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

triplicate, or more replicas of each interval's data to

be output in the TYPE78IO per-IOPIQID dataset. I think

the SMF78QDS data should have been written only in the

first split record, but now that I see what IBM does, the

change now only outputs TYPE78IO when SMF78RSQ is zero or

one; zero is the value in an un-split record, one is the

first record when records are split.

P.S. Because the split records have slightly different

values in SMFTIME, these duplicates cannot be

automatically recognized as dupes with the NODUP

option in PROC SORT after the fact; they must be

prevented up front.
-No one has noticed/reported the duplicate TYPE78IO data,

and only two sites encountered the ABEND in their daily

BUILDPDB due to the invalid SMF78HIX, but IBM immediately

accepted the problem and created the APAR to correct the

error. But RMF 78s are automatically processed by MXG's

BUILDPDB job, so the behind-the-scene installation of

HyperPAV might cause a fine-running-daily-accounting-JOB

to ABEND, so do not overlook this Change and/or APAR.

Thanks to Sam Knutson, Geico, USA.

Thanks to Mike Lewellen, Charles Schwab & Co., USA.

Change 25.140 Support for z/OS 1.9 (very minor additions, COMPATIBLE).

ADOC6 - Variables SMF22ETY and SMF22FNC values are formatted

EXTY8223 and decoded by existing MG022FN, MG022ET formats.

EXTY9214 - Variable R723TPDP='CPU TIME AT*PROMOTED*DISPATCH*PRTY'

IMAC82 is added to TYPE72GO dataset.

IMAC92 - Variables

VMAC1718 R742PSTM='MORE*PATH*STATUS*FLAGS'

VMAC22 R742PRCT='PATH*RETRY*COUNT'

VMAC30 R742PPND='CURR*PENDING*SIGNALS*OUTBOUND'

VMAC7072 R742PUSE='CURRENT*BLOCKS OF*BUFF SPACE*USED'

VMAC71 are added to the TYPE74PA dataset.

VMAC74 - Variables

VMAC79 R742MST1='EXTENDED*MEMBER*STATE*ONE'

VMAC82 R742MST2='EXTENDED*MEMBER*STATE*TWO'

VMAC92 R742MINT='STATUS*CHECKING*INTERVAL'

VMXGINIT R742MJOB='JOB THAT*JOINED*MEMBER'

Jul 26, 2007 are added to the TYPE74ME dataset.

- Variables

R744FPSN='NUMBER OF*SHARED*CF*PROCESSORS'

R744FPDN='NUMBER OF*DEDICATED*CF*PROCESSORS'

CFPTYP01='1ST CF*CPU*PROCESSOR*TYPE' thru

CFPTYP16='16TH CF*CPU*PROCESSOR*TYPE'

CFPWGT01='1ST CF*SHARED*PROCESSOR*WEIGHT' thru

CFPWGT16='16TH CF*SHARED*PROCESSOR*WEIGHT'

are added to the TYPE74CF dataset.

- Variable

R744SETM='STRUCTURE*EXECUTION*TIME'

are added to the TYPE74ST dataset.

- Variable

R791EXCW='EXCP*COUNT'

are added to the TYPE791 dataset.

- Variable

R792EXCW='EXCP*COUNT'

are added to the TYPE792 dataset.

- Variable

SMF82TKS='TKDS NAME'

are added to the TYPE8201 dataset.

- New dataset TYPE8223 dataset contains

SMF82TKF='TKDS*BITS'

SMF82TKH='TKDS*HANDLE*BEING*PROCESSED'

SMF82TKN='TKDS*NAME'

- New dataset TYPE9214 dataset contains

SMF92DDN='UNIQUE*DEVICE*NUMBER*FOR FILE'

SMF92DFG='FLAGS*80X=*RENAME'

SMF92DFN='NAME OF FILE*DELETED OR*RENAMED'

SMF92DFS='FILE*SYSTEM*NAME/

SMF92DFT='DATETIME OF*DELETE'

SMF92DIN='FILE*SERIAL*NUMBER'

SMF92DIP='FILE*SERIAL*NUMBER*OF PARENT'

SMF92DNL='LENGTH OF*FILE NAME*FOR DELETE'

SMF92DNR='LENGTH OF*FILE NAME*FOR RENAME'

SMF92DRN='NEW NAME*THAT WAS*RENAMED'

SMF92DTY='FILE TYPE*AS DEFINED IN*BPXYFTYP'

new variables.

