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

CPUIFATM='TOTAL*CPU TIME*ON*IFA'

CPUDFATM='DEPENDENT*ENCLAVE*CPU TIME*ON IFA'

CPUEFATM='INDEPENT*ENCLAVE*CPU TIME*ON IFA'

CPUIFETM='IFA-ELIGIBLE*CPU TIME*ON*CP'

CPUDFETM='IFA-ELIGIBLE*DEP ENCLAVE*CPU TIME*ON CP'

CPUEFETM='IFA-ELIGIBLE*IND ENCLAVE*CPU TIME*ON CP'

and these variables are added to the PDB.JOBS & PDB.STEPS

in the BUILDPDB/BUILDPD3 logic.

Thanks to Pat Curren, SuperValu Inc, USA.
Change 22.220 All 275 ADOC members were updated with new variables and

ADOC.... the variable lengths were corrected back to their MVS

Aug 22, 2004 lengths. The text in these documentation members was not

touched by the elegant utility that Freddie implemented,

which uses the text in DOCVER to update the variables.

Thanks to Freddie Arie, TXU, USA.

====== Changes thru 22.219 were in MXG 22.08 dated Aug 20, 2004=========
MXG 22.08 was re-dated Aug 20, 2004, when I discovered seven members

were missing from the 3480 tapes,including VMAC0 which caused JCLTESTx

to fail. The re-dated version included these additional changes:
Change 22.219 -Variable DISKRATE was misspelled in two KEEP= lists ad

VMACMWUX DISKRT.

Aug 20, 2004 -Variable DISKNAME was increased to $64 as $20 was too

short; only $35 were seen thus far, but this will avoid

yet another change later, hopefully!

-Comments revised to caution that a dollar sign should not

be used as a delimiter, as dollar signs appear in the tty

field in the MeasureWare PROC records.

Thanks to Roman Gudz, Penske Logistics, USA.

Thanks to Albert Jacobs, KBC, BELGIUM.

Change 22.218 The PDB exit _EDBJOBS was relocated so it precedes the

BUILD005 OUTPUT ACCOUNT; statement; if you used the _EDBJOBS exit

BUIL3005 to populate ACCOUNTn variables for Started Tasks that do

Aug 20, 2004 not have accounting parameters on their "JOB" card, your

code worked just fine for the PDB.JOBS dataset, but the

observations in PDB.STEPS and especially in PDB.SMFINTRV

contained blanks for the ACCOUNTn variables, because the

ACCOUNT (temporary) dataset is used to "back fill" those

other PDB datasets. With this code relocation, any exit

changes to ACCOUNTn variables will be propagated into the

PDB.STEPS, PDB.PRINT, and PDB.SMFINTRV datasets.

Note: It used to be that you HAD to use a table lookup

based on JOB name to assign ACCOUNTn variables

for STCs, but for all current OS/390 and z/OS,

you can put account fields on the "JOB" card for

the STC, and they will automatically be carried

into the PDB.JOBS/STEPS/PRINT/SMFINTRV datasets.

Thanks to Ken Jones, Xwave, CANADA.

Change 22.217 The extraneous IF in IF BYTEAVAIL=MAX(...) was removed,

VMACNTSM but it had no impact since the prior test for AVAILMEK

Aug 20, 2004 was never satisfied.

Thanks to Xiaobo Zhang, ISO, USA.

Change 22.216 The utility to convert character-variables-containing-hex

UTILCVRT after you move SAS datasets from MVS to ASCII worked fine

Aug 20, 2004 but added temp variables LENGTH and _I_ to the output.

Now it doesn't, and LENGTH was respelled as _L_.

Thanks to Al Sherkow, I/S Management Strategies, Ltd., USA.
Change 22.215 Creating CPU graphs of NT SMF data caused errors VARIABLE

BLDNTPDB STARTIME NOT FOUND and INVALID VALUE FOR SORTEDBY; the

Aug 20, 2004 sumby= statement was corrected and the DROP was removed.

Thanks to Matthew Chappell, Queensland Transport, AUSTRALIA.

Change 22.214 Commas were missing after XX='EN' and YY='WO' in lines

MXGSASV9 15 and 16 in the JCL procedure for SAS V9.

Aug 20, 2004

Thanks to Wade Peterson, McMaster-Carr Supply Company, USA.

Change 22.213 Variable R723CRTF was not kept in the TRND72GO dataset

TRND72GO because it was misspelled as R723CRFT in the ID=

Aug 19, 2004 statement.

