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

non-zero only if VNO='FF'x. ADOC94 was updated to

identify the "Note 1 and Note 2" variables in TYPE94.

Thanks to Art Hunter, Penn Mutual, USA.

Change 21.190 Support for APAR PQ77633, which corrects FTPREPLY code in

VMAC119 FTP client records; it should have been a three digit RFC

Sep 29, 2003 code in left-justified EBCDIC, but the SMF 119 record had

Mar 2, 2004 a binary value prior to this APAR. See Change 22.019.

Change 21.189 CICS SMF 110 subtype 1 MNSEGCL=5 records caused ERROR:

VMAC110 INVALID DO LOOP CONTROL INFORMATION because I forgot that

Sep 23, 2003 DO _I_=1 TO MNR5LEN3; fails if MNR5LEN3=., and the third

triplet is not always present. Change to IF MNR5LEN3 GT

0 THEN DO _I_=1 TO MNR5LEN3 to fix.

Thanks to Pat Curren, SuperValu Inc, USA.

Change 21.188 NDM SMF user record subtypes 'IP','JI','QH','QW', and

VMACNDM 'WS' are still not undecoded, until an MXG user wants

Sep 23, 2003 those data, but the UNKNOWN RECORD messages are now no

longer printed for those subtypes.

Thanks to Trevor Ede, EDS, NEW ZEALAND.
Change 21.187 Data values in PDB.DB2GBPST were occasionally way too

VMACDB2 large; the logic to deaccumulate did not test the

Sep 23, 2003 SEQCHECK variable, and thus missed restarts/resets. The

test for SEQCHECK was also added to the logic for

PDB.DB2STATR and PDB.DB2STATB, although neither had any

reports of bad data values.

Thanks to Lori A. Masulis, Fidelity Systems, USA.
Change 21.186 Another major enhancement for RMF III VSAM support now

ASMRMFV adds support for six new RMF III data tables:

Sep 23, 2003 CFI (Coupling Facility Information Table)

Nov 18, 2003 RCD (Resource Collection Data Table)

SVP (Service Policy Data)

SHD (Sample Header Data)

RED (Resource Data Record)

UWD (Use/Wait Record)

See change 21.228 for updates to VMACRMFV to support the

new data added by this change to ASMRMFV.

See Change 21.236.

Thanks to Jerry Urbaniak, Acxiom, USA.

Change 21.185 INTERVAL=SECOND in VMXGSUM, VMXGDUR, and ANALCNCR is

ANALCNCR added to MINUTE HOUR etc values to set the interval in

ANALCNCR summarization and concurrency analysis (e.g. counting

VMXGSUM concurrent NT processes in each second). But use the

VMXGDUR INTERVAL=SECOND value with care, as it can require

Sep 23, 2003 workspace that is many times the size of the input data.

Sep 30, 2003

Thanks to Art Morelock, CheckFree, USA.

Change 21.184 Support for 3592 Tape Drives already exists in MXG, as

none they are still 3590s (UCBTY34='8083'x) in SMF records.

Sep 23, 2003 The specific type and model is contained in the TYPE74

for each device in variable SMF74NID='003592J....'.

Thanks to MP Welch, SPRINT, USA.
====== Changes thru 21.183 were in MXG 21.05 dated Sep 22, 2003=========
Change 21.183 SMF 115 Subtype 2 INPUT STATEMENT EXCEEDED because the

VMAC115 protection in Change 21.161 was incomplete.

Sep 21, 2003

Thanks to Bruce Widlund, Merrill Consultants, USA.

Change 21.182 Support for z/OS 1.4 Feature 1 (COMPATIBLE) that adds

VMAC89 SMF89LP3 variable with one or two digit LPAR ID value.

Sep 19, 2003 Feature 0 is "z/OS V1.4 z999 Compatibility Support".

Feb 13, 2004 Feature 1 is "z/OS V1.4 z990 Exploitation Support".

Feature 2 is "z/OS V1.4 Consoles Enhancements".
====== Changes thru 21.181 were in MXG 21.05 dated Sep 18, 2003=========
Change 21.181 TMS/CA1 Release 5.2 (finally!) lets you block the TMC

TYPETMS5 file to half-track blocksize of its 340 byte records, so

Sep 18, 2003 the MXG recommendation to use BUFNO=200 on //TMC is

withdrawn when you have enabled half-track blocksize.

That large a BUFNO with half-track blocksize could cause

an 80A out-of-virtual-memory ABEND. Reblocking the TMC

to half-track reduced the MXG run time of TYPETMS5 from

6.5 minutes to only 2.5 minutes.

Thanks to Chuck Hopf, MBNA, USA.
Change 21.180 -Example to build Weekly PDB from Daily PDBs on tape to

JCLWEEKD eliminate backspace/rewind/remount wasted time, using the

WEEKBLDD same technique as MONTHBLD: copy the selected PDB

WEEKBL3D datasets from tape to temporary DASD and build WEEKLY

Sep 17, 2003 from those disk copies rather than by reading tapes.

Thanks to Ron Lundy, AHOLDUSA, USA.

Thanks to Bruce Widlund, Merrill Consultants, USA.
Change 21.179 -TYPE50 records contain interval data, but there is no

FORMATS interval value in the record; this enhancement creates

VMAC50 new variable DELTATM as the delta between SMFTIMEs for

Sep 16, 2003 the same instance; however, DELTATM will be missing in

the first record for each instance, even though the rest

of the data is valid in those first-per-day data.

-Format MG050AT was updated for 3:Read Write Subchannel

Thanks to Dr. Alexander Raeder, Itellium, GERMANY.

Change 21.178 Variable STARTIME is now created in TPXINTRV dataset in

VMACTPX the _STPXINT (DIFFTPX) deaccumulation logic.

Sep 16, 2003

Thanks to Dr. Alexander Raeder, Itellium, GERMANY.

Change 21.177 Cosmetic; suppressed printing of temporary datasets that

VMXGPRAL were created by a prior run of VMXGPRAL.

Sep 12, 2003
Change 21.176 -Support for CICS APAR PQ76703 new Transaction Resource

EXCICDSR Class Monitoring TSQUEUE segment in SMF 110 Subtype 1

EXCICDST with MNSEGCL=5, which caused INPUT STATEMENT EXCEEDED

EXCICRDQ error: the new triplet and class were unexpected. You

VMAC110 can confirm this error by seeing MNSEGCL=5 in the list of

VMXGINIT variables after the error is printed. The new data

Sep 12, 2003 tracks Temporary Storage at the transaction level with ts

queue name, and is output in new CICSRDQU dataset. The

DFHMCT TYPE=INITIAL TSQUEUE=4 default captures four

queues, a maximum of TSQUEUE=32 is allowed, and if

TSQUEUE=0, no data segments will be created and the

CICSRDQU dataset will have zero observations.

-New STID=64 and STID=65 for DSR and DST DSECTS create two

new CICDSR and CICDST Statistics Datasets. These

segments were added by IBM APAR PQ76697.

Thanks to David Klein, DOITT - City of New York, USA.

Thanks to Eileen Barkow, DOITT - City of New York, USA.
Change 21.175 Support for DB2 102 IFCID=217 (Storage Manager Pool)

EX102T17 creates four datasets:

EX102U17 T102S217 QW0217HE DSECT - Storage Available

EX102V17 T102T217 QW02172 DSECT - Storage each QW0217PH

VMAC102 T102U217 QW02173 DSECT - Storage each QW02173H

VMXGINIT T102V217 QW02174 DSECT - Dictionary Storage

Sep 10, 2003 for initial investigation of the new data. These data

records are strange, with the 2nd,3rd,& 4th triplet's

segments containing different DSECTs in different SMF

records, so logic using Segment Length is required to

determine exactly which DSECT is in which triplet! The

only unique variable in T102T217 and T102U217 is the

undocumented "Serviceability" token, QW0217PH and

QW02173H. Revision is possible if a better way seen!

-Support for IFCID=254 (Group Buffer Pool) adds 22 new

variables to the T102S254 with only common variables.

existing T102S254

Thanks to Phil Parker, TNT Post Group, ENGLAND.

Change 21.174 The PROC FREQ in _RPDBID failed, if you used _STYID in

BUILD001 EXPDBOUT (to sort WORK.ID to PDB.ID), because MXG hard

BUIL3001 coded the DATA=_WTYID token, so SAS only looked for ID in

BUILDPDB the Work Lib. Instead, the %VMXGWORL macro is used to

BUILDPD3 determine where you left that dataset, populating macro

Sep 9, 2003 variable &MXGWORL ("W or L") with its location:

%VMXGWORL(DDDDDD=TYID);

PROC FREQ DATA=&MXGWORL;

This should have been done when %VMXGWORL was created.

Thanks to Hugh Lapham, RCMP, CANADA.

Change 21.173 SMF 116 Subtype 0 records with QWHS header LENQWHS=36

VMAC116 caused INVALID PRODUCT SECTION message, and were then

Sep 9, 2003 deleted (and hence not output in MQMACCT dataset).

LENQWHS=36 segment is now decoded, up thru QWHSMTN, but

QWHSLOCN/NID/LUNM/LUUV/LUCC/QWHSLUCN are all missing

values in MQMACCT short QWHS segment observations.

-Input of QWHCCV was changed to $CHAR12 from $EBCDIC12.,

to support execution under ASCII. See Change 22.089.

Thanks to Jacob Nudel, Thompson BETA Systems, USA.
Change 21.172 Variable PROCNAME in HPSUPROC dataset was not in the

VMACMWSU LENGTH $32 statement, so it defaulted to only 8 bytes.

Sep 9, 2003 It is now $32 bytes, as process names can be long.

Thanks to Chris Morgan, UFI, ENGLAND.

Change 21.171 ASCII execution only. RACF variables INTENT and ALLOW

VMAC80A were incorrect when MXG executed under ASCII SAS; the

Sep 7, 2003 INPUT of INTENTCH and ALLOWCH should be $CHAR1.

Thanks to Matthew T. Chappell, Queensland Transportation, AUSTRALIA.

Change 21.170 -TYPE70 variable NRCPUS is redefined to be the Average

VMAC7072 Number of CPUs that were Online, because IRD can vary

VMXGRMFI CPUs On and Off dynamically, and the original NRCPUS was

Sep 7, 2003 the integer number of CPUs online at interval end that

had not been varied, which is incorrect for IRD.

SMF70ONT, the Online Time from PR/SM segment is now

summed across all of the LCPUs in this MVS System, and

that sum is then used to recalculate NRCPUS

NRCPUS=SMF70ONT/DURATM

to now be the Average Number of Online CPUs, and its

Label changed. SMF70ONT is also now output in TYPE70,

and MXG percentages based on NRCPUS in TYPE70 were moved

to use the Average NRCPUS value in calculations. One

visible impact will be that NRCPUS will no longer be an

integer value!

-RMFINTRV was revised, adding new SMF70ONT variable and

with NRCPUS recalculated to be Average Number of CPUs.

Thanks to Cheryl Watson Walker, Watson & Walker, USA.

Change 21.169 Specifying KEEPALL=YES in %VMXGSUM invocation in these

BUILD005 three members improves performance, and eliminates

BUIL3005 confusing messages if you happen to test with OBS=1.

SPUNJOBS

Sep 7, 2003

Thanks to Chuck Hopf, MBNA, USA.

Change 21.168 ERROR: DATASET WORK.MXGSUM2 IS NOT SORTED IN ASCENDING

AUTOEXEC may occurs if your SAS installation's SORTSEQ= option is

CONFIGV8 not SORTSEQ=EBCDIC or SORTSEQ=ASCII, as documented in SAS

Sep 7, 2003 Note SN-07151, which has no fix yet. So to eliminate the

exposure with MXG jobs, the appropriate SORTSEQ=EBCDIC or

SORTSEQ=ASCII is now set in MXG's defaults.

