* copyright (C) 1984-2019 merrill consultants dallas texas usa
Yüklə 28,67 Mb.
|
|
Nov 15, 2009 -READDB2 corrected logic that was suppressing the T102S106 Nov 24, 2009 dataset when WANTONLY was not used, even when requested. -READDB2 ACCOUNT and STATISTICS now work as expected, to create only those subsets of datasets; examples revised to show the use of optional DB2ACTB=NO, etc. Thanks to Jim Kovarik, AEGONUSA, USA. Change 27.296 Variable CECSER added to datasets RMFMSUSE and SMFMSUSE. ASUMMIPS
Oct 28, 2009 Thanks to Nicholas Ward, Centrelink, AUSTRALIA Change 27.295 PRO/SMS SMF records were misaligned after the MSGED text VMACPROS which was increased from 256 to 292 bytes; the first 42 Oct 27, 2009 are now GWMSGECH $CHAR42 with $HEX84 format, while the actual text is GWMSGED input now as $EBCDIC250. Thanks to Scott Barry, SBBWorks Inc, USA.
VMXG70PR and DURATM were missing values, and IFLCPUS was wrong if Oct 26, 2009 the summary interval was larger than original DURATM. Thanks to Kenneth D. Jones, Bell Aliant, CANADA. Thanks to Bruce Perry, Bell Aliant, CANADA.
EXCICIDD CICS/TS 4.1 & APARs PK95579,PK83741,PK98426, creates new EXCICIDN SMF 110 Subtype 1 MNSEGCL=6 record (if MNIDN=ON in SIT), FORMATS from which MXG creates two new datasets: IMAC110 dddddd Dataset Description VMAC110 CICIDN CICSIDNT IDENTITY Transaction Information VMXGINIT CICIDD CICSIDND Identity Realm/Distinguished Name Oct 26, 2009 This new Identity Propagation facility sends a user's Nov 16, 2009 security identity information (the distributed identity) from a client system across a network, preserving the distributed identity for use during CICS authorization and for subsequent auditing purposes. Following text was revised Feb 1, 2010: -The new MNSEGCL=5 record was "supported" by MXG 20.20 but no records were ever read; this SMF file contained those records and exposed errors in that MXG code, including a possible INPUT STATEMENT EXCEEDED RECORD LENGTH error. -The MNSEGCL=5 subtype record can be compressed by CICS, but the new MNSEGCL=6 record can not.
VMXG70PR TOTSHARC and TOTSHARE in the observations for the first Oct 26, 2009 LPAR (only in the first interval). Setting LPARNUM=. was changed to LPARNUM=9999 so the descending sort corrected the missing values. Thanks to Horst Noerenberg, GaVI Gesellshaft .. Informatik, GERMANY. Change 27.291 Variable S42DSRDD was incorrect by a factor of 10**6; the VMAC42 calculation should have multiplied by 1000 to convert to Oct 23, 2009 milliseconds, but it instead divided by 1000. The right statement for all three instances now is S42DSRDD=128*S42DSRDD*1000; Thanks to Jim Horne, Lowe's Companies, USA. Thanks to Sridher Nelliyappan Manivel, Lowe's Companies, USA.
VMXGUOWT had not been updated in four years, suggesting not much Oct 23, 2009 usage, but now it is updated to match ASUMUOW. Thanks to Chris Weston, SAS ITRM Development, USA. Change 27.289 Support for CPUZIPTM in SYNCSORT records; additionally, VMACSYNC the new 4-byte fields in the SORTWORK extension are used Oct 22, 2009 for the primary and secondary track allocations and the Nov 18, 2009 primary tracks released. Syncsort acknowledges that not all sorts can use a zIIP: -Eligible sorts (not Merge or copies) are seeing 50 to 60 percent use of the zIIP eligible algorithms which result in a 4 to 10 percent improvement in CPU time. -Sorting jobs which employ the following features are currently not eligible for zIIP processing: SUM Exit routines other than E15 or E35 Using SORTIN with an E15 exit OUTREC with CONVERT Relative length greater than 32756 VBS truncation TSO invoked sorts SVC=0
VB INCLUDE/OMIT OUTFIL INCLUDE TRUNCATION OF KEYS SYNCSORT TPF SY66820 enables use of: INCLUDE/OMIT with STOPAFT -Although these limitations are being reviewed by SYNCSORT development staff, there is no schedule nor time frame if or when any of these will become eligible for zIIP processing. Thanks to Jerry Urbaniak, Acxiom, USA. Change 27.288 New format $MGCICCN is created for variable CMQCNSTA in FORMATS CICS Statistics dataset CICIMQ dataset to identify if the VMAC110 Queue is connected or not. Oct 22, 2009 Thanks to Dale Slaughter, Aegonusa, USA.
DAILYDSC -ERROR: New DAILYDSC, DAILYDSN, and DAILYDSR. DAILYDSN The new VMXGDSN's argument was changed to TAPEDATA= from DAILYDSR TMC= during testing, but the three DAILYDSx members still VMAC74 had the original TMC= argument in their VMXGDSN calls. CONFIGV9 If you have an old, tailored DAILYDSx member, it is NOT Oct 21, 2009 impacted by this new design that uses VMXGDSN. VMXGDSN -COSMETIC: VMAC74. Nov 13, 2009 DEBUGGING PUT statement was not removed. -COSMETIC: CONFIGV9 Option NOMAUTOSOURCEDISPLAY replaces MAUTOSOURCEDISPLAY. Many lines printed on log, no impact. See Change 27.284. Option was changed in 27.08. -Nov 13: VMXGDSN typo PDN corrected to PDB. Thanks to Paul Naddeo, FISERV, USA. Thanks to Robert Hamilton, Fifth Third Bank, USA. Thanks to Rachel Holt, Fidelity Systems, USA.
VMACPRPR as OFFSETS, and new variable INTERPOS is INPUT and KEPT. Oct 18, 2009 Thanks to Nik Marien, KBC Global Services NV, BELGIUM. Change 27.285 Label, formats, and spelling corrections: TYPEIMSA TYPEIMSA: VMACNTSM variable ELAPSTM formatted TIME12.2 VMACQACS VMACNTSM: VMACVMXA labels added Oct 17, 2009 MLMDINST='MLMD*INSTANCE*NAME' VMAC71 MLLOINST='MLLO*INSTANCE*NAME' VMACTMNT QLBLINST='QLBL*INSTANCE*NAME' Oct 20, 2009 labels changed SQBTATWB='AVG*TIME TO*WRITE*BATCH' QLBLATWB='AVG*TIME TO*WRITE*BATCH' QLBLTBBA='AVG*TIME*BETWEEN*BATCHES' SQBTTBBA='AVG*TIME*BETWEEN*BATCHES' formatted QLBLTBBA QLBLATWB TIME13.3 by list changes VMWGURES dataset, VWGRINST inserted VMWHODIS dataset, VMWHINST inserted MLCONS dataset, MLCOINST inserted VMACVMXA: label corrected VIMXCD20='DEVTYPE 20*MAXIMUM*VIRTUAL*CPU TIME*ACB' variable name spelling corrections in DIF() logic: ASCHLRC ASCHLLC spelling was wrong: ASCCHLRC/ASCCHLLC VMACQACS: labels added SYSIUL='USER*AUTHORIZATIONS*AVAILABLE' SYSCIU='USER*AUTHORIZATIONS*NEEDED' VMAC71: Dataset TYPE71. Variable SMF71GIE was added to the keep list for TYPE71 by Change 27.178 but the change was not documented. Now it is labeled and formatted. VMACTMNT: Dataset TYPESYMT. Variables SYSLENCR, SYSLBLKS labeled. VMXGRMFI: Dataset RMFINTRV. Variable NRINTRV now labeled. Variables IFATM, ZIPTM are now TIME12.2 formatted. VMXG70PR: Variables IFAACTTM, ZIPACTTM labeled and TIME12.2'd. All LnnPAT variables are formatted TIME12.2 All LnnIFKTM and LPnILWTM in ASUM70PR TIME12.2'd. VMAC30: Variable ASID label corrected. Thanks to Chris Weston, SAS ITRM Development, USA.
CONFIGxx member in MXG 27.08, but not documented. That option Oct 16, 2009 prints a message when an autocall %MACRO is invoked, to identifies the source member that was autocalled. That option is very useful for MXG Technical Support: having a back-level member in your tailoring library causes most errors, and so we can diagnose directly from your log, without requesting a second run to enable diagnostics. However, the message is NOT printed for each compile of an autocalled %MACRO, but instead is printed for each invocation of an autocall, which causes lots of messages, so I have now changed CONFIGV9 to NOT set that option. You can reset the option with an OPTIONS= parameter: // EXEC MXGSAS,OPTIONS='NOMAUTOLOCDISPLAY' ====== Changes thru 27.283 were in MXG 27.09 dated Oct 14, 2009======== Change 27.283 CPU Activity Report updates. ANALRMFR
Oct 14, 2009 Thanks to Kim Westcott, New York State Office of Technology, USA. Change 27.282 -TMON for VTAM 'SE' record average values are no longer VMACTMVT contained in the record, causing an INPUT EXCEEDED error. Oct 12, 2009 These variables are now calculated rather than INPUT: SEHRTA SENRTA SEINA SEOTA. -Some 'SE' records are compressed, while most are not, in the test file, so MXG now skips the compressed records, printing a message for the first five, but outputting the non compressed records. Thanks to Paul Volpi, UHC, USA.
EXXAMAPM and STOASI. Those segments can be repeated in XAMSYS EXXAMASI records, and thus must be output in separate datasets. EXXAMIOP -New XMMTRPAG dataset created from MTRPAG segment in the EXXAMPAG XAMDEV file, which is quite different than the MTRPAG EXXAMVSW segment in the XAMSYS file. Variables PGS1-PGS3,CALCYLPA, IMACXAM CALCYLPR and CALCYLSP that were previously kept in XAMDEV VMACXAM are no longer kept in that dataset as they exist only in VMXGINIT the MTRPAG segment in XAMSYS. Oct 13, 2009 -IMACXAM is updated with the new datasets. EXXAMSEK -MANY new variables in release 3.7 that were previously Oct 17, 2009 overlooked are now output; the KEEP= list for each of Oct 20, 2009 those datasets lists this Change Number before the list of new variables. -Additional MXG corrections for alignment corrected these dataset's variables: XAMSYS SYSBUSY XAMUSR IDLETIME; variable VMTRUSR was removed. XMFALACB INDEX -Variable LASTCHNG (a TODSTAMP) is sometimes populated with 8-bytes of blanks, which cause a datetime value of '27OCT1935:08:26:40.59'. Now, blanks set it to missing. -Additional problems with data values that are observed: -Variable SMOOTHTM in XMSUBNET, XMTCPAPP, XMTCPCON has many (not all) instances of negative values. -Variables PVMAJTFA in datasets XMVSISFT and XMVSINAP and variable PVMINTFA in dataset XMVSISFT are sometimes negative values. -Variable RSASXRPM in XAMSYS dataset can be negative. -Variable PTYPE in XMVSISFT is numeric, usually 0 or 4, but it contains blanks; four blank ('40'x) input as PIB4 is the decimal value 1077952576. -Variable PPATH in XMVSISFT should contain path name in EBCDIC, but values starting with 'init ' have hex value 'BAF3BB'x in the last three bytes. -Variable PERFNICE in XMVSISFT can be negative. -Oct 17 updates after Barton updated his PL/1 Declares: -XAMDEV dataset variables PAGUSED SCMWORK SPLUSED no longer exist. variables CALSPOOL CALPAGE RDEVDRAN are added. MDISK segment, length 92, 8+4+16+16+8+16=68 documented CONFIG - corrected USEACT - corrected -XAMSYS dataset RLEASE - corrected SUBSUM - corrected MONITR - corrected MTRMEM - corrected MTRSYS - corrected -XAM TCP record FALCPU - updated. TCP - updated. HSTSFT - updated. VSINAP - updated. VSIUSR - updated. -Oct 20 status: updates pending documentation DEV - IODQDS - Segment not documented. DEV - MDISK - Extra data not documented. DEV - SEKSEK - Decoded, but disagrees with Declare. USR - SERVRS - Segment not documented. Thanks to Chris Morgan, CITIBANK, ENGLAND.
VMACDCOL DDCSPECC='DDCSPECC*ADDITIONAL*SPECIFICATION*FLAGS' Oct 9, 2009 DDCFATTR='EATTR*SPECIFIED?' DDCSPECD='DDCSPECD*ADDITIONAL*SPECIFICATION*FLAGS' DDCVBYT1='VSAM*EXTENSION' DDCEX255='OVER 255*EXTENTS*ALLOWED?' DDCEATTR='EXTENDED*ATTRIBUTE' Variable DDCSFLG was incorrectly located in z/OS 1.10 IBM documentation in Access Method Services for Catalogs; the +3 preceding it was really only +2. These new variables are decoded from DDCSFLG: DDCOVRD ='BWO*SPECIFIED?' DDCSDB ='SPHERE*RECOVERABILITY*SPECIFIED?' Thanks to Mike Blocker, Fidelity Investments, USA. Change 27.279 IBM has documented when Subtype 24 and 25 records are VMAC42 created in APAR OA30503: Oct 8, 2009 Only those applications which issue STOW and DESERV calls for PDS or PDSE directory processing will generate Subtype25 records. Some applications, such as IEBCOPY, do not issue a STOW or DESERV issue STOW when updating the directory.
CONFIGxx Change 27.124, but that error can also occur if you have Oct 8, 2009 changed the S2=0 option in MXG's CONFIGV9 to S2=72.
VMAC42 used in many places, where &MXGWORK macro variable should VMAC80A have been the text. Oct 8, 2009 -Variables W2RCHGAG and W2RINSAG in dataset CICW2R are now Oct 9, 2009 formatted MG110AG. -Variable SMFA2GZ2 label corrected to "SMS" versus "SMF" in dataset TYPE42D3. -Variable SMFA2NRS label in TYPE42X3/X4 to match SMF42NRS. -Variables TOKCHARV, TOKLDAP, TOKNUMRV, TOK80FLG, TOK80LN2 in dataset TYPE80TK are now labeled. Thanks to Chris Weston, SAS ITRM Development, USA.
VMACVMXA more than 65K tracks; four larger fields are added at the Oct 7, 2009 end of the SEKSEK record, and are input by this change into the four existing MXG variables. Change 27.275 Variable PCTIOPBY='IOP/SAP*PERCENT*BUSY' is created and VMAC78 kept in TYPE78IO. Oct 6, 2009 Thanks to Karl Lasecki, Chemical Abstracts Service, USA. Change 27.274 -ERROR: CICDS DISPATCH INTERVAL LARGER THAN INTERVAL can VMXGCICI occur when creating PDB.CICINTRV Statistics Summary if VMAC110 a region was Shutdown (creating an EOD record) and then Oct 7, 2009 restarted, and the first INT record is created such that Oct 21, 2009 both the EOD and INT records have the same Summary (recalculated) COLLTIME. For example, with an requested INTERVAL=THREEHR summary: INTERVAL START ACTUAL SUMMARY EVENT CICSSTCK COLLTIME COLLTIME EOD 00:00 01:26 03:00 INT 02:00 03:00 03:00 the two events with the same Summary COLLTIME were NOT combined, because variables CICSSTCK and COLLTIME were both in the BY list for summarization. CICSSTCK has been removed from both BY and KEEPs; only COLLTIME is correct. -Variables A17DSIXP/A17DTRDS were incorrectly deaccumed, causing negative values for them in PDB.CICINTRV. -The SORT order for CICFCR (A17 variables) was corrected to insert A17FNAM ahead of A17DSNAM in both VMAC110 and in VMXGCICI. While I hate changing the SORT ORDER, this change is required for duplicate removal. -Oct 21: CICSSTCK added in KEEPIN= for all steps that have CICSSTCH in the BY statement in their INCODE=, so that KEEPALL=NO can be used if needed. Thanks to Lynn Hong, UCLA, USA.
BLDSMPDB using AUTOALOC (which create output directory names that DAILYDSC contain todays date), it ignores ROLLWEEK (copying of the DAILYDSN week4-week5 etc) since AUTALOC takes care of that. DAILYDSR -New REROUTE= argument let's you change the destination IMACZDAT directories of individual PDB datasets. VMXGALOC -New macro variable &MXGZDATE created in IMACZDAT so that VMXGDSN ZDATE can be set for rerun without hard EDIT of IMACZDAT. VMXGDSN -New macro VMXGDSN reads DCOLLECT, TMC, RMM and CONTROLT VMXGINIT data files for "Daily Dataset Billing", but separates Oct 14, 2009 reading of the raw data files from the creation of the Oct 21, 2009 "DATASETS" dataset, so those phases can be separately Nov 19, 2009 executed by BLDSMPDB. Members DAILYDSN, DAILYDSC, DAILYDSR members now invoke VMXGDSN. -Some confusing WTD/WEEK & MTD/MONTH messages eliminated. Thanks to Cletus McGee, Alfa Mutual Insurance Company, USA. Change 27.272 READDB2 support for WANTONLY is enhanced (i.e., now it READDB2 works, and with new options!). Sep 28, 2009 When WANTONLY is used, the datasets that are NOT listed in the WANTONLY= argument that would be created by other arguments (e.g., ACCOUNT) will be created with zero obs. Example 1: IFCIDS=ACCOUNT,WANTONLY=DB2ACCT IFCIDS=ACCOUNT,DB2ACCT=YES//keep/x y z Only the DB2ACCT dataset will be populated; all other DB2ACCTx datasets will be created, but with zero obs, and no other datasets are created. Example 2: %READDB2(PDBOUT=PDBOUT,DB2ACTB=NO); If you do NOT want a dataset to be created, then you can use NO as the value of the Dataset Name Argument and the dataset will NOT exist. For example, will suppress the creation of DB2ACCTB. Thanks to Alyona Bertneski, JPMorgan, USA. Change 27.271 DB2 IFCID=173 support is updated to detect "truncated" VMAC102 names, and supports UNICODE text data. Sep 28, 2009 Thanks to Ervin L. Claxon, IBM Global Services, USA. Change 27.270 CICS Statistics STID=115 record is not supposed to exist, VMAC110 so when it did, MXG didn't handle it's non-non-existence Sep 28, 2009 correctly. Thanks to Leendert Keesmaat, UBS, SWITZERLAND. Change 27.269 Change 27.158 still caused QWHDRQNM and QWHDSVNM to be VMACDB2H incorrect, if SMF data was read on ASCII platforms, due Sep 25, 2009 to spelling QWHDxxxx as QWHSxxxx. Thanks to Bill McDonald, Kimberly-Clark, USA. Change 27.268 Example to summarize PDB.TYPE72GO Service/Report dataset, ASUM72GO creating 10 minute output intervals in PDB.ASUM72GO from Sep 25, 2009 5 minute input intervals. Using DURATM=INTERVAL forces the output DURATM value to be the INTERVAL=TENMIN value, even if the Service Class was only active during one of the two 5-minute original intervals. -Originally, the DURATM from the first "period" was kept in this INCODE logic: INCODE = PROC SORT DATA=MXGSUM1 OUT=MXGSUM1A; BY SYSTEM STARTIME SRVCLASS RPRTCLAS; DATA MXGSUM1; SET MXGSUM1A; BY SYSTEM STARTIME SRVCLASS RPRTCLAS; IF NOT FIRST.RPRTCLAS THEN DURATM=.; , but that not only cost the sort and extra data pass, it also removed the /VIEW. And, perhaps worst, the value of DURATM in the output dataset was not always the 10 min desired - if a Service Class was only used for one 5 min period that interval had DURATM of 5 minutes with