data extraction program (cp3kextr) for zcp3000 · ... program – the rmf interval and the extract...
TRANSCRIPT
Data Extraction Program
CP3KEXTR v3.82 07/02/18
For zBNA, zCP3000, and zPCR
Tom Harper
Valerie Spencer
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
Copyright ©2000-2018, IBM Corporation Data Extraction Program 3
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
Copyright ©2000-2018, IBM Corporation Data Extraction Program 4
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
Copyright ©2000-2018, IBM Corporation Data Extraction Program 5
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.
Copyright ©2000-2018, IBM Corporation Data Extraction Program 6
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.
Copyright ©2000-2018, IBM Corporation Data Extraction Program 7
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).
Copyright ©2000-2018, IBM Corporation Data Extraction Program 8
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.
Copyright ©2000-2018, IBM Corporation Data Extraction Program 9
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:
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 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
Copyright ©2000-2018, IBM Corporation Data Extraction Program 10
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.
Copyright ©2000-2018, IBM Corporation Data Extraction Program 11
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
Copyright ©2000-2018, IBM Corporation Data Extraction Program 12
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.
Copyright ©2000-2018, IBM Corporation Data Extraction Program 13
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.
Copyright ©2000-2018, IBM Corporation Data Extraction Program 14
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) /*
Copyright ©2000-2018, IBM Corporation Data Extraction Program 15
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.
Copyright ©2000-2018, IBM Corporation Data Extraction Program 16
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.
Copyright ©2000-2018, IBM Corporation Data Extraction Program 17
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.
Copyright ©2000-2018, IBM Corporation Data Extraction Program 18
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.
Copyright ©2000-2018, IBM Corporation Data Extraction Program 19
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.
Copyright ©2000-2018, IBM Corporation Data Extraction Program 20
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.
Copyright ©2000-2018, IBM Corporation Data Extraction Program 21
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. <>
Copyright ©2000-2018, IBM Corporation Data Extraction Program 22
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
Copyright ©2000-2018, IBM Corporation Data Extraction Program 23
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
Copyright ©2000-2018, IBM Corporation Data Extraction Program 24
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.
Copyright ©2000-2018, IBM Corporation Data Extraction Program 25
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
Copyright ©2000-2018, IBM Corporation Data Extraction Program 26
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
Copyright ©2000-2018, IBM Corporation Data Extraction Program 27
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
4 Messages of note have been issued.
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.
20 CP3KEXTR couldn't write to the PRINT001 DD.
225 No data was found that meets the input specification.
Copyright ©2000-2018, IBM Corporation Data Extraction Program 28
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)
Copyright ©2000-2018, IBM Corporation Data Extraction Program 29
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=*
Copyright ©2000-2018, IBM Corporation Data Extraction Program 30
//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)
Copyright ©2000-2018, IBM Corporation Data Extraction Program 31
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
Copyright ©2000-2018, IBM Corporation Data Extraction Program 32
//* 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
Copyright ©2000-2018, IBM Corporation Data Extraction Program 33
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) /*
Copyright ©2000-2018, IBM Corporation Data Extraction Program 34
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
Copyright ©2000-2018, IBM Corporation Data Extraction Program 35
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
Copyright ©2000-2018, IBM Corporation Data Extraction Program 36
//*-------------------- # 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
Input parameters for
the second subrun
Input parameters for
the third subrun
Copyright ©2000-2018, IBM Corporation Data Extraction Program 37
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
Copyright ©2000-2018, IBM Corporation Data Extraction Program 38
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
Copyright ©2000-2018, IBM Corporation Data Extraction Program 39
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
Copyright ©2000-2018, IBM Corporation Data Extraction Program 40
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.
Copyright ©2000-2018, IBM Corporation Data Extraction Program 41
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
Copyright ©2000-2018, IBM Corporation Data Extraction Program 42
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)
Copyright ©2000-2018, IBM Corporation Data Extraction Program 43
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
Copyright ©2000-2018, IBM Corporation Data Extraction Program 44
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.
Copyright ©2000-2018, IBM Corporation Data Extraction Program 45
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
Copyright ©2000-2018, IBM Corporation Data Extraction Program 46
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
Copyright ©2000-2018, IBM Corporation Data Extraction Program 47
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
Copyright ©2000-2018, IBM Corporation Data Extraction Program 48
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".
Copyright ©2000-2018, IBM Corporation Data Extraction Program 49
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=*
Copyright ©2000-2018, IBM Corporation Data Extraction Program 50
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.
