data extraction program (cp3kextr) for zcp3000 · ... program – the rmf interval and the extract...

Data Extraction Program

CP3KEXTR v3.82 07/02/18

For zBNA, zCP3000, and zPCR

Tom Harper

Valerie Spencer

[email protected]

Copyright ©2000-2018, IBM Corporation Data Extraction Program 2

Contacts:

Comments, requirements, and questions may be directed to us at any of the electronic

addresses.

Tools Team

Lotus Notes: CPS PC Tools/Gaithersburg/IBM@IBMUS

Internet: [email protected]

Tom Harper

Lotus Notes: Tom Harper/Des Moines/Contr/IBM@IBMUS

Internet: [email protected]

You can download this document (CP3KEXTR.pdf) and the program package (CP3KEXTR.Zip)

from one of the following sources: http://w3-03.ibm.com/support/techdocs/atsmastr.nsf/WebIndex/PRS4144

or: ftp://cpstools.washington.ibm.com (click on the CP3KEXTR folder)

Business Partners may obtain the program from the PartnerWorld system: https://www-304.ibm.com/partnerworld/wps/servlet/ContentHandler/tech_PRS4144

Customers may download the program from this site:

http://www.ibm.com/support/techdocs/atsmastr.nsf/WebIndex/PRS4229

mailto:[email protected]

mailto:[email protected]

http://w3-03.ibm.com/support/techdocs/atsmastr.nsf/WebIndex/PRS4144

ftp://cpstools.washington.ibm.com/

https://www-304.ibm.com/partnerworld/wps/servlet/ContentHandler/tech_PRS4144

http://www.ibm.com/support/techdocs/atsmastr.nsf/WebIndex/PRS4229


Table of Contents

Table of Contents ................................................................................... 3

Recent Program Versions ....................................................................... 4

1. Introduction ...................................................................................... 5

2. Quick Start Guide ............................................................................. 8

2.1 Preparation ................................................................................... 8 2.2 Research ....................................................................................... 8 2.3 Install the Program ........................................................................ 9 2.4 Customize the JCL ...................................................................... 10 2.5 Setup the Input Parameters ......................................................... 11 2.6 Run the Extract .......................................................................... 13

3. Program Input Parameters .............................................................. 14

3.1 Required Parameters ................................................................... 14 3.2 Reporting Interval Parameters ..................................................... 15 3.3 Other Parameters ........................................................................ 17 3.4 BCU - Parameters ....................................................................... 19 3.5 PGN - Workload Parameters ........................................................ 22 3.6 Convert Tape to Disk ................................................................... 23 3.7 Summary Output ........................................................................ 25

4. JCL ................................................................................................. 26

4.1 DD Statement Summary ............................................................. 26 4.2 Return Codes .............................................................................. 27 4.3 JCL for zCP3000 ......................................................................... 28 4.4 JCL for zPCR ............................................................................... 29 4.5 JCL for zBNA .............................................................................. 31 4.6 JCL for SORT .............................................................................. 33 4.7 JCL to Create Multiple Output Files in One Run .......................... 34 4.8 JCL Generalized for All Job Types ................................................ 37

5. SMF Input Data Requirements ........................................................ 39

6. Map Files ......................................................................................... 41

6.1 BCU Map .................................................................................... 41 6.2 PGN Map .................................................................................... 44 6.3 T30 Map ..................................................................................... 48

7. “Data” Output File .......................................................................... 50

7.1 Data File – Type 30 Records......................................................... 51 7.2 Data File – Type 42 Records......................................................... 53 7.3 Data File – Type 14 & 15 Records ................................................ 56

8. Errors .............................................................................................. 58

9. EDF Output ..................................................................................... 59

9.1 EDF Format ................................................................................ 59 9.2 Sample EDF output ..................................................................... 60 9.3 EDF Sections ............................................................................. 63 9.4 EDF Fields ................................................................................. 64


Recent Program Versions

3.82 07-02-18 Do ChpID selection based on R744FPCM bitmask

Add FIDP ChpID list (taken from R744FIDP)

Delete cluster name vector, LPCL

Fix CPU-type matching logic for CPUPWAIT calculation

Fix case where E224, E225, & E226 vectors could have the wrong count

Fix the wrong count issue for LPWLMCx and LPWLMDx

3.79 05-28-18 New CEC section vector MaxPU

New 113 vectors: GHZ, E224, E225, E226

New 113 vector: LPARCPU = 1/CPSP/1mil * B0/ IntSec * 100

New 70(1) vectors: PAGEMB, SCMTL, SCMMA, FR1RN, FRCAM, FRL7M

Change CSMINV vector to minimum, measured in MB

Remove unused 70(1) SWTOT item

Remove unused SIO, SPAGE, and SSTOR items from the SAMP section

3.75 04-09-18 Fix divide error when BCURDBR was 0

3.74 03-23-18 Fix case where EFD vector would fail to include zero elements

3.73 02-28-18 Add new vectors to report hardware group capping

Fix 42(6) DATA file case where certain hyperlink fields were repeated

Add new FCPI and PRBT vectors

Detect case where type 70 CPU Data Section count is zero

Add code to suppress EDF null SR parameter to space

Add vector ABSMSUCP to indicate when absolute MSU capping is active

3.65 01-05-18 Fix case where type 113 vectors could get an invalid number of entries

Fix 42(6) DATA file DS name corruption

Collect IPL information and write it to SYSINFO DD

Fix CSECT entry point name

3.62 12/04/17 Add new SMTEFFTDP vector for average thread density

Fix case where counts were repeated in the SysId list

Make adjustments to the FICON Aggregation analysis

Adjust the CPU MF L4LP formula

Refactor code in the cp2kExtr module

Add HyperLink and micro-second fields to DATA file


1. Introduction

The Extract program, CP3KEXTR, runs on a z/OS system where it reads and summarizes one

or more SMF datasets for subsequent analysis by one of the CPS Tools Applications (zCP3000,

zBNA, or zPCR). It is intended to be used by IBM Employees and Business Partners.

The program performs these functions:

Parses the complex structure of the SMF data

Extracts selected fields needed by the Tools applications

Combines the data into a series of Extract “reporting periods”

Produces an Enterprise Data File (EDF) that is read by the analysis applications

Optionally writes a separate file for zBNA with selected SMF fields

Generates and/or uses several supplemental mapping files

Here the basic data flow:

When the Extract program is run the default is that all input SMF records will be examined, but you can specify particular dates and times to select only certain data. You can also adjust

the duration of the reporting period.


Program Inputs

The program uses makes use of several different inputs as follows:

SMF data SMF (the System Management Facility) is a z/OS mechanism that records system and job-related information. The data is represented by a series of different record

types each specializing in a different kind of information. The program selects

specific record types based on the requirements of the run (more in part 5).

RMF Data RMF (the Resource Measurement Facility) is a performance monitor for z/OS that

collects operational data for later analysis. RMF data is placed into the SMF record

stream using record types 70 to 79 (i.e., the SMF dataset includes the RMF data).

User

Parameters

User inputs let you select a variety of processing options and identify what part of

the input SMF data you want to examine. Most of the parameters are optional and

many have default values that provide standard processing (see part 3).

BCU Map Typically control units with identical channel configurations are combined for

reporting. This map identifies how (details in part 6.1).

PGN Map The program combines service classes for reporting using the PGN map. Usually the program generates the map automatically and then uses it in processing, but if

needed, you can manually edit the file to adjust the combining (see part 6.2).

T30 Map The T30 map provides a way for you to break out workloads in greater detail than

the WLM service classes (details in part 6.3 ).

IOCP File The IOCP (I/O control program) produces the IOCDS (I/O Control dataset). This

contains details of directors, links, channel paths, units, and subsystems along

with other system information. Here “IOCP File” refers to the systems IOCDS.

Program Outputs

These are the program output datasets:

EDF The Enterprise Data File (EDF) is the primary output. It is a flat text file containing summarized data extracted from the SMF scan. The EDF is the input to other

analysis applications like zBNA, zCP3000, and zPCR. An example is in part 8.2.

Print Output The print listing shows the input parameters specified for the run along with any

messages that were generated by the program.

BCU Map The program will generate a BCU map if you specify the BCU=AUTO parameter,

otherwise it will an existing BCU file (details in Part 6.1).

PGN Map The program will generate a PGN map if you specify the PGN=GOAL parameter,

otherwise it will an existing PGN file (details in part 6.2).

DATA File The program can extract fields from the specific SMF record types and format

the data into a flat text file for use by zBNA. The output file is specified by the

DATA001 DD and can include fields from type 30, 42, 14, and 15 records. Part 7

gives details.


Reporting Interval

There are two different time intervals that must be considered when working with the Extract

program – the RMF interval and the Extract reporting interval. If you are familiar with these

skip this introductory page.

The RMF Interval is the time period (typically 15 minutes) during which z/OS accumulates and summarizes both software and hardware activities. At the end of

this interval the system writes the collected information to the SMF dataset and, for

many items, clears the accumulators to start again. The RMF interval is set by the

system programmers and usually does not change.

The Extract Reporting Interval is the time period the Extract program uses to

group the incoming SMF data. This is specified by the DURATION parameter and

often is set to one hour (which gives more attractive graphs in zCP3000), but setting

the reporting interval to be the same as the RMF interval is also common. Data from

the SMF records that fall within an Extract reporting interval are combined and

reduced to a single value (often an average).

For example, consider one specific measurement item, the “start subchannel count”. Processing for this item will flow through the following steps (over-simplified for discussion):

During the 15 minute RMF interval while the original applications are executing, this counter is incremented by the system for each start subchannel request.

At the end of the RMF interval that count is written by the system to an RMF record.

Later, when the Extract program is run it scans the RMF records.

The “start subchannel count” from each RMF record (type 74) is totaled within the duration of an Extract reporting interval (e.g., records from 4 RMF intervals would be

encountered for a 1 hour reporting interval).

At the end of the Extract reporting interval the total is divided by the number of values giving an average.

That average is appended to the string of values reported as the “start subchannel count” vector (DASDIOV) in the output EDF (resulting in one value in the vector for each reporting period).

Summary -- for many data values the progression is to combine execution-time measurements

into an RMF interval and then further combine these into an Extract reporting interval.

SMF Record Selection

It may be that the input SMF dataset contains information for only one partition for one

particular time period. This is a simple case where you may want to accept and process

everything in the input.

A more complicated situation is when the SMF input includes data for several partitions,

perhaps generated over many days. In this case you can ask the program to select just certain records belonging to a particular partition or a specific time period. I.e., you can extract and

summarize a subset of the original data. (see part 3.2).


2. Quick Start Guide

This section provides a checklist of the basic steps needed to get started running the Extract

program. This is a simplified approach and more details are provided in Part 3 of this

document.

2.1 Preparation

a. Have a TSO user ID and password on the z/OS system where the Extract will be run.

b. Identify the necessary parameters for the JES JOB card for that installation.

c. Get the SMF SysId of the system(s) for which we are Extracting data (see also part 2.5).

d. Decide when and for how long you wish to sample and set the Extract reporting interval

accordingly (with the DURATION parameter). For performance analysis using the 15

minute intervals may be reasonable. However, for doing capacity planning one-hour

intervals are a minimum. Depending upon the amount of data available different durations may be recommended (but nothing less than 5 minutes). If you don’t know

where to start use a week of one hour samples. More details are in part 2.5 and part 3.2.

e. Know the dataset names of the SMF files for the samples you have chosen.

2.2 Research

If possible, sit down with the system programmer and discuss these points:

a. Review the SMFPRMxx member of 'SYS1.PARMLIB'. Verify that you are collecting the

SMF record types you will need.

For zPCR types 70 through 75, 78 and 113 are required.

For zBNA also include 14, 15, 30 (2,3,4,5), 42 (6)

For zCP3000 use the zPCR types and also optionally 42 (6).

b. Instructions for collecting 113 records are at this link: http://www.ibm.com/support/techdocs/atsmastr.nsf/WebIndex/TC000066 Note: zCP3000 requires Basic and Extended counters.

c. Verify that you are doing INTERVAL processing.

d. Ask about the job that dumps off the SMF datasets. Sometimes that job will exclude

certain types and they will not make it to the output dataset. Verify that the ones you will need are present. You can use either the standard SMF dump program, IFASMFDP, or if

the installation uses log streams for the SMF data you can use IFASMFDL to select a

subset of SMF record types.

e. Ask whether the SMF records are in sequence by date and time.

f. If possible, obtain an I/O configuration diagram for the installation.

g. If you will be using the FICON aggregation function of zCP3000, learn the name of the

IOCP dataset and the name of the LPAR partition you will be studying.

http://www.ibm.com/support/techdocs/atsmastr.nsf/WebIndex/TC000066


2.3 Install the Program

a. Download the CP3KEXTR Extract Program to your PC

CP3KEXTR is an assembler language program that is distributed in LOADER format.

The file CP3KEXTR.zip contains both the program and the JCL. Download this from one of the following sources:


or: ftp://cpstools.washington.ibm.com (click on the CP3KEXTR folder)

Business Partners may obtain the program from the PartnerWorld system: https://www-304.ibm.com/partnerworld/wps/servlet/ContentHandler/tech_PRS4144

Customers may obtain from: http://www-03.ibm.com/support/techdocs/atsmastr.nsf/WebIndex/PRS4229

b. On your PC, unpack CP3KEXTR.zip to obtain the CP3KEXTR.bin file. This is the file that will be uploaded to the TSO system.

c. Sign on to TSO

Log in to the TSO system (where you will run the Extract program)

d. In TSO, Allocate the upload dataset

This is the dataset into which you will upload the CP3KEXTR.bin file.

In ISPF/PDF 3.2 choose Allocate. Then,

Name the new dataset CP3KEXTR.UPLOAD.

Set the RECFM to FB (or F) and the LRECL to 80.

Specify a BLKSIZE of 0 for system determined block size. If system determined block size is not implemented zero will give you an error. In this

case use 6400.

Suggest a primary extent of three cylinders with a secondary extent of one and no directory blocks.

e. Upload the Extract program

The CP3KEXTR.bin file in the PC contains binary characters and must be uploaded in

binary mode. Use FTP to transfer the file. Open a command prompt window and switch

to the directory that holds the CP3KEXTR.bin file. Then enter:

FTP wscmvs.washington.ibm.com <-- (substitute your systems IP address) Enter your userid and password

User (wscmvs.washington.ibm.com:(none)): YourUid 331 Send password please.

Password: YourPw ftp: binary 220 Representation type is Image

ftp: put CP3KEXTR.bin CP3KEXTR.upload

ftp: quit


ftp://cpstools.washington.ibm.com/

https://www-304.ibm.com/partnerworld/wps/servlet/ContentHandler/tech_PRS4144

http://www-03.ibm.com/support/techdocs/atsmastr.nsf/WebIndex/PRS4229


f. In TSO, “Receive” the file

The CP3KEXTR.bin file from your PC is now in the TSO system. This step reformats it with the RECEIVE command into the CPSTOOLS.JCL dataset.

In ISPF/PDF 6 type: RECEIVE INDSN(CP3KEXTR.UPLOAD)

The command will prompt you for the dataset name.

Type: DA(CPSTOOLS.JCL) SPACE(2,2) CYLINDERS

This will create the CPSTOOLS.JCL dataset (within your high-level qualifier).

2.4 Customize the JCL

In the CPSTOOLS.JCL dataset there is a member already set up for each of the tools. Edit and

use the appropriate member as described below.

member used for

EXTRALL zPCR

zBNA

zCP3000

Extracts SMF 14, 15 30, 42, 7x, 89, and 113 records into file(s)

that can be loaded into PC for zPCR, zBNA or zCP3000. Includes

all options, although some less used may appear as comments.

EXTRCP3K zCP3000 Extracts SMF 7x, 89, 113 Records into file(s) that can be loaded to PC for zCP3000. This JCL includes all options, although some

less used may appear as comments.

EXTRZPCR zPCR This is simplified JCL for running the Extract to create all of the

data that is used by the zPCR tool.

EXTRZBNA zBNA This is simplified JCL for running the Extract to create all of the

data that is used by zBNA (z Batch network Analyzer).

a. Change the JOB card to meet the standards of your installation.

b. Change the SMFIN DD statement to point to your input SMF dataset.

c. Specify the correct prefix in the dataset name for the EDF001 dataset and (if they are

needed) the PGN001 and BCU001 datasets. Note: If you are using another dataset it

must have a record format of F or FB and an LRECL of 80.

d. Advanced zCP3000 users: If you are using the IOCP option of the BCU=AUTO option specify the correct prefix in the dataset name for the IOCP001 dataset.


2.5 Setup the Input Parameters

The Extract procedure is flexible and can support a number of variations in processing. These

are selected using input parameters you provide via the SYSIN001 DD.

Most of the parameters are optional and default to functional values if not specified. However,

the SysId of the system you want to study must be provided.

If you don’t know the SysId for the SMF data you can specify a dummy SYSID=XXXX (which won’t match anything). When there is no match the program skips its regular

processing and instead prints a summary of the contents of the SMF dataset. This

includes, among other things, a list of the SysId values found in the data. Put the

following parameters (and only these) in the SYSIN001 input stream.

//SYSIN001 DD * ENT='Customer Name' (substitute a descriptive name) SYSID=XXXX (specify a value of ‘XXXX’) /* There is an output example in part 3.7

The following input parameters are the most commonly used. Additional parameters and

variations are shown in part 3.

These 4 parameters are normally provided. The first 2 are always required:

ENT=’xxxxxx’ The Enterprise name, specified as a character string in single quotes (50 characters maximum). Imbedded blanks are permitted. Required.

SYSID=xxxx Specifies the four-character JES SYSID of the system to be studied. Required

BCU=AUTO Specified to generate a BCU Mapping file. Normally you will include this.

If you are interested in FICON Mapping you also want to specify an IOCP001

DD name pointing to the current IOCDS. This will generate a more complete

BCUMAP suitable for FICON analysis. For the FICON analysis you should also use BCU=FICONAG and PART= to specify the LPAR partition name.

PGN=GOAL Used to generate the Workload Mapping. Also normally included


The following parameters can be left out and the default for the run will be to process all SMF

input data with an Extract reporting interval of 1 hour. Or you can specify the range of SMF

data you want to process and adjust the reporting interval.

DURATION Specifies the Extract reporting interval in hours. This parameter defaults to 1 if

not provided. The minimum is value is 5 minutes.

E.g., DURATION=1 for one hour (the default)

DURATION=2 for 2 hours

DURATION=00:15 for 15 minutes

TIME Specifies the times (the hours of the day) you want to study. If not specified, all

times will be used. Be sure that the duration fits evenly into this time range.

E.g., TIME=(00-24) to select all hours of the day (the default)

TIME=(08-11) to select data from the 8th, 9th, and 10th hours.

You may have multiple TIME= statements. For example, if you are using 2-hour

samples you may have TIME=(08-12) and another TIME=(13-17).

DATE Specifies the days of interest. If it is not specified all dates in the input SMF file will be used. There are several variations of the DATE statement – it can

include a single date, a range of dates, for a list of separate dates.

E.g., DATE=(09/25/15)

DATE=(09/25/15-09/29/15)

DATE=(09/25/15,09/26/15,09/27/15,09/28/15,09/29/15)

The RMF interval must fit evenly within the Extract reporting DURATION and the

Extract reporting DURATION must fit evenly within the specified TIME period.

See part 4 for an example of the JCL and the input statements.


2.6 Run the Extract

a. Submit the Job

Use the JCL member you just customized and submit the job. This will generate the output

EDF file.

The following return codes will be issued:

0 No messages.

4 Messages of note have been issued.

8 Either important data (RMF types) were missing from the file or there is

questionable data in some of the intervals (that you may want to exclude).

12 Something is critically wrong with the parameters, or no data fit within the

SYSID and timeframe requested.

16 Something is drastically wrong and the program failed.

20 CP3KEXTR couldn't write to the PRINT001 DD.

225 No data was found that meets the input specification. This could happen if the specified SysId was not present in the SMF input dataset or if the date

and time selections did not cover the time periods of the data.

Note: If no data was selected it might be because:

The specified SysId wasn’t present.

No records were in the specified date/time range.

The data is not sorted (add the sort work datasets to the JCL)

If the data is not sorted you have to specify SORT=YES to get it to sort. You may or may not need sort work datasets based on the size of the input SMF file.

b. Check for Messages

Review the messages produced by the program -- they go into the PRINT001 DD statement and

will be with your output JCL.

c. Download the EDF (and optionally the DAT file)

The EDF file will be in the member referenced by the EDF001 DD statement. For zBNA the

DAT file will be in the dataset referenced by the DATA001 DD statement.

