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

Change 24.136 created DB2UNICD='Y' when UIFCIDS=YES, and

DB2UNICD was used to conditionally INPUT these variables

QLACLOCN QLSTLOCN QMDALOCN QPACCOLN QPACLOCN QPACPKID

QWHCAID QWHCOAUD QWHCOPID QWHCROLE QWHCTCXT QWHSLOCN

(because only these fields were ASCII in test data).
However, when the below circumvention for an IBM ABEND

required UIFCIDS=YES to be set, these header variables

QWHDRQNM QWHDSVNM

were seen to be in ASCII, and VMACDB2H was updated.

There are at least 224 %U fields in DB2 V9 DSECTS that

need update or at least verification, and they will all

(EVENTUALLY!!) be updated/supported.
As of Jul 13, these variables are now "%U" Supported:
Still to be done:

These are the only "ACCOUNT" IFCIDS that need updating:

QPACASCH QPACAANM

and these are all of the SMF 102 IFCIDs that need update:

QW0022CI QW0022CN QW0022CR QW0022PG QW0022QO QW0022TN

QW0029CI QW0029LN QW0029PI

QW0030CI QW0030LN QW0030PI

QW0031CI QW0031LN QW0031PI

QW0053LN QW0053PC QW0053PN

QW0055NI QW0055OI

QW0058LN QW0058PC QW0058PN

QW0059CN QW0059LN QW0059PC QW0059PN

QW0060LN QW0060PC QW0060PN

QW0061CN QW0061LN QW0061PC QW0061PN

QW0064CI QW0064CN QW0064LN QW0064PN

QW0065CN QW0065LN QW0065PC QW0065PN

QW0066CN QW0066LN QW0066PC QW0066PN

QW0083SA

QW0087SA

QW0096PC QW0096PN

QW0108NC QW0108NI QW0108NL QW0108OH QW0108OW QW0108QL

QW0110PC QW0110PI QW0110PL

QW0112OH

QW0113OH

QW01247S QW0124CI QW0124LN QW0124PN QW0124SP

QW0125PC QW0125PN

QW0140SC QW0140SN QW0140TC QW0140TN QW0140UR

QW0141OR

QW0142CR QW0142OW QW0142TN

QW0145LN QW0145PC QW0145PN

QW01488L QW01489L QW0148CI QW0148LN QW0148PN QW0148SP

QW0157LN QW0157PN

QW0158PN

QW0159LN

QW0162LN

QW0169AL QW0169AU QW0169LO QW0169NE

QW0173CN*QW0173ID*QW0173PC QW0173PK

QW0177CO QW0177LO QW0177OH QW0177PI

QW0183CO QW0183LN QW0183PN

QW0185CN QW0185CR QW0185TB

QW0191LN

QW0192LN [192cs 192ec]

QW0193LN

QW0194LN

QW0195LN

[195hb]

QW0203CO QW0203LO QW0203PA

QW0204LO

QW0205LO

QW0206LO

QW0207HN QW0207TN QW0207UN

QW0208LO

QW0209LO

QW0210LO

QW0221LN

QW0221PC QW0221PN

QW0222LN QW0222PC QW0222PN

QW0224CI QW0224PN

QW0233LN QW0233PC QW0233PN QW0233PR QW0233SC

QW0235LO

QW0236LO

QW0247LN QW0247PC QW0247PN

QW0248LN QW0248PC QW0248PN

QW0269PR QW0269RA QW0269RC QW0269RU QW0269SA QW0269TC

QW0272LN QW0272LP QW0272PC QW0272PG QW0272PN QW0272QN

QW0273CN QW0273LN QW0273LP QW0273PC QW0273PG QW0273PN

QW0305CN

QW0311CN QW0311LN QW0311PC QW0311PN QW0311QN QW0311TN

QW0313AI

QW03141N QW03142N QW0314BN QW0314LN QW0314MN QW0314NN

QW0316QD QW0316SC QW0316T1 QW0316T3 QW0316TD QW0316X4

QW0324CI QW0324CV QW0324FI QW0324FN QW0324FS QW0324NV

QW0325CO QW0325NM QW0325PR QW0325SC QW0325TX

QW0334LN

QW0341LN QW0341PC QW0341PN

QW0343ID QW0343PC QW0343PK QW0343PL

QW148SCH

QWP1RLFA

QWP4ADM2 QWP4DFID QWP4OPR1 QWP4OPR2 QWP4OZUS QWP4REGA
An ABEND 0C4 in DB4PMSTR (takes down DB2 Subsystem) can

occur if zparm UIFCIDS is set to "NO", and IFCIDs 316,317

are written. IBM suggested PTF UX37196, but that is PE.