Copyright ©2000-2018, IBM Corporation Data Extraction Program 51
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)
Copyright ©2000-2018, IBM Corporation Data Extraction Program 52
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
Copyright ©2000-2018, IBM Corporation Data Extraction Program 53
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)
Copyright ©2000-2018, IBM Corporation Data Extraction Program 54
42 Data File (continued)
Field Name SMF Field Contains Size Type
S42JDRSD S42JDRSD Date reader recognized the
JOB card for this job
6 YYMMDD
S42JDRST S42JDRST Time reader recognized the
JOB card for this job
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
Copyright ©2000-2018, IBM Corporation Data Extraction Program 55
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
Copyright ©2000-2018, IBM Corporation Data Extraction Program 56
7.3 Data File – Type 14 & 15 Records
Field Name SMF Field Contains Size Type
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
JOB card for this job
6 cyyddd
SMF14RST SMF14RST Time reader recognized the
JOB card for this job
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)
Copyright ©2000-2018, IBM Corporation Data Extraction Program 57
14 & 15 Data File (continued)
Field Name SMF Field Contains Size Type
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
Copyright ©2000-2018, IBM Corporation Data Extraction Program 58
8. Errors
The Extract Program can return different return codes:
0 No errors or warnings
4 Messages of note have been issued.
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.
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.
Copyright ©2000-2018, IBM Corporation Data Extraction Program 59
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.
Copyright ©2000-2018, IBM Corporation Data Extraction Program 60
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
Copyright ©2000-2018, IBM Corporation Data Extraction Program 61
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
Copyright ©2000-2018, IBM Corporation Data Extraction Program 62
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
Copyright ©2000-2018, IBM Corporation Data Extraction Program 63
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.
Copyright ©2000-2018, IBM Corporation Data Extraction Program 64
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
Copyright ©2000-2018, IBM Corporation Data Extraction Program 65
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
Copyright ©2000-2018, IBM Corporation Data Extraction Program 66
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
Copyright ©2000-2018, IBM Corporation Data Extraction Program 67
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
Copyright ©2000-2018, IBM Corporation Data Extraction Program 68
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.
Vector sum of SMF70ONT (or whole
interval if field is not present.)
LPPRZn Average number of ‘zAAP’
processors for LPAR n.
Vector
LPPRLn Average number 'IFL' processors for
LPAR n.
Vector sum of SMF70ONT (or whole
interval if field is not present.)
Copyright ©2000-2018, IBM Corporation Data Extraction Program 69
LPPRPn Average number of “zIIP”
processors in LPAR n
Vector
LPPROn Average number other(?) processors
for LPAR n.
Vector sum of SMF70ONT (or whole
interval if field is not present.)
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
Vector SMF70HW_CAP_Limit
LPRWTAPn Vector of absolute capping value in
hundredths – for zIIP
Vector SMF70HW_CAP_Limit
LPRWTAIn Vector of absolute capping value in
hundredths – for ICF
Vector SMF70HW_CAP_Limit
LPRWTALn Vector of absolute capping value in
hundredths – for IFL
Vector SMF70HW_CAP_Limit
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
Copyright ©2000-2018, IBM Corporation Data Extraction Program 70
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
Vector SMF70HWGR_Cap_Limit
LPPGVn Hardware Group Capping –
Hardware Cap limit for zIIPs
Vector SMF70HWGR_Cap_Limit
LPIGVn Hardware Group Capping –
Hardware Cap limit for ICFs
Vector SMF70HWGR_Cap_Limit
LPLGVn Hardware Group Capping –
Hardware Cap limit for IFLs
Vector SMF70HWGR_Cap_Limit
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
Copyright ©2000-2018, IBM Corporation Data Extraction Program 71
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 (*)
Copyright ©2000-2018, IBM Corporation Data Extraction Program 72
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
E225 113 counter 225 Vector
E226 113 counter 226 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
Copyright ©2000-2018, IBM Corporation Data Extraction Program 73
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
Copyright ©2000-2018, IBM Corporation Data Extraction Program 74
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)
Copyright ©2000-2018, IBM Corporation Data Extraction Program 75
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
Copyright ©2000-2018, IBM Corporation Data Extraction Program 76
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
Copyright ©2000-2018, IBM Corporation Data Extraction Program 77
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
Copyright ©2000-2018, IBM Corporation Data Extraction Program 78
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
Copyright ©2000-2018, IBM Corporation Data Extraction Program 79
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
Copyright ©2000-2018, IBM Corporation Data Extraction Program 80
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
Copyright ©2000-2018, IBM Corporation Data Extraction Program 81
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
entire BCU. (All devices cached or
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
entire BCU. (All devices cached or
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
Copyright ©2000-2018, IBM Corporation Data Extraction Program 82
BCUFV Percent Fast Write Destages for
entire BCU. (All devices cached or
not.)
Vector CRR: "Cache to DASD Transfers - Total"
divided by "Total I/O" times 100
BCUKV Percent Count Key Data Writes for
entire BCU. (All devices cached or
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
Copyright ©2000-2018, IBM Corporation Data Extraction Program 83
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
Copyright ©2000-2018, IBM Corporation Data Extraction Program 84
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"
divided by "Total I/O" times 100
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
Copyright ©2000-2018, IBM Corporation Data Extraction Program 85
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
BCUID BCU identification char User Parm
CTYPE CU type char User Parm or SMF74CU
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
Copyright ©2000-2018, IBM Corporation Data Extraction Program 86
BCUO Section -- Basic Configurable Unit for Other
BCUO Basic Configurable Unit for Other
BCUID BCU identification char User Parm
CTYPE CU type char User Parm or SMF74CU
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