-The default example TRND72GO summarizes at the SRVCLASS

level of detail, but you may prefer to summarize at the

PERIOD level, since Goals and Importance can be set at

the lower level. You can add PERIOD at the end of the

SUMBY= statement in your tailored TRND72GO if you want

the PERIOD level included in the trending summary.

Thanks to Rick Mansfeldt, IBM Global Services, USA.

Change 22.212 The FTPREPLY field changed from a binary value to EBCDIC

VMACTCP numeric; APARs PQ92769 and PQ83055 documents this change

Aug 18, 2004 in the format of the field; MXG code supports both the

old and the new format transparently.

Thanks to Tom White, SPRINT, USA.
Change 22.211 The JCL example for testing with SAS Version 9 needed the

JCLTEST9 //MONITASK and //SPIN DDs in step TESTOTHR.

Aug 17, 2004

Thanks to Paul Gillis, Pacific Systems Management Pty. Ltd

Change 22.210 JES3 JMF SMF 84 caused INPUT STATEMENT EXCEEDED RECORD;

VMAC84 the statement LOCJSTOF+LOCJSTOF+48; in VMAC84 should

Aug 17, 2004 have been: LOCJSTOF=LOCJSTOF+48;

Thanks to Paul Gillis, Pacific Systems Management Pty. Ltd

Change 22.209 While all INFORMAT xxx $NOTRAN statements were removed in

BUILD005 Change 22.192, using PROC SYNCSORT still caused BUILDPDB

BUIL3005 to fail, because the SPIN.SPIN30_4 dataset had variables

Aug 17, 2004 with that INFORMAT. This change removes the informat for

the two variables ABNDRSNC TERMIND as that dataset is

read, eliminating that (hopefully!) final exposure.

Alternatively, you could just use

DATA SPIN.SPIN30_4;

SET SPIN.SPIN30_4;

INFORMAT ABNDRSNC TERMIND ;

to replace the SPIN30_4 data set and remove the INFORMAT.

-I discovered the INFORMAT statement must be AFTER the SET

statement to remove the informat from the variables!

Thanks to Michael L. Kynch, International Paper, USA.

Thanks to Peter Giles, Centrelink, AUSTRALIA.
Change 22.208 -Support for TOKDANAM=GID creates new variable TOKGID in

IMAC6ESS TYPE80xx datasets that can contain the 53 segment.

VMAC80A -Removed unneeded PUT statement in IMAC6ESS member that

Aug 13, 2004 was observed in the log.

Thanks to Engelbert Smets, Provinzial Rheinland Versicher., GERMANY.
====== Changes thru 22.207 were in MXG 22.08 dated Aug 13, 2004=========
Change 22.207 FLASH. For MVS SAS V9.1 and V9.1.2. SAS Note SN-12943

CONFIGV9 reports incorrect results with PROC MEANS,SUMMARY,REPORT

Aug 13, 2004 and TABULATE only on "MVS", if threading is used, and the

SAS default option is THREADS. Specifying NOTHREADS does

corrects the error, which is fixed in SAS V9.1.3, and SAS

will have a Hot Fix very soon for V9.1 and V9.1.2, but I

still decided this was sufficiently critical to re-date

the MXG 22.08 release and include NOTHREADS in CONFIGV9

member, since you could easily change that option in your

CONFIGV9 later, or use OPTIONS= on your // EXEC statement

when you have the fix and find threading is of value.

Thanks to MP Welch, SPRINT, USA.

Change 22.206 Cosmetic. Some fields caused 'Note 49' and blanks were

VMACEREP inserted, to support that future version possible change.

IMACICBB

Aug 13, 2004

Thanks to Thomas Heitlinger, FIDUCIA IT AG, GERMANY.
====== Changes thru 22.205 were in MXG 22.08 dated Aug 12, 2004=========
Change 22.205 New CICS Statistics datasets were not listed in IMACCICS,

IMACCICS and worse, were not invoked in the _CICSTAT nor _S110ST

VMAC110 macros, and the example to use both in Change 18.152 did

Aug 12, 2004 not work. The new datasets are now added/listed.

-The _S110ST macro sorts all CICS Statistics datasets from

the "Wdddddd" work default (WORK) to their "Pdddddd" PDB

library.

-The _CICSTAT changes both the _W and _Pdddddd names, to

the value (ddname) you stored in macro variable CICSTAT,

but could not be used with the _S110ST macro because the

same DD would be used for both _W and _P datasets.