Setting zparm UIFCIDS to "YES" bypassed the conversion.

ERROR DESCRIPTION:

Rapid stack storage increase about 200M per day was

observed by customer. This occurs when IFCID 316, 317

were in effect and the zparm UIFCIDS was set to no. The

stack storage was used for conversion from UNICODE to

EBCDIC.

LOCAL FIX:

Set zparm UIFCIDS to yes to bypass conversion or turn

off IFCID 316 and 317.

Thanks to Barry Pieper, Blue Cross of Minnesota, USA.

CHANGE 27.157 SMF 116 variables WQGETJCE/WQPUTJCE/WQPUT1JE/WQSET1JE in

VMAC116 the MQMQUEUE dataset were not divided by 4096, so their

Jul 7, 2009 values were incorrectly large without that divide.

Version 7.0 SMF 116 records have been read without error;

there are no new fields in the SMF 116 (nor 115) records.

in Version 7.0, so the MXG support was already in place.

Thanks to Frank Debree, Dexia, BELGIUM.
CHANGE 27.156 Support for z/VM 6.1.0 is already in place in MXG 27.01+,

TYPEVMXA as there were no changes to the MONWRITE data records,

Jul 7, 2009 except for the new version number.
CHANGE 27.155 -Type 42 Subtype 16 RLS Record INPUT STATEMENT EXCEEDED

VGETUTKN because the two length tests LEN42D2 and LEN42D4 should

VMAC42 have been 1652 instead of 1504.

Jul 6, 2009 -Type 42 Subtype 21 and 24 with invalid ICHRUTKN segment

Jul 8, 2009 had all of UTKNxxxx variables wrong or truncated. This

Jul 13, 2009 change detects the mis-located segment (starts in 187 vs

185) and correctly populates the UTKNxxxx variables,

except that the last variable in the record, UTKNGRUP,

is only 6-bytes long when the segment in invalid.

-IBM APAR OA27563 corrects the errors; search NEWSLTRS for

the details.

Thanks to Karl Lasecki, Chemical Abstracts Service, USA.

CHANGE 27.154 USS dataset TYPE9201 variable SMF92PPN (pathname to the

VMAC92 filesystem) was not input because SMF92PPL, its length

Jul 6, 2009 field was INPUT from the wrong offset in line 640,

which now contains

If SMF92MPF GT 0 THEN INPUT @SMF92MPF+1

Thanks to Dan Squillace, SAS Institute Cary, USA.

CHANGE 27.153 Dataset TYP11921 variable NTBTRNBE should have been INPUT

VMAC119 as &PIB.4. instead of &PIB.4.3 since it is a count and is

Jul 5, 2009 note a duration.

Thanks to Paul Volpi, UHC, USA.

CHANGE 27.152 MXG 27.01-MXG 27.05. DCOLLECT variable UBALLSP is missing

VMACDCOL always; its INPUT statement was lost when Change 27.034

Jul 2, 2009 updates were made.

Thanks to Thomas Heitlinger, FIDUCIA IT AG, GERMANY.

Thanks to Sieghart Seith, FIDUCIA IT AG, GERMANY.

Thanks to Isabelle Diagremont, FIDUCIA IT AG, GERMANY.

CHANGE 27.151 XAM TCP record caused INPUT EXCEEDED RECORD LENGTH ERROR.

VMACXAM A TCPAPP segment with only Segment Name and SEGLEN=8 with

Jul 2, 2009 no data can be created for internal XAM purposes. Only

the TCPAPP segment can have these "null segments", so the

SEGLEN is now tested after the header is INPUT, and the

segment is NOT output if there is no data.

Thanks to Chris Morgan, CitiCorp, ENGLAND.
CHANGE 27.150 -MXG 27.05. COPYONLY option of READDB2 didn't work for the

READDB2 ID=102 IFCIDs selection.

Jul 2, 2009

Thanks to Dan Almagro, Automobile Club of Southern California, USA.

CHANGE 27.149 Labels for Counts of ZIP, IFA, CPs, etc are clarified:

VMAC7072 In ASUM70PR ASUM70LP TYPE70PR - PER SYSTEM-CEC DATASETs

Jul 1, 2009 and ASUMCEC ASUMCELP - PER CEC DATASETS:

NRICFCPU='ICF CPUS*AVAILABLE*IN THIS CEC'

NRIFACPU='IFA CPUS*AVAILABLE*IN THIS CEC'

NRIFLCPU='IFL CPUS*INSTALLED*IN THIS CEC'

NRZIPCPU='ZIIP CPUS*AVAILABLE*IN THIS CEC'

because RMF can see everything in the CEC in the 70PR SMF

data.

But the TYPE70 and RMFINTRV DATASETS - PER SYSTEM DATA:

NRIFAS ='IFA CPUS*AVAIL TO*THIS ZOS'

NRZIPCPU='ZIIP CPUS*AVAIL TO*THIS ZOS'

NRIFACPU='IFA CPUS*AVAIL TO*THIS ZOS'

NRIFLCPU='IFL CPUS*INSTALLED*IN THIS CEC'

NRCPUS ='AVERAGE*ONLINE*NON-PARKED*CP ENGINES'

are not the "hardware" counts, but only the count of the

engines that this z/OS system can use and knows about.

Thanks to Scott Wigg, USBank, USA.

CHANGE 27.148 Support for APAR OA29428 adds diagnostic bits that create

VMAC1415 these new MXG variables and labels:

Jun 30, 2009 SMF14TCL='TASK*TERMINATION*CLOSED*THE DCB?'

SMF14ABD='TASK IS*TERMINATIONG*TCBFA*IS ON?'

MXG Variables whose label ends with a question mark are

one byte character variables whose value will be 'Y' if

true, and should be blank if not true (although some old

variables might contain 'N' for false).

The APAR text states:

For debugging purposes, customers need to know if a

dataset has either been closed by a program or RTM

cleanup processing. The SMF records (types 14 or 15)

do not currently provide a way to identify this

information. The addition of flag(s) are required to

indicate normal versus abnormal CLOSE. One methodology

that can add this information is to add a bit to

indicate that task termination closed the dcb and add

a second bit to indicate whether or not the task is

abending.
CHANGE 27.147 -MXG ANALZPCR didn't expand variable SCP to 10 bytes so

ANALZPCR z/OS 1.10 was not correctly set in the external study

Jun 30, 2009 file.

-MXG incorrectly set SCP from MVSLEVEL to be the same as

the MVSLEVEL of the TYPE70 that wrote the record, instead

instead of using the SMF70STN from each LPAR to define

that system's SCP. Now, TYPE70 is read to create the new

$TEMPSTN format, mapping SYSTEM to MVSLEVEL, and then

SMF70STN value in TYPE70PR is looked up to set MVSLEVEL.

Thanks to Karl Lasecki, Chemical Abstracts Service, USA.

CHANGE 27.146 These OMCI datasets built by TYPE112

VMAC112 T112ADA T112ADAT T112DLI T112DLIT T112IDMT T112IDMS

Jun 30, 2009 T112VSAM T112VSAT T112SUPR T112SUPT T112DTCO T112DTCT

did not keep variables OMCIJOB OMGAPPL OMSAPPL.

Thanks to Richard Schwartz, State Street Bank, USA.
====== Changes thru 27.145 were in MXG 27.05 dated JUN 29, 2009========
CHANGE 27.145 Support for TMON for MQ record "QA" (APPLICATION STATS)

EXTMQQA creates two new datasets, TMMQQA and TMMQQAA. All of the

EXTMQQAA QAxxxxx variables with TIME12.2 format are guesses as to

IMACTMMQ their internal format, and TMON reports for validation

VMACTMMQ weren't available when MXG 27.05 was ready; please check

VMXGINIT with [e-mail] for any updates prior to using those

Jun 26, 2009 variables in these two new datasets.

Jul 1, 2009 Jul 1: QA record's 32 Latch Class times/counts created.

Thanks to Paul Volpi, UHC, USA.
CHANGE 27.144 New documentation member ADOCRMFR tabulates the name of

ADOCRMFR each MXG variable that is printed in IBM RMF reports.

Jun 26, 2009

Thanks to George Canning, Produban, ENGLAND.

CHANGE 27.143 Variable REDIRECT in PRPR1010 dataset contains text, in

VMACPRPR spite of the documentation that it is a numeric field.

Jun 26, 2009 Previous test data had blanks. Now INFORMAT $16.

Thanks to Siegfried Trantes, Gothaer Systems GmbH, GERMANY.

CHANGE 27.142 Using the %VMXGPRAL utility to print SUSE datasets caused

VMXGPRNT errors because of the long-length-variable-names in SUSE,

Jun 26, 2009 and the need for LRECL=255 on the temporary FILENAME.

Thanks to James J. Flanagan, ISO, USA.

CHANGE 27.141 Variables R792FLG and R792FLG2 are now kept in TYPE792.

VMAC79

Jun 25, 2009

Thanks to Bruce Widlund, Merrill Consultants, USA.

CHANGE 27.140 Change 27.062 incorrectly input S42DSRDD/S42DSRDT as &RB

VMAC42 but the fields are &PIB format.

Jun 25, 2009

Thanks to Siegfried Trantes, Gothaer Systems GmbH, GERMANY.

CHANGE 27.139 Line 1024 should be BO SCANLOOP Y. TRY ANOTHER VOLUME

ASMVTOC instead of BZ. This caused no volumes to be returned when

Jun 22, 2009 SELECT vol1 vol2 .. was used.

Thanks to Paul Gillis, Pacific Systems Management Pty Ltd, AUSTRALIA.

CHANGE 27.138 -z/OS 1.10 SMF 85 OAM record INPUT EXCEEDED RECORD ERROR.

VMAC85 MXG Version/Release/ModLevel variable R85PRVM='1*0' was

Jun 18, 2009 created as it didn't expect a 2-digit Release Number.

Jul 17, 2009 Fortunately, R85PRVM is not kept so it was expanded to

now supports 2 digit Version so '1100' is GT '1 30'.

-TYPE85AC dataset. New variables

R85STY='RECORD SUBTYPE'.

R85STOK ='OSREQ*STOKEN'

R85RC2 ='OSREQ*RETURN*CODE*TWO'

-Subtypes 8 thru 10 now are also output to TYPE85AC.

-But the R85PRVM value is not consistent. These values

are found in my past test data files:

1999 1. 4.0

2003 1. 3.0

2003 2.10.0

2007 1. 8.0

2009 1.10.0

-Jul 17: New variables added to TYPE85ST dataset:

R85PUDK ='BYTES*DELETED*TAPE*SUBLEVEL 2*/

R85PUDO ='PRIMARY*OBJECTS*DELETED*TAPE*SUBLEVEL 2'

R85PURK ='BYTES*READ*TAPE*SUBLEVEL 2'

R85PURO ='PRIMARY*OBJECTS*READ*TAPE*SUBLEVEL 2'

R85PUWK ='BYTES*WRITTEN*TAPE*SUBLEVEL 2'

R85PUWO ='PRIMARY*OBJECTS*WRITTEN*TAPE*SUBLEVEL 2'

-Corrections to handling of subtype 78,79,and 88 for

test data from versions 1100, 1 30, and 1 40.

Thanks to Joachim Sarkoschitz, DATEV eG, GERMANY.

Thanks to Harald Seifert, HUK-COBURG, GERMANY.

CHANGE 27.137 Analysis of interval zIIP usage by JOB/PROGRAM/etc using

ANALZIPU PDB.SMFINTRV. You define the ZIPGROUP variable to group

Jun 18, 2009 that is then summed for each interval, and the percentage

of total zip usage by that ZIPGROUP for that interval is

calculated. The analysis does demand that your SMF 30

data is synchronized.