These files should be downloaded using normal ASCII CRLF protocol. The suggested extension

is EDF, so zCP3000 can find it.


3. Program Input Parameters

This section describes the input parameters accepted by the Extract program (CP3KEXTR)

through the SYSIN001 dataset. Note that if you are processing more than one LPAR in this

Extract run there will be a SYSIN DD statement for each (e.g., SYSIN001, SYSIN002, etc.).

The ENT and SYSID parameters are required for all jobs while the others are used only if needed. The following conventions apply to coding parameters:

One parameter per line

Parameter text must begin in column 1

The text may run up to column 72

Comment lines are indicated by an asterisk ‘*’ in column 1

You can review a sample of this type of input in part 4.3.

Note that any of the Extract input streams may have comment lines. This includes:

SYSIN, PGNMAP, BCUMAP or T30MAP. A comment is identified by an asterisk * in

the first column (causing the remainder of the line to be ignored).

3.1 Required Parameters

There are two required parameters that must always be specified in each SYSIN dataset.

ENT=’xxx xxx xxx’

The Enterprise name is specified as a character string in single quotes (up to 50 characters).

Imbedded blanks are permitted. Multiple EDF files from different partitions and CECs for the

same enterprise should have the same ENT parameter.

SYSID=xxxx

Specify the four-character JES SYSID as it appears in the SMF records. This is used to select only one particular SYSID from the SMF input.

Remember, if you don’t know the SysId you can do a run with just the following two

parameters and the program will print a summary of the contents of the SMF dataset showing

the SysIds that are present.

//SYSIN001 DD * ENT='Customer Name' (substitute a descriptive name) SYSID=XXXX (specify a value of XXXX) /*


3.2 Reporting Interval Parameters

These parameters let you select or exclude certain dates and times. You can also specify the

duration of the Extract reporting period.

DATE=(MM/DD/YY)

The specified date or date range will be used to select RMF records. Multiple DATE parameters

can appear separately. If no DATE parameter is specified all dates in the input file will be used.

The parameter can be a single date, a list of dates, or a date range as follows:

DATE=(MM/DD/YY)

DATE=(MM/DD/YY-MM/DD/YY)

DATE=(MM/DD/YY,MM/DD/YY,MM/DD/YY)

DURATION=hh

The DURATION value specifies the Extract reporting interval -- the time period in hours the

Extract program uses to group the incoming SMF data. One or more RMF intervals will be

summarized into each DURATION reporting interval.

If the DURATION is not specified the program uses a default of one hour, DURATION=1.

This parameter can be specified in whole hours or in hours and minutes:

DURATION=hh

DURATION=hh:mm

Remember, the RMF interval must fit evenly within the Extract reporting DURATION

and the Extract reporting DURATION must fit evenly in the specified TIME period

(next parameter).

For example, if your RMF has a 15-minute interval set your DURATION to 15 minutes or some

multiple of 15 minutes. Specifying DURATION=01 (one hour) would evenly include 4 RMF

intervals.


TIME=(hh-hh)

This parameter gives a range of hours. SMF data within this range will be accepted as input

and data outside of this range will be ignored. The selection begins with the first value and

runs up to the second value. Note that it is important that the DURATION of the Extract

reporting interval fits evenly in the specified TIME period.

For example, if you specify TIME=(09-11) SMF data from the 9th and 10th hour will be included

and would evenly contain a 1 hour reporting DURATION.

Multiple TIME parameters can be specified. The duration value must be specified first.

The time parameter can be given in whole hours or in hours and minutes:

TIME=(hh-hh)

TIME=(hh:mm-hh:mm)

The decision whether to include an RMF record is based on a time that represents the midpoint

of the RMF interval. For example, if TIME=(09-11) is specified,

An RMF record for an interval beginning at 10:55 with a duration of 15 minutes would not be included (because the midpoint time is after the 11:00 cutoff).

A record beginning at 8:55 also with a duration of 15 minutes would be included (because its midpoint time is after 9:00).

With either form be sure that the DURATION fits evenly in this time window.

If you are using the SELECT statement you don’t need to also use the TIME parameter.

If neither TIME nor SELECT is specified the program defaults to TIME=(00-24), i.e., select all

times.

SELECT DATE=MM/DD/YY,TIME=HH-HH

EXCLUDE DATE=MM/DD/YY,TIME=HH-HH

These statements may be used to include or to exclude certain periods from analysis.

For example, if you are studying a customer's prime shift during a week but they are open late

on Thursdays, you can add additional hours to the study for Thursday. In addition, if on

Tuesday morning there was a hardware failure that made the data from 9-10 invalid you can

EXCLUDE that data.

It is also possible to simply look at discrete times. For example, if the study of a week's data reveals that there are three peak hours it is possible to SELECT only those three hours for

study.


3.3 Other Parameters

CECID/CPCID=cccccccc

Up to eight characters for the CECID/CPCID. If omitted the SYSID will be used for machines

in basic mode. For machines in LPAR Mode CPCnnnnn will contain the 5-digit correct serial

number.

NOTE: It is best to let this default and rename the CEC once the data has been loaded into zCP3000.

CPUMOD=cccccccccc

A 10-character CPU model. If this is not specified a table lookup will be used with data in the

RMF type 70 record.

PROCESS=YES/NO

This specifies whether you want the program to perform its full SMF processing.

NO The program will accept the input parameters, generate the BCU and/or

PGN maps, then stop without further processing. This is used in the

case where you want to edit a BCU and/or PGN map before using it.

YES The program will proceed through all the normal processing of the SMF

data. This is the default.

Note: If you have done manual editing of the BCU or PGN maps be sure

to remove the BCU=AUTO and/or PGN=GOAL statements.

RMFINT=nn

Number of RMF intervals to be used when calculating the peak to average ratio. Any consecutive nn intervals will be considered. Default is one.

SORT=YES/NO

Coding YES will cause the program to do an internal sort on only the SMF data required for

this study. In most cases the SMF data is already in time sequence and this sort is not

required. The default is NO.

If you should see interval starts or durations which are strange or if CP3KEXTR should give

you a message to sort the data then a sort is in order.

When using the internal sort the REGION= parameter should be adjusted as high as

reasonable. Also, some thought should be given to the size of the sort work datasets.


SAVE=SORT=ALL/SysId

This is used to take an input (presumably tape) dataset and copy it to disk while sorting the

data and eliminating duplicate records. The destination is described by the SMFSAVE DD

statement.

ALL All records (including every SysId) are written to disk

SysId Only records for the specified SysId are written to disk

A report of the SysIds encountered, record types and subtypes encountered, and earliest and

latest date and times will be written to the SYSPRINT DD statement.

Note that no other processing is permitted with this parameter. If this parameter is specified no other parameters are allowed except TYPES.

SUB30=n,n,n...

This is used with SMF records 30 processing to determine which type 30 subtypes are included

in the DATA001 file. The default is SUB30=4,5. This parameter does not apply to the T30MAP

processing which always uses type 2 and 3.

TYPES=70(1,2),71(1),72(1,3),73(1),74(1,4,5,8),75(1),78(1,3) TYPES=30(2,3,4,5),42(6),113

This parameter defines the records that are to be written to the disk dataset. It can only be

specified with SAVE=SORT=. If this statement is omitted the default includes the types and

subtypes shown above.

If you are not interested in using a particular record type you would include the TYPES statement as shown above but leave out that one type.


3.4 BCU - Parameters

BCU (basic control unit) is a term used in the Extract program to represent a physical I/O

control unit -- the “box” that contains the LCUs (logical control units). The BCU map provides

a mechanism to combine LCUs that are part of the same physical control unit.

BCU=AUTO/AUTOADD

The BCU=AUTO or AUTOADD option causes the program to create a new BCU Map and write it to the BCU001 dataset. This will combine BCUs that have the same channels. If not specified,

an existing BCU map will be used. BCU=AUTO is the typical processing case, but since this is

not the default it must be specified.

If IOCP data (IOCDS) is used it will combine BCUs with the same switches and links.

You must have type 74 subtype 1 and type 78 subtype 3 records. If you specify BCU=AUTO

and these types aren’t present the run will stop with an error message.

The mapping process is described in part 6.1. (Note that the map will be more complete if 74

subtype 5 records are available). If the IOCPnnn DD statement is found it will also be read to

create a more complete BCUMAP.

Note that for shared DASD the BCUID may not be unique. This should be adjusted to be the

same value on all systems.

Specification BCUID in EDF

BCU=AUTO The BCU manufacturer (IBM) plus a hyphen (-) plus the 5-digit device serial number.

If no VOLSER is found the name will default to "L" plus the LCU number.

BCU=AUTOADD "A" concatenated with the 4-digit hex address of the lowest

numbered device on this BCU.


BCU=NO/DASD/TAPE/OTHER/ALL

These choices are used to request processing of an existing BCU map that will be read in from

the BCU001 DD (i.e., you did not specify BCU=AUTO). This maps the volume data to BCUs by

SYSID and address. Default is ALL.

You can specify the type of the BCU entries you want to process as follows:

NO Skip BCU processing

DASD Only process the BCU macros in the BCU map

TAPE Only process the BCUT macros in the BCU map

OTHER Only process the BCUO macros in the BCU map

ALL Process BCUD, BCUT, and BCUO

DASD,TAPE Process both DASD and TAPE. (Or any combination of

DASD, TAPE, or OTHER separated by commas)

BCU=LCU/COM

You can specify BCU=COM as an additional option for BCU=AUTO to choose an alternate

combining strategy. Put this specification on a separate line either before or after the

BCU=AUTO statement. The processing distinction is:

Specification Method for Combining

BCU=LCU (default) Combine different LCUs that the program can determine are part of

the same physical control unit.

BCU=COM Combine different control units that share the same set of paths (i.e., the same CHPIDS and links).

ESCON control units that share the same directors are combined.

Control units on copper channels where the control units are daisy chained are combined.

If you will be using zCP3000 to analyze shared DASD, it will expect that a given BCU will have

the same BCUID from each system. This will more likely happen automatically with

BCU=AUTO than with BCU=AUTOADD.


BCU Summary

The BCU map can be generated automatically and used in the same run (case 1 below) or it

can be generated, manually edited, and then used in subsequent processing (2 & 3 below). These are the parameters needed to handle these cases:

Case Specify these parameters

1 Fully Automatic

Generate the map and then use it

BCU=AUTO to generate the BCU map

PROCESS=YES generate the map and then

do the normal SMF processing

BCU=ALL use everything in the new map

2 Generate a map

to manually edit

BCU=AUTO to generate the BCU map

PROCESS=NO to stop after making the map

3 Use existing map

after editing

omit BCU=AUTO don’t make a new map

(it would write over the edited one)

PROCESS=YES do the normal SMF processing

and use the edited BCU map

BCU=ALL use everything in the map

BCU=FICONAG

There may be cases in FICON studies where it is desired that parts of physical boxes not be combined. These cases are when different groups of LCUs connect to different groups of

CHPIDs. If you specify BCU=FICONAG these LCU groups will be kept separate in the

BCUMAP.

PART=nnnnnnnn

This parameter specifies the partition name which is used to process the IOCP file in an LPAR

environment. It is only used with BCU=AUTO when an IOCPnnn DD statement is provided.

The default is the SYSID specification.

SHOWACT=ALL/NONE

Specifying NONE will suppress all actuator data. zCP3000 does not yet support files produced by using a value here but SHOWACT=.001 will cause only those actuators which have I/O

rates of .001 or greater to be output. NOT YET TO BE USED WITH zCP3000. <>


3.5 PGN - Workload Parameters

zCP3000 users frequently want to combine service classes to a smaller number for executive

reporting. The Extract program provides a facility to do this. It can also, probably more

simply, be done once in zCP3000.

The PGN Mapping mechanism provides a way to combine service classes for reporting. The program does this automatically and you can make manual adjustments if needed. However,

the zCP3000 application handles this so manual adjustments are usually not necessary.

The PGN001 DD statement must be specified with the JCL to process workload data. If it is

not present, there will be no PGN processing regardless of the input parameters. See also

part 6.2

PGN=GOAL

This option asks the program to automatically build a PGN map into the PGN001 DD file. It will scan the SMF data, read the SMF Type 72 subtype 3 records, do the combining, and write

the map. If record type 72 subtype 3 data is not present the program will issue an error

message and terminate the run.

Normally you include this parameter. But when you leave it out the program uses an existing

PGN map, giving you the opportunity to make manual adjustments. See also part 6.2c.

Case 1 below is where the PGN map is generated automatically and then used in the same run. Case 2 & 3 below is where the map is generated, manually edited, and then used in later

processing. These are the parameters needed to handle these cases:

Case Specify these parameters

1 Fully Automatic

Generate the map

and then use it

PGN=GOAL generate the PGN map

PROCESS=YES do the normal SMF processing and use the new PGN map

2 Generate a map

to manually edit

PGN=GOAL generate the PGN map

PROCESS=NO stop after making the map

3 Use existing map

after editing

omit PGN=GOAL don’t make a new map

(it would write over the edited one)

PROCESS=YES do the normal SMF processing

and use the new PGN map


TYPE42=YES/NO

This option will build records for each of the workloads describing the dataset usage by

business unit. This data will be built from type 42 subtype 6 records. The default is YES.

There are some assumptions for the type 42 records processing.

A PGNMAP was supplied.

BCU=DASD or ALL was specified.

SHOWACT=NONE was not specified

3.6 Convert Tape to Disk

The Extract Program has a facility to copy an input dataset (usually from tape), sort the records

removing duplicates, and write them to an output dataset (usually on disk). The destination is described by the SMFSAVE DD statement.

As part of this processing a report will be written to the PRINT001 DD statement which lists:

All SYSIDs encountered

For each SYSID, the record types and subtypes encountered

For each record type and subtype the earliest and latest time and date

No other processing can take place in a tape to disk run.

The input dataset must be allocated to the SMFIN DD statement. The output (disk) dataset will

be written to the SMFSAVE DD statement. Give the job as much region and SORTWKnn’s as

you can.

SAVE=SORT=ALL

SAVE=SORT=SysId

No other statements except for TYPES described below are permitted with the SAVE=SORT

option.

=ALL All records (including every SysId) are written to disk. Use this if you have data with multiple SYSIDs and you would like to keep the data

for distinct SYSIDs separate.

=SysId Only records for the specified SYSID are written to disk

For example, if you received a tape with data for SYSA, SYSB and SYSC, you would need to

make three separate passes:

one with SAVE=SORT=SYSA producing the first SMFSAVE dataset, then

another pass with SAVE=SORT=SYSB producing the 2nd different SMFSAVE dataset

and a final pass with SAVE=SORT=SYSC


By default the program will collect and sort the following record types:

TYPES=14,15,30(2,3,4,5),42(6),70(1,2),71(1),72(1,3),73(1),74(1,4,5),75(1),78(1,3),113

You may change this list adding or subtracting record types. This is the list of all records that

the Extract Program can process so if you are not interested in using one particular record type

you would include the TYPES statement shown above but leave out that one type.

For example, if you are pressed for space and know that you will not be processing the type 30

records you could place this statement in the input stream and remove the “30(2,3,4,5)” part.

Then the output dataset would not contain type 30 records. This is the only other statement

permitted with the SAVE=SORT= parameter.


3.7 Summary Output

If you use a SysId that doesn’t match anything in the incoming SMF data the program skips its

regular processing and instead prints a summary of the contents of the SMF dataset. (Note

that you can force this by deliberately specifying SYSID=xxxx). Here is an example of that

output:

ENT='Customer Name' SYSID=XXXX <----------- SYSIN001 DATASET PROCESSED CP2KEXT: NO BCUMAP PROCESSING REQUESTED CP2KPR10: NO PGNMAP PROCESSING REQUESTED CP2KT30M: NO T30M001 PROCESSING REQUESTED CP3KEXTR: NO DATA FOUND TO PROCESS CP2KUTIL: CEC 2097-742 S/N (LAST 4 DIGITS): 9F30 LPAR AQFT SYSID AQFT GMT OFFSET -4:00 LPAR VICTEST LPAR VMTOOL1 LPAR AQCF1 LPAR AQHO LPAR AQLINX LPAR HOCF4 LPAR GDLVM7 LPAR LNXVM14 LPAR POKVMXA1 LPAR PHYSICAL CP2KUTIL SYSID: AQFT MVS: ZV010900 RMF: V7R1M9 CP2KUTIL RMF DURATION: 30 MINUTES CP2KUTIL: GOAL MODE DATA. POLICY PRIMSHFT DESC PLPSC Production policy TYPE 002 1 RECORDS FROM: 06/23/08 09:10 - 06/23/08 09:10 TYPE 003 1 RECORDS FROM: 06/23/08 09:26 - 06/23/08 09:26 TYPE 023 10 RECORDS FROM: 06/20/08 09:17 - 06/20/08 13:47 TYPE 070 01 10 RECORDS FROM: 06/20/08 09:00 - 06/20/08 13:30 TYPE 070 02 10 RECORDS FROM: 06/20/08 09:00 - 06/20/08 13:30 TYPE 071 01 10 RECORDS FROM: 06/20/08 09:00 - 06/20/08 13:30 TYPE 072 03 360 RECORDS FROM: 06/20/08 09:00 - 06/20/08 13:30 TYPE 072 04 10 RECORDS FROM: 06/20/08 09:00 - 06/20/08 13:30 TYPE 073 01 10 RECORDS FROM: 06/20/08 09:00 - 06/20/08 13:30 TYPE 074 01 370 RECORDS FROM: 06/20/08 09:00 - 06/20/08 13:30 TYPE 074 02 10 RECORDS FROM: 06/20/08 09:00 - 06/20/08 13:30 TYPE 074 03 10 RECORDS FROM: 06/20/08 09:00 - 06/20/08 13:30 TYPE 074 04 20 RECORDS FROM: 06/20/08 09:00 - 06/20/08 13:30 TYPE 074 05 1390 RECORDS FROM: 06/20/08 09:00 - 06/20/08 13:30 TYPE 074 06 10 RECORDS FROM: 06/20/08 09:00 - 06/20/08 13:30 TYPE 075 01 70 RECORDS FROM: 06/20/08 09:00 - 06/20/08 13:30 TYPE 076 01 120 RECORDS FROM: 06/20/08 09:00 - 06/20/08 13:30 TYPE 077 01 10 RECORDS FROM: 06/20/08 09:00 - 06/20/08 13:30 TYPE 078 02 10 RECORDS FROM: 06/20/08 09:00 - 06/20/08 13:30 TYPE 078 03 20 RECORDS FROM: 06/20/08 09:00 - 06/20/08 13:30


4. JCL

Modify the sample JCL in this document to meet the requirements of you installation.

4.1 DD Statement Summary

DD Name Purpose Required

SYSIN001 Input parameters to the Extract program. Always

SMFIN The input SMF dataset Always

EDF001 Is the EDF output file. Logical record length is 80 Always

PRINT001 CP3KEXTR messages Always

BCU001 The BCU mapping. LRECL is 80.

PGN001 The service class mapping file. LRECL is 80.

DATA001 Dataset for extracts of type 30, 42, 14, & 15 records for zBNA

T30M001 The workload decomposition group mapping to be used

with type 30 record processing. LRECL is 80.

IOCP001 The IOCP dataset, also known as the IOCDS (I/O Control

Dataset). LRECL is 80.

SYSLOUT LOADER messages

SORTMSGS Sort Messages if SORT=YES

SORTWK0x Sort Work Space if SORT=YES

SMFSAVE SMF Records to be saved for analysis in future runs if SAVE= is specified


4.2 Return Codes

These are the Extract Program return codes. Non-zero return codes can be found in the

SYSOUT files along with the message generating the non-zero return code.

0 No notifications


8 A warning that messages have been issued that you should review. Either important

SMF/RMF record types were missing from the input file or there is questionable data

in some of the intervals (that you may want to exclude).

12 Something is critically wrong with the parameters or no data fit within the SYSID and

timeframe requested.

16 Something is drastically wrong and this run has failed.


225 No data was found that meets the input specification.


4.3 JCL for zCP3000

The JCL is dependent upon the options selected. The following sample is the EXTRACT

member that is shipped in the package. It builds the PGNMAP and BCUMAP files. Follow the

directions in part 2 to create the EDFI and Disk Magic files given below.

