the net software was blocked because the receiver could not implicitly enter the group. • NetSamples can be used to normalize all aggregated sampled statistics to an average queue-length basis. Number of times an IOP task encountered an empty row-block buffer on all Bynets. • All of the aggregated sampled statistics columns in the following table are net-specific. Number of times the net software could not accept a message and caused a transmitter to block.

Number of times entry into a lockable resource was blocked. Aggregated sample broadcast high priority queue length on all networks. The maximum value of NetHPBrdQueue over all gather intervals in this reporting interval. LockEnters LockBlocks count count Number of times entry into a lockable resource was requested. (requests . regardless of which net during each log interval. A merge may be queued for reasons such as: • the local IOP memory is saturated • system memory is trashing. Total number of services on the BlockableServices queue. The maximum value of NetHPPtPQueue over all gather intervals in this reporting interval. The current count of pending merges. CacheAccess count Total number of accesses. GENERAL CONCURRENCY CONTROL COLUMNS Data Type FLOAT FLOAT FLOAT FLOAT FLOAT FLOAT Invalid Platform NetBlockQueueTotal NetBlockQueueMax NetPendMrgQueue count max count FLOAT FLOAT FLOAT Operating System Lock Management Columns Identify database locking activities for internal multiprocessing operating system concurrency control.Appendix B: ResUsageIpma Table Column Name NetBrdQueueMax NetHPPtPQueue NetHPPtPQueueMax NetHPBrdQueue NetHPBrdQueueMax NetBlockQueueSum Type of Data max count max count max count Description The maximum value of NetBrdQueue over all gather intervals in this reporting interval. Total number of services on the BlockableService queue.) FLOAT FLOAT ALL ALL Secondary Cache Misses Columns Identify the percentage of time accesses were not in the cache. Maximum number of services on the BlockableServices queue in this log interval. requiring the requestor to spin until the resource is unblocked. Aggregated sample point-to point high priority queue length on all Bynets. FLOAT ALL Resource Usage Macros and Tables 235 . and daemon services. receiver resource usage.blocks = immediate grants. regardless of which net. Services can be blocked for a variety of reasons including receiver flow control.

Appendix B: ResUsageIpma Table Spare Columns Column Name CacheMiss CacheWrites CacheWriteThrus Type of Data count count count Description Number of times accesses were not in the cache. Total number of writes to the cache. so a displayed value of 500 represents a COD value of 50. The spare column fields expand to values 00 . Number of cache write through accesses (write bypasses the cache and goes straight to main memory). delayed writes of data previously written to the cache by the CPU. SpareCount02. so that column names would be SpareCount01.0%. that is. SpareTrack01. The SpareTmon00 field contains the COD value. 236 Resource Usage Macros and Tables . Spare tracked statistic. Column Name SpareCount[00-02] SpareTrack[00-02] SpareTmon00 Type of Data count track n/a Description Spare counted statistic. SpareTmon[01-02] count Spare time monitored statistic. and so on. Note: This value is valid only on SUSE Linux Enterprise Server 10 systems and is a single value for the entire system. The value represents the COD value in one tenths of a percent.02. Data Type FLOAT FLOAT FLOAT Invalid Platform ALL ALL ALL CacheWriteBacks count FLOAT ALL Spare Columns The ResUsageIpma table has nine spare columns (one of which is being used) as shown in the table below. Number of cache write back accesses from cache to main memory.

but this field will still contain the time value when the entry should have been logged. by no more than one or two seconds). Invalid Platform Column Name Type of Data Description HOUSEKEEPING COLUMNS Data Type RELATIONAL PRIMARY INDEX COLUMNS These columns taken together form the primary index. 5600C or 5555H. Type of node. See the Secs and NominalSecs columns. For example. entries may be logged late (typically. This table is intended for internal use only. Nominal time of the log entry. virtual processor information. representing the per node system family type.APPENDIX C ResUsageIvpr Table The ResUsageIvpr table includes resource usage data for system-wide. NodeId n/a Identifies the Node upon which the vproc resides. a node in chassis 9 of cabinet 3 has a node ID of ‘003-09’. Note: SMP nodes have a chassis and cabinet number of 1. MISCELLANEOUS HOUSEKEEPING COLUMNS GmtTime n/a Greenwich Mean Time is not affected by the Daylight Savings Time adjustments that occur twice a year. FLOAT INTEGER DATE FLOAT NodeType n/a CHAR(8) Resource Usage Macros and Tables 237 . the node ID of an SMP node is ‘001-01’. For example. where CCC denotes the three-digit cabinet number and MM denotes the two-digit chassis number of the node. TheDate TheTime n/a n/a Date of the log entry. For example. The Node ID is formatted as CCC-MM. The following table describes the ResUsageIvpr table columns. Note: Under conditions of heavy system load.

Identifies the summarization status of this row. Data Type INTEGER Invalid Platform VprType Secs n/a n/a The values can be NODE. This field is useful when performing data calculations with small elapsed times where the difference between centisecond-based data and whole seconds results in a percentage error. The VprId can be any of the following depending on the type: • AMP vprocs: numbered upward from 0. Actual number of seconds in the log period represented by this row. 5=VSS). AMP. CHAR(4) SMALLINT CentiSecs n/a Number of centiseconds in the logging period. 3=GTW. and 'S' if the row is a summary row. INTEGER NominalSecs SummaryFlag n/a n/a SMALLINT CHAR 238 Resource Usage Macros and Tables . RSG. • VSS vprocs: numbered downward from 10238. 1 = AMP. Each existing vproc type range should not overlap into the range of another existing vproc type on the system. but can be different in three cases: • The first interval after a log rate change • A sample logged late because of load on the system • System clock adjustments affect reported Secs Useful for normalizing the count statistics contained in this row. GTW. The vproc numbers within each type range are contiguous. • NODE vprocs: numbered upward from 16384. 0 = NODE. A specified or nominal number of seconds in the logging period. • PE vprocs: numbered downward from 16383. to a persecond measurement. Normally the same as NominalSecs. • RSG vprocs: numbered downward from 9215. 4=RSG. PE. • GTW vprocs: numbered upward from 8192.Appendix C: ResUsageIvpr Table Column Name VprId Type of Data n/a Description Identifies the vproc number (non-Summary Mode) or the vproc type (Summary Mode. for example. Possible values are 'N' if the row is a nonsummary row. 2 = PE. or TVS (see Teradata Virtual Storage).