Thanks to David Carr, Blue Cross Blue Shield of Kansas, USA.
CHANGE 27.136 Unused Change.
CHANGE 27.135 The (VERY OLD) ANALRMFI example's selection criteria had

ANALRMFI test for a series of variables ANDed to be GE 0, but some

Jun 17, 2009 of the variables are, now, always missing (logical swaps)

causing nothing to be selected. The WHERE clause tests

now are ORed, and always-missing variables were removed.

However, like all ANALxxxx report examples, this is only

an example, as a starting report to be tailored into your

very own report, if the report is of interest.

Thanks to Cletus McGee, Alfa Mutual Insurance Company, USA.
CHANGE 27.134 Further support for IBM i, i5/OS, iSeries, or AS400 V6R1

EXQAPHDW (to include all old and new names of these products).

IMACQACS New dataset created:

VMACQACS -QAPMHDWR Hardware Configuration

VMXGINIT Existing datasets updated:

Jun 17, 2009 -QAPMBUS - BUIOPB increased from 2 to 3 bytes.

Jun 18, 2009 -QAPMJOBM- New variables added:

JBJTHDT ='JVM*THREAD*TYPE'

JBJVMF ='JVM*STARTED'

JBJVMT ='JVM*TYPE'

JBMDYR ='FILE*SYSTEM*DIRECTORY*READS'

JBMLCH ='FILE*SYSTEM*LOOKUP*CACHE HITS'

JBMLCM ='FILE*SYSTEM*LOOKUP*CACHE MISSES'

JBMNDC ='FILE*SYSTEM*NON-DIR*CREATE*OPS'

JBMNDD ='FILE*SYSTEM*NON-DIR*DELETE*OPS'

JBMOPN ='FILE*SYSTEM*OPEN*OPERATIONS'

JBMSLR ='FILE*SYSTEM*SYMBOLIC*LINK READS'

JBPASE ='I5/OS*PASE*RUNTIME*ACTIVE'

JBPGRL ='PAGE*FRAMES*RELEASED'

JBPGRQ ='PAGE*FRAMES*REQUESTED'

JBSCPU ='SCALED*CPU*MICROSECONDS'

JBSTCPU ='TOTAL*SCALED*JOB*CPU'

There were also a number of fields labeled RESERVED that

are INPUT and kept, in case they become populated.

-QAPMSYST New variables added:

SYSIUL ='USER*AUTHORIZATIONS*AVAILABLE'

SYSCIU ='USER*AUTHORIZATIONS*NEEDED'

Thanks to Clayton Buck, UniGroup, USA.

CHANGE 27.133 Variable CFBUSYTO is created in TYPE74CF as the sum of

VMAC74 all CF engine busy time (CFBUSY01-CFBUSY16).

Jun 16, 2009
CHANGE 27.132 DB2 V8 IFCID=22 Record with truncated name fields caused

VMAC102 INPUT STATEMENT EXCEEDED RECORD LENGTH some of the time.

Jun 16, 2009 -MXG 27.05 was still wrong for some cases; Jul 2 revision

Jul 2, 2009 with tested with both V8 and V9, but the V8 records do

not exactly agree with the V8.1 DSECTS, so counters were

added to ensure alignment (I hope!).

Thanks to Tom Buie, Southern California Edison, USA.
CHANGE 27.131 Enhancements to DB2PM-like STATISTICS LONG report have

ANALDB2R caused updates to create/correct needed DB2 variables:

ASUMDBSS -Variables QWHADSGN,QWHAMEMN kept in all DB2ACCTx/DB2STATx

READDB2 datasets, as they are in the header of all DB2PM reports.

TRNDDBSS -Variables QDSTQMIT QTMAXPB QDSTQIN2 QBSTXIS QTGSPEMX

VMAC102 QXMAXDEG QXSTXMLV QXSTLOBV

VMACDB2 were DIF()'d but they do not contain accumulated values.

Jun 26, 2009 -Variables QDSTQIN2 QTPACOW1 QTPACOW2

were NOT DIF()'d, now are, contain accumulated values.

-Variables QTGACSLM QTGALSLM QTGAUSLM QTGSCSLM QTGSLSLM

and QTGSUSLM are actually spelled QTG..LSM in DB2 DSECT,

but once named in MXG there only pain and anguish if I

rename a variable, and these aren't important enough to

have me create correctly spelled duplicate variables.

-T102S106 System Parameters dataset was updated with these

new variables added for V8 and V9:

QWP1ACCU QWP1ACID QWP1CDB QWP1IDBP QWP1IXPX QWP1IXQT

QWP1LVA QWP1LVS QWP1LWCK QWP1MOFR QWP1RLFA QWP1SYFL

QWP1SYMV QWP1TP16 QWP1TP32 QWP1TP8 QWP1TPLB QWP1TPXM

QWP1TSQT QWP1WLME QWP1XVA QWP1XVS

QWP4APS QWP4FRLC QWP4IAST QWP4MIS6 QWP4MS4C QWP4MXAB

QWP4MXDC QWP4NUPT QWP4PMGT QWP4RSDC QWP4RSMT QWP4SKLC

QWP4SRTN QWP4WFAL

QWP5DCYC QWP5DLOK QWP5FLG QWP5HASH QWP5MCSA QWP5PHSH

QWP5RLE QWP5TVAL

QWP9TCKA QWP9INAC

QWPBAPSC QWPBDB2S QWPBDDRM QWPBLCTP QWPBLNM QWPBNUFN

QWPBPADN QWPBUGID QWPBUMID QWPBUSID

-New ASUMDBSS Statistics Summarization summarizes:

Default INPUT dataset Default OUTPUT dataset

PDB.DB2STATS PDB.ASUMDBSS

PDB.DB2STATB PDB.ASUMDBBS

PDB.DB2GBPST PDB.ASUMDBSG

The new ASUMDBSS eliminates the need for old ASUMDBSB as

only ASUMDBSS is needed to create all three stat ASUMs.

-New TRNDDBSS Statistics Summarization summarizes:

Default INPUT dataset Default OUTPUT dataset

WEEK.ASUMDBSS TREND.TRNDDBSS

WEEK.ASUMDBBS TREND.TRNDDBBS

WEEK.ASUMDBSG TREND.TRNDDBSG

The new TRNDDBSS eliminates the need for old TRNDDB2S and

TRNDDB2X, as TRNDDBSS creates all three stat TRNDs.

If you are currently using TRNDDB2S or TRNDDB2X, read the

comments in TRNDDBSS with the one-time migration example.

-Statistics reports can be selected by DB2 subsystem

(QWHSSSID), local location (QWHSLOCN), BEGIN and END time

using the DB2 DB2LOCAL BEGTIME and ENDTIME parameters.

-READDB2 revised with new parameter RUNASUMS, default NO.

Set RUNASUMS= interval desired (HOUR/DATE/WEEK/etc.) and

the ASUMDB2S will create the statistics summary datasets

in PDBOUT library.

-ANALDB2R these reports have been brought up to date with

DB2 Performance Expert/Performance Monitor documentation:

PMACC01/PMACC03=Accounting short report

PMSTA01/PMSTA03=Statistics long reports

PMSTA02/PMSTA04=Statistics short reports

PMSPR01=System Parameter Report

Thanks to Chuck Hopf, Independent Consultant, USA.

CHANGE 27.130 These variables for dataset QAPLPAR

VMACQACS LPCAP LPAVL LPBSY LPRSP LPRDS LPWRTS

Jun 15, 2009 LPDISK LPMEM

were INPUT but not kept.

Thanks to David Bixler, FISERV, USA.
CHANGE 27.129 Support for changes to CONTROL-D TYPE C User SMF record.

VMACCTLL Eight LOGCODE values have existing JOBNAME RECIPIENT and

Jun 15, 2009 REPORT variables in different locations, plus two new

ACTIONCT and VALUESCT variables.