//#PATTRSN JOB (????,????),MSGLEVEL=1,MSGCLASS=O,NOTIFY=???????? //* //* DATA EXTRACTION PROGRAM FOR ZCP3000 //* //EXTR EXEC PGM=LOADER //PRINT001 DD SYSOUT=* //SMFIN DD DISP=SHR,DSN=PATTRSN.RMF.CTS51T //EDF001 DD DISP=SHR,DSN=PATTRSN.CPSTOOLS.JCL(EDFI) //PGN001 DD DISP=SHR,DSN=PATTRSN.CPSTOOLS.JCL(PGNMAP) //BCU001 DD DISP=SHR,DSN=PATTRSN.CPSTOOLS.JCL(BCUMAP) //*IPS001 DD DSN=SYS1.PARMLIB(IEAIPS00),DISP=SHR //*IOCP001 DD DISP=SHR,DSN=PATTRSN.CPSTOOLS.JCL(IOCP) //*MAGIC001 DD DISP=SHR,DSN=PATTRSN.CPSTOOLS.JCL(MAGIC) //*SORTMSGS DD SYSOUT=* //* //SYSOUT DD SYSOUT=* //SORTWK01 DD UNIT=SYSDA,SPACE=(CYL,(200)) //SORTWK02 DD UNIT=SYSDA,SPACE=(CYL,(200)) //SORTWK03 DD UNIT=SYSDA,SPACE=(CYL,(200)) //SORTWK04 DD UNIT=SYSDA,SPACE=(CYL,(200)) //* //* # OUTPUT DATASET FOR TYPE30 RECORDS //* NOTE: DCB INFORMATION FOR THESE DATASETS IS SUBJECT TO CHANGE. //* SO IT IS BEST TO LET THE PROGRAM ALLOCATE THEM. //*DATA001 DD SPACE=(CYL,(20,10),RLSE),DISP=(,CATLG),UNIT=SYSDA, //* DSN=PATTRSN.TYPE30.DATA //* # OUTPUT DATASET FOR SAVE= PARAMETER (SAVE CP3K'S SMF RECORDS) //*SMFSAVE DD SPACE=(CYL,(20,10),RLSE),DISP=(,CATLG),UNIT=SYSDA, //* BLKSIZE=8192,DSN=PATTRSN.SMF.SAVE //* # OPTIONAL USER DUMP DATASET //*SYSUDUMP DD SYSOUT=* //SYSIN001 DD * ENT='YOUR COMPANY' YOUR COMPANY NAME SYSID=AQFT SMF SYSID DURATION=01:00 DURATION FOR SUMMARY SAMPLES * DURATION MUST BE A MULTIPLE OF RMF INTERVAL *TIME=(00:00-24:00) CONTINGUOUS PERIOD OF INTEREST *DATE=(08/05/09-08/12/09) CONTINGUOUS DAYS PGN=GOAL GENERATE PGNMAP BCU=AUTO GENERATE BCUMAP *SHOWACT=NONE SHOWACT=NONE REDUCES OUTPUT SIZE. SORT=YES IF RECORDS ARE IN SORTED ORDER * OMITTING SORT=YES WILL SPEED THINGS UP. /* //* MESSAGES FROM THE MVS LOADER //SYSLOUT DD SYSOUT=* //* THE OBJECT MODULE OF THE PROGRAM //SYSLIN DD DISP=SHR,DSN=PATTRSN.CPSTOOLS.JCL(ZOBJEXTR)


4.4 JCL for zPCR

The JCL is dependent upon the options selected. The following sample is the EXTRSMLL

member that is shipped in the package. It builds the PGNMAP and BCUMAP files. Follow the

directions in part 2 to create the EDFI files given below.

//EXTRSML1 JOB (????,????),MSGLEVEL=1,MSGCLASS=O,NOTIFY=???????, // REGION=64M //* //* LICENSED MATERIALS - PROPERTY OF IBM //* DATA EXTRACTION PROGRAM FOR THE ZCP3000 TOOL //* COPYRIGHT IBM CORP. 1986, 2010 ALL RIGHTS RESERVED. //* US GOVERNMENT USERS RESTRICTED RIGHTS - USE, //* DUPLICATION OR DISCLOSURE RESTRICTED BY GSA ADP //* SCHEDULE CONTRACT WITH IBM CORP. //* //* //* THIS IS JCL TO RUN THE IBM EXTRACT PROGRAM FOR ZCP3000 //* TO PRODUCE AN ABBREVIATED OUTPUT SUITABLE FOR THE ZPCR //* TOOL. //* SEE THE USER'S GUIDE FOR A DESCRIPTION OF //* PARAMETERS AND DD STATEMENTS //* //* //* IBM EXTERNAL CUSTOMERS FIND THE USER'S GUIDE AT //*HTTP://WWW-03.IBM.COM/SUPPORT/TECHDOCS/ATSMASTR.NSF/WEBINDEX/PRS1381 //* //EXTR EXEC PGM=LOADER //* PRINT DATA SET FOR MESSAGES FROM YOUR RUN //PRINT001 DD SYSOUT=* //* INPUT SMF DATASET - REQUIRED 70, 113 OPTIONAL 74 //SMFIN DD DISP=SHR,DSN=CUSTMER.SMF.DATASET //* //* DATASET FOR OUTPUT FILE - REQUIRED //EDF001 DD DISP=SHR,DSN=CUSTMER.CPSTOOLS.JCL(EDFISM) //* WORK DATASET FOR AUTO BCU MAP (DASD I/O RATES) //* COMMENT THIS DD IF YOU DO NOT HAVE SMF 74S. //BCU001 DD DISP=SHR,DSN=CUSTMER.CPSTOOLS.JCL(BCUMAPSM) //* DDS FOR OPTIONAL INTERNAL SORT OF SMF RECORDS //* //*SORTMSGS DD SYSOUT=* //SYSOUT DD SYSOUT=* //SORTWK01 DD UNIT=SYSDA,SPACE=(CYL,(200)) //SORTWK02 DD UNIT=SYSDA,SPACE=(CYL,(200)) //SORTWK03 DD UNIT=SYSDA,SPACE=(CYL,(200)) //SORTWK04 DD UNIT=SYSDA,SPACE=(CYL,(200)) //* # OPTIONAL USER DUMP DATASET //*SYSUDUMP DD SYSOUT=*


//SYSIN001 DD * ENT='CUSTOMER DATA' YOUR COMPANY NAME (REQUIRED) SYSID=SYSD SMF SYSID (REQUIRED) * DURATION=01:00 DURATION FOR SUMMARY SAMPLES. DURATION * MUST BE A MULTIPLE OF THE RMF INTERVAL * BCU=AUTO PROCESS THE 74 RECORDS. * IF YOU DO NOT HAVE 74 RECORDS * REMOVE THIS STATEMENT. * SORT=YES IF SMF RECORDS ARE ALREADY IN SORTED ORDER * OMITTING SORT=YES WILL SPEED THINGS UP. * *TIME=(00:00-24:00) DEFAULT IS TO COLLECT FOR ALL DATA IN INPUT *DATE=(07/16/10) FILE. USE THESE PARAMETERS TO REDUCE THE * PERIOD OF STUDY. /* //* MESSAGES FROM THE MVS LOADER //SYSLOUT DD SYSOUT=* //* THE OBJECT MODULE OF THE PROGRAM //SYSLIN DD DISP=SHR,DSN=PATTRSN.CPSTOOLS.JCL(ZOBJEXTR)


4.5 JCL for zBNA

The JCL is dependent upon the options selected. The following sample is the EXTRZBNA

member shipped in the package. It builds the PGNMAP. The highlighted fields must be

modified as needed for your installation. Follow the directions in part 2 to create the EDF and

DATA files given below.

//VSPENCE1 JOB (????,????),MSGLEVEL=1,MSGCLASS=O,NOTIFY=???????, // REGION=64M //* //* LICENSED MATERIALS - PROPERTY OF IBM //* DATA EXTRACTION PROGRAM FOR THE ZCP3000 TOOL //* COPYRIGHT IBM CORP. 1986, 2016 ALL RIGHTS RESERVED. //* US GOVERNMENT USERS RESTRICTED RIGHTS - USE, //* DUPLICATION OR DISCLOSURE RESTRICTED BY GSA ADP //* SCHEDULE CONTRACT WITH IBM CORP. //* //* NOTE: THIS JCL IS FOR RELEASE 2 AND LATER OF ZBNA. //* IF YOU ARE EXTRACTING FOR THE OLDER RELEASE 1 //* (NOT RECOMMENDED), PLEASE USE THE MEMBER EXTRZBR1 //* //* THIS IS JCL TO RUN THE IBM EXTRACT PROGRAM FOR ZCP3000 //* TO PRODUCE AN ABBREVIATED OUTPUT SUITABLE FOR THE ZBNA TOOL. //* SEE THE USER'S GUIDE FOR A DESCRIPTION OF //* PARAMETERS AND DD STATEMENTS //* //* IBM EXTERNAL CUSTOMERS FIND THE USER'S GUIDE AT //* HTTP://WWW-03.IBM.COM/SUPPORT/TECHDOCS/ATSMASTR.NSF/WEBINDEX/PRS1381 //* // EXEC PGM=IEFBR14 //DD1 DD DSN=CUSTMER.CPSTOOLS.SMFDAT,DISP=(MOD,DELETE), // SPACE=(TRK,(1)),UNIT=SYSDA //DD2 DD DSN=CUSTMER.CPSTOOLS.RMFDAT,DISP=(MOD,DELETE), // SPACE=(TRK,(1)),UNIT=SYSDA //* //EXTR EXEC PGM=LOADER //* PRINT DATA SET FOR MESSAGES FROM YOUR RUN //PRINT001 DD SYSOUT=* //* //* INPUT SMF DATASET - REQUIRED TYPE 70, 72, 113 //SMFIN DD DISP=SHR,DSN=CUSTMER.SMF.DATASET //* DATASET FOR OUTPUT FILE – REQUIRED //* //EDF001 DD DSN=CUSTMER.CPSTOOLS.RMFDAT,DISP=(,CATLG), // UNIT=SYSDA,SPACE=(CYL,(5,1),RLSE),LRECL=80,RECFM=FB //* TYPE 30 EXTRACT DATA //* //DATA001 DD DSN=CUSTMER.CPSTOOLS.SMFDAT,DISP=(,CATLG), // UNIT=SYSDA,SPACE=(CYL,(50,50),RLSE)

Output EDF

goes to zBNA

Output DATA file goes to zBNA


//* PGNMAP DATA //* //PGN001 DD DSN=CUSTMER.CPSTOOLS.JCL(PGNMAP),DISP=SHR //* DDS FOR OPTIONAL INTERNAL SORT OF SMF RECORDS //*SORTMSGS DD SYSOUT=* //SYSOUT DD SYSOUT=* //SORTWK01 DD UNIT=SYSDA,SPACE=(CYL,(200)) //SORTWK02 DD UNIT=SYSDA,SPACE=(CYL,(200)) //SORTWK03 DD UNIT=SYSDA,SPACE=(CYL,(200)) //SORTWK04 DD UNIT=SYSDA,SPACE=(CYL,(200)) //* # OPTIONAL USER DUMP DATASET //*SYSUDUMP DD SYSOUT=* //SYSIN001 DD * * ENT='CUSTOMER DATA' YOUR COMPANY NAME (REQUIRED) * SYSID=SYSD SMF SYSID (REQUIRED) * DURATION=01:00 HH:MM DURATION FOR SUMMARY SAMPLES. DURATION * MUST BE A MULTIPLE OF THE RMF INTERVAL * PGN=GOAL CREATE A PGNMAP AND GATHER WORKLOAD DATA SORT=YES IF SMF RECORDS ARE ALREADY IN SORTED ORDER * OMITTING SORT=YES WILL SPEED THINGS UP. * *TIME=(00:00-24:00) DEFAULT IS TO COLLECT FOR ALL DATA IN INPUT *DATE=(07/16/15) FILE. USE THESE PARAMETERS TO REDUCE THE * PERIOD OF STUDY. /* //* MESSAGES FROM THE MVS LOADER //SYSLOUT DD SYSOUT=* //* THE OBJECT MODULE OF THE PROGRAM //SYSLIN DD DISP=SHR,DSN=CUSTMER.CPSTOOLS.JCL(ZOBJEXTR)

Creates this DS

and writes I/O data

Be sure this SYSID is in

the data set defined in

//SMFIN above

Points to the DS

containing the ZOBJEXTR member


4.6 JCL for SORT

This JCL is required to run RMF Post Processor. The MODS= statement is required only for

the post processor.

//WICKSS JOB (????,????),MSGLEVEL=1,MSGCLASS=O,NOTIFY=WICKS, // REGION=4000K //SORT EXEC PGM=ICEMAN //SYSOUT DD SYSOUT=* //SYSPRINT DD SYSOUT=* //SORTMSGS DD SYSOUT=* //* //SYSLMOD DD UNIT=SYSDA,SPACE=(3600,(20,20,1)) //SYSLIN DD UNIT=SYSDA,SPACE=(80,(10,10)) //SORTLIB DD DSNAME=SYS1.SORTLIB,DISP=SHR //SYSUT1 DD UNIT=(SYSDA,SEP=(SORTLIB,SYSLMOD,SYSLIN)), X // SPACE=(1000,(60,20)) //* //SORTIN DD DISP=SHR,DSN=SYS1.MANDATA //SORTOUT DD DISP=OLD,DSN=WICKS.RMF.CTS51 //*ORTOUT DD DSN=&&SORTED,DISP=(NEW,PASS), //* SPACE=(CYL,(100,50)),UNIT=SYSDA //SORTWK01 DD DSN=&&SORT1,SPACE=(CYL,(125,125)),UNIT=SYSDA //SORTWK02 DD DSN=&&SORT1,SPACE=(CYL,(125,125)),UNIT=SYSDA //* SORT BY DATE, TIME, RECTYPE //* //SYSIN DD * SORT FIELDS=(11,4,CH,A, 7,4,BI,A, 6,1,BI,A), EQUALS, FILSZ=E400000 MODS E15=(ERBPPE15,500),E35=(ERBPPE35,500) /*


4.7 JCL to Create Multiple Output Files in One Run

The program has the ability to create multiple output EDF files from a single pass of the SMF

input data. For example, you could create an EDF for prime shift, second shift, and all shifts

with a single pass of the data, creating three EDF files with different time parameters in a

single run.

To do this you will use two different groups of DD statements:

DD statements used for all “subruns”

DD statements specific to a particular subrun.

An example of the JCL to accomplish this task is on the next page.

4.7a DDs used for all “subruns”

These are specified once and apply to all subruns.

SORTMSGS Sort Messages. Required for SORT=YES

SORTWK0x Sort Work Space. May be required for SORT=YES

SMFIN This is the input SMF dataset. All subruns will process this same data.

SMFSAVE SMF Records to be saved for analysis in future runs.

Required if SAVE= is specified

SYSLOUT LOADER messages

4.7b DDs Specific to a Particular Subrun

These DD statements are associated with each separate subrun. Each DD name ends in a

three-digit number the program uses to correlate them. I.e., all the DD statements related to subrun 1 end in “001”, the ones for subrun 2 end in “002”, etc.

SYSIN001 Input parameters to the Extract program for a subrun

EDF001 Is the EDF output file. Logical record length is 80

PRINT001 CP3KEXTR messages

BCU001 The BCU mapping. LRECL is 80.

PGN001 The service class mapping file. LRECL is 80.

DATA001 Dataset for extracts of type 30, 42, 14, & 15 records

T30M001 The workload decomposition group mapping to be used with type 30

record processing. LRECL is 80.

IOCP001 The IOCP dataset. LRECL is 80.

The program will load the parameters for all of your subruns. If errors are found in any of the

inputs the run will stop. If all parameters are successfully processed the PGNMAPs and


BCUMAPs will be generated and processed. Assuming no errors are found in this stage the

program will continue.

The input parameters will apply only to the “subrun” in which they are specified with one exception -- “SORT=YES” will cause the input to be sorted for all “subruns”.

4.7c JCL for Multiple Outputs

This is an example of a job with 3 subruns.

//VS1PASS JOB (????,????),MSGLEVEL=1,MSGCLASS=O,NOTIFY=VSPENCE, // REGION=64M //* //* THIS IS JCL TO RUN THE EXTRACT PROGRAM //* TO PRODUCE MULTIPLE EDF OUTPUT FILES //* IN A SINGLE RUN //*-------------------- //EXTR EXEC PGM=LOADER //* PRINT DATA SETS FOR MESSAGES FROM YOUR RUN //PRINT001 DD SYSOUT=* //PRINT002 DD SYSOUT=* //PRINT003 DD SYSOUT=* //* INPUT SMF DATASET //SMFIN DD DISP=SHR,DSN=A11PTG.ACXIOM.SMF //* //*---------------- OUTPUT DATASETS FOR EDFI FILES //EDF001 DD DISP=SHR,DSN=VSPENCE.CPSTOOLS.JCL(EDF1S) //EDF002 DD DISP=SHR,DSN=VSPENCE.CPSTOOLS.JCL(EDF2S) //EDF003 DD DISP=SHR,DSN=VSPENCE.CPSTOOLS.JCL(EDF3S) //* //*---------------- OUTPUT DATASETS FOR WORKLOAD GROUPING DATASET //PGN001 DD DISP=SHR,DSN=VSPENCE.CPSTOOLS.JCL(PGNEDF1S) //PGN002 DD DISP=SHR,DSN=VSPENCE.CPSTOOLS.JCL(PGNEDF2S) //PGN003 DD DISP=SHR,DSN=VSPENCE.CPSTOOLS.JCL(PGNEDF3S) //* //*---------------- OUTPUT DATASETS FOR AUTO BCU MAP //BCU001 DD DISP=SHR,DSN=VSPENCE.CPSTOOLS.JCL(BCUEDF1S) //BCU002 DD DISP=SHR,DSN=VSPENCE.CPSTOOLS.JCL(BCUEDF2S) //BCU003 DD DISP=SHR,DSN=VSPENCE.CPSTOOLS.JCL(BCUEDF3S) //* //*---------------- COMPAT MODE PROCESSING OF IEAIPS TO GENERATE PGNMAP //*IPS001 DD DSN=SYS1.PARMLIB(IEAIPS00),DISP=SHR //* INPUT DATASET FOR IOCP PROCESSING (FICON) //*IOCP001 DD DISP=SHR,DSN=PATTRSN.CPSTOOLS.JCL(IOCP) //* //*SORTMSGS DD SYSOUT=* //SYSOUT DD SYSOUT=* //* //SORTWK01 DD UNIT=SYSDA,SPACE=(CYL,(200)) //SORTWK02 DD UNIT=SYSDA,SPACE=(CYL,(200)) //SORTWK03 DD UNIT=SYSDA,SPACE=(CYL,(200)) //SORTWK04 DD UNIT=SYSDA,SPACE=(CYL,(200))

One EDF output for each subrun


//*-------------------- # OUTPUT DATASET FOR TYPE30 RECORDS //* NOTE: DCB INFORMATION FOR THESE DATASETS IS SUBJECT TO CHANGE. //* SO IT IS BEST TO LET THE PROGRAM ALLOCATE THEM. //*TYP30001 DD SPACE=(CYL,(20,10),RLSE),DISP=(,CATLG),UNIT=SYSDA, //* DSN=PATTRSN.TYPE30.DATA //* # OUTPUT DATASET FOR SAVE= PARAMETER (SAVE CP2K'S SMF RECORDS) //*SMFSAVE DD SPACE=(CYL,(20,10),RLSE),DISP=(,CATLG),UNIT=SYSDA, //* BLKSIZE=8192,DSN=PATTRSN.SMF.SAVE //* # OPTIONAL USER DUMP DATASET //*SYSUDUMP DD SYSOUT=* //* //*-------------------- //SYSIN001 DD * ENT=SHIFT 1 DATA IN A SINGLE PASS ### YOUR COMPANY NAME SYSID=Z390 ### JES SYSID DURATION=00:15 RMF/SMF DURATION FOR GATHER * DURATION >= RMF INTERVAL TIME=(08:00-15:00) CONTINGUOUS PERIOD OF INTEREST DATE=(03/28/10-04/03/10) ### CONTINGUOUS DAYS PGN=GOAL BCU=AUTO *SHOWACT=NONE SHOWACT=NONE REDUCES OUTPUT SIZE. SORT=YES IF RECORDS ARE IN SORTED ORDER * OMITTING SORT=YES WILL SPEED THINGS UP. PROCESS=YES /* //*-------------------- //SYSIN002 DD * ENT=SHIFT 2 DATA IN A SINGLE PASS ### YOUR COMPANY NAME SYSID=Q390 ### JES SYSID DURATION=00:15 RMF/SMF DURATION FOR GATHER * DURATION >= RMF INTERVAL TIME=(15:00-23:00) CONTINGUOUS PERIOD OF INTEREST DATE=(03/28/10-04/03/10) ### CONTINGUOUS DAYS PGN=GOAL BCU=AUTO *SHOWACT=NONE SHOWACT=NONE REDUCES OUTPUT SIZE. SORT=YES IF RECORDS ARE IN SORTED ORDER * OMITTING SORT=YES WILL SPEED THINGS UP. PROCESS=YES /* //*-------------------- //SYSIN003 DD * ENT=SHIFT 3 DATA IN A SINGLE PASS ### YOUR COMPANY NAME SYSID=F390 ### JES SYSID DURATION=00:15 RMF/SMF DURATION FOR GATHER * DURATION >= RMF INTERVAL TIME=(00:00-08:00) CONTINGUOUS PERIOD OF INTEREST DATE=(03/28/10-04/03/10) ### CONTINGUOUS DAYS PGN=GOAL BCU=AUTO *SHOWACT=NONE SHOWACT=NONE REDUCES OUTPUT SIZE. SORT=YES IF RECORDS ARE IN SORTED ORDER * OMITTING SORT=YES WILL SPEED THINGS UP. PROCESS=YES /* //*-------------------- MESSAGES FROM THE MVS LOADER //SYSLOUT DD SYSOUT=* //* THE OBJECT MODULE OF THE PROGRAM //SYSLIN DD DISP=SHR,DSN=PATTRSN.CPSTOOLS.JCL(ZOBJEXTR)

Input parameters for

the first subrun


the second subrun


the third subrun


4.8 JCL Generalized for All Job Types

The EXTRALL version of the JCL is a general purpose solution to running the Extract program.