then the row contains modified data fields. This is especially important in coexistence systems where the number of CPUs can vary across system nodes. Note: To calculate the average number of processes. STATISTICS COLUMNS PROCESS SCHEDULING COLUMNS SMALLINT Work Type Summary Columns Identify the distribution of work types among allocated processes. where [i] expands to the values 0-15.Appendix C: ResUsageIvpr Table Column Name NCPUs Type of Data n/a Description Number of CPUs on this node. FLOAT CollectIntervals n/a The number of gather periods per reporting period. For more information. ProcWorkType[i]Sum count Total number of processes of work type i during each log interval. the value is the number of Gather operations that have been performed during the period. Each entry below represents 16 columns. FLOAT ALL FLOAT ALL Resource Usage Macros and Tables 239 . then none of the data fields in the row have been updated during the logging period. Data Type SMALLINT Invalid Platform Active count Controls whether or not the rows will be logged to the ResUsage tables if Active Row Filter Mode is enabled. then the rows that have a zero Active field value will not be logged to the ResUsage tables. If Active is set to: • a non-zero value. see the CollectIntervals column. In the Collect Buffer and Log Buffer. This field is useful for normalizing the CPU utilization field values for the number of CPUs on the node. divide this value by the CollectIntervals value. The total of the following average columns approximately equals total Process Allocations in ResUsageSpma table. for example. ProcWorkType[i]Max max Maximum number of processes of work type i during each log interval. The CollectIntervals value is the number of gather periods per reporting period. ProcWorkType2Sum. • a zero value. For example. if Active Row Filter Mode is enabled. This number can vary from one period to the next.

FLOAT 240 Resource Usage Macros and Tables . MsgHashReads MsgHashWrites MsgProcReads MsgProcWrites MsgGroupReads MsgGroupWrites MsgLocalReads MsgLocalWrites MsgChanReads MsgChanWrites MsgMboxReads MsgMboxWrites count count count count count count count count count count count count Number of hash messages read by this vproc. since Hash + Proc + Group + Local messages = Chan + MBox messages. Two times are provided: message transmission to mailbox delivery (MDelivery) and mailbox delivery to process delivery (PDelivery). MsgHashMDelivery count Total amount of time read hash messages took for mailbox delivery. Number of processor messages written by this vproc. and so on. where: • • • • • hash messages (Hash) are data sent to a destination through its primary or fallback hash value processor messages (Proc) are data sent to a destination through a vproc ID group messages (Group) are broadcasted messages to be received by members of a group local messages (Local) are messages communicated locally within the node channel messages (Chan) are data sent between vprocs through channel IDs for purposes of a private conversation to perform functions such as row redistribution. • mailbox messages (Mbox) are data sent between vprocs through mailbox IDs for similar purposes as channel messages. Number of group messages read by this vproc. Number of channel messages written by this vproc. group and local messages to reach their destination.Appendix C: ResUsageIvpr Table Column Name Type of Data Description NET COLUMNS Data Type Invalid Platform Message Type Columns Subdivide all messages sent and received into the type of message. Number of local messages read by this vproc. A duplicated accounting is done with two different perspectives. Number of local messages written by this vproc. Number of hash messages written by this vproc. Number of mailbox messages written by this vproc. Number of processor messages read by this vproc. FLOAT FLOAT FLOAT FLOAT FLOAT FLOAT FLOAT FLOAT FLOAT FLOAT FLOAT FLOAT Message Delivery Times Columns Identify the time it took for hash. Number of channel messages read by this vproc. Number of group messages written by this vproc. Number of mailbox messages read by this vproc. processor.

Number of times a monitor yield was requested. Total amount of time read processor messages took for process delivery. Total amount of time read group messages took for process delivery. Each cylinder migration event implies 1 logical read and 3 logical writes of the cylinder index.blocks = immediate grants.) Only logical I/Os and the amount moved (KBs) for data blocks are identified.) Number of times entry into a monitor was deadlocked. Total amount of time read hash messages took for process delivery. Total amount of time read group messages took for mailbox delivery. FLOAT FLOAT FLOAT FLOAT FLOAT FLOAT FLOAT ALL Resource Usage Macros and Tables 241 .Appendix C: ResUsageIvpr Table Column Name MsgProcMDelivery MsgGroupMDelivery MsgLocalMDelivery MsgHashPDelivery MsgProcPDelivery MsgGroupPDelivery MsgLocalPDelivery Type of Data count count count count count count count Description Total amount of time read processor messages took for mailbox delivery. Number of times entry into a monitor was requested. Total amount of time read local messages took for mailbox delivery. FileDbCylMigrIO FileDbCylMigrKB count count Number of data block logical I/Os due to cylinder migration. Data Type FLOAT FLOAT FLOAT FLOAT FLOAT FLOAT FLOAT Invalid Platform GENERAL CONCURRENCY CONTROL COLUMNS Monitor Management Columns Identify monitor activities for Teradata Database concurrency control. (Event counts are found in ResUsageSvpr. KBs moved by FileDbCylMigrIOs. Only permanent tables (including append and transient journal tables) are migrated. FILE SYSTEM COLUMNS Cylinder Overhead Columns Further identify file system cylinder split/migrate (CylMigr) overhead performed when cylinders can not accommodate new data. Number of times entry into a monitor was blocked. MonAllocates MonEnters MonBlocks MonDeadlocks MonYields count count count count count Number of monitors allocated. (requests . Total amount of time read local messages took for process delivery.

