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

-All four datasets default output token are now consistent

(all use the _LSUxxxx macro token, which has the same

default value, so this should be a transparent change,

but now the actual output agrees with the revised doc in

the comments.

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

Thanks to Ken Jones, Xwave, CANADA.

Change 24.123 The debugging PROC PRINT DATA=R73HOUR; statement was

VMXGRMFI removed.

Jul 13, 2006

Thanks to Tom Kelman, Commerce Bank - Missouri, USA.

Change 24.122 Variable R85FLGS is now decoded for the TYPE85RQ,TYPE85TP

FORMATS datasets in new R85ST74F,R85ST78F variables by formats

VMAC85 $MG085DM and $MG085TM for optical disk and tape mounts:

Jul 13, 2006 VALUE $MG085DM

'80'X='80X:MOUNT NOT REQUIRED'

'40'X='40X:FLIP DISK REQUIRED'

'20'X='20X:MOUNT ON EMPTY'

'10'X='10X:DISMOUNT REQUIRED'

;

VALUE $MG085TM

'80'X='80X:MOUNT NOT REQUIRED'

'40'X='40X:MOUNT ON EMPTY'

'10'X='10X:ATL TAPE USED'

'20'X='20X:DISMOUNT REQUIRED'

;

Thanks to Betra Reeves, Infocrossing, USA.

Change 24.121 Support for IFCID=350, the SQL statement produced by the

VMAC102 parser written during the bind of dynamic or static SQL;

Jul 7, 2006 this IFCID can be activated with Trace Class 32 for 350.

Jul 13, 2006 The IFCID=350 replaces the IFCID=63 record because the

SQL text in the IFCID=63 record was limited to 5000 bytes

and any additional text was truncated. Instead now, DB2

writes multiple IFCID=350 records with 5000 bytes each if

the SQL text exceeds 5000 bytes.

Thanks to Christian Vandenbossche, Societe Generale, FRANCE.
====== Changes thru 24.120 were in MXG 24.05 dated Jul 3, 2006=========
Change 24.120 Internal utility used only in MXG QA revised to exeute on

UTILVREF ASCII or EBCDIC systems; previously, a manually edited

Jul 2, 2006 version was required to run on ASCII.
Change 24.119 Typos introduced in Change 24.064 had _WKBLD instead of

WEEKBL3D _WEEKBLD for the new datasets, causing 180 syntax errors.

WEEKBL3T

WEEKBLDD

WEEKBLDT

Jul 1, 2006

Thanks to Art Hunter, Penn Mutual, USA.
Change 24.118 The INTERVAL=DURSET default is restored in ASUM70PR, as

VMXG70PR it was in MXG 24.03 and earlier, so PDB.ASUM70PR/ASUM70LP

Jul 1, 2006 will again have the same STARTIME as RMFINTRV when both

have DURSET defaults. I changed the default in MXG 24.04

to QTRHOUR because with DURSET, the ASUMCEC/ASUMCELP

datasets sometimes had wrong STARTIME/SMF70GIE/DURATMs,

but 24.04 was needed for its zIIP engine support. Now,

with the other enhancements in 24.111 to support SYNC59

along with DURSET, VMXG70PR has again been revised so the

original MXG default can be restored safely.

Defaults are: INTERVAL=DURSET,CECINTRV=HOUR.
Change 24.117 Variables IFAUNITS IFEUNITS CPUIFATM and CPUIFETM are now

TRND72GO created in the TREND.TRND72GO dataset.

Jun 30, 2006

Thanks to Stan Dylnicki, Royal Bank of Canada, CANDAD.

Change 24.116 z990/2084 CPUs don't have 'IFA' in LPARNAME='PHYSICAL'

VMAC7072 segments, which caused NRIFACPU count in LPAR datasets

VMXGINIT (TYPE70PR,ASUM70PR,ASUMCEC) to be zero. The IFA variables

Jun 29, 2006 in System datasets TYPE70 and RMFINTRV is fine, because

they count the IFAs assigned to each SYSTEM, but RMF 70s

for z990s do not contain the number of installed IFAs in

your CEC (later hardware/software does). So, if you have

IFAs on z990s, you will have to tell MXG how many IFAs

you have, by setting a text value in new macro variable

with a %LET statement, in IMACKEEP or //SYSIN DD stream:

- If you have the same number of IFA engines in all of

your CECs, then you would use this statement:

%LET CECIFANR= %QUOTE( CECIFANR=1; );
which stores the text "CECIFANR=1;" (which is a SAS

statement to set variable CECIFANR to 1), in the macro

variable &CECIFANR. When executed, CECIFANR=1 causes

MXG to change SMF70CIN='ICF' to SMF70CIN='IFA' in that

many LPARNAME='PHYSICAL' segments in each interval;

these are output in the PDB.TYPE70PR dataset.