Thanks to Brian Sanger, Zurich Financial Services, ENGLAND.
Change 21.167 Support for multiple uniquely-named CICS user fields is

EXUTILEX added by the new _UNIQUE macro defined in member

UTILEXCL EXUTILEX, allowing you to create your own named variables

Sep 19, 2003 in CICSTRAN from your user data fields.

Thanks to Terry Fox, Principal Insurance, USA.
Change 21.166 IIS Web Log data caused UNRECOGNIZED HEADER and ARRAY

VMACWWW SUBSCRIPT OUT OF RANGE because MXG expected all 21 IIS

Sep 4, 2003 fields were present. Logic to recognize the end of

fields and protect for non-existent fields was added.

Thanks to Mike Penlington, Westpac, NEW ZEALAND.
Change 21.165 CICS/TS 2.2 STID=126 UNEXPECTED DATA due to STILEN=284

VMAC110 because MXG didn't INPUT variable S6RSP9CT, but now it

Sep 3, 2003 does and the variable is kept in dataset CICCFS6D.

Thanks to Allen Mayer, Prudential Securities, USA.

Change 21.164 Typos caused unprintable characters in MXG 21.04. In

DOCLRMFV member DOCLRMFV ASCII '85'x "dashes" were converted to

IMAC6ESS EBCDIC '25'X, but 'real dash' is '2D'x ASCII, '60'X on

Sep 3, 2003 EBCDIC. In member IMAC6ESS, there is a '0D'x that should

have been '40'X on EBCDIC, '20'x on ASCII.

Thanks to Scott Barry, SBBWorks, USA.

Change 21.163 Enhancement permits un-sorted datasets in MONTHBLD and

MONTHBLD circumvention for the (very rare) NOT-SORTED ERROR if you

MONTHBL3 build a monthly from daily/weekly PDBs that were created

MONTHBLS with two different MXG versions that spanned a change in

Sep 3, 2003 MXG sort order (last time this happened was when SYSPLEX

had to be added to RMF sort order!). Now, inserting this

statement:

MACRO _BY % MACRO _SORTBY %

before the _MNTHBLD invocation, both the expectation that

the input dataset is sorted, and the assertion that the

output dataset is sorted, are both disabled.
The building-in-sort-order by MXG is only for the

performance of your report programs, so that if you PROC

SORT in the same (recommended) order, SAS will bypass the

unnecessary sort. Nothing else in MXG requires datasets

be sorted.
Once you have nulled those macros, all subsequent

invocations of _MNTHBLD will be unsorted in and out.

Thanks to Caron Knox, Willis Corroon, ENGLAND.
Change 21.162 Typo of IFDURTM/496 instead of IFDURTM/4096 caused

VMAC119 invalid value for IFDURTM.

Sep 3, 2003

Thanks to Elisabeth Ballet, Michelin, FRANCE.

Change 21.161 MQ V5.3 CF dataset MQMCFMGR with MQ 5.3 has very large

VMAC115 values. MXG assumed QESTLL of 64, as documented, but

Aug 27, 2003 protection for longer QESTLL was added to see if that

could account for the problem. Await SMF test data.

Thanks to Ruud van Zundert, UBS, ENGLAND.
Change 21.160 Support for TNG Version 7 (INCOMPAT, due to insertion of

VMACTNG one field in the CAPMPCM HEADER record).

Aug 27, 2003

Thanks to Colin Bowen, CSC, SOUTH AFRICA.

====== Changes thru 21.159 were in MXG 21.04 dated Aug 25, 2003=========
Change 21.159 Variable SYNCTIME in TYPE23 is on the GMT time zone, so

VMAC23 it should have had ",GMT" at the end of its

Aug 21, 2003 %VMXGTIME(SYNCTIME,SYSTEM,GMT); statement.

Thanks to Chuck Hopf, MBNA, USA.