-APAR OA17070 Additions to RMF 74-4 Coupling Facility CF

records provides the RMF support for accounting and the

measurement extensions of CF level 15. The COMPATIBLE

record change to RMF 74 records added these new data:

- Variables

R744FPSN='NUMBER OF*SHARED*CF*PROCESSORS'

R744FPDN='NUMBER OF*DEDICATED*CF*PROCESSORS'

CFPTYP01='1ST CF*CPU*PROCESSOR*TYPE'

thru

CFPTYP16='16TH CF*CPU*PROCESSOR*TYPE'

CFPWGT01='1ST CF*SHARED*PROCESSOR*WEIGHT'

thru

CFPWGT16='16TH CF*SHARED*PROCESSOR*WEIGHT'

are added to the TYPE74CF dataset.

- Variable

R744SETM='STRUCTURE*EXECUTION*TIME'

are added to the TYPE74ST dataset.
Change 25.139 Support for IMF Version 4.3 (INCOMPATIBLE). Part of that

VMACCIMS incompatibility was because MXG tested IMSVERS LT '13'

Jul 14, 2007 (because CIMS/Boole made major record format changes when

IMS 1.3 replaced IMS 1.2 years ago), but now, IMSVERS has

'1010' for IMS Version 10, causing MXG to try to read the

current records with ancient 1.2 code. However, many new

new fields were inserted by IMF 4.3, which relocated the

DBD segments incompatibly; this won't happen again, as

one of the new fields added is the OFFDBTRL, the offset

to the DBD segments. These new variables are created

in the CIMSTRAN dataset:

TRNARVTH TRNARVTJ TRNE1STD TRNEAPPL TRNEAPPU TRNEDB2

TRNEDB2U TRNEDLDB TRNEDLTM TRNEF2 TRNELDLI TRNEMQS

TRNEMQSU TRNEOESS TRNEOESU TRNEOPCL TRNESYNC TRNETFLG

TRNETIME TRNEWLMC TRNMODET TRNMSCTH TRNMSCTJ TRNMSCUT

TRNNUOWP TRNOTCLF TRNOTCLN TRNOTCON TRNOTCOR TRNOTIP

TRNOTMAM TRNOTMAP TRNOTPRT TRNOTSCF TRNOTSLF TRNOTSOC

TRNOTSTC TRNOTSTF TRNOTSYF TRNOTUSR TRNRSENM TRNSMQGN

TRNSTCKE TRNSTOPA TRNSTOPU TRNSTRTA TRNSTRTU TRNTCPU

TRNUOW TRNW1OTH TRNW2IOO TRNW2IOV TRNW2LCH TRNW2OTH

TRNW3LCH TRNW3OTH TRNW4DBR TRNW4IO TRNW4OTH TRNW5IOD

TRNW5IOO TRNW5IOV TRNW5LCH TRNW5LCK TRNW5OTH TRNWLMRE

TRNWLMRT TRNWLMSC TRNWLMZR TRNXCKPC TRNXCKPM TRNXCKPT

Thanks to Thomas Heitlinger, FIDUCIA IT AG, GERMANY.

Thanks to Sieghart Seith, FIDUCIA IT AG, GERMANY.
Change 25.138 Support for APAR OA20314 which adds two variables

VMAC89 SMF89LPN='LPAR*NAME'

Jul 14, 2007 SMF89ZNA='LICENSE*ZNALC*IN*EFFECT?'

to both TYPE89 and TYPE892 datasets.

Thanks to Al Sherkow, I/S Management Strategies, USA.
Change 25.137 INPUT STATEMENT EXCEEDED RECORD LENGTH error with SMF 80

VMAC80A record that contained new TOKDANAMs CPUTIMEMAX,

Jul 13, 2007 ASSIZEMAX and FILEPROCMAX segments, because MXG did not

protect for unknown tokens. This fix protects so the

record will be read and the new tokens will be skipped.

This change and text will be revised to support those

three new tokens.

Thanks to Gerald Nagao, University of California, USA.

Change 25.136 Two utility programs to count PDS Directory Blocks used

TYPELPDS and defined.

UTILLPDS The first example, in TYPELPDS, uses IBM LISTVTOC and

Jul 4, 2007 LISTPDS programs to create flatfiles that are then read

Aug 7, 2007 in a second SAS step (so this could be used by sites with

SAS only on ASCII), and it captures all PDS on a volume,

with no prior knowledge of their DSNAMES.

The second example, in UTILLPDS, is a SAS example that

gets the blocks for a single dataset.