-The new _CICSTAS (S=sort!) changes only the Pdddddd for

the output of the SORTs, so it can be used in EXPDBOUT

ahead of _S110ST to sort all CICS statistics datasets to

the DDname specified in macro variable CICSTAT.

To COPY all cics stats datsets to DDname "CICSSTAT", use

%LET CICSTAT = CICSSTAT;

_CICSTAS;

_S110ST

in your EXPDBOUT member, and all of the CICS statistics

datasets will be CICSSTAT.CICxxxxx.

Thanks to Nori Abakah, Progress Energy, USA.

Change 22.204 Support for new fields added by DB2 V7 and V8 in T102S172

VMAC102 for IFCID=172 (all units of work involved in a deadlock).

Aug 12, 2004 The test for QWHSRELN EQ 7.1 was changed to GE 6, as the

Aug 13, 2004 "V7" data existed starting with DB2 Version 6.

Thanks to Andy Creet, Department of Defence, AUSTRALIA.

Thanks to Brad Kiemann, Department of Defence, AUSTRALIA.

Change 22.203 Cosmetic. VGETJESN decodes JCTJOBID to create JESNR and

VGETJESN TYPETASK, but if JCTJOBID was hex zeros or blank, as for

Aug 12, 2004 some records for JOB='INIT', unnecessary messages that

WARNING - TYPETASK NOT DECODED were printed on the log.

Now, VGETJESN tests to see that JCTJOBID is non-blank,

so the unnecessary warning messages won't be printed.

Thanks to Ng Kin, Acxiom, USA.
Change 22.202 Cosmetic. Invalid SMF 42 subtype 21 record cause MXG to

VMAC42 print an MXG error message that APAR OA02184 corrects the

Aug 11, 2004 bad records, but that APAR only corrects errors when the

delete used ISPF STOW command; if the delete was done by

DESERV FUNC=DELETE, the record is wrong and IBM has just

created APAR OA08693 to correct that error. The text of

MXG's error message said the record was DELETED, but at

the point of the error, MXG had already output TYPE4221.

The bad segment prevents an output to TYPE422A dataset,

so the text of the error message was clarified that it is

the rest of the record that is being deleted.

Thanks to Scott Wiig, U.S. Bank, USA.

Change 22.201 Aren't you glad SAS developers use MXG for testing new

VMXGSUM releases of SAS? Their early tests of SAS V9.2 found an

Aug 11, 2004 MXG syntax error that had previously been absorbed with

no compiler error. The text %ELSE %THEN %DO; was

corrected to %ELSE %DO;

Thanks to Rick Langston, SAS Institute Cary, USA.

Change 22.200 Variables SASVEREL SASUSER and SASJOBID are created by a

VMACSASU SUBSTR() of a $64 variable, so they defaulted to length

Aug 11, 2004 of 64. They are now defined in a LENGTH $8 statement.

These fields were added by Change 22.251, but, now, see

Change 22.303 for the (final?) correction.

Thanks to Dr. Alexander Raeder, Itellium, GERMANY.

Change 22.199 Major revision to IMS0708 dataset, created by TYPEIMS7,

EXIMS78 from the IMS 07/08 log records. Previously, transactions

EXIMSMRY with the same DTOKEN were summed into only one output

TYPEIMS7 observation, but that destroyed the detail "servictm" of

VMACIMS each program schedule event, and there can be thousands

VMXGINIT of Quick Restart program schedules with the same DTOKEN.

Aug 12, 2004 And each program schedule event could have serviced many

(NMSGPROC) transactions for each detail event. This

redesign eliminates the summarization by DTOKEN, creating

one observation for each program schedule event, with the

detail NMSGPROC, IMSCPUTM, etc., for each schedule event.

-If both 08 and 07 are found, STRTTIME and ENDTIME will

be non-missing; if only an 07 end event with no matching

08 start is found, STRTTIME will be missing, and for an

08 only with no 07 start (or back-to-back 08s) ENDTIME

will be missing.

-For events with the same DTOKEN, new variable INSTANCE

counts individual event instances.

-New variables SYSABEND and USRABEND decode the COMPCODE

into more meaningful variables in IMS0708 detail dataset.

-VMACIMS had to be revised to increase the stored length

of IMSRECNO to 8 bytes; it is now used to sequence events

within each DTOKEN. Originally it was a PIB4. field that

was stored in 5 bytes, but it was changed to a PIB8 field

in IMS Version 6, and should have been increased then.

-VMXGINIT was updated to support the new tokens PIMS78 and

PIMSMRY which can be used to set the DDNAME of IMS0708

and IMSSUMRY datasets. The default value for the output

DD is the _IMSTRAN old-style macro, so if you have set it

in your IMACIMS7 member to "IMSTRAN", the new TYPEIMS7

will still create IMSTRAN.IMS0708, or the example in the

TYPEIMS7 member shows you can use %LET PIMS78=IMSTRAN; in

your //SYSIN stream to set the destination DDname.

-But in case you really want the summary by DTOKEN, new

IMSSUMRY data set can be created by the _IMSUMRY macro,

as documented in comments in TYPEIMS7 member.

-And the original TYPEIMS7 program is preserved in the MXG

member ZTYPIMS7, just in case you feel a need to revert.

-The requestor of this change also wanted to create a new

dataset, IMSABEND, with only ABENDing transactions, but I

decided against making that an "MXG dataset", because the

IMACKEEP member and the _KIMS78 and _EIMS78 macros can

easily be used to create the new dataset, as shown below

in the example that also creates a new ORIGUSID variable

at the same time:

MACRO _KIMS78

ORIGUSID )

IMSABEND (KEEP=COMPCODE ORIGUSID USRABEND SYSABEND

%

MACRO _EIMS78

ORIGUSID=IMSUSID;

IF '0' LE IMSUSID LE '9' THEN IMSUSID='INTERNAL';

OUTPUT _LIMS78;

IF COMPCODE NOT

IN('00000000'X,'40404040'X,'20202020'X)

THEN OUTPUT IMSABEND;

%

The use of the _Kdddddd and _Edddddd macros to create new

datasets is documented in the DOCMXG member.

Thanks to Glenn Lundgren, SEB IT Service, SWEDEN.

Change 22.198 The PCTWFLOW calculation was revised to match the WFL %

ZRBRPT3 value shown on the RMF III panels, using the revised

Aug 6, 2004 equation from the RMF Programmers Guide that had been

overlooked since 1996.

Thanks to Randy Shumate, LexisNexis, USA.

Thanks to Steven L. Hankins, LexisNexis, USA.

Change 22.197 The Ficon Open Exchange analysis program was revised to

ANALFIOE keep only Online channels, and to delete Ficon Bridge

Aug 5, 2004 channels, to which the analysis does not apply.

The example in comments to create the datasets from SMF

was revised again, now using the three _Sdddddd dataset

sort macros instead of the _Sxxxx product sort macros.

Thanks to Dr. H. Pat Artis, Performance Associates, USA.

Thanks to Betty Wong, Bank of America, USA.

Change 22.196 Support for longer length DB2 identification variables:

VMACDB2H QWHSLOCN QWHSAID QWHCOPID QWHCEUID QWHDRQNM QWHDSVNM

Aug 5, 2004 is prepared to read in the up-to-128-byte character text,

but the length of those variables is not changed, yet.

Only when you actually increased-length data would you

want to increase their stored length in each observation

in your DB2ACCT dataset, and that can be done in your

//SYSIN in the job that creates your DB2ACCT data, using:

%LET MACFILE= %QUOTE( LENGTH QWHSLOCN $128;);

The longer fields can also now be in UNICODE, but until I

have test data, and until there is a need, I chose not to

to add that alternative code (UNICODE requires SAS V8,

and because the DB2 code is used in BUILDPDB, the longer

I hold off adding the UNICODE support, the longer those

few sites still running SAS V6 can limp along).

Thanks to Charles Harbour, John Deere, USA.

Change 22.195 INVALID DATA FOR YYYY warning for SMF 102 IFCID=22 when

VMAC102 the QW0022TS field contained nulls; the inputs with NUM

Aug 4, 2004 informat were protected with ?? syntax.

Thanks to Jim Poletti, Edward Jones, USA.

Change 22.194 Typo. SUN.ASUM79PR should have been SUN.ASUM70PR.

WEEK70PR

Aug 4, 2004

Thanks to Ingegerd Jansson, Volvo Information Technology AB, SWEDEN.

Change 22.193 Support for many new NTSMF objects, including DTS CPU

EXNTDTCP configuration records, Exchange 2003, Exchange 2000

EXNTDTLP instant messaging, Outlook 2003, and RSVP objects create

EXNTIACP these new datasets:

EXNTIAUP dddddd dataset description

EXNTMEII NTDTCP DTSCPU NT DTS.CPU

EXNTMETF NTDTLP DTSLOGPR NT DTS.LOGICALPROCESSOR

EXNTMSAS NTIAUP IASAUTPR NT IAS AUTHENTICATION PROXY

EXNTMSDC NTIACP IASACCPR NT IAS ACCOUNTING PROXY

EXNTMSGC NTMEII MSEXIM NT MSEXCHANGEIM

EXNTOUTL NTMETF MSEXTRFS NT MSEXCHANGE TRANSPORT FILTER SINK

EXNTRSVI NTOUTL OUTLOOK NT OUTLOOK

EXNTRSVP NTRSVI RSVPINTR NT RSVP INTERFACES

EXNTSMSM NTRSVP RSVPSERV NT RSVP SERVICE

IMACNTSM NTMSAS MSEXCHAS NT MS EXCHG ACTIVESYNCNOTIFY OMAPUSH

VMACNTSM NTMSDC MSEXCHDC NT MS EXCHG DSACCESS DOMAIN CONTROLLER

VMXGINIT NTMSGC MSEXCHGC NT MS EXCHG DSACCESS GLOBAL COUNTERS

Aug 3, 2004 NTSMSM SMSMSE NT SMS SMSMSE 4.5

Aug 5, 2004 -Most of the sample data records from the test system have

zero values, so it is impossible to know which, if any,

will need to be de-accumulated; if you have real data for

any of these objects, please contact [e-mail] for

instructions on sending us your file for validation.

-The new DTSCPU and DTSLOGPR objects indicate HyperThread

Support and Status. MXG merges the two objects into an

enhanced PDB.DTSLOGPR dataset (in the _SNTDTLP sort macro

when you use TYPSNTSM) with the CPU information from the

DTSCPU object and the logical processor information from

the DTSLOGPR object. HyperThreading is determine by:

DTCPNLPS - Number of Logical Processors Supported

DTCPNLPA - Number of Logical Processors Active.

DTCPNLPS DTCPNLPA

Hyperthreading Unsupported: missing one

Supported/Disabled: not missing one

Supported/Enabled: not missing more than 1

-The new HyperThread objects were added structurally by

NTSMF 3.4.7, so this change supports that release.
Change 22.192 All INFORMAT xxxx $NOTRAN.; statements are removed from

DOC all source code. That informat has never actually done

Aug 2, 2004 anything; it was planned to "mark" the hex-containing

character variables, so SAS could detect NOTRAN and not

translate them, but that was never implemented in SAS,

and the circumvention (to use UTILCVRT instead) has been

provided in MXG since Change 19.271.

Note that the $NOTRAN. format will always be created in

the MXG FORMATS member, to protect old MXG datasets.

The motivation for this change was the error in Syncsort

PROC SYNCSORT product under SAS Version 9.1.2, which does

not correctly handle INFORMATs, and since the $NOTRAN was

no longer needed, removing all will eliminate one more

possible error, if you have the PROC SYNCSORT product and

they have not fixed their error when you install 9.1.2.

Search NEWSLTRS for PROC SYNCSORT for details.