so no attempt is made to separate permanent and spool data segments. • Each cylinder defragment event implies 1 logical cylinder index read and 1 logical cylinder index write. • Only logical I/Os and the amount moved (KBs) are identified. • Only logical I/Os and the amount moved (KBs) are identified. except the amount moved for cylinder indexes because they can be calculated by multiplying the current cylinder index fixed size and the I/Os. Number of data block logical I/Os performed for a block update operation requiring no new data blocks. KBs moved by the FileDbCylDefragIO. FileDbCylDefragIO FileDbCylDefragKB count count Number of permanent data block logical I/Os due to cylinder defragmentation. except for the amount moved for cylinder indexes because they can be calculated by multiplying the current fixed cylinder index size by the I/Os. FileCiUpd0IO count Number of cylinder index logical I/Os performed for a block update operation requiring no new data blocks. it can be ‘in place’ and requires no new data blocks. Number of cylinder index logical I/Os performed for a block update operation requiring 1 new data blocks. FLOAT FilePDbMCylPackIO FilePDbMCylPackKB count count FLOAT FLOAT Cylinder Defragmentation Overhead Columns Identify background file system overhead associated with fragmented free space to achieve one large free space within that cylinder (CylDefrag). KBs moved by FilePDbCylPackIOs. KBs moved by FileDbUpd0IO.Appendix C: ResUsageIvpr Table Column Name Type of Data Description Data Type Invalid Platform Cylinder MiniCylPack Overhead Columns Identify file system overhead associated with MiniCylPacks (MCylPack) that get performed to make available a free cylinder when one is needed but not available. When a block is updated. Cylinder defragments are done on cylinders containing permanent tables (including append and transient journal tables) only. • MiniCylPacks are done on cylinders containing permanent tables (including append and transient journal tables) only. FilePCiMCylPackIO count Number of permanent cylinder index logical I/Os due to performing MiniCylPack operations. Data block updates should only be performed on permanent tables (including append and transient journal tables). Only logical I/Os and the amount moved (KBs) are identified. or it could spill over the current data block and require 1. FLOAT FileDbUpd0IO count FLOAT FileDbUpd0KB FileCiUpd1IO count count FLOAT FLOAT 242 Resource Usage Macros and Tables . 3 or more new data blocks in addition to the current data block. • Event counts are found in ResUsageSvpr. FLOAT FLOAT Data Block Update Operations Columns Identify the file system operations required when a data block is being updated (BlkUpd). • Event counts are found in ResUsageSvpr. Number of permanent data block logical I/Os due to performing MiniCylPack operations. 2.

FileTJBufUpdates count Number of transient journal buffer updates. Number of spool data blocks created. KBs created by FilePDbCreates. Number of cylinder index logical I/Os performed for a block update operation requiring over 3 new data blocks. KBs created by FileSDbCreates. Number of cylinder index logical I/Os performed for a block update operation requiring 3 new data blocks. Number of cylinder index logical I/Os performed for a block update operation requiring 2 new data blocks. FLOAT FilePDbCreateKB FileSDbCreates FileSDbCreateKB count count count FLOAT FLOAT FLOAT Transient Journal Overhead Columns Identify file system overhead associated with maintaining a transient journal (TJ). KBs moved by FileDbUpd3IO. Data Type FLOAT Invalid Platform FileDbUpd1KB FileCiUpd2IO count count FLOAT FLOAT FileDbUpd2IO count FLOAT FileDbUpd2KB FileCiUpd3IO count count FLOAT FLOAT FileDbUpd3IO count FLOAT FileDbUpd3KB FileCiUpdNIO count count FLOAT FLOAT FileDbUpdNIO count FLOAT FileDbUpdNKB count FLOAT Data Block Creations Columns Identify the file system operations required when a data block is being created (BlkCreate). KBs moved by FileDbUpdNIO.Appendix C: ResUsageIvpr Table Column Name FileDbUpd1IO Type of Data count Description Number of data block logical I/Os performed for a block update operation requiring 1 new data blocks. FilePDbCreates count Number of permanent table (including append and transient journal tables) data blocks created. KBs moved by FileDbUpd1IO. It does not include data blocks created due to any of the new data blocks created when a data block was updated as described in the Data Block Update Operations Columns description. Number of data block logical I/Os performed for a block update operation requiring 2 new data blocks. KBs moved by FileDbUpd2IO. Number of data block logical I/Os performed for a block update operation requiring over 3 new data blocks. Number of data block logical I/Os performed for a block update operation requiring 3 new data blocks. FLOAT Resource Usage Macros and Tables 243 .