- If you have different numbers of IFAs in different CECs

you will need to tell MXG how many IFAs you have on

each of your CECs, by CECSER, using this logic:
%LET CECIFANR =

%QUOTE (

CECIFANR=0;

IF CECSER='ABCD' THEN CECIFANR=1;

ELSE IF CECSER='1234' THEN CECIFANR=2;

);
so that MXG will change 'ICF' to 'IFA' in the correct

number of TYPE70PR observations.

-Once you have created PDB.TYPE70PR with this change,

you can rerun your %INCLUDE SOURCLIB(ASUM70PR) program

to create the corrected ASUM70PR/ASUMCEC/etc summaries.

-VMXGINIT GLOBALs macro variable CECIFANR with blanks.

Thanks to Jan Bigalke, Allianz, GERMANY.

Change 24.115 CICSTRAN variable WTUNIOTM was not kept when UTILEXCL was

UTILEXCL executed; the logic to calculate it was correct, but the

Jun 28, 2006 added code that is needed in UTILEXCL to KEEP a cacluated

variable was missing in UTILEXCL.

Thanks to Robert Blackburn, Dominion Resources, USA.
Change 24.114 The "bucket" variables S94ADV00-S94ADV95 with the number

ASUM94 of volumes containing 0-5 pct, 5-10, etc, are shifted so

VMAC94 their values match S94AD00-S94AD95, and labels now are:

Jun 28, 2006 S94ADV00='VOLUMES*CONTAINING*0-5PCT*ACTIVE'

S94ADV95='VOLUMES*CONTAINING*95-100PCT*ACTIVE'

Those distribution values were incorrectly SUMmed in the

ASUM94 program; now, their MAX value is output in the

summary dataset.

Thanks to Hugh Lapham, Royal Canadian Mounted Police, CANADA
Change 24.113 The DTF segment was changed, causing invalid data values

VMACOMCI for DTFCKLnn and DTFCNTnn variables (but no error message

Jun 27, 2006 was printed. Segment increased from 140 to 172 bytes,

which now is the expected length.

Thanks to Karl B. Schweitzer, State Street Bank, USA.

Thanks to Richard Schwartz, State Street Bank, USA.

Change 24.112 Variable RESIND/SMF77DFG, a bit map, is now decoded into:

VMAC77 RESIND00='RESOURCE*STILL IN*CONTENTION?'

Jun 27, 2006 RESIND01='SCOPE OF*SYSTEMS(A)*OR SYSTEM(B)?'

RESIND02='OWNER*HAD EXCLUSIVE*OR SHARED?'

RESIND03='FIRST JOB*WAITING*EXCLUSIVE*OR SHARED?'

RESIND04='SECOND JOB*WAITING*EXCLUSIVE*OR SHARED?'

RESIND05='RESOURCE*IS*GLOBAL?'

Thanks to Chuck Hopf, Bank of America, USA.

Change 24.111 RMFINTRV is enhanced to support SYNC59 if INTERVAL=DURSET

VMXGRMFI is specified, provided that your IMACRMFI member _DURSET

VMXG70PR logic does NOT alter the value of STARTIME (i.e., the MXG

VMAC7072 default IMACRMFI is used). With INTERVAL=DURSET the

VMXGINIT STARTIME/SMF70GIE timestamps in RMFINTRV, ASUM70PR,

Jun 27, 2006 ASUM70LP, ASUMCEC, and ASUMCELP will be the original time

Jun 29, 2006 values of your RMF data, and you can now use SYNC59= in

Jul 1, 2006 both RMFINTRV and ASUM70PR members to shift times so the

times in the summary datasets are "exact and pretty".
Change 24.110 Support for z9BC processor CPUTYPE='2096'x, COMPATIBLE if

FORMATS your z/OS is 64-bit. If your z/OS is NOT 64-bit, this

VMAC7072 change is required so that variables SMF70CPA, which

VMXGRMFI becomes CECSUSEC, the SU_SEC value of the hardware CEC,

Jun 27, 2006 will be set by table lookup in the MG070CP format, which

was also updated by this change to contain the 2096 SUs.

Thanks to Al Sherkow, I/S Management Strategies, Ltd., USA.
====== Changes thru 24.109 were in MXG 24.04 dated Jun 22, 2006=========
Change 24.109 -ASUMTAPE enhancement for MXGTMNT with IBM Volume Mount

ASUMTAPE Exit enabled now populates these job-level variables:

VMACTMNT ASID INITTIME JOBCLASS LOCLINFO PGMRNAME RACFTERM

Jun 22, 2006 RACFUSER READTIME STEPNAME STEPNR TMNTEXIT

from the first-volume mount, which goes thru that exit

and is captured by MXGTMNT, into subsequent multi-volume

mounts, that do not go thru IBM's exit, and thus are not

captured by MXGTMNT. This enhancement populates the job

variables, but the TAPMNTTM (MXGTMNT monitor mount time

duration) will still be missing in the second and later

multi-volume mounts. However, the new TOTMNTTM added by

Change 24.102, created from MAX of the SYSLOG or MXGTMNT

mount delay, is populated for these multi-volume mounts,

so the new BEGTMNT and ENDTMNT datetime values should now

be used for the start and end times of mount events.

The SYSLOG mount time is earlier than the mount start the

MXGTMNT monitor captures, and the SYSLOG mount verifiy is

later than the MXGTMNT end time, so the new TOTMNTTM is

a more accurate measure of tape mount delay to each job.

-The SORT order of PDB.ASUMTAPE is changed

from BY DEVNR DESCENDING EVENTIME;

to BY DEVNR EVENTIME;

but as the expected SORT order in WEEKBLD/MONTHBLD is

just BY DEVNR

the change from DESCENDING to ASCENDING order should have

no impact and is the more righteous order, anyhow.

-Note that the IBM Volume Mount Exit still misses mounts

issued by some programs: DFHSM, OPC, and DMS jobs have

mounted tapes that MXGTMNT did not see in the IBM exit,

and many of those missed mounts created a type 21 record

with a blank volume serial, and these missed do not have

the standard SYSLOG mount messages. But with those few

exceptions, this iteration of the ASMTAPEE/MXGTMNT

monitor and ASUMTAPE summary program captures all

possible IBM controlled tape mounts, to virtual or real

devices, and populates all possible job-related variables

in ASUMTAPE.

-VMACTMNT correction; SYSLOG records with INVALID MSGID

notes were due to incorrect processing of the IEC235D

WAITING FOR VOLUMES message, so those events were not

output in TYPESYMT dataset. However, as the IEC235D

message does not contain a DEVNR, it cannot be stored

in the PDB.ASUMTAPE dataset.

Thanks to Normand Poitras, IBM Global Services, CANADA.

Thanks to Paul Naddeo, FISERV, USA.
Change 24.108 Variable TSUDRECD and TSUDXMDA, Datagrams Received/Sent

VMAC119 were incorrectly read as 4-bytes, but they are documented

Jun 22, 2006 as 8 byte binary fields.

Thanks to Al Smye, IBM Canada at PWGSC, CANADA.

Change 24.107 Variable CPUCEPTM is now kept in PDB.STEPS and PDB.JOBS;

BUILD005 it was already included by IBM in CPUTCBTM, but should

BUIL3005 have been kept when created, since BUILDPDB intends to

Jun 22, 2006 keep all of the CPU time component variables.

Thanks to Chuck Hopf, Bank of America, USA.
Change 24.106 Heuristic revisions to calculation of WAITTOTM total wait

VMXGUOW time of the constructed PDB.ASUMUOW event; the original

Jun 22, 2006 estimate included WTDISPTM and WTLMIOTM but data suggests

those wait durations are overlapped with other delays.

The total count of wait events WAITTOTL was also revised.

And RCT time should not be included; consider the MRO

TOR/AOR/DOR combination; the transaction starts in the

TOR, moves to AOR, then moves to DOR. While the tran

is not in TOR, the TOR CICSTRAN still clocks RCT wait,

even while the AOR may be waiting on something else.

Thanks to Hugh Lapham, Royal Canadian Mounted Police, CANADA

Thanks to Tom Kelman, Commerce Bank - Missouri, USA.

Change 24.105 Corrections, enhancements for ASUM70PR zIIP and zAAPs.

ASUM70PR -If you had more than one IFA or more than one ZIP, the

VMXG70PR counts NRIFACPU/NRZIPCPU and uptimes IFAUPTM/ZIPUPTM

VMXGDUR in PDB.ASUM70PR and PDB.ASUMCEC summary datasets were

Jun 20, 2006 wrong, but all other IFA/ZIP CPU times were summarized

Jun 22, 2006 correctly, and all of the individual LPAR data values

(LPnXXXX variables in PDB.ASUM70PR/ASUMCEC datasets, and

all of PDB.ASUM70LP and PDB.ASUMCELP datasets) were fine.

-SYNC59 logic (for ex-MICS sites!) was imperfect and did

not always create the expected shifted exact datetime.

-Note: See Change 24.118, which reversed the change in the

default for INTERVAL: The discussion is correct, but

the default was only changed in 24.04, then restored to

the original INTERVAL=DURSET in MXG 24.05.

-The INTERVAL=DURSET default for the summary interval for

PDB.ASUM70PR and PDB.ASUM70LP datasets only works if the

_DURSET macro does NOT change the value of STARTIME. If

the value is changed, there is no way that I can tell the

desired interval, and the logic to build the

PDB.ASUMCEC/ASUMCELP datasets requires knowledge of their

INTERVAL.

DURSET says to use the _DURSET macro definition in

the IMACRMFI member to redefine the STARTIME value,

but if you do change STARTIME therein, I cannot tell

what summary interval you wanted, and if you leave

STARTIME unchanged (MXG's IMACRMFI default) I still

can't tell what is your minimum RMF interval.

The DURSET can still be used to create PDB.ASUM70PR

and PDB.ASUM70LP datasets with their original STARTIME

values, but the PDB.ASUMCEC and PDB.ASUMCELP datasets

may not be completely valid. It may be possible to

figure out a way to safely use DURSET and still have

the correct DURATM values in ASUMCEC, but for now the

safe resolution is to require you to specify the

INTERVAL= and CECINTRV= values in your ASUM70PR member

and then I can build all four datasets without errors.

-Internal VMXGDUR invocations invoke SYNC59=YES only for

the ASUM70PR/ASUM70LP creation; the second invocations

for ASUMCEC/ASUMCELP specify SYNC59=NO because times have

already been shifted.

-Change: Variables IFAWSTTM and ZIPWSTTM are now set to

a missing value in the PDB.ASUM70PR/PDB.ASUMCEC datasets

as they are overlapped wait times across all LPARs that

had an IFA or a ZIP, and really are meaningless totals.

-LPMSUHR was actually per second rather than per hour,

not fixed until Jun 22.

-LPCTBY and LPCTOV were wrong in PDB.ASUM70LP as they were

not divided by LPARCPUS.

-Enhancement: Variables NRIFACPU and NRZIPCPU are added

to the PDB.ASUM70LP and PDB.ASUMCELP dataset, calculated

from IFAUPTM or ZIPUPTM divided by DURATM.

-Some Labels and Formats were revised, including remvoal

of *PCT from the Current and Initial SHARE WEIGHTs.

-This change has been extensively tested but only with Z9

processors that populate SMF70CIN. There may still be

a problem counting IFAs on earlier hardware that did not

populate SMF70CIN.

-VMXGDUR: Variable MXGDURTM is created/dropped for time

intervals and used to reset SMF70GIE after STARTIME has

been SYNC59'd/FLOORed back to start of final interval.

-Jun 22: Variables LPSHARE, LPSHARC, TOTSHARE, TOTSHARC

were sometimes incorrect but the LPnSHARE/LPnSHARC values

were correct.

-SMF70LAC was wrong (zero) in PDB.ASUM70PR and PDB.ASUMCEC

datasets (but LPnLAC values were valid for all LPARs).

Now, the LPnLAC values are summed to create the total of

the 4-hour Average MSU values back into SMF70LAC. But

SMF70LAC isn't an LPAR variable; it comes from the "this

system" segment of the TYPE70 record, and while it is

always valid in per-system TYPE70 & RMFINTRV datasets,

in the LPAR data in PDB.TYPE70PR, SMF70LAC is non-zero

ONLY in the "this system" observations, so this is just

one more variable that requires your ASUM70PR program to

read the PDB.TYPE70PR with data from ALL of your SYSTEMs

to create correct data in the CEC-level summary datasets

PDB.ASUMCEC and PDB.ASUMCELP.

Thanks to Nathan Loewenthal, CitiGroup, USA.

Thanks to Douglas C. Walter, CitiGroup, USA.

Thanks to Michael Salyer, CitiGroup, USA.

Thanks to Brent Turner, CitiGroup, USA.

Thanks to Tom Koelle, CitiGroup, USA.

Thanks to Hugh Lapham, Royal Canadian Mounted Police, CANADA

Thanks to Tom Kelman, Commerce Bank - Missouri, USA.

Thanks to Deborah L. Soricelli, CIGNA, USA.

Thanks to Ray Dunn, CIGNA, USA.
Change 24.104 BMC CMRDETL data file is created by their CMRCMPW program

TYPE110J but the output format is determined by a control card.

TYPEMVCI When FORMAT=CMR is specified, the output format is the

Jun 19, 2006 decompressed T6E record that MXG's TYPEMVCI processes.

When FORMAT=2.2 or FORMAT=2.2Y is specified, the output

format is an SMF 110 record, but written as RECFM=U that

does NOT contain valid BDW/RDWs, and instead has only a

4-byte record length field; this is the old "DOS JOURNAL"

type 110 format, which is processed by MXG's TYPE110J.

Thanks to Pat Perecca, Pershing, USA.

Change 24.103 Variable QWHCCV was misspelled in READB2 as QWHCCCV,

ADOCDB2 which caused selection by CORRID to fail. The variable

READDB2 was also misspelled in ADOCDB2 documentation.

Jun 17, 2006

Thanks to Julain Smailes, Experian, ENGLAND.
Change 24.102 Documentation: PDB.ASUMTAPE will have zero observations

ASUMTAPE if you do not have any observations in SYSLMNT dataset.

Jun 15, 2006 You must install ASMTAPEE (MXGTMNT) monitor at ML-38 or

later, which captures the SYSLOG mount information, now

required by the ASUMTAPE logic to create PDB.ASUMTAPE.

-New variables BEGTMNT, ENDTMNT, the begin and end times

of each tape mount event, and TOTTMNTM, total duration of

the tape mount are created:

BEGTMNT is SYLMTIME (SYSLOG Mount Start), which is

usually fractions of a second earlier than the

TMNTTIME when MXGTMNT saw the mount start, but

TMNTTIME is used if SYLMTIME is missing.

ENDTMNT is SYL5TIME (SYSLOG IEC705I Label Written) that

is usually fractions of a second later than the

TENDTIME when MXGTMNT saw the mount end, but

TENDTIME is used if SYL5TIME is missing.

TOTMNTTM=ENDTMNT-BEGTMNT, the mount delay to the job.
Change 24.101 Flag variable FSRFALT='WRITTEN*IN*DUPLEX*MODE?' is added

VMACHSM to the HSMFSRTP dataset.

Jun 14, 2006

Thanks to Michael E. Friske, Fidelity Systems, USA.

Change 24.100 Utility to get the ENGINE type of a SAS data library was

VGETENG revised to use PROC SQL instead of VGETOBS, to eliminate

Jun 14, 2006 the annoying MXGWARN that MXGENG datasets doesn't exist.

Thanks to Chuck Hopf, Bank of America, USA.

Change 24.099 Format MG116TY for SMF 116 MQSeries type was updated to

FORMATS add 0:INTERNAL TASK and 8:IGQ AGENT values, and to revise

Jun 14, 2006 so now 2:BATCH OR TSO.

Thanks to Scott Barry, SBBWorks, Inc, USA.

Change 24.098 Support for the DCE segment in RACFTYPE 301 segment will

VMAC80A populate TOKDCE variable in the many events that can

Jun 14, 2006 have a DCE segment.

Thanks to Colin Wessels, Unicible, SWITZERLAND

Thanks to Pierre Beda, Unicible, SWITZERLAND
Change 24.097 More than six RACFTYPE=42 CLASS NAME segments caused MXG

VMAC80A to get INPUT STATEMENT EXCEEDED error, the "TRUNCATED

Jun 13, 2006 RECORD FOUND" that was added by Change 24.021. This

change added protection to skip over extra segments for

RACFTYPE 33, 42, and 43 segments.

Thanks to Michael Yuan, University of California Office of Pres, USA.

Change 24.096 Revised STK Exit UX01 protects a local UX01 exit that

ASMHSCEX does not save registers and status. This enhancement is

Jun 13, 2006 for sites that specify SYSPARM(Y) to include their local

UX01 exit in our link-edit step.

Change 24.095 PDB.ASUM70LP variables LPCTBY LPCTOV PCTLPBY PCTLPOV

VMXG70PR were always missing values because the DURATM=SYSDUR

Jun 10, 2006 statement was mis-located, and LPARDUR for PHYSICAL LPAR

was zero, now is set to DURATM*NRPHYCPS.

