* copyright (C) 1984-2019 merrill consultants dallas texas usa

WPS. And since I needed to find any other exposures, I

removed a number of archaic tests for &SASVER 5 and 6 in

these members:

BUIL3001 BUILD001 BUILDPD3 BUILDPDB DOCPDB FORMATS

IMACPDBX JCLUXRE6 MONTHBL3 MONTHBLD MONTHBLS MONTHWEK

PRODTEST PRODTESW TYPSIMS TYPSIMS7 VFMT102 VMAC120

VMAC7072 VMACIMS VMXGINIT WEEKBL3D WEEKBL3T WEEKBLDD

WEEKBLDT

Thanks to Erling Andersen, SMT, DENMARK.
Change 33.076 Support for IMS56FA Chained/Switched INPQUETM correction.

TYPEIMST INPQUETM for chained transactions (have same ARRVTIME) is

TYPSIMST revised to use the prior transaction's HELDENDTIME as the

VMACIMS the "ARRVTIME" for the next transaction. Variable SWITCH

Mar 24, 2015 identifies if this was the first, or subsequent, switched

transaction, and HELDENDTIME is kept in IMS56FA. Variable

RESPNSTM is also recalculated with the new INPQUETM.

The correction is made in the _SIMS56G dataset sort macro

so it requires a full sort of the IMS56FA dataset; the

TYPSIMST member has the JCL and SYSIN example to use to

correct INPQUETM and create the IMS56FA.IMS56FA dataset.

The TYPEIMST member creates WORK.IMS56FA, unsorted and

uncorrected.

Or, to create ALL of the IMS log datasets and correct

the INPQUETM, you can use

%INCLUDE SOURCLIB(TYPEIMS7);

which invokes the _SIMS46X sort macro to sort both the

46x and 56x IMS log records and fix IMS56FA's INPQUETM,

writing all IMS datasets to //WORK, but you can use

%LET PIMSxxx=YOURDD; to send individual datasets to your

chosen DDNAME.

Thanks to Thomas Peiper, TIETO, FINLAND.

Change 33.075 When VMXGDBAA was run, the KEEPIN= logic in VMXGSUM could

VMXGSUM fail, resulting in UNINITIALIZED VARIABLEs notes for

VMXGDBAA QWACESC and other variables. Exposed when removing the

Mar 22, 2015 AUTOCALL macros from VMXGDBAA, was not actually reported.

Change 33.074 Support for Mainview for MQ version 5.2 BBMQQUES 'E6'

VMACBBMQ with these new variables added:

Mar 22, 2015 QSAC6NXG ='NEXT*AC6 RECORD*WITH THIS*HASH'

QSIBGETC ='CONSUMED*MQGET*BYTES'

QSIBGTCF ='GET*BYTES*CONSUMED*RATE'

QSIFPUT ='FAST*PUT*COUNT'

QSILMPUT1='LARGEST*MQPUT1*SESSION'

QSISMPUT1='SMALLEST*MQPUT1*INTERVAL'

QSISTCKB ='INTERVAL*START*DATETIME*STAMP'

QSQUDPTH ='UNCOMMITTED*MESSAGES'

QSSBGETC ='CONSUMED*MQGET*BYTES'

QSSBGTCF ='GET*BYTES*CONSUMED*RATE'

QSSBGTCR ='GET*BYTES*CONSUMED*RATE'

QSSFPUT ='FAST*PUT*COUNT'

QSSFPUT1 ='FAST*PUT1*COUNT'

QSSLMPUT1='LARGEST*MQPUT1*SESSION'

QSSSMPUT1='SMALLEST*MQPUT1*SESSION'

QSSSTCKB ='SESSION*START*DATETIME*STAMP'

QWIFPUT1 ='FAST*PUT1*COUNT'

and these old variables are no longer input and will be

missing values in BBMQQUES:

QSIFBGTF QSIFBGTR QSIFBP1F QSIFBP1R QSIFBPT1 QSIFBPTF

QSIFBPTR QSIFBPUT QSINDSCR QSSFBGTF QSSFBP1F QSSFBPT1

QSSFBPTF QSSFBPTR QSSFBPUT

Thanks to Paul Volpi, UHC, USA.
Change 33.073 New variable FCBYTERATE added to TYP11903 and TYP11907

VMAC119 datasets.

