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

Thanks to Brian LeBlanc, Candle Corporation, USA.

Change 11.289 AS/400 data sets are now fully documented in this new

ADOCQAPM ADOC chapter, although it has not been final-pass edited.

Jan 8, 1994
Change 11.288 This example of a prime-time cross-system DASD report

ANALDASD from PDB.TYPE74 may be of use in tracking DASD activity

Jan 8, 1994 across multiple systems.

Thanks to Dan Kaberon, Hewitt Associates, USA.

Change 11.287 Preliminary testing under CMS of MXG 11.09 found that the

REXXTES6 REXX sample did not include VMXGINIT nor did it contain

Jan 6, 1994 the correct FILEDEFs to concatenate USERID and SOURCLIB

MACLIBs. See also the circumvention in Change 11.067 to

remove CICS and DB2 processing to run BUILDPDB. This

text will be revised when testing under CMS is completed.

Thanks to Jerry Maier, NBD Bank, USA.
Change 11.286 DB2PM-like reports have been revised to correct all

ANALDB2R reported errors and now exploit fields added by DB2 2.3.

Jan 6, 1994 Many of the TRACE reports (notably AUDIT, LOCKING, SQL)

were heavily modified, and all reports now use ANALDBTR

pairing where appropriate. Input can now be from a PDB

library on tape, and one new report has been added:

PMLOK04 - Lock Detail Trace report.

Thanks to Mike Skopec, Platinum Technology, USA.

Thanks to Wai Choong Mak, Development Bank of Singapore, SINGAPORE.

Thanks to Andy Vick, Allied Dunbar Life Assurance, ENGLAND.

Thanks to H. Lugert, Datav, GERMANY
Change 11.285 Pairing logic for IFCIDs 114-115 was added, the T102S054

ANALDBTR dataset was added to the S044S045 pairing. Also, all

Jan 6, 1994 observations of T102S063 (SQL text) are now output (only

the first 200 bytes of text was previously output - this

change in ANALDBTR will now cause the ANALDB2R SQL trace

report to now print all of the SQL text). Also, if the

input datasets are from a PDB library, the intermediate

T102Sxx datasets in the WORK file are now deleted.

Change 11.284 Some variables needed for ANALDB2R were not kept in some

FORMATS T102Sxxx datasets, and compressed data in IFCID 21 and 54

VMAC102 are now decoded, using the new $MGDB2XX format. Three

Jan 6, 1994 labels that exceeded 40 bytes were shortened.

Change 11.283 A minor correction: if IFCID=ALL was specified, the

READDB2 output T102Sxxx datasets were not copied to the output

Jan 6, 1994 PDB library, but now they are.
Change 11.282 New dataset PDB.DB2STATS is now created by merging the

DIFFDB2 three DB2 statistics datasets (DB2STAT0, DB2STAT1, and

Jan 6, 1994 DB2STAT2) for ease in statistics reports. Eventually,

the new dataset will be used by ANALDB2R reports, and you

should use DB2STATS in your own reports. As these

statistic datasets are small, there is little cost for

the redundancy.
Change 11.281 Major reduction in VMXGSUM resources (CPU,DISK) are now

VMXGSUM completed. Originally planned as Change 11.245, those

Jan 6, 1994 changes were not implemented until now, because the NODUP

SAS option and the MXG _KMXGSUM operand were in conflict;

NODUP dropped observations that were not duplicate, but

looked like dupes because of the reduced variables kept.

So the NODUP option on the SORTs in VMXGSUM were removed

(although a new option NODUP= lets you reinstate them if

you need them and know what you are doing!), and these

changes in the architecture of VMXGSUM were validated:

- The first data step is bypassed if there is no INCODE=,

no INTERVAL=, no SHIFT=, and no NORMx= arguments, and

if there is only one input dataset.

- The PROC SORT is bypassed if there is no SORTBY=.

In another enhancements, three new symbolic arguments

TEMP01,TEMP02,TEMP03 (which default to WORK) are used as

the high-level node of the work datasets, which permits

placing one or more of the work datasets on tape,

reducing the DASD requirement for VMXGSUM processing. If

all three TEMPxx parameters point to the same DDNAME,

then a PROC DATASETS is run at the end of processing to

clean up the temporary space, but if they are different,

we cannot clean up as we have no way to tell if this is a

tape or disk data set. VMXGSUM now uses DKRICOND=NOWARN

as the default (so input datasets with non-existent

variables do not raise a warning), but DKRICOND=WARN is

reset at the end of VMXGSUM execution.
Yet another enhancement is the new argument TIMERNGE=.

If TIMERNGE= is specified, four macro variables are then

constructed to determine the date, time, or datetime

range of the TIMERNGE= argument across the entire input

datasets. The macro variables created are:

MAXINDT - Maximum datetime value for the input data

MAXSLDT - Maximum datetime value for the output data

MININDT - Maximum datetime value for the input data

MINSLDT - Maximum datetime value for the output data

(You must supply the appropriate FORMAT if you print the

value of these four new variables.)
Note: The VMXGSUM architecture was further enhanced in

Change 11.309. This text was edited after that change.

Change 11.280 TYPE39_8 (Netview Session Awareness) LSAWxxxx variables

VMAC39 are wrong because the INPUT statement was off by 4 bytes.

Jan 6, 1994 Also, variables LSAWPSPU and LSAWPSPM were not input.

Change the LSAW section logic to read:

INPUT @OFFSAW REVISION &PIB.2.

+2

LSAWASBC &PIB.4.

- - - - - - - - - - - - - - 17 lines not displayed

LSAWRDQM &PIB.4.

@;

SKIP=LENSAW-80;

IF SKIP GE 8 THEN DO;

INPUT LSAWPSPU &PIB.4.

LSAWPSPM &PIB.4.

@;

SKIP=SKIP-8;

END;

IF SKIP GT 0 THEN INPUT +SKIP @;

END; /* SAW SECTION */

and add LSAWPSPM and LSAWPSPU to the KEEP= list for

TYPE39_8.

Thanks to Hr. Schulte, GRZ, GERMANY.

Change 11.279 MXG 11.07 thru MXG 11.09 did not process TPX Version 3.0

VMACTPX records (but prior MXG versions did), because the LEGENT

Jan 6, 1994 change in format of TPXVER did not document the 3.0 value

to be expected. The conversion code now reads:

IF TPXVER EQ '100 ' THEN TPXVER=' 1.0';

ELSE IF TPXVER EQ '200 ' THEN TPXVER=' 2.0';

ELSE IF TPXVER=:'3.0' THEN TPXVER=' 2.0';

ELSE IF TPXVER=:'3.5' THEN TPXVER=' 3.5';

ELSE IF TPXVER=:'1.0' THEN TPXVER='10.0';

Thanks to Jan Decuypere, Gemeentekrediet N.V., BELGIUM

Change 11.278 Landmark CICS variable TIESDATE (start date) can be null

TYPEMON8 if PTF U200952 is not installed. MXG uses TIESDATE to

Jan 5, 1994 validate that DUMP CONVERT was specified when you dumped

Feb 16, 1994 the Landmark data (see their examples TMON9DBU/TMON9FSU),

and MXG generates the "ERROR3.LANDMARK.MONITOR" error

message "INVALID FORMAT FOR TIESDATE" when TIESDATE is

null. Install the PTF. However, if you encounter that

error message, and know that your data was converted,

you can suppress the ABEND by commenting out the "ABORT

ABEND 1099" statement in member TYPEMON8, until the PTF

is installed. This change text was revised Feb 16, 1994.

Thanks to Mark Holland, State Energy Commission, AUSTRALIA

Change 11.277 If ASUMDB2A is executed separately from BUILDPDB, there

ASUMDB2A will be errors because %INCLUDE SOURCLIB(IMACDB2); did

Jan 5, 1994 not exist in member ASUMDB2A, but now it does.

Thanks to Steve Bryant, Belk Stores Services, USA.

Change 11.276 Support for ZARA Release 1.1 added 36 bytes to the sort

VMACZARA key and header area. You will have zero observations in

Dec 15, 1993 the ZARAVOL and ZARADSN datasets until you change all of

the @xxx values that are now greater than 100 by adding

36 to the present value. Thus the @172 VRCDTYP became

@208 VRCDTYP, the @183 VOLCHAIN became @219 VOLCHAIN,

and so forth.

Thanks to David Childress, Lowe's Companies, USA.

