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

Apr 29, 2015 by dubious programming techniques where a task has

May 1, 2015 modified storage owned by another task. These ABENDs

will need application programmer design changes; the APP

should have been fixed long ago - IBM is looking forward!

-New variables (INCOMPATIBLE) inserted in CICSTRAN dataset

NCGETCN ='NAMED*COUNTER*SERVER*REQUESTS'

TSGETSCN='SHARED*TEMP*STORAGE*GETS'

TSPUTSCN='SHARED*TEMP*STORAGE*PUTS'

May 1: Original Note wrong here, NO ERROR in TSTOTCN.

-New variables inserted in CICSRDQU dataset (MNSEGCL=5):

TSQGESTM='SHARED*TSQUEUE*GET*TIME'

TSQGESCN='SHARED*TSQUEUE*GET*COUNT'

TSQPUSTM='SHARED*TSQUEUE*PUT*TIME'

TSQPUSCN='SHARED*TSQUEUE*PUT*COUNT'

TSQGESBY='SHARED*TSQUEUE*GET*BYTES'

TSQPUSBY='SHARED*TSQUEUE*PUT*BYTES'

-"WRITING OF PERFORMANCE DATA IS MORE EFFICIENT WHEN

FIELDS ARE NOT EXCLUDED." was noted in a comment.
CHANGE 33.111 -The ANALID report that tabulates SMF record types is

ANALID enhanced to optionally invoke %UTILBLDP to create MXG

SMFSRCH code to build all possible MXG datasets from ALL records

UTILBLDP in the input SMF file, and to optionally run the program,

VMXGPRAL -This change originally documented macro variables UTILBLD

VMXGINIT VMXGPRA but they were removed in Change 33.117, unused.

Apr 27, 2015 -VMXGPRAL revised to print the LIB.DSNAME + LABEL without

May 27, 2015. extra blanks and periods.

Thanks to MP Welch, Bank of America, USA.
CHANGE 33.110 -Support for TS7700 HYDRA Version 3.2A (COMPATIBLE)

EXBVR302 creates new BVIR302 dataset with the Extended HSM Cache

FORMATS Partition Container Statistics from subtype 30 record.

IMACBVIR -Variables added to BVIR01 dataset:

VMACBVIR THRUPUT AHEADCNT BEHNDCNT

VMXGINIT -Changed records have BVIRVERS=6.

Apr 24, 2015

Thanks to Tim Campbell, TELUS, CANADA.

CHANGE 33.109 -Support for TYPE 23 zEDC variables SMF23BBC/SMF23BAC has

VMAC23 been corrected and validated. Those two variables plus

Apr 23, 2015 SMF23PFT SMF23PFM SMF23PFH SMF23CWN SMF23NCN are now

correctly aligned and are now output in dataset TYPE23LS

and removed from dataset TYPE23.

-Two new fields were documented in the z/OS 2.1 SMF Manual

but IBM has confirmed they only exist in z/OS 2.2:

SMF23BBC='ZEDC*UNCOMPRESSEDSED*BYTES*TOTAL'

SMF23BAC='ZEDC*COMPRESSED*BYTES*TOTAL'

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

CHANGE 33.108 Support for APARs OA45944 and OA45897 which adds new data

VMAC42 for contention counts for record locks, component1 and

Apr 23, 2015 component2 locks:

-New variables in dataset TYPE42S1:

SMF42FUA ='ACCUMULATION*WAITERS FOR*RECORD LOCK'

SMF42FUB ='RECORD LOCKS*HASHED TO*SAME ENTRY(/

SMF42FUD ='ACCUMULATION*WAITERS*FOR DIWA*LOCK'

SMF42FUE ='ACCUMULATION*WAITERS*FOR UPGRADE*LOCK'

SMF42FUF ='ACCUMULATION*WAITERS*FOR COMP2*LOCK'

SMF42FUG ='LOCKS*HASHED TO*SAME*ENTRY'

-New variables in dataset TYPE42S2:

SMF42FVA ='ACCUMLATION*WAITERS*FOR RECORD*LOCK'

SMF42FVB ='RECORD LOCKS*HASHED TO*SAME ENTRY'

SMF42FVD ='ACCUMULATION*WAITERS*FOR DIWA*LOCK'

SMF42FVE ='ACCUMULATION*WAITERS*FOR UPGRADELOCK'

SMF42FVF ='ACCUMULATION*WAITERS*FOR COMP2*LOCK'

SMF42FVG ='LOCKS*HASHED TO*SAME*ENTRY'

-New variables in dataset TYPE42S3:

SMFA2FUA ='ACCUMULTION*WAITERS*FOR RECORD*LOCK'

SMFA2FUB ='LOCKS*HASHED*TO SAME*ENTRY'

SMFA2FOD ='ACCUMULTION*WAITERS*FOR DIWA*LOCK'

SMFA2FOE ='ACCUMULTION*WAITERS*FOR UPGRADE*LOCK'

SMFA2FOF ='ACCUMULTION*WAITERS*FOR COMP2*LOCK'

SMFA2FOG ='LOCKS*HASHED*TO SAME*ENTRY'

-New variables in dataset TYPE42S4:

SMFA2FVA ='ACCUMULATION*WAITERS*FOR RECORD*LOCK'

SMFA2FVB ='LOCKS*HASHED*TO SAME*ENTRY'

SMFA2FVD ='ACCUMULATION*WAITERS*FOR DIWA*LOCK'

SMFA2FVE ='ACCUMULATION*WAITERS*FOR UPGRADE*LOCK'

SMFA2FVF ='ACCUMULATION*WAITERS*FOR COMP2*LOCK'

SMFA2FVG ='LOCKS*HASHED*TO SAME*ENTRY'

-New variables in dataset TYPE42D1:

SMF42GUA ='ACCUMULATION*WAITERS*FOR RECORD*LOCK'

SMF42GUB ='LOCKS*HASHED*TO SAME*ENTRY'

SMF42GUD ='ACCUMULATION*WAITERS*FOR DIWA*LOCK'

SMF42GUE ='ACCUMULATION*WAITERS*FOR UPGRADE*LOCK'

SMF42GUF ='ACCUMULATION*WAITERS*FOR COMP2*LOCK'

SMF42GUG ='LOCKS*HASHED*TO SAME*ENTRY'

-New variables in dataset TYPE42D2:

SMF42GVA ='ACCUMULATION*WAITERS*FOR RECORD*LOCK'

SMF42GVB ='LOCKS*HASHED*TO SAME*ENTRY'

SMF42GVD ='ACCUMULATION*WAITERS*FOR DIWA*LOCK'

SMF42GVE ='ACCUMULATION*WAITERS*FOR UPGRADE*LOCK'

SMF42GVF ='ACCUMULATION*WAITERS*FOR COMP2*LOCK'

SMF42GVG ='LOCKS*HASHED*TO SAME*ENTRY'

-New variables in dataset TYPE42D3:

SMF42GUA ='ACCUMULATION*WAITERS*FOR RECORD*LOCK'

SMF42GUB ='LOCKS*HASHED*TO SAME*ENTRY'

SMF42GUD ='ACCUMULATION*WAITERS*FOR DIWA*LOCK'

SMF42GUE ='ACCUMULATION*WAITERS*FOR UPGRADE*LOCK'

SMF42GUF ='ACCUMULATION*WAITERS*FOR COMP2*LOCK'

SMF42GUG ='LOCKS*HASHED*TO SAME*ENTRY'

-New variables in dataset TYPE42D4:

SMF42GVA ='ACCUMULATION*WAITERS*FOR RECORD*LOCK'

SMF42GVB ='LOCKS*HASHED*TO SAME*ENTRY'

SMF42GVD ='ACCUMULATION*WAITERS*FOR DIWA*LOCK'

SMF42GVE ='ACCUMULATION*WAITERS*FOR UPGRADE*LOCK'

SMF42GVF ='ACCUMULATION*WAITERS*FOR COMP2*LOCK'

SMF42GVG ='LOCKS*HASHED*TO SAME*ENTRY'

-New variables in dataset TYPE42L1:

SMF42HUA ='ACCUMULATION*WAITERS*FOR RECORD*LOCK'

SMF42HUB ='LOCKS*HASHED*TO SAME*ENTRY'

SMF42HUD ='ACCUMULATION*WAITERS*FOR DIWA2*LOCK'

SMF42HUE ='ACCUMULATION*WAITERS*FOR UPGRADE*LOCK'

SMF42HUF ='ACCUMULATION*WAITERS*FOR COMP2*LOCK'

SMF42HUG ='LOCKS*HASHED*TO SAME*ENTRY'

-New variables in dataset TYPE42L2:

SMF42HVA ='ACCUMULATION*WAITERS*FOR RECORD*LOCK'

SMF42HVB ='LOCKS*HASHED*TO SAME*ENTRY'

SMF42HVC ='ACCUMULATION*WAITERS*FOR DIWA2*LOCK'

SMF42HVD ='ACCUMULATION*WAITERS*FOR UPGRADE*LOCK'

SMF42HVF ='ACCUMULATION*WAITERS*FOR COMP2*LOCK'

SMF42HVG ='LOCKS*HASHED*TO SAME*ENTRY'
CHANGE 33.107 -DB2PM-like reports PMACC02/4 class 3 delay section was

ANALDB2R missing TCP/IP LOB and the global contention corrected

Apr 26, 2015 from only parent locks to include all contention delays.

-In PMAUD02 there were uninitialized variables corrected.

-The z/OS external sort can not be used if the length of

an observation is greater than 32760 bytes, and many of

the DB2 102 records contain 32K SQL Text variables.

For the AUDIT reports, the SQL text is reduced to 80

bytes, since that is all IBM's DB2PM reports, and the

full text is still in the T102Snnn dataset.

-If you asked for AUTHCHG in the AUDIT= parameter and

there were any IFCID 83 records the selection criteria

were not applied and that caused missing timestamps in

the report headers.

-If there were 0 OBS in the IFCID 142 dataset then a

a work file was not cleaned up.

Thanks to Neil Ervin, Wells Fargo, USA
Change 33.106 New variable EXCMTYPE in CICSEXCE (Exception) dataset

FORMATS is created to describe the exception event:

VMAC110 1='DATABASE REQUEST POLICY RULE'

Apr 22, 2015 2='FILE REQUEST'

3='PROGRAM REQUEST'

4='START REQUEST'

5='STORAGE REQUEST'

6='SYNCPOINT REQUEST'

7='TD QUEUE REQUEST'

8='TIME REQUEST'

9='TS QUEUE REQUEST'

10='WAIT FOR CF LOCKING REQUEST'

11='WAIT FOR CF NON-LOCKING REQUEST'

12='WAIT FOR UDSA'

13='WAIT FOR EUDSA'

14='WAIT FOR CDSA'

15='WAIT FOR ECDSA'

16='WAIT FOR SDSA'

17='WAIT FOR ESDSA'

18='WAIT FOR RDSA'

19='WAIT FOR ERDSA'

20='WAIT FOR GCDSA'

21='WAIT FOR GUDSA'

22='WAIT FOR GSDSA'

23='WAIT FOR TEMPORARY STORAGE'

24='WAIT FOR STRING FOR FILE'

25='WAIT FOR STRING FOR LSRPOOL'

26='WAIT FOR STRING FOR DFHTEMP'

27='WAIT FOR BUFFER FOR LSRPOOL'

28='WAIT FOR BUFFER*FOR DFHTEMP'

Thanks to Bahman Nejad, Union Bank, USA.

Thanks to Perry Lim, Union Bank, USA.

Change 33.105 Using PROC CPORT to copy z/OS data libraries to ASCII can

IMACFMTS corrupt values in character variables with $HEX format if

UTILCVRT those variables did not have the TRANSCODE=NO attribute

Apr 21, 2015 set, or they were built with an ancient SAS version prior

to that attribute. If you have downloaded PDBs and see

$HEX variables are wrong (check CPUTYPE in TYPE70), the

UTILCVRT program can be used to convert those $HEX vars

back to their correct values. Maybe. The default mapping

table is the $MGAS2EB format in FORMATS, using a table

created in 2009 for PROC CPORT translation, replacing an

earlier table created from IND$FILE. However, now using

the CPORT table under both Windows and Linux does NOT map

characters correctly; the alternative $MGAS2EB table from

IND$FILE must be used. That alternative format is defined

in Example 2 in a comment block in the IMACFMTS member.

Copy IMACFMTS to your USERID.SOURCLIB, remove the comment

block in Example 2, and run the FORMATS step of JCLINSTL

(%INCLUDE SOURCLIB(FORMATS); with DISP=OLD on //LIBRARY)

and the alternative table values will be used.

Thanks to Warren Cravey, Fidelity, USA.

Thanks to Rachel Holt, Fidelity, USA.
Change 33.104 NMON CPU records can now contain three digits, CPU001, so

VMACNMON the logic to create CPUNR now decodes two or three digits

Apr 21, 2015 for CPUnnn, PCPUnnn and SCPUnnn.

Thanks to Raissa Moussu, METROLOGFIE AIX EN PROVENCE, FRANCE.

Change 33.103 VMXGPRAL (PRINT/MEAN/FREQ ALL datasets &ALL variables in

VMXGPRAL a "PDB" data library, with both Label and Variable Name)

Apr 15, 2015 enhanced for FREQ with TABLE= option to choose what

variables are tabulated. TABLES=_CHAR_, will only show

the character variables. Without the TABLES= argument,

all variables will be tabulated.

Thanks to MP Welch, Bank of America, USA.
Change 33.102A Support for SYNCSORT Release 2.1 (INCOMPATIBLE, test for

VMACSYNC SYNCLVL for ancient Release 3.7 conflicted) causing many

Apr 15, 2015 variables to be missing values with no error message,

notably, CPUZIPTM is a missing value without this change.

Thanks to Roger Foreman, Transunion, USA.
Change 33.102 DB2 V10 INVALID DATA FOR QLACCPUR and QLACDBAT when field

VMACDB2 QLACLOCN was relocated because it's longer than 8 bytes,

Apr 14, 2015 but the length of the relocated field in XLENLOCN is NOT

in LENLOCN causing misalignment when XLENLOCN GT 20.

Thanks to Betty Wong, Bank of America, USA.
Change 33.101 Support for iSeries 7.2 (COMPATIBLE with new LRECLS).

VMACQACS Comments in VMACQACL list LRECLs and 7.2 records tested.

Apr 13, 2015 -New variables in QAPMDISK:

DSLVLMP ='LEVEL OF*MIRRORED*PROTECTION'

DSLSCNDS='LOG*SENSE*COMMANDS*ISSUED'

DSLSRT ='LOG*SENSE*RESPONSE*TIME'

-New variables in QAPMJOBM (QAPMJOBMI):

JBTNAME ='THREAD*NAME'

JBSLTCNT ='SHORT*LIFESPAN*ENTRY*COUNT'

JBSACPU ='ACCUMULATED*JOB*SCALED*CPU TIME'

JBINDCPU ='THREAD*UNSCALED*CPU TIME*USED'

JBSINDCPU ='THREAD*SCALED*CPU TIME*USED'

JBCPUWC ='PROCESSOR*ELAPSED*TIME'

JBVPDLY ='VIRTUAL*PROCESSOR*DELAY*TIME'

JBSEIZECNT ='SIEZE*COUNT'

JBPSLCKCNT ='PROCESS*SCOPED*LOCK*COUNT'

JBTSLCKCNT ='THREAD*SCOPED*LOCK*COUNT'

JBPSRCDLCK ='PROCESS*SCOPED*DATABASE*LOCK*COUNT'

JBTSRCDLCK ='THREAD*SCOPED*DATABASE*LOCK*COUNT'

JBNFOGDT ='OFF-GROUP*DISPATCH*TIME'

JBNFOGMA ='OFF-GROUP*PAGE*FRAMES'

JBPGEZSTL ='PAGES*MARKED*EASY TO*STEAL'

JBSQLCLK ='SQL*CLOCK*TIME'

JBSQLCPU ='THREAD*UNSCALED*SQL*TIME'

JBSQSCPU ='THREAD*SCALED*SQL*CPU TIME'

JBSQLDBR ='SQL*SYNCHRONOUS*DATABASE*READS'

JBSQLNDBR ='SQL*SYNCHRONOUS*NONDATABASE*READS'

JBSQLDBW ='SQL*SYNCHRONOUS*DATABASE*READS'

JBSQLNDBW ='SQL*SYNCHRONOUS*NONDATABASE*READS'

JBSQLADBR ='SQL*ASYNCHRONOUS*DATABASE*READS'

JBSQLANDBR ='SQL*ASYNCHRONOUS*NONDATABASE*READS'

JBSQLADBW ='SQL*ASYNCHRONOUS*DATABASE*WRITES'

JBSQLANDBW ='SQL*ASYNCHRONOUS*NONDATABASE*WRITES'

JBAGCRT ='ACTIVATION*GROUPS*CREATED'

JBPGMSACT ='PROGRAM*ACTIVATIONS*CREATED'

JBCURTMP ='CURRENT*TEMPORARY*STORAGE'

JBPEAKTMP ='PEAK*TEMPORARY*STORAGE'

JBMAXTMP ='MAXIMUM*TEMPORARY*STORAGE'

JBTMPPGA ='4K UNITS*ALLOC*TEMPORARY*STORAGE'

JBTMPPGD ='4K UNITS*DEALLOC*TEMPORARY*STORAGE'

JBHSQLSTMT ='HIGH*LEVEL*SQL*STATEMENTS'

JBTICC ='THREAD*INSTRUCTION*COUNT*CHARGED'

JBTICU ='THREAD*INSTRUCTION*COUNT*USED'

JBTTMBU ='THREAD*VIRTUAL*TIME'

JBPICC ='PROCESS*HDW INSTR*COUNT*CHARGED'

JBPRRSCPTY ='PROCESSOR*RESOURCE*PRIORITY'

-New variables in QAPMJOBO (QAPMJOBOS):

JBFSOPN 'FILE*SYSTEM*OPENS''

JBFSDC 'FILE*SYSTEM*DIRECTORY*CREATES'

JBFSNDC 'FILE*SYSTEM*NON-DIRECTORY*CREATES*

JBFSDD 'FILE*SYSTEM*DIRECTORY*DELETES'

JBFSNDD 'FILE*SYSTEM*NON-DIRECTORY*DELETES*

JBSPLFC 'SPOOLED*FILES*CREATED*BY THIS JOB*

JBSBMJOBS'JOBS*SUBMITTED*OR SPAWNED.'

JBSQLSTMT'SQL*STATEMENTS'

JBLWTSQL 'SQL*RELATED*DATABASE*WRITES*(LOGIC

JBLRDSQL 'SQL*RELATED*DATABASE*READS*(LOGICA

JBDBUSQL 'SQL*RELATED*DATABASE*MISCELLANEOUS

JBPASCMP 'SQL*PAS*COMPRESSIONS*ALL THREADS'

JBPKGCMP 'SQLPKG*COMPRESSIONS'

JBCTHD 'CONNECTED*THREAD*IDENTIFIER'

JBCJNAM 'CONNECTED*JOB*NAME'

JBCJUSR 'CONNECTED*JOB*USER'

JBCJNBR 'CONNECTED*JOB*NUMBER'

-New variables in QAPMPOLB (QAPMPOOLB):

POPGS4 ='POOL*4K*PAGES'

POPGS64 ='POOL*64K*PAGES'

PODBF4 ='POOL*DATABASE*4K*PAGE*FAULTS'

PODBF64 ='POOL*DATABASE*64K*PAGE*FAULTS'

PONDBF4 ='POOL*NONDATABASE*4K*PAGE*FAULTS'

PONDBF64 ='POOL*NONDATABASE*64K*PAGE*FAULTS'

PODBPG4 ='POOL*DATABASE*4K*PAGES*READ'

PODBPG64 ='POOL*DATABASE*64K*PAGES*READ'

PONDPG4 ='POOL*NONDATABASE*4K*PAGES*READ'

PONDPG64 ='POOL*NONDATABASE*64K*PAGES*READ'

POUNAL4 ='UNALLOCATED*4K*PAGES'

POUNAL64 ='UNALLOCATED*64K*PAGES'

POAGED4 ='POOL*4K*PAGES*AGED/

POAGED64 ='POOL*64K*PAGES*AGED'

POSTLN4 ='POOL*4K*PAGES*STOLEN'

POSTLN64 ='POOL*64K*PAGES*STOLEN'

POUNUSE4 ='POOL*4K*PAGES*UNUSED'

POUNUSE64 ='POOL*64K*PAGES*UNUSED'

POSYNC4 ='POOL*4K*PAGE*SYNC I/O'

POSYNC64 ='POOL*64K*PAGE*SYNC I/O'

POASYNC4 ='POOL*4K*PAGE*ASYNC I/O'

POASYNC64 ='POOL*64K*PAGE*ASYNC I/O'

POPGOUT4 ='POOL*4K*PAGE*OUTS'

POPGOUT64 ='POOL*64K*PAGE*OUTS'

POPGABLE4 ='POOL*4K*PAGES*PAGEABLE'

POPGABLE64='POOL*64K*PAGES*PAGEABLE'

POATMPT4 ='POOL*4K*ALLOCATION*ATTEMPTS

POATMPT64 ='POOL*64K*ALLOCATION*ATTEMPTS

POAPS4 ='POOL*4K*AFFINITY*SUCCESSES

POAPS64 ='POOL*64K*AFFINITY*SUCCESSES

POAPMIG4 ='POOL*4K*AFFINITY*MISSES

POAPMIG64 ='POOL*64K*AFFINITY*MISSES

POAPMOG4 ='POOL*4K*AFFINITY*MISSES*OFFGROUP'

POAPMOG64 ='POOL*64K*AFFINITY*MISSES*OFFGROUP'

POCRTPG64 ='POOL*4K*PAGES*CREATED'

POBRKPG64 ='POOL*64K*PAGES*BROKEN*UP'

POPOSIZB ='POOL*SIZE*KB'

POUNALB ='UNALLOCATED*POOL*SPACE*KB'

-New variables in QAPMSYST (QAPMSYSTEM):

SYVPID ='VIRTUAL*SHARED*POOL ID'

SYVPCAP ='VIRTUAL*SHARED POOL*ENTITLED*CAPACITY'

SYPPLU ='PHYSICAL*SHARED POOL*CPU TIME*USED'

SYPPLA ='PHYSICAL*SHARED POOL*CPU TIME*AVAILABLE'

SYPTHV ='HYPERVISOR*CPU*TIME'

SYPTINT ='INTERRUPT*CPU*TIME'

SYPTWS ='WAITTASK*TIME'

SYPTDN ='DONATED*CPU*TIME'

SYSSPTU ='SCALED*CPU*TIME*USED'

SYUCAPF ='PARTITION*UNCAPPED*FLAG'

SYDONF ='PARTITION*DONATION*FLAG'

SYPTWAIT ='VIRTUAL*PROCESSOR*THREAD WAIT*EVENT TIME'

SYPTREADY ='VIRTUAL*PROCESSOR*THREAD WAIT*READY TIME'

SYPTLATEN ='VIRTUAL*PROCESSOR*THREAD DISPATCH*LATENCY'

SYPTACT ='VIRTUAL*PROCESSOR*THREAD*ACTIVE TIME'

SYPTIDLE ='VIRTUAL*PROCESSOR*THREAD*IDLE TIME'

SYPTINTR ='VIRTUAL*PROCESSOR*THREAD*INTERRUPT TIME'

SYFRMCPU ='PROCESSOR*FIRMWARE*TIME USED'

SYFRMSCPU ='PROCESSOR*SCALED*FIRMWARE*TIME USED'

SYPFOLDSW ='PROCESSOR*FOLDING*SWITCH*STATE'

SYPFOLDST ='PROCESSOR*FOLDING*STATE'