Mar 22, 2015
Change 33.072 SMF 103 Documentation. For the Apache Server (which must

VMAC103 be used in z/OS 2.2 where the prior Domino Server is not

Mar 22, 2015 supported), the JOB name can be passed into the D record

with a variable - export JOBNAMEH="WEBSRV02" - in the

shell scripts that start the Apache Web Server and using

this JOBNAMEH variable as ServerName ${JOBNAMEH} in the

httpd.conf file.

The SMF 103 record has been written by the same product

since its inception, but the IBM product names have

included these:

"HTTP Web Sphere Server"

Lotus Domino GO Webserver R4

Internet Connection Secure Server (R3,R2)

and now, in 2015, Apache Server.

Thanks to Joseph J. Faska, DTCC, USA.
Change 33.071 z13 SUPPORT. MXG 33.03 REQUIRED ONLY FOR PROCVIEW CORE.

VMAC7072 For PROCVIEW CPU non-SMT, NO CHANGES WERE MADE TO RMF 70.

Mar 21, 2015

For SMT PROCVIEW CORE Mode, MXG Change 33.046 in 33.02

updated the TYPE70 dataset, but this Change 33.071 in MXG

33.03 is required to update the new SMT metrics correctly

in the TYPE70PR dataset, to get the CPUID, LCPUADDR, and

CORE_ID from the four segments that don't have the same

number of segments: OFFCPUD and SMF70COS have 20 for the

6 online CPs, 4 offline CPs, and 5 zIIPs with CPU_NUM=2,

while SMF70BDN/LPARCPUX has only 18 segments (with the

CORE_NUM needed to look-up the LCPUADDR), and there are

only 14 Core_ID values.
This was a complex update to a CRITICAL MXG MEMBER, with

500+ lines of code inserted lines into the 27,000 lines.

The SMT Mode data has been validated with RMF records,

with a wide range of LPAR configurations. When in SMT

mode, please examine the new data carefully and contact

[e-mail] if you have questions.

Note: If you read the changed SMT mode RMF 70s with an

old MXG, RMFINTRV may have NEGATIVE CPUOVHTM values and

the %PCTCPUBY values may be over 100%.
Specifically, if z/OS is IPL'd with LOADxx PROCVIEW CORE, on

a processor that is SMT capable, then whether or not MT is

Active, and even if MT=1 is specified, then the RMF 70 record

is restructured with CORE_ID.

This support was validated with z/OS 2.1 and z13 data.

Change 33.070 New INCLAI, three $LIST_L and $RESTAR TOKENIDs, added by

VMACTPMX ThruPut Manager AE+ V7R1.0 create five new variables

Mar 20, 2015 INCLA1 JBL54043 JBL54L44 JBL54L45 RESTAR

Thanks to Scott Barry, SBBWorks, Inc., USA.
Change 33.069 SAS 9.4 has a dictionary entry for DESTINATIONS but

ANALDB2R older releases do not. On zOS, report MXGDB2B1 used this

Mar 20, 2015 to see if the active destination for ODS was listing and

to bypass the graphics code in that report. The PROC SQL

is now only used when we find SAS 9.4 or greater.

-Spurious message were produced indicating reports were

requested when you specified report=NO because the code

was checking the length and not the value.

-If the DB2STATB dataset was empty or non-existent the

macro variable GOTRMF did not exist and raised an

unresolved MACRO variable error.

Thanks to Randy Hewitt, HP, USA.

