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

(CICREQRV), or at USS (CICUSSRV). The CICINTRV dataset

replaces and enhances the old CICSYSTM dataset that has

zero observations under CICS 3.1+. The CICEODRV dataset

is brand new and provides Shutdown & EOD statistics. In

fact, CICS 3.1 no longer prints ANY of the shutdown stats

on the Job's listing. IBM provides the DFHSTUP utility to

extract Shutdown stats from SMF data. MXG has CICEODRV!

Printing the CICEODRV data set wil give your CICS folks a

report for each shutdown/end of day evolution. There are

the nineteen CICS datasets that are currently used to

create the four (INT/EOD/REQ/USS) interval datasets:

CICAUTO CICDMG CICDQG CICDQT CICDS CICDTB

CICFCT CICLDG CICM CICSDG CICSMDSA CICST

CICTC CICTCT CICTDG CICTM CICTSQ CICTST

CICVT
Note that there are a few other CICS Statistics datasets

that are not currently included in these four interval

MXG data sets. As we learn more about CICS/ESA, these

four datasets built by CICINTRV member may be enhanced.
These four CIC...RV interval datasets are automatically

created in the PDB library by BUILDPDB/BUILDPD3 if you

do nothing; member IMACCICS defaults the destination DD

name to "PDB". The sort order of these four datasets is:

BY SYSTEM APPLID SMFPSSPN COLLTIME DURATM SMFSTINO;
Change 08.250 PR/SM "overhead" is explicitly captured when APAR OY36668

VMAC7072 (and associated microcode) is installed in ES/9000 or S/J

Mar 4, 1991 models of the 3090 (this APAR is not available on earlier

hardware platforms). The APAR captures the "Effective

Dispatch" duration, LCPUEDTM, for each partition. The

difference between LCPUPDTM, the existing LPAR Dispatch

time, and this new LCPUEDTM "Effective" LPAR Dispatch, is

the time that this LPAR was dispatched for management of

this partition. In addition to capturing the partition

LPAR management time, there will be a new observation in

TYPE70PR with a partition name of PHYSICAL that will have

only a value for LCPUPDTM. This "pseudo" partition does

not actually exist, but is the capture mechanism for the

total additional time that LPAR was dispatched but could

not be charged to a specific partition. See the technical

discussion and schematic in Newsletter NINETEEN.

Note that with this APAR installed, all of the IBM RMF

reports will now use the "Effective" dispatch time as

their 100% CPU busy denominator.
Although unrelated to the primary purpose of this APAR,

two new fields (LCPUCAP and LCPUCAPC) identify if each

partition had Capping in effect, or if Capping status

changed during the interval.

Change 08.249 DB2 trace records are now "paired" (event start and end

ANALDBTR are merged) in this new analysis routine for serious DB2

Feb 28, 1991 questions. (Note that most sites with DB2 do not need the

detail of SMF 102 tracing; member ANALDB2R seems to meet

the reporting needs of the vast majority of MXG sites).

But if you need to trace SQL calls, I/Os, etc., this

routine will sort and match the pairs and builds the many

data sets from your DB2 trace. The instructions for its

use and the datasets created are documented in the member

itself.
Change 08.248 DB2ACCT variable QWHDUNIQ is now formatted $HEX12. Three

VMACDB2 variables, QLACCPUL, QLACCPUR, QLACDBAT, added for DB2

Feb 28, 1991 distributed data durations, are not in 8-byte Timer Units

used every where else in DB2, but are microseconds stored

in an 8-byte floating point number! Change lines 189800

thru 190000 to use RB8.6 instead of MSEC8.!

Thanks to Bob Otis, Rockwell International, USA.

Change 08.247 Support for Boole and Babbage's CMF Monitor's "Type 240"

ANALCM27 SMF record is added by this MAJOR user contribution! The

ANALCM29 subtypes 0, 1, 19, and 23 are processed by the standard

EXCMFASM architecture MXG members TYPECMF/VMACCMF into 16 datasets

EXCMFCPQ CMFASMQ CMFDEVIC CMFEXTRT CMFPG0V

EXCMFCPS CMFCACHE CMFDOM CMFIPS CMFPRIOS

EXCMFDEV CMFCPUQ CMFEXTCC CMFOBJ CMFSRMC

EXCMFDOM CMFCPUS CMFEXTPG CMFPG CMFTRACE

EXCMFID

EXCMFIPS In addition, two specific processing/reporting members

EXCMFOBJ for the subtype 27 (cache sampler) and subtype 29 (the

EXCMFPG CS monitor) are provided in ANALCM27 and ANALCM29.