FilePRowReadInit FilePRowReadCont FilePRowReplace FilePRowInsert FilePRowDelete FilePRowAppend FileSRowReadInit FileSRowReadCont FileSRowReplace FileSRowInsert FileSRowDelete FileSRowAppend FileAPtRowReadInit FileAPtRowReadCont FileAPtRowReplace FileAPtRowInsert FileAPtRowDelete FileAPtRowAppend count count count count count count count count count count count count count count count count count count Number of requests for an initial permanent row read. Number of requests for a permanent row delete. Number of requests for a permanent row insert. spool (S) or user append table / permanent journal table (APt). Number of requests for a spool row delete. FLOAT FLOAT FLOAT FLOAT FLOAT FLOAT FLOAT FLOAT FLOAT FLOAT FLOAT FLOAT FLOAT FLOAT FLOAT FLOAT FLOAT FLOAT File System Multi-Row Requests Columns Identify the significant multi-row requests made by application software on the file system. FLOAT 244 Resource Usage Macros and Tables . Rows are distinguished as permanent data (P). Number of requests for an append row replace/ update. Rows are distinguished as permanent data (P). Number of requests for a row append. Number of requests for an initial append row read. Number of requests for a continued permanent row read. Number of requests for a continued spool row read. spool (S) or user append table / permanent journal table (APt). Number of requests for an append row append. Number of requests for an append row insert. FilePBlkRead count Number of requests for a permanent data block read.Appendix C: ResUsageIvpr Table Column Name Type of Data Description Data Type Invalid Platform File System Single-Row Requests Columns Identify the significant single-row requests made by application software on the file system. Number of requests for a row append. Number of requests for a spool row replace/ update. Number of requests for an initial spool row read. Number of requests for an append row delete. Number of requests for a spool row insert. Number of requests for a permanent row replace. Number of requests for an continued append row read.

Number of requests for a spool data multi-row update. Number of requests for spool table sort. Number of requests for a spool data multi-row delete. Number of requests for a spool data block read. Number of requests for a permanent table multi-block read. Number of requests for a permanent data multi-row insert. Number of requests for a spool table delete. Number of requests for a spool multi-row delete. Number of requests for a spool data block replace. Data Type FLOAT FLOAT FLOAT FLOAT FLOAT FLOAT FLOAT FLOAT FLOAT FLOAT FLOAT FLOAT FLOAT FLOAT FLOAT FLOAT FLOAT FLOAT FLOAT FLOAT FLOAT Invalid Platform Resource Usage Macros and Tables 245 . Number of requests for a spool data multi-row insert.Appendix C: ResUsageIvpr Table Column Name FilePBlkReplace FilePRowNDel FilePRownins FilePRowNUpd FilePSortable FilePTabdelete FilePTabdelra FilePTabmrows FilePTabrblocks FileSBlkRead FileSBlkReplace FileSRowNDel FileSRownins FileSRowNUpd FileSSortable FileSTabdelete FileSTabdelra FileSTabmrows FileSTabrblocks FileAPtBlkRead FileAPtBlkReplace Type of Data count count count count count count count count count count count count count count count count count count count count count Description Number of requests for a permanent data block replace. Number of requests for a spool table multiblock read. Number of requests for a permanent data multi-row update. Number of requests for a multi-row delete. Number of requests for permanent table sort. Number of requests for a permanent table delete. Number of requests for a spool table modification. Number of requests for an append data block replace. Number of requests for an append data block read. Number of requests for a permanent table modification. Number of requests for a permanent data multi-row delete.

The system uses full-cylinder read mode. but the block count would still be roughly proportionate to the I/O load. This is a reasonable approximate measure of the I/O load. The modification can either be an update or a delete of an existing WAL or TJ record. Number of requests for an append table multiblock read. Number of requests for an append data multirow update. TRANSIENT JOURNAL MANAGEMENT COLUMNS Transient Journal Purge Overhead Columns Identify the background overhead associated with the occasional transient journal purge operation. Number of requests for an append data multirow delete.Appendix C: ResUsageIvpr Table Column Name FileAPtRownins FileAPtRowNDel FileAPtRowNUpd FileAPtSortable FileAPtTabdelete FileAPtTabdelra FileAPtTabmrows FileAPtTabrblocks Type of Data count count count count count count count count Description Number of requests for an append data multirow insert. Data Type FLOAT FLOAT FLOAT FLOAT FLOAT FLOAT FLOAT FLOAT Invalid Platform File System Transient Journal Requests Column Identifies the significant transient journal requests made by application software on the file system. TJPurges TJDbPurgeReads count count The number of purge passes in which a blockby-block scan is done. FileTJCalls FileTJDbUpdates count count Number of transient journal calls. The number of blocks actually mapped in during the purge scan. Number of requests for an append table sort. Number of requests for an append table delete. Number of requests for an append table modification. FLOAT FLOAT FLOAT FLOAT 246 Resource Usage Macros and Tables . Number of requests for an append multi-row delete. Number of WAL data blocks modified.

Appendix C: ResUsageIvpr Table Column Name TJDbPurgeDeletes Type of Data count Description The number of blocks mapped in during the scan that were included in the ranges of blocks that were deleted. However. then the quick purge method is not effective. the WAL log. as they are automatically filtered out. a pointer to the oldest active transaction is maintained which is a quicker method. This does not require any scanning and the system cannot definitely determine how many blocks actually get deleted. the ratio cannot be interpreted quite so simply because: 1 The range of deleted blocks could include Data Type FLOAT Invalid Platform blocks that were not actually mapped in (and therefore not counted). neither TJDbPurgeReads nor FLOAT TJDbPurgeDeletes gets incremented during a normal purge pass. with WAL. PurgeTJ() can simply compute the bounds of the range of records that can be deleted in the part of the WAL/TJ that precedes the start of the oldest transaction. Before WAL. Therefore. Under typical conditions. TJDbPurgeDeletes (continued) count 2 Post-WAL. there are probably relatively few such blocks. so the proportion of WAL-only blocks would probably be significant. the system reverts back to the full scan method. If the oldest transaction remains open for a long time. Instead of scanning the active data blocks. WRITE AHEAD LOGGING COLUMNS Identify the log-based file system recovery scheme in which modifications to permanent data are written to a log file. there will be relatively few TJ records written. Therefore. Resource Usage Macros and Tables 247 . But during periods when the system workload is dominated by MultiLoad/FastLoad work. TJ and WAL records are typically generated in an interleaved sequence by regular SQL transactions. The TJDbPurgeReads and TJDbPurgeDeletes are only incremented during a full scan. the ratio of deletes to reads would have been a useful measure of the effectiveness of the purge processing. Blocks that contain only WAL records are not mapped in during the scan.

Number of times a request to force transient journal records within the WAL log to be written to disk has been issued. A single append call can append multiple transient journal rows. summarizing the vprocs of that type on that node. The block can contain either TJ records. Data Type FLOAT Invalid Platform FileTJAppends count FLOAT FileTJFlush count FLOAT FileWDBCreates count FLOAT FileWFlush count FLOAT FileWRowDelete FileWTabDelRa count count FLOAT FLOAT Summary Mode When Summary Mode is active for the ResUsageIvpr table. A single append call can append multiple rows. Subtracting FileTJAppends from this counter results in the number of times nontransient journal rows were appended to the WAL log. one row is written to the database for each type of vproc on each node in the system. Number of times rows were deleted from the WAL log. Number of WAL data blocks created. Number of requests for a WAL multi-row delete. A transient journal append by itself does not imply a write of a WAL block. Number of times transient journal records were appended to the WAL log. for each log interval. You can determine if a row is in Summary Mode by checking the SummaryFlag column for that row. Subtracting FileTJFlush from this counter results in the number of times a non-transient journal WAL flush was issued. An increment of this counter may or may not result in an I/O depending on whether the request was to flush records that were already on disk. 248 Resource Usage Macros and Tables . WAL records or both Number of times a request to force any record in the WAL log to be written to disk has been issued. An increment of this counter may or may not result in an I/O depending on whether the request was to flush records that were already on disk. nor a WAL Cylinder Index (WCI) modification.Appendix C: ResUsageIvpr Table Summary Mode Column Name FileWAppends Type of Data count Description Number of times a record was appended to the WAL log.

and so on.0%.09. Type of Data count track n/a Column Name SpareCount[00-09] SpareTrack[00-09] SpareTmon00 Description Spare counted statistic. Resource Usage Macros and Tables 249 . Spare Columns The ResUsageIvpr table has 30 spare columns (one of which is being used) as shown in the table below. SpareTrack02. Note: This value is valid only on SUSE Linux Enterprise Server 10 systems and is a single value for the entire system. The SpareTmon00 field contains the COD value. so that column names would be SpareCount01. normally. The value represents the COD value in one tenths of a percent. SpareTmon07. SpareTmon[01-09] count Spare time monitored statistic. Spare tracked statistic.Appendix C: ResUsageIvpr Table Spare Columns IF the SummaryFlag column value is… ‘S’ ‘N’ THEN the data for that row is being logged… in Summary Mode. so a displayed value of 500 represents a COD value of 50. The spare column fields expand to values 00 .

Appendix C: ResUsageIvpr Table Spare Columns 250 Resource Usage Macros and Tables .

there is more than one definition of partition. Each AMP owns a portion of the overall database storage. Partitions are the subdivision of vproc software processes into 32 semi-isolated domains. Partitions are the primary mechanism used by Teradata Database for managing parallel programs. The partitions here refer to the following Parallel Database Extensions (PDE) and vproc definition: • A partition is a collection of tasks and associated resources grouped within a virtual processor according to the function of the tasks. PE RSG VSS Resource Usage Macros and Tables 251 . and do not appear in the output of the Vproc Manager utility. such as executing database queries. security validation. In all other vproc types. Partition reservation is as follows: • • Partitions 0 through 6 are reserved by PDE Partitions 7 through 47 are for use by Teradata Database The table listed under “Partition Assignment Listing” on page 252 describes the individual partitions. Partition 11 is the AWT Partition. and for relaying dictionary changes to the Teradata Meta Data Services utility. Partition 11 is unused. Another partition description is only meaningful in a dialog between client programs and Teradata Database. specifically the CONNECT parcel. query parsing. in an AMP vproc. This concept of partitions is centered in the CLIv2 interface. • For example. Teradata Database uses the following vprocs: Vproc Type AMP GTW Node Description Access module processors perform database functions. query optimization. There are multiple partitions within a single virtual processor. It has nothing to do with PDE vproc partitions. Node vprocs cannot be externally manipulated. The node vproc handles PDE and operating system functions not directly related to AMP and PE work. Parsing engines perform session control. Manages Teradata Database storage. AMPs acquire their portions of database storage through the TVS vproc. Relay Services Gateway provides a socket interface for the replication agent. Gateway vprocs provide a socket interface to Teradata Database. but is a way of enforcing rules about what a client session is allowed to do and of keeping client sessions isolated from each other.APPENDIX D Partition Assignments With regards to Teradata Database. and query dispatch.

252 Partition: No.Partition usage is also discussed under “CPU Utilization Columns” on page 127 in the Chapter 13: “ResUsageSvpr Table” chapter. Partition Assignment Listing The following table lists the Node. The symbol used in the Partition Assignment Listing… —— ? Indicates… partition is unused. GTW. and VSS (allocator/node agent) usage of PDE vproc partitions by PDE and Teradata Database. 0 Resource Usage Macros and Tables 1 2 3-6 Name Kernel 7 Service Table Conventions The following table describes the table symbols used in the partition assignments table below. activity has been observed but not identified. AMP. and PE (Parsing Engine). Usage in Vprocs of Type: Node PDE daemons AMP —— PE GTW RSG VSS System Debugger Console Interactive 1 through 4 System Debugger tasks Console Supervisor —— —— Console interactive partition programs —— Console utilities . and RSG.