Change 33.068 Change 30.133 typo caused Q8STCCPU to not be kept.

VMACDB2

Mar 20, 2015

Thanks to Tim King, Blue Cross Blue Shield of South Carolina, USA.

Change 33.067 DB2 Trace ID=102 IFCID=220 misaligned, causing ILLEGAL

VMAC102 ARGUMENT TO MDY function. The QW0220xx variables are

Mar 17, 2015 now kept.

Thanks to Randy Hewitt, HP, USA.

Change 33.066 VMXGPPDS failed when used as documented, unable to find

VMXGPPDS MYPDS because the last few lines of the macro incorrectly

Mar 17, 2015 tried to execute the macro with the defaults.
Change 33.065 If %VGETLIBS found no LIBNAMEs were allocated, it failed

VGETLIBS with MACRO variables not found. Now if there are none

Mar 17, 2015 found, it dies a graceful death and tells you that it

found no matching LIBNAMES.

Change 33.064 Format $MGSTCRT new values for SMF19RTM/RTE variables:

FORMATS '07'X='07X:VTV MOVE'

Mar 16, 2015 '08'X='08X:ALTERNATE RECALL'

Thanks to Mike Jacques, BB&T, USA.

Change 33.063 In parsing a quoted literal such as '15MAR2015'D, to make

VMXGPARS the text more readable, the D could be separated from the

Mar 26, 2015 Quote, resulting in a 180 Syntax Error. VMXGPARS now

recognizes a DT following a quote and/or any non-blank

character and extends the string.
Change 33.062 -In Change 31.209, VMXGSUM's %TRIM AUTOCALLed function was

VMXGSUM replaced by the internal %SYSFUNC function, to circumvent

VMXGINIT errors when sites don't have AUTOCALL/MAUTOSOURCE/etc set

Mar 16, 2015 up correctly, but %TRIM crept back in VMXGSUM and is now

VMXGPARS again replaced with %SYSFUNC.

-MXGDEBUG=FULL option executes PROC OPTIONS VALUE, to show

each option's value and how it was set, PROC PRINTs the

SASHELP tables VLIBNAM to show all LIBNAMES, and VEXFL to

show all External FILENAMES - (is there a SASAUTOS??),

and enables these full source code diagnostic options

OPTIONS SOURCE SOURCE2 MACROGEN MPRINT SYMBOLGEN;

to print full source with line numbers on the SAS log.

Invoke the FULL option in //SYSIN with this syntax:

%LET MXGDEBUG=FULL; %VMXGINIT;

-The %LEFT in %VMXGPARS was replaced.
Change 33.061 GDPS SMF 105 INPUT STATEMENT EXCEEDED when XVMX segment

VMAC105 was not present; the bit test was insufficient to confirm

Mar 16, 2015 the segment was extended so a length test was added.

Thanks to Paul Volpi, UHC, USA.

Change 33.060A Change text was lost, added May 2016.

ANAL113 Dataset ASUM113 was replaced by ASUM1131 for reports.

Mar 11, 2015
Change 33.060 -DCOLLECT Cluster dataset DCOLCLUS now contains variables

VMACDCOL DCDDATCL,DCDMGTCL,DCDSTGCL, Data/Management/Storage Class

Mar 11, 2015 that are retained from the immediately preceding type D

Mar 21, 2015 record prior to the type A record, when the "D" record's

DSNAME matches the "A" records' Cluster Name.

-In DCOLDSET dataset, multi-volume records now contain

those three Class variables in all observations; in the

DCOLLECT records, those fields are only populated in the

DCDVOLSQ=1 records. This change revised the _SDCODSN

Data Set Sort to propagate from the first to subsequent

volume's observations. You MUST USE TYPSDCOL to invoke

that _SDCODSN data set sort or classes will be blank in

the 2nd and subsequent volumes.

-Mar 21: DCOLDSET dataset label added to PROC SORT.