EXCMFPG0 The actual SMF record ID is set in member IMACCMF.

EXCMFPRI

EXCMFSRM The implementation includes a CALLed function, that is

EXCMFTCC yet to be added in MXG. Also, actual testing with the

EXCMFTPG BUILDPDB has not been performed yet.

EXCMFTRA

EXCMFTRT

IMACCMF

TYPECMF

VMACCMF

Feb 21, 1991

Thanks to Richard L. Gimarc, Boole and Babbage Austin, USA.

Change 08.246 SAS 6.06 Compatibility. LENGTH DEFAULT=4 is positional.

VMACSMF Variables ID and SUBTYPE in TYPE30_V and TYPE1718 were

Feb 21, 1991 stored in 8-bytes instead of 4-bytes because they precede

the LENGTH DEFAULT=4 statement. The LENGTH statement in

VMACSMF now includes LENGTH DEFAULT=4 preceding INPUT.

Thanks to Siegfried Trantes, Gothaer Versicherungsbank VVAG, GERMANY.

Change 08.245 Support for NETMASTER V2.1.0 has been added. The type 39

EXTY39 records written by NETMASTER are generally identical to

FORMATS IBM type 39 records. Two TOD variables exist at the end

VMAC39 of the SCS secion, SMFNCSTM/SMFNCETM, but their values

Feb 20, 1991 are identical to ACTSTIME/ACTETIME in the ACCT section,

and are not kept. The subtype 255 creates a new dataset,

TYPE39FF that is unique to NETMASETR and contains:

SMFNRCFA='RESOURCE*AVAILABLE*?'

SMFNRCLN='RESOURCE*LINK NAME'

SMFNRCNI='RESOURCE*NETWORK*ID'

SMFNRCNM='RESOURCE*NAME'

SMFNRCPU='RESOURCE*PU NAME'

SMFNRCR ='CONFIG DATA REVISION LEVEL'

SMFNRCSP='RESOURCE*SUBAREA PU NAME'

SMFNRCSS='RESOURCE*OWNING/ADJACENT SSCP'

SMFNRCST='END OF*NETMASTER*INTERVAL'

SMFNRCTP='RESOURCE*TYPE'

This change has been bench validated with printed dumps

of several NETMASTER records.

Thanks to Gerry Binas, DCSD, Manila Electric Company, PHILIPPINES.

Change 08.244 VM/XA and VM/ESA MONWRITE records can cause an erroneous

VMACVMXA MXG message "Probable Data Loss Due to MONWRITE Failure".

Feb 20, 1991 MXG detection of internally spanned MONWRITE records was

imperfect. Line 005970 was IF (COL-4+MRHDRLEN) GT LENGTH

but should have been (COL-3+MRHDRLEN) and line 006000 was

SKIPTHIS=LENGTH-COL-1 but should have been LENGTH-COL+1.

Thanks to Bob Small, Grumman Data Systems, USA.

Thanks to Mike Masella, G.T.E. Services Corp, USA.

Change 08.243 The test IF X NE 0 may not produce the expected result.

SAS coding If variable X can be missing, the test wil be TRUE. This

Feb 15, 1991 suggests that a (generally) safer algorithm, which does

ensure that X is both non-zero AND non-missing, is to use

the test IF X GT 0 for this purpose. The example below

tested X (an offset to a data field) to ensure that the

FIELD was present in the record:
wrong: IF X NE 0 THEN INPUT @X FIELD $CHAR1.;

right: IF X GT 0 THEN INPUT @X FIELD $CHAR1.;

Because X was in a conditional INPUT statement that had

not been executed, the value of X was missing, but the

NE 0 test (instead of GT 0) executed the INPUT statement

with @X where X was missing! What did SAS do with missing

pointer? For either a zero or missing pointer value in

an INPUT statement, SAS moves the pointer to byte 1 of

the record and beings normal INPUT. There is no problem

with using X GE 0 if you can guarantee that X will be non

missing, and in all instances of its use in MXG it does

have adequate protection. Fields unconditionally input as

PIBn (full binary) can never be missing, but variables

input as PDn and other packed formats can be missing.

Now that I realize GT 0 is safer than NE 0, I will cease

using NE 0. I'll bet this already written up somewhere.

Change 08.242 An MRO CICS transaction in an AOR that is waiting for a

VMAC110 TOR user to type enter will find the AOR transaction's

Feb 14, 1991 IRESPTM=(ELAPSED-WTTCIOTM) will include all of the time

that the AOR was waiting on the (user) TOR, and the time

waiting is captured in the WTIRIOTM, the time that the

AOR was waiting on inter-region communications. As long

as this AOR has no FOR, then all of the WTIRIOTM can be

recognized as "user think time" and should be subtracted

from IRESPTM to capture the "real" internal response time

for this type of AOR transaction. However, if the AOR

waits for any other inter-region communications, such as

and FOR, then the WTIRIOTM would contain ambiguous delay

time. The IRESPTM=IRESPTM-WTIRIOTM; statement could be

added in your EXCICTRN exit; it cannot be added to MXG

code because it might not apply to all CICS installations

and it might not even apply to all CICS regions at an

installation.

Thanks to John Nursey, Commercial Union Assurance.

Change 08.241 Unused.
Change 08.240 Variable R79GTOD was added to all TYPE79xx datasets, as

VMAC79 that timestamp may be needed in analysis.

Feb 14, 1991

Thanks to Diane Eppestine, Southwestern Bell, USA.

Change 08.239 If CICS variables are EXCLUDEd (i.e, the CICS MCT has the

VMAC110 EXCLUDE=(DFHaaaa,DFHbbbb) or INCLUDE=(DFHaaaa,DFHbbbb),

Feb 12, 1991 MXG member VMAC110 can be edited and the excluded fields

can be commented out as described in instructions in this

member (FIND occurrence of EXCLUDE and read comments).

The comments suggested using the UTILCICS program in the

MXG library, but you can just as easily look directly at

the EXCLUDE/INCLUDE statements in your MCT to know which

fields to comment out in VMAC110. This change takes note

of that user suggestion, and protected BMSTOTCN & TDTOTCN

variables against "Un-Initialized Variable" messages.

Thanks to Joe Naussbaum, Morgan Stanley, USA.

Change 08.238 Validation of CICS 3.1 statistics records uncovered some

VMAC110 incorrect input formats. In CICFCR, variables A17OPENT

Feb 12, 1991 and A17CLOST were incorrect. Instead of PD4. format, the

four bytes must be read HH PK1. MM PK1. SS PK1. TH PD1.1

and then SS=SS+TH; and A17OPENT=HMS(HH,MM,SS); and this

is repeated for A17CLOST. In CICSLDG, variables LDRFT,

LDGLLT, and LDGTTW were incorrectly input as TU4., when

they are actually CICS internal clock values and must

be INPUT as PIB4.6 and then multiplied by 16 to convert

to seconds. See Technical Discussion in Newsletter 19

for a table of time formats and how to read them.
Change 08.237 A short SMF record can be created by STC's 4400 Silo when

VMACSTC a silo is varied on/off line. STC ZAP L1H00o8 (yes, that

Feb 12, 1991 is two zeros and an alphabetic o) corrects the bad record

that caused MXG to ABEND (INPUT STATEMENT EXCEEDED ...).

The following circumvention fix has been added to MXG so

that it will not fail even with the short record. Lines

022700 thru 023000 now read:

INPUT @21+OFFSMF SMLSFLAG $CHAR1. @;

IF LENGTH GT 26 THEN

INPUT +3 SMLSATID PIB2. @;

Thanks to Raymond Zieverink, Belk Stores Services, USA.
Change 08.236 Support for Legent's ASTEX ("Automated Storage Expert"

VMACDMON replacement for DASDMON SMF records. A number of new data

Feb 6, 1991 variables have been added to all three datasets. Some of

the variables that were common in all records now exist

only in the DMONDSN and DMONJOB datasets under ASX, but

since MXG now supports either old DASDMON or new ASX SMF

records, these variables are still kept in both places

for compatibility; the ones moved will exist but have

missing values with ASTEX. The following are changed:
New in all three (DMONVOL,DMONDSN,DMONJOB) datasets:

ASXFLAG ='SPANNING FLAG'

ASXFLAG2='SYSTEM*TYPE*FLAG'

ASXFLAG3='AUTHORIZED*FLAG*DASD'

RALBSY -'TOTAL*CON/DISC*OUTBOARD TIME'

RBYCNT ='TOTAL IO-S FOR BYPASSES'

RCCCNT ='TOTAL IO-S FOR CACHE CANDIDATES'

RCFCNT ='TOTAL IO-S FOR CFW'

RDFCNT ='TOTAL IO-S FOR DFW'

RDLCNT ='TOTAL*DFW*LOAD COUNT'

RHTDFW ='TOTAL DASD FAST WRITE HITS'

RHTRD ='TOTAL*READ*HITS'

RHTRDS ='TOTAL READ SEQUENTIAL HITS'

RLDCNT ='TOTAL*LOAD*COUNT'

RNRCYLS ='NR OF CYLINDERS ON THE DEVICE'

RNWCNT ='IO-S THAT WERE NORM WRITES'

ROICNT ='OPTIMIZER INHIBIT IO-S'

RSTCFW ='TOTAL CACHE FAST WRITE HITS'

RTICNT ='TOTAL IOS INHIBIT IO-S'

RXSCNT ='CROSS*SYSTEM*SEEK COUNT'

New in DMONVOL dataset:

RVCACHFL='CACHE*FLAG*HVLFLAG4'

RVCFLAG2='CACHE*FLAG2 '

RVDMTOBJ='DEMOTION*TIME (SECS)*OBJECTIVE '

RVDSNRBA='RBA OF DSN RECORDS'

RVDSNBKS='NUMBER PHYS. BLOCKS FOR DSN RECS'

RVDSNOFF='OFFSET INTO RBA TO 1ST DSN REC'

RVJOBBKS='NUMBER OF PHYS. BLOCKS FOR JOB RECS'

RVJOBOFF='OFFSET INTO BLOCK TO 1ST JOB REC'

RVJOBRBA='RBA OF JOB RECORDS'

RVMXNBR ='DSNRBA'

RVSIIX ='VSAM IMBEDDED*INDEX I-O*COUNT'

New in DMONDSN and DMONJOB dataset:

RPRGMNM ='PROGRAM NAME*ASSOCIATED*WITH DSN'

RTRATME ='TOTAL*INTRA-DSN*SEEK TIME '
Variables removed from all three DMON.... datasets.

RCHCNT ='TOTAL*CACHE*HITS'

RSQRSP ='SUM*RESP*SQUARED'
Variables removed from DMONVOL dataset

RVMXNBR ='LONGEST*NON-BUSY*RESERVE'

Now missing values (ASX only) in DMONVOL dataset:

RFTCNT ='TOTAL*FETCH*IO-S'

RLRCNT ='TOTAL*LONG*RESERVES'

RSEEKTM ='TOTAL*SEEK*TIME'

RSRCNT ='TOTAL*BLDL-FIND*IO-S'

Thanks to Joseph J. Faska, Depository Trust, USA.

Change 08.235 The support to read HSM data sets needed correction for

VMACHSM HSM 2.4 and 2.5. Two occurrences of 177 for offset P2

Feb 6, 1991 changed to 185, and the two occurrences of 297 for P3 are

changes to 405.

Thanks to Milt Weinberger, Information Services International, USA.
Change 08.234 The Hiperbatch counters for VSAM added to the type 64 SMF

VMAC64 record were wrong, because MXG failed to skip over the 3

Feb 5, 1991 undocumented bytes in the record. Change the test for 193

to 196, and insert +3 between INPUT and SMF64SIO.

Thanks to Bob Brosnan, Goldman Sachs, USA.
Change 08.234A Preliminary support for VPD, "Vital Products Data" aka

ADOCVPD NAM, "Network Asset Management" NETVIEW "hardware monitor

EXTYVPDE records" written to SMF (usually) as type 37 containing

EXTYVPDF "VPD" as SUBSYS and SUBTYPE=22. Member IMACVPD sets the

EXTYVPDM record ID MXG expects, and defaults to 37. (If VP writes

EXTYVPDP the record, I would not be surprised to see the value 0!)

EXTYVPDS Seven internal record subtypes, E,F,M,P,S,T and W create

EXTYVPDT seven VPD..... datasets documented in DOCVPD. An extra

EXTYVPDW part of this user contribution is the INFOLOAD member,

IMACVPD which shows how an MXG SAS data set (in this example, the

LOADINFO PDB.VPDPUHAR) can be "loaded" into an INFO/SYS database!

TYPEVPD All seven subtypes have been coded and syntax checked,

VMACVPD but only "P", VPDPUHAR PU Hardware subrecords have been

VMAC37 tested with data records. Member VMAC37 was updated to

Feb 4, 1991 NOT process type 37 records with SUBSYS='VPD'.

Thanks to Bill Border, Apollo Computer Inc, USA.

Change 08.233 "Invalid data for MSOCOEFF in line ..." error and a hex

VMAC7072 dump is printed with MVS/ESA 4.1 type 72 records, but the

XMAC7072 data sets built by MXG are NOT in error. IBM increased

Feb 1, 1991 the width of the MSOCOEFF field from 4 to 8 bytes in

MVS/ESA 4.1 (because sites wanted to set a small value

and the 4-byte EBCDIC value was limited to .001 for the

memory service coefficient) by adding a new 8-byte field

to the end of the record, and putting a zero in the old

4-byte field. Unfortunately, I expected an EBCDIC zero,

but IBM changed the format to binary zeros, causing the

invalid data message. Since MXG went on to then input a

value from the new 8-byte field, MSOCOEFF was correct in

the TYPE72 data set, in spite of the note on the log!

The correction is simple. Change the line now reading:

@OFFWRKC+34 MSOCOEFF 4.

to read:

@OFFWRKC+34 MSOCOEFF ?? 4.

(the ?? added suppresses the error message and the dump

and sets the variable to missing value on invalid data).

Thanks to Ann Slaughter, Airline Tariff, USA.

Change 08.232 The CICS Dictionary utility had misleading titles and was

UTILCICS accidentally altered when CICS 3.1 support was added. The

Feb 1, 1991 report was valid, however.

Thanks to Phil Shelley, Jewish Hospital HealthCare Services, USA.

Change 08.231 VM/ESA testing found an unprotected divide by zero that

VMACVMXA is now corrected in de-accumulation of the user records.

Feb 1, 1991 (It had no impact on the MXG-build data sets, because the

observation causing zero-divide was the first for a user,

and the first record is always deleted). Additionally,

variable VMDELIST is now formatted $HEX2.

Change 08.230 Duplicate Job Name Hold delay time for batch scheduling

ASUMJOBS is now detected and eliminated in this summarization part

Jan 31, 1991 of the MXG Trend Data base that builds the PDB.JOBSKED

data set and calculates IWT (Initiation Wait Time) by

job class. See the MVS Technical Note in Newsletter 19.

Thanks to Freddie Arie, Lone Star Gas, TEXAS.

Change 08.229 SAS 6.06 Compatibility? SAS 5.18 handled "PIB2 ." in

VMXGHSM line 370300, but SAS 6.06 requires the "PIB2." syntax

Jan 31, 1991 (i.e., no blank between PIB2 and its period)!

Thanks to Andrew Davies, TAB of W.A., AUSTRALIA.

Change 08.228 Support for SESAME, a Volvo (Europe) VTAM Monitor, which

EXTYSESA creates an SMF record. This is preliminary support, that

IMACSESA has only been syntax checked after editing this user

TYPESESA contribution to MXG Software. Formats are defined in

VMACSESA FORMATS but are not yet associated with their variables.

Jan 30, 1991

Thanks to Olli Rautajuuri, SAS Institute Oy, FINLAND.
Change 08.227 Support for MEMO, a European Electronic Mailsystem which

EXTYMEMO creates an SMF "accounting" record. This is preliminary

IMACMEMO support, that has only been syntax checked after editing

TYPEMEMO this user contribution to MXG Software. Formats are

VMACMEMO defined in FORMATS but are not yet associated with their

Jan 30, 1991 variables.

Thanks to Pertti Russcanen, Carelcomp Oy, FINLAND.
Change 08.226 Variable PMHTSKID should not have been KEEPed in these

VMACIDMS non-task-related MXG datasets: IDMSARA,BUF,CPM,INS,INT,

Jan 29, 1991 JRL,LNE,PGM,RUS,STG,YPE. Only caused confusion.

Thanks to Bob Rutledge, Sherwin Williams Paint, USA.

Change 08.225 Support for Landmark's The Monitor for DB2 type 100 & 101

TYPEDB2 records requires a fix from Landmark. ZAP U900266 exists

Jan 29, 1991 now, and that fix will be included in Landmark's planned

maintenance release level ML911, due out this Spring. The

DB2ACCT data is good without that fix, but DB2STAT0/1 are

wrong until the fix is installed. (The statistics record

sequence number was not updated by Landmark, and MXG uses

it to validate the de-accumulation of interval records,

and to detect skipped intervals and multiple startups).

Landmark does not yet write type 102 SMF records.

Change 08.224 Support for MVS/ESA 4.2 (RMF 4.2.1).

EXTY22IO 1.New dataset TYPE22IO reports I/O configuration changes as

EXTY33U1 a result of the ACTIVATE funcation (add, delete, modify)

IMACPDB with these 33 new variables from the new subtype of the

TYPE33 type 22 SMF record:

VMAC6 FORCEOPT='ACTIVATE*FORCE*OPTION?'

VMAC22 SMF22CCH='CHANNEL*PATH*ID'

VMAC26J2 SMF22CFI='OPSYS*CONFIGURATION*ID'

VMAC26J3 SMF22DCH='ARRAY OF*CHPIDS*ADDED OR DELETED'

VMAC30 SMF22DCM='MASK OF*CHPIDS*ADDED OR DELETED'

VMAC32 SMF22DPC='ARRAY OF*PHYS CU-S*ADDED OR DELETED'

VMAC33 SMF22DPM='MASK OF*PHYSICAL CU-S'ADDED-DELETED'

VMAC4345 SMF22DVN='DEVICE*NUMBER'

VMAC7072 SMF22EDT='ID OF*NEW*EDT'

VMAC71 SMF22EFL='ENTRY*FLAGS'

VMAC73 SMF22EHD='ENTRY*HEADER'

VMAC74 SMF22ELN='LENGTH OF EACH ENTRY'

VMAC78 SMF22ER ='ENTRY*TYPE*(MXG FORMATTED)'

Dec 6, 1990 SMF22ERQ='ENTRY*REQUEST'

Jan 28, 1991 SMF22ETY='ENTRY*TYPE'

SMF22FLS='FLAGS'

SMF22FNC='ACTIVATE*FUNCTION*REQUESTED'

SMF22HCT='HARDWARE*CONFIGURATION*TOKEN'

SMF22IDN='DSNAME OF IODF*WITH NEW*I/O CONFIGURA'

SMF22NE ='NUMBER OF ENTRIES IN THIS RECORD'

SMF22NUA='NUMBER OF*UCBS ADDED*FOR THIS CHANGE'

SMF22NUD='NUMBER OF*UCBS DELETED*FOR THIS CHANGE'

SMF22OFF='OFFSET TO FIRST ENTRY'

SMF22PCH='ARRAY OF*CHPIDS*ADDED OR DELETED'

SMF22PCM='MASK OF*CHANNEL PATH IDS*ADD-DELETED'

SMF22PCN='PHYSICAL*CONTROL UNIT*NUMBER'

SMF22PSU='STARTING*UNIT*ADDRESS'

SMF22PUA='RANGE OF UNIT ADDRESSES*ADD-DELETED'

SMF22PUC='COUNT*OF UNIT*ADDRESSES'

SMF22RNR='RECORD NUMBER OF THIS RECORD'

SMF22TNE='TOTAL NUMBER OF ENTRIES FOR THIS CHANGE'

SMF22TR ='TOTAL RECORD NUMBERS FOR THIS CHANGE'

SOFTOPT ='ACTIVATE*SOFT*OPTION?'

2.New variables have been added to the TYPE30_V, TYPE30_4,

TYPE30_5, and TYPE30_6 datasets built from type 30 SMF

records, and to datasets PDB.SMFINTRV, PDB.STEPS and

PDB.JOBS built by BUILDPDB/BUILDPD3. The BUILDPDB change

required modification to member IMACPDB, so that MXG

users with a modified IMACPDB will need to retrofit its

changes onto the new IMACPDB member.
a. Ten new variables describe pages and blocks moved to

and from Auxiliary and Expanded storage for this job

due to the new Blocked Paging enhancements. Variables

in upper case are actual data fields, while lower case

variables are calculated from actual data fields:
Blocks Blocked Unblocked Total

Pages Pages Pages

In from AUX BLKSAUIN PGBKAUIN

In from ESTORE BLKSESIN PGBKESIN PGUNESIN PGTOESIN

In total BLKSTOIN PGBKTOIN

Out to AUX BLKSAUOU PGBKAUOU

Out to ESTORE BLKSESOU PGBKESOU PGUNESOU

Out total BLKSTOOU PGBKTOOU

Type 71 variables: Blocks Blocked Unblocked Total

Pages Pages Pages

In total BLKSTOIN PGBKTOIN
Type 72 variables: Blocks Blocked Unblocked Total

Pages Pages Pages

In from AUX BLKSAUIN pgbkauin

In from ESTORE BLKSESIN PGBKESIN pgunesin PGTOESIN

In total blkstoin PGBKTOIN
c. Eight new variables describing APPC activity by this

address space were added (but the APPC resources are

not currently contained in the subtype 2 & 3 interval

records, so these variables are not kept in TYPE30_V