Can be used for any of the CPS Tools programs -- zPCR, zBNA or zCP3000.

Generates the files that any of the programs might need.

Outputs the EDF, the PGNMAP, the BCUMAP, and the DATA file.

Extracts from SMF type 14, 15, 30, 42, 7x, 89, and 113 records

Includes all options, although some that are less used may appear as comments.

The highlighted fields must be modified as needed for your installation.

//PATTRSN JOB (????,????),MSGLEVEL=1,MSGCLASS=O,NOTIFY=???????? //* // EXEC PGM=IEFBR14 CLEAN UP //DD1 DD DISP=(MOD,DELETE),UNIT=SYSDA,SPACE=(TRK,1), // DSN=PATTRSN.OTHRTYPS.DATA //DD2 DD DISP=(MOD,DELETE),UNIT=SYSDA,SPACE=(TRK,1), // DSN=PATTRSN.DATA.TRS //* ---------------------------------------- //EXTR EXEC PGM=LOADER //* PRINT DATA SET FOR MESSAGES FROM YOUR RUN //PRINT001 DD SYSOUT=* //* INPUT SMF DATASET //SMFIN DD DISP=SHR,DSN=PATTRSN.INPUT.SMF //* OUTPUT DATASET FOR EDFI FILE //EDF001 DD DISP=SHR,DSN=PATTRSN.CPSTOOLS.JCL(EDFI) //* OUTPUT DATASET FOR WORKLOAD GROUPING DATASET //PGN001 DD DISP=SHR,DSN=PATTRSN.CPSTOOLS.JCL(PGNMAP) //* OUTPUT DATASET FOR AUTO BCU MAP //BCU001 DD DISP=SHR,DSN=PATTRSN.CPSTOOLS.JCL(BCUMAP) //* INPUT DATASET FOR IOCP PROCESSING (FICON ANALYSIS) //*IOCP001 DD DISP=SHR,DSN=PATTRSN.CPSTOOLS.JCL(IOCP) //* DDS FOR OPTIONAL INTERNAL SORT OF SMF RECORDS //*SORTMSGS DD SYSOUT=* //SYSOUT DD SYSOUT=* //* //SORTWK01 DD UNIT=SYSDA,SPACE=(CYL,(500)) //SORTWK02 DD UNIT=SYSDA,SPACE=(CYL,(500)) //SORTWK03 DD UNIT=SYSDA,SPACE=(CYL,(500)) //SORTWK04 DD UNIT=SYSDA,SPACE=(CYL,(500)) //* //* OUTPUT DATASET FOR TYPE14, 15, 30 & 42 RECORDS //* NOTE: DCB INFORMATION FOR THIS DATASETS IS SUBJECT TO CHANGE. //* SO IT IS BEST TO LET THE PROGRAM ALLOCATE THEM. //DATA001 DD SPACE=(CYL,(300,100),RLSE),DISP=(,CATLG),UNIT=SYSDA, // DSN=PATTRSN.OTHRTYPS.DATA //* //* OUTPUT DATASET FOR SAVE= PARAMETER (SAVE CP3K'S SMF RECORDS) //*SMFSAVE DD SPACE=(CYL,(20,10),RLSE),DISP=(,CATLG),UNIT=SYSDA, //* BLKSIZE=8192,DSN=PATTRSN.SMF.SAVE //* //* # OPTIONAL USER DUMP DATASET //*SYSUDUMP DD SYSOUT=*

This step removes the 2 data sets that will

be created below

Input SMF data set

Output EDF goes to

other applications

PGN map:

Workload data

DATA file:

info for zBNA

BCU map:

I/O configuration


EXTRALL -- Generalized JCL (continued) //* ---------------------------------------- //SYSIN001 DD * ENT='YOUR COMPANY' YOUR COMPANY NAME SYSID=SY01 SMF SYSID DURATION=01:00 DURATION FOR SUMMARY SAMPLES * DURATION MUST BE A MULTIPLE OF RMF INTERVAL *TIME=(00:00-24:00) CONTINGUOUS PERIOD OF INTEREST *DATE=(08/05/09-08/12/09) CONTINGUOUS DAYS PGN=GOAL GENERATE PGNMAP BCU=AUTO GENERATE BCUMAP SORT=YES IF RECORDS ARE IN SORTED ORDER * OMITTING SORT=YES WILL SPEED THINGS UP. /* //* ---------------------------------------- //* MESSAGES FROM THE MVS LOADER //SYSLOUT DD SYSOUT=* //* THE OBJECT MODULE OF THE PROGRAM //SYSLIN DD DISP=SHR,DSN=PATTRSN.CPSTOOLS.JCL(ZOBJEXTR) //* //* ---------------------------------------- //* THIS STEP TERSES THE DATA FILE TO MINIMIZE NETWORK BANDWIDTH //* IT WILL ISSUE RC=10, IF THERE ARE NO RECORDS IN THE INPUT DSN. //* //T EXEC PGM=TRSMAIN,PARM=SPACK //SYSPRINT DD SYSOUT=* //INFILE DD DISP=SHR,DSN=*.EXTR.DATA001 //OUTFILE DD DSN=PATTRSN.DATA.TRS,DISP=(,CATLG,DELETE), // SPACE=(CYL,(50,15),RLSE),UNIT=SYSDA

Input parameters

for the Extract run

Be sure this SYSID is in

the SMF input data set

TERSE output

data set


5. SMF Input Data Requirements

The following SMF record types and subtypes are used by the Extract program:

Record

Type

Type of Data

Required for

Subtypes Used

14 Input DS Activity For zBNA

15 Output DS Activity For zBNA

30 Common Address

Space Work

(accounting information)

For zBNA, and

for T30MAP

processing or creating the

DATA001 dataset

2 - Activity since previous interval end

3 - Activity for interval before step end

4 - Step totals 5 - Job termination

42 DFSMS Statistics

(Data Facility Storage

Management

Subsystem)

For zBNA

If TYPE42=YES

also for zCP3000

6 - DASD data set level I/O statistics

70 CPU activity record Required, except for type 30

processing

1 - CPU, LPAR, coupling facility data 2 - Data for cryptographic coprocessors

71 Paging Activity Only for system

paging statistics.

1 - Central storage page Information

72 Workload Activity for PGN=GOAL 3 – Workload activity data

73 Channel Path Activity for BCU=AUTO

74 Device Activity for BCU=AUTO 1 – Device activity

4 – Coupling facility

5 – Cache subsystem 8 - Enterprise Disk System

75 Page Data Set Activity For auxiliary

storage

information

78 I/O Queuing Activity for BCU=AUTO 2 – Virtual storage

3 – I/O queuing & HiperPAV

113 Hardware capacity all tools Contains hardware capacity, reporting,

and statistics for IBM System z10 or later machines.

2 – Hardware data event counters


SMF records are expected to be sorted by DATE and Time. If they are not, use SORT=YES

control statement. Or use the JCL in part 4.6 to rectify that situation.

While at it, the file can be reduced in size by selecting only the required records. Do this with these input parameters: Date, Time, Select, and Exclude.

Record Types 76, 77, and 79 are currently ignored.

Note: the Extract Program will eliminate duplicate records in the input stream. Occasionally

customer procedures are such that records are duplicated in the SMF input stream. If records

are discarded a message will indicate how many have been discarded.

Note: If you have CMF (Boole and Babbage's equivalent of monitor I) data it is successfully processed by Extract Program as well.


6. Map Files

There are three supplemental map files that can help provide additional control in processing.

Map Use

BCU Combine control units with identical channel configurations

PGN Combine service classes for reporting

T30 Break out workloads in greater detail using type 30 records

6.1 BCU Map

BCU (basic control unit) is a term used in the Extract program to represent a physical control

unit -- the “box” that may contain multiple LCUs (logical control units). The BCU map provides

a mechanism to combine LCUs that are part of the same physical control unit. I.e., a number of LCUs of a single current technology control unit will be reported as only one control unit.

6.1a How the Program Builds the BCU Map

When you request that a BCU Map be built by specifying BCU=AUTO the processing is as follows (type 74 subtype 1 records are required to do BCU processing. If not present you

should omit the BCU DD statement and the BCU=AUTO parameter).

1. If you provide a DD for the IOCP001 DD the associated dataset with IOCP data is read

in. Normally the IOCDS is only needed for the FICON aggregation function.

2. Then in all cases the SMF data is read in.

Type 74 subtype 1 - for device information (required)

Type 74 subtype 5 - for control unit caches and names

Type 78 subtype 1 - for LCU to path information

3. Next combining of Disk LCUs into BCUs takes place. The first combining takes place using NED (node element descriptor) data. NED data is control unit/device data that a

device sends to the host in response to a particular command. Control unit NED data is

in 74.5 records and device NED data is in 74.1 records.

4. Using the NED data the program can determine what a physical box is (i.e., the BCU)

and will combine all LCUs and channels for a given NED Id into that one BCU.

5. If BCU=FICONAG is specified only LCUs with identical channel configurations are combined based on NED data. This normally results in 2 BCUs for a shark control unit.

The result of this is that we should have essentially one BCU per box (unless we have used

BCU=FICONAG). If BCU=COM has been specified additional combining will take place. Since

we are combining control units with identical channel configurations it means that control

units on copper channels where the control units are daisy chained are combined.

Also if BCU=COM is specified ESCON control units that share the same directors are

combined. (This does not happen with the IOCDS option because channel information also

includes the ESCON switch and port. These cannot be identical and are not combined.)

If there are 74 subtype 5 records the NED ID determines the type/model, manufacturer, and

serial number. Here accurate combining is assured. Without 74 subtype 5 data there is device


level NED data in the 74 subtype 1 records but it is at the device level. If they are all reporting

the same for all devices on a control unit it is possible to correctly combine the devices.

The BCU map provides an address map for z/OS data which identifies the BCU characteristics. Three types of statements can be in the BCU map:

TIMESTMP has environmental information and will be first

BCUx where x is either blank, T or O defines the BCU characteristics.

ADDRS defines address ranges.

It is important that the BCUx statement labels match across the different systems so that

zCP3000 can correlate them. An example is in part 4.3. Here are the OpCodes and operands

used in the BCU map file:

OpCode BCU Operands

TIMESTMP Defines Environmental information,

Generated for:

SYSID= BCU=IOCP and BCU=AUTO. Required

PART= BCU=IOCP. The partition name in the IOCP definition

IOCPDATE= BCU=IOCP. The date from the IOCP definition if present

IOCPTIME= BCU=IOCP. The time from the IOCP definition if present

IODFDATE= BCU=IOCP. Last date of the IODF file from type 78 records

IODFTIME= BCU=IOCP. The time of the IODF file from type 78 records

BCU

BCUT

BCUO

The BCU identifier ends with: blank for DASD, T or Tape, and O for other.

These opcodes will each have a label -- a character string starting in position 1

to uniquely identify this BCU across the enterprise. For shared DASD this

string should be the same for each SYSID mapping.

CTYPE= Required for RMF before 4.2.0. Device type, e.g. 3990-3

CACHE= Cache size if controller is cached. Defaults to 0

NVS= Non-volatile storage size if controller is cached. Defaults to 0

CHPID= The chpids used by this BCU

SWCHS= The switches corresponding to the CHPIDs

LINKS= The switch links corresponding to the CHPIDs

CHPTYP= The CHPID types corresponding to the CHPIDs

ADDRS The address identifier includes these sub-fields:

AD= The range of addresses defined by this line

LCU= The LCU number for these addresses

RMF= Y or N whether RMF data was found for this range of devices

DEVT= The device type (e.g. 3380K)


6.1b BCU Map – Sample Output with BCU=AUTO

TIMESTMP SYSID=PRD0 00010001 VSPSSD0 BCUD CTYPE=3990-2,CACH=0,NVS=0, X00020002 MFR=HTC,SER=03111, X00020003 CHPID=(4C,4D,7C) 00020004 ADDRS AD=0200-0202,LCU=009C,RMF=Y,DEVT=33903 00020005 VDEVTB7 BCUD CTYPE=9393-2,CACH=2048,NVS=8, X00030006 MFR=IBM,SER=24325,PLNT=91, X00030007 CHPID=(50,51,52,53,6E,6F,80,81) 00030008 ADDRS AD=D600-D63F,LCU=01AC,RMF=Y,DEVT=33903,SSID=0061 00030009 ADDRS AD=D640-D67F,LCU=01B0,RMF=Y,DEVT=33903,SSID=0062 00030010 ADDRS AD=D680-D6BF,LCU=01B4,RMF=Y,DEVT=33903,SSID=0063 00030011 ADDRS AD=D6C0-D6FF,LCU=01B8,RMF=Y,DEVT=33903,SSID=0064 00030012 VDEVT29 BCUD CTYPE=9393-2,CACH=1024,NVS=8, X00040013 MFR=IBM,SER=23774,PLNT=91, X00040014 CHPID=(50,51,52,53,6E,6F,80,81) 00040015 ADDRS AD=D700-D73F,LCU=01BC,RMF=Y,DEVT=33903,SSID=0071 00040016 ADDRS AD=D740-D77F,LCU=01C0,RMF=Y,DEVT=33903,SSID=0072 00040017 ADDRS AD=D780-D7BF,LCU=01C4,RMF=Y,DEVT=33903,SSID=0073 00040018 ADDRS AD=D7C0-D7FF,LCU=01C8,RMF=Y,DEVT=33903,SSID=0074 00040019


6.2 PGN Map

The PGNMAP (Performance Group Name Mapping) is a working file for the WLM (workload

management) description of the workload and service class (business units). (The earlier

performance group name convention has given way to goal mode operation but the old name is

still being used here).

The PGN Mapping mechanism includes a way for you to manually combine service classes for

reporting. However, the newer zCP3000 application will handle this automatically so manual

adjustments are usually not necessary.

6.2a PGN Map – Reporting Classes

Reporting classes are included in the EDF and can be viewed in zCP3000. Analysis of most

aspects of workloads can be done with these.

Reporting classes can be used in place of a service class but be sure to understand how the

reporting classes are set up. A reporting class must correspond to a given workload class.

Since this does not always happen, problems can occur.

If you are going to make manual adjustments to the service classes the following steps outline

a procedure to verify that you have good data. (An example is in part 6.2b).

1. Do one standard Extract run with PGN=GOAL. This will build a PGNMAP file.

2. Open the PGNMAP file and find the entries for the reporting performance groups – they

will contain “RPT=1”.

3. Change the reporting classes you want to “RPT=0”. Then comment out the

corresponding service classes.

4. Also comment out the workload called “THE REST” – it will be at the very end of the list

of workloads.

5. Remove PGN=GOAL and run the Extract again.

6. Load the two files separately into zCP3000. If everything lined up correctly the capture ratio will be identical in each of the files. If this is not the case the reporting

performance groups do not correspond directly with the service class.

Note: The reporting classes that have been updated in the PGNMAP file will not be available in

the reporting class view in zCP3000. In GOAL Mode it is defined by the WORKLOAD name

(R723MWNM) and the Service Class name (R723MCMN). The Workload and Service Class data is mapped to a unique description.

(PGN=GOAL) builds a description of WORKLOAD.SCLASS as a default. Each description

defines a unique business unit. To combine data from different workload and service class,

make their descriptions be the same.

In the example in part 6.2b both “SSYSTEM.SYSTEM” and “SYSTEM.SYSSTC” are combined

into one unit called “STUFF”.

A mapping for each SYSID can be provided or where a performance group maps to a business

unit for all SYSIDs to be processed an "*" is permitted indicating any SYSID. One workload

name can now appear across multiple SYSIDs in zCP3000.


The PRTY (priority) field is the relative priority of the workload. This is not so easy to establish

in GOAL mode. CP3KEXTR will artificially construct a relative priority value from the

R723CIMP (the importance) field. It will be 11-R723CIMP. It is an optional field.

The PGN mapping file contains the following parameters. Two lines are allowed for each. An

asterisk for a workload name means put anything not yet processed into this bucket. (See “The

Rest” in the example).

SYSID= A character string to identify the SMF ID of the system image corresponding

to this data. If this is valid for all images an '*' may be used.

PGN= This is ignored. PGN=0 is provided by PGN=GOAL as a default.

PER= An optional specification of performance group period. A specification of

PER=* will group all periods. This parameter is optional.

PRTY= Relative priority of the performance group. Parameter is optional and

defaults to 0.

DESC= A unique character string describing the business unit. Field length for CP3KEXTR is 48, but zCP3000 will only display the first 26 characters.

WORKLOAD= A unique character string describing the workload name as it would have

appeared in an RMF workload report. The field length is eight characters.

If this is WORKLOAD=* then all data not already processed will be collected

here (this should be the last entry in the PGNMAP file). See the example for

“The Rest”.

SCLASS= A unique character string describing the service class name as it would have appeared in an RMF workload report. Field length is 8 characters.

Example of a single PGN map entry:

SYSID=X001 PGN=0 PRTY=9 DESC='CICSTEST.FBTEST ' PER=* WORKLOAD=CICSTEST SCLASS=FBTEST


6.2b PGN Map – Sample Using Goal Mode

******************************************************************* * SYSID= MUST BEGIN IN COL 1 TO INDICATE START OF RECORD * DESC= THE DESCRIPTION IS THE WORKLOAD NAME.SERVICE CLASS * PGN= IS UNUSED IN GOAL MODE (GOAL MODE IS WHEN WORKLOAD IS NOT BLANK) * PRTY= IN GOAL MODE IS 10-S723CIMP RELATIVE IMPORTANCE OF GOAL * SCLASS= IS SERVICE CLASS NAME S723MCNM * REPORT CLASSES ARE COMMENTED OUT ******************************************************************* * SYSID=X001 PGN=0 PRTY=9 DESC='CICSTEST.FBTEST ' PER=* WORKLOAD=CICSTEST SCLASS=FBTEST * SYSID=X001 PGN=0 PRTY=8 DESC='TSOTEST.TSOT1 ' PER=1 WORKLOAD=TSOTEST SCLASS=TSOT * SYSID=X001 PGN=0 PRTY=8 DESC='TSOTEST.TSOT2 ' PER=2 WORKLOAD=TSOTEST SCLASS=TSOT * SYSID=X001 PGN=0 PRTY=7 DESC='TSOTEST.TSOT3 ' PER=3 WORKLOAD=TSOTEST SCLASS=TSOT* SYSID=X001 PGN=0 PRTY=6 DESC='TSOTEST.TSOT4 ' PER=4 WORKLOAD=TSOTEST SCLASS=TSOT * SYSID=X001 PGN=0 PRTY=0 DESC='TSOTEST.TSOT* ' PER=* WORKLOAD=TSOTEST SCLASS=TSOT * SYSID=X001 PGN=0 PRTY=0 DESC='STUFF' PER=* WORKLOAD=SYSTEM SCLASS=SYSTEM * SYSID=X001 PGN=0 PRTY=0 DESC='STUFF' PER=* WORKLOAD=SYSTEM SCLASS=SYSSTC * SYSID=X001 PGN=0 PRTY=0 DESC='The Rest' PER=* WORKLOAD=* SCLASS=any * .....

These will be combined for

summarization because the descriptions match


6.2c PGN Map – Sample Using Reporting Classes

******************************************************************* * SYSID= MUST BEGIN IN COL 1 TO INDICATE START OF RECORD * DESC, THE DESCRIPTION IS THE WORKLOAD NAME || SERVICE CLASS * PRTY IN GOAL MODE IS 10-S723CIMP RELATIVE IMPORTANCE OF GOAL * SCLASS IS SERVICE CLASS NAME S723MCNM * RPT = 1 IS REPORTING PERFORMANCE GROUPS ******************************************************************* * SYSID=PROD RPT=0 PRTY=33 DESC='BATCHPRD.BATCHHOT ' PER=* WORKLOAD=BATCHPRD SCLASS=BATCHHOT * SYSID=PROD RPT=0 PRTY=21 DESC='BATCHPRD.BATCHPRD ' PER=* WORKLOAD=BATCHPRD SCLASS=BATCHPRD * SYSID=PROD RPT=0 PRTY=50 DESC='DDFDB2.DDFDB2.1 ' PER=1 WORKLOAD=DDFDB2 SCLASS=DDFDB2 SYSID=PROD RPT=0 PRTY=34 DESC='DDFDB2.DDFDB2.2 ' PER=2 WORKLOAD=DDFDB2 SCLASS=DDFDB2 SYSID=PROD RPT=0 PRTY=22 DESC='DDFDB2.DDFDB2.3 ' PER=3 WORKLOAD=DDFDB2 SCLASS=DDFDB2 SYSID=PROD RPT=0 PRTY=0 DESC='DISCRET.DISCRET ' PER=* WORKLOAD=DISCRET SCLASS=DISCRET * *YSID=PROD RPT=0 PRTY=46 DESC='STC.STCDB2HI ' *PER=* WORKLOAD=STC SCLASS=STCDB2HI * SYSID=PROD RPT=0 PRTY=54 DESC='STC.STCDB2LO ' PER=* WORKLOAD=STC SCLASS=STCDB2LO * *YSID=PROD RPT=0 PRTY=54 DESC='STC.STCDB2MD ' *PER=* WORKLOAD=STC SCLASS=STCDB2MD * SYSID=PROD RPT=0 PRTY=50 DESC='TSO.TSO.1 ' PER=1 WORKLOAD=TSO SCLASS=TSO * SYSID=PROD RPT=0 PRTY=255 DESC='SYSTEM.SYSTEM ' PER=* WORKLOAD=SYSTEM SCLASS=SYSTEM * *YSID=PROD RPT=0 PRTY=254 DESC='SYSTEM.SYSSTC ' *PER=* WORKLOAD=SYSTEM SCLASS=SYSSTC * SYSID=PROD RPT=1 PRTY=21 DESC='.BATCHPRD ' PER=* WORKLOAD= SCLASS=BATCHPRD * SYSID=PROD RPT=0 PRTY=46 DESC='.STCRTC ' PER=* WORKLOAD= SCLASS=STCRTC * SYSID=PROD RPT=1 PRTY=50 DESC='.TSO.1 ' PER=1 WORKLOAD= SCLASS=TSO * *YSID=* PRT=0 PRTY=0 DESC='THE REST ' *PER=* WORKLOAD=* SCLASS=* RPT=0


6.3 T30 Map

There are cases where the breakdowns by workloads is not as granular we would like. In cases

like these we can create a mapping using parameters in Type 30 (work unit accounting) records

to break out workloads in greater detail than is possible with WLM Service Classes. To do this

the customer must be collecting type 30 interval records (Subtype 2 and 3).

There are some differences between the workload data generated by using SMF Type 72 via a

PGNMAP and using SMF Type 30 via a T30MAP.

The T30 data does not include a lot of the SYSTEM information since SMF 30 is only recording address spaces started after SMF. That’s not too much of a problem for the

business unit capture ratio but can be for the overall system CR. If you use T72 data,

the SYSTEM work is often deleted anyway.

The T30 data offers you the opportunity to break down the workload in zCP3000 to a much finer granularity. Often all of CICS is put into one service class which means

that it will appear as one workload in zCP3000. You can split the CICS workload by

using T30 data and jobname as the classification of work.

You should remember to make the DUR parameter in larger than the type 30 interval.

This option will generate an additional set of workloads like those resulting from the PGNMAP

with a different header in the EDF file. Once in zCP3000 you can only process one of these at a

time. By default it will be the PGN001 data, but you can set the profile option “TYPE 30” in zCP3000 so it will read the T30M001 instead.

This file is very similar to the PGN map except there is no way to automatically generate one.

Therefore, you will need to code this mapping file yourself. Refer to part 6.3a . The program

looks through this map and the first match it finds will be the entry where the data is

accumulated.

Each entry begins with SYSID= in column one and can be at most two lines long. It can have at most one field for job name or RACF userid etc. (Note: WLM= and SCN= must both be

specified together.)

For example JOBN=PROD* or JOBN=TEST%% will collect either all jobs which begin with

PROD or all jobs that begin with TEST and two characters. To combine these entries in the

output the DESC fields for both of these should be the same.

A mapping for each SYSID can be provided or where a performance group maps to a business

unit for all SYSIDs to be processed, an "*" is permitted indicating any SYSID.

One workload name can now appear across multiple SYSIDs in zCP3000.

The PRTY (priority) field is a relative priority of the workload. You will need to somehow arrive

at values for the different business units. It is an optional field.

If a type 30 record is found which does not satisfy the criteria of one of the entries, the program will create a new entry for this data. This entry will have a description of "workload

name.service class name".


The PGN mapping file contains the following parameters. Two lines are allowed. SYSID in

column 1 begins the entry.

SYSID= A character string to identify the SMF ID of the system image corresponding

to this data. If this is valid for all images, an '*' may be used.

PRTY= Relative priority of the performance group. Parameter is optional and

defaults to 0.

DESC= A unique character string describing the business unit. Field length for

CP3KEXTR is 48, but zCP3000 will only display the first 26 characters. Workload descriptions must be unique for each SYSID.

JOBN= The job name indicated in the type 30 record.

CLASS= The job class indicated in the type 30 record.

USER= The RACF userid in the type 30 record.

RACF= The RACF group in the type 30 record.

PROG= The program name in the type 30 record.

WLM=/SCN= The workload manager name and the service class name. Both must be

specified. An example is shown in part 6.3a.

6.3a T30 MAP – Sample Data Set

SYSID=ASYS DESC='AP' JOBN=PAP* SYSID=ASYS DESC='BE' JOBN=PBE* SYSID=ASYS DESC='GL' JOBN=PGL* SYSID=ASYS DESC='PC' JOBN=PPC* SYSID=ASYS DESC='PO' JOBN=PPO* SYSID=ASYS DESC='TR' JOBN=PTR* SYSID=ASYS DESC='REST' JOBN=*


7. “Data” Output File

The Extract Program can create the DATA001 file containing selected fields from SMF record

types 30, 42, 14, & 15. This is a flat text file used by the zBNA program.

When these records are encountered in the SMF input the program checks for a DATA001 DD

statement. If present the SMF fields are extracted and written to this file. The same data may also be included in the EDF file.

The first character of each record is a number that indicates the data type (30, 42, 14, or 15)

and all records and self defining records will start with this same code.

code record type contains

1 30 Accounting information

2 42 DFSMS (Data Facility Storage Management Subsystem)

statistics and configuration

3 14, 15 Dataset activity – Input (14) & Output (15)

Each line for a self-defining field entry includes the following fields, separated by commas:

code The record type code (1,2, or 3). Always position 1.

* Constant – marker for self-defining entry. Always position 2.

NAME= The name of the field

N or C Whether the output is numeric or character.

(Note the output in the file will always be in character format.)

COLS= The column in the output in which this field will start

COLE= The column in the output in which this field will end.

Here is an example of the self defining record data:

1*NAME=SMF30TRS,N,COLS=155,COLE=164 1*NAME=SMF30DTE,N,COLS=1,COLE=6 1*NAME=SMF30TME,N,COLS=7,COLE=12

1*NAME=SMF30STP,C,COLS=13,COLE=13

By default only type 30 subtype 4 records (step totals) are processed. However by using an

additional parameter, “SUB30=2,3,4,5” you may specify which other subtypes are to be

included.

For example, SUB30=2,3 indicates that you want just subtypes 2 and 3 and will exclude

subtypes 4 and 5 (i.e., specifying the SUB30 parameter overrides the default subtype 4). The

subtype of the record is included in the output so the post processing program can know the

source.


7.1 Data File – Type 30 Records

Field Name SMF Field Contains Type

SMF30DTE SMF30DTE Date of record creation

SMF30TME SMF30TME Time of record creation

SMF30STP SMF30STP Record subtype

SMF30JBN SMF30JBN Job Name

SMF30PGM SMF30PGM Program name

SMF30WLM SMF30WLM Workload Name

SMF30SCN SMF30SCN Service Class

SMF30RUD SMF30RUD RACF Userid

SMF30TCN SMF30TCN Connect Time Units unchanged

SMF30ABD SMF30ABD Completion A = Abend

N = Normal

SMF30PTY SMF30PTY Dispatching priority Not valid in goal

mode

CPUTIME SMF30 ICU, ISB, IIP, RCT,

HPT, CPT, CPS

CPU Time 0.01 sec

SMF30TEP SMF30TEP Blocks Transferred

SMF30PGI SMF30PGI Pages paged in

SMF30NSW SMF30NSW Swap Sequences

SMF30HPI SMF30HPI Hiperspace page ins

CS SMF30 ARB, EAR, URB,

EUR

Central Storage Kilobytes

SMF30TAT SMF30TAT Transaction Active Time 0.01 sec

SMF30RES SMF30RES Transaction Residency Time

0.01 sec

SMF30PGN SMF30PGN Compatibility Mode Performance Group

SMF30TRS SMF30TRS Number of transactions

SMF30CLS SMF30CLS Job Class

SMF30GRP SMF30GRP RACF Group

SMF30SQT SMF30SQT Waiting for Initiator 0.01 sec

SMF30WID SMF30WID Subsystem identifier

(continued)


30 Data File (continued)

Field Name SMF Field Contains Type

SMF30SRV SMF30SRV Total Service Units Units

SMF30IO SMF30IO I/O Service Units Units

SMF30CSU SMF30CSU (4-byte) or

SMF30CSU_L (8-byte)

CPU Service Units Units

SMF30SRB SMF30SRB SRB Service Units Units

SMF30SCC SMF30SCC Completion code Hex

SMF30STD SMF30STD Initiator select date YYMMDD

SMF30SIT SMF30SID Initiator select time HHMMSShh

SMF30SID SMF30SID JES SysId

SMF30_TIME_ ON_ZIIP

SMF30_TIME_ON_ZIIP zIIP Time 0.01 sec

SMF30_TIME_ ON_ZAAP

SMF30_TIME_ON_ZAAP zAAP Time 0.01 sec

SMF30ZNF SMF30ZNF zAAP Normalization Factor

SMF30SNF SMF30SNF zIIP Normalization Factor

SMF30CPT SMF30CPT CPU Step Time 0.01 sec

SMF30JNM SMF30JNM JES Job Identifier

SMF30STN SMF30STN Step Number

SMF30_Highest_

Task_CPU_Percent

SMF30_Highest_

Task_CPU_Percent

Highest Task Percent

SMF30_Highest_

Task_CPU_Program

SMF30_Highest_

Task_CPU_Program

Highest Task Program

Name

SMF30RCN SMF30RCN Reporting Class

SMF30RSD SMF30RSD Reader Date YYMMDD

SMF30RST SMF30RST Reader Time HHMMSShh

WAITINIT SMF30SIT minus

SMF30RST

Initiator Wait seconds seconds

SMF30ICU SMF30ICU Initiator time under TCB

SMF30RQT SMF30RQT Job Prep ineligible time 0.01 sec

SMF30HQT SMF30HQT Job Prep ineligible time,

other

0.01 sec

SMF30ACL SMF30ACL Accounting section Length SMF30ACL

SMF30ACT SMF30ACT Accounting Data SMF30ACT


7.2 Data File – Type 42 Records

Field Name SMF Field Contains Size Type

SMF42DTE SMF42DTE Date of record creation 6 YYMMDD

SMF42TME SMF42TME Time of record creation 8 HHMMSShh

DURATION SMF42PTE -

SMF42PTS

Interval Duration 10 numeric

S42JDJNM S42JDJNM Job name 8 alpha

S42JDWLM S42JDWLM Workload class 8 alpha

S42JDWSC S42JDWSC Service class 8 alpha

S42DSVOL S42DSVOL Volume serial number 6 alpha

S42DSN S42DSN Data set name 44 alpha

S42DSION S42DSION Total number of I/Os 10

S42DSIOR S42DSIOR Average response time 10 0.1 msec

S42DSIOC S42DSIOC Average I/O connect time 10 0.1 msec

S42DSIOP S42DSIOP Average I/O pending time 10 0.1 msec

S42DSIOD S42DSIOD Average I/O disconnect time 10 0.1 msec

S42DSIOQ S42DSIOQ Average control unit queue time 10 0.1 msec

S42DSBSZ S42DSBSZ Block size 10 numeric

S42BLKRD S42AMSRB +

S42AMDRB

Total blocks read,

sequential & direct

10 numeric

S42BLKWR S42AMSWB +

S42AMDWB

Total blocks written,

sequential & direct

10 numeric

S42DSEF S42DSEF Extended format flag 1 = 0 or 1

S42DSEFC S42DSEFC Compressed format flag 1 = 0 or 1

S42DSTYP S42DSTYP Data set type 3 = 0 - 22

(continued)


42 Data File (continued)


S42JDRSD S42JDRSD Date reader recognized the

JOB card for this job

6 YYMMDD

S42JDRST S42JDRST Time reader recognized the


8 HHMMSShh

S42DSIOSE S42DSIOS Flag to indicate S42DSIOS

field is present

1 = 0 or 1

S42DSIOS S42DSIOS Total number read/write ops to Metro

Mirror secondary

10 numeric

S42FICON S42DSHRD

S42DSHWR

Flag to indicate that either of these FICON

fields are nonzero

1 = 0 or 1

S42DSENC S42DSFL1 Set to 1 when S42DSFL1 x’20’ bit is on

indicating an Encrypted DASD data set

1 = 0 or 1

S42DSENT S42DSENT Encryption type: AES-256 5 numeric

S42DSCMT S42DSCMT Data Set Compression type 3 numeric

S42AMRIB S42AMRIB Number of bytes read 10 numeric

S42AMWIB S42AMWIB Number of bytes written 10 numeric

S42AMRBD S42AMRBD For encrypted data sets, number of bytes

decrypted when reading.

For non-encrypted data sets, number of

bytes that would be eligible for decryption

when reading if the data set was allocated

as encrypted

10 numeric

S42AMWBE S42AMWBE For encrypted data sets, nbr of bytes

encrypted when writing For non-encrypted data sets, nbr of bytes

that would be eligible for encryption when

writing if the data set was allocated as

encrypted.

10 numeric

S42AMRCI S42AMRCI Number of VSAM CIs read.

For non-VSAM, this is the number of

physical blocks

10 numeric

S42AMWCI S42AMWCI Number of VSAM CIs written.

For non-VSAM, this is the number of

physical blocks.

10 numeric


S42DSRDT S42DSRDT Total number of read operations 10 numeric

S42SNERD S42SNERD Number of read requests eligible for

synchronous I/O

10 numeric

S42SNERH S42SNERH Number of read hits eligible for

synchronous I/O

10 numeric

S42SNEWR S42SNEWR Number of write requests eligible for

synchronous I/O

10 numeric

S42SNRDT S42SNRDT Number of synchronous I/O read

attempts

10 numeric

S42SNROS S42SNROS Number of synchronous I/O read successes

10 numeric

S42SNWTT S42SNWTT Number of synchronous I/O write attempts

10 numeric

S42SNWOS S42SNWOS Number of synchronous I/O write successes

10 numeric

S42DSA1U S42DSA1U Average I/O device-active-only time 1us 10 numeric, usec

S42DST1U S42DST1U Average read disconnect time 10 numeric, usec

S42DSB1U S42DSB1U Average device busy time 10 numeric, usec

S42DSM1U S42DSM1U Average initial command response time 10 numeric, usec

S42DSHRD S42DSHRD zHPF Read count 10 numeric

S42DSHWR S42DSHWR zHPF Write count 10 numeric

S42DSRRU S42DSRRU Average random read cache hit response

time

10 numeric

S42DSRSU S42DSRSU Average random read cache hit service

time

10 numeric

S42SNRDU S42SNRDU Average sync_io Read Time in 1us 10 numeric, usec

S42SNWTU S42SNWTU Average sync_io Write Time in 1us 10 numeric, usec


7.3 Data File – Type 14 & 15 Records


SMF14DTE SMF14DTE Date of record creation 6 YYMMDD

SMF14TME SMF14TME Time of record creation 8 HHMMSShh

SMF14RTY SMF14RTY Record type 2 = 14 or 15

SMF14JBN SMF14JBN Job name 8 alpha

SMF14RSD SMF14RSD Date reader recognized the


6 cyyddd

SMF14RST SMF14RST Time reader recognized the


8 HHMMSShh

SMF14DSN JFCBDSNM

from SMFJFCB1

Dataset Name 44 alpha

SMF14DSG JFCDSRG1 from SMFJFCB1

DSORG 2 hex

SMF14RCF JFCRECFM from SMFJFCB1

Record format 2 hex

SMF14LRE JFCLRECL from SMFJFCB1

Logical record length 8 numeric

SMF14BLK JFCLRECL

from SMFJFCB1

Block size 10 numeric

SMF14DCL SMF14DCN Data class name 8 alpha

SMF14STP SMF14STP Step name 8 alpha

SMF14CMP SMF14STY = 1 ‘Y’ if compression segment 1 alpha

SMF14DSV SMF14DSVER Data set version 1 numeric

SMF14XF1 SMF14XF1 Compression indicator flags 2 hex

SMF14RIN SMF14RIN Record and DS indicator 4 hex

SMFDCBMF SMFDCBMF Type of I/O macro instruction

and options. Mapped by

DCBMACRF in the DCB

mapping

4 hex

(continued)


14 & 15 Data File (continued)


SMF14CDL SMF14CDL Number of bytes of compressed

data read or written

8 numeric

SMF14UDL SMF14UDL Data length prior to compression 8 numeric

SMF14CDS SMF14CDS Number of compressed user data

bytes

8 numeric

SMF14UDS SMF14UDS Size of compressed format data

set (number of uncompressed

user data bytes)

8 numeric

SMF14CIS SMF14CIS Physical block size of extended

format data set

8 numeric

SMFEXCP SMFEXCP Step EXCP count -- sum of

SMFEXCP for all UCBs

8 numeric

SMF14EXT SMF14NEX Number of extents -- sum of

SMF14NEX for all UCBs

8 numeric

SMF14EDI SMF14EDI Enhanced Data Integrity (EDI)

flag indicator

2 hex

SMF14CMPTYPE SMF14CMPTYPE Compression type indicators 2 hex


8. Errors

The Extract Program can return different return codes:

0 No errors or warnings


8 A warning that messages have been issued that you should review. Either important

SMF/RMF record types were missing from the input file or there is questionable data

in some of the intervals (that you may want to exclude).

12 Something is critically wrong with the parameters, or no data fit within the SYSID and timeframe requested.

16 Something is drastically wrong and the program failed.


225 No data was found that meets the input specification. This could happen if the

specified SysId was not present in the SMF input dataset or if the date and time

selections did not cover the time periods of the data.


9. EDF Output

The EDF (Enterprise Data File) is the primary output of the Extract program. It is a flat text

file that contains the summarized information drawn from the SMF scan. It is an input to

other analysis applications like zBNA, zCP3000, and zPCR.