Thanks to John Kim, Worker's Compensation Board of Alberta, CANADA.
Change 33.059 Support for Thales Security Resource Manager RG1100.

EXTHALRT -New variables added to THALSUMD dataset:

IMACTHAL THAASID ='SSID OF*ORIGIN*APPL-AS'

VMACTHAL THADBUSY='BUSY*WITH*USER*REQUESTS'

VMXGINIT THADINTV='INTERVAL*DURATION'

Mar 10, 2015 THADNCNT='REQUESTS*PROCESSED'

Mar 24, 2015 THADOVER='BUSY*WITH*SRM*REQUESTS'

Apr 15, 2015 THADUFLG='DEVICE*CONTINUOUS*USE*OR NOT'

-New THALRESP Response Time dataset, with variables:

THAASID ='SSID OF*ORIGIN*APPL-AS'

THAJBNM ='JOBNAME*OF ORIGIN*USER APPL'

THARSTCK ='USER*MESSAGE*INITIATION*TIME'

THARFDBK ='FEEDBACK*CODE*SET BY*SRM'

THARPRI ='MESSAGE*PRIORITY'

THAMTYPE ='MESSAGE*TYPE*HSM*CODE'

THAERRCD ='ERROR*CODE*SET BY*HSM'

THARINTV1='1ST*INTERVAL*ELAPSED*TIME'

. . .

THARINTV12='12TH*INTERVAL*ELAPSED*TIME'

-Note: TYPETHAL replaced TYPESRHS.

-Mar 24: THAHSM formatted $HEX16, THAHSMR $HEX32.

-Apr 15: THALRESP updated and validated with data.

Datasets THALEXCE,THALVIOL,THALSUMD,THALRESP have these

variables decoded from THAHAM/THAHAMR:

THAHSMCH='HSM*CHANNEL*DEVICE*NAME'

THAHSMIP='HSM*IP*ADDRESS'

THAHSMAX='HSM*IP*AUXILLARY*PORT'

THASMVT ='HSM*VTAM*DEVICE*NAME'

Thanks to Yves Cinq-Mars, IBM Global Services, CANADA.
Change 33.058 Support for CICS User Field DNDBKR/DNDBKR in CICSTRAN.

IMACICVK

UTILEXCL

VMAC110

Mar 8, 2015

Thanks to Michael Creech, Black Knight Financial Services, USA.

Change 33.057 ECHO= option added in Change 32.154 but not documented:

UTILBLDP If you want to see the code that was created by UTILBLDP,

VMXGPARS it can be printed on the LOG by specifying either the new

Mar 10, 2015 ECHO=YES (or ECHO=Y) argument, or with MXGEXIMSG=YES.

Mar 20, 2015 -%LEFT replaced in VMXGPARS.
Change 33.056 Using OPTIONS OBS=0 to create zero-observation datasets

SAGANAL from SMF data, read via the ftp access method, executing

UTILBLDP on ASCII, failed with these messages:

Mar 11, 2015 NOTE: <<< 451 Transfer aborted: send error.

NOTE: >>> QUIT

ERROR: Bad request. Use the debug option for more info

SAS Note http://support.sas.com/kb/14/679.html documents

that OBS=0 can not be used with ftp access method.

Tailoring to build RMFINTRV from only 70s and 72s used

%INCLUDE SOURCLIB(TYPS7072);

to create the TYPE70 and TYPE72GO datasets for input to

create the RMFINTRV dataset with CPU metrics, but using

OPTIONS OBS=0;

%%INCLUDE SOURCLIB(TYPS71,TYPS73,TYPS74,TYPS75,TYPS78);

RUN;

to create the other datasets required for RMFINTRV, but

creating them with zero observations so they take no disk

space, preventing B37 if high volume ID=74 SMF records

were accidentally in the input SMF file. Each TYPS member

opens the SMF filename, although no records are read.