Thanks to Steve Olmstead, Northwestern Mutual Company, USA.
Change 24.094 Support for APAR OA12857 which adds PDSE Caching stats to

VMAC1415 the type 14/15 SMF records, with these new variables:

Jun 9, 2006 SMF14DRD ='PDSE*DIRECTORY*READ*REQUEST*COUNT'

SMF14DRDH='PDSE*DIRECTORY*READ*HIT*COUNT'

SMF14MCE ='PDSE*MEMBER*CACHE*ELIGIBLE*COUNT'

SMF14MCF ='PDSE*MEMBER*ELIGIBLE*CACHE FULL*COUNT'

SMF14MNC ='PDSE*MEMBER*ELIGIBLE*NOT CACHED*COUNT'

SMF14MRD ='PDSE*MEMBER*READ*REQUEST*COUNT'

SMF14MRDH='PDSE*MEMBER*READ*HIT*COUNT'

SMF14MST ='PDSE*MEMBER*CACHE*STOLEN*COUNT'

But see Change 24.196; it is required to avoid ABENDs.
Change 24.093 New variables added to EDALOGOF dataset:

VMACEDA EDACOMPL='COMP'

Jun 6, 2006 EDAOFA40='FULL*ACCOUNT*NUMBER'

EDAOFID1='SYSPLEX*ID1'

EDAOFID2='SYSPLEX*ID2'

EDAOFPID='POOLED*USERID'

EDAPRTY ='PRIORITY'

Thanks to Andreas von Imhof, Rabobank, THE NETHERLANDS.

Change 24.092 z/OS 1.8 support (COMPATIBLE).

FORMATS -TYPE30 new EXSRMERR flag variable if SRM is unable to

VMAC108 return the SMF30AIx and SMF30EIx variables, and the

VMAC30 EXCPERR flag for over 4 Billion EXCP logic was revised.

VMAC7072 -SMF30MLS new value 10 decoded by MG030ML format.

VMAC71 -TYPE70 new variables:

VMAC74 SMF70CSC='SEQUENCE*CODE*OF*CONFIGURATION'

VMAC80 SMF70GJT='TIME WHEN*PARTISHN*JOINED*CAPACITY GROUP'

Jun 5, 2006 SMF70POM='PLANT*CODE*OF*MANUFACTURE'

Jun 22, 2006 -TYPE71 new variables for UIC statistics:

Aug 13, 2007 SMF71ULM LOWEST*MINIMUM*SYSTEM*UIC

SMF71ULC LOWEST*CURRENT*SYSTEM*UIC

SMF71UHC HIGHEST*CURRENT*SYSTEM*UIC

SMF71UHX HIGHEST*MAXIMUM*SYSTEM*UIC

SMF71UAM AVERAGE*MINIMUM*SYSTEM*UIC

SMF71UAC AVERAGE*CURRENT*SYSTEM*UIC

SMF71UAX AVERAGE*MAXIMUM*SYSTEMI*UIC

-TYPE72GO new Resource Group Capacity variables

R723GMLP='PERCENT*CAPACITY*OF LPAR?'

R723GMSS='PERCENT*CAPACITY*OF SINGLE*SYSTEM?'

-TYPE72DL new Resource Manager state samples

R723BPMI='BUFFER*POOL*IO*MISSES*SAMPLES'

R723RW05-R723RW15 'RESOURCE TYPE NN SAMPLES'

-TYPE74 new flag variable

IOTMERR ='CONNECT*TIME*IN ERROR?'

-TYPE74CF variable R744VLVL corrected to R744FLVL.

New variable R744FMPC with Plant Code of CF.

New variable R744FSEQ with CF Sequence Number.

-TYPE88 new variable

SMF88GRP='GROUP*VALUE*FOR THE*LOG STREAM'

-TYPE108, new transaction type values in MG108TR format.

-Jun 22: IBM confirmed SMF70GMU is binary, and not the

EBCDIC format in the pre-GA documentation.

-Aug 13, 2007: These group variables are in TYPE70PR and

should not have been output in TYPE70 dataset:

SMF70GNM='CAPACITY*GROUP*NAME'

SMF70POM='PLANT*CODE*OF*MANUFACTURE'

See Change 25.163.

Change 24.091 Revision to RMF III VSAM decompression utility corrects

ASMRMFV processing of the ENCG3 table for the ENC option, adds

DOCLRMFV reporting of the RMF III Index usages, and contains the

Jun 2, 2006 CPUG3 record size reduction.