9.1 EDF Format

The EDF is a text file with 80-character fixed-size records divided into named sections. A section name begins in the first position of a line while subsequent section content always

starts in position 7. Each record is organized as follows:

Column

1-4 Section name.

5 "S" when a section is specified. (This value will change when zCP3000 uses this format for output.)

6 Always blank

7-72 Information in the form KEYWORD=data appropriate to the section.

The section content consists of a series of expressions, one for each target SMF field. An

expression has a key name followed by an equal sign followed by one or more values. Some

have just a single value but more often there will be a series of multiple values (a vector) that is made up of the key name, the equal sign, and the following sequence of items:

Average of the individual values

Minimum value

Maximum value

Standard deviation of the individual values

Count of values: 5 summary items + number of individual values (see below)

List of individual values – one for each reporting interval (see below)

The first 5 items, the summary values, are always present in a multi-item vector. But if all the

individual values are all the same the list of individual values will be suppressed. In this case

the minimum and the maximum will be equal and the count will be 5 (the number of summary

items). If the individual values are not all the same then the entire list of individual values will

be present and the count will be 5 + the number of individual items.

Issues With Averaging

The average calculation includes all of the individual values, one for each reporting interval,

with equal weight. This means that certain measures may result in an average that gives a

misleading impression. In such a case the user may wish to examine the detail data.

For example, the DASDRSPV (DASD Response Time) vector may include intervals with very few

I/Os along with an interval that has many. The resulting average may not be useful.

Other examples are: DELAYV (the average wait for a tape mount) or ALLOCV (the average

allocation time for a tape drive). Intervals which have no mounts will have zero for these

intervals and will distort the average calculation.


9.2 Sample EDF output

Here is an example of EDF output:

HEAD ENT=zCP3000 Inc. R70INT=30 SMFDSN=MVSSPT.AQFT.SMF.D08158 SOURCE=CP3KEXTR02/16/16 CEC S CECID=CEC9F30 CPUMOD=2097-742 SUPVR=LPAR VC=00 PR=46 SR=09F30 CPV=42 42 42 0 10 IFLV=2 2 2 0 10 ZIIPV=2 2 2 0 10 CPUMODV=2097-742 CMIND=1 1 1 0 10 HWCMODV=2097-E56 HWIND=1 1 1 0 10 SYS S SYSID=AQFT SCP=Z/OS VERSION=ZV010900 HPTSID=MCLXCF01 NSAMPS=10 GL=1 WC=0 BIT=64 RMFV=RMF0719 HIPDSP=1 PAR=1.85 RMFINT=1 DTSRC=RMF RMFINTL=30 SRM=35476.72 GMTOF=-4:00 SYSN=AQFT DASDIO=1553.4 PAGE=1 CS=25601 ES=0 CSAVAIL=12824.1 ESAVAIL=0 SCPCS=275.7 SCPES=0 CECUTILV=18.4 14.5 21.3 1.7 10 18.1 19.8 17 18.8 21.3 19.6 17.6 19.4 18.1 14.5 PAGEV=1 0 3.5 1.2 10 0 0.2 0.7 0.5 3.5 2 2.4 0.5 0 0 PAGEDS=7 7 7 0 10 CSV=25601 25601 25601 0 10 CSAVAILV=12824.1 11745.8 14676.4 880.2 10 14676.4 14083.1 12989.6 13114.4 12629.9 12346.7 12460.6 12132.7 11745.8 12061.8 SCPCSV=275.7 274.6 276.3 0.6 10 274.7 275.2 274.6 276.1 276 276.2 276.3 276 275.9 275.5 AUIC=65535 65535 65535 0 10 MUIC=65535 65535 65535 0 10 MXUIC=65535 65535 65535 0 10 AINR=2.2 1.7 3 0.4 10 3 2.8 1.7 2.3 2 2.2 2.1 2.1 1.8 1.7 AIN=156.2 131.1 172.8 13.4 10 131.1 136.9 142.9 161.8 162 165.8 166.5 172.8 165 156.9 LOGSWAP=100 100 100 0 10 LOGEFF=100 100 100 0 10 BATCHV=5 2.6 7.6 1.4 10 6.1 7.6 6.6 5 5.2 3.9 4.9 4.7 3.8 2.6 STCV=201.8 193.8 220.5 7.9 10 199.5 193.8 195.1 200.2 195.2 202.9 197.9 201.4 211.9 220.5 TSOV=121 44.5 153.7 35.8 10 44.5 72.4 113.6 135.5 151.2 153.7 153.1 149.4 134.4 102.6 OCPU=0 0 0.2 0.1 10 0 0 0.2 0 0 0 0 0 0 0 OCPU2=0 0 0.2 0.1 10 0 0 0.2 0 0 0 0 0 0 0 LREV=0.1 0 0.1 0 10 0 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0 DASDIOV=1553.4 981.5 2078.1 307.9 10 1319.7 1297.9 1964.5 2078.1 1519.9 1400.6 1659.2 1653.1 1659.4 981.5 DASDRSPV=1.7 1.4 2.2 0.3 10 1.4 1.4 2 1.6 1.7 2 1.8 2.2 1.4 1.6 DASDSERV=1.2 1 1.5 0.2 10 1 1 1.5 1.2 1.3 1.3 1.1 1.5 1 1.2 TAPEIOV=12.8 0.7 42.6 13.2 10 42.6 7.6 3.3 19 0.7 9.1 30.8 7.8 5.6 1.7 TAPEALC=10.5 6 14 2.3 10 14 10 11 10 10 12 10 8 6 14 IFANF=256 256 256 0 10 IIPNF=256 256 256 0 10 WLA=718 718 718 0 10 LAC=86.6 67 107.5 13.1 10 96.5 107.5 102 99 87 76.5 77 77 76 67 SMF23NFR=24587.1 11934.9 37324.7 7912.2 10 11934.9 18047.3 37324.7 31933.1 24180 26533.7 22917.6 35236.9 22175.6 15587.5 SMF231RF=12127.7 5588.6 18041.3 3368.3 10 5588.6 10803.6 12779.8 15498.2 12369.2 14437 12095.1 18041.3 11600.2 8064.1 SMF23NIO=393.8 159.7 510.1 109.7 10 159.7 325.8 440.3 488.2 510.1 450.3 476 481.7 336.7 269.6 SMF23SRB=16174 8312.1 24332 5036.4 10 11633.8 8312.1 18990 24332 19838.6 17560.6 14029.8 22198.6 14907.7 9937.2 LPARNO=1 LPAR=AQFT VICTEST VMTOOL1 AQCF1 AQHO AQLINX HOCF4 GDLVM7 LNXVM14 POKVMXA1 PHYSICAL LPWGT=0 215 0 0 3 0 0 315 0 -1 -1 -1 0 2 0 0 1 0 -1 -1 -1 0 32 0 0 4 0 0 32 0 LPCAP=0 0 0 0 0 0 0 0 0 0 LPCL=AQFT 0 0 0 AQHO 0 0 0 0 0 LPCS=25600 5120 65536 3072 6144 2048 3072 8192 24576 131072 LPES=0 5120 65536 0 0 1024 0 2048 8192 65536 LPPRC1=11 11 11 0 10 LPPRP1=2 2 2 0 10 LPPCTM1=4740.1 2624 7705.2 1573.7 10 7705.2 7245.5 2624 4953.9 4217.2 5348 4198.3 4508.7 3433.3 3167.1 LPPPTM1=13.3 5.4 51.4 13 10 11.6 12.8 13.9 6.2 9.2 8.3 5.4 7.1 7.1 51.4 LPWT1=215 215 215 0 10 LPWP1=215 215 215 0 10 LPWTT1=215 215 215 0 10 LPPRC2=5 5 5 0 10 LPPCTM2=14.3 13 17.4 1.1 10 13 14.4 13.5 13.6 13.9 14.1 14.4 14.5 17.4 14.4 LPWT2=3 3 3 0 10 LPWTT2=3 3 3 0 10 LPPRC3=16 16 16 0 10 LPPCTM3=13681.6 9660.2 18242.7 2191.7 10 10894.6 13420.5 14072.6 14065.3 18242.7 14230.9 12961.8 15016.7 14250.7 9660.2 LPWT3=315 315 315 0 10 LPWTT3=315 315 315 0 10 LPPRC4=1 1 1 0 10 LPPCTM4=3599.8 3599.8 3599.9 0 10 3599.8 3599.8 3599.8 3599.9 3599.9 3599.8 3599.9 3599.8 3599.9 3599.8 LPWT4=-1 -1 -1 0 10 LPWTN4=-1 -1 -1 0 10 LPWTT4=-1 -1 -1 0 10


LPWTX4=-1 -1 -1 0 10 LPPRC5=5 5 5 0 10 LPPRP5=2 2 2 0 10 LPPCTM5=83.1 79.7 85.8 1.8 10 81.1 79.7 82 82.4 83.7 83.8 85.4 84.2 85.8 82.9 LPWT5=2 2 2 0 10 LPWP5=2 2 2 0 10 LPWTT5=2 2 2 0 10 LPPRC6=2 2 2 0 10 LPWT6=1 1 1 0 10 LPWTT6=1 1 1 0 10 LPPRC7=1 1 1 0 10 LPPCTM7=3584.6 3571.2 3596.1 9 10 3592.7 3595.1 3596.1 3582.6 3578.7 3571.2 3588.4 3572.4 3576.6 3591.9 LPWT7=-1 -1 -1 0 10 LPWTN7=-1 -1 -1 0 10 LPWTT7=-1 -1 -1 0 10 LPWTX7=-1 -1 -1 0 10 LPPRC8=3 3 3 0 10 LPPCTM8=591 464.5 775.5 102 10 522.8 775.5 464.5 628.2 586.1 489.5 663.2 482.3 732.5 565.3 LPWT8=32 32 32 0 10 LPWTT8=32 32 32 0 10 LPPRL9=2 2 2 0 10 LPPLTM9=23.2 22.6 24.3 0.5 10 22.7 23 22.6 22.9 23.4 24.1 23.2 24.3 23.2 22.9 LPWT9=4 4 4 0 10 LPWTT9=4 4 4 0 10 LPPRC10=3 3 3 0 10 LPPCTM10=229.1 187 287.6 28.8 10 187 196.6 217.2 226 225.4 211.5 287.6 256.8 257.5 225 LPWT10=32 32 32 0 10 LPWTT10=32 32 32 0 10 LPPCTM11=1315.6 807.2 2113.8 387.5 10 807.2 963.5 960.8 1231.8 1636.8 2113.8 1262.5 1751.9 1339.6 1087.9 LPPLTM11=35.1 29.8 40.9 3.2 10 29.8 31.4 32.4 34.6 37.5 40.9 35.8 38.6 36.1 33.6 LPPPTM11=0.2 0.2 0.4 0.1 10 0.2 0.2 0.2 0.2 0.3 0.3 0.2 0.2 0.2 0.4 SAMPS DATE=06/06/08 TIME=07:00 DUR=01:00 SAMPS DATE=06/06/08 TIME=08:00 DUR=01:00 ..... SAMPS DATE=06/06/08 TIME=16:00 DUR=01:00 WORKS WDESC=PRIMEAPP.APPPRIME.1 WIO=0 WCS=0 WES=0 WPAGE=0 WTRANSRS=0 WPRTY=50 WMPL=0 WRESP=0 WIORESPV=0.7 0.5 1.2 0.2 10 1.2 0.6 0.6 0.5 0.7 0.5 0.9 0.6 0.8 0.5 WCPUTM=0.1 0 0.1 0 10 0 0 0 0.1 0.1 0 0.1 0.1 0.1 0.1 WIMP=2 2 2 0 10 WORKS WDESC=PRIMEBAT.COMBUILD WIO=255.3 WCS=1.87 WES=0 WPAGE=0 WTRANSRS=0.01 WPRTY=32 WMPL=0.22 WRESP=197.49 MPLV=0.2 0 0.6 0.2 10 0 0 0.1 0.6 0.1 0.1 0.1 0.5 0.3 0.1 TRANV=0.01 0 0.01 0 10 0 0 0.01 0 0.01 0 0 0.01 0.01 0 RESPV=131.81 1.3 755.26 217.32 10 1.8 1.3 755.26 229.86 18.27 28.92 42.86 82.48 95.44 61.91 WIOV=83.3 0.1 212.6 59.3 10 0.2 0.1 67.7 212.6 54.1 86.7 86.4 118.1 130.4 76.4 WIORESPV=1.9 1 3.1 0.7 10 2.5 3.1 1.1 2 1.3 1 1.2 3 1.9 1.4 WCPUTM=65.6 0.2 179.6 57.5 10 0.2 0.2 29.1 179.6 52.8 38.9 57.1 122.8 137.8 37.4 WSRBTM=5.9 0 19.4 6.6 10 0 0 2 19.4 2.1 3.5 3.9 17.3 8.1 3.2 WCSV=1.9 0.1 5.2 1.9 10 0.1 0.1 0.6 5.2 0.7 0.9 1.5 4.1 4.5 1 WEXCPV=255.3 0.1 854.7 262.2 10 0.2 0.1 163 854.7 115.2 206.9 178.6 634.7 275.3 124.5 WDIV=2 0 3.2 1.1 10 0 0 2.9 3.2 1.5 2.3 1.9 2.8 3 2 WDCV=0.1 0 0.2 0.1 10 0 0 0.1 0.1 0.1 0.1 0.1 0.2 0.2 0.1 WPIV=0.4 0 1.5 0.4 10 1.5 0 0.3 0.3 0.3 0.3 0.3 0.3 0.3 0.3 WPITYP=1 1 1 0 10 WPIVAL=25 25 25 0 10 WPIVAT=69.3 0 89.6 30.9 10 16.7 0 81.5 83.2 87.9 86.6 89.6 79.2 82.3 86.4 WIMP=3 3 3 0 10

WDSNS WDSNV=C90PSL WDSNN=*C90PSL WDR=0.1 WDHR=59.6 WDHRT=1.2 WDC=1 WDD=0 WDP=0.2 WDRT=1.4 ..... WDSNS WDSNV=ZF8F1F WDSNN=OMVS.ZFS.SHUT.VAULT.TST.DATA WDR=0.1 WDHR=2.9 WDHRT=7.4 WDC=1.2 WDD=5.8 WDP=0.3 WDRT=7.4 WORKS WDESC=SYSTEM.SYSOTHER WIO=0 WCS=0 WES=0 WPAGE=0 WTRANSRS=0.11 WPRTY=250 WMPL=8.71 WRESP=72.4 MPLV=8.7 6.8 10.8 1.3 10 6.8 6.8 7.5 8.2 9.1 8.9 9.4 9.8 9.8 10.8 TRANV=0.11 0 0.42 0.13 10 0 0.01 0 0.07 0.04 0.07 0.07 0.29 0.42 0.1 RESPV=2339.89 0.02 22692.51 6786.2 10 22692.51 0.7 0.04 0.16 562.58 60.21 12.04 0.08 70.55 0.02 WIOV=1.1 0.1 5.2 1.5 10 0.4 1.3 0.1 0.6 5.2 2.4 0.1 0.3 0.6 0.1 WIORESPV=5.1 0.7 10.8 3.1 10 0.7 5.7 3.8 2.2 6.1 6.8 9 10.8 1.2 4.4 WCPUTM=25.9 10.2 96.1 23.9 10 20.9 23.7 25.2 11.4 16.7 15 10.2 18.7 20.7 96.1 WIIPV=13.1 5.3 50.6 12.8 10 11.4 12.6 13.7 6 8.9 8 5.3 6.9 7 50.6 WIIPCPV=0.2 0 1.3 0.4 10 0 0.1 0.1 0 0.1 0.1 0 0 0 1.3 WPITYP=4 4 4 0 10


CF S HPTSID=MCLXCF01 CFNAME=AQCF2 CFTYPE=2094-S38 CFNCPS=01 FLVL=15 CFSTOR=2983 CFDMPSTOR=6 CFSTORAV=2515.5 FPAM=2 FSCL=14 FTAP=CBP CBP CFREQ=2960.54 FSCCV=0 0 0 0 10 CFLINKS=14 CFUT=11.5 3 21.7 7 10 5.1 3.5 3 12.3 15.9 21.7 8.4 21.5 17.9 5.7 EFFCP=1 1 1 0 10 CFREQV=2960.5 708.9 6104.7 2010.4 10 1144.8 831.2 708.9 2793.3 4111.4 6104.7 1912.7 6000 4722.4 1276 PEER=AQCF1 PRTYPE=2097-E56 PRSR=19F30 CFS S STRNAME=DSNDB2L_GBP0 STRTYPE=CACH STRSIZE=11 FLG=0 CFE S SYS=AQFT AREQV=0.03 0 0.06 0.02 10 0.01 0.01 0.01 0.05 0.06 0.04 0 0.04 0.05 0.01 ASRVV=410.22 151.38 745.25 228.8 10 636.47 745.25 551.41 151.38 153.36 165.6 527.5 158.34 345.8 667.07 SREQV=1.23 0.03 2.27 0.55 10 1.38 0.91 1.74 1.04 2.27 1.34 1.24 1.36 1.01 0.03 SSRVV=11.17 8.82 16.07 2.53 10 9.88 9.94 9.27 10.94 11.02 16.07 8.82 9.25 10.45 16.01 SQV=0 0 0 0 10 SQTMV=0 0 0 0 10 SCNV=0 0 0 0 10 SFCNV=0 0 0 0 10 SSTAV=0 0 0 0 10 SHTOV=0 0 0 0 10 SLTOV=0 0 0 0 10 STRCV=1.26 0.04 2.33 0.56 10 1.39 0.92 1.75 1.09 2.33 1.38 1.24 1.4 1.06 0.04 .... BCU S BCUID=VFV1611 CTYPE=9393-2 CACHE=1024 NVS=8 NOAD=5 BCUSLO=0 BCUSKEW=5 BCUIO=0.02 BCURESP=4.95 BCUCONN=4.55 BCUDISC=0.02 BCUPEND=0.38 BCUQUE=0 BCUIOV=0.02 0.02 0.02 0 10 BCURESPV=4.95 4.72 5.94 0.34 10 4.9 5.94 4.92 4.91 4.73 4.83 4.83 4.72 4.91 4.8 BCUCONNV=4.55 4.55 4.56 0 10 4.55 4.55 4.55 4.55 4.55 4.55 4.55 4.55 4.56 4.55 BCUDISCV=0.02 0.02 0.02 0 10 BCUPENDV=0.38 0.16 1.37 0.34 10 0.33 1.37 0.36 0.34 0.16 0.27 0.26 0.16 0.33 0.23 BCUQUEV=0 0 0 0 10 BCUR=0 BCUH=0 BCUW=0 BCUG=0 BCUC=0 BCUF=0 BCUK=0 BCUD=0 BCUB=0 BCURV=0 0 0 0 10 BCUHV=0 0 0 0 10 BCUWV=0 0 0 0 10 BCUGV=0 0 0 0 10 BCUCV=0 0 0 0 10 BCUFV=0 0 0 0 10 BCUKV=0 0 0 0 10 BCUDV=0 0 0 0 10 BCUBV=0 0 0 0 10 BCUDASD1=33903 BCUDASDN1=5 PATHS PID=P72 PTYPE=E PBUSYV=0 0 0.01 0 10 0 0 0 0 0 0 0 0.01 0 0 PBYV=0 0 0.01 0 10 0 0 0 0 0 0 0 0.01 0 0 PIOV=0.01 0.01 0.01 0 10 PATHS PID=P73 PTYPE=E PBUSYV=0 0 0.01 0 10 0 0 0 0 0.01 0 0 0 0 0 PBYV=0 0 0.01 0 10 0 0 0 0 0.01 0 0 0 0 0 PIOV=0.01 0.01 0.01 0 10 PATHS PID=P74 PTYPE=E PBUSYV=0 0 0.01 0 10 0 0 0 0 0 0 0 0.01 0 0 PBYV=0 0 0.01 0 10 0 0 0 0 0 0 0 0.01 0 0 PIOV=0.01 0.01 0.01 0 10 PATHS PID=P75 PTYPE=E PBUSYV=0 0 0.01 0 10 0 0 0 0 0.01 0 0 0 0.01 0 PBYV=0 0 0.01 0 10 0 0 0 0 0.01 0 0 0 0.01 0 PIOV=0.01 0.01 0.04 0.01 10 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.04 0.01 ACT S V=FV1611 A=1611 T=33903 R=0.02 SDS=0 Q=0 P=0.38 D=0.02 C=4.55 SDR=0.35 SG= DS=0 RWR=0 RDHT=0 FWHT=0 SQSTG=0 PC=0 PFWD=0 PCW=0 CWH=0 ST=N CFWB=0 .... ACT S V=FV163A A=163A T=33903 R=0 SDS=0 Q=0 P=0 D=0 C=0 SDR=0 SG= DS=0 RWR=0 RDHT=0 FWHT=0 SQSTG=0 PC=0 PFWD=0 PCW=0 CWH=0 ST=N CFWB=0 BCUTS BCUID=A6D70 CTYPE= CACHE=0 NVS=0 NOAD=4 AAD=4 BCUIOV=0 0 0 0 10 ALLOCV=12.3 0 68.3 24.7 10 0 68.3 54.3 0 0 0 0 0 0 0 MOUNTV=0.2 0 1 0.4 10 0 1 1 0 0 0 0 0 0 0 DELAYV=42.7 0 273 89.5 10 0 273 154.1 0 0 0 0 0 0 0 CONNV=0.1 0 0.2 0 10 0 0.1 0.1 0.2 0.1 0.1 0 0 0 0.1 DISCV=5.8 0 57.1 17.1 10 0 0.2 57.1 0.2 0.1 0.1 0.1 0 0 0.1 BCUTAPE1=3590 BCUTAPEN1=4


9.3 EDF Sections

The following EDF sections may be produced:

HEAD This contains the description of the account or ENTerprise. A required

section.

CEC This contains the CECID/CPCID and the number of System Images in

the EDF. A required section.

SYS This is the System Image. This section contains the variables

describing the system image and the number of subsections. A

required section.

SAMP The sample section. A brief description of the file. One section for

each sample.

SOFT

WORK Each of these sections describes a workload running on the system

image. One section for each workload. Optional

WK30

WDSN For each workload, this section represents the Data Set data found in

the Type 42 data. For DSNs with a low rate (<1) the data is

accumulated by VOLSER. This will appear as *VOLSER.

CF For each Coupling Facility.

CFS

For each Coupling Facility Structure .

CFE For each CFS with request data for the structure. For each Coupling

Facility Structure there will be an entry for each connected system

CFL Information from the Channel Path Data section for each CHPID

listed in the Local Coupling Facility Data section.

DCM For each DCM channel in a SYS there will be a section. Optional

BCU For each DASD BCU (Controller) there will be a section, one for each

BCU. Optional

PATH The path section will contain data for those paths connected to this BCU, one section for each Path. Optional even if BCU is selected.

ACT The actuator section will contain the data for the actuators running on this controller; one section for each actuator. Optional even if BCU

is selected. Required if path is present.

BCUT For each TAPE BCU (Controller) there will be a section, one for each

BCU. Optional

PTHT The path section will contain data for those paths connected to this BCU, one section for each Path. Optional even if BCU is selected.

BCUO For each other BCU (Controller) there will be a section, one for each

BCU. Optional

PTHO The path section will contain data for those paths connected to this

BCU; one section for each Path. Optional even if BCU is selected.


9.4 EDF Fields

Head Section

ENT Enterprise Name Char (50) User input

SMFDSN SMF Data Set Name Char (44) RDJFCB

SOURCE Program creation data Char CP3KEXTRmm/yy

VER Version number Char n.nn

CEC Section

CECID or

CPCID

CEC Identifier Char User

SUPVR Supervisor Char

CPUMOD CPU Model Char SMF70MOD SMF70VER

VC Version Code Char SMF70VER

PR Number of Processors including

ICFs

Number SMF70BNP. Of last sample.

SR Serial Number of the processor Number SMF70SER, SMF70CSC

MAXPU Number of processor cores

available in htis machine

Number SMF70MAXPU

PRV Number of Processors including

ICFs etc.

Vector SMF70BNP. Appears only if PR=

changed during run.

CPV Number of CP processors Vector Count of CP processors in

PHYSICAL partition

ICFV Number of ICF processors Vector Count of ICF processors in

PHYSICAL partition

IFLV Number of IFL Vector Count of IFL processors in

PHYSICAL partition

ZAAPV Number of zAAPs Vector Count of IFA or AAP processors in

PHYSICAL partition

ZIIPV Number of zIIPs Vector Count of IIP processors in

PHYSICAL partition

CPUMODV Names of the CPCs Multiple

text strings

SMF70MOD SMF70MDL

CMIND Index of current CPC in CPUMODV Vector

HWCMODV Name of the hardware CPC Multiple text strings

SMF70MOD SMF70HWM


HWIND Index of current hardware CPC in

HWCMODV

Vector

SYS Section

SYSID SYSID Char SMF70SID

HPTSID Parallel Sysplex name Char SMF70XNM, only if SMF 74

subtype 4 (CF) records are

present

NSAMPS Number of Samples number

SCP z/OS software level char SMF70MVS

VERSION SCP level char SMF70RLS

SRM SRM Constant number SMF72ADJ

GL Goal or Compat Mode 1/0 1 if 72.3 records are present

WC Wait Complete. 1 if Wait Complete

is set for this partition.

1/0 (SMF70VPF bit SMF70WSA)

BIT 64 bit mode indicator 1 SMF70EME. Appears only if in 64

bit mode.

PAR Peak:Avg. Ratio number

RMFINT same as input parameter number

DTSRC Source of Data RMF or

CMF

SMF70RV2 bit x’0080’ on means

CMF

HDV HiperDispatch Vector 0=off , 1=on SMF70HHA

ABSMSUCP Indicates when absolute MSU

capping is active

Vector SMF70HAM bit in SMF70HHF

RMFINTL Average length of RMF interval in

Minutes

number

GMTOF offset GMT number

(-)hh:mm

SMF70LGO

SNAM System name Char SMF70SNM

CPUWAIT Seconds of CPU Wait Time Vector SMF70WAT

CPUPWAIT Seconds of CPU Wait Time on zIIP

processors

Vector SMF70WAT

CPUZWAIT Seconds of CPU Wait Time on zAAP

processors

Vector SMF70WAT

DASDIO DASD I/O start subchannel rate number SMF74SSC

PAGE System I/O Rate to Page DSNs number SMF75SIO

PAGEDS Number of Page Datasets vector SMF75PSN


CS CS installed MB stats number SMF71TFC + SMF71FIN

ES ES installed MB stats number SMF71OLE

CSAVAIL Amount of available central storage

in MB. stats

number SMF71AVF+SMF71L2A

ESAVAIL Amount of available expanded

storage in MB. stats

number SMF71ASA

SCPCS SCP pages (Nucleus, SQA, LPA,

CSA)

number SMF71FIN + ASR + ALP + AVP

SCPES SCP pages (Nucleus, SQA, LPA,

CSA)

number SMF71ASE + LAE + CAE

IOCPDT Date & time from IOCP file Char

IODFDT Date & time from IOQ data Char R783TDT & R783TTM

CECUTILV Statistics from samples for CPU% for CEC.

Vector SMF70WAT SMF70PDT

PAGEV System I/O Rate to Page DSNs Vector SMF75SIO

DMDPGV Demand paging rate Vector SMF71PIN and SMF71POT

HICOM

High Common Vector SMF71CRA

HISHARE High Shared Vector SMF71SRA

CSV CS installed MB stats Vector SMF71TFC+SMF71FIN

ESV ES installed MB stats Vector SMF71OLE

CSAVAILV Amount of available central storage

in MB. stats

Vector SMF71AVF

ESAVAILV Amount of available expanded

storage in MB. stats

Vector SMF71ASA

NONPG Non-pageable frames Vector SMF71AVX

SCPCSV SCP pages (Nucleus, SQA, LPA,

CSA)

Vector SMF71FIN + ASR + ALP + AVP

SCPESV SCP pages (Nucleus, SQA, LPA,

CSA)

Vector SMF71ASE + LAE + CAE

PGTOES Total pages to ES Vector SMF71PES

PGFROMES Total pages from ES Vector SMF71RES

ESAUX Total pages migrated ES to AUX Vector SMF71PEA

AUIC Average UIC stats Vector SMF71ACA

MUIC Minimum UIC stats Vector SMF71LIC

MXUIC Maximum UIC stats Vector SMF71HIC

PAGEMB Minimum total number of local

page data set slots, in MB

Vector SMF71MNA


SCMTL Minimum total number of 4K SCM

blocks available to ASM, in MB

Vector SMF71TSM

SCMMA Minimum number of available

(not in-use) SCM blocks, in MB

Vector SMF71ASM

FR1RN Number of first reference faults

taken per second

Vector SMF711RN

FRCAM Minimum number of avail CS

Frames (does include 1M frames)

Vector SMF71CAM

FRL7M Minimum number of 1 MB not in

use frames in the LFAREA

Vector SMF71L7M

CSMINV Minimum number of Available

page frames (doesn't include 1M

frames), in MB

Vector SMF71MNF

AMIGR Average migration age stats Vector SMF71AMA

MMIGR Minimum migration age stats Vector SMF71LMA

AINR Average in and ready Vector SMF70RTT

AIN Average in Vector SMF70ITT

AOUTR Average out and ready Vector SMF70OTT

LOGSWAP Logical Swap % Vector SMF71OTT - AXD -ESD

LOGEFF Logical Swap effective % Vector LOGSWAP - LES - LAX

BATCHV Average Batch users Vector SMF70BTT

STCV Average Started Task users Vector SMF70STT

TSOV Average TSO users Vector SMF70TTT

ASCHV Average ASCH users Vector SMF70PTT

OMVSV Average OpenMVS users Vector SMF70XTT

OCPU Percent of the time that the

in/ready count is greater than # of

processors

Vector SMF70Rnn

OCPU2 Percent of the time that the

in/ready count is greater than # of processors plus 1

Vector SMF70Rnn

LREV Logical Ready Vector SMF70LTT

DASDIOV Total DASD I/O stats Vector SMF74SSC

DASDRSPV Total DASD Response Time Vector computed from SMF74SSC +

QUE + CNN + DSC + PEN + MEC

DASDSERV Connect + Disconnect DASD Time Vector computed from SMF74SSC +

CNN + DSC + MEC

TAPEIOV Total TAPE I/O stats Vector SMF74SSC

TAPEALC Max Allocated Tape Devices Vector SMF74MLC


IFANF IFA Normalization Factor (mult this

number times time on IFA divided

by 256 for equivalent time on CP)

Vector R723NFFI

IIPNF IIP Normalization Factor (multiply

this number times time on IIP

divided by 256 for equivalent time on CP)

Vector R723NFFS

WLA Defined Capacity Vector SMF70WLA

LAC Rolling 4 Hour Average Vector SMF70LAC

LPARNO Index our LPAR in the LPAR array (next field)

Number SMF70PTN

LPAR Names of the LPARS Vector SMF70LPM

LPWGT Vector of weights for the partitions.

3 per partition: min, def & max. A

value of -1 for the defined weight

means dedicated processors. (Physical partition omitted.)

Vector SMF70MIS, SMF70BPS,

SMF70MAS

LPCS Vector of the central storage sizes for the LPARS.

Vector SMF70CSF

LPES Vector of the expanded storage sizes for the LPARS.

Vector SMF70ESF

LPCL -- Deleted in v3.82 -- SMF70SPN

LMSUn Defined Capacity Vector SMF70MSU

LPPCTMn Seconds of 'CP' CPU time for LPAR

n.

Vector SMF70PDT

LPPITMn Seconds of 'ICF' CPU time for LPAR

n.

Vector SMF70PDT

LPPZTMn Seconds of “zAAP CPU time for

LPAR n.

Vector SMF70PDT

LPPLTMn Seconds of 'IFL' CPU time for LPAR

n.

Vector SMF70PDT

LPPPTMn Seconds of ‘zIIP” CPU time for LPAR

n.

Vector SMF70PDT

LPPOTMn Seconds of other (?) CPU time for

LPAR n.

Vector SMF70PDT

LPPCSn Average MB of central storage for

LPAR n.

Vector SMF70CSF

LPPRCn Average number 'CP' processors for

LPAR n.

Vector sum of SMF70ONT (or whole

interval if field is not present.)

LPPRIn Average number 'ICF' processors

for LPAR n.



LPPRZn Average number of ‘zAAP’

processors for LPAR n.

Vector

LPPRLn Average number 'IFL' processors for

LPAR n.




LPPRPn Average number of “zIIP”

processors in LPAR n

Vector

LPPROn Average number other(?) processors

for LPAR n.



LPWTn Average GCP LPAR weights for

LPAR n.

Vector SMF70ACS

LPWIn Average ICF LPAR weights for LPAR

n.

Vector SMF70ACS

LPWLn Average IFL LPAR weights for LPAR

n.

Vector SMF70ACS

LPWZn Average ZAAP LPAR weights for

LPAR n

Vector SMF70ACS

LPWPn Average ZIIP LPAR weights for

LPAR n

Vector SMF70ACS

LPWLMCn Vector of percentage of time that

the WLM considers to cap the

LPAR.

Vector SMF70NSW

LPWLMDn Vector of percentage of time that

the WLM actually did cap the LPAR.

Vector SMF70NCA

LPCAPn Vector 1 and 0 corresponding to whether the GCP in LPAR n are

capped. 1=capped (field is absent if

not capped)

Vector . SMF70VPF bit SMF70CAP Vector for each partition (Physical

is omitted). One entry per sample

LPRWTAn Vector of absolute capping value in

hundredths – for GCP

Vector SMF70HW_CAP_Limit

LPRWTAZn Vector of absolute capping value in

hundredths – for zAAP


LPRWTAPn Vector of absolute capping value in

hundredths – for zIIP


LPRWTAIn Vector of absolute capping value in

hundredths – for ICF


LPRWTALn Vector of absolute capping value in

hundredths – for IFL


LPCAPIn Vector of whether ICFs in LPAR are

capped.

Vector SMF70CAP

LPCAPLn Vector of whether IFLs in LPAR are

capped.

Vector SMF70CAP

LPCAPZn Vector of whether zAAPs in LPAR

are capped.

Vector SMF70CAP

LPCAPPn Vector of whether zIIPs in LPAR are

capped.

Vector SMF70CAP

LPGGRPn Hardware Group Capping –

Name list for GCPs

List SMF70HWGr_Name

LPZGRPn Hardware Group Capping –

Name list for zAAPs

List SMF70HWGr_Name

LPPGRPn Hardware Group Capping –

Name list for zIIPs

List SMF70HWGr_Name


LPIGRPn Hardware Group Capping –

Name list for ICFs

List SMF70HWGr_Name

LPLGRPn Hardware Group Capping –

Name list for IFLs

List SMF70HWGr_Name

LPGLn Hardware Group Capping –

Index into name list for GCPs

Vector

LPZLn Hardware Group Capping –

Index into name list for zAAPs

Vector

LPPLn Hardware Group Capping –

Index into name list for zIIPs

Vector

LPILn Hardware Group Capping –

Index into name list for ICFs

Vector

LPLLn Hardware Group Capping –

Index into name list for IFLs

Vector

LPGGVn Hardware Group Capping –

Hardware Cap limit for GCPs

Vector SMF70HWGR_Cap_Limit

LPZGVn Hardware Group Capping –

Hardware Cap limit for zAAPs


LPPGVn Hardware Group Capping –

Hardware Cap limit for zIIPs


LPIGVn Hardware Group Capping –

Hardware Cap limit for ICFs


LPLGVn Hardware Group Capping –

Hardware Cap limit for IFLs


CRYTM Total crypto coprocessor time in

secs

Vector R7023TO*R7023SF

CRYRA Total crypto rate Vector R7023CO

CRYNU Number of coprocessors Vector SMF7023N

CRYTY Coprocessor type Vector R7023CT

LPCRSK Coprocessor utilization skew Vector (R7023TO*R7023SF)/duration

LPGRPN Names of partition groups -- a text

string with multiple LPAR names

separated with a space

Text string SMF70GNM

LPLnn Index of current group name (nn = LPAR #)

Vector computed

LPGVnn Group Value for partition nn Vector SMF70GMU

LPPRKC Total parked GCP time for Dur in seconds

Vector SMF70PAT

LPPRKA Total parked zAAP time for Dur in seconds

Vector SMF70PAT

LPPRKI Total parked zIIP time for Dur in seconds

Vector SMF70PAT

ZEP1 thru ZEP4

Vector SMF70ZEP, ZER, ZEE,ZEC


IFAHP zAAP priority Vector R723MHPO

IIPHP zIIP priority Vector R723MHPZ

SMTEFFP Multithreading capacity numerator

for zIIP

Vector SMF70CFS

SMTEFFMP Multithreading maximum capacity

numerator for zIIP

Vector SMF70MCFS

SMTEFFTDP zIIP Average thread density Vector SMF70ATDS

INSTR Instructions in MIPS Vector SMF 113 B1

CPI Cycles per Instruction Vector SMF 113 B0,B1

FCPI Finite Cycles per Instruction Vector SMF 113

PRBS % problem state Vector SMF 113 B1, P1

PRBT % problem state time Vector SMF 113 P32, B0

GHZ CPU speed in GHz Vector 113_1_CPUSPEED

or 113_2_CPSP

LPARCPU = 1/CPSP/1mil * B0/ IntSec * 100 Vector

L1MP % Level 1 Miss Vector SMF 113 B1,B2,B4

L15P % sourced from L1.5 cache Vector B2, B4, E128, E129

L2LP % sourced from L2 cache Vector B2, B4, E130, E131

L2RP % sourced from L2 remote cache Vector B2, B4, E132, E133

MEMP % sourced from Memory Vector B2, B4, E128, E129, E130, E131, E132, E133, E134, E135

L2P % sourced from level 2 cache Vector E128,E129,B2,B4

L3P % sourced from level 3 cache on

same chip cache

Vector E150,E153,B2,B4

L4LP % sourced from level 4 cache on

same book

Vector E135,E136,B2,B4

L4RP % source from level 4 cache

different book

Vector E138, E139,B2,B4

RNI Relative Nest Intensity Vector Computed.

SCPL1M Est. sourcing cycles/L1 miss Vector

TLBM TLB1 miss % of total CPU Vector (*)

TLBC TLB1 cycles per TLB miss Vector (*)

TLBP PTE % of all TLB1 misses Vector (*)

TLBMG TLB1 miss % of total GCP only Vector (*)

TLBCG TLB1 cycles per TLB miss GCP Vector (*)

TLBPG PTE % of all TLB1 misses GCP Vector (*)

TLBMZ TLB1 miss % of total CPU Vector (*)


TLBCZ TLB1 cycles per TLB miss Vector (*)

TLBPZ PTE % of all TLB1 misses Vector (*)

TLBMP TLB1 miss % of total GCP only Vector (*)

TLBCP TLB1 cycles per TLB miss GCP Vector (*)

TLBPP PTE % of all TLB1 misses GCP Vector (*)

TLBR TLB Miss Rate – All processors Vector (*)

TLBRG TLB Miss Rate - GCP only Vector (*)

TLBRZ TLB Miss Rate - zIIP only Vector (*)

TLBRP TLB Miss Rate - zAAP only Vector (*)

E224 113 counter 224 Vector



CAI Capacity adjustment indicator Vector SMF70CAI

CCR Capacity change reason Vector SMF70CAI

IOPRTY I/O Priority used 1= Yes, 0=No Constant Velocity and R723MOVE

CMM GCP work unit max Vector SMF70CMM

CTT GCP work unit average Vector SMF70CTT/SRM70SRM

DMM zAAP work unit Max. Vector SMF70DMM

DTT zAAP work unit Avg. Vector SMF70DTT/SRM70SRM

EMM zIIP work unit Max. Vector SMF70EMM

ETT zIIP work unit Avg. Vector SMF70ETT/SMF70SRM

PAV 1=PAV used, 2=Hiper Constant SMF74PAV or SMF74HPV

(*) see CPU MF – 2013 Update and WSC Experiences

SAMP Section -- Sample data profile

DATE Start date mm/dd/yy SMF70DAT

TIME Start Time HH:MM SMF70IST

DUR Duration (MINIMUM is five minutes)

HH:MM SMF70INT

SOFT Section -- Software used

http://www-03.ibm.com/support/techdocs/atsmastr.nsf/WebIndex/TC000139


OWN Owner Character SMF89UPO

NAME Work name Character SMF89UPN

VERSION Version Character SMF89UPV

PRODID Product ID Character SMF89UPI

SAMPCNT Sample Count Character Number of samples with this

record

WORK Section -- Workloads

WDESC Description char User parm

WTYPE Workload type char User Parm

WIO "EXCP" Rates num SMF72ITS

WSTOR Storage Avg (xSTOR will be used for

versions prior to RMF421 and xCS

and xES otherwise.

num MB SMF72FT1,2

WCS Central Stor num MB SMF72FT1,2

WES Expanded Stor num MB SMF72ER1,2

WMPL MPL num Calculate

MPLV MPL vector Calculate

WRESP Avg Response time (secs) num SMF72TTM

RESPV Avg Response time (secs) vector SMF72TTM

WTRANSRS Trans Rate/sec num RMF72TTX

TRANV Trans Rate/sec vector(n) SMF72TTX

WPAGE Demand Page Rate num RMF72PIN

WPRTY Workload priority num IEAIPSnn

WHPTSID Sysplex ID char SMF70XNM

WIOV Start SubChannel count vector R723CIRC

WIORESPV DASD I/O Response Time vector SMF72IRC, ICT, IWT, IDT, IOT &

R723Cnnn

WCPUTM Workload CPU Time in Seconds

(Includes CP & IFA time.)

vector SMF72CTS + SMF72STS &

R723Cnnn

WSRBTM Workload SRB Time in Seconds vector SMF72STS

WPAGEV Workload Paging vector SMF72PIN + R723CPIN

WCSV Workload Central Storage vector SMF72FT1,2

WESV Workload Expanded Storage vector SMF72ER1,2

WEXCPV Workload Excp Count vector SMF72ITS + R723CIOC


WCDUR Period duration in weighted service

units

WDCV Workload CPU Delay Vector vector R723CCDE / R723CTSA-

R723CIDL

WDCAP Workload capping delay Vector R723CCCA

WDZIP Workload zIIP Delay Vector Vector R723SUPD

WDZAP Workload zAAP Delay Vector R723IFAD

WDCCP Workload Crypto coprocessor

Delay

Vector R723IFAD

WDIV Workload I/O Delay Vector vector R723CIOD / R723CTSA-R723CIDL

WIFAV Workload zAAP time in seconds Vector R723IFAT

WIFACPV Workload time on CP that could

have executed on zAAP in seconds.

Vector R723IFCT

WIIPV Workload zIIP time in seconds Vector R723CSUP

WIIPCPV Workload time on CP that could have executed on zIIP in seconds.

Vector R723CSUC

WPIV Workload Performance Index (-1 if missing)

Vector Calculate

WPITYP PI type 1=velocity goal 2=percentile response 3=discretionary 4=System

5=Average response

Vector R723CRGF

WPIVAL Goal Value

(based on PI type)

Vector R723CVAL

WCPCT Goal percentage Vector R723CPCT

WPIVAT Goal Attained (Velocity) Vector Calculate

WPROT 1=CPU, 2=STOR, 3=Both

0=Not protected

Vector R723MCPP and R723MSTP

WIMP Workload Importance Vector R723CIMP

WRKR Section -- Reporting Workloads (Same fields as WORK)


WK30 Section -- Workloads when T30MAP is specified

WDESC Description char User parm

WTYPE Workload type char User Parm

WMPLV MPL vector Calculate

RESPV Avg Response time (secs) vector SMF30TRS

TRANV Trans Rate/sec vector SMF72TRS

WPRTY Workload priority num User parm

WIOV Start SubChannel count vector SMF30AIS & SMF30EIS

WIORESPV DASD I/O Response Time vector SMF30AIC, AID & AIW

WCPUTM Workload CPU Utilization vector SMF30CPT, CPS HPT, RCT, IIP, ISB & ICU

WPAGEV Workload Paging vector SMF30PGI, PGO, PSI, PSO, VPI, VPO, CPI, HPI, LPI & HPO

WCSV Workload Central Storage vector SMF30PRV & SYS

WESV Workload Expanded Storage vector SMF70ERS

WEXCPV Workload EXCP Count vector SMF30TEP

WIFAV Workload zAAP time in seconds Vector SMF30_TIME_ON_IFA & ENCLAVE

& DEP_ENCLAVE

WIFACPV Workload time on CP that could

have executed on zAAP in seconds

Vector SMF30_TIME_IFA_ON_CP &

ENCLAVE & DEP_ENCLAVE


WSDN Section -- Workload type 42

DSDD Read only disconnect num S42DSRDD

DSDT Read rate num S42DSRDT (total number of

Read/IOs)

ROD Intensity Num Product of DSDD and DSDT

WDSNV VOLSER Char S42DSVOL

WDSNN Data set name Char S42DSN

WDR I/O Rate num S42DSION

WDHR Hot spot rate num S42DSION

WDHRT Hot spot response time num S42SIOR

WDC Connect time num S42DSIOC

WDD Disconnect time num S42DSIOD

WDP Pend time num S42DSIOP

WDRT Response Time num S42DSIOR


CF Section -- Coupling Facility

HPTSID Parallel Sysplex name Char SMF70XNM, only if SMF 74

subtype 4 (CF) records are present.

CFNAME Coupling Facility name Char R744FNAM

CFTYPE Coupling Facility Device Char User Parm or R744FMOD &

R744FVER

CFSTOR Total Storage num MB R744GTSD

CFDMPSTOR Dump Storage num MB R744GDSA

CFSTORAV Available Storage num MB R744GTSF

CFREQ Coupling Facility requests per second

num 2 dec R744FTOR

FSCCV Subchannel contention vector R744FSCC divided by R744FTOR times 100

FPBCV Request failure count vector R744FPBC

CFLINKS Number of subchannels currently

in use

num R744FSCU

CFUT CPU Utilization of the CF vector R744PBSY

CFPWGT CF Partition weight num 0= record is downlevel,

-1=dedicated,

any other value = weight

EFFCP Effective number of CPUs vector R744PBSY+R744PWAI

CFREQV Coupling Facility requests per

second

vector R744FTOR

CFNCPS Number of processors in the CF num SMF744PN

FLVL Coupling Facility Level num R744FLVL

DYNDISP Dynamic Dispatching num 0: R744FLVL <= 14,

2: R744FLVL >14 and

R744FFLG x’10’ bit is off,

3: R744FLVL >14 and

R744FFLG x’10’ bit is on

FPSNV # shared processors Vector R744FPSN

FPDNV # dedicated processors Vector R744FPDN

CFSYSN Names of systems connected to this

CF

vector R744XSYS

PEER Name of peer CF Char R744RNAM

PRSR Serial of peer CF hex R744RNDE

PRTYPE Type of peer CF char R744RNDE


FPAM Path available num R744FPAM

FSCL Number of subchannels that can be

used (limit)

num R744FSCL

RTAP CHPID acronyms Vector R744RTAP

FTAP CHPID acronyms Vector R744FTAP

SR Seq num of CF Char R744FSEQ

LPN Log part num Num R744FLPN (x’6F’)

FPCM Composite path mask Hex R744FPCM

FIDP List of up to 8 ChpIDs Hex R744FIDP

CFS Section -- Coupling Facility Structure

STRNAME CF Structure Name Char R744SNAM

STRTYPE CF Structure type Char R744STYP

‘LIST’, ‘LOCK’, ‘CACH’

STRSIZE Storage Allocated to this Structure num MB R744SSIZ

FLG Duplex primary or secondary num R744QFLG, 0=not duplex,

1= primary, 2= secondary.

CFL Section -- Channel Path Structure Entry

Information from the Channel Path Data section for each CHPID listed in the Local Coupling

Facility Data section.

HCPI Channel path Identifier Hex (2) R744HCPI

HFLG Validity bit mask Hex (4) R744HFLA

HOPM Channel path operation mode Hex (2) R744HOPM

HPCP Physical channel ID (PCHID) Hex (4) R744HPCP

HAID Host channel adapter ID Hex (4) R744HAID

HAPN Host channel adapter port nbr. Hex (2) R744HAPN

HCHF Status flags Vector R744HCHF

HLAT Channel path latency time Vector R744HLAT


CFE Section -- Coupling Facility Structure Entry

For each Coupling Facility Structure there will be an entry for each connected system.

SYS SYSID of connected system Char SMF74SID

AREQV Asynchronous Requests per second vector R744SARC

ASRVV Asynchronous Requests average service time in microseconds

vector R744SATM divided by R744SARC

SREQV Synchronous Requests per second vector R744SSRC

SSRVV Synchronous Requests average service

time in microseconds

vector R744SSTM / R744SSRC

SQV Number of requests queued per second vector R744SQRC

SQTMV Average time queued in microseconds vector R744SQTM / R744SQRC

SCNV Lock contention vector R744SCN / Interval Sec

SFCNV False lock contention vector R744SFCN / Interval Sec

SSTAV requests changed from SYNC to ASYNC vector R744SSTA / Interval Sec

SHTOV requests waiting on high priority queue vector R744SHTO / Interval Sec

SLTOV requests waiting on low priority queue vector R744SLTO / Interval Sec

STRCV total requests to this structure from this system

vector R744STRC / Interval Sec

SETMV Structure execution time metric vector R744SETM / Interval Sec

SWDRV Number of requests to write data to the

CF structure

vector R744SWDR

SWACV Number of adjunct areas written to the

CF structure

vector R744SWAC

SRDRV Number of requests to read data from

the CF structure

vector R744SRDR

SRACV Number of adjunct areas read from the

CF structure

vector R744SRAC

SWECV Number of data entries with data

elements that have been written to the

CF structure. Includes both single and multi entry write requests

vector R744SWEC

SRECV Number of data entries with data elements that have been read from the

CF structure. Includes both single and

multi entry read requests

vector R744SREC

SWEDV Sum of 256 byte increments

accumulated for entry data with data

elements written to the CF structure

vector R744SWED

SREDV Sum of 256 byte increments

accumulated for entry data with data

elements read from the CF structure

vector R744SRED


SENCV Set to 1 if R744SFLG x'02' bit is set

indicating that structure data is

encrypted

vector value =0 or 1

DCM Section -- DCM channel for DASD

For each DCM channel there will be a complete section.

PID Path ID char SMF78CPID

PTYPE Channel Type char SMF73ACR C'P' parallel, C'E' ESCON,

C'V' or C'F' for FICON

PBUSYV Path Busy vector SMF73BSY

PBYV Path Busy for this partition vector SMF73PBY / SMF73PTI

BCU Section -- Basic Configurable Unit for DASD

For each SYSID BCU there will be a complete description

BCUID BCU identification char User Parm

CTYPE CU type char User Parm or SMF74CU

CACHE Cache Size num User Parm or CSCONF in CRR record

NVS Non Volatile Storage size num User Parm or CNCONF in CRR record

NOAD number of addresses under this

BCU

num From BCU map (May include offline

volumes for which there will be no ACT

section.)

BCUDASDi type of DASD char User input or SMF74DEV

BCUDASDNi number of this type num computed

BCUIO Total I/O Rate num SMF74SSC

BCURESP Average Response Time num computed

BCUSLO Service level objective num Maximum response time for devices with

an I/O rate > 1

BCUCONN Average connect time for BCU num SMF74CNN

BCUDISC Average disconnect time for BCU num SMF74DIS

BCUPEND Average disconnect time for PEND num SMF74PEN

BCUQUE Average IOS queue time num SMF74QUE

BCUSKEW Maximum device busy to average. num computed

BCUIOV Total I/O Rate Vector SMF74SSC


BCURESPV Average Response Time Vector computed

BCUCONNV Average connect time for BCU Vector SMF74CNN

BCUDISCV Average disconnect time for BCU Vector SMF74DIS

BCUPENDV Average pend time for BCU Vector SMF74PEN

BCUQUEV Average IOS queue time Vector SMF74QUE

BCURDRT Read Rate Vector (R745DRCR +R745DRSR +R745DRNR)

/ Interval Seconds

BCUWRRT Write Rate Vector (R745DWRC +R745DWSR R745DWNR)

/ Interval Seconds

BCURDBR Read bytes per request Vector R7451CT1 *128 *1024 /

(R745DRCR +R745DRSR R745DRNR)

BCUWRBR Write bytes per request Vector (R7451CT2 *128 *1024 / (R745DWRC +R745DWSR +R745DWNR)

BCUR Read Write Ratio for entire BCU. (All devices cached or not.)

num CRR: "Total (Cache) R/W Ratio"

BCUH Read Hit Ratio for entire BCU. (All devices cached or not.)

num CRR: "Total Read H/R"

BCUW Fast Write Hit Ratio for entire BCU. (All devices cached or not.)

num CRR: "Total F/W H/R"

BCUG Sequential Stage Ratio for entire BCU. (All devices cached or not.)

num CRR: "DASD to Cache Transfers - Sequential" divided by "Total I/O

Requests"

BCUC Percent of I/Os eligible to be cached

for entire BCU.

num CRR: "Total Cacheable I/Os" divided by

"Total I/O Requests" times 100

BCUF Percent Fast Write Destages for

entire BCU. (All devices cached or

not.)

num CRR: "Cache to DASD Transfers - Total"

divided by "Total I/O" times 100

BCUK Percent Count Key Data Writes for


not.)

num CRR: "Count Key Data Writes" divided by

"Total I/O" times 100 (Not supported

with CRR 1.4)

BCUD Count Key Data Write Hit Ratio for


not.)

num CRR: "Count Key Data Write Hits"

divided by "Count Key Data Writes" (Not

supported with CRR 1.4)

BCUB Cache Fast Write Bypass num CRR: “Cache Fast Write Bypass” divided by “Total I/O” times 100

BCURV Read Write Ratio for entire BCU. (All devices cached or not.)

Vector CRR: "Total (Cache) R/W Ratio"

BCUHV Read Hit Ratio for entire BCU. (All devices cached or not.)

Vector CRR: "Total Read H/R"

BCUWV Fast Write Hit Ratio for entire BCU. (All devices cached or not.)

Vector CRR: "Total F/W H/R"

BCUGV Sequential Stage Ratio for entire BCU. (All devices cached or not.)

Vector CRR: "DASD to Cache Transfers - Sequential" divided by "Total I/O

Requests"

BCUCV Percent of I/Os eligible to be cached

for entire BCU.

Vector CRR: "Total Cacheable I/Os" divided by

"Total I/O Requests" times 100


BCUFV Percent Fast Write Destages for


not.)

Vector CRR: "Cache to DASD Transfers - Total"


BCUKV Percent Count Key Data Writes for


not.)

Vector CRR: "Count Key Data Writes" divided by

"Total I/O" times 100.

BCUDV Count Key Data Write Hit Ratio for

entire BCU. (All devices cached or not.)

Vector CRR: "Count Key Data Write Hits"

divided by "Count Key Data Writes"

BCUBV Cache Fast Write Bypass Vector CRR: “Cache Fast Write Bypass” divided by “Total I/O” times 100

MFR CU Manufacturer Char SMF74DCT

BCUSER Serial number from the NED data num SMF74DCT

ERBi ECKD read in KB/sec Vector R748LERB with i as index= R748LTYP

EWBi ECKD write in KB/sec Vector R748LEWB

EROi ECKD read ops/sec Vector R748LERO

ERTi ECKD Read time in MS/sec Vector R748LERT

EWTi ECKD Write time/sec Vector R748LEWT

PSBi Send bytes/sec Vector R748LPSB

PRBi Receive bytes/sec Vector R748LPRB

PSOi Send ops/sec Vector R748LPSO

PROi Receive ops/sec Vector R748LPRO

PSTi Send time in MS/sec Vector R748LPST

PSBi Send bytes/sec Vector R748LPSB

SRBi SCSI Read in KB/sec Vector R748LSRB

SWBi SCSI Write in KB/sec Vector R748LSWB

SROi SCSI Read Ops/sec Vector R748LSRO

SWOi SCSI Write Ops/sec Vector R748LSWO

SRTi SCSI Read Time in Ms/sec Vector R748LSRT

SWTi SCSI Write Time in Ms/sec Vector R748LSWT


PATH Section -- BCU Path Data

FOE Ficon operations Vector SMF73EOS

HOE Ficon transports Vector SMF73ETS

PID Path ID (if P* then DCM

and only the PIOV will be present)

char chpid SMF78CPID

PTYPE Channel Type Char SMF73ACR C'P' parallel, C'E' ESCON, C'V' or C'F' for FICON

PSW Escon Director Number From the BCUMAP(IOCP input)

PLK Link on Escon Director Number From the BCUMAP(IOCP input)

PCHID Physical channel ID Number R744HPCP

OFFL Offline num

(0 or 1)

If no RMF data for path, assumed

offline

CL FICON Chpid Level 1.0,

1.5, or

1.75

9672 is 1.0, 2064 is 1.5 or 1.75

(SMF73MBC > 40,000 is 1.75), 2066

& later 1.75

GEN Type Generation Number SMF73GEN

PBUSYV Path Busy vector SMF73BSY

PBYV Path Busy for this partition vector SMF73PBY / SMF73PTI

PIOV Chpid taken vector R783PT

PBBY Ficon Bus Busy vector SMF73TBC and SMF73MBC

PTRD Ficon Total read MB/Sec vector SMF73TRU and SMF73US

PTWR Ficon Total write MB/Sec vector SMF73TWU and SMF73US

PLRD Ficon LPAR read MB/Sec vector SMF73PRU and SMF73US

PLWR Ficon LPAR write MB/Sec vector SMF73PWU and SMF73US

FRATE Ficon command mode ops vector SMF73EOC

HRATE Ficon transport mode ops Vector SMF73ETC


ACT Section -- Actuators

Detail needed for performance analysis

V Volser char SMF74SER

A Address char SMF74NUM

T DASD Type char User Parm or SMF74DEV

R I/O Rate num SMF74SSC

SDS Standard deviation for service

num calc

Q IOSQ num calc

P PEND num SMF74PEN

D DISC num SMF74DIS

C CONN num SMF74CNN

SG Storage Group char SMF74SGN

DS Total Number of Allocations

in Effect for the Device

num SMF74NDA

SDR Standard deviation for

response

num calc

RWR Read Write Ratio num CRR: "Total (Cache) R/W Ratio"

RDHT Read Hit Ratio num CRR: "Total Read H/R"

FWHT Fast Write Hit Ratio num CRR: "Total F/W H/R"

SQSTG Sequential Stage Ratio num CRR: "DASD to Cache Transfers - Sequential" divided by "Total I/O

Requests"

PC Percent Cached num CRR: "Total Cachable I/Os" divided

by "Total I/O Requests" times 100

PFWD Percent Fast Write Destages num CRR: "Cache to DASD Transfers -

Total" divided by "Total I/O" times

100

PCW Percent Count Key Data

Writes

num CRR: "Count Key Data Writes"


CWH Count Key Data Write Hit

Ratio

num CRR: "Count Key Data Write Hits"

divided by "Count Key Data Writes"

CFWB Cache Fast Write Bypass num CRR: “Cache Fast Write Bypass”

divided by “Total I/O” times 100


ST Actuator Status char CRR: "Device Status":

N - Caching Activated,

DASD FW Allowed,

D - Caching Deactivated,

DASD FW Deactivated, C - Caching Activated,

DASD FW Deactivated,

F - Caching Deactivated,

DASD FW Allowed.

BCUT Section -- Basic Configurable Unit for Tape



CACHE Cache Size num User Parm or 0

NVS Non Volatile Storage size num User Parm or CNCONF in CRR

record

AAD Number of active addresses num Number of addresses on this BCU

with RMF data

NOAD number of addresses under

this BCU

num From BCU map (May include offline

volumes that have no ACT section).

BCUTAPEi type of Tape Device char User input or SMF74DEV

BCUTAPENi number of this type num computed

BCUIOV I/O Rate this BCU Vector SMF74SSC

ALLOCV Average Allocation Time in seconds

Vector computed from SMF74ALC + SAM

MOUNTV Number of Mounts for this BCU

Vector computed from SMF74MTC

DELAYV Average Wait for mount in seconds

Vector computed from SMF74MTP + SAM + MCT

CONNV Total Connect time for this

BCU in seconds

Vector computed from SMF74CNN

DISCV Total Disconnect time for

this BCU in seconds

Vector computed from SMF74DIS

PTHT Section -- See PATH above


BCUO Section -- Basic Configurable Unit for Other

BCUO Basic Configurable Unit for Other



CACHE Cache Size num User Parm or 0

NVS Non Volatile Storage size num User Parm or CNCONF in CRR

record

NOAD number of addresses under

this BCU

num From BCU map (May include

offline volumes that have no ACT

section).

PTHO See PATH above

PTHO Section -- See PATH above

data extraction program (cp3kextr) for zcp3000 · ... program – the rmf interval and the extract...

Documents