This ancient and inefficient syntax was replaced with

%UTILBLDP(OUTFILE=INSTREAM,

BUILDPDB=NO,

USERADD=7072 71 73 74 75 78,

ZEROOBS=71 73 74 75 78);

%INCLUDE INSTREAM;

-But using UTILBLDP under Linux raised errors because some

macros (%QCMPRES, %LEFT) were not found because SASAUTOS

is not allocated by default under Linus (CH 33.051) and

were unresolved. But Change 31.209 had stated that those

macros were replaced by SYSFUNC calls to avoid the long

standing problem trying to use SAS-supplied SASAUTOS

macros, so UTILBLDP is now free of SASAUTO macros.

Thanks to David F. Salsieder, American Family Insurance, USA.

Change 33.055 Support for CICS User Field USER/MEBTRAN in CICSTRAN.

IMACICVK

UTILEXCL

VMAC110

Mar 8, 2015

Thanks to Jeff Fracas, WiPro, USA.

Change 33.054 Reserved Change Number.
Change 33.053 -Support for z13 updates to type 99 subtype 14, STRONGLY

VMAC99 RECOMMENDED by IBM to be enabled for PROCVIEW CORE to

Mar 8, 2015 allow you to see which LPARs end up where.

Variables were added to both TYPE99EM and TYPE99EP.

Change 33.052 z13 Support updates.

VMAC113 -With PROCVIEW CORE, SMF 113 was INCOMPATIBLY changed to

ASUM113 add COREID etc to support SMT, with changes to MANY of

Mar 14, 2015 the calculated variables that were not in Change 33.023.

Apr 10, 2015 -The SPEED value for the zIIP can incorrectly be zero but

MXG sets SM1132SP=5 for the z13 to circumvent the known

occasional incorrect value.

-With PROCVIEW CPU, non-SMT mode, RNI was incorrect, with

a negative value, that was also corrected by this change.

-Apr 10: zEC12 RNI 1st factor changed from 2.2 to 2.3.

Change 33.051 MXG initialization revised so the FILENAME SASAUTOS isn't

VMXGINIT listed in the list of input source libraries, as Linux by

Mar 3, 2015 default does not have a SASAUTOS filename allocated.
Change 33.050 Graphics code had a not sorted condition caused by the

ANALHSM order of the variables in the BY list. Protection added

Mar 3, 2015 for all missing values in a variable, which caused ODS

graphics to print warning and suppressed changes in the

tick value formats by using XAXIS YAXIS code.

Thanks to Lindsay Oxenham, IBM, AUSTRALIA.

Change 33.049 Three new UTILEXCL reports may be printed. REPORT 00 will

UTILEXCL print all dictionary records read in today from SMF.

Mar 2, 2015 REPORT 00 prints any Transaction Records that do NOT have

a dictionary record, and REPORT 000 shows each of the

APPLIDs that did not have a dictionary record. Both 00

and 000 reports show the SYSTEM and READTIME of those

Transaction Records, since THAT is when the dictionary

record was written, so you know which day's SMF to read

to find those missing dictionary triplet records.

Thanks to Scott Barry, SBBWorks, Inc., USA.

Change 33.048 -When a GETDSAB error occurred the Reason Code was

ADOCRMFV displayed as NNNN in severe error message RMFV007S.

ASMRMFV The Reason Code is only valid if the Return Code is

Feb 27, 2015 12. In all other cases the Reason Code for this

situation will display as blanks.

-Update documentation for message RMFV007S.

-REQUIREMENT: In order to implement these features the

ASMRMFV utility program from this MXG change must be

installed. See MXG SOURCLIB member JCLASM3 for sample

JCL for the assembly and link-edit install steps.

Thanks to Wayne Bell, UniGroup, Inc.
====== Changes thru 33.047 were in MXG 33.02 dated Feb 27, 2015=========
Change 33.047 -UTILEXCL updated to skip the CICS/TS 3.2 OMEGAMON CMRDATA

IMACICM0 segment when it is the incorrect length of 200 bytes.