SYEMMAJCDE='ENERGY*MANAGEMENT*MAJOR CODE'

SYEMMINCDE='ENERGY*MANAGEMENT*MINOR CODE'

SYEMATTR ='ENERGY*MANAGEMENT*ATTRIBUTES'

SYEMPWRLMT='ENERGY*MANAGEMENT*POWER DRAW*LIMIT WATTS'

SYSQLCPU ='UNSCALED*SQL*CPU TIME*USED'

SYSQLSCPU ='SCALED*SQL*CPU TIME*USED'

SYOSTMP ='CURRENT*TEMPORARY*STORAGE*ALLOCATED*NONDB'

SYDBTMP ='CURRENT*TEMPORARY*STORAGE*ALLOCATED*DB'

SYAJOBTMP ='CURRENT*TEMPORARY*STORAGE*CHARGED*ACTIVE'

SYEJOBTMP ='CURRENT*TEMPORARY*STORAGE*CHARGED*EMDED'

SYUSERTMP ='CURRENT*USER*TEMPORARY*STORAGE'

SYPSLPU ='PHYSICAL*SHARED POOL*SCALED*CPU TIME*USED'

SYTRUNIC ='HARDWARE*INSTRUCTIONS'

SYTRUNVTB ='NON-IDLE*PROCESSOR*VIRTUAL*TIME'

SYTITUIC ='INTERRUPT*INSTRUCTION*COUNT'

SYTFRMIC ='FIRMWARE*INSTRUCTION*COUNT'

Thanks to Raymond J. Smith, United Health Group, USA.

Change 33.100 -Protect RMF III for invalid ASI table index in UWD table

ASMRMFV entry that caused ABEND S0C4 in FINDAS when the invalid

Apr 12, 2015 index value was really bad, i.e., it exceeded the number

of entries in the ASI table and an attempt was made to

access data from well beyond the ASI table's end.

Thanks to Randy Schlueter, First Data Corporation, USA

Change 33.099 MXG 32.13-MXG 33.03. TMON/CICS Version 3.4 AND EARLIER.

VMACTMO2 -Dataset MONITASK CPU time (TASCPUTM) and ALL DURATION

Apr 11, 2015 variables are WRONG, TOO SMALL, by a factor of 4096.

In TMON Version 4, all durations were changed from the

original microseconds to TODSTAMP, which requires the

division by 4096, but the MXG code inserted for MONITASK

was mis-located and applied to all TMON versions.

-Less critical, and wrong ONLY with Version 4 records: ALL

duration variables in all OTHER MONIxxxx datasets were

NOT divided by 4096, so they are TOO LARGE by 4096.

Thanks to Andrew Petersen, CSC, AUSTRALIA.
Change 33.098 NMON (AIX/LINUX) CPU variables in PDB.NMONCPUD and in the

VMACNMON PDB.NMONINTV were revised and missing values corrected.

Apr 10, 2015 The individual CPUNR data in PDB.NMONCPUD from the three

CPUnn, PCPUnn, and SCPUnn records are the variables

prefixed with CPUNxxxx,PCPUNxxxx,SCPUNxxxx with suffixes

BUSY,IDLE,SYS,USER,WAIT.

The interval CPU_ALL, PCPU_ALL, and SCPU_ALL record's

in PDB.NMONINTV are prefixed PCPUxxxx,APCPUxxxx,ASCPUxxxx

with the same suffixes (plus variable APCPUENTCAP).

The BUSY variables are calculated for all observations.

Thanks to Florent Boulesteix, INOVANS partenaire CAAGIS, FRANCE.
Change 33.097 -SMFSRCH utility (select SMF records containing text and

SMFSRCH create all MXG datasets from those selected records) adds

Apr 9, 2015 new PRINT= option with YES/NO/nnn to print all, none, or

the first nnn observations of each dataset. The PRINTIT=

argument's YES or NO are still supported, but the PRINT=

argument is the more common spelling in MXG macros.

-Set USERADDS=NOUSERID, and all matching SMF types are

reported, but the 128-255 USER SMF records are skipped,

This avoids an abend if not all of your USER records are

mapped in IMACKEEP, MACKEEP, OR USERADDS= argument.

Thanks to MP Welch, Bank of America, USA.
Change 33.096 -z13 with SMT PROCVIEW=CORE, dataset TYPE70PR, variable

VMAC7072 SMT_NUM, the count of threads for each core, wasn't kept.

Apr 12, 2015 -TYPE70PR variable SMF70MTTT, MULTI-THREADING*IDLE*TIME,

which is actually SMF70MTIT mis-spelled, is accumulated,

requiring an additional sort and DIF() to deaccumulate.

As is ALWAYS the case when IBM writes ACCUMULATED values,

the value for the first instance of each LCPUADDR in each

LPAR will be a missing value. Some cases when the next

accumulated value was slightly smaller than the prior are

also set to missing value, but if they occur in your data

a PMR should be raised with IBM support. The label for

SMF70MTTT now contains 'SMF70MTIT' to compensate.

Thanks to Don Deese, Computer Management Sciences, USA.
Change 33.095 Optional SUMMARY=SMFBYTES/SMFRECNT argument will produce

ANALID a report sorted by DESCENDING &SUMMARY to print largest

Apr 3, 2015 first.
Change 33.094 Format MGSMFID did not describe SMF 115 records subtypes

FORMATS 5, 6, and 7.

Apr 3, 2015

Thanks to Paul Bennett, Euroclear, BELGIUM.

Thanks to Edmond Dierickx, Euroclear, BELGIUM.
Change 33.093 MXG 33.03. Disabled debug statement *PUTLOG +2 'DEBUG..;

VMAC7072 was followed by an unmatched */, but there was no syntax

Apr 3, 2015 error, because SAS recognized the asterisk as the start

of another comment, that ended with the semicolon of the

next statement, HOLICFTM=ICFACTTM;, so that statement was

absorbed and not executed. Fortunately, only ICF time in

some unusual circumstances, when the time had to be held,

might have been impacted.

Thanks to MP Welch, Bank of America, USA.
Change 33.092 -Some BBBPnnn variables were mis-assigned values: BBBP040

VMACNMON and BBBP049 because the BBBP40 'MAXIMUM CAPACITY' text

Apr 8, 2015 overlapped with BBBP49 'MAXIMUM CAPACITY OF POOL'. All

text tests are expanded to prevent the overlap.

-The MERGE of CPUD/CPUNP/CPUNS datasets had DUPLICATE

MERGE VARIABLES when the number of CPUs (observations)

in those datasets were not equal. Since CPUD has the

the primary CPU metrics, the MERGE was split into two

steps, first the three CPUD/CPUNP/CPUNS with CPUNR added

to the BY list, and now OUTPUT only if CPUD is present,

whether or not the other two datasets are present for

this CPUNR. Then the second step MERGES the output with

SNAPMAP to create PDB.NMONCPUD with the CPU metrics for

each CPUNR.

-Diagnostic tests for DISKBUSY counter errors corrected.

Thanks to Raissa Moussu, METROLOGFIE AIX EN PROVENCE, FRANCE.

Change 33.091 Support for CICS User fields USERAGT, ORIGUOW, and

IMACAAAA ORIGTGID.

IMACICVM

IMACICVN

IMACICVO

UTILEXCL

VMAC110

Apr 2, 2015

Thanks to Rob Hollingum, HSBC, ENGLAND.
Change 33.090 Mainview for MQ PTFs corrected the Offset ENTO value from

VMACBBMQ the always incorrect prior value of 28 for BBMQQUES 'E6'

Apr 2, 2015 to the correct value of 32, but this is an INCOMPATIBLE

change which mis-aligned MXG input, because MXG has known

to skip over those four bytes. Logic revised to support

either value since the records are otherwise unchanged.

BPL2458(MVMQ51) and BPL2459(MVMQ52) were the PTFs.

Thanks to Jim Swinarski, Credit-Suisse, USA.

Change 33.089 Not Used.
Change 33.088 Some duration variables in RMF III ZRBxxxx datasets still

VMACRMFV had TIME12.2 format but have millisecond resolution, so