Christa requested the facility. Chuck wrote TYPELPDS.

Geert contributed UTILLPDS. I wrote the change/comments.

Thanks to Chuck Hopf, Bank of America, USA.

Thanks to Christa Neven, KBC Bankverzekerinngsholding, BELGIUM.

Thanks to Geert De Batselier KBC Bankverzekerinngsholding, BELGIUM.

Change 25.135A The label for TYPE70PR variables LCPUCAP and LCPUCAPC now

TYPE7072 include "Hard CAP" because those flag variables are set

Jul 4, 2007 'Y' when Hard Capping was specified when the LPAR was

defined (LCPUCAP) or if the Hard Capping status of the

LPAR was changed during this interval (LCPUCAPC). Hard

Capping tells PR/SM to cap resources delivered to that

LPAR at that LPAR's Designated Share (which is derived

from the weight given to this LPAR versus the total of

the weights you gave all of the other LPARs in the CEC).

Once an LPAR specified as hard capped reaches its

designated share, that LPAR is then "hard capped" and

normally does not receive more than its share of total

CEC resources.
Soft capping of an LPAR is done by WLM in response to

license manager determining that the LPAR's rolling

4-hour average MSU has exceeded the MSU specified for the

LPAR. A Soft Capped LPAR will have PDB.TYPE70PR variable

LPARNSW='PERCENT*LPAR SOFT*CAPPED' nonzero in soft capped

intervals (and LPnNSW variables nonzero in PDB.ASUM70xx

ASUMCExx summary datasets).
To further confuse the issue, there can be resource group

capping of service classes based on Resource Group

specifications.
Even more confusing is that Discretionary Management

algorithms can cap a service class period that is a

candidate for Discretionary Mangement if it significantly

exceeds its performance goal (PI LT 0.71), and this

Discretionary Management cap will show up even though the

service class is not assigned a Resource Group cap.

Thanks to ???, ???, FRANCE, for asking why LCPUCAP was never 'Y'.

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

the above answers.
Change 25.135 %ANALRMFR(PDB=PDB,REPORT=DEVC,DEVICEC=20) caused several

ANALRMFR errors (variables CPUTYPE CPUVERSN, variable CAIM NOT

Jul 4, 2007 FOUND). A list with dashed names CPUSER0-CPUSER9 was

enumerated and CAIM was removed (no such variable), and

UNINITIALIZED VARIABLES SMF70CPA and SMF70WLA were fixed

by their addition to ID statements where needed. But the

removal of CPUTYPE and CPUVERSN from the TEMP74 ID list

was the correct revision, because the don't exist when

a report for only RMF 74 records is requested.

Thanks to Esti Sas, Bank Lemui, ISRAEL.

====== Changes thru 25.134 were in MXG 25.06 dated Jul 4, 2007=========
Change 25.134 Support for IRRDBU00 record types 0560, 0561, 0562 and

EXRAC560 05C0 create four new datasets (all of which, unlike prior

EXRAC561 IRRDBU00 records, contain multiple segments per record).

EXRAC562 dddddd Dataset Description

IMACRACF RAC560 RACF0560 GEN RES CERTIFICATE DATA

VMACRACF RAC561 RACF0561 GEN RES CERTIFICATE REFERENCE

VMXGINIT RAC562 RACF0562 GEN RES KEY RING DATA

Jul 3, 2007 RAC5C0 RACF05C0 GEN RES CDTINFO DATA

Thanks to Aimee Steele, Euroclear, BELGIUM.
Change 25.133 Support for Williams Data Systems FERRET product user SMF

EXFERTCP creates three new datasets:

EXFERTEE

EXFERTRT DDDDDD MXG MXG

IMACFERT DATASET DATASET DATASET

TYPEFERT SUFFIX NAME LABEL

TYPSFERT

VMACFERT FERTCP FERRETCP FERRET CP SUBTYPE 1

VMXGINIT FERTEE FERRETEE FERRET EE SUBTYPE 2

Jun 30, 2007 FERTRT FERRETRT FERRET RTP SUBTYPE 3

Change 25.132 AS400 TYPECONF variables GDESCL,GDESCN,GDESED,GDESET, and

VMACQACS GDES91 are now kept, GDES2 is formatted TIME8, and GDES3

Jun 30, 2007 is now INPUT as $7 (Model+System-TYPE - e.g., 8709406).

Date/Time variables GDESED, GDESET are character text, so

they cannot be formatted, but new GDESEDD and GDESETT are

numeric variables with DATE9 and TIME8 formats.

Thanks to Urs Kugler, Zurich Insurance Company, SWITZERLAND.
Change 25.131 -Support for RACFEVNTs 52 (SET UID) and 79 (CRL PUBLISH)

EXTY8052 creates two new datasets from SMF 80 records:

EXTY8079 DDDDDD MXG MXG

IMAC80A DATASET DATASET DATASET

VMAC80A TY8052 TYPE8052 RACF EVT 52 SET UID

VMXGINIT TY8079 TYPE8079 RACF EVT 79 CRL PUBLISH

Jun 30, 2007 -Support for TOP SECRET records with RACFVRSN '90'x.

Jul 2, 2007 The '90'x records contain ONLY the "standard" relocate

sections (they have SMF80DTP=RACFTYPE values 1-255, there

are NRSET sections, and they are located at OFFSET), and

the header ends at SMF80VER/RACFVRSN, the "short" CA SMF

80 header, which is different than the "standard" IBM 80

record header; MXG has always detected IBM vs TOP SECRET

format based on RACFVRSN and found the short header.

-Support for TOP SECRET records with RACFVRSN '00'x is an

INCOMPATIBLE change that required new logic to detect the

new and different format. The '00'x records contain ONLY

the "extended" relocate sections (they have segment type

SMF80TP2=EXTLNTYP values of 256-65535, there are SMF80OC2

sections, and they are located at SMF80OC2), but the '00'

records now have the "full" IBM header (thru SMF80RSV).

And, of course, CA changed the SMF80OC2 offset value in

the '00'x, but not in the '90'x, to match the way IBM

has always (incorrectly!) set OFFSET and SMF80OC2:

In all other IBM SMF records, offsets are from the RDW

and the INPUT is @OFFSET-3, but the IBM SMF 80s have

always required @OFFSET+1. The original and '90'x Top

Secret records require -3, the new ones require +1.

-The MXG code that formerly created TYPE80A and TYPE80CM

datasets has been changed, and there now should never be

any observations written to those datasets. Previously,

TYPE80A was OUTPUT if NRSET=0 (e.g., RACFEVNT=79 that

has only extended relocate segments), and TYPE80CM was

OUTPUT when an unknown SMF80DTP segment type was found.

TYPE80A now will contain obs only if the SMF 80 records

has no segments: OUTPUT only if NRSET=0 AND SMF80OC2=0.

TYPE80CM will contain obs only if SMF 80 records contain

relocate sections (standard or extended) that MXG does

not (yet) decode, but if you send your SMF records to

[e-mail], those segments will be decoded and those

observations will go away.

-Top Secret Documentation: MXG decodes the Top-Secret-Only

SMF80DTP segment values of 103, 104, 105, 109, and 114,

and outputs one observation in the TYPE80TS for each

segment (i.e., multiple observations from one SMF 80

record). MXG also outputs one observation per record to

the appropriate TYPE80xx dataset for that RACFEVNT=xx,

with all the "IBM" variables from the relocatable DTP/DT2

segments.

-New Top Secret DTP=115 and DTP=116 segments will be

decoded and output to TYPE80TS when documentation is

received; this text will be updated when completed.

Thanks to Greg Burt, Fifth Third Bank, USA.
Change 25.130 Support for Clarion Disk Array flat files creates new

EXCLARDI datasets:

EXCLARLU DDDDDD MXG MXG

EXCLARPO DATASET DATASET DATASET

EXCLARRA SUFFIX NAME LABEL

EXCLARSA

EXCLARSN CLARDI CLARDISK CLARION ARRAY DISK DATA

EXCLARSP CLARLU CLARLUN CLARION ARRAY LUN DATA

IMACCLAR CLARPO CLARPORT CLARION ARRAY PORT DATA

TYPECLAR CLARRA CLARRAID CLARION ARRAY RAID DATA

TYPSCLAR CLARSA CLARSAN CLARION ARRAY SAN DATA

VMACCLAR CLARSN CLARSNAP CLARION ARRAY SNAP DATA

VMXGINIT CLARSP CLARSP CLARION ARRAY SP DATA

Jun 29, 2007 Jul 5: Variable OBJECT removed from BY lists as it is

Jul 5, 2007 subordinate to their OWNERNAM, and it's parsed pieces

are already available in the other BY variables.

Thanks to Jim Quigley, ConEd, USA.
Change 25.129 Variable QW0224PN and QW0224CI are now created in the

VMAC102 T102S224 DB2 Trace IFCID=224 dataset.

Jun 29, 2007

Thanks to Christa Neven, KBC Bankverzekerinngsholding, BELGIUM.

Change 25.128 Variable UBRECRD is now formatted with $MGDCORF, and is

VMACDCOL INPUT as $CHAR1 instead of $EBCDIC1 so it is valid on the

Jun 27, 2007 ASCII platform as well as on EBDIC, and it is set to

LENGTH $1 (without the LENGTH set to $1, the length of

the format item, $MGDCORF, is used by SAS for the length

of the variable, wasting space).

Thanks to Trevor Ede, EDS New Zealand Limited, NEW ZEALAND.
Change 25.127 Support for Oracle Version 10 has been in place since MXG

VMACORAC Version 24.06, as there were no changes in the physical

Jun 27, 2007 contents of their user SMF record. However, V10 data for

Oct 16, 2008 the Kernel fields (MXG variables LOGREADS PHYREADS

LOGWRTS DMLCOMIT DMLROLLS DEADLOCK) is trashed with very

large values in the initial test records; a problem has

been opened with Oracle Technical Support and this note

will be updated when a fix is available from Oracle.

Oct 2008 update:

The incorrect I/O counts appear to have been fixed in

Oracle 10.2.0.3. The SMF fix 6725982 also requires fix

6138068 and 6646861.

Thanks to Sudie Wulfert-Shcickendanz, Anheuser-Busch, USA.
Change 25.126 Circumvention to skip new TMC 'FF20'x Volume Definition

VMACTMS5 record, awaiting doc and data to decode.

Aug 7, 2007
Change 25.125 VGETOBS added conditional use a DATA step for testing,

VGETOBS but that revision was removed and would not be executed.

Aug 7, 2007
Change 25.124 Preliminary changes to support testing of MXG under WPS.

AUTOEXEW -Member CONFIGW2 defines the SAS options for WPS execution

CONFIGW2 on z/OS platforms. Some SAS options do not yet exist in

MXGWPSV2 WPS, and some are no longer needed or are not changeable.

VMXGINIT It must be the last dataset in the //CONFIG DD in your

Jun 25, 2007 MXGWPSV2 JCL procedure.

Aug 8, 2007 -Member AUTOEXEW defines the SAS options for WPS execution

Sep 25, 2007 on ASCII platforms.

VMACSMF -VMXGINIT set new GLOBALed macro variable &WPSVER=2 when

execution under WPS V2 is detected, and sets the existing

macro variable &SASVER=8, so that MXG will generate code

for the SAS V8 environment that WPS V2 replicates.

Sep 25: test for SASVER EQ 2.00 in VMXGINIT was changed

to LT 6 when WPS changed version to 2.01.

-VMACSMF updated to protect FILENAME= option not in WPS.

Change 25.123 An undetected DIVIDE BY ZERO in VMXGRMFI occurred in the

VMXGRMFI PGPERBLK=PGPERBLK/BLKSAUIN calculation, but I do not know

Jun 25, 2007 why SAS did not detect and report the error (which is

inside the OUTCODE= of a PROC MEANS inside VMXGSUM). The

division is now protected.

Change 25.122 If BUILDPDB=NO, USERADD didn't function correctly for the

UTILBLDP TMNT, 5, 26J2, or 30 "products".

Jun 25, 2007

Thanks to Scott Barry, SBBWorks, Inc, USA.

Change 25.121 PDB.ASUMUOW dataset is enhanced to keep the response time

VMXGUOW of each of the CICS segments in the Unit of Work.

Jun 22, 2007
Change 25.120 Utility for counting CICS record, bytes, transactions for

UCICSCNT each region and subtype now separates CICS Resource field

Jun 22, 2007 segments from CICS Transaction segments in counts, and

report titles and labels were clarified.

Change 25.119 SMF 119 with (new in z/OS 1.8) IPV6 ICMP segments caused

VMAC119 INVALID DATA error messages due to blanks where &PIB.4.

Jun 20, 2007 should have been for the INFORMAT of several variables

in the TYP11905 dataset (after @OFF11907).

Thanks to Marty Rhodes, Hertz Corporation, USA.
Change 25.118 DEVMODEL='3380K' caused INVALID ARGUMENT error as noted

ASUMCACH in Change 23.073; logic revised to set MODEL=3380 when

Jun 20, 2007 DEVMODEL=:'3380' to circumvent.

Thanks to Tom Heller, Cincom Systems, Inc, USA.

Change 25.117 INVALID ARGUMENT TO FUNCTION INPUT reading SYNCSORT SMF

VMACSYNC records were due to incorrect HEX4 and HEX3 informats

Jun 19, 2007 that should have been HEX8 and HEX6, and (on ASCII only)

SYNSRTOU was input as $CHAR instead of $EBCDIC. The

variables that were wrong were Device Addresses

SIDEVNR and SODEVNR.

Thanks to Patrick Holloman, Zions Management Services Co., USA.
Change 25.116 MXG 25.05 GMTOFF30 was still wrong because Change 25.089

VMAC30 for TYPE30U6 GMTOFF30/INTETIME/INTBTIME was still wrong

VMAC30 in some cases; those variables are now created in the

Jun 19, 2007 _STY30U6 logic based on SMFTIME and the EXECTM; variable

SYNCTIME sometimes exists, and if it does, it is reset

from GMT; otherwise, it is set to INTETIME. With 25.05

TYPE30_V/SMFINTRV has negative EXECTM and ITRVLTM values.

With MXG 25.05,

Thanks to Rudolf Sauer, T-Systems Enterprise Services, GERMANY.
Change 25.115 Incorrect values for memory variables in XMUCDSYS were

VMACXAM the result of my inability to recognize PL/1 source, so

Jun 15, 2007 I overlooked several fields in this Linux under VM data.

Thanks to Robert Obee, IMS Health, USA.

Change 25.114 -The XCOM support is revised to conform to Change 18.338

VMACNDM by defining the MACRO _VTYXCOM for this product. Over

VMACSOLV time, all members will be revised to provide this option,

VMACXCOM but I tend to get motivated only when someone asks.

Jun 14, 2007 -Jun 27: SOLV enhanced to define _VTYSOLV, _VTYSOL1 macros

Jun 27, 2007 -Jun 27: NDM now defines _VNDMxx macro for all datasets.

Thanks to Trevor Ede, EDS, New Zealand.
Change 25.113 The HSM support is enhanced to create an "HSM COMPLEX"

ASUMHSM variable, HSMPLEX, for the HSM Summarization logic to

Jun 14, 2007 support sites with multiple DFHSM-managed storage

environment. HSMPLEX is blank by default, but the new

MACRO _ESUHSM has a commented exmaple on how you might

create values in HSMPLEX for your DFHSM data.

Thanks to Scott Barry, SBBWorks, Inc, USA.
Change 25.112 The CA IDMS Performance Monitor support is enhanced with

IHDRIDMS the insertion of an INCLUDE for IHDRIDMS member after the

VMACIDMS IDMS header variables have been input. New &MACIDMSH

VMXGINIT macro variable is defined so either the IHDRIDMS member

Jun 14, 2007 or it can be used to access the header variables.

Thanks to Scott Barry, SBBWorks, Inc, USA.

Change 25.111 A misplaced end comment after PUT 'DEBUG TWO' statement

VMAC80A after Change 25.024 caused variable TYPSTRNG to be wrong.

Jun 13, 2007 TYPSTRNG is an MXG-created variable that identifies which

RACFTYPE subtypes were found for this RACF event.

Thanks to John Hammonds, Texas Comptroller of Public Accounts, USA.
Change 25.110 -When there are more than some number of disks, NMON will

VMACNMON create additional objects DISKBUSY1/DISKREADn etc; they

Jun 12, 2007 are now supported.

-Inconsistent sort order BY statement were revised.

-JFSFILE name was increased in size to $128 to prevent

false duplicates

Change 25.109 z/OS 1.8 changed the meaning of UIC values in SMF71Uxx

VMAC71 variables to now represent the time for one steal cycle

Jun 14, 2007 through the whole storage area, with a range of values

from 0 to 65535 (18 hours). As before, low UIC implies

higher storage contention. The new UIC metrics for

Current UIC are calculated every second, while Min and

Max are the lowest/highest UIC during the last walk thru

the storage area. IBM has kept the old form of UIC

(values 0-2054) in SMF71LIC/HIC/ACA, MXG variables

HIUICMN/MX/AV, but this change now stores the new

SMF71Uxx values in the three old UIC MXG variables

HIUICMN/HIUICMX/HIUICAV variables, as their value has no

meaning with z/OS 1.8.

See Change 26.069, which corrected HIUICMN and HIUICMX.

Thanks to Karl Lasecki, Chemical Abstracts Service, USA.

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