UTILEXCL Change 28.027 discusses, but only told you to install the

Feb 27, 2015 correction; this change circumvents by telling you to use

the IMACICM0 (instead of the normal IMACICMR) to skip.

Thanks to Donald Blaszka, Wipro, USA.

Thanks to Jeff Fracas, Wipro, USA.
Change 33.046 New variables in TYPE70 for z13 MULTITHREADING:

FORMATS

VMAC23 MXG 33.01 CHANGES noted one z13 site had NEGATIVE CPUOVHTM,

VMAC26J2 but that was ONLY IF z/OS on z13 is in MULTI-THREADING MODE.

VMAC26J3 MXG was unaware of IBM's restructuring the RMF type 70 SMF

VMAC7072 record's calculation of CPU BUSY time for the new MT mode,

VMAC71 but this MAJOR CHANGE restructured TYPE70 processing in MXG

VMAC75 to order by CORE_ID and CPU_NUM rather than CPUID/LCPUADDR.

Feb 26, 2015

If NOT Multi-Threading, new metrics were COMPATIBLY ADDED.

Mar 21, 2015

Specifically, if z/OS is IPL'd with LOADxx PROCVIEW CORE, on

a processor that is SMT capable, then whether or not MT is

Active, and even if MT=1 is specified, then the RMF 70 record

is restructured with CORE_ID. While this change in MXG 33.02

updated the TYPE70 dataset, MXG 33.03 is REQUIRED now for the

complete support for the TYPE70PR changes as well.
SMF70MCF ='MULTITHREADING*MAXIMUM*CAPACITY*GP'

SMF70MCFS='MULTITHREADING*MAXIMUM*CAPACITY*ZIIP'

SMF70MCFI='MULTITHREADING*MAXIMUM*CAPACITY*ZAAP'

SMF70CF ='MULTITHREADING*CAPACITY*GP'

SMF70CFS ='MULTITHREADING*CAPACITY*ZIIP'

SMF70CFI ='MULTITHREADING*CAPACITY*ZAAP'

SMF70ATD ='AVERAGE*THREAD*DENSITY*GP'

SMF70ATDS='AVERAGE*THREAD*DENSITY*ZIIP'

SMF70ATDI='AVERAGE*THREAD*DENSITY*ZAAP'

New variables in TYPE70PR

SMF70MTID='MAXIMUM*THREAD*IDENTIFICATION'

SMF70MTTT='MULTI*THREADING*IDLE*TIME'

Variables SMF70CAN/SMF70CTN are now correctly missing

values for an LPAR with no engines (LCPUADDR=. also).

These variables were previously incorrectly populated.

The order of these offline LPARS may be different now.

New variables in TYPE71

SMF71MCF='MULTITHREADING*MAXIMUM*CAPACITY'

And several PIB4 variables are now INPUT as RB8.

TYPE75 SMF75AVL now input as floating point.

New variables in TYPE799 dataset:

R799CUQ ='CU*QUEING*TIME'

R799CN2 ='DEVICE*FLAG*ESTENSION*2'

R799CSC ='SUBCHANNEL*SET*IE'

New variables in TYPE89 dataset:

SMF89CR ='0=CP CREDITS*1=IB CREDITS'

New variables in TYPE9005 dataset:

SMF90ESWT='SWT*VALUE'

SMF90ETWT='TWT*VALUE'

New variables in TYPE23 dataset:

SMF23BBC='ZEDC*UNCOMPRESSEDSED*BYTES*TOTAL'

SMF23BAC='ZEDC*COMPRESSED*BYTES*TOTAL'

New variable in TYPE26J2 and TYPE26J3

SMF26JCR='JOB*CORRELATOR'

In multi-threading mode, the TYPE70PR data is changed

from a one-to-one mapping of CPUID and LCPUADDR to the

CORE_ID, and a core can have 2 CPUs. The CPUID value from

the CPU Data Section is used to select the suffix for

those datasets that have a unique set of variable names

for each LPAR (TYPE70,ASUM70PR,ASUMCEC).

