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

QWP4ESTC='EDM*POOL*STATEMENT*CACHE SIZE'

QWP4HINT='OPTIMIZATION*HINTS*ALLOWED?'

QWP4INTE='RTS*STATISTICS*TIMER*INTERVAL'

QWP4LEM ='MAXIMUM*NUMBER OF*LE TOKENS'

QWP4LRTH='UNCOMITD*READ*WARNING*THRESHOLD*(MINS)'

QWP4MDEG='MAX DEGREE*OF PARALLELISM'

QWP4MDSC='MIN SCALE*FOR DECIMAL*DIVISION*RESULT'

QWP4MNTY='MAINTAINED*TABLE*TYPE*BITMAP'

QWP4MXNC='MAX*CURSORS*OPEN PER*THREAD'

QWP4MXSP='MAX*ACTIVE*STORED PROCS*PER THREAD'

QWP4NPAG='NPAGES*THRESHOLD*FOR OPTIMIZER'

QWP4OJEH='OUTER*JOIN*PERFORMANCE*ENHANCEMENT?'

QWP4OZCI='ONLINE*ZPARM*CORRID*MONITOR'

QWP4OZTM='ONLINE*ZPARM*TIME*CHANGED'

QWP4OZTP='ONLINE*ZPARM*TYPE'

QWP4OZUS='ONLINE*ZPARM*USERID*MONITOR'

QWP4PDIX='PAD INDEXES BY DEFAULT?'

QWP4PKYU='ALLOW*UPDATE*PARTITIONING*KEY?'

QWP4RAC ='ROUTINE*AUTHORIZATION*CACHE SIZE'

QWP4RFSH='REFRESH*AGE'

QWP4SJRT='STAR*JOIN*RATIO'

QWP4SJTB='STAR*JOIN*THRESHOLD*(#TBLS/QB)'

QWP4SJXP='MAX SIZE*MEMORY POOL*FOR STARJOIN'

QWP4VDTY='DEVICE*TYPE FOR*TEMPORARY*DATA SETS'

QWP4WAIT='RETAINED*LOCK*TIMEOUT*MULTIPLIER'

QWP4EDMX='EDM POOL MAX DATA SPACE SIZE'

QWP4EDDS='EDM POOL DATA SPACE SIZE'

Thanks to Ron Alterman, Pacific Gas and Electric, USA.

Change 23.159 Yet another INPUT STATEMENT EXCEEDED RECORD for PSF SMF 6

VMAC6 record with SMF6LN6=24 (See Changes 23.091, 22.309).

Jun 13, 2005 The logic for SUBS=7 and SMF6LN6=24 was revised based on

the current SMF manual so that when SMF6LN6=24, there is

no longer an attempt to input SMF6PRTQ name, since there

cannot be roome for a name field when SMF6LN6=24.

Thanks to Diane Eppestine, Southwestern Bell, USA.
Change 23.158 MXG RMF-like reports printed whole numbers for AVG RESP

ANALRMFR TIME and AVG IOSQ TIME but IBM reports now print one

Jun 13, 2005 decimal place, so MXG format was changed to match.

Thanks to Bob Anders, McMaster-Carr Supply Co., USA.

Change 23.157 SMF file might not have been found, depending on use of

BLDSMPDB SMFIN= argument, and whether FILENAME SMF existed. The

Jun 10, 2005 &SMF was changed to FILENAME SMF "&SMFIN"; to correct.

Thanks to Joe Key, BOC Gasses, ENGLAND

Thanks to Alex DaSilva, BOC Gasses, ENGLAND
Change 23.156 ASUMUOWT had incorrect syntax for _LMONTSK definition.

ASUMUOWT Two periods are required.

Jun 9, 2005

Thanks to Chris Weston, SAS Institute Cary, USA.

Change 23.155 -First 23.05 only; incorrect syntax for default options in

ASMTAPEE "OPTIONS DC" statement, with a comma instead of plus.

Jun 9, 2005 -One site with ASM option CASE as their default had to use

ASM option COMPAT(NOCASE) because ASMTAPEE contains both

upper and lower case text, for readability.

Thanks to Dan Squillace, SAS Institute Cary, USA.

Thanks to Shirley Fhng, Hong Kong Shanghai Bank, Hong Kong.

====== Changes thru 23.154 were in MXG 23.05 dated Jun 7, 2005=========

Change 23.154 First 23.05 only. Line 3138 extended into column 3138.

ASMTAPEE Comments clarified.

Jun 7, 2005

Thanks to Dan Squillace, SAS Institute Cary, USA.

Change 23.153 -Utility to read SMF to count IDs and Subtypes did not

VMXGGETM count ID=102 correctly, when run on ASCII platform; those

Jun 7, 2005 PIB4. and PIB2. should have been &PIB.4. and &PIB.2. so

that the code was transparently portable across all SAS

platforms. With PIB4. on ASCII, extremely large values

were created, which caused the ID=102 sub-logic to never

be executed, so all ID=102 had SUBTYPE=0 on ASCII.

-The realization that I had not checked ALL MXG members to

be transparently portable cause me to search and find

these members with either PIB instead of &PIB., or with

$ instead of $CHAR or $EBCDIC; all are now corrected:

ANALCM27 ANALCM29 ANALIPAC ANALSNAP CHANGES EXIMRACF

IDMSJRNL IDMSLO57 IDMSLOG JCLCIMS JCLPDBXX MAKECODE

TYPECIMS TYPEIMSB TYPEMONX TYPESLRI UDB2GTF UDEBLOCK

USMFRATE USTKVOL UTILDBLK UTILGEVB UTILGEVM UTILSPAC

VAXPDS VMACCMFV VMACCOMP VMACMONI VMXGGETM XADALOG

XGTFANAL XIBMFDP XLOGREC XMACACHE XMACVMXP XPAII

XTYPEMVT XTYPEVS1

Fortunately, all are pretty obscure, non-mainstream code,

or someone would have certainly noticed the incorrect

data values if they were exected on ASCII platforms.

Thanks to Alfred Holtz, Medco Health Solutions, USA.

Change 23.152 First MXG 23.05 Only. Debug PUTs at lines 2147 and 2148:

VMACTPMX PUT _N_= COL= FLENGTH= VARNAME= TPMFIELD=

Jun 7, 2005 TOKENID= TOKFIELD= TOKFIELD= $HEX16.;

printed unneeded lines of text on the SAS log.

Thanks to Lawrence Jermyn, Fidelity Systems, USA.
Change 23.151 MXG 23.03-First 23.05. Debug PUT statement at line 1425

VMACNDM PUT _N_= COL= NDMPNODE= NDMSNODE= ;

Jun 6, 2005 printed unneeded lines of text on the SAS log.

Thanks to Michael Creech, Fidelity Information Systems, USA.

Thanks to David Kaplan, DTCC, USA.
Change 23.150 First MXG 23.05 only. VARIABLE NOT FOUND ERROR due to

ASUMMIPS typo: BY RPPTCLAS ... should be BY RPRTCLAS ....

Jun 6, 2005

Thanks to Pat Curren, Supervalu Inc, USA.

Change 23.149 Cosmetic. Examples of defining start and end of SHIFTs

IMACSHFT was clarified to make it easier for neophytes to change

Jun 6, 2005 the values to define their installation's shifts.
====== Changes thru 23.148 were in MXG 23.05 dated Jun 5, 2005=========
Change 23.148 General cleanup and revision to QA job stream:

ANALVM -Uninitialized xxxIFE and xxxIFA in RMFINTRV corrected.

QASAS -Specific reference to VMPDB removed in ANALVM, test of

TESTANAL ANALVM removed from TESTANAL as it was tested in TESTVM.

TRNDNTIN -QASAS code revised to PROC DELETE all LIBREFs.

TRNDNTLD -LENGTH/FORMAT added to MAXTIME/MINTIME for PRINTER data

TRNDPRTR set built in TRNDPRTR.

VMXGRMFI -TRNDNTIN,TRNDNTLD revised to use only &Pdddddd, and

Jun 5, 2005 eliminated the confusing _L macro definitions.

Thanks to Freddie Arie, Merrill Consultants, USA.

Thanks to Bruce Widlund, Merrill Consultants, USA.
Change 23.147 The response time buckets for the PDB.CICS dataset that

ASUMCICS is built by ASUMCICX from PDB.ASUMUOW dataset (which is

ASUMCICT the recommended way to measure CICS response with MRO),

ASUMCICX (or for the PDB.CICS dataset that can be built by the

VMXGINIT not-recommended ASUMCICS/ASUMCICT members from CICS

Jun 3, 2005 transaction segments in datasets CICSTRAN/MONITASK)

are defined as macro variables but they could not be

overridden without tailoring the ASUM member. Now, the

macro variables are externally defined in VMXGINIT, and

can be overridden in your //SYSIN DD *, before your

%INCLUDED SOURCLIB(ASUMCICx); statement. The default

values remain unchanged:

%LET SUCIVAL1=1;

%LET SUCIVAL2=2;

%LET SUCIVAL3=3;

%LET SUCIVAL4=4;

%LET SUCIVAL5=5;

%LET SUCIVAL6=8;

%LET SUCIVAL7=10;

%LET SUCIINTV=HOUR;

Thanks to Frank Lund, EDB, NORWAY.
Change 23.146 Support for new subtype 70 "FTP Server Security Section"

FORMATS adds these variables to TYP11970 dataset.

VMAC119 FSPSCCPL='CONTROL*CORRECTION*PROTECTION*LEVEL'

Jun 1, 2005 FSPSCIPH='CIPHER*SPECIFICATION'

FSPSDCPL='DATA*CORRECTION*PROTECTION*LEVEL'

FSPSLGME='LOGIN*METHOD'

FSPSMECH='PROTECTION*MECHANISM'

FSPSPRBF='NEGOTIATED*PROTECTION*BUFFER*SIZE'

FSPSPROT='PROTOCOL*LEVEL'

and new formats were created to decode four of them.

Thanks to Randy Shumate, LexisNexis, USA.
Change 23.145 The sample RMF III report had typos; ASISUPR should have

ZRBRPT3 been spelled ASISUCPR, and ASISCDOP should be ASISCDQP.

Jun 1, 2005

Thanks to Mike Obrien, Bank of America, USA.

Change 23.144 The "expanded length" format values introduced for TNG in

FORMATS Change 23.113 were most definitely non-trivial to code.

May 31, 2005 Initially, PROC FORMAT on ASCII SAS V9.1.3 ran without an

error, but values were not-found if a spanned-line had a

blank in column 72; those lines were then shifted right

so column 72 was non-blank, and values were then found.

But on z/OS V8.2, the new code failed to execute, with an

error 22-322 which underlined the equal sign on a line

that had 72 characters. Meanwhile, the new code executed

without error on z/OS 9.1.3, AND all values were found,

even for the spanned-text lines!
Noting that the V8.2 execution errors were on lines that

were exactly 72 positions, the code was revised to delete

a blank on the left of the '=' where possible, or the

line was split at the equal sign, and now the PROC FORMAT

executed without the 22-322 error on V8.2 on z/OS.
But now, values were not-found for spanned-text lines on

z/OS V8.2, while they were found on z/OS V9.1.3! This

problem had been opened with SAS Technical Support, and

their tests suggested that OPTIONS S2 might be involved,

as they could replicate the error when OPTIONS S2=71 was

used. Our real breakthrough was to compare execution on

z/OS V8.2 using %INCLUDE, versus copying the code into

the //SYSIN DD * and submitting the code; the instream

test had no errors AND properly found all values, while

the %INCLUDE-built format had values not-found!

More tests proved that the PROC FORMAT execution on z/OS

requires S2=72, while PROC FORMAT under ASCII requires

S2=0, so the end result is that member FORMATS now tests

its execution platform and sets S2=72 for PROC FORMAT on

EBDCIC, but sets OPTION S2=0 when executed on ASCII.

For z/OS, MXG's CONFIGV8 member has always had S2-72,

but two of the test sites were using SAS default S2=0,

which was known only in hindsight! MXG's CONFIGV9 for

z/OS has S2=0 (as extensively discussed in CHANGES; see

"Major Enhancements in MXG 22.08" and Newsletter 45),

but the AUTOEXEC for ASCII execution does not specify

either S= nor S2= option, so the S2=0 option on ASCII

accounts for my early tests not failing in execution.

This problem is still open at SAS, to understand what

is the righteous way to code formats that have values

that won't fit on one line, but the revised FORMATS

member now appears to work on all platforms with the

internal setting of S2 above.

Change 23.143 IMPLEX subtype 4 record's IMPLEXRT dataset did not have

VMACMPLX all of the possible variables kept, and the deaccumulate

May 31, 2005 logic was incomplete, causing zero observations in the

PDB.IMPLEXRT dataset; groups of the new variables will be

missing, as they are only input for specific values of

the EXMFRTYP variable. Some of the MXG tests for other

DIF()'d variables that were missing the "LT 0" for their

final test are now corrected.

Thanks to David Bixler, FISERV, USA.
Change 23.142 Support for Oce's Prisma Print log file creates these

EXPR1000 forty-five new datasets:

EXPR1010

EXPR1011 DDDDDD MXG MXG

EXPR1012 DATASET DATASET DATASET

EXPR1020 SUFFIX NAME LABEL

EXPR1030

EXPR1031 PR1000 PRPR1000 PR1000:INPUT FILTER

EXPR1032 *PR1010 PRPR1010 PR1010:PRINT JOB MANAGER JOB

EXPR1040 *PR1011 PRPR1011 PR1011:PRINT JOB MANAGER FILE

EXPR1041 *PR1012 PRPR1012 PR1012:ODS

EXPR1042 *PR1020 PRPR1020 PR1020:UI-MANAGER

EXPR1043 *PR1030 PRPR1030 PR1030:SPOOL PARAMETER READ

EXPR1110 *PR1031 PRPR1031 PR1031:JOB FINISHED OR DELETED

EXPR1120 *PR1032 PRPR1032 PR1032:USER DATA

EXPR1130 *PR1040 PRPR1040 PR1040:SNIPDS BACKEND

EXPR1140 *PR1041 PRPR1041 PR1041:BINS USED

EXPR1200 *PR1042 PRPR1042 PR1042:INPUT BIN USED

EXPR1201 *PR1043 PRPR1043 PR1043:OUTPUT BIN USED

EXPR1202 *PR1110 PRPR1110 PR1110:HOST DOWNLOADED

EXPR1400 PR1120 PRPR1120 PR1120:LP TRANSMISSION

EXPR1410 PR1130 PRPR1130 PR1130:HOT DIRECTORY PRINT JOB

EXPR1411 PR1140 PRPR1140 PR1140:LU 6.2 PRINT JOB

EXPR1412 PR1200 PRPR1200 PR1200:XFILTER LCDS REPORT INFO

EXPR1413 PR1201 PRPR1201 PR1201:XFILTER LCDS ADDITIONAL

EXPR1420 PR1202 PRPR1202 PR1202:XFILTER LCDS MORE ADDITI

EXPR1421 PR1400 PRPR1400 PR1400:B1 START SUBSYSTEM

EXPR1422 PR1410 PRPR1410 PR1410:E1 VOLUME HANDLING

EXPR1423 PR1411 PRPR1411 PR1411:F5 HDR1 HANDLING

EXPR1424 PR1412 PRPR1412 PR1412:F6 HDR2 HANDLING

EXPR1425 PR1413 PRPR1413 PR1413:F7 HDR3 HANDLING

EXPR1426 PR1420 PRPR1420 PR1420:N5 PARAMETER SECTION 1

EXPR1430 PR1421 PRPR1421 PR1421:N6 PARAMETER SECTION 2

EXPR1451 PR1422 PRPR1422 PR1422:N7 PARAMETER SECTION 3

EXPR1452 PR1423 PRPR1423 PR1423:P4 PARAMETER SECTION 4

EXPR1453 PR1424 PRPR1424 PR1424:P5 PARAMETER SECTION 5

EXPR1454 PR1425 PRPR1425 PR1425:P6 PARAMETER SECTION 6

EXPR1455 PR1426 PRPR1426 PR1426:P7 PARAMETER SECTION 7

EXPR1456 PR1430 PRPR1430 PR1430:PRINTING RUN

EXPR1457 PR1451 PRPR1451 PR1451:U1 USER SPECIFIC

EXPR1458 PR1452 PRPR1452 PR1452:U2 USER SPECIFIC

EXPR1459 PR1453 PRPR1453 PR1453:U3 USER SPECIFIC

EXPR1600 PR1454 PRPR1454 PR1454:U4 USER SPECIFIC

EXPR1601 PR1455 PRPR1455 PR1455:U5 USER SPECIFIC

EXPR1602 PR1456 PRPR1456 PR1456:U6 USER SPECIFIC

EXPR1603 PR1457 PRPR1457 PR1457:U7 USER SPECIFIC

FORMATS PR1458 PRPR1458 PR1458:U8 USER SPECIFIC

IMACPRPR PR1459 PRPR1459 PR1459:U9 USER SPECIFIC

TYPEPRPR PR1600 PRPR1600 PR1600:FTP BACKEND

TYPSPRPR *PR1601 PRPR1601 PR1601:FTP JOB FINISHED

VMACPRPR *PR1602 PRPR1602 PR1602:FTP TRANSMISSION PRINT FIN

VMXGINIT *PR1603 PRPR1603 PR1603:FTP TRANSMISSION GEN OCT

May 30, 2005 All of the datasets are created, but only those fifteen

asterisk-flagged datases have been decoded and validated

with test data records; all others have only one variable

kept. As there are multiple records for a single job, it

may be necessary to post-process these data to combine to

get job totals, so this is preliminary support.

Thanks to Krista Neven, KBC Bankverzekeringsholding, BELGIUM.
Change 23.141 Enhancement adds MIPFACTR= as a parameter to convert MSU

GRAFWRKX to MIPS, added graph of MIPS used, supported LPARSHAR as

May 30, 2005 a percentage to fix the LPAR percentages.

Thanks to Chuck Hopf, MBNA, USA.

Change 23.140 DB2 Version 8.1 Support: More IBM INCOMPATIBLE changes.

VMACDB2 Errors INVALID DATA FOR QLACCPUR, QLACCPnt, causing

May 30, 2005 INVALID DATA FOR QLACCPUL, QLACCPUR, QLACDBAT messages.

Jul 17, 2005 Change 23.111 discovered that DB2 8.1 sets QPACLEN=zero

to indicate that the real QPACLEN is in the first 2 bytes

at OFFQPAC, but when a QLACLEN=0 segment caused errors,

I realized this "feature" apparently can apply to all DB2

segments, so VMACDB2 was revised to test each segment for

LENxxxx=0 and NRxxxx non-zero, with this inserted code

OFFxxxx=OFFxxxx

IF LENxxxx EQ 0 AND NRxxxx GT 0 THEN DO;

INPUT @OFFxxxx LENxxxx &PIB.2. @;

OFFxxxx=OFFxxxx+2;

LENxxxx=LENxxxx+2; <<= See Change 23.182

END;

to pick up the new LEN and modify the OFFSET.

-Note: The LENQLAC=0 SMF records did NOT produce that same

INVALID DATA message when those records were read by SAS

under Windows; instead, the misaligned RB8. fields were

input as very large negative numbers, but there was no

error message, and I had not used PROC MEANS to see that

that invalid values existed in the output dataset. This

is one of few real differences in executing MXG on ASCII

platforms; ASCII is more tolerant of bad data than z/OS!

-Change 23.181 updated this text and is required for V8.1.

Thanks to Ingvar Rosenberg, SEB, SWEDEN.

Thanks to Glenn Lundgren, SEB, SWEDEN.
Change 23.139 Support for CyberFusion user SMF record has been revised,

FORMATS now that DSECTS that match the data records have been

VMACCYFU received. All of the bit mapped fields are now decoded

May 29, 2005 by 43 new MGCYFxx formats, although there are a few cases

where the hex value is not described in the DSECT EQU's,

so some final tweaking will be required.

Thanks to Robin Murray, ManuLife, CANADA.
Change 23.138 Cosmetic. Labels for SMF88SC1/2/3 are clarified to read:

ADOC88 SMF88SC1='TYPE 1*COMPLETIONS'

VMAC88 SMF88SC2='TYPE 2*COMPLETIONS'

May 30, 2005 SMF88SC3='TYPE 3*COMPLETIONS'

and the explanation from the SMF manual Section 9.1.1.2

was added in the ADOC88 member for those three variables.

Thanks to Helmut Rose, COM Software, GERMANY.
Change 23.137 Support for CA SYSVIEW CICS XPFCMCR DSECT user SMF data

EXSYSVCI record creates these new datasets:

EXSYSVPC -Subtype 17: Documented in XPFCMRC, creates these data:

EXSYSVFC dataset dddddd description

EXSYSVTS SYSVCICS SYSVCI SYSVIEW CICS PERFORMANCE (CICSTRAN)

EXSYSVIN SYSVPROG SYSVPC SYSVIEW CICS PROGRAMS

IMACSYSV SYSVFILE SYSVFC SYSVIEW CICS FILES

TYPESYSV SYSVTEMP SYSVTS SYSVIEW CICS TEMP STORAGE

TYPSSYSV The SYSVCICS dataset is the same in structure as CICSTRAN

VMACSYSV and almost all of the variable names are also the same,

VMXGINIT although not all of the (newer) CICSTRAN fields exist in

May 29, 2005 the SYSVCICS dataset, and there are a few CICSTRAN vars

that are output instead in SYSVPROG,SYSVFILE or SYSVTEMP.

There are additional segments in the SYSVIEW subtype=17

record, but they were not populated in the test data, so

only the subtypes that could be validated were decoded in

this initial support. If data from the other segments is

really needed, updates will be made to this support.

-Subtype 23: Documented in XPFCSDCR, creates this data:

dataset dddddd description

SYSVINTV SYSVIN SYSVIEW SYSTEM DATA COLL INTERVAL

In a radical departure from previous MXG code, the names

of the variables in SYSVINTV are longer than 8-bytes. The

DSECT names are repeated inside structured names, and to

compact the names and avoid duplicates would have taken

more time that it was worth, at least for this first pass

as support for the SYSVIEW product, and especially for

this SYSVINT dataset, neither of which I expect will have

very much usage. I think long names, in general, are more

confusing than useful, so if it turns out that SYSVINTV

is of significant value, I'll revisit and compact the MXG

variable names.

- The DOCVER member will still only show the first eight

characters of the variable names, as that format is

limited so the format and label fit on one line. You

can always use a PROC CONTENTS to see the full name of

each variable in SYSINTV.

Thanks to James D. Lieser, United Health Group, USA.

Change 23.136 -Updates to BMC CMF "240" USER SMF record:

VMACCMF -New variables in CMF27C93 and CMF27CSC for 2105's:

May 26, 2005 C276RCRM='RECORD*CACHE*READ*MISSES'

C276LRED='LOC RCD*DMN/REC*CA*WRITES'

C2795AID='DEVICE ADAPTER ID'

C2795HDD='NO. OF HDD'S IN RAID RANK'

C2795HSS='HDD sector size'

C2795NVS='NVS space allocation'

C2795RFL='FLAGS BYTE, THUS:'

C2795RID='RAID RANK ID'

C2795RMR='record mode read requests'

C2795RRQ='RAID rank read requests'

C2795RRT='RAID RANK READ RESPONSE TIME'

C2795RSV='LOWER*INTERFACE*I/O*RESPONSE'

C2795RTY='RAID RANK TYPE, THUS:'

C2795SR ='RAID rank FB sectors read'

C2795SW ='RAID rank FB sectors written'

C2795TSP='TRACKS TRANSFERRED TO PPRC VO'

C2795WRQ='RAID rank write requests'

C2795WRT='RAID RANK WRITE RESPONSE TIME'

C2795XCW='XRC or CC contaminated writes'

C2795XSF='XRC or CC sidefile read requests'

C279XRSV='LOWER*INTERFACE*I/O*RESPONSE'

Thanks to John Kington, Convergys, USA.

Thanks to Dr. H. Pat Artis, Performance Associates, USA.
Change 23.135 -Top Secret Version 5.2 and 5.3 are now supported, simply

VMAC80A by expanding the test for RACFVRSN to include 52x,53x.

May 26, 2005 -Top Secret creates its own user record with SMF80DTP=114,

but that record is really in SMF 80 format, so the user

SMF record can be processed with MXG TYPE80A code, using:

//SMF DD

//SYSIN DD *

%LET MACFILE= %QUOTE ( IF ID=231 THEN ID=80; ) ;

%INCLUDE SOURCLIB(TYPS80A);

to change the (example) ID=231 records back to ID=80 so

they are processed by TYPE80 code.

-The SMF80DTP=114 record is not yet decoded, so the above

code will only output the RACF header variables into the

TYPE80CM dataset; when documentation from CA is received,

the DTP=114 record will be decoded.

Thanks to Brent Turner, CitiGroup, USA.

Change 23.134 -RACF Extended TP2=301 with TOK80LN2=0 was not expected,

VMAC80A so it generated an error message on the SAS log stating

May 26, 2005 the record was deleted, but now, a valid TP2=301 segment

only TOK80FLG and TOKSUBSY='OMVS' is created, for NOOMVS

for RACFEVNT=13:ALTUSER record, although the record is

still does not agree with the IBM documentation:

MXG TOK80LN2 is byte 11 of the TP2=310 segment, which

is documented:

Byte 11: Length of subkeyword; 0 if byte 1 bit 1 is set

==> but byte 1 bit 1 is not set.

Byte 1 bit 4 "keyword has no subfield"

==> should be set, but it isn't.

-However, the documentation did allow the creation of two

new variables in datasets TYPE8009,8010,8012,8013:

KW301IGN='KEYWORD*IGNORED*INSUFFIENT*AUTHORITY?'

KW301DEL='SEGMENT*DELETED*VIA A*NOXXX*KEYWORD?'

so that you could detect the NOOMVS event.

Thanks to Erling Andersen, SMT, DENMARK.

Change 23.133 Support for Iway Software's IWAY (formerly EDA/SQL) SMF

EXIWAY1 record, creates these four datasets:

EXIWAY2 DDDDDD MXG MXG

EXIWAY4 DATASET DATASET DATASET

EXIWAY5 SUFFIX NAME LABEL

FORMATS

IMACIWAY IWAY1 IWAY01 IWAY SERVER START OF TASK

TYPEIWAY IWAY2 IWAY02 IWAY SERVER END TASK

TYPSIWAY IWAY4 IWAY04 IWAY SERVER BEGIN QUERY

VMACIWAY IWAY5 IWAY05 IWAY SERVER END QUERY

VMXGINIT

May 26, 2005

Thanks to Luis Mendoza, ABNAMRO, USA.
Change 23.132 Support for z/OS 1.7 (COMPATIBLE) and APAR OA10556.

VMAC1415 -TYPE70 dataset

VMAC7072 -New MACHTIME='MACHINE*TOD*DATETIME*STAMP' is created by

VMAC88 addition of new SMF70HOF='HYPERVISOR*DATETIME*OFFSET' to

VMAC94 SYNCTIME (SMF70IET + SMF70LGO). MACHTIME will be the

VMAC74 true GMT datetime value, independent of the IPL time and

May 21, 2005 site GMT offsets, and will be used by SCRT to know the

Oct 5, 2005 true time for IBM bills, since it is possible to IPL

Dec 12, 2005 with repeated/future/past datetimes, causing double

Oct 1, 2012 accounting for MSUs, etc. However, this value has not

yet been validated with SMF70LGO populated.

-SMF70HWM='CPC*PHYSICAL*MODEL*IDENTIFIER' is created, and

STFBIT04='SMF70MDL*MODEL*CAPACITY*ONLY?' flag, if true,