IMACICBB IMACICOC IMACICSA TYPEMOND TYPETMON TYPEVM

VMAC1415 VMAC21 VMAC22 VMAC28 VMAC33 VMAC36

VMAC37 VMAC38 VMAC39 VMAC41 VMAC42 VMAC60

VMAC6156 VMAC64 VMAC78 VMAC79 VMAC80 VMAC84

VMAC88 VMAC89 VMAC8911 VMAC90 VMAC90A VMAC91

VMAC92 VMACACF2 VMACACHE VMACAICO VMACAICS VMACAIM6

VMACBETA VMACBGSI VMACCIAO VMACCIMS VMACCMF VMACCMFV

VMACCOMP VMACCRAY VMACCTLG VMACDALO VMACDLMN VMACDMON

VMACEPIL VMACEPMV VMACFILA VMACFOCU VMACFTEK VMACFTP

VMACHBUF VMACHIPR VMACIAM VMACIDMJ VMACIDMS VMACILKA

VMACILKV VMACIMS1 VMACITRF VMACLMS VMACMDF VMACMIM

VMACMON8 VMACMPT VMACNAF VMACNNAV VMACNOAM VMACODS

VMACOMAU VMACOMCI VMACOMVT VMACOPC VMACPKMN VMACPRFS

VMACPROS VMACQAPM VMACRDS VMACROSC VMACRTE VMACSAMO

VMACSAR VMACSARS VMACSARX VMACSFTA VMACSMF VMACSNAP

VMACSOLV VMACSPMS VMACSTX VMACTCP VMACTIE VMACTIRS

VMACTMNT VMACTMS5 VMACTMV2 VMACTPX VMACTSOM VMACVMON

VMACVMXP VMACVSME VMACWSF VMACX37 VMACXCOM VMACXPSM

VMACXSTR VMACXXXX VMXGHSM VMXGVTOC VMXGVTOF

Subnote: in my class, discussing "sub-second" response

response time, I compare measures of EDITing under TSO

with three (okay, old!) connections: 2400 and 9600 baud

and local channel attached, and measured a max of 900

"TGET"s (inputs) per hour. This PC updated 120 members

took 20 minutes; there were at least one sequence of

SELECT, FIND, MARK, DELETE, FIND, SAVE for each member,

so, without the delay due to TSO swapping, this EDIT

session exceeded 2100 transactions per hour.

Change 22.191 Support for ASG/TMON TCE for CICS/ESA 2.3: COMPATIBLE, as

EXMONAMQ all new fields were added at the end of records, so MXG

IMACTMO2 21.21 or later will read the new records without error.

VMACTMO2 The new version added many fields now supported in MXG in

VMXGINIT the many TMON datasets, and one new dataset is created:

Aug 2, 2004 Dataset MONITASK dataset new variables:

TAWRDCT ='WAIT FOR*REDISPATCH*COUNT'

TAWRDTM ='WAIT FOR*REDISPATCH*TIME'

TAUSRWCT ='USERWAIT*TIME'

TAUSRWTM ='USERWAIT*COUNT'

Dataset MONISYST new variables:

TIMIDNSC='MIDNIGHT*NON-SEGMENTED*TASKS*COUNT'

TIUSRWCT='WAIT FOR*REDISPATCH*COUNT'

TIUSRWTM='WAIT FOR*REDISPATCH*TIME'

TIWRDCT ='WAIT FOR*REDISPATCH*COUNT'

TIWRDTM ='WAIT FOR*REDISPATCH*TIME'

Dataset MONIAWT new variables:

TAAWTWRC='WAIT FOR*REDISPATCH*COUNT'

TAAWTWRT='WAIT FOR*REDISPATCH*TIME'

Dataset MONICMX new variables:

TICMXHWM='CLASS*HIGH*WATER*MARK'

TICMX ='TICMX*SEGMENT*COUNT'

Dataset MONIIWT new variables:

TIIWTWRC='WAIT FOR*REDISPATCH*COUNT'

TIIWTWRT='WAIT FOR*REDISPATCH*TIME'

Dataset MONITJ new variables:

TJSREQCT='JVM*REQUESTS*WITH*RESET'

TJCREQCT='JVM*REQUESTS*WITH*CLASSCACHE'

Dataset MONIPA new variables:

PACBRNM ='CORBASERVER NAME'

PADSCWC ='STG CONSTRAINT WAIT COUNT'

PADSCWT ='STG CONSTRAINT WAIT TIME'

PADSMWC ='TCB MISMATCH WAIT COUNT'

PADSMWT ='TCB MISMATCH TIME'

PADSTHW ='PEAK OPEN TCBS'

PAEJBACT='BEAN ACTIVATION REQUESTS'

PAEJBCCT='BEAN CREATION REQUESTS'

PAEJBMCT='EJB METHOD CALLS'

PAEJBPCT='BEAN PASSIVATION REQUESTS'

PAEJBRCT='BEAN REMOVAL REQUESTS'

PAEJBTCT='TOTAL EJB REQUESTS'

PAHTDLYC='HOT POOL DELAYS'

PAHTDLYT='HOT POOL DELAYS TIME'

PAJ9CPUC='J9 CPU COUNT'

PAJ9CPUT='MODE J9 CPU TIME'

PAJTDLYC='JVM TCB DELAYS TIME'

PAJTDLYT='JVM TCB DELAYS'

PAKY9CPC='KEY 9 CPU COUNT'

PAKY9CPT='KEY 9 CPU TIME'

PAKY9DSC='KEY 9 DISPATCH COUNT'

PAKY9DST='KEY 9 DISPATCH TIME'

PAPTPWTC='PARTNER WAIT COUNT'

PAPTPWTT='PARTNER WAIT TIME'

PAROCPUC='RO CPU COUNT'

PAROCPUT='RO CPU TIME'

PARODSPC='RO DISPATCH COUNT'

PARODSPT='RO DISPATCH TIME'

PASOI1C ='SOCKET CHARS RECEIVED'

PASOIMC ='SOCKET RECEIVE REQUESTS'