28 29 30 Filesys Gateway AWT Session Dispatch Unused Startup [unused] RSS Startup DDF Server [unused] [unused] [unused] [unknown] [unused] —— gtw processes —— —— —— —— —— —— File system rss startup DDF services —— —— —— —— —— —— MDS program rsgmain —— .22 23 .Resource Usage Macros and Tables 253 Partition: No. 8 Name CnsProc Usage in Vprocs of Type: Node —— AMP Host Utility console procedures File System processes —— AMP Worker Tasks —— Session Control tasks Dispatcher tasks Unused Startup tasks PE —— GTW RSG VSS 9 10 11 12 13 14 15 16 17 18 19 20 .

254 Partition: No. 31 32 33 Name Allocator Node Agent Clique Coordinator Usage in Vprocs of Type: Node —— —— —— AMP PE GTW RSG VSS Allocator services Node Agents services Clique Coordinator services 34 .47 [unused] —— Resource Usage Macros and Tables .

Glossary AG AMP Allocation Group Access Module Processor API Application Programming Interface AWT AMP Worker Task Banyan Network (high-speed connection) Database Window Data Definition Language File Segment Teradata Gateway BYNET DBW DDL FSG GTW I/O Input/Output LAN Local Area Network MPP Massively Parallel Processing NUPI Nonunique Primary Index PDE Parallel Database Extensions. This refers to a node. It provides Teradata Database the ability to run in a parallel environment. and more. execute vprocs. The subsystem that logs Resource Usage data from RSS to the ResUsage tables. PMPC APIs Performance Monitor and Production Control Application Programming Interfaces PP Performance Period Relational Database Management System RDBMS ResUsage Resource Usage. PDE is a software interface layer between the operating system and the Teradata Database software. PE PG PMA Parsing Engine Performance Group Processor Module Assembly. RSG Relay Services Group Resource Usage Macros and Tables 255 .

and provides the ability to access the statistics through an API.Glossary RSS Resource Sampling Subsystem. ResUsage uses the RSS data from the RSS API to log data to the selected ResUsage tables. The RSS provides a method to gather statistics from across the Teradata Database system. Virtual Processor WD Workload Definition WCI WAL Cylinder Index 256 Resource Usage Macros and Tables . Virtual BYNET for a single-node. SMP TCHN Symmetric Multi-Processing Teradata Channel Teradata Active System Management Teradata ASM TPA VNET vproc Trusted Parallel Application Virtual Network.

description of 31 C Clearing old resource usage data 36 Co-existing node macros. meaning in macro outputs 176 A AMP information macros 177. See DBW Deleting old resource usage data 36 DISABLE LOGONS effects on logging 36 E Example executing ResAmpCpuByGroup macro 35 executing ResCPUByAmp macro 34 executing ResCPUByAmpOneNode macro 35 Resource Usage Macros and Tables 257 . See ByGroup macros Collecting resource usage data 17 CollectIntervals using to determine average 42 CPU use by AMP macro output columns Awt User Exec% 182 Awt User Serv% 182 Misc User Exec% 182 Misc User Serv% 182 Total Busy% 182 Total User Exec% 182 Total User Serv% 182 CPU use by AMPs macros function 182 input format examples 182 normalized viewing 184 output column descriptions 182 ResAmpCpuByGroup 182 ResCPUByAMP 182 ResCPUByAMPOneNode 182 usage notes 182 CPU use by each node macros function 189 input format examples 189 output examples 189 ResCPUByGroup 189 ResCPUByNode 189 ResCPUOneNode 189 usage notes 189 CPU use by nodes macro output columns I/O Wait % 190 Total Busy % 190 Total User Exec % 190 Total User Serv % 190 CPU use by PEs macro output columns Disp User Exec% 186 Disp User Serv% 186 Misc User Exec% 186 Misc User Serv% 186 Pars User Exec% 186 Pars User Serv% 186 Ses User Exec% 186 Ses User Serv% 186 Total Busy% 187 Total User Exec% 187 Total User Serv% 187 CPU use by PEs macros function 186 input format examples 186 normalized viewing 188 output column descriptions 186 output examples 186 ResCPUByPE 186 ResCPUByPEOneNode 186 ResPeCpuByGroup 186 usage notes 186 D Database commands SET LOGTABLE 28 SET RESOURCE 28 SET SUMLOGTABLE 28 Database Window Supervisor setting logging rates 28 Database Window. 182 table 125 view 152 Average values determining for gather period 42 B ByGroup macros.Index Symbols ?.

Index ResAmpCpuByGroup macro report 183 ResAWT macro report 180 ResAWTByAMP macro report 180 ResAWTByNode macro report 181 ResCPUByAMP macro report 183 ResCPUByAMPOneNode macro report 183 ResCPUByGroup macro report 191 ResCPUByNode macro report 190 ResCPUByPE macro report 187 ResCPUByPEOneNode macro report 187 ResCPUOneNode macro report 190 ResHostByGroup macro report 195 ResHostByLink macro report 194 ResHostOneNode macro report 195 ResLdvByGroup macro report 198. See Macro input M Macro input 29 Macro output format general format 175 ID 175 statistics 175 Macro statistics. 201 ResLdvByNode macro report 197. 218 table 85 view 158 F format. description 41 L Logging costs 23 optimizing 23 rates 21 resource usage data 17 which tables to enable 20 Logging rates definition 21 minimum 22 recommended values 22 using SET RESOURCE 28 Logical device information macros 196. datatypes 176 Macro syntax element for all nodes 33 for co-existing nodes 33 for multinodes 33 FromDate 33 FromNode 34 FromTime 33 Node 34 G GmtTime 38 H Host communication traffic macro output columns Avg ReqQ Len 193 Blk Read Fail % 193 Blk Write Fail % 193 Blks Read/Sec 193 Blks Write/Sec 193 Host Type 193 KBs Read/Sec 193 KBs Write/Sec 193 258 Resource Usage Macros and Tables . 200. 201 ResLdvOneNode macro report 197 ResMemByGroup macro report 204 ResMemMgmtByNode macro report 204 ResMemMgmtOneNode macro report 204 ResNetByGroup macro report 207 ResNetByNode macro report 206 ResNetOneNode macro report 206 ResNode macro report 211 ResNodeByGroup macro report 212 ResNodeByNode macro report 212 ResOneNode macro report 211 ResPeCpuByGroup macro report 187 ResPsByGroup macro report 216 ResPsByNode macro report 215 ResPsByNodeWDJoin macro report 216 ResVdskByGroup macro report 219 ResVdskByNode macro report 219 ResVdskOneNode macro report 219 EXECUTE MACRO. syntax elements 32 KBs/Blk Read 193 KBs/Blk Write 193 Max ReqQ Len 193 Msgs/Blk Read 193 Msgs/Blk Write 193 Host communications traffic information macros function 192 macros input format examples 192 macros usage notes 192 table 79 view 157 Host communications traffic macros output column descriptions 193 ResHostByGroup 192 ResHostByLink 192 ResHostOneNode 192 I Invalid platform columns.

See Summary macros P Parameters. 218 KB/ I/O 196. meaning in macro outputs 176 R Rates logging rate definition 21. macros use of 32 Partition Assignments listing 252 reserved for 251 table convention 252 Partitions. types of all-node 29 ByGroup 29 multiple-node 29 one-node 29 Memory management by node macros function 202 input format examples 202 output column descriptions 202 ResMemByGroup 202 ResMemMgmtByNode 202 ResMemMgmtOneNode 202 usage notes 202 Memory management macro output columns # Proc Swp 203 % Mem Free 202 Ages/Sec 203 Aloc Fail % 203 KB/Swp Drp 203 KB/Swp Rd 203 KB/VPR Aloc 203 P+S Drps/Sec 203 P+S I/O % 203 P+S Rds/Sec 203 P+S Wrts/Sec 203 Pg Drps/Sec 203 Pg Rds/Sec 203 Pg Wrts/Sec 203 Swp Drps/Sec 203 Swp Rds/Sec 203 Text Alocs/Sec 202 VPR Alocs/Sec 202 MULTISET table why resusage tables are created as 38 Node network traffic macros function 205 input format examples 205 output column descriptions 205 ResMemMgmtByNode 205 ResNetByGroup 205 ResNetOneNode 205 usage notes 205 Nonunique primary index 38 NUPI. See Nonunique primary index O Occasional event data 38 One-Node macros. description of 31 One-Node macros.Index ToDate 33 ToNode 34 ToTime 33 Macros logging rates for tables 176 usage notes 176 Macros. 189 view 148 Node network traffic macro output columns % Brd 206 % PtP 206 % Retries 205 KB/IO 206 Total IOs/Sec 206 Total Reads/Sec 205 Total Writes/Sec 206 Q Question marks. description 31 Overall resource usage information. 200 Resource Usage Macros and Tables 259 . definition 251 PE information macros 186 table 125 view 154 Priority Scheduler information macros function 213 macros input format examples 213 macros usage notes 213 Priority Scheduler macros ResPsByGroup 213 ResPsByNode 213 ResPsByNodeWDJoin 213 Purging old resource usage data 36 N Node information macros 177. 199 Out Rqst Time % 197. 199. 22 recommended values 22 rules for setting 22 Raw disk drive traffic macro output columns Avg I/O Resp 196.

199 ResLdvByGroup 196 ResLdvByNode 196 ResLdvOneNode 196 usage notes 196. 199 input format example 196. definition listing 152 ResCPUUsageByPEView. 199 260 Resource Usage Macros and Tables . 200. definition listing 154 ResGeneralInfoView. 199 what it reports 196. 199 Relational Primary Index 38 ResAmpCpuByGroup macro column descriptions 182 input format example 182 sample output 184 usage notes 182 what it reports 182 ResAWT macro column description 177 input format example 177 output column descriptions 178 sample output 180 usage notes 177 what it reports 177 ResAWTByAMP macro column description 177 input format example 177 output column descriptions 178 sample output 180 usage notes 177 what it reports 177 ResAWTByNode macro column description 177 input format example 177 output column descriptions 178 sample output 181 usage notes 177 what it reports 177 ResCPUByAMP macro column descriptions 182 input format example 182 sample output 183 usage notes 182 what it reports 182 ResCPUByAMPOneNode macro column descriptions 182 sample output 184 usage notes 182 what it reports 182 ResCPUByGroup macro column descriptions 189 sample output 191 usage notes 189 what it reports 189 ResCPUByNode macro column descriptions 189 input format example 189 sample output 190 usage notes 189 what it reports 189 ResCPUByPE macro column descriptions 186 input format example 186 sample output 187 usage notes 186 what it reports 186 ResCPUByPEOneNode macro column descriptions 186 sample output 188 usage notes 186 what it reports 186 ResCPUOneNode macro column descriptions 189 sample output 190 usage notes 189 what it reports 189 ResCPUUsageByAMPView. definition listing 148 ResHostByGroup macro column descriptions 193 sample output 195 usage notes 192 what it reports 192 ResHostByLink macro column descriptions 193 input format example 192 sample output 194 usage notes 192 what it reports 192 ResHostOneNode macro column descriptions 193 sample output 195 usage notes 192 what it reports 192 ResLdvByGroup macro column descriptions 196. 199 Raw disk drive traffic macros function 196. 199 ResLdvByNode macro column descriptions 196. 199 ResLdvOneNode macro column descriptions 196. 199 sample output 197. 199 input format examples 196. 201 usage notes 196.Index Reads/Sec 196. 199 what it reports 196. 199 sample output 198. 199 Writes/Sec 196. 199 output column descriptions 196. 201 usage notes 196.

