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

VMACBETA thru 4 have been tested; there are some relatively minor

Aug 10, 1994 questions that were sent to the vendor for answers that

will cause another change, but with this change, the data

for the tested subtypes seems valid. Note, however, that

subtypes 2, 3, and 4 are pairs of records, one for Start

and one for End, with most activity counts identical in

both records - be careful not to double account! Feb 95:

Beta 93 Records for Version 1.6.22 have now been tested.

Thanks to Manfred Thomas, BHF-Bank, GERMANY.
Change 12.131 In support of the SAS/CPE Product, variable DATETIME has

BUILDPDB been kept in datasets PDB.JOBS, PDB.STEPS, and PDB.PRINT.

BUILD005 DATETIME is the datetimestamp that was used to calculate

Aug 10, 1994 the shift of the job, step, or print event.

Thanks to Tricia Dudley, SAS Cary, USA.
Change 12.130 Dataset MONIMRO variable TAMRTIME was incorrect because

TYPEMON8 the statement TAMRTIME=64*TAMRTIME should have been

Aug 9, 1994 before the %%INCLUDE SOURCLIB(EXMONMRO) instead of just

after it.

Thanks to Dave Konz, Dreyfus, USA.
Change 12.129 ASMIMSLG program may ABEND 002 on DDNAME of IMSMPRS if

ASMIMSLG the IMS log being read contains records greater than 248

Aug 8, 1994 bytes (some Fast Path records are), because the DCB for

this output file was specified in the ASMIMSLG program.

To correct, delete the line MVC DCBLRECL,IMSTY7L

(so that there is no LRECL is specified in the program);

since the JCLIMSLG example has always had a DCB parameter

of RECFM=VB,LRECL=23572,BLKSIZE=23576, that DCB will be

used instead.

Thanks to Ronnie Stake at NORDBANKEN, SWEDEN.

===Changes thru 12.128 were in MXG PreRelease 12.03 dated Aug 4, 1994===

===Changes thru 12.128 were printed in MXG Newsletter 26 Aug 4, 1994===

Change 12.128 Continued corrections and validation of IBM RMF reports.

ANALRMFR Major change was to revert back to use the Partition pct

Aug 3, 1994 CPU busy (instead of Effective CPU busy) because IBM has

again changed back, (and I want to match their reports!).

Some minor corrections in spelling and labelling too!
Change 12.127 Continued corrections and validation of CICS DFHSTUP-like

ANALCISH reports. Fixed problems are:

Aug 3, 1994 No "Terminal" Report.

WORK.CICAUSS.DATA does not exist.

Missing System and Transaction Dump Stats.

Missing Interval Report for "Task Subpools and DSA"

Thanks to Kusol Voratanitkitkul, Health Care Service Corp, USA.
Change 12.126 MXG 12.01/MXG 12.02 only. DB2 Buffer Pool data for DB2

VMACDB2 2.3 pools BP2 and BP32 and for DB2 3.1 pools with BPID=3

Aug 3, 1994 or greater (these are MXG variables QB3Caaaa/QB4Caaaa in

DB2ACCT, and variables QB3Taaaa/QB4Taaaa in DB2STAT1) are

missing because Change 12.033 did not initialize the 3rd

and 4th sets of variables. Many lines were involved in

this change: the new logic for the 3rd and 4th sets of

variables now tests if QB3CPID=. or QB4CPID=. (for

DB2ACCT logic) or QB3TPID or QB4TPID=. (for DB2STAT1

logic), and if true, a new DO group initializes the full

set of variables (including storing into QB3CPID QB4CPID,

QB3TPID or QB4TPID) to initialization only once, and the

existing sum logic is wrapped in a new ELSE DO group.

This should have been caught in my testing, but I did not

have test data with these higher buffer pool IDs.

Thanks to Siegfried Trantes, Gothaer Versicherungsbank, GERMANY.

Thanks to Waldemar Schneider, SAS Europe, GERMANY.

Thanks to Lori A. Masulis, Fidelity Systems, USA.

Change 12.125 Support for BETA93 Report Distribution & Print Management

EXTYBET0-6 System 1.6.0 user SMF record creates six datasets:

IMACBETA BETA0 - Subtype 0 - Input of Lists

TYPEBETA BETA1 - Subtype 1 - Output of Lists

VMACBETA BETA2 - Subtype 2 - Bundling

Aug 3, 1994 BETA3 - Subtype 3 - Recipient Packeting

BETA4 - Subtype 4 - Address Packeting

BETA5 - Subtype 5 - Action Performed Against List

BETA6 - Subtype 6 - Completion of Print Request

Thanks to Manfred Thomas, BHF-Bank, GERMANY.

Change 12.124 DB2 Trace Paring for IFCIDS 15,16,17,18 mis-matched scans

ANALDBTR when there were multiple BEGIN SCANs before an END SCAN.

VMAC102 This was an extensive logical changed; the first three

Aug 3, 1994 bytes of the CUB (QW0015AC,16AC,17AC,18AC) is now stored

into new variable CUBTOKEN and kept in the four T102S0xx

datasets in VMAC102, and ANALDBTR logic for _015PAIR was

revised to insert CUBTOKEN in all BY statements after

QWHCCV, and CUBTOKEN was added to the six KEEP= lists for

the datasets inside the _015PAIR macro. The CUBTOKEN was

created because the final byte of QW0018AC was a hex zero

while the final byte of the other QW001xAC fields was

found to contain a wide range of non-zero values.

(In addition, QW0018AC's INPUT in VMAC102 was changed to

$CHAR4. instead of &PIB.4.; all other CUB fields had been

input as characters. And finally, while cleaning up the

CUB fields that end in AC, I noted that other IFCIDs have

a field ending in AC (the ACE Token) that was mostly CHAR

but sometimes &PIB, so all occurrences were revised to be

input $CHAR4., formatted $HEX8., and added to $NOTRAN.)

Thanks to Dave Witkowski, Amdahl, USA.

Change 12.123 Support for Omegamon II for SMS V100 and V110. Subtype

VMACOMSM 2 record was restructured (compatibly), and new subtype 3

Aug 3, 1994 record was added, but since the subtype 3 replaces the

subtype 1 record in V110, you will need to install the

new MXG version before you process V110 records.
Change 12.122 Support for APAR OW00157 adds variables SMF62MGT,SMF62STR

VMAC62 and SMF62DAT (Management Class, Storage Class, and Data

Aug 3, 1994 Class names) to the VSAM Open Record. Unfortunately, the

fields would be far more useful in TYPE64 Close Records!

Change 12.121 NETSPY 4.4 added four fields to the NCP record that i did

VMACNSPY not decode; the four fields are all PIB4. and replace the

Aug 2, 1994 first 16 bytes of the +26 toward the end of the NCP code:

CFBQUELN='CURRENT*FREE BUFFER*QUEUE LENGTH'

FBQLNHWM='FREE BUFFER*QUEUE LENGTH*HIGH WATER MARK'

FBQLNLWM='FREE BUFFER*QUEUE LENGTH*LOW WATER MARK'

TFQBANCP='TOTAL*FREE QUEUE BUFFS*ALLOC IN THE NCP'

Note that LEGENT told Roger that for NCP Version 6.2, the

MAXFREEB (Max Free Buffers) is now moved to MXG variable

TFQBANCP (at offset 47). I chose to keep both variables.

Thanks to Roger Zimmerman, Kemper Service Company, USA.
Change 12.120 Support for ASTEX 2.0 adds several fields; the new record

VMACDMON can be read by prior versions of MXG, but variable RDUA

Aug 2, 1994 (device unit address) will be unreadable without the MXG

changes (it was EBCDIC, now it's binary, so MXG decodes

the binary back to EBCDIC). Several other new variables

were added to the record (compatibly). I replicated the

six "May/Must Cache" variables that were misspelled with

"H" to be correctly spelled, but also keep the old named

misspelled variables.
Change 12.119 The REXX example statement containing AND FORMAT IT AS

REXXPDB needs to be shortened and an ending quote added to the

Aug 2, 1994 line.

Thanks to Waldemar Schneider, SAS Europe, GERMANY.

Change 12.118 MVS/ESA 5.1 TYPE72GO dataset RTSTRNnn variables may have

VMAC7072 counts for a Service Class Period that does not have any

Aug 2, 1994 response time goal; MXG logic did not properly match the

SCS section with the appropriate RTS section, instead it

carried the RTSTRNnn variables forward. The observations

with bad counts can be identified; they have R723CRTX=0,

but this change corrects the logic, and also corrected a

related problem with R723TYPE for these observations.

Thanks to Don Deese, (CPExpert), Computer Management Sciences, USA.
Change 12.117 Support for SMF type 99 (Workload Manager Trace) Subtype

EXTY99U2 2 creates dataset TYPE99_2 for each Service Class Period

IMAC99 with lots of useful detail about what's really happening

TYPE99 inside the new Workload Manager (for example, you can see

VMAC99 the changes in dispatch priority for each workload with a

Aug 1, 1994 "PROC PLOT DATA=TYPE99_2; PLOT PBDP*SMFTIME=SRVCLASS);".

There are many other subtypes in the type 99, many are so

detailed as to be questionable in value, but as the need

arises, additional subtypes will be decoded.

Thanks to Peter Enrico, IBM Poughkeepsie, USA.

Change 12.116 If ANALDB2R was run with only PMACC04=YES specified, an

ANALDB2R unresolved macro reference for MAXINDT was produced. The

Aug 1, 1994 code was corrected. If any other PMACCnn was requested,

along with PMACC04, the error did not occur.

Thanks to Perry Lim, AVCO, USA.
Change 12.115 Monthly trending members MNTHxxxx were revised to take

MNTH70PR advantage of the enhancements made to VMXGSUM by earlier

Aug 1, 1994 Change 12.084; minor cosmetic improvements also made.
Change 12.114 Minor change to this macro now can select ranges of DB2

READDB2 IFCID values.

Aug 1, 1994
Change 12.113 UTILGETM is used to create the SMFSMALL subset file with

UTILGETM JCLTEST6; this revision was really for me, so I can give

VMXGGETM you a trivial invocation of VMXGGETM when I need to see a

ZMXGGETM particular SMF record. UTILGETM now just invokes the new

Aug 1, 1994 VMXGGETM macro with default parameters. You might find

the new macro useful if you are chasing raw SMF records!

Because the new UTILGETM won't work with SAS Version 5,

the original UTILGETM was copied/renamed into ZMXGGETM.

Change 12.112 Support for UniKix Release 4.1 now creates three datasets

FORMATS from UniKix Technologies "CICS-like-under-UNIX"

VMACUNIA product's accounting files:

VMACUNIK UNIKTRAN - Transaction record, one per transaction for

ZMACUNIA those transactions that are being monitored

ZMACUNIK UNIKUSER - User record, one per transaction for those

Jul 29, 1994 users whose transactions are being monitored.

UNIKEVNT - Event record (STARTUP/SHUTDOWN/etc.)

These three datasets are created from either the Unix

Binary file (using TYPEUNIK to read INFILE UNIKIXBI under

UNIX/WINDOWS/OS/2 versions of SAS), or the three datasets

are also created from the UniKix Account Conversion

Program's output ASCII file (using TYPEUNIA to read the

INFILE UNIKIXAS, which can execute under any version of

SAS, so that the ASCII file could be uploaded first and

then processed with mainframe SAS; however, the ASCII

file is twice as large as the binary file). Note that it

is possible to do double accounting; if a monitored user

executes a monitored transaction, there will be two obs

created for that event, one in UNIKUSER and in UNIKTRAN.

I have used common CICS variable names where possible.

See Chapter 10 of the UniKix Administrator's Guide for

more documentation. data fields. The MXG code for the

earlier UniKix Release 4.0 was copied into the members

ZMACUNIA/ZMACUNIK (see comments on their usage); note

that that old code creates only a single dataset,

TYPEUNIK or TYPEUNIA, and all three types of observations

(event, transaction, and user accounting) are contained

in that dataset. Last minute note for MXG 12.03: Only

the Binary support was finished. ASCII (VMACUNIA) will

be completed in next MXG release.
Change 12.112A VARIABLE BCRTBLA HAS ALREADY BEEN DEFINED AS NUMERIC

VMXGHSM results from insufficient testing after Change 12.030!

Jul 29, 1994 Variables BCRTBLA,BCRTLAB, and BCRTCAB must be spelled

BCRTBLX,BCRTLAX, and BCRTCAX in three lines where they

INPUT as CHAR4., and in three lines where they are inside

INPUT functions (i.e., on right hand side of the equality

but they are correctly spelled on the left hand side).

Also, LENGTH BCRTBLA BCRTLAB BCRTCAB 8; was added.

