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

as well as for CPs.

Thanks to Don Deese, (CPExpert), Computer Management Sciences, USA.
Change 24.183 Variables Q3STHWIB, Q3STHWIF, and Q3STHWCT are high water

VMACDB2 mark values and should not have been de-accumulated.

Sep 16, 2006 Oct 5: Also, variable QDSTMIN2.

Oct 5, 2006

Thanks to Rachel Holt, Fidelity Systems, USA.

Thanks to Ralph Baechle, John Deere, USA.

Change 24.182 -NDMCPUTM (created from the text string CPUTIME=) had a

VMACNDM few small negatives, because the BY list was insufficient

Sep 12, 2006 to force the correct order for de-accumulation.

Sep 13, 2006 The time sequence within NDMPRCNO was different when

Sep 14, 2006 sorted by NDMTIME vs SMFTIME; each created a different

group of observations with negative NDMCPUTM. Using

the raw NDMCPUTM value in place of time of day appears

of have corrected the negative values; but check your

own data to be sure.

-Change 24.144 created new variable NDMCPU when the DSECT

showed a four-byte "CPU TIMEUSED" field added in NDM 4.3,

but it took us several iterations to INPUT the field from

the right place with the (undocumented) correct decimal,

in part because with NDM 4.5, the first-bit of NZMZFMT is

off, indicating an 8-byte UID, but the record has the

64-byte Expanded UID. I assume it is always present in

the current versions, so MXG now always INPUTs 64-bytes.

And one site's network group validated the MXG NDMCPU

value, so the NDMCPU variable may be valid. Sterling

says the field was populated in Version 4.4.

Thanks to Rodger Foreman, Acxiom, USA.

Thanks to David Kaplan, Depository Trust & Clearing Corporation, USA.

Thanks to Rob Hollingum, HSBC, ENGLAND.
Change 24.181 -Support for OPDG3 and SPGG3 RMF III segments required

ASMRMFV updates to ASMRMFV and revision to MXG code to create:

EXZRBOPD ddddd Dataset Descriptino0

EXZRBSPG ZRBSPG ZRBSPG RMFIII STORAGE GROUP AND VOLUME DATA

IMACRMFV ZRBOPD ZRBOPD RMFIII OMVS PROCESS DATA TABLE

VMACRMFV -Support for zIIP data added to ASIG3 segment:

VMXGINIT ASIMCDLY='MULTI STATE*PROCESSOR*DELAY*PCT'

Sep 12, 2006 ASIMCUSE='MULTI STATE*PROCESSOR*USING*PCT'

Sep 16, 2006 ASIPHTZA='PREEMPTABLE*SRB*FOR ZAAPS'

Sep 19, 2006 ASIPHTZI='PREEMPTABLE*SRB*FOR ZIIPS'

ASISDCCP='PCT*DELAYED*BY CP*PROCESSOR'

ASISDCSP='PCT SINGLE STATE*SAMPLES*DELAYED*BY ZIIP'

ASISUCCP='PCT SINGLE STATE*SAMPLES*USING*CP'

ASISUCSC='PCT SINGLE STATE*SAMPLES*USING*ZIIP*ON CP'

ASISUCSP='PCT SINGLE STATE*SAMPLES*USING*ZIIP'

ASITIIP ='ACCUMULATED*ZIIP*TIME'

ASITIIPC='ACCUMULATED*ZIIP*ON CP*TIME'

-Support for new zIIP/zAAP data in RCDG3 segment:

RCDHST ='HIPERSPACE*CPU*TIME'

RCDIFACP='ZAAP*SERVICE*UNITS*ON CP'

RCDIFASU='ZAAP*SERVICE*UNITS'

RCDIFAT ='ZAAP*SERVICE*TIME'

RCDIFCT ='ZAAP*SERVICE*TIME*ON CP'

RCDIIT ='IO*INTERRUPT*CPU*TIME'

RCDRCT ='REGION*CONTRAL*TASK*CPU*TIME'

RCDSUPCP='ZIIP*SERVICE*UNITS*ON CP'

RCDSUPSU='ZIIP*SERVICE*UNITS'

-Support for new zIIP/zAAP data in CPUG3 segment:

CPUIFCOL='ACCUM*ONLINE*ZAAPS'

