* copyright (C) 1984-2019 merrill consultants dallas texas usa
Yüklə 28,67 Mb.
|
|
SYSPLEX names, ANALZPCR sometimes accidentally had the right SYSTEM in the right SYSPLEX, but not always. Now, PDB.TYPE70, whose SYSTEM-SYSPLEX pairing is always right, is read to create a format mapping SYSTEM to its SYSPLEX (in addition to the existing format that maps the SYSTEM to MVSLEVEL). Then, PDB.TYPE70PR is read, values of MVSLEVEL/SYSPLEX are set from SMF70STN format lookups, and the WORK.TYPE70PR dataset has correct SYSPLEX and MVSLEVEL for LPARNAME, so ANALZPCR now always is right. -zPCR load fails with MXG-created .zpcr file if incomplete SMF/PDB data input was read and SCP The input SMF or PDB must have TYPE70 observations for every SYSTEM to get the MVSLEVEL, which is used to set the SCP tag from SMF70STN in TYPE70PR. If a system's data doesn't have TYPE70 data and is only in the TYPE70PR LPAR data, the SCP tag value 'z/OS-MXG**' has always been created in the .zpcr file, but not documented, and that tag value must be changed for zPCR to load the text file. Now, there are ERROR messages when you have missing data telling you MUST change those SCP tag values before using the .zpcr file. Output reports also are enhanced to show the MVSLEVEL and SMF70STN for each LPAR. -ANALZPCR program fails with overlapping FORMAT values if the SMF/PDB input has data with multiple MVSLEVELS from the same SYSTEM. ANALZPCR will now detect and continue to execute, and will use the LAST MVSLEVEL for SCP tag, but will also print ERROR messages when this is detected. ANALZPCR can't easily be redesigned to support this multiplicity, but you can split your input and run the ANALZPCR twice to create each of the two .zpcr files. -When ANALZPCR is run on ASCII, to read a PDB.TYPE70 that was CIMPORTed from z/OS and that had been created on z/OS "before" the TRANSCODE attribute was added by MXG 27.01 (Change 27.014) to all HEX-containing-$CHAR-variables, variable CPUTYPE will have been converted to '886D'x for a true CPUTYPE='2094'x. ANALZPCR now tests CPUTYPE for these "wrong" values: 886D/886F/8870/8871x in CPUTYPE and corrects them to: 2094/2096/2097/2098x so that the NAME tag that zpcr expects is created in the .zpcr file. -When PDB=SMF was used, DASDIORT counted only DASD I/O by selecting only DEVCLASS=20X when SMF was read. But when PDB=PDB was used, but only if your PDB.TYPE74 dataset had other DEVCLASS's captured, DASDIORT was the total I/O. Now, the PDB=PDB logic selects only DEVCLASS=20X counts. -MXG uses LCPUSHAR, the Initial Shared LPAR Weight, but if IRD is active (LPARWLMG='Y'), zPCR expects SMF70ACS, the Current Shared Weight. -The message for Linux LPARs that you have to manually put the SCP type printed the SYSTEM instead of the LPARNAME. -When PDB=CECTIME was used, specialty engine and ICF's could have been miscounted. -SELECT PEAKTIME or PEAKPCT only summarize zOS SYSTEMs, and they do NOT contain IFLs, nor ICF partitions, and thus are unlikely to be used. CECTIME is DEFAULT. -Mar 4, 2010. zPCR Release 6.3a failed to load a model with CPUTYPE created an updated zPCR program with same-day-service! With that new release, the Book Configuration, which is now VERY important, can be specified in the pull-down. Thanks to Frank Bereznay, IBM Global Services, USA. Thanks to Raff Rushton, IBM Global Services, USA. Thanks to Karl Lasecki, Chemical Abstracts Service, USA. Change 28.020 Variable QAQSGDSP, Sharing Group Dispositions, in dataset FORMATS TMMQQAA is now decoded by new FORMAT $MGTMQDI. VMACTMMQ Feb 12, 2010 Thanks to Paul Volpi, UHC, USA.
VMACEDGR IBM writing a question-mark character instead of a count Feb 10, 2010 of errors as an &NUM4. field. This change eliminates that NOTE and the HEX record dump for all four xxxxERR variables (by changing the INPUT to use ?? &NUM.4.), but IBM will be contacted for an explanation; perhaps they store a question mark when the error count is larger than the maximum of 99999 that fits in that 4-byte field. These instance of invalid data values can be identified because RVTRERR will be a missing value in the EDGRXEXT or EDGRVEXT dataset. Thanks to Paul Volpi, UHC, USA. Change 28.018 All of the duration/clock values are in 1024 microsecond VMACTMVT units and not the 64 microsecond units MXG had coded; the Feb 10, 2010 correct 1024 multiplier is now used. The FORMAT TIME13.3 will still display decimals the maximum resolution of one millisecond (0.001 seconds), since 1024 microseconds is just at teeny bit more than one millisecond. Thanks to Paul Volpi, UHC, USA.
UTILEXCL 'PSB SCHD' field name, is now recognized and supported. Feb 8, 2010 Thanks to Thomas E. Porta, TYCO Electronics, USA. Change 28.016 SAS Version 9.2 changed the PROC GCHART's PATTERN default DOC to a terrible choice that produces unreadable patterns. Feb 6, 2010 Feb: Investigating alternatives. Aug 9, 2010 Aug: Better definition of color/patterns now defined as default in VMXGINIT, but if you choose your own, yours will instead be used. Change 28.015 Support for z/TPF SMF 89 record; z/TPF put the subtype in VMAC89 byte 19 rather than bytes 19-20 as documented for z/OS. Feb 3, 2010 MXG protects either z/OS or z/TPF SMF ID=89 records now. Thanks to Paul J. Westerhout, KLM, THE NETHERLANDS. Change 28.014 MXG 27.10-MXG 27.27. The wildcard colon-suffix in the VGETDDS DDNAMES= argument, to allocate all DDNAMEs starting with Feb 3, 2010 those characters (%VGETDDS,DDNAMES=PDB:); worked ONLY for DDNAMES=PDB:). Any OTHER prefix characters ahead of that colon caused syntax errors. And, DDNAMES=ALLDAYS or even a list of specific DDNAMES=MON TUE WED names also failed, with LIBNAME PDBx NOT ASSIGNED because a separate error sent VMXGDDS to try to allocate LIBNAMES starting with PDBn, no matter what names you used in your DDNAMES=. Note: If the DATASET you will look for in VMXGSET might not exist in all of the LIBNAMES you specify, you can use OPTIONS NODSNFERR; and only the found datasets will be read by VMXGSET, and the SAS log will indicate which DDnames had the DATASET. Thanks to Paul Naddeo, FISERV, USA. Thanks to Jim Horne, Lowe's Companies, USA. Change 28.013 Variables QW0127FG/QW0127PG/QW0128FG/QW0128PG are INPUT FORMATS and kept in T102S127 and T102S128 datasets. The "PG" VMAC102 variables are four-byte replacements for the three-byte Jan 30, 2010 existing "PN" page number variables. Thanks to Chuck Hopf, Independent Consultant, USA.
VMXGFIND FIND all observations in all datasets that meet your test VGETVAR criteria, outputs those selected obs into a separate SAS VMXGPRNT data library, and prints all selected obs (all variables, Jan 31, 2010 with the variable's name AND label as heading). For example: %VMXGFIND( PDB=PDB,
PDBOUT=PDBOUT, KEEPIN=JOB, FIND= IF JOB='MXGBUILD'; , PRINT=YES); finds all obs with JOB='MXGBUILD' in all of the datasets in the PDB input library, and outputs those obs into their datasets in the PDBOUT data library, and prints.
that will be used in the SAS code in your FIND= argument. Only the datasets with one or more of those variables are read, and the FIND= logic is used to output the selected. You can test with complex logic with multiple variables that don't exist in all of the dataset. For example: %VMXGFIND( PDB=PDB,
PDBOUT=PDBOUT, KEEPIN=STARTIME STRTTIME INTBTIME, FIND= IF ('31JAN2010:15:00:00'DT LE STARTIME LT '31JAN2010:16:00:00'DT ) OR ('31JAN2010:15:00:00'DT LE STRTTIME LT '31JAN2010:16:00:00'DT ) OR ('31JAN2010:15:00:00'DT LE INTBTIME LT '31JAN2010:16:00:00'DT ) ;, PRINT=100);
hour, from RMFINTRV plus all detail RMF datasets, from CICINTRV and any other CICxxxxx interval datasets, from DB2STATx interval datasets, from the SMFINTRV dataset, and from ANY other PDB dataset with ANY of those three variables meeting the test. There will be log messages UNINITIALIZED VARIABLE printed when a dataset being read doesn't contain all KEEPIN= variables, but they are harmless and unavoidable.
KEEPIN=RACFUSER QWHCAID FSRUID, FIND=IF RACFUSER=::'JOE" OR QWHCAID=:'JOE' OR FSRUID=:'JOE';, PRINT=YES ); will find all observations from user "JOE". New %VGETVAR(DDNAME=,DATASET=,VARNAME=), used in VMXGFIND determines if variable VARNAME exists in DDNAME.DATASET. VMXGPRNT now deletes WORK.TMPPRNT (previously WORK.SP_L). Thanks to Chuck Hopf, Independent Consultant, USA. Change 28.011 Reading VSAM SMF for type 119 records failed; the OFFSMF VMAC119 was not added to all of the offsets. Jan 29, 2010 Thanks to Kim Westcott, State of New York, USA. Change 28.010 Variable SHIFT is created from the QWHSSTCK (END TIME) in VMACDB2H the DB2 Header for all DB2 records, and SHIFT is now kept VMACDB2 in all of the datasets created from SMF 100 and 101 data. Jan 28, 2010 Thanks to Atle Mjelde, ErgoGroup, NORWAY
VMAC0 to absence of &IB.4. in its INPUT statement, but as this Jan 28, 2010 new variable is not used elsewhere, this had no impact, except the waste of your time chasing this message down. I missed this because there were no IPL records in any of my z/OS 1.11 test SMF data, and because the absence of an INFORMAT in an INPUT statement is not a syntax error. Thanks to Jim Horne, Lowe's Companies, USA. Change 28.008 The SPIN.SPINMOUN and SPIN.SPINSYSL dataset could grow to ASUMTAPE massive size (1.8 million obs in SPINSYSL) because there Jan 26, 2010 was no test to stop their spinning after some number of days. Now, IMACSPIN is read and its value of _SPINCNT is used to determine when a still-unmatched syslog message should be "spun" again. When observations are deleted due to their age, the count is printed on the SAS log. The default IMACSPIN has _SPINCNT of zero, because if you failed to read INSTALL's instructions on how to EDIT your IMACSPIN member, at least then when you run your first BUILDPDB, all of the jobs in SMF are output to the PDB library, with none output to SPIN, so you will find all your jobs, complete and incomplete, in that first PDB. Then, when you ask about all those incomplete jobs, tech support will point you back to read INSTALL and IMACSPIN to set _SPINCNT. But for ASUMTAPE, I will always spin the incomplete mount events at least one day, using _SPINCNT+1, which should allow most incomplete mounts today to match up tomorrow, even if you haven't changed IMACSPIN. Thanks to Jim Horne, Lowe's Companies, Inc, USA.
ANALRANK
Jan 25, 2010 Change 28.005 -Support for Sun StorageTek Enterprise Library Software VMACSTC Version 6.2 and Version 7.0. Version 6.2 adds new field Jan 25, 2010 STC14N4K, the number of 4K blocks, used to re-calculate Feb 2, 2010 STC14VSZ, which was previously limited to a max of 4GB. Sep 16, 2010 Variables marked with * below, are only in Version 7. -New variables added to STCVSM13 dataset; *STC13PLX='TAPEPLEX*FROM WHICH*VTV RECEIVED' -New variables added to STCVSM14 dataset; STC14SRS='SYNCHRONOUS*REPLICATION*STATUS' STC14RUN='REWIND*UNLOAD*RECEIVED*DATETIME' *STC14PLX='TAPEPLEX*FROM WHICH*VTV RECEIVED' -New variables added to STCVSM16 dataset; STC16INF='RTD*CHANNEL*INTERFACE*ID' STC16ADR='MVS*ADDRESS*OF*RTD' -New variables added to STCVSM17 dataset; STC17INF='RTD*CHANNEL*INTERFACE*ID' STC17ADR='MVS*ADDRESS*OF*RTD' *STC17DFL='DISMOUNT*FLAG' -New variables added to STCVSM18 dataset; STC18INF='RTD*CHANNEL*INTERFACE*ID' STC18ADR='MVS*ADDRESS*OF*RTD' -New variables added to STCVSM19 dataset; STC19INF='RTD*CHANNEL*INTERFACE*ID' STC19ADR='MVS*ADDRESS*OF*RTD' -New variables added to STCVSM25 dataset; STC25SCL='MVS*STORAGE*CLASS' STC25TUS='SPACE*USED BY*CURRENT*VTVS' STC25NDV='NUMBER*OF HOLES*DELETED*VTVS' STC25LUT='MVC*LAST*USED*DATETIME' STC25LWT='MVC*LAST*UPDATED*DATETIME' -New variables added to STCVSM26 dataset; STC26MGT='VTV*MANAGEMENT*CLASS' -New variables added to STCVSM27 dataset; STC27CTP='CARTRIDGE*TYPE' STC27MV3='VOLSER OF*MVC3*CONTAINING*THE VTV' STC27MV4='VOLSER OF*MVC4*CONTAINING*THE VTV' -New variables added to STCVSM28 dataset; STC28RUN='REWIND*UNLOAD*RECEIVED*DATETIME' *STC28PLX='TARGET*TAPEPLEX*FOR*EXPORT' -Sep 16: No code change, but with this change there were obs created from subtype 7 records in STCHSC dataset with the Oracle SL8500 hardware, whereas MXG 27.05 created zero observations in STCHSC. Thanks to Davide Marone, SGS S.p.a. ITALY Thanks to Carlo Gavinelli, SGS S.p.a. ITALY Thanks to Gianvittorio Negri, SAS Institute, ITALY. Change 28.004 The EXTREMELY DETAIL DB2 Trace Report PMTRC01 is revised ANALDB2R for efficiency, keeping track of which of 350 possible Jan 25, 2010 trace datasets are actually populated, and only using Feb 22, 2010 their variables. Some uninitialized variables messages and character to numeric conversions were eliminated. Note that your DBA needs to enable all of these IFCIDs to cover all I/O and SQL calls: IO 6 7 8 9 10 29 30 34 35 36 37 38 39 40 41 105 107 114 115 116 119 120 SQL 6 7 8 9 10 11 12 15 16 17 18 19 20 22 44 45 53 55 58 59 60 61 62 63 64 65 66 68 69 70 71 73 74 75 76 77 78 79 86 87 88 89 95 96 105 107 125 157 158 159 160 163 174 175 177 183 ACCOUNT With CONNTYPE=4 in the argument list, only records from CICS Connection will be reported, so if both I/O and SQL traces are enabled, you can see what DB2 Tables/DBIDs are touched for each transaction, and can what types of SQL calls were made for each transaction. Unfortunately, you can NOT map SQL calls to each DB2 Table that was used. %ANALDB2R(PDB=SMF,PMACC01=NO,PMACC02=NO, PMSTA02=NO,PMSPR01=NO, PMTRC01=YES,CONNTYPE=4); See correction in Change 28.083. Thanks to Paul Volpi, UHC, USA. Change 28.003 Summary of (archaic) SMF 118 records in ANALTCP and of ANALTCP current SMF 119 records failed if PDB was on TAPE. ANAL119 Jan 21, 2010 Thanks to Chuck Hopf, Independent Consultant, USA. Change 28.002 Variables CPUZIPTM and CPUIFATM added to this example ASUMSMFI summarization of PDB.SMFINTRV. Jan 21, 2010 Thanks to Frank Lund, EDB Business Partner Norge AS, NORWAY Change 28.001 Unused Change. LASTCHANGE: Version 28. =========================member=CHANGE27================================ /* COPYRIGHT (C) 1984-2010 MERRILL CONSULTANTS DALLAS TEXAS USA */
MXG Version 27.27 is the 2010 "Annual Version". MXG Newsletter FIFTY-FIVE is dated Jan 20, 2010 MXG Version 27.11 is dated Dec 31, 2009, thru Change 27.337 MXG Version 27.10 is dated Dec 6, 2009, thru Change 27.325 MXG Version 27.09 was dated Oct 14, 2009, thru Change 27.283 MXG Version 27.08 was dated Sep 1, 2009, thru Change 27.229 MXG Version 27.07 was dated Aug 11, 2009, thru Change 27.195 MXG Newsletter FIFTY-FOUR is dated Aug 10, 2009 MXG Version 27.06 was dated Jul 20, 2009, thru Change 27.167 MXG Version 27.05 was dated Jun 29, 2009, thru Change 27.145 MXG Version 27.04 was dated May 27, 2009, thru Change 27.107 MXG Version 27.03 was dated May 4, 2009, thru Change 27.083 MXG Version 27.02 was dated Apr 13, 2009, thru Change 27.066 MXG Version 27.01 was dated Mar 17, 2009, thru Change 27.042 MXG Version 26.26 was dated Feb 12, 2009, thru Change 26.326 MXG Version 26.26 was the 2009 "Annual Version". MXG Newsletter FIFTY-THREE was dated Feb 3, 2009 Instructions for ftp download can be requested by using this form: http://www.mxg.com/ship_current_version Your download instructions will be sent via return email.
II. SAS Version requirement information. III. WPS Version requirement information. IV. MXG Version Required for Hardware, Operating System Release, etc. V. Incompatibilities and Installation of MXG 27.27. VI. Online Documentation of MXG Software. VII. Changes Log
and is updated with new notes frequently. All Newsletters are online at http://www.mxg.com in the "Newsletters" frame.
Member CHANGESS contains all changes that have ever been made to MXG. All MXG changes are also online at http://www.mxg.com, in "Changes".
I. MXG Version 27.27 dated Jan 20, 2010, thru Change 27.361. MXG Version 27.27 is the "Annual Version" for 2010. Major enhancements added in MXG 27.27, dated Jan 20, 2010 VMXGCNFG 27.356 The standard SAS JCL Proc can be used for MXG. // EXEC SAS,CONFIG='MXG.SOURCLIB(CONFIMXG)' //MXGNAMES DD DSN=MXG.USERID.SOURCLIB(MXGNAMES),DISP=SHR TYPE112 27.358 Support for OMEGAMON ONDV SMF 112 SUBTYPE 0100X TYPEEDGR 27.349 Support for RMM APAR OA28930, RDBLKCNT/RDTOTAL. TYPEEDGR 27.349 Support for RMM APAR OA24025 RDDESKEY. TYPEEDGR 27.339 RVTxERR RVPxERR variables now numeric, incompatible. TYPE1415 27.361 MXG 27.06-27.11. BUFNO always zero due to typo. TYPESRMH 27.340 SRM Thales Security PTF SL24010 INCOMPATIBLE support. VGETDDS 27.359 WAIT=N option protects for DSNAME already in use. VGETDDS 27.330 New DATEJUL= correctly generates julian dsnames. VMXGALOC 27.355 NOWAIT added, create/allocates are now conditional. READDB2 27.351 READDB2 didn't always invoke its EXdddddd members. TYPE80A 27.357 TYPE8066 dataset enhanced. TYPETMVT 27.347 TMON/VTAM "SI" record Interval variables now INPUT. ANAL307X 27.346 Analysis compares Hourly CPU in 70, 72, 30 interval. JCLCIDB2 27.345 UTILBLDP JCL example for CICS, DB2, and ASUMUOW+. ASUM70PR 27.344 Variables LPCTBY/PCTLPBY missing for PHYSICAL LPAR. VMXGSET 27.343 VMXGSET permits multiple datasets with APPEND=YES. VMACSMF 27.341 WARNING: SUBTYPE GT 255 message now not defaulted. TYPEIMSA 27.354 MXGNOTE now prints the value of _IMSVERS on the log.
(Must remove &EPDBINC from EXPDBINC if you used the %UTILBLDP to create your BUILDPDB source SYSIN code.)
ANALRMFR 27.333 RMF Summary Report (REPORT=SMRY) was incorrect. ANALJOBE 27.332 Analysis of Job Events revised, JESNR test removed. TYPE80A 27.331 Protection for unknown TOKDANAM eliminates STOPOVER. TYPE82 27.330 SMF82PDK, SMF82RKN, SMF82PTA increased, corrected. TYPE110 27.329 MXG 27.10, CICIDNNR NOT FOUND, only if _S110ST used. ASUMMIPS 27.327 DUR70 created with actual vs expected DURATM. TYPEDB2 27.326 Variable QDSTQCIT now not deaccumulated. Major enhancements added in MXG 27.10, dated Dec 6, 2009 TYPE0 27.325 Updated Support for z/OS 1.11, ID=0 now supported. TYPECDC 27.324 Support for IBM's InfoSphere Change Data Capture CDC TYPEDB2 27.319 Support for QMDAPTYP='JCC' (Type 4 JDBC Driver). TYPE7072 27.317 Support for APAR OA28670 RMF 70 Crypto Express3 VMACORAL 27.306 Support for restructured ORACLE SMF records. TYPESYNC 27.289 Support for CPUZIPTM added in SYNCSORT SMF records. READDB2 27.322 IFCIDS= features (STATS, DB2ACCT, Dataset Name, etc) MRGDB2 27.321 MERGE of PDB.DB2ACCT and PDB.DB2ACCTP for DB2PARTY JCL40GIG 27.316 JCL example to split/parallelize BUILDPDB job. VGETDDS 27.310 New DATEJUL= created DSNAMES with Julian YYYDDD. VMXGDUR 23.308 SYNC59-NO/VMXGDUR/VMXGSUM (final?) enhancements. ASUMMIPS 27.305 ASUMMIPS now inflates MSU by system CAPTURE RATIO. TYPE110 27.303 CICS Total I/O and Total Other Wait updates. ALOCGOVO 27.299 Allocate GDG's with GOOVOO to minimize DSