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

Dec 16, 2014 variables LCLDSN, XCOLDSN1, and in the FILE= argument in

Dec 22, 2014 the LASTMSG variable.

-Dec 16: This support extended back to version 11.5.

-Dec 22: Lines longer than 72 characters reduced.

Thanks to Rosa Maria Martinez Alonso, Bustia, SPAIN.

Change 32.275 Variable NDMRTCPU in NDMRT dataset was incorrectly INPUT.

VMACNDM The +2 after NDMGPE1D no longer exists.

Nov 20, 2014

Thanks to Michael Oujesky, DTCC, USA.

Change 32.274 -The PDBAUDIT report of sizes of datasets in your PDB's is

PDBAUDIT enhanced with SORTEDBY option so that you can choose

Nov 27, 2014 the printed report order as deeded. The default order is

SORTBY=LIBNAME DESCENDING FILESIZE MEMNAME

but you can alter the order any way you wish as:

SORTBY=DESCENDING FILESIZE MEMNAME

-Numeric values are now formatted with commas and a new

column is added with FILESIZE in MGBYTES, and labels

rather than variable names are printed.
Change 32.273 ASCII execution only. Variable LIOBTIME in RACF0200

VMACRACF dataset was incorrect for times from midnight to 1 am.

Nov 20, 2014 The EBCDIC logic was copied to the ASCII code area.

Thanks to Matthew Chappell, QLD Dept Transport Main Roads, AUSTRALIA

Change 32.272 ERROR FOUND IN XAMSYS FILE. RECORD WAS NOT OUTPUT.

VMACXAM _N_=1 COL=8987 SEGNAME=CMS SEGSTART=8979 SEGLEN=16448

Nov 18, 2014 A second INPUT +SKIP in the STOSHR record misaligned.

Thanks to Randall Springs, BB&T, USA.

Change 32.271 The "true" PDB.SMFINTRV dataset built by BUILDPDB, with

SMFINTRV consolidated MULTIDD='Y' observations, ACCOUNTn's added,

VMAC30 and with the true count of tape devices allocated for

Nov 19, 2014 each interval, can now be built directly from SMF data,

or from these datasets in //WORK and/or //PDB:

TYPE30TD TYPE30_1 TYPE30_5 and TYPE30_V=SMFINTRV.

Use %INCLUDE SOURCLIB(TYPE30,SMFINTRV); to build ONLY the

PDB.SMFINTRV dataset from SMF; use TYPS30 to also create

all TYPE30xx datasets. This %UTILBLDP example shows how

other SMF records can be added to the processing in one

pass of the SMF file, with USERADD= adding SMF type 42,

and INCLAFTR=SMFINTRV adding creation of PDB.SMFINTRV:

%UTILBLDP(BUILDPDB=NO,

USERADD=30 42,

INCLAFTR=SMFINTRV,

OUTFILE=INSTREAM);

%INCLUDE INSTREAM;
(The _STY30UV Data Set Sort macro in TYPS30 to sort the

WORK.TYPE30_V dataset to the PDB does not name the sort

output to PDB.TYPE30_V, but instead, it simply sorts the

unconsolidated TYPE30_V dataset, naming it PDB.SMFINTRV.

Fortunately, by design, the %VMXGWORL invocation looks

first for the _WTY30UV token, which is WORK.TYPE30_V, but

if that does not exist, then then the _LTY30UV token,

which is PDB.SMFINTRV is used, so SMFINTRV can be used

with either TYPE30 or TYPS30 programs transparently.)
To identify the source of your PDB.SMFINTRV dataset,

the PROC CONTENTS DATA=PDB.SMFINTRV; will show

Built By Data Set Label text

TYPS30(_STY30UV) TY30UV: TYPE 30 INTERVAL EVENT

SMFINTRV DB30UV: SMFINTRV FROM SMFINTRV

BUILDPDB/3 DB30UV: BUILDPDB - SMFINTRV

-"Missing value" messages during _STY30U6 execution were

observed and eliminated with VMAC30 "cosmetic" changes.

Thanks to Cesar V. Cocco, Verisk, USA.
Change 32.270 The IHDRBVIR "BVIR Header" exit member was overlooked but

IHDRBVIR it now exists, and is taken after these header variables

VMACBVIR BVIRHIST BVIRLEN BVIRTYPE BVIRVERS CLUSTER DLIBSEHX

VMXGINIT DLIBSEQN DURATM ENDTIME GLIBSEHX GLIBSEQN GMTOFFBV

Nov 19, 2014 MACHMODL MACHSERL MACHTYPE NODEID SMFTIME STARTIME

TMZOFF VECDLEVL VECDLEVL

have been input and can be used to select which BVIR data

records are processed. As with all IHDRxxxx exits, when

EDITed with your criteria, that will apply to ALL jobs

that process the BVIR data, but each IHDRxxxx member also

also defines a unique macro variable, here MACBVIRH, that

can be used in the //SYSIN text to tailor a specific job.

See comments in the IHDRBVIR and IMACFILE exit examples.

Thanks to Ravi Kumar, Bank of America, USA.

Change 32.269 Documentation of two common ODS errors on z/OS:

ADOCODSO -NOTE: IMAGE NAME PREFIX WILL BE TRUNCATED TO 5

Nov 19, 2014 CHARACTERS TO COMPLY WITH Z/OS FILENAME RULES.

ERROR: CANNOT WRITE IMAGE TO UXMCBH.CPUS0.GIF. PLEASE

ENSURE THAT PROPER DISK PERMISSIONS ARE SET.

No output destination was defined so ODS is trying to

create datasets with userid.SGPLO.PNG and will almost

always fail, or ODS LISTING CLOSE; was not used.

-ERROR: THE JAVA PROXY IS NOT RESPONDING.

ERROR: THE JAVA PROXY'S JNI CALL TO START THE VM FAILED.

ERROR: UNABLE TO LOAD THE JAVA VIRTUAL MACHINE.

PLEASE SEE THE INSTALLATION INSTRUCTIONS OR

SYSTEM ADMINISTRATOR.

Region size too small

Change 32.268 The new ANALDSCK ("DSNAME Check") reads all SMF records

ANALDSCK (SMF 6 14,15,17,18,42.6,60,61,65,66,62,64) that contain

Nov 19, 2014 a DSNAME or ENTRNAME and selects the observations that

contain the full DSNAME, or the High-Level-Index, or any

text you choose, since the INDEX(DSNAME,'text') function

is used for the selection criteria. %VMXGPRAL is invoked

in this example to print all of the found observations.

Thanks to Jack Newburn, Los Angeles County Education, USA.

Change 32.267 Variables CRYAC1U and CRAM3U (1024/4096 utilizations) in

VMAC7072 the TYPE70X2 (PCI CRYPTO ACCELERATOR PCICA) need to be

Nov 19, 2014 multiplied by 100, like the other percentages.

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

Change 32.266 For CICS, DB2 variable QWHCCV is an alternate source for

VMXGUOW the CICS transaction name, so it is now added to the list

Nov 19, 2014 of transaction names in PDB.ASUMUOW dataset.
Change 32.265 Support for a second CMODNAME='USER' CMODHEAD='USER' CICS

IMACICDV optional segments for dataset CICSTRAN. The first segment

IMACICVS will be tailored in the IMACICUS member and the second

UTILEXCL segment' CMODNAME/HEAD values are set to CMODNAME='USER2'

Nov 18, 2014 and the new IMACICVS member is listed to be tailored for

the second segment.

Thanks to Tom MacCabe, Dominion Resource Services, USA.
Change 32.264 WPS ONLY. The ancient test for &SYSVER GE 6 (to ensure

WEEKBLDD the "new in V6" CLEAR function was available, to then

WEEKBLDT CLEAR the LIBNAME) failed under WPS, with a spurious

WEEKBL3D error message about BLKSIZE mismatch with LRECL=133!

WEEKBL3T The test needs to use &SASVER which is 8 or greater for

Nov 18, 2014 both SAS and WPS; SYSVER returns the current SAS version

which is also 8 or greater, but WPS returns a 3.

Thanks to Simone Niemczura, Sungard, USA.

Change 32.263 SAS fails to PROC COPY a DATA library with an ITEMSTORE,

DOC to a tape or sequential format library on disk, with:

ERROR: An item store cannot be created in a sequential library.

ERROR: File PDBINTRV.REGSTRY.ITEMSTOR has not been saved

Nov 14, 2014 because copy could not be completed.

All PROC COPY statements now include the MEMTYPE=DATA

option to prevent the error; most COPY's did have either

MT=DATA or MEMTYPE specified.

Thanks to Cesar V Cocco, Verisk, USA.
Change 32.262 -If you are running BLDSMPDB with Week-To-Date WTD, the

BLDSMPDB WEEKKEEP logic in BLDSMPDB failed if you used a wild card

VMXGALOC syntax (ASUM: for example).

Nov 12, 2014 -If you used WTD logic with AUTOALOC on ASCII and used

Dec 15, 2014 FORCEDAY to do a rerun the data went into the WEEK-1

dataset rather than the correct week. Dec 12: Typo fixed.

Change 32.261 If a tailored program compiled _CDE102 first, variable

VMACDB2H JOB could be length $12, because there was no LENGTH

Nov 11, 2014 statement, causing JOB=(SUBSTR(QWHCCV,1,8) to set the

length from the length of QWHCCV. A LENGTH statement is

added in VMACDB2H to protect the expected length $8.

Thanks to Paul Volpi, UHC, USA.

Change 32.260 Executing MXG on Linux to process AS400 data, a forward

VMXGDSNL slash in the path name was fine on windows but was taken

Nov 12, 2014 to be a divide symbol under linux, causing a numeric

operator to be looked for and not found. The use of the

%QUOTE() was required to resolve the intended meaning of

the slash, but the code used ONLY for TYPEQACS processing

has been rewritten to be sensitive to the execution

platform.

FOR ZOS IT WILL USE THE PENULTIMATE NODE OF THE DSNAME

Thanks to H. Sterling James, DST Systems Inc.

Change 32.259 DB2 IFCID=220: "NOTE: Invalid Argument to function INPUT"

VMAC102 prints a hex dump of the first two instances. MXG's INPUT

Nov 10, 2014 statement had not been validated with IFCID=220s and its

misalignment had not been previously observed.

Thanks to Rachel Holt, Fidelity Systems, USA.
Change 32.258 Variable SHIFT (the shift of INTBTIME) is now added to

ASUMSMFI the PDB.ASUMSMFI dataset which summarizes PDB.SMFINTRV.

Nov 10, 2014

Thanks to Frank Lund, EVRY, NORWAY.

Change 32.257 RMF Type 78 Subtype 2 record always contain the Virtual

FORMATS Storage Section, and can optionally contain Private Area

VMACID and Subpool Statistics (enabled by listing JOB names to

VMACSMF be monitored in the VSTORE options in your ERBRMFxx).

VMAC78 The %ANALID report now recognizes and reports the records

NOV 7, 2014 78.2.1:VIRTUAL STORAGE ACTIVITY

78.2.2:VIRT+PRIVATE AREA/SUBPOOL

so you can tell if there were any of those JOBS that were

initiated during this interval.

Thanks to Ray Dunn, CIGNA, USA.

Thanks to Keith C. Shaffer, CIGNA, USA.
Change 32.256 These duration variables in dataset TYPEMVAO

VMACMVAO ACTIVETM CPUTOTAL CPUMAX ELAPSTOT ELAPSMAX TAKETIME

Nov 7, 2014 SUCCTIME SUCRETTM FAILTIME FAIRETTM

were all incorrectly multiplied by 60; I think the first

data in 2010 was in minutes, but now is in seconds, so

that multiply is now removed.

Thanks to Christa Neven, KBC Group, BELGIUM.
Change 32.255 The HIGHCONTRAST style that was used as the parent for

MXGSTYL1 MXGSTYLE1 was changed (unexpectedly, unclear when) to

Nov 6, 2014 have a black background and all of the colors were also

modified. Changing the parent to PRINTER resolved the

problem.

Thanks to Richard Way, Office Depot, USA.

Change 32.254 z/VM Variable INTERVAL and HFRATE can be wrong if there

FORMATS is no SAMPLE PROFILE (1.09) MONWRITE record written at

VMACVMXA startup. They are missing values if there is only one

Nov 4, 2014 system's MONWRITE data being read, but if MONWRITE data

from multiple systems is concatenated, the variables will

be non-missing and wrong values, as they were retained

from the prior systems 1.09 record. Now, both variables

are set to a missing value in the 1.04 (Monitor Start) to

clear the retained values. Fortunately, neither variable

is used in any MXG processing.

-Cosmetic, in FORMATS, the RECTYPE format MGVXATY had not

been updated for recently added MONWRITE records.

Thanks to Giuseppe Giacomodonato, EPVTECH, ITALY.
Change 32.253 zOS only and only if daily datasets are GDGs, MONTHBLD:

BLDSMPDB It is possible that unless the GDGs are in precisely the

MONTHASC right order (today is 0, yesterday -1, etc,) that the

MONTHBL3 MONTHBLD logic could cause you to have duplicate and/or

MONTHBLD missing data in your MONTHLY datasets. All of these

MONTHDSK members now look for daily data that is in the current

VSETMNTH week - that is the last week that started before

Nov 1, 2014 month-end and if the ZDATE is in that week will continue

Nov 6, 2014 to process that record. On zOS, all 7 days are always

set by VSETMNTH to ensure we get all of the data from the

current week into the process.
Change 32.252 Support for EzSM 4.2.0 creates two new datasets, one each

EXEZSMT2 from splitting subtype 2 and 3, and subtypes 1, 2, 3, 7

EXEZSMV3 are all now decoded and variables created. Only header

IMACEZSM variables are output for subtypes 4, 5, and 6, pending

VMACEZSM test data for those subtypes.

VMXGINIT dddddd dataset description

Oct 31, 2014 EZSM02 EZSM02 EZSM02: USER START OR END OR END

EZSMT2 EZSM02T EZSMT2: TRANSACTION START OR END

EZSM03 EZSM03 EZSM03: MVS CONSOLE COMMAND

EZSMV3 EZSMV3 EZSMV3: MVS VARY COMMAND

Thanks to Perry Lim, Union Bank, USA.
Change 32.251 TYPE30 variables EXCPNODD/TODD/TOTL,IOTMNODD/TODD/TOTL

VMAC30 were inconsistent when NUMDD=0. MXG set EXCPNODD=0 if

Oct 30, 2014 it was missing, to prevent MISSING VALUE messages, but

when NUMDD=0, there ARE no DD segments, so leaving

EXCPNODD=. when NUMDD=0 makes better sense. And while

NUMDD=0 did set IOTMTODD=., that caused the calculated

IOTMTOTL to be missing when IOTMTOTL=IOTMTOTL-IOTMTODD

was executed. All are now correct and consistent; when

NUMDD=0 the two NODDs will be missing and TOTL correct

in both TYPE30_V and PDB.SMFINTRV datasets. While MULTIDD

observations (30's with ONLY DD segments) don't have a

Unit Record segment (so IOTMTOTL in TYPE30_V will be a

missing value), the BUILDPDB/ONLYINTV consolidation of

all of those MULTIDD='Y' observations from TYPE30_V into

the PDB.SMFINTRV observation sums those variables.

Thanks to David Campbell, SunTrust, USA.

Change 32.250 -RMF III Data Set selection enhancements, fixes, and

ADOCRMFV documentation upgrades.

ASMRMFV -New parameter SYSPLEX= (alias PLEX=) provides for the

CLRMFV selection of RMF Monitor III data sets based on the

JCLRMFV originating Sysplex Name and is explained more below.

JCLCRMFV -New parameter SYSTEM= (aliases SYSID=, SID=) provides for

JCLDRMFV the selection of RMF Monitor III data sets based on the

Nov 29, 2014 originating System Id and is explained more below.

-New parameters SHOWMATCH (alias SM) and NOSHOWMATCH

(alias NOSM) control display results of SYSPLEX= and

SYSTEM= value matching. The default is NOSHOWMATCH.

-New parameters NODUPDSN (alias NODUP) and DUPDSN (alias

DUP) control detection of duplicate RMF Monitor III data

set names as input.

-The default for NODUPDSN/DUPDSN is DUPDSN that provides

the previous ASMRMFV behavior of no duplicate detection.

-New parameter DUPERR=ABEND/WARN/IGNORE controls what

happens when NODUPDSN is in effect and an RMF III data

set name duplicate is detected. The default is

DUPERR=WARN. DUPERR is ignored with DUPDSN in effect.

-SYSPLEX= and SYSTEM= do NOT allocate any RMF Monitor III

data sets, but just filter those already allocated

whether by JCL or INDSNAME= patterns.

-The SYSPLEX= and SYSTEM= parameters may be useful for

large sites with many Sysplexes and/or System Ids that

have an established comprehensive set of RMF III data

sets used for MXG PDB builds. But a complete set may not

be needed for all uses such as for certain ad hoc

studies.

-Use of SYSPLEX= and/or SYSTEM= avoids having to set up

special groups of RMF III data sets in JCL or as

INDSNAME= patterns just for special needs. The standard

configuration set of RMF III data sets can always be

used, but filtered as needed with SYSPLEX= and/or

SYSTEM=.

-Note that SYSPLEX= and SYSTEM= when specified do add some

overhead because every data set has to be opened to

examine the originating Sysplex Name and System Id

whether it will be filtered or not.

-However, that drawback is possibly offset with the

avoidance of inconvenience and errors associated with

tailoring of JCL or INDSNAME= patterns for special PDB

builds.

-Defaults are SYSPLEX=ALL and SYSTEM=ALL meaning that no

RMF Monitor III data set filtering occurs based on

Sysplex or Sysid origin as in prior ASMRMFV versions.

-SYSPLEX= and SYSTEM= support any of the following formats

for values:

SYSPLEX=Range1 SYSTEM=Range1

SYSPLEX=Range1:Range2 SYSTEM=Range1:Range2

SYSPLEX=Pattern1 SYSTEM=Pattern1

SYSPLEX=Pattern1:Pattern2 SYSTEM=Pattern1:Pattern2

SYSPLEX=Pattern1:Range1 SYSTEM=Pattern1:Range1

SYSPLEX=Range1:Pattern1 SYSTEM=Range1:Pattern1

-A Range is one or more alphanumeric (A-Z,0-9) or national

(@,#,$) characters allowed in a Sysplex Name or System

Id.

-Range1:Range2 specifies a low/high Range Pair of values

where the source Sysplex or Sysid string must be equal or

greater than Range1 and also equal to or less than Range2

for the data set to be selected.

-Patterns provide generic matching with the Sysplex and

System fields in the Data Set Header (DSH) table for

flexibility.

-A Pattern can include any of the characters in a Range

and MUST also include at least 1 special pattern

character to be recognized as a Pattern. Otherwise it is

handled as a Range.

-Supported pattern characters are:

Pattern Character Matches In Source String

----------------- ------------------------------

'*' (Asterisk) 0 or more characters

'%' (Percent sign) 1 Non-blank character

'+' (Plus sign) 1 Numeric character (0-9)

'_' (Underscore) 1 Alphabetic character (A-Z)

'.' (Period) 1 National character (@,#,$)

-The colon ':' character is a required delimiter when two

values are provided for SYSPLEX= or SYSTEM=.

-Either or both SYSPLEX= and SYSTEM= parameters may be

coded and each repeated as needed (subject to internal

table size limits).

-New severe error message RMFV034S is displayed when a

Range or Pattern filter table is exhausted. The limit

for SYSPLEX= Ranges or Patterns is 16 each. The limit

for SYSTEM= Ranges or Patterns is 64 each. If higher

limits are needed please contact MXG Technical Support.

-When both SYSPLEX= and SYSTEM= are specified the match

results are logically ANDed so that matches must occur

for BOTH Sysplex Name and System Id for the data set to

be selected.

-For more details and examples on Range and Pattern usage

with SYSPLEX= and SYSTEM= for RMF Monitor III data set

selection see the ASMRMFV, ADOCRMFV, JCLRMFV, JCLCRMFV,

and JCLDRMFV Sourclib members.

-New information message RMFV034I is displayed when

SHOWMATCH is in effect. If a match occurs for SYSPLEX=

and/or SYSTEM= the original source string as well as the

Range Pair or Pattern that matched are displayed.

-When no matches occur for all Ranges and/or all Patterns

this condition is also shown.

-SHOWMATCH can be useful for testing the new SYSPLEX= and

SYSTEM= parameters. SHOWMATCH has very low overhead and

can be used freely without concern for a performance

impact.

-Add new variable severity message RMFV036* (*=I,W,E) when

a duplicate data set is found and NODUPDSN is in effect.

-NODUPDSN adds some CPU overhead because the current RMF

III data set name being opened is compared with all the

previous data set names already processed.

-Although duplicate data set name detection is a simple

compare with a basic counter loop, the number of

comparisons rises rapidly with the number of RMF III DD

allocations:

Maximum

Number RMF III Duplicate

Data Sets Input Comparisons

--------------- -----------

2 1

3 3

4 6

5 10

10 45

50 1,225

100 4,950

500 124,750

1,000 499,500

1,500 1,124,250

1,635 1,335,795 Default 32K TIOT size

2,000 1,999,000

2,500 3,123,750

3,000 4,498,500

3,273 5,354,628 Maximum 64K TIOT size

-Users who are confident they do not have any duplicate

RMF III data sets being input to ASMRMFV do not need to

code the NODUPDSN option.

-However, new ASMRMFV users may want to use NODUPDSN at

least once to insure their ASMRMFV run does not contain

duplicate data sets.

-Duplicates are also more likely using the Dynamic Method

with the INDSNAME= parameter due to possible overlapping

of data set name patterns.

-With the addition of the NODUPDSN, SYSPLEX=, and SYSTEM=

parameters RMF III data set filters are applied in the

following order (if coded):

1) NODUPDSN Prevent duplicate dsname input

2) SYSPLEX= Select Sysplex dsname origin(s)

3) SYSTEM= Select LPAR dsname origin(s)

4) Date/Time FROMDATE=,TODATE=,FROMTIME=,TOTIME=,WINDOW

5) DOW= Day of the Week

-Filter messages RMFV006I have been reordered in the log

to correspond with the general order that filters are

applied during an ASMRMFV run.

-A new summary RMFV014I message is added with RMF III data

set filter (bypass) counts for Total, Duplicates, Sysplex

Name, System Id, Date/Time, and Day of the Week filters.

-Options message RMFV036I showing RMF III table selections

is now an RMFV006I filter message and is now part of that

multi-line message group. These are properly filters not

just options.

-Several filters in an RMFV037I option message are moved

to an RMFV006I message as these are also properly filters

not just options.

-Add new Abend Reason code 84 when DUPERR=ABEND and a

duplicate data set name is found with NODUPDSN in effect.

-Message RMFV050S for the internal Dsname Table full

condition is now message RMFV079S.

-DDNAMEs dynamically allocated by the TSO Clist CLRMFV

will now have a RMFC DDNAME prefix that is also supported

by ASMRMFV instead of an RMFV DDNAME prefix as before.

-Thus the DDNAME prefix in messages now indicates:

Prefix Allocation Method

------ -------------------------------

RMFC CLRMFV Clist Dynamic Allocation

RMFD INDSNAME= Dynamic Allocation

RMFV Static JCL DD statement

-Split message RMFV049* into RMFV049* for CSI Catalog

errors only and RMFV050* for CSI Dsname entry errors

only. Before both errors used the same message skeleton.

-All variable severity messages with ids RMFVnnn* (*=

I,W,E,S) are now initialized only once during ASMRMFV

startup according to the settings of the various

*ERR=ABEND/WARN/IGNORE options.

-This improves performance because in prior ASMRMFV

versions some messages were reused and were reset

according to severity multiple times during execution.

-There is a new variable severity message RMFV056*

(*=I,W,E) issued just for pattern errors. Before these

were part of RMFV004* messages.

-CSI severe error message RMFV043S was not being built

with variable error text correctly when the condition

occurred.

-Issue Index messages RMFV028I and RMFV029* only if the