INCODE. -Originally, originally, the INCODE kept PERIOD=1 obs, but there are PDB.TYPE72GO observations with only PERIOD=2; MXG only creates observations if the service class was in active use during an interval. Thanks to Scott Weiner, Wisconsin Physicians Service, USA. Change 27.267 -These variables in TYPE42X1 and TYPE42X2 are now MGBYTES. VMAC42 formatted: Sep 23, 2009 SMF42JON SMF42JOO SMF42JOP SMF42JOR SMF42JOS SMF42JOT SMF42JOV SMF42JOW SMF42JOX SMF42JOY SMF42JOZ SMF42JQG SMF42JQH SMF42JQI SMF42JQJ SMF42JQK SMF42JQL SMF42JQM SMF42JQN SMF42JRA SMF42JRB SMF42NSZ SMF42SAP SMF42SPR Dataset Labels for TYPE42Xn now match Change 27.062. Thanks to Karl Lasecki, Chemical Abstracts Service, USA. Change 27.266 VMXGALOC did not create a new weekly directory when it VMXGALOC should have, and it did not create the WTD (week to date) Sep 24, 2009 folder at the beginning of the week, nor did it create Oct 2, 2009 the MTD (month to date) folder at the beginning of the month, but now it does. Thanks to Stan Dylnicki, Royal Bank of Canada, CANDADA Change 27.265 -ASMRMFV incorrectly issued RMFV005E ERROR INVALID PARM=0 ASMRMFV when sequence numbers were in columns 73-90 of SYSIN. Sep 22, 2009 -Notes that PARM values are needed only once were added. Thanks to Doug Medland, IBM Global Services, USA. Thanks to Jerry Urbaniak, Acxiom, USA.
VMACVMXA data, and to create only PDB.VXBYUSR and PDB.VXSUMUSR VMXGINIT output datasets required updates to VMACVMXA because the Sep 21, 2009 VXSUMUSR's output MACRO _SVMSUMU wasn't undefined. -This example can then be used to read only Domain 4 and create only PDB.VXBYUSR and PDB.VXSUMUSR in //PDB: //MWINPUT DD DSN=YOUR.MONWRITE.DATA.FILE,DISP=SHR //PDB DD DSN=YOUR.ZVM.OUTPUT.PDB,DISP=OLD //SYSIN DD * %VMXGINIT(DEFAULT=WORK); %LET PVMBYUS=PDB; %LET PVMSUMU=PDB; %LET MACKEEP= MACRO _VMRPT % ; %LET MACVMXH= %QUOTE( IF MRHDRDM NE 4 THEN DO; INPUT +SKIP @; SKIP=0;
MRHDRDM=-99; MRHDRRC=-99; GOTO MWENDIT; END;
); %INCLUDE SOURCLIB(VMACVMXA,IMACKEEP); _TESTMW; RUN;
Thanks to Doug Wells, First National Bank of Omaha, USA. Change 27.263 Variable R744FNAM, the Coupling Facility Name, was not VMAC74 kept in TYPE74DU (Duplex Coupling Facility). Any ITRM DOC sites using dataset XTY74DU created by MXG macros need to Sep 21, 2009 update their ITRM dictionary. Fortunately, MXG's R744FNAM already exists as ITRM name R744FNA (in XTY74CF,XTY74ST), and it will be added in ITRM's next dictionary update, but it can easily be added (with help from ITRM techie!):
(the data dictionary utility) to add the variable to the table. The example uses the CPCAT macro twice to create the catalog entry with the necessary control statements that are input to the cpddutl macro: %cpcat;
cards4; SET TABLE NAME=XTY74DU ; CREATE VARIABLE NAME=R744FNA EXTNAME=R744FNAM LABEL='Name of Coupling Facility' DESCRIPTION='Name of Coupling Facility' TYPE=CHARACTER LENGTH=8
INTERPRET=STRING KEPT=YES ; ;;;; %cpcat(cat=WORK.ddutl.ddutl.source); %cpddutl(entrynam=work.ddutl.ddutl.source); Run the above in a regular ITRM batch job that uses CPSTART to point to your PDB with DISP=OLD, but do not include other CMPROCES, CPPROCES or CPREDUCE macros. Thanks to Yacine Amraoui, Banque Postale, FRANCE.
Yüklə 28,67 Mb. Dostları ilə paylaş:
|