definition listing 157 ResSldvGroupView.Index usage notes 196. 199 what it reports 196. definition listing 166 ResUsage tables columns ending in "Avg" 42 enabling Summary Mode 28 invalid platform columns 41 naming convention 37 primary index 38 reporting Summary Mode 42 Resource Usage Macros and Tables 261 . definition listing 155 ResShstGroupView. 199 ResMemByGroup macro column description 202 sample output 204 usage notes 202 what it reports 202 ResMemMgmtByNode macro column description 202 input format example 202 sample output 204 usage notes 202 what it reports 202 ResMemMgmtOneNode macro column description 202 sample output 204 usage notes 202 what it reports 202 ResNetByGroup macro column description 205 sample output 207 usage notes 205 what it reports 205 ResNetByNode macro column description 205 input format example 205 sample output 206 usage notes 205 what it reports 205 ResNetOneNode macro column description 205 sample output 206 usage notes 205 what it reports 205 ResNode macro column description 208 input format example 208 sample output 211 usage notes 208 what it reports 208 ResNodeByGroup macro column description 208 sample output 212 usage notes 208 what it reports 208 ResNodeByNode macro column description 208 sample output 212 usage notes 208 ResOneNode macro column description 208 sample output 211 usage notes 208 what it reports 208 Resource usage data collecting 17 definition 15 deleting old data 36 functions of 15 logging 17 saving old data 32 Resource usage macros definition 18 example of executing a ByGroup macro 35 example of executing a Multinode macro 34 example of executing a One-Node macro 35 executing 32 syntax for 32 Resource Usage tables. definition listing 158 ResSpsView. definition listing 159 ResSvprView. See ResUsage tables Resource usage views ResCPUUsageByAMPView 152 ResCPUUsageByPEView 154 ResGeneralInfoView 148 ResShstGroupView 157 ResSldvGroupView 158 ResSvprView 166 ResPeCpuByGroup macro column descriptions 186 sample output 188 usage notes 186 what it reports 186 ResPsByGroup macro column description 213 sample output 216 usage notes 213 what it reports 213 ResPsByGroup macro column description 213 ResPsByNode macro column description 213 input format example 213 sample output 215 usage notes 213 what it reports 213 ResPsByNode macro column description 213 ResPsByNodeWDJoin macro sample output 216 ResSawtView.

See One-Node Stopped logging how to re-enable logging 36 Summary macro output columns A+R % of IOs 210 CPU Bsy % 210 CPU Eff % 210 Fre Mem % 210 Ldv Eff % 210 Ldv IOs/Sec 210 Ldv KB/IO 210 Mem Age/Sc 210 Mem Aloc/Sec 210 Mem Fai % 210 ms/Blk 211 Net Rtry % 211 P+S % of IOs 210 Prc Blks/Sec 211 Read % of IOs 210 TMIt IOs/Sec 210 TPtP IOs/Sec 210 WIO % 210 Summary macros function 208. 213 input format examples 208. 213 output column descriptions 210 ResNode 208 262 Resource Usage Macros and Tables . 124 Summary Mode 124 ResUsageSvpr column names 125 spare columns 143 Summary Mode 142 ResVdskByGroup macro column descriptions 218 sample output 219 usage notes 218 what it reports 218 ResVdskByNode macro column descriptions 218 input format example 218 sample output 219 usage notes 218 what it reports 218 ResVdskOneNode macro column descriptions 218 sample output 219 usage notes 218 what it reports 218 RSS logging enabling from ctl and xctl 27 enabling from DBW 28 RSS table settings enabling from Database Window 28 S Saving old resource usage data 32 SET LOGTABLE command 28 SET RESOURCE command 28 SET SUMLOGTABLE command 28 Single-Node.Index ResUsageIpma 20 ResUsageIvpr 20 ResUsageSawt 20 ResUsageShst 20 ResUsageSldv 20 ResUsageSpdsk 91 ResUsageSpma 20 ResUsageSps 20 ResUsageSvdsk 20 ResUsageSvpr 20 types of statistics reported 39 which to enable 20 ResUsageIpma column names 227 gathering method 227 ResUsageIvpr column names 237 gathering method 237 spare columns 249 Summary Mode 248 ResUsageSawt column names 73 Summary Mode 77 ResUsageScpu column names 45 spare columns 48 Summary Mode 48 ResUsageShst spare columns 84 Summary Mode 83 ResUsageSldv column names 79. 116. 85 gathering method 85 spare columns 89 Summary Mode 88 ResUsageSpdsk column names 91 gathering method 91 spare columns 97 Summary Mode 97 ResUsageSpma column names 51 ResUsageSps column names 99 ResUsageSvdsk column names 119 gathering method 119 spare columns 78.

how to read 221 System information macros 177 T Table naming conventions 37 Tables which to enable 19 Teradata ASM and Priority Scheduler macros 213.Index ResNodeByGroup 208 ResNodeByNode 208 ResOneNode 208 usage notes 208. 216 Teradata Virtual Storage 41 V Vdisk device traffic information macros 218 Vdisk drive traffic macro output columns Avg I/O Resp 218 Out Rqst Time % 219 Rd KB/ I/O 218 Read Cnt/ Sec 218 Write Cnt/ Sec 218 Wrt KB/ I/O 218 Vdisk drive traffic macros output column descriptions 218 ResVdskByNode 218 ResVdskOneNode 218 Vdisk logical drive traffic macros ResVdskByGroup 218 Views. 213 Summary Mode description 42 enabling a table 28 using 22 Syntax for deleting old resource usage data 36 executing macros 32 reporting dates in macros 33 Syntax. resource usage data 147 Resource Usage Macros and Tables 263 .

Index 264 Resource Usage Macros and Tables .