Thanks to Josep Miquel, La Caixa, SPAIN.
CHANGE 27.128 Change 27.046 made major update to EDGHSKP type D,V,X but

VMACEDGR invalid syntax of 'GT . GT . ' was not detected, causing

Jun 12, 2009 MANY datetime variables to have missing values.

Thanks to Rudolf Sauer, T-Systems, GERMANY.

CHANGE 27.127 Support for CONTROLT for "daily dataset" billing, and

DAILYDSC other "DAILYDSN" job enhancements:

DAILYDSN -New DAILYDSC created for CONTROLT.

DAILYDSR -DAILYDSR modified to handle SPACE6 DAYS6 (bytes on

IMACVTS virtual and days)

JCLDAYDS -DAILYDSN modified to handle SPACE6 DAYS6 and to use the

Jun 11, 2009 same constructs as DAILYDSC overriding the _L for the

output TMS datasets to send them to the DATASETS LIB.

-IMACVTS used by all of the above to set a variable

VIRTREAL to REAL or VIRT. Comments on usage in member.

Default is REAL.

-JCLDAYDS modified to add a CTLT step

Thanks to Chuck Hopf, Independent Consultant, USA.
CHANGE 27.126 zVM Velocity Software variable SYTLPNAM/LCPUNAME is no

VMACXAM longer kept in dataset XAMSYU, which has an observation

Jun 11, 2009 with the LPAR management time for each hardware Engine,

and is not a per-LPAR dataset. MXG had incorrectly kept

SYTLPNAM from the last SYTCUP segment.

Thanks to Douglas C. Walter, Citigroup, USA.

CHANGE 27.125 -DB2 V8 Only, DB2STATS variables QISESKCT, QISESKPT were

VMACDB2 zero in SMF 100 subtype 1 records with LENQISE=148 bytes.

Jun 10, 2009 The QISE V8 DSECT was 96 bytes, ending at QISECFRE. In V9

QISE increased to 148 bytes. MXG tested length to input

new fields, so when a DB2 V8 record LEN=148 was read, the

new fields were input. However, in V9 IBM relocated the

two SKCT/SKPT fields into the first 8 bytes of the new

data, but this V8 record has zeros there, and the two

original field locations contain the number of SKCT and

SKPT pages. The MXG Logic continues to read new fields

QISEKFAL QISEKPGE QISEKFRE QISECTA QISEKTA

QISESFAL QISESPGE QISESFRE QISEKNFM QISEKNFA

QISEKNFR

with LEN=148 records, but no longer replaces the already

QISESKCT or QISESKPT values if the record is from V8.

-The MERGE of TEMPSTAT (ST0+ST1) and DB2STAT4/T102S225

now outputs PDB.DB2STATS if TEMPSTAT obs exist; Change

27.097 incorrectly also output unmatched 225s, including

the initial IFCID=225 for each subsystem, which has no

no matching ST0+ST1 observation. If only 225's exist,

then PDB.DB2STATS will have no observations, but the 225s

data will still be found in DB2STAT4 and T102S225.

-Change 27.097 incorrectly kept QWHSSTCK in PDB.DB2STATS;

BEGTIME ENDTIME are the datetime variables.

Thanks to John Shuck, SunTrust, USA.

Thanks to Chuck Hopf, Independent Consultant, USA

CHANGE 27.124 "WARNING: Apparent invocation of macro TRIM not resolved"

VMXGOPTR is serious, but ONLY happens if your //SASAUTOS does not

Jun 9, 2009 include the SAS-supplied AUTOCALL library, or if your

CONFIGVx in use doesn't have this required statement

OPTIONS MAUTOSOURCE SASAUTOS=(SOURCLIB SASAUTOS);

because %TRIM() is delivered by SAS as a %MACRO in their

AUTOCALL library. Previously, only a few ANALxxxx used

%TRIM(), but in MXG 27.04, it was added to %VMXGOPTR by

Change 27.092, and %VMXGOPTR is used throughout MXG, so

several sites with defective SASAUTO allocations failed

when they installed MXG 27.04 (but they would have failed

earlier if they had used any of those ANALxxxx's!). But

with closer examination, we have eliminated the need for

the %TRIM() use in %VMXGOPTR, as the string is already