Change 33.045 The SMF 110 ST 1 record descriptions in the ANALID report

VMACID incorrectly stored 110.001 in variable IDANDSUB instead

VMACSMF of values 110.101-110.106, so only '110.001' was printed

Feb 23, 2015 instead of these possible descriptions;

110.1.1:CICS DICTIONARY MNSEGCL=1

110.1.3:CICS TRANSACTION MNSEGCL=3

110.1.4:CICS EXCEPTION MNSEGCL=4

110.1.5:CICS RESOURCE MNSEGCL=5

110.1.6:CICS IDENTITY MNSEGCL=6

While this is only a cosmetic error, these are important

enough that I refreshed 33.01 to add this change.

Thanks to Andre G Amoretto, IBM Global Services, BRAZIL.

====== Changes thru 33.044 were in MXG 33.01 dated Feb 20, 2015=========
Change 33.044 See Change 33.155. These data are now in TYPE74ST.

VMAC74 Corrections for TYPE74MO SCM Structure dataset. The test

Feb 19, 2015 for length had incorrect syntax, but fortunately did not

cause a false positive. There was a missing */ in the

INPUT that misaligned data. Also, it appears, from this

note 2 on page 366 of z/OS V2R1.0 RMF Report Analysis

that TYPE74MO can have zero observations because:

"SCM statistics are only included in the SCM Structure

Summary for those structures that make use of the SCM

storage extension and have set a non-zero maximum SCM

size. If none of the structures is configured to exploit

SCM, the SCM Structure Summary displays message: No

storage class memory data available."

Thanks to Otto Burgess, ATPCO, USA.

Thanks to Scott Barry, SBBWorks Inc., USA.
Change 33.043 DATASET VXUSEACT/VXUSEINT NOT SORTED ERROR because the BY

VMACVMXA lists for the MERGE were not identical.

Feb 19, 2015

Thanks to Chris Weston, SAS Institute, USA.

Change 33.042 Using INCLAFTR=BUIL3005 for JES3 failed with PDB.TYPE25 ,

UTILBLDP not found because UTILBLDP incorrectly blanked _STY25.

Feb 19, 2015

Thanks to Paul Volpi, UHC, USA.

Change 33.041 SMFSRCH did not read the SMFOUT file of selected records,

FORMATS due to an undocumented change in VMXGINIT that replaced

SMFSRCH &SMF with SMF, preventing changing the input DDNAME and

VMXGINIT causing the full input SMF file to be read by TYPESMF,

Feb 18, 2015 287 times! TYPESMF is replaced by a tailored UTILBLDP

Feb 23, 2015 invocation that only processes the selected record types,

Feb 24, 2015 and in only one step. And a PROC FORMAT that could fail

with overlapped values was replaced in the new design.

The selected record types are reported, and if you have

user SMF records selected, you will need to use USERADDS

%SMFSRCH(USERADDS=xxxx/nnn yyyy/mmm) to tell MXG which

xxxx product has which nnn SMF record type.

-Format MGSMFPR maps SMF Record Type to Product Suffix for

the UTILBLDP USERADD= argument.

-Feb 23: Harmless APPARENT UNRESOLVED ADDIDS if no records

were selected corrected.

-If SMFSRCH failed to find your text in SMF, it generated

a call to UTILBLDP with BUILDPDB=NO and USERADD= to null

which caused UTILBLDP to fail with MACRO variables not

found. Now if no records are found, it terminates with:

MXGNOTE: NO RECORDS WERE FOUND CONTAINING &LOOKFOR;

MXGNOTE: SMFSRCH WILL BE TERMINATED;

and if UTILBLDP has BUILDPDB=NO and USERADD=blank it will

terminate with: