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

May 31, 1996 your MSOCOEFF value (in IPS) to make it a very small

Mar 10, 2003 value - perhaps .001 to .0001 - so that MSOUNITS are

insignificant (less than .1 percent?) in the total

service units calculated by the SRM. You can use a

PROC MEANS DATA=PDB.TYPE72 SUM; to get the percent of

SERVICE that is recorded in MSOUNITS and change your

MSOCOEFF, because MSOUNITS are not a measure of work,

and should not be included in the SRM's measurement of

work/service. Work is CPU and I/O only. Memory is

like air conditioning and floor space (you have to

have enough, and it only comes in big chunks) but

memory occupancy, floor space occupancy, and cooling

occupancy are not measures of computer work; they are

the fixed costs that makes computing possible!

Why then make this minor change? Because Freddie's QA

stream saw the minor difference in values between 13.13

and 14.03 and so for the sake of perfection (rather than

real impact), the change was extended to these members.

Mar 10, 2003: Original text had .000001 but .0001 is the

minimum permitted value, other than zero.

Feb 21, 2005: See MXG Technical Note in Newsletter 47;

IBM Discovered that the smallest value is actually .0122.

Any value less than that value will be reported in the

MSOCOEFF in TYPE72GO, but only 0.0122 will be used!

Thanks to Freddie Arie, Lone Star Gas, TEXAS.
Change 14.116 Connect-Direct (originally NDM) V 1 R 4 Mod 1 uses two

FORMATS existing reserved fields to compatibly add variables

VMACNDM NDMLUOTH (Other*node LUNAME) and NDMNETYP (Network Type).

May 31, 1996 New MXG format $MGNDMTY was created to decode NDMNETYP.

Thanks to Steve Beaver, Texas Commerce Bank, USA.
Change 14.115 The date part of SYNCSORT variables SORTBEGN and SORTEND

VMACSYNC are off by one day for sorts that spanned midnight. The

May 30, 1996 statement

May 31, 1996 IF SORTBEGN GT SORTEND THEN

SORTBEGN=SORTBEGN-DHMS(1,0,0,0);

should have been

IF SORTBEGN GT SORTEND THEN

SORTEND=SORTEND+DHMS(1,0,0,0);

Also, the +32 reserved bytes INPUT after HPUSED are now

replaced with +22 TSTSYNCS &PIB.8. +2

and IF TSTSYNCS GT 0 THEN CHARACTS=TSTSYNCS; was added

after the @; to support the (as yet undocumented) 8-byte

value so more than 2.4GB of characters can be counted.

Thanks to Linda Berkley, Amdahl, USA.

====Changes thru 14.114 were included in MXG 14.03 dated May 27, 1996===
Change 14.114 Support for RACF 1.10 records (compatible). MXG version

VMAC80 14.03 and later tolerate RACF 1.10 records (i.e., will

VMAC80A not fail), but the new "extended length relocate sections

May 26, 1996 are not yet decoded, as I need test data records to

understand how they can be useful. When test data has

been received and the new section is decoded, this note

will be revised.
Change 14.113 Support for Thruput Manager #V041238 "Job Extract"

IMACTPM extension (INCOMPATIBLE) restructured the fixed segment

VMACTPM and creates these new variables in TYPETPMF dataset:

May 26, 1996 CPUTMINT='CPU USED*BY INTER*AND JOB ANAL'

CPUTMJAT='TOTCPU USED BY JOB ANALYZER'

JCTJOBID='JES*JOB*IDENTIFIER'

TPMTMCTA='CATALOGS*ALLOCATED*IN JOB ANAL'

TPMTMLCA='CATALOG*LOCATES*VIA*CAT LOOKASIDE'

TPMTMLCM='CATALOG*LOCATES*VIA*CAT MANAGEMENT'

TPMTMLRQ='CATALOG*LOCATES*REQUIRED'

The default in MXG is for the new record; if you still

process the base record you must set macro _TPMVERS to 0

in member IMACTPM.

In addition, the processing of the extract segments was

redesigned. Previously, an observation in TYPETPMV was

created for each 16 bytes of the TPM data value in each

TPMFIELD segment. Now, there will be one observation for

each TPMFIELD segment, with TPMVALUE now containing the

first 64 bytes of the data value. $ACCT segments have

been found with 161 bytes, but all other data lengths are

less than 64, so the new design should decrease the size

and number of observations in TYPETMMV dataset.

Thanks to Andy Chandler, Eagle Star, ENGLAND.
Change 14.112 IDMS Performance Monitor dataset IDMSBUFF was incorrectly

VMACIDMS decoded. CA uses ALIGN in their assembly, causing bytes

May 26, 1996 to be inserted in the SMF record that do not exist in the

Jul 11, 1996 DSECT, and two fields were increased from 2 to 4 bytes.

-Insert after INPUT BUFNAME $EBCDIC18. +2 to skip the

undocumented bytes, and change the SKIP=SKIP-18; to -20;

-May 26 change said: Change INPUT of BUFNDEFN and BUFNINUS

from &PIB.2. to &PIB.4., and 28 lines later change the

SKIP=SKIP-104 to -108. However, Jul 11, change puts two

fields back as &PIB.2., and instead inserts +2 after each

of those halfwords. The SKIP change of May 26 is coorect.

Thanks to Terry Heim, Ecolab Inc. USA.

Thanks to Martin Wieland, Neckermann B.V., THE NETHERLANDS
Change 14.111 Support for NETSPY Release 4.7 compatibly added variables

VMACNSPY to several datasets:

May 25, 1996 Dataset Variables added

NSPYNCPY CBPIDTID,CBPCBFLG

NSPYTIC3 NSPNTMBR NSPNTMBS NSPVTMFR NSPVTMFS

NSPYVIRT VRBPIINB VRBBYINB

Change 14.110 Variables BADTAP EDMTAP and DYNAM were created but not

VMACTMS5 kept in dataset TMSREC; they have now been added to the

TYPETMS5 KEEP= lists in both VMACTMS5 and TYPETMS5.

May 24, 1996

Thanks to Terry Duchow, U.S. Postal Service, USA.
Change 14.109 Type 110 SMF record with subtype=2818 ('B02'X) is written

UTILGETM by Boole and Babbage's CICS Manager, but UTILGETM has but

May 24, 1996 255 buckets to count subtypes. Now, UTILGETM recognizes

this record, counts it as a subtype 2 record, and the

UTILGETM FOUND A RECORD WITH SUBTYPE GREATER THAN 255

message is not printed for this statistics record.

Thanks to Trevor Ede, Bank of New Zealand, NEW ZEALAND.
Change 14.108 MXG 14.02 only. BY VARIABLES ARE NOT PROPERLY SORTED.

DIFFDB2 The PROC SORT for DATA = _LDB2STR for new PDB.DB2STATR

May 23, 1996 dataset should have BY SYSTEM QWHSSSID QLSTLOCN QWHSSTCK;

(QLSTLOCN was missing from that BY statement).

Thanks to Tony Sandora, CIGNA, USA.
Change 14.107 MXG 14.02 only. INPUT STATEMENT EXCEEDED with type 37

VMAC37 because BRFLOCAL and BRFRMT are not separate new fields.

May 23, 1996 (The type 37 documentation moved from the NetView Admin

Guide into the Application Programming Guide in NetView

2.4, but I then misread Table 33 and thought there were

two new 30-byte fields (BRFLOCAL, BRFRMT), but they are

actually only redefinitions of table 34 whose fields are

already decoded by MXG, so there truly was NO CHANGE in

the Netview 2.4 type SMF record, and the text of 14.090

has been revised.) This change removed the DO group

(IF BRFREVLT GE 2.4 THEN DO; ... END;) that had been

incorrectly added by 14.090. The other correction in

14.090 was valid.

Thanks to Tony Sandora, Cigna, USA.

Change 14.106 Variable A14EQPCT (MAXQTIME Purge Count) was not kept in

VMAC110 dataset CICCONSR (ISC/IRC data), but now is. Also, the

May 23, 1996 variables MCTMNTAD and SMFPSRVR were not kept in dataset

CICSEXCE, but now have also been added to the KEEP= list.

Thanks to Don Deese, (CPExpert), Computer Management Sciences, USA.

Thanks to Dawn Medwid, Aetna, USA.

Change 14.105 Variable QWSDLR (High used RBA address of log) was added

VMACDB2 to the LENGTH statement as length 8, and formatted HEX12.

May 23, 1996 as it is an address, not a number; this was supposedly

fixed by Change 13.129 but was not actually made til now.

Also, the input of QWSCIID after IF NRQWSC GT 15 THEN DO;

was changed to XXXCIID, so that the real value of QWCSIID

is not corrupted when there are more than 15 segments.

Thanks to Tom Parker, Hogan Systems, USA.

Change 14.104 Variables BEGTIME and ENDTIME were not created in the DB2

DIFFDB2 interval datasets DB2STATB (buffer pool) and DB2STATR

May 23, 1996 (remotes), but now they are!

Thanks to Susan Marshall, SAS Institute Cary, USA.

Change 14.103 Variables ACTLNGH, NETLNGH, and USELNGH should be INPUT

VMACORAC as &PIB.2. instead of &PIB.4., so that accounting data

May 23, 1996 is correctly input.

Thanks to Dave Harris, TVA, USA.

Change 14.102 MXG 14.02 only. INPUT STATEMENT EXCEEDED RECORD LENGTH

VMAC7072 with MVS/ESA 5.2 type 72 record because of MXG coding

May 22, 1996 errors (and insufficient testing of Change 14.085).

-Find the statement IF LENSCS GE 316 THEN DO;

Change the following INPUT statement for variable

R723CSRS from &PIB.4.6 to &PIB.8.6, and for variables .

R723CSPA and R723CSPE from &PIB.4. to &PIB.8.

-Find the statement SKIP=LENSCS-292; (which was 11 lines

after the IF LENSCS GE 316 THEN DO; statement), and move

it immediately before the IF LENSCS GE 316 THEN DO;

-After R723CSRS=1024*R723CSRS; statement, insert new line

SKIP=SKIP-24;

Thanks to Tony Sandora, CIGNA, USA.
Change 14.101 The error message "JOB WAS NOT EXECUTED ON THE FIRST DAY'

MONTHBLD was made more readable, and now points to MXG Newsletter

May 22, 1996 TWENTY-FIVE for instructions for "Running the MONTHBLD

program on a day other than the 1st day of the month".

Thanks to Bernie Levy, CUC International, USA.
Change 14.100 In fifteen members, the datetime literal value of

IMACEXCL '01JAN00:00:00:00'DT was used to initialize variables to

TYPEMOND IBM's epoch datetime value, but that 00 year value can be

TYPEMONI interpreted as 1900 or 2000, depending on the value of

TYPEMONX your SAS option YEARCUTOFF, so MXG was changed to use

TYPEMON8 '01JAN1900:00:00:00'DT to always set the correct value.

TYPETMON MXG uses this value only when decoding raw records into

VMACCIMS SAS datasets, so it is highly unlikely that you would

VMACHURN have this literal value in your own SAS programs;

VMACOMCI nevertheless, it would certainly be wise to search you

VMACVMON SAS source libraries for '01JAN00's existence and change

VMAC110 to the 1900 value to thereby support the year 2000.

VMAC110M

VMAC110S

VMAC116

VMACDB2H

YEAR2000

May 22, 1996

Thanks to Graeme Yeandle, British Telecom, ENGLAND.
Change 14.099 The @137,141,143,145 for the input of OFFARMS,LENARMS,

VMAC30 NRARMS and EXTRARMS should have been 149,153,155 and 157

May 22, 1996 for the Auto Restart section. I still do not have a dump

of a real type 30 record with Auto Restart section, but

I just happened to notice this error.
Change 14.098 Support for AS/400, OS/400 Release 3.6 INCOMPATIBLY added

VMACQAPM fields to the QAPMDISK, deleted fields in QAPMCONF,

May 1, 1996 increased the size of one field in QAPMSYS, and added and

deleted fields from QAPMJOBS.

Added to QAPMDISK:

DMDRN ='MIRRORED*DEVICE*RESOURCE*NAME'

DMIRN ='MIRRORED*IOP*RESOURCE*NAME'

DSCCFW ='CONTROLLER*CACHE*FAST*WRITES'

DSCCRH ='CONTROLLER*CACHE*READ*HITS'

DSCCWH ='CONTROLER*CACHE*WRITE*HITS'

DSDRN ='DEVICE*RESOURCE*NAME'

DSPCPH ='CONTROLLER*CACHE*PARTIAL*READ HITS'

IOPRN ='IOP*RESOURCE*NAME'

and deleted variables DS/DMARMP DS/DMIOPA DS/DMIOPB

and DS/DMPORT

Added to QAPMJOBS:

JBDRN ='DEVICE*RESOURCE*NAME'

JBIRN ='IOP*RESOURCE*NAME'

JBSJNB ='SUBMITTORS*JOB*NUMBER'

JBSJNM ='SUBMITTORS*JOB*NAME'

JBSJUS ='SUBMITTORS*JOB*USER'

JBTTYE ='TASK*TYPE*EXTENDER'

JBTTYP ='TASK*TYPE'

Only these four structures have been enhanced for 3.6.0

as of this date. Further updates may be expected.

Thanks to Clark Jennings, Reynolds Metal, USA.

Change 14.097 MXG 14.02 only. Zero obs in dataset NSPYLU due to typo.

VMACNSPY In line number 15610061 which read:

May 1, 1996 INPUT LRSPHOST &PIB.4.6 /*

that extraneous "/*" must be removed from that line.

Thanks to Freddie Arie, Lone Star Gas, TEXAS.
====Changes thru 14.096 were included in MXG 14.02 dated Apr 25, 1996===
Change 14.096 Duplicate INPUT of the same variable in the same INPUT

VMACNSPY statement did not cause errors, but to avoid confusion,

Apr 26, 1996 those INPUT statements were cleaned up in these members:

VMACIMSA SYNCUOWC Duplicate line deleted

VMACLMS OPTION1 Duplicate renamed OPTIONLM

VMACNSPY CBPPOOL New variable CBPPOOLU created

VMACNSPY NSPBEXPN New variable NSPBEXTN created

VMACQAPM xxIOPA,B Input @20, @21 removed

VMACRMFV SSHCPUX Duplicate renamed SSHCPUX

VMACX37 VOLSER Duplicate renamed VOLSERX

VMACXAM DASDCACH Duplicate renamed DASDCACX

VMACXCOM UNDOC04 Duplicate renumbered.

Thanks to Freddie Arie, Lone Star Gas, TEXAS.
Change 14.095 NETSPY type S record from R4.4 caused STOPOVER; the MXG

VMACNSPY tests for NSPYENTL=108 and NSPYENTL=160 must be changed

Apr 26, 1996 to NSPYENTL=112 and NSPYENTL=164 respectively.

Thanks to Freddie Arie, Lone Star Gas, TEXAS.

Change 14.094 Variable SYNCTIME was not kept in TYPE30_0 dataset, but

VMAC30 now it is.

Apr 26, 1996

Thanks to Dr. Alexander Raeder, Karstadt AG, GERMANY.

Thanks to Harmuth Beckmann, Karstadt, AG, Germany
Change 14.093 Support for IBM's Cache RMF Reporter (CRR) 1.7 adds new

VMACACHE variables and a new CACMODEL='02'X, for Emulated 3990s

Apr 25, 1996 (such as the RAMAC Array Subsystem). The IBM changes are

compatibly made, but there will be no observations in

dataset CACHE90 for the new Emulated 3990s records until

you install this change. New variables in CACHE90 are:

CACMODEL='CACHING*SUBSYS*MODEL'

CWCC ='CONCURRENT*COPY*CONTAM*WRITES'

DCTC ='DEVICE CACHE TRACKS CONFIGURED'

DCTR ='DEVICE CACHE TRACKS REMOVED'

DVRD ='DEVICE READS*BY*CONTROL UNIT'

DVRH ='DEVICE READ HITS*BY*CONTROL UNIT'

DVWH ='DEVICE WRITE HITS*BY*CONTROL UNIT'

DVWR ='DEVICE WRITES*BY*CONTROL UNIT'

PPRC ='PPRC*WRITE*COUNT'

RSV ='LOWER*INTERACE*IO*(MILLISEC)'

SFRD ='TRACKS*READ*FROM SIDEFILE'

TSRR ='TRACKS*STAGED*FOR RAID RECONSTRUCION'

CSCSDFM0='CACHE 0*DISABLED*FOR*MAINTENANCE?'