Change 21.158 Support for new DFSMS/rmm EDGRXEXT data eliminates MXG

VMACEDGR "LESS THAN 545 BYTES" messages and incorrect data. New

Aug 20, 2003 variables in 'X' record, EDGRXEXT dataset:

RDLABNO ='LABEL NUMBER LABEL=(XX,LL) NEW'

RDLABNOX

RVALVERS='ANSI*LABEL*VERSION'

RVDCRSID='FIRST*FILE*CREATION*SYSTEM ID'

RVDSNNO ='LABEL*NUMBER OF*FIRST FILE*ON VOLUME'

RVEXPTKN='EXPORT*TOKEN'

RVLABNO1='LABEL*NUMBER OF*FIRST FILE*ON VOLUME'

RVOLTYPE='VOLUME*TYPE'

RVPERCNT='VOLUME*FULL*PERCENTAGE'

RVRBYSET='VOLUME*RETAINED*BY SET?'

RVSTKVCN='STACKED*VOLUME*COUNT'

RVSTKVLE='STACK*VOLUME*RECORD*ENABLED?'

New variables in 'D' & 'X' record, EDGRDEXT & EDGRXEXT

datasets: (SEE CHANGE 27.046 CORRECTIONS).

RDDEFRET='DEFAULT RETPD USED'

RDDSNSEQ='DATA SET SEQUENCE NUMBER NEW'

RDEXPDT ='DATA SET EXPIRATION DATE'

RDEXPDTO='ORIGINAL D/S EXPIRATION DATE'

RVDSTBIN='DESTINATION BIN NUMBER'

RVDSTMED='DESTINATION BIN MEDIA NAME'

RVMVDSN1='FIRST DSNAME OF A VOLUME SET'

RVVOL1 ='VOL1 LABEL VOLSER'

The two separate code blocks for the DEXT/RDEXT data was

consolidated into a single block and linked to.

Thanks to James D. Lieser, United Health Group, USA.

Change 21.157 DB2PM-like reports DB2 Accounting Summary had totals

ANALDB2R instead of averages for Class 2 Elapsed, Class 2 CPU, and

Aug 20, 2003 Lock Suspensions. Now they match DB2PM averages.

Thanks to Dr. Yao-Chun Rickert, Bank One, USA.

Change 21.156 Support for Control-D Log file creates 9 new datasets

EXCTLOGA reading their log records from the INFILE name CTLLLOG to

EXCTLOGB create a dataset for each record type:

EXCTLOGC MXG MXG

EXCTLOGD DATASET DATASET LOG

EXCTLOGE NAME LABEL TYPE

EXCTLOGF

EXCTLOGL CTLDOGA CTLD CONTROL REGION A

EXCTLOGM CTLDOGB CTLD ARCHIVE B

EXCTLOGX CTLDOGC CTLD USER C

FORMATS CTLDOGD CTLD DAILY D

IMACCTLL CTLDOGE CTLD AMF E

TYPECTLL CTLDOGF CTLD DECOLLATE F

TYPSCTLL CTLDOGL CTLD AUX REGION L

VMACCTLL CTLDOGM CTLD MIGRAION M

VMXGINIT CTLDOGX CTLD OTHER LOG TYPES ?

Aug 20, 2003 The "C" and "F" datasets contain the variables decoded

Nov 27, 2003 from the Log Message Text; for now, the other datasets

just contain the full LOGMSG text. If the CTLLLOG file

is to be sent to an ASCII platform for MXG execution, the

file must be sent as binary, as it contains both EBCDIC

text and binary values that cannot be translated to

ASCII. The correct DCB is RECFM=F,LRECL=6400, and that

is specified in the MXG program.

-Nov 27: Missing STARTIME in CTLDLOGF when log on/off time

stamps were identical corrected by revised SORT order.

Thanks to Elisenda Masacana, la Caixa, SPAIN.
Change 21.155 Support for Netspy Versions 6 and 7 already exists in MXG

VMACNSPY since MXG 20.10. Test V7 data was read without errors

Aug 18, 2003 nor any indications of inserted fields.

This change is documentation only; no change was made.

Thanks to Ron Byers, British Energy, ENGLAND.

Change 21.154 WARNING:LABELS EXCEED LENGTH 40 NOT SUPPORTED BY V6SEQ

ASUMCICS AND ARE BEING TRUNCATED had no impact, as it was only the

ASUMCICT DSNLABEL= argument revised in Change 21.090 that was too

ASUMCICX long.

Aug 18, 2003

Thanks to Pat Curren, SuperValu Inc, USA.
Change 21.153 -Support for BMC's MVIMS 3.3.0 (COMPAT) adds new data in

VMACCIMS reserved fields in 'FA'x IMS log record that are created

Aug 18, 2003 in these new variables in TYPECIMS dataset:

ABENDSYS='SYSTEM*ABEND*CODE'

ABENDUSR='SYSTEM*ABEND*CODE'

SAPEXTP1='SAP*EXIT*PASS*1?'

SAPEXTP2='SAP*EXIT*PASS*2?'

SAPTRNCD='SAP*TRANSACTION*CODE'

SMQFLAG ='SHARED*MESSAGE*QUEUE*ENVIRONMENT?'

SMQGROUP='SHARED*MESSAGE*QUEUE*GROUP NAME'

SRV

UOWTRANS='UOW*FOR*TRANSACTION'

The ABENDSYS/ABENDUSR variables already existed in the

CIMSPROG (program deschedule) dataset, they were added to

record Transaction Abends in CIMSTRAN dataset.

-Variables COREALOC/COREUSED are converted to bytes,

labeled 'BYTES*AVAILABLE/BYTES*USED' and formatted as

MGBYTES to "print pretty". Previously they contained the

count of 2K blocks.

Thanks to Fabio Massimo Ottaviani, DTS Italia, ITALY.

Change 21.152 Support for CICS/TS 2.3 (INCOMPATIBLE):

CICINTRV - New variables were inserted into CICSTRAN Subtype 1.

EXCICPGR - Two new CICS TCBs, 'J9' and 'JM' added, supported in

EXCICSJR Subtype 2 CICDS and CICINTRV datasets.

IMACEXCL - New Subtype 2 vars added to STID=2 and STID=117.

UTILEXCL - New Subtype 2 STID=118 and STID=119 create datasets.

VMAC110 -While you can use TYPE110/TYPS110/BUILDPDB to create the

VMXGCICI CICSTRAN dataset, we strongly recommend that you use the

VMXGINIT UTILEXCL program to read your CICS Dictionary records to

Aug 22, 2003 create an IMACEXCL tailoring member that will input and

Sep 1, 2003 keep only the CICSTRAN variables in your CICS records,

Sep 9, 2003 as that can reduce MXG runtime and save disk space.

Sep 10, 2003 UTILEXCL is required if you have excluded any fields

Nov 11, 2003 from your Performance Class (Subtype 1) records.

-New fields are inserted in SMF 110 Subtype 1 creating 17

new variables in CICSTRAN dataset:

CBSRVRNM='CORBA*SERVER*NAME'

DSMMSCCN='MVS*STORAGE*NO TCB*WAIT*COUNT'

DSMMSCTM='MVS*STORAGE*NO TCB*WAIT*TIME'

DSTCBMCN='TCB*MISMATCH*ELAPSED*WAIT*COUNT'

DSTCBMTM='TCB*MISMATCH*ELAPSED*WAIT*TIME'

EJBCRECT='EJB*BEAN*CREATION*CALLS'

EJBMTHCT='EJB*BEAN*METHOD*CALLS*EXECUTED'

EJBREMCT='EJB*BEAN*REMOVAL*CALLS'

EJBSACCT='EJB*BEAN*ACTIVATIONS'

EJBSPACT='EJB*BEAN*PASSIVATIONS'

EJBTOTCT='EJB*TOTAL*REQUESTS'

J9CPUTCN='USER TASK*J9 MODE*CPU TCB*COUNT'

J9CPUTTM='USER TASK*J9 MODE*CPU TCB*TIME'

KY9CPUCN='USER-TASK*KEY 9*TCB CPU*COUNT'

KY9CPUTM='USER-TASK*KEY 9*TCB CPU*TIME'

KY9DISCN='USER-TASK*KEY 9*TCB DISPATCH*COUNT'

KY9DISTM='USER-TASK*KEY 9*TCB DISPATCH*TIME'

-Two new TCBs, 'JM' and 'J9' are created in the CICDS and

CICINTRV datasets from the STID=60 Dispatcher Statistics

110 subtype 2 record, creating new DSExxxxx and DSFxxxxx

variables for those 14th and 15th TCBs. MXG input based

on TCB Number is no longer valid, as IBM uses the same

TCB number for different named TCBs; the TCB name is now

used to store TCB data into the correct set of variables

and the MXG Variable label accurately identifies what

TCB is stored in each set of MXG variables:

Prefix: DSG DS2 DS3 DS4 DS5 DS6 DS7 DS8 DS9

TCB Name: QR RO CO SZ RP FO SL SO J8

Prefix: DSA DBB DSC DSD DSE DSF

TCB Name: L8 S8 H8 D2 JM J9

-The VMXGCICI member was revised to include the two new

TCBs in the PDB.CICINTRV dataset.

-New variables added from STID=2 DFHSMSDS to CICSMDSA:

SMSWAITM='TOTAL*TIME*WAITING FOR*MVS STORAGE'

SMSWAICN='NUMBER OF*REQUESTS*CAUSING*WAITS'

-New variables added from STID=117 DFHSJGDS to CICTCPSJ:

SJGCUWCJ='JVM*CURRENT*WORKER CACHE*JVMS'

SJGMXWCJ='JVM*PEAK*WORKER CACHE*JVMS'

SJGRQWCJ='JVM*REQUESTS*CLASS*CACHE'

-New CICSJR dataset created from STID=118 DFHSJRDS with

the JVMPROFILE Resource Statistics, one observation per

Storage Key for each Profile/Path:

SJRJVCRE='SJR*NEW*JVMS*CREATED'

SJRJVDES='SJR*JVMS*DESTROYED*SOS'

SJRJVHWM='SJR*JVM*HEAP*HWM'

SJRJVRES='SJR*JVMS*RESETTABLE'

SJRLEHWM='SJR*LE*HEAP*HWM'

SJRMMSTE='SJR*MISMATCH*STEALER'

SJRMMVIC='SJR*MISMATCH*VICTIM'

SJRPRCLC='SJR*PROFILE*CLASS*CACHE*SETTING'

SJRPRCUR='SJR*CURRENT*PROFILE*USE COUNT'

SJRPRFRQ='SJR*PROFILE*REQUESTS'

SJRPRHWM='SJR*PEAK*PROFILE*USE*COUNT'

SJRPRNAM='SJR*PROFILE*NAME'

SJRPRPTH='SJR*PROFILE*PATH*NAME'

SJRPRXMX='SJR*PROFILE*XMX*VALUE'

SJRSTOKY='SJR*STORAGE*KEY'

-New CICPGR dataset created from STID=119 DFHPGRDS with

the JVM Program Resource Statistics, one obs per record

Storage Key:

PGRJVCLS='JVMPROGRAM*JVMCLASS*NAME'

PGRJVKEY='JVMPROGRAM*CICS/USER*KEY'

PGRJVPGM='JVMPROGRAM*NAME'

PGRJVPRO='JVMPROGRAM*PROFILE'

PGRJVUSE='JVMPROGRAM*USE*COUNT'

-The new and changed STIDs have been validated with data.

-Note: Several new fields in the SMF records can have up

to 255 characters; for now, MXG only inputs the first

200 characters, so that MXG can still execute under SAS

Version 6, with its 200-byte limit. But as soon as any

MXG customer actually has a name field that is 201 bytes

of text, I will revise those INPUTs to $EBCDIC256, and

the SMF 110 code will have to be executed under SAS V8.

P.S. New stuff, like WebSphere, that is not in the

default BUILDPDB already requires SAS V8!
CICS/TS 2.3 Documentation Errors:
1. In the Data Areas, the SJGDS description does not