Thanks to Peter MacDonald, USF&G, USA.
Change 12.111 APAR PN56441 corrects DB2 IFCID 239 record, which was

VMAC102 erroneously written as an SMF ID=102 record; it is fact

Jul 27, 1994 the ID=101 SUBTYPE=1 record that creates DB2ACCTP, and

already supported in MXG. Only comments were revised.

Change 12.110 APAR OW01522 adds new value 7 for Extended Sequential to

FORMATS format MG042DS (for variable DSTYPE); the APAR text also

Jul 27, 1994 states that all values will be documented in IGWDSSBC.
Change 12.109 The PAGRT calculation did not match IBMs Paging Report.

VMAC71 Variable PGMIEXAU was added to PAGRT and should not have

Jul 27, 1994 been, and variable PGBKAUIN should have been added but

wasn't! The MXG code changes were to delete the line

PAGRT=PAGRT+PGMIEXAU; and insert PAGRT=PAGRT+PGBAKUIN;

after the line reading DPR=DPR+PGBKAUIN;

Thanks to Ms Foo Hooi Ping, Maybank, MALAYSIA.
Change 12.108 Variable MULTIDD was left out of the KEEP= list for the

BUILD005 PROC SORT of TYPE30_5.

Jul 27, 1994
Change 12.107 MXG CRITICAL ERROR. UNEXPECTED TCP/IP DATA VALUES will

IMACTCP result if you specify a value of 118 in member IMACTCP

VMACTCP for either macro _IDTCPF or _IDTCPT. The comments in the

Jul 26, 1994 IMACTCP member were strengthened that these macros are

only for the archaic version of TCP/IP, but also MXG is

now robust against this error - the logic in VMACTCP now

tests automatically for type 118 before examining what

you specified in those macros, so this won't recur!

Thanks to Dov Brosh, Solomon, Inc, USA.

Thanks to Rich Warren, Solomon, Inc, USA.

Change 12.106 Continued validation of DFHSTUP-like reports uncovered

ANALCISH more minor glitches. Following the first occurrence of

Jul 25, 1994 &CAUS and &CAUT, the macros CICAUTO and CICAUSS need to

be reversed - CICAUSS is invoked after &CAUS and CICAUTO

is to be invoked after &CAUT. The four tests for INOBS=

SDG,SDR,TDG, and TDR must be changed to CSGD,CSDR,CTDG,

and CTDR respectively. After both occurrences of

PROCESS(MXGDS=CICTCR); insert %RPTTCR(MXGDS=CICTCR);

Thanks to Kusol Voratanitkitkul, Health Care Service Corp, USA.
Change 12.105 MXG Tape Mount and Tape Allocation Monitor, ASMTAPES, is

ASMTAPES now healed of all known glitches! This change corrects

Jul 25, 1994 the MIM flag (mounts were always indicated as from MIM),

and correctly identifies PRIVAT/DEMAND mounts (once a UCB

has a PRIVAT mount, all subsequent mounts were flagged as

PRIVAT). With this change, I believe ASMTAPES is safe to

install in production, in place of the old ASMTMNT that

only tracked tape mounts.

- after the line

MNTPEND OI DEVFLGS,DEVPEND FIRS REMEMBER ITS PENDING

insert this new line;

NI DEVFLGS,X'FF'-DEVPVT-DEVSCR RESET PRIVAT/SCRTCH

- the third line after label SETJOB, now containing:

LH R4,UCBASID-UCBCMEXT(,R4) ALL SO WE CAN GET THE ASID

is replaced by these three lines:

ICM R4,B'0011',UCBASID-UCBCMEXT(R4) GET THE ASID

ICM R4,B'0100',=X'00' ZERO OUT HIGH-ORDER HALF OF R4

STH R4,DEVASID STORE THE ASID FOR MIM CHECK

Thanks to Bob Kinney, Kaiser Permanente, USA.

Thanks to Mike McClean, Travelers, USA.

Change 12.104 Support for EMPACT's HIPER-CACHE Version 1.1.1 (INCOMPAT)

VMACHIPR inserted 24 bytes (JCTJOBID,INITTIME,READTIME) in the

Jul 21, 1994 header. You can process the new version's SMF records

(but not the old!) by changing the +2 after the INPUT of

SMFVERS to +26. The actual change creates the three new

variables.

Thanks to Neil Ervin, Huntington National Bank, USA.
Change 12.103 DB2 Trace IFCID=141 (dataset T102S141) was not correctly

FORMATS decoded for DB2 2.3/3.1, and the SQL text was not read.

VMAC102 Change the statement:

Jul 20, 1994 IF QWHSRELN GT 2.2 THEN INPUT QW0141CO &IB.4.

to read:

IF QWHSRELN GT 2.2 THEN INPUT +2 QW0141CO &IB.4.

After INPUT of QW0141HO $CHAR8. and before the @; insert

QW0141LL &PIB.2.

so that the length of SQL text will be read in.

Thanks to Ricky Wafford, MBNA, USA.

Change 12.102 APAR OW04453 corrects SMF type 50 VTAM Tuning Statistics

FORMATS documentation and I discovered that the type 50 record

VMAC50 is actually four different structures, described in VTAM

Jul 20, 1994 Network Implementation Guide (SC31-6419) Table 29 (SNA

Controller), Table 30 (Channel-to-Channel) and Table 31

(Multipath Channel Adapter, MPNCB or CPNCB). All fields

are now decoded, and variable VERSN50 identifies whether

the record is for SNA, CTC, or MPC, and variable ATTCHTYP

identifies if an MPC record is an MPNCB or a CPNCB.
Change 12.101 Documentation of common error messages is now provided in

INSTALL a new section zero at the beginning of member INSTALL.

Jul 20, 1994 These messages are now described, with more detail:

APPARENT SYMBOLIC REFERENCE PIB NOT FOUND

(missing or old dsname for CONFIG= or //CONFIG DD)

INFORMAT $NOTRAN or FORMAT $MGxxxxx or MGxxxxx NOT FOUND

(missing or old dsname for //LIBRARY DD)

LENGTH OF CHARACTER VARIABLE HAS BEEN SET

(has no effect, to be eliminated by SAS, disregard)
Change 12.100 Documentation of the output reports produced by this CICS

UTILCICS utility that prints the contents of your CICS Dictionary

Jul 19, 1994 records is now extensive, with examples, in comments.

Thanks to Danilo Gipponi, SAS Institute, ITALY.

Thanks to Werner Bundsuhuh, SAS Institute Europe, GERMANY.
Change 12.099 Protection for invalid expiration dates (96366,97366) in

VMACDCOL both DCOLLECT and HSM dataset processing now sets the

VMXGHSM xxEXPDT variables to '31DEC2099' when the day-part of the

Jul 18, 1994 Julian date is 366 or greater. (See Change 14.124).

Thanks to Norbert Korsche, OMV, AUSTRIA.
Change 12.098 DB2 report PMSQL01 or PMSQL02 cause MISSING BY STATEMENT

ANALDB2R because IFCID 54 was not included in the list defined in

Jul 18, 1994 MACRO _PMSQL01. Add 54 to that list and the syntax error

will go away.

Thanks to Dave Witkowski, Amdahl, USA.
Change 12.097 NPM Release 2.0 records cause INVALID DATA FOR MGBYTES

VMAC28 and/or INPUT STATEMENT EXCEEDED RECORD LENGTH and/or

Jul 18, 1994 NONZERO COF SEGMENT for NPMSUBTY=214 thru 219, because

no test data was available for validation to catch my

careless errors in the new VTAM monitoring subtypes:

-Remove both occurrences of " MGBYTES " (without a period)

but do not remove the one occurrence of " MGBYTES." (with

a period).

-Remove one occurrence of "TIME 13.3" (with blank between

TIME and 13.3), but do not remove the one occurrence of

TIME13.3 (without the blank).

-The line now reading:

ELSE IF 0D0X LE NPMSUBTY LE 0D5X THEN COFTYPE='VCD';

should read:

ELSE IF 0D0X LE NPMSUBTY LE 0DBX THEN COFTYPE='VCD';

Also, now with real data, variable VADSTIME in dataset

NPMVSVAD is now properly decoded into a datetimestamp.

Thanks to Tom Parquette, Mutual of New York, USA.

Change 12.096 Support for RSD's WSF Release 3.5.1 (INCOMPATIBLE)! The

FORMATS new records do not cause any MXG error, but the dataset

VMACWSF WSFEVTPR was trashed because ACCPRNAM was expanded from 8

Jul 17, 1994 to 16 bytes! For a workaround to read the new records

(but not the old ones), you can change then INPUT of

WSFEVTPR from $EBCDIC8. to $EBCDIC16. The actual change

added a new variable and its FORMAT, corrected MXG logic

that did not read the Account Extension segment, and now

blanks variables that exist only in the ERD records.

Thanks to M. Kirassian, GIA Rhone Alpes, FRANCE.

Change 12.095 DB2 report PMACC02 variables QWACAWTW/QWACAWTR/QWACAWTE/

ANALDB2R QWACALOG were incorrectly normalized. In two places,

Jul 15, 1994 in the ELSE statement following IF QWACAWTR = ., change

both occurrences of QWACAWTW to QWACAWTR. In two places,

in the ELSE statement following IF QWACAWTE = ., change

both occurrences of QWACALOG to QWACAWTE.

Thanks to Caron Knox, Willis Corroon Group, ENGLAND.
Change 12.094 Dataset TYPE30_V was not deleted (due only to oversight)

BUILDPDB after it was copied and renamed into PDB.SMFINTRV in the

BUILDPD3 BUILDPDB/BUILDPD3 logic, so it has now been added to the

BUILD005 DELETE statement of the PROC DATASETS following its last

BUIL3005 reference.

Jul 15, 1994

Thanks to Diane Eppestine, Southwestern Bell, USA
Change 12.093 Cosmetic correction to spelling in labels for variables

VMACNSPY NSPETHR and NSPLFCNT AND LUGTRGT1-LUGTRGT4.

Jul 15, 1994

Thanks to Warren Hayward, TJX Companies, USA.

Change 12.092 OMEGAMON/CICS 550/551 User SMF record datasets OMCIDTCO

VMACOMCI and OMCIDTCT (DATACOM counts) had zero observations and

Jul 13, 1994 datasets OMCISUPR/OMCISUPT (SUPRA) had extra obs (with

missing variables) because DATACOMM records were not

output into the correct datasets. After "MACRO _COMCDTC"

change %INCLUDE SOURCLIB(EXOMCSUT); to (EXOMCDTT) and

change %INCLUDE SOURCLIB(EXOMCSUP); to (EXOMCDTC).

Thanks to Freddie Arie, Lone Star Gas, TEXAS.

Change 12.091 Variable QPENRECA should have been spelled OPENRECA in

VMACVMXA VM/ESA Monitor dataset VXAPLSRV.

Jul 13, 1994

Thanks to Mitch McKenna, SAS Institute Europe, GERMANY.

Change 12.090 TCP/IP variables APILCLPO.APIRMTPO,FTPLCLPO,FTPRMTPO,

VMACTCP TELLCLPO, and TELRMTPO (port addresses) should have been

Jul 13, 1994 FORMATted $HEX4.; now they are.

Thanks to Jan Van Kemenade, Univ. Rekencentrum Nijmegen, HOLLAND.

Change 12.089 A cosmetic change; the comment named YOUR.SMF.DATA as the

JCLTEST6 sample SMF input DSNAME, but the actual JCL used a name

Jul 11, 1994 of YOUR.SMF.DUMPED.DATA. Now both use the latter name!

Thanks to Philip Pearson, Ansett Australia, AUSTRALIA.

Change 12.088 DB2 Trace IFCIDS=180,181,182,186,190,192,193,194,195, and

FORMATS 198 are now decoded, although the code has only been

VMAC102 syntax checked as of this date. More subtypes and update

Jul 11, 1994 to this text are planned.

Thanks to Bob Brosnan, Goldman, Sachs & Co., USA.
Change 12.087 New analysis routing which matches CICSTRAN with DB2ACCT.

ANALDB2C See comments in the member for how multiple CICSTRAN and

Jul 8, 1994 multiple DB2ACCT transactions for a single unit-of-work

are handled in this initial analysis.

Thanks to Dan Null, Caterpiller, USA.
Change 12.086 Label for variable DVLVSER was incorrect; it is now

VMACDCOL VOLUME*SERIAL*NUMBER.

Jul 8, 1994

Thanks to Joseph J. Faska, Depository Trust, USA.

Change 12.085 Change 12.059 set SUBSYS='VPS ' for VPS output, but did

VMAC6 not also set SUBSYS6='VPS', so the change was not carried