CSCSDFM1='CACHE 1*DISABLED*FOR*MAINTENANCE?'

CSCSTAT0='CACHE 0*STATUS'

CSCSTAT1='CACHE 1*STATUS'

CSNSBAT0='NVS 0*BATTERY*IS*DEFECTIVE?'

CSNSBAT1='NVS 1*BATTERY*IS*DEFECTIVE?'

CSNSDFM0='NVS 0*DISABLED*FOR*MAINTENANCE?'

CSNSDFM1='NVS 1*DISABLED*FOR*MAINTENANCE?'

CSNSPND0='NVS 0*PENDING*DUE TO*ERROR?'

CSNSPND1='NVS 1*PENDING*DUE TO*ERROR?'

CSNSTAT0='NVS 0*STATUS'

CSNSTAT1='NVS 1*STATUS'

Note that variables CSCSDFM, CSCSTAT, CSNSBAT, CSNSDFM,

CSNSPND and CSNSTAT were replaced by the above pairs.

Change 14.092 IBM's RMDS Version 2.2 is supported by MXG, as there was

VMACRMDS no change in the contents of their SMF record.

Apr 25, 1996
Change 14.091 Boole's STOPX37's replacement version records are trash.

VMACX37 The version 3.5 SMF record was replqced by a record with

Apr 25, 1996 only text fields. Many users complained to Boole, and

Boole was finally persuaded to create the SMF record that

should not have been replaced in their next version. As

yet I have no documentation by which to update MXG.

Update: See Change 14.207 for PRO/SMS replacement.
Change 14.090 The original text of this change said that the NETVIEW

VMAC37 2.4 type 37 SMF record was INCOMPATIBLY changed, but that

Apr 25, 1996 is false - see Change 14.107 - there was no change.

May 23, 1996 However, 14.090 did correct MXG's logic error that had

caused INPUT STATEMENT EXCEEDED errors due to incorrect

detection of the existence/insertion of OFFETHD:

The line IF OFFPROD-3-COL GE 16 THEN DO; was changed to

IF OFFPROD-3-COL EQ 16 THEN DO; and the line

IF OFFPROD-3-COL GE 24 THEN DO; was changed to

ELSE IF OFFPROD-3-COL GE 24 THEN DO;

Thanks to Joe Schwartz, CIGNA, USA.
Change 14.089 Support for APAR PN69653, which changes COLLTIME in IBM

VMAC110 type 110 record from YY to YYYY for year 2000 compliance.

YEAR2000 Member YEAR2000 updated.

Apr 25, 1996

Change 14.088 BLKSIZE=23364 values of Change 12.248 were implemented in

JCLIMSLG the sample JCL in member ASMIMSLG, but only now have the

Apr 25, 1996 DCB parameters been corrected in member JCLIMSLG.

Thanks to Waldemar Schneider, SAS Institute Cary, USA.

Change 14.087 Variable QWHCATYP exists both in DB2 datasets and in MQM

FORMATS dataset TYPE116, but with different values for Attachment

VMAC116 Type, so the variable QWHCATYP in TYPE116 dataset is now

Apr 25, 1996 renamed QWHCXTYP which is FORMATed MG116TY to decode

the MQM attachment types. This is an INCOMPATIBLE change,

if any of your reports used variable QWHCATYP from the

dataset TYPE116. Sorry for the inconvenience!

Thanks to Rick Apel, Square-D Company, USA.

Change 14.086 ASMTAPES MAINTLEV 9 supresses the stoppage of writing the

ASMTAPES Allocation subtype after TMNT013E error message. See the

Apr 25, 1996 MXG Technical Note in CHANGES in MXG 14.02 (to be printed

in Newsletter 30 this summer) for further discussion.

Find the TMNT013E message WTO and change that code to

look like this: (change text of WTO, comment NI and MVC):

WTO 'TMNT013I A LOCK OBTAIN LOOP HAS EXCEEDED MAXIMUM TIME AX

LLOWED - POSSIBLE SYSTEM STALL'

* NI OPTIONS,255-DOALLOC TURN OFF ALLOCATION BIT

* MVC PARSSAL,=CL3'NO ' UPDATE STATUS MESSAGE

B MONITOR AND LEAVE

LOCKED EQU *

Thanks to Mike ONeill, Dun and Bradstreet, USA.
Change 14.085 Variables added by IBM in MVS/ESA 5.2.0 but overlooked:

VMAC7072 Dataset TYPE72GO (Goal Mode Service/Report Class data):

VMAC74 R723CSRS='SHARED*PAGE*RESIDENCY*TIME'

Apr 23, 1996 R723CSPA='SHARED*PAGEINS*FROM*AUXSTORE'

R723CSPE='SHARED*PAGEINS*FROM*ESTORE'

Dataset TYPE74OM: (Open Edition MVS):

OMVSCAMN='MIN*MEM MAP STORAGE PAGES*PER CYCLE'

OMVSCAMX='MAX*MEM MAP STORAGE PAGES*PER CYCLE'

OMVSCGMN='MIN*SHARED MEM PAGES*PER CYCLE'

OMVSCGMX='MAX*SHARED MEM PAGES*PER CYCLE'

OMVSCHMN='MIN*SHARED MEM IDS*PER CYCLE'

OMVSCHMX='MAX*SHARED MEM IDS*PER CYCLE'

OMVSCMAP='AVG*MEM MAP STORAGE PAGES*PER CYCLE'

OMVSCMMN='MIN*MSG QUEUE IDS*PER CYCLE'

OMVSCMMX='MAX*MSG QUEUE IDS*PER CYCLE'

OMVSCMSG='AVG*MSG QUEUE IDS*PER CYCLE'

OMVSCPAG='AVG*SHARED STORAGE PAGES*PER CYCLE'

OMVSCSEM='AVG*SEMAPHORE IDS*PER CYCLE'

OMVSCSHM='AVG*SHARED MEM IDS*PER CYCLE'

OMVSCSMN='MIN*SEMAPHORE IDS*PER CYCLE'

OMVSCSMX='MAX*SEMAPHORE IDS*PER CYCLE'

OMVSCSPG='AVG*SHARED MEM PAGES*PER CYCLE'

OMVSCXMN='MIN*SHARED STORAGE PAGES*PER CYCLE'

OMVSCXMX='MAX*SHARED STORAGE PAGES*PER CYCLE'

OMVSMMAP='MAX*MEM MAP PAGES*PER CYCLE'

OMVSMMSG='MAXIMUM*MESSAGE*QUEUE*IDS'

OMVSMPAG='MAX*SHARED STORAGE PAGES*CONSTANT'

OMVSMSEM='MAXIMUM*SEMAPHORE*IDS'

OMVSMSHM='MAXIMUM*SHARED*MEMORY*IDS'

OMVSMSPG='MAXIMUM*SHARED*MEMORY*PAGES'

OMVSOAMN='MIN*EXCEED MEM MAP PAGES*PER CYCLE'

OMVSOAMX='MAX*EXCEED MEM MAP PAGES*PER CYCLE'

OMVSOGMN='MIN*EXCEED MAX SHR PAGES*PER CYCLE'

OMVSOGMX='MAX*EXCEED MAX SHR PAGES*PER CYCLE'

OMVSOHMN='MIN*EXCEED MAX SHRMEM IDS*PER CYCLE'

OMVSOHMX='MAX*EXCEED MAX SHRMEM IDSPER CYCLE'

OMVSOMAP='AVG*EXCEED MEM MAP PAGES*PER CYCLE'

OMVSOMMN='MIN*EXCEED MAX MSGS*PER CYCLE'

OMVSOMMX='MAX*EXCEED MAX MSGS*PER CYCLE'

OMVSOMSG='AVG*EXCEED MAX MSGS*PER CYCLE'

OMVSOPAG='AVG*EXCEED STORAGE PAGES*PER CYCLE'

OMVSOSEM='AVG*EXCEED MAX SEMA IDS*PER CYCLE'

OMVSOSHM='AVG*EXCEED MAX SHRMEM IDS*PER CYCLE'

OMVSOSMN='MIN*EXCEED MAX SEMA IDS*PER CYCLE'

OMVSOSMX='MAX*EXCEED MAX SEMA IDS*PER CYCLE'

OMVSOSPG='AVG*EXCEED MAX SHR PAGES*PER CYCLE'

OMVSOXMN='MIN*EXCEED STORAGE PAGES*PER CYCLE'

OMVSOXMX='MAX*EXCEED STORAGE PAGES*PER CYCLE'

OMVSCYCT='CYCLE*TIME*VALUE'
Change 14.084 INPUT STATEMENT EXCEEDED for SUBTYPE=41. The line with

VMACBETA DATE &PD.4 must be DATE &PD.4. (i.e., the final

Apr 22, 1996 period was missing). Originally, this text also said

May 26, 1996 "there are other changes in the subtype=41 record that

are not documented", but that was not true; there were no

other changes to that subtype of the BETA93 SMF record.

Thanks to Mr. Bronner, Versandhaus Wenz, GERMANY
Change 14.083 Variables QW0045SR and QW0045XR added to T102S045 dataset

VMAC102 by DB2 4.1, with "Suspend Reason" and "Extent of Global

Apr 22, 1996 Contention" flags.

Thanks to Dan Kaberon, Hewitt Associates, USA.

Change 14.082 Labels for QBGxxx variables were changed from "SES ..."

VMACDB2 to "CF ...". These variables report Coupling Facility

Apr 16, 1996 contentions and false contentions. What was originally

called Shared Expanded Storage (SES) is now known as the

Coupling Facility (CF).
Change 14.081 The OPTIONS MAUTOSOURCE SASAUTOS=SOURCLIB; statement in

TYPEIMSA TYPEIMSA was removed. Those options have been set by the

Apr 16, 1996 CONFIG option for many versions, and so this statement is

redundant; it also caused SAS/CPE problems in the way

they modify MXG code on the fly.

Thanks to Carl Sommer, SAS Institute Cary, USA.

Change 14.080 VM/370 (archaic!) monitor records with RELEASVM='01' were

VMACVMON not recognized in the Device (6.02) section. Before the

Apr 16, 1996 line reading ELSE LENDV2=1; insert this line:

ELSE IF RELEASVM EQ '01' THEN LENDV2=2;

to correct the error.

Thanks to Jim Cheesman, NWNL Corporation, USA.

Change 14.079 Date variables which are zero printed as 01JAN1960, which

VMACTLMS is the SAS date value epoch, but this caused confusion on

Apr 16, 1996 reports (which previously printed only a 0). All of the

statements ... THEN variable='01JAN1960'D; were replaced

with THEN variable =.; (to set zero dates to missing).

Thanks to Rey Marquez, General Accident, USA.

Change 14.078 Reserved Change Number.

Apr 12, 1996

Change 14.077 OPC Release 3.0 and 3.1 may cause INVALID MTD SUBTYPE

EXOPC24Y messages to fill the SAS log, and observations were not

IMACOPC output. New fields that were not in the 3.0 manual are

VMACOPC now in the 3.1 manual, but also now show up in records

Apr 11, 1996 created by OPC Release 3.0! Since I can't get straight

Apr 16, 1996 answers from IBM as to whether the fields were really

there in 3.0 and I missed them, or whether PTFs created

the new fields, I have added support for the new data but

you will have to update macro _OPCVER in member IMACOPC

to a value of 3.1 for MXG to read in the new fields. In

otherwords, try processing OPC 3.0 with _OPCVER 3, and if

that fails, change _OPCVER to 3.1 and try again!