Change 11.275 IBM APAR OY67002 reports that if the number of LCPUs is

VMAC7072 changed in an LPAR, the data for that RMF interval

Dec 15, 1993 contains the accumulated values from IPL, and there are

no PR/SM segments in the type 70 for the interval.

This corrupts TYPE7072 variables LCPUEDTM,LCPUPDTM,

PCTCPBYx,PCTTIPx, corrupts TYPE70 variables CPUEDTMx,

CPUPDTMx,PCTCPBYx, and PCTTPIx, and corrupts essentially

all ASUM70PR variables for that interval. MXG detects

that the number of LCPUs changed, and sets LPARCHRN='Y'

in TYPE70PR (the Label for LPARCHNR now reads NUMBER OF

LCPUS*IN PARTITION*CHANGED?), but you MUST install the

PTF associated with the APAR if you change the number of

LCPUs in an LPAR while the system is up.

Thanks to Gary Hoover, American Express IPS, USA.

Thanks to Roger Zimmerman, Kemper Services Company, USA.
Change 11.274 Three new variables were added to the SAS User SMF record

VMACSASU in SAS 6.07/6.08. Variables SASVAFF (Vector Affinity CPU)

Dec 15, 1993 SASVUSE (Vector Usage CPU), and SASHSP (Hiperspace CPU)

are now input and formatted TIME12.2.

Thanks to Bruce Lietz, Cessna Aircraft Company, USA.
Change 11.273 Some TRND71 variables were not calculated. In the NORM2=

TRND71 list, the first occurrence of SLOTNVAV should be SLOTNGAV

Dec 15, 1993 and variables ASMNVSC ASMSLOTS ASMSLOTX and ASMVSC were

added to the NORM2= list.

Thanks to Carl Tosetto, E-Systems Garland Division, USA.
Change 11.272 TYPE64 records with SITUATN='COMPONENT CLOSED' are the

ADOC64 only valid records to use for counting EXCPS, etc. The

ANALDSET type 64 records with NO SPACE AVAIL or VOL SWITCHED do

Dec 15, 1993 not contain the "delta" EXCP count, but instead have the

accumulated EXCP count from open to the time of that

event. As a result of this discovery, member ANALDSET

now keeps only TYPE64 observations for Component Closed.

Thanks to Scott Ashby, Wachovia Operational Services Corp, USA.

Change 11.271 Statement DEVASID DS XL2 has an asterisk in column

ASMTAPES 72 which causes the Assembly to fail. This member is

Dec 15, 1993 still in internal testing. See Change 11.360.
Change 11.270 Support for Landmark CICS/ESA Version 1.1 records. That

TYPEMON8 new version added 44 fields to the end of the MONISYST

Dec 14, 1993 interval record, causing "INVALID DO LOOP CONTROL" error.

This error can be circumvented by inserting "HITI=1;"

just before the "DO _I_=1 TO HITI;" statement, but you

will then get "INVALID DATA FOR TASKCPTM" and hex dumps

twice, because Landmark's field TIAPLCPU (MXG variable

TASKCPTM) is invalid until you install Landmark ZAP for

Problem Number U201102. (You can still process without

their ZAP as all other variables in MONISYST are valid,

and only the MONISYST data set was affected by the new

version!). In the actual change, the DO group to read

the history segment was removed, as that segment has not

existed for some time, and MONIHIST always has zero

observations. However, I did not delete the MONIHIST

dataset nor its _L,_K macro names, since that might have

caused a syntax error if you reference MONIHIST in your

report programs. Take heed, though, and remove MONIHIST

references so that I can deleted it in a later version.

Setting HITI=1 has the same effect as skipping the

history record; the TI record is now 1464 bytes long.

Thanks to Mark Holland, State Energy Commission of Western Australia.

Change 11.269 Dataset PDB.JOBS variables ACCOUNTn may be blank, and

BUILDPDB variables RESTARTS, JINLTIME, and NRTRANS can be missing

BUILDPD3 for jobs with MULTIDD='Y' (that is, multiple type 30s due

BUILD005 to many DD segments, typical only for long running jobs

Dec 13, 1993 like SAR, RMDS, CA-DISPATCH, etc., that have lots of

dynamic DDs). The MULTIDD='Y' observations must be

deleted during the second occurrence of the creation of

GOOD30_5 dataset. Find this block of code:

ELSE IF IN30_5 THEN DO:

IF MULTIDD5=' ' THEN RESTART+1;

OUTPUT GOOD30_5;

END;

And change it to read:

ELSE IF IN30_5 THEN DO:

IF MULTIDD5='Y' THEN DELETE;

RESTART+1;

OUTPUT GOOD30_5;

END;

The actual change also cleaned up references to MULTIDD

and TEMP3 which are no longer needed, but which are now

superfluous and safe to leave until you get a tape!

Thanks to Mark van der Eynden, Ferntree Computer Services, AUSTRALIA.

Change 11.268 CICS DL/I counts with CICS/ESA Version 3.3 Omegamon 110

IMACICDL records may be wrong. First, the test in IMACICDL needs

IMACICDA to be changed to IF SMFPSRVR GE 3 instead of EQ 3.

VMAC110 Second, if you have not added a "USERCHAR" optional data

Dec 14, 1993 field in your type 110 record, you must comment out the

statement "%INCLUDE SOURCLIB(IMACICDU);" in member

IMACICDA (because if that statement is present, MXG

always read at least one byte, causing wrong alignment).

Commenting out that %INCLUDE will generate an

"UNINITIALIZED VARIABLE USERCHAR" message, which is

not a problem, but in member VMAC110, I have added

"IF USERCHAR=" " THEN USERCHAR=" "; after both of the

%%INCLUDE SOURCLIB(IMACICDA); statements to correct.

The notes in IMACICDA and IMACICDL were updated to list

the CMODHEAD field names of the DL/I counters to make it

easier for you to know what sections you do or do not

have in your SMF type 110 record from Omegamon, using MXG

program UTILCICS to examine your dictionary.

This text was revised Dec 14, 1993, after MXG 11.09.

Thanks to Khin Zaw, Nordstrom, USA.

Change 11.267 APAF V2.1 dataset APAFCHAN is trashed, because the offset

VMACAPAF had a spelling error. Inside the SUBTYPE=5 DO group, the

Dec 13, 1993 statement LOC=OFFLP-3+OFFSMF should have been written as

LOC=OFFCHPID-3+OFFSMF;

Thanks to George Scott, Rockwell International Corporation, USA.
==Changes thru 11.266 were included in MXG 11.09 dated Dec 17, 1993=====
Change 11.266 Dataset TYPE1415 variable TEMP='TEMP' was incorrectly set

VMAC1415 if DISP1='.......1'B, so that part of the test to set

Dec 12, 1993 TEMP has been removed.

Thanks to Bruce Hudson, Payless Shoe Source, USA.

Change 11.265 Boole and Babbage CMF Type 72 Subtype 2 record is invalid

VMAC7072 and causes INPUT STATEMENT EXCEEDED error. There is only

Dec 10, 1993 one SWAP section, instead of one SWAP section for each

Performance Group, and other fields in TYPE72MN dataset

may be trashed. Boole anticipates a fix in their 9402

maintenance release. MXG changed IF OFFSWP GT 0 THEN DO;

to IF OFFSWP GT 0 AND OFFSWP+60 LE LENGTH+1 THEN DO; to

prevent the STOPOVER condition.

Thanks to Gavin Ring, Alcatel Australia, Australia
Change 11.264 Variable PGPERBLK in RMFINTRV was incorrectly normalized.

RMFINTRV The two lines in which PGPERBLK was multiplied/divided by

Dec 8, 1993 DURATM were deleted, and the division by BLKSAUIN was

protected for zero divide.

Thanks to Dan Squillace, SAS Institute Cary, USA.
Change 11.263 Members CHANGE01,CHANGE02... are redundant with CHANGESS,

CHANGESS which contains all changes ever made to MXG in one place,

Dec 4, 1993 so the individual CHANGEnn members now only point to the

member CHANGESS, reducing the size of the MXG library.

Searching for all MXG enhancements now involves looking

only two members: CHANGESS or NEWSLTRS. Member CHANGES

will continue to exist, repeated in CHANGESS.
Change 11.262 Support for Softworks' Performance Solution SMF records.

ADOCPRFS Three new datasets are created:

EXPRFSHL Dataset SMF Subtype Description

EXPRFSVI PRFSVIO 00 I/O Plus for VSAM Close

EXPRFSXI PRFSXIO 01 I/O Plus for XSAM Close

IMACPRFS PRFSHLP 02 HiperLoad Plus for VSAM

TYPEPRFS These data records describe the before and after buffers,

VMACPRFS etc., to track the effectiveness of this product.

Dec 2, 1993

Thanks to Yao-chun Rickert, First Chicago, USA.

Change 11.261 VM/ESA dataset VXSYTCUP (PR/SM and MLPF LPAR measurement)

VMACVMXA did not keep LCUCLPTM (LPAR Management time), so it was

Dec 1, 1993 added to the KEEP= list in macro _VSYTCUP, and is DIF'ed

in macro _DSYTCUP. Also, for inactive LPARs, MXG output

an observation, but variables LCUCPUID LCUCWGHT LCUCFLGS

LCUCACTM LCUCLPTM LCPUDED LCPUWAIT LCPUCAPD are not input

and could contain values from a prior LPAR; now they are

initialized. Similarly, VXSYTCUM logic for inactive LPARs

now initialized the un-read variables.

Thanks to Angie Olver, PERSETEL, South Africa.

Change 11.260 SAS Usage note 6886 identifies a SAS error when blank

ANALRACF characters appear between non-blank characters in an ID

Dec 1, 1993 value with PROC TRANSPOSE. In 6.06 the variables were

incorrectly left as blanks instead of being replaced by

underscores (as they were in 5.18, and are now in 6.07

and 6.08), and MXG code was actually modified between

5.18 and 6.06 to compensate. Now, however, UNINITIALIZED

variable messages result, so MXG has been re-modified to

correct. Variables _5_DATA, _6_RACF, _7_DATA,_8_NAME,

and _52_DSN are now spelled with an ending underscore:

_5_DATA_,_6_RACF_,_7_DATA_,_8_NAME_, and _52_DATA_.

Thanks to David Vaughan, Shropshire County Council, ENGLAND.

Change 11.259 A collection of data value errors have cropped up in the

VMXGHSM processing of the HSM BCDS and MCDS datasets:

Dec 3, 1993 -"Error 180" results because a semicolon is in the wrong

place. The semicolon after RECTYPE='REC TYPE ...' must

be removed, and instead a semicolon must be added to the

following line, after the label for TYPE='MCKTYPE ...'.

-"Invalid data for INPUT" results because "INPUT" should

removed from the line now reading "INPUT @65 BCRFLAGS..."

-Invalid argument for DATEJUL occurs if MCDEXPDT=99366,

95366,98000,99999,9999999 (which are invalid dates). The

calculation for MCDEXPDT was changed to set these invalid

data values to '31DEC2099'D (i.e., far into the future):

IF MCDEXPDT=99366 OR MCDEXPDT=95366 OR ...

THEN MCDEXPDT='31DEC2099'D;

ELSE IF MCDEXPDT GT 0 THEN MCDEXPDT=DATEJUL(MCDEXPDT);

ELSE MCDEXPDT=.;

-MHCRBGNT,MHCRENDT,and MHCRNXTT are now input &PIB.4.2 and

not &PK.4, the three PK1 fields after each time were

deleted, the +1 before the three dates was removed, and

the three HMS() calculations were deleted.

-DATEJUL(MHCRENDD) was changed to DATEJUL(MHCRNXTD).

-Two lines calculating DSRDATE were moved to before the

line DO I=1 to 11;

-Change DO C=1 TO MCPDGNCT; to read

IF MCPDGNCT GT 0 THEN DO C=1 TO MCPDGNCT;

-Change input of BCRTCAB from &PK.4. to BCRTCABC $CHAR4.,

delete input of BCRTCABH,BCRTCABM,BCRTCABH and following

+1, replace the BCRTCAB=HMS() line with

BCRTCAB=INPUT(BCRTCABC!!'00000000'X,TODSTAMP8.);

remove BCRTCAB from TIME8. format and instead format it

BCRTCAB DATETIME21.2.

Note: Dec 18, after 11.09 built: used similar logic for

BCRTBLA and BCRTLAB datetime variables.

-Change input of DCRCLEN from &PIB.4. to &PIB.2.

Thanks to Norbert Korsche, OMV-AG, AUSTRIA.
Change 11.258 Variables ACTDLYTM,DSPDLYTM,RESDLYTM were added by Change

IMACPDB 10.031 to TYPE30_4, but were not carried into PDB.STEPS

Nov 30, 1993 nor PDB.JOBS. They have now been added to the _PDB30_4

and _SUMSTP macros defined in IMACPDB, so they will now

exist in PDB.STEPS and PDB.JOBS.

Thanks to Bob Eastlake, Alternative Marketing Systems, USA.

Thanks to Tom Elbert, John Alden Life Insurance, USA.
Change 11.257 ASMVTOC may produce no output under MVS/ESA 3.1.3, 4.1.0,

ASMVTOC or possibly even 4.2.0. APAR OY39355 points out that DFP

Nov 22, 1993 3.3 (which introduced the new UCBSCAN COPY service) does

not run with earlier levels of MVS/ESA. If you have this

problem (no error, but no output from ASMVTOC), you can

simply comment out (by adding an asterisk in column one)

these two lines in ASMVTOC:

TM FLAG,MVSESA42 MVS/ESA V4.2 OR BEYOND?

BO MVS42AAB YES, GO USE UCBSCAN SERVICE

ASMVTOC has also failed with 213-04 when it encounters a

TPF volume. That problem is still under investigation.

Thanks to Ron Willingham, Fina Oil, USA.

Change 11.256 RMF Reports from PDB data are further corrected and

ANALRMFR enhanced. CPU percentages could be wrong, and type 74

Nov 21, 1993 reports were corrected.
Change 11.255 Support for DB2 Release 3.1.0 incompatibly altered SMF

type 100, 101, and 102 SMF records.

== added 11Jan95=======

== Messages on the log "ID=100 SUBTYPE=2 QWHSIID=202"

== tell you that you are reading DB2 3.1 SMF records

== with an old Version of MXG (MXG 11.09 or later is

== required).

ADOCDB2 100, 101, and 102 SMF records. New Datasets created are:

DIFFDB2 Dataset SMF Record Description

EXDB2ACP DB2ACCTP 101 Subtype 1 Package/DBRM executions

EXDB2ST2 DB2STAT2 100 Subtype 2 Hiperpool interval statistics

IMACDB2 Many new variables were also added to existing datasets.

VMACDB2 New variables are listed in member DOCVER11, and ADOCDB2

VMAC102 was updated with which DSECT is in which record and which

Sep 6, 1993 MXG dataset is created from which records. Details:

Nov 30, 1993 Dataset DB2ACCT:

Now contains 312 variables (was 215). New variables in

existing sections QWAC,QXST,QBAC,QTXA,and QLAC, and

new section QIFA. Variables QBACBPX and QBACSWU are

now reserved.

Dataset DB2ACCTP:

New dataset for Package/DBRM executions is built from

QPAC section(s) added to type 101 subtype 0 record (up

to ten sections will create ten observations), or from

new type 101 subtype 1 record (up to 10 per record, for

overflow if sections do not fit in existing subtype 0).

Dataset DB2STAT0:

Now contains 319 variables (was 277). New variables in

existing sections Q9ST,QWSD,QLST,QJST and QDST.

Variable QJSTWTL is now reserved.

Dataset DB2STAT1:

Now contains 447 variables (was 297). New variables in

existing sections QXST,QTST,QBST,QIST, and QTXA.

Variables QTCURPB,QTOPNOK,QTOPNNO,QTTTBRN,QTEXDRN,

QTSTDRN,QTPUBDD,QBSTBPX,QBSTSWU,and QBSTPUW are now

reserved.

Dataset DB2STAT2:

New dataset from new type 100 subtype 2 record.

Dataset T102S148:

New variables: QBnCDPF,QBnCHPG,QBnCHRE,QBnCHRF,QBnCHWF,

QBnCHWR,QBnCNGT,QBnCSIO n=1,2,3,4

Dropped vars: QBnCBPX,QBnCSWU n=1,2,3,4

Change 11.254 Support for AS/400 Version 2.3 Performance Data adds four

EXQAPDDI new datasets:

EXQAPFRL Dataset Description

EXQAPSTD QAPMDDI Distributed Data Interface (DDI) data