CPUIFCON='ZAAPS*ONLINE*AT END'

CPULOGIF='ZAAP*LOGICAL*CPU*TIME'

CPULOGZI='ZIIP*LOGICAL*CPU*TIME'

CPUONTIF='ACCUM*ZAAP*ONLINE*TIME'

CPUONTZI='ACCUM*ZIIP*ONLINE*TIME'

CPUPHYIF='ZAAP*PHYSICAL*CPU*TIME'

CPUPHYZI='ZIIP*PHYSICAL*CPU*TIME'

CPUZICOL='ACCUM*ONLINE*ZIIPS'

CPUZICON='ZIIPS*ONLINE*AT END'

Thanks to Jerry Urbaniak, Acxiom CDC, USA.
Change 24.180 Labels for IFATYPnn now contain IFA, ZIP, and CP text.

VMAC7072 Labels for SMF70Q01-Q11 were misleading, implying the.

Sep 11, 2006 values were cumulative, but they are discrete percents

when In-Ready WAS N+1, rather than In-Ready LE N+1.

Thanks to Don Deese, (CPExpert), Computer Management Sciences, USA.
Change 24.179 Cosmetic. Labels for HSM datasets HSMWWFSR and HSMWWVOL

VMACHSM were not propagated in their _Sdddddd dataset sort macro.

Sep 11, 2006

Thanks to Christa Neven, KBC Bankverzekeringsholding, BELGIUM.

Change 24.178 Support for IRRHFSU unload utility; Unix System Services

EXRAC900 for z/OS file-permissions are a "IRRDBU00-format" RACF

EXRAC901 0900-0903 records. While these new records contain only

EXRAC902 ASCII data, and the IRRDBU00 records contain only EBCDIC,

EXRAC903 this implementation supports either file or concatenation

IMACRACF of both record types, and can be executed under ASCII or

VMACRACF EBCDIC SAS systems. New datasets created are these:

VMXGINIT dddddd Dataset Description

Sep 10, 2006 RAC900 RACF0900 USS RACF BASIC RECORD

Sep 26, 2006 RAC901 RACF0901 USS RACF FILE ACCESS

RAC902 RACF0902 USS RACF DEFAULT ACCESS

RAC903 RACF0903 USS RACF DIRECTORY DEFAULT ACCESS

-Sep 26 Variables RECNR/RECTYPE added to RAC09xx datasets.

Variables GRNAME GRMEMBAL INTRNVOL input length expanded.

Variable ATTRIBS created for RACF0200

Variable TYPE0901 rename GRPORUSR.

Thanks to Bill Arrowsmith, Euroclear, BELGIUM.

Thanks to Aimee Steel, Euroclear, BELGIUM.

Change 24.177 DB2 Statistics ID=100 SUBTYPE=0 caused INPUT EXCEEDED if

VMACDB2 there is more than one QLST segment, and LENQLST=0 in the

Sep 8, 2006 triplets, which is IBM's new way to read the length field

at the OFFQLST offset. But this record is in error; its

length field contains 176, when the actual segment length

is 178 bytes; this LENxxxx field does not contain itself,

but in all other LENxxxx=0 segments, the length at the

OFFxxxx contains the 2 byte of that field. Circumvention

code was added to protect the QLST segment for the

unexpected (incorrect?) length field value.

Thanks to Ray Dunn, CIGNA, USA.
Change 24.176 DB2 Statistics variable QISTRHIG was incorrectly DIF()'d;

VMACDB2 it is a maximum value, and is not accumulated, so that

Sep 8, 2006 variable is no longer deaccumulated.

Thanks to Steve Morris, State of Ohio BWC, USA.

Change 24.175 For PDB.ASUMTAPE with STATUS='TY21ONLY', the DSNAME field

ASUMTAPE should be blank, but it was populated with a DSNAME from

Sep 7, 2006 a prior job. Now, it will be blank as expected.

Thanks to Geoges Rondeau, Amicam, FRANCE.

Change 24.174 The new Release 4.3 fields, including NDMCPU and NDMRIP

VMACNDM were incorrectly INPUT due to undocumented alignment data

Sep 7, 2006 bytes in the record. See Change 24.182.

Thanks to Rob Hollingum, HSBC, ENGLAND.

Thanks to David Kaplan, Depository Trust & Clearing Corporation, USA.
Change 24.173 Documentation only. IBM VMA product incorrectly decoded

IMACACCT SMF ACCOUNTn fields that contained an underscore. APAR

Sep 6, 2006 OA17684 list the ACCOUNTn characters they consider valid:

Letters A thru Z, numbers 0 thru 9, space, period, dollar

sign, asterisk, dash, slash, comma, at-sign, pound-sign

(a/k/a hash mark), equal sign, and now, with that APAR,

an underscore.
Change 24.172 Using %UTILBLDP with EXPDBOUT= that has a %INCLUDE caused

UTILBLDP a syntax error if the output was directly executed; there

Sep 6, 2006 was no error with the output, so running it as a two-step

build-and-then-execute circumvented. See Change 24.190.

Thanks to Robert Carballo, Office Depot, USA.
Change 24.171 CopyCross+HSC caused MXGTMNT Tape Mount Monitor to stop

ASMHSCEX writing SMF records. Apparently, CopyCross alters the

Sep 4, 2006 JFCB, which we use to get the DSNAME of the tape mount,

and apparently the JFCB address in the TIOT is not valid;

MXGTMNT took five internal S0B0 ABENDS, the error from

the IBM service that reads data from SWA (IEFQMREQ) when

the data we've pointed to is not in SWA, and assumed we

had a real problem, and turned the monitor off.

We know that HSC mounts do not go thru the IBM Volume

Mount Exit; we assume in the HSC exit that if we didn't

see it in the IBM exit that it must be HSC controlled,

so we were seeing the CopyCross mounts in the HSC exit.

There doesn't seem to be a way to identify these mounts

as CopyCross, but "asmguy" as figured a way to obtain the

DSNAME so that the abend won't occur, and thus MXGTMNT

will now capture the CopyCross mounts thru HSC exit.

Thanks to Brian Felix, Wachovia Corporation, USA.
Change 24.170 A debugging PUT statement at line 664: IF SMF14TY NOT ...

VMAC1415 is deleted. It was also truncated and had no semi-colon,

Sep 1, 2006 which caused NO MATCHING IF error in TESTIBM in JCLTEST.

Thanks to Bernd Klawa, Stadtwerke Bielefeld, GERMANY.

Change 24.169 Format $MGTMDAC was not built, and $MGTIVSA was wrong;

FORMATS the semi-colon to end the $MGTIVSA format was missing.

Aug 31, 2006 Inserting the semicolon corrected both formats.

Thanks to Nick Johns, Sainsbury's Supermarkets LTD, ENGLAND.

Change 24.168 Variable CPUTYPE was blank in PDB.ASUMCEC (because the

VMXG70PR HOLD7CPT temp variable was not in the RETAIN statement).

Aug 31, 2006 Fortunately, variable CPFCNAME='2084-308' does have the

CPUTYPE in character, but CPUTYPE is now corrected.

Thanks to Curdin Salis Gross, Credit Suisse, SWITZERLAND.
Change 24.167 Support for MEMORY object with NRDATA=35 caused message

VMACNTSM KNOWN MEM OBJECT UNEXPECTED NRDATA/NRNAMES. DELETED

Aug 31, 2006 and MXG skipped the record. NRDATA=35 is now supported.

Thanks to Roger Zimmerman, Hewitt Associates, USA.

Change 24.166 Support for BMC Mainview for CICS Optional DB2 and CMR.

IMACICDA -The existing IMACICMR for the CMRDATA segment was updated

IMACICMR with four new pairs of counts/times variables for ADABAS,

IMACICMD DATACOM, IDMS, and MQSeries.

UTILEXCL -The optional CMRDB2 data enabled by BMC BBSAMP CMR$2MCTs

VMAC110 is supported in IMACICMD optional member.

Aug 31, 2006 -UTILEXCL was updated to support the CMRDB2 segment and

Nov 7, 2006 the eight new variables in the CMRDATA segment.

-IMACICDA, used only for non-UTILEXCL segment processing

was updated to add the include for IMACICMD after the

existing IMACICMR, an assumed order that may need to be

reversed, if you don't use UTILEXCL.

-Unrelatedly, variables in Change 24.140 were still in the

FORMAT statement, causing UNITIALIZED VARIABLE ARZGEOS

(harmless), but they are now removed in VMAC110.

Thanks to Jane Dickenson, Santander Produban UK, ENGLAND.

====== Changes thru 24.165 were in MXG 24.06 dated Aug 30, 2006=========
Change 24.165 Variables CPUIFATM and CPUZIPTM are added to TRNDSMFI.

TRNDSMFI

Aug 29, 2006

Thanks to Stan Dylnicki, Royal Bank of Canada, CANADA.

Change 24.164 IMACJBCK (the JOB-check exit for selection of SMF records

VMAC6 by JOB/READTIME/SYSTEM/etc.), for SMF 6 and 26 records,

VMAC26J2 was called before the JESNR had been INPUT. The %INCLUDE

VMAC26J3 was moved until after JESNR has been created, so it can

Aug 28, 2006 be used for selection, as documented in IMACJBCK.

Thanks to Kris Ferrier, State of Washington DIS, USA.

Change 24.163 Support for z9EC processors. MXG 23.09+ supported 64-bit

FORMATS z/OS on z9 and z9EC; this change is required ONLY if you

Aug 27, 2006 are using a 32-bit z/OS (See Change 24.110 for impact).

Thanks to Al Sherkow, I/S Management Strategies, USA.

Change 24.162 Support for NTSMF Version 3.0.0, new objects & variables.

EXNTARMA -Dataset SYSTEM new variable:

EXNTARMP RDYTHPER='READY*THREADS*PER*PROCESSOR'

EXNTASPA -Dataset MEMORY new variables:

EXNTASPN PCTAVLBY='PCT*AVAILABLE*BYTES'

EXNTASPS VTORATIO='V TO R*RATIO'

EXNTCCMQ -Dataset NETWINTR new variable

EXNTCIPS PCTNETBY='PCT*NETWORK*UTILIZATION'

EXNTIPV6 -Dataset DTSCPU new variables

EXNTNECD DTCPNCOS='SUPPORTED*CORES'

EXNTNECE DTCPNCOA='ACTIVE*CORES'

EXNTNECI -Dataset BLKBERRY has new instance variable, if there is

EXNTNECJ more than one blackberry server:

EXNTNECL BLKBINST='INSTANCE*NAME*FOR*ANTIGEN SCAN'

NOTE: SEE CHANGE 24.015 for INCOMPATIBILITY note.

EXNTNECM -Datasets MSQBUFMG and SQLBUFMG no longer populate the

EXNTNECR variable CACHSIZE, although it will continue to exist

EXNTNECS in MXG with a missing value.

EXNTNECT -Dataset MSQGENST and SQLGENST new variables:

EXNTPSPI ACTMPTBL='ACTIVE TEMP TABLES'

EXNTSAAL TMTACRDT='TEMP TABLES CREATION RATE'

EXNTSAJO LOGLCONN='LOGICAL CONNECTIONS'

EXNTSAJS TRANSACT='TRANSACTIONS'

EXNTSAST NATOMYRT='NON-ATOMIC YIELD RATE'

EXNTSQBA MARSDEAD='MARS DEADLOCKS'

EXNTSQBN HTTPAURQ='HTTP AUTHENTICATED REQUESTS'

EXNTSQBR SOAPMTRQ='SOAP EMPTY REQUESTS'

EXNTSQBS SOAPSQRQ='SOAP SQL REQUESTS'

EXNTSQCA SOAPMEIN='SOAP METHOD INVOCATIONS'

EXNTSQCL SOAPWSRQ='SOAP WSDL REQUESTS'

EXNTSQCT SOAPSEIR='SOAP SESSION INITIATE REQUESTS'

EXNTSQCY SOAPSETR='SOAP SESSION TERMINATE REQUESTS'

EXNTSQES PROCBLKD='PROCESSES BLOCKED'

EXNTSQPC TMTADEST='TEMP TABLES FOR DESTRUCTION'

EXNTSQSP EVNODLDR='EVENT NOTIFICATIONS DELAYED DROP'

EXNTSQSR TREVBIQU='TRACE EVENT NOTIFICATION QUEUE'

EXNTSQST SQTRPRLW='SQL TRACE IO PROVIDER LOCK WAITS'

EXNTSQWS -Dataset MSQLOCKS and SQLLOCK new variable:

EXNTTCV6 LOKTIMEO='LOCK*TIMEOUTS'

EXNTUDV6 -Dataset MSQLATCH and SQLLATCH new variables

IMACNTSM SUPRLACN='NUMBER OF*SUPERLATCHES'

VMACNTSM SUPRPRRT='SUPERLATCH*PROMOTIONS*PER SEC'

VMXGINIT SUPRDERT='SUPERLATCH*DEMOTIONS*PER SEC'

Aug 26, 2006 -Dataset MSQACCES and SQLACCES new variables

DEDRROST='DEFERRED DROPPED ROWSETS'

DRROCLRT='DROPPED ROWSET CLEANUPS/SEC'

DRROSKRT='DROPPED ROWSETS SKIPPED/SEC'

DEDRAUS ='DEFERRED DROPPED AUS'

AUCLUPRT='AU CLEANUPS/SEC'

AUCLBART='AU CLEANUP BATCHES/SEC'

FACLBART='FAILED AU CLEANUP BATCHES/SEC'

USTRPACO='USED TREE PAGE COOKIE'

FATRPACO='FAILED TREE PAGE COOKIE'

USLEPACO='USED LEAF PAGE COOKIE'

FALEPACO='FAILED LEAF PAGE COOKIE'

LOPRCRCN='LOBSS PROVIDER CREATE COUNT'

LOPRDECN='LOBSS PROVIDER DESTROY COUNT'

LOPRTRCN='LOBSS PROVIDER TRUNCATION COUNT'

LOBHCRCN='LOBHANDLE CREATE COUNT'

LOBHDECN='LOBHANDLE DESTROY COUNT'

BYLOCRCN='BY-REFERENCE LOB CREATE COUNT'

BYLOUSCN='BY-REFERENCE LOB USE COUNT'

PUOFROCN='COUNT PUSH OFF ROW'

PUINROCN='COUNT PULL IN ROW'

LOREAHCN='COUNT LOB READAHEAD'

-Dataset MSQSTATS and SQLSTATS new variables:

FRCPRMRT='FORCED*PARAMETERIZATIONS*PER SEC'

SQLATTRT='SQL*ATTENTION*RATE'

-And support for these 37 new objects:
DDDDDD DATASET OBJECT
NTARMA ASPNET ARMTECH APPLICATION

NTARMP ASPNETAP ARMTECH PROCESS

NTASPN ASPNET ASP.NET

NTASPA ASPNETAP ASP.NET APPLICATIONS

NTASPS ASPNETSS ASP.NETSTATESERVICE

NTCCMQ CCMSGQUE CCM MESSAGE QUEUE

NTCIPS CITRIXPS CITRIX METAFRAME PRESENTATION SERVER

NTNECD NETCLRDT .NET CLR DATA

NTNECI NETCLRIN .NET CLR INTEROP

NTNECJ NETCLRJI .NET CLR JIT

NTNECL NETCLRLO .NET CLR LOADING

NTNECT NETCLRLK .NET CLR LOCKSANDTHREADS

NTNECM NETCLRME .NET CLR MEMORY

NTNECS NETCLRSE .NET CLR SECURITY

NTNECE NETCLREX .NETCLREXCEPTIONS

NTNECR NETCLRRE .NETCLRREMOTING

NTIPV6 IPV6 NT IPV6

NTSAAL SAALERTS SQLAGENT:ALERTS

NTSAJS SAJOBSTP SQLAGENT:JOBSTEPS

NTSAJO SAJOBS SQLAGENT:JOBS

NTSAST SASTATS SQLAGENT:STATISTICS

NTSQBA SQLBRKAC SQLSERVER:BROKER ACTIVATION

NTSQBN SQLBUFND SQLSERVER:BUFFER NODE

NTSQBR SQLBRDBM SQLSERVER:BROKER/DBM TRANSPORT

NTSQBS SQLBRSTA SQLSERVER:BROKER STATISTICS

NTSQCA SQLCATMD SQLSERVER:CATALOG METADATA

NTSQCL SQLCLR SQLSERVER:CLR

NTSQCT SQLCURTO SQLSERVER:CURSOR MANAGER TOTAL

NTSQCY SQLCURTY SQLSERVER:CURSOR MANAGER BY TYPE

NTSQES SQLEXECS SQLSERVER:EXEC STATISTICS

NTSQPC SQLPLNCA SQLSERVER:PLAN CACHE

NTSQSR SQLSQLER SQLSERVER:SQL ERRORS

NTSQSP SQLSPIPE SQLSERVER:SSIS PIPELINE

NTSQST SQLTRANS SQLSERVER:TRANSACTIONS

NTSQWS SQLWAITS SQLSERVER:WAIT STATISTICS

NTTCV6 TCPV6 NT TCPV6

NTUDV6 UDPV6 NT UDPV6
-The _UNTDISC logic to recognize new DISCOVERY record was

revised to protect multiple systems (data with very old

and new NTSMF interleaved caused variable DISCOVRY to not

always be incremented; now, it will always be, but it can

skip a value, which is ok, as it is only for grouping).

-NTCONFIG dataset variable SUMRYVER could be incorrectly

carried forward if you have new and then old NTSMF data.

Thanks to Jim Quigley, CONED, USA.

Change 24.161 Support for i/Series QACS/QAPM AS/400 Release 5.4.0, is

VMACQACS INCOMPATIBLE, because when new data is added, the LRECL

Aug 26, 2006 in your JCL/FILENAME must be changed to read the new data

records. The comments in VMACQACS list the new LRECLs:

File LRECL Change

QAPMDISK 376 +3

QAPMJOBL 1116 +31

QAPMJOBM 542 +31

QAPMLPAR 172 +92

QAPMPOLB 83 +5

QAPMSYSL 3367 +23

QAPMSYST 555 +20

-Dataset QAPMCONF supports GKEY='21' to create new vars:

GDES21 ='GDES21*ASP*CAPACITY*IN*BYTES'

New GDKEYs 'B1' thru 'B5' create GDESB1N-GDESB5N numeric

and GDESB1C-GDESB5C character variables, awaiting doc to

properly label them.

-Dataset QAPMDISK new variables:

DSRDT ='RAID TYPE?*0=RAID 5*1=RAID 6'

DSIOPF ='MANAGED*BY IOP?*0=NO*1=YES'

DSCAT ='DISK*UNIT*CATEGORY?*0=NO*1=EXTERNAL'

-Dataset QAPMJOBM and QAPMJOBL new variables:

JBACPU ='ACCUMULATED*JOB CPU*TIME'

JBIPAF ='IP TYPE*02X=IPV4*18X=IPV6'

JBIPAD ='IP ADDRESS BINARY'

JBIPPT ='IP PORT*NUMBER'

-Dataset QAPMSYS and QAPMSYSL and QAPMSYST new variables:

SYVCPU ='VIRTUAL*PROCESSOR*TIME*CONFIGURED'

SYDPCH ='TOTAL*DISPATCH*TIME'

SYSHRF ='SHARED*PROCESSOR*FLAG 0=NO*1=YES'

-Dataset QAPMPOLB new variables:

POUNAL ='UNALLOCATED*POOL*SPACE'

-Dataset QAPMSYSL variables previously overlooked:

SYNUAL SYIFTA SYSPTU SYCTA SYUTA SYNUTC SYNPLU SYNPLA

-Dataset QAPMLPAR new variables decoded, but not yet

tested with data.

LPDDTM ='DATETIME*WHEN DISK*DATA WAS*COLLECTED'

LPCAP ='TOTAL*DISK*CAPACITY'

LPAVL ='TOTAL*DISK*CAPACITY*AVAILABLE'

LPBSY ='DISK*BUSY*TIME'

LPRSP ='DISK*RESPONSE*TIME'

LPRDS ='DISK*READ*COMMANDS'

LPWRTS ='DISK*WRITE*COMMANDS'

LPDISK ='NUMBER OF*SELECTED*DISKS'

LPMEM ='TOTAL*MEMORY*IN SYSTEM'

Thanks to Jim Wertenberger, Antares Management Solutions, USA.

Thanks to Tim Follen, Antares Management Solutions, USA.
Change 24.160 -NDMCPUTM could still be wrong, if the NDMLENPA/NDMLENSA

VMACNDM field lengths were more than the arbitrary $VARYING48 I

Aug 25, 2006 chose, thinking ACCT data would be an MVS 44-byte account

field, but the "ACCT" data is text. With lengths of 89

I arbitrarily increased the length to 256 bytes, but,

more importantly, longer lengths are protected.

-The INPUT for NDMCPUTM itself was revised to validate

that CPUTIME= text exists immediately prior to the INPUT

location.

-Code was revised to be independent of the order of the

four optional segments (SVOL,PVOL,SACCT,PACCT).

-NDMOFF43 and NDMLEN43 logic was removed, as the segment

is always present in current NDM-Connect Direct records.

-Some CT records contain nonzero NDMCTDOF (*P.DTOTAL) that

is the offset to an 80-byte "TOTALS (IF RESTARTED)" area,

from which I create NDMTOT01-NDMTOT20 variables, pending

finding documentation of what these totals are.

-Invalid date/times in 'MC' record is under investigation.

Thanks to David Kaplan, Depository Trust & Clearing Corporation, USA.
Change 24.159 Support added for CICS Statistics Intervals option values

VMXGCICI of TWOHOUR, FOURHOUR, and EIGHTHR.

Aug 25, 2006

Thanks to Tom Kelman, Commerce Bank of Kansas City, USA.

Change 24.158 Attempting to use MACRO _VTY30UV to DROP variables caused

EXPDB30V ERROR: VARIABLE EXCP2305 NOT FOUND; you cannot use that

Aug 23, 2006 macro to drop variables from PDB.SMFINTRV. Furthermore,

you must use BUILDPDB, BUILDPD3, or ONLINTV to create the

PDB.SMFINTRV dataset, as only those programs have logic

that combines the MULTIDD='Y' observations into a single

PDB.SMFINTRV observation; while %INCLUDE of TYPS30 does

create a dataset named PDB.SMFINTRV, that dataset will

still have multiple MULTIDD='Y' observations.

But, you can instead use the EXPDB30V exit member and use

a DROP statement to drop variables from the summarized

PDB.SMFINTRV dataset.

Thanks to Colin Bowen, CSC, SOUTH AFRICA.
Change 24.157 Analysis example to identify all jobs/STCs that allocated

ANALDEVN a DEVNR range or a DEVICE types, by creating only the

Aug 23, 2006 TYPE30_D dataset for selected DEVNR/DEVICEes.

Thanks to Yaohua Hu, ISO, USA.

Change 24.156 Corrections to reported ANALDB2 errors/omissions:

ANALDB2R -"STMT #: 4040 was printed instead of the correct 16448

Aug 21, 2006 for the statement number.

Thanks to ???, ???, ???

Change 24.155 The CICSBAD dataset added by Change 23.312 could contain

EXCICBAD legitimate transactions; the test for PROGRAM='########'

Aug 21, 2006 previously identified invalid CICS transaction names that

Sep 27, 2007 did not have a program name (i.e., typo'd TRANNAME).

However, some sites have chosen to use ######## to make

their transaction routing easier, which caused millions

of transactions to be output to CICSBAD, when this site

wanted them output to CICSTRAN. MXG's test for EXCICBAD

IF PROGRAM='########' OR

SUBSTR(TRANFLAG,6,1)='......1.'B THEN DO;

will %INCLUDE the EXCICBAD exit, so you can tailor the

code in your EXCICBAD member to decide which, if any,

transactions are output to CICSBAD or to CICSTRAN.

In this particular case, the site observed that all of

their TOR regions' transactions had PROGRAM='########',

but they also all had RSYSID NE '00000000'X, so the site

could use this logic in their EXCICBAD tailoring member:

IF PROGRAM='########' AND RSYSID NE '00000000'X THEN DO;

OUTPUT _WCICTRN;

END;

ELSE DO;

OUTPUT _WCICBAD;

END;

to output real bad to CICSBAD and the rest to CICSTRAN.

Sep 27, 2007:

-And if you DON'T want CICSBAD to be created, but instead

want these "bad" transactions output in CICSTRAN, you can

copy EXCICBAD into your tailoring library and change

_WCICBAD to _WCICTRN.

-And what about that bit test that sends transactions to

CICSBAD instead of CICSTRAN, that test for:

SUBSTR(TRANFLAG,6,1)='......1.'B THEN DO;

It was obscurely documented back in Change xx.yyy:

"When a CICS task is executing on an OPEN TCB, and is

then purged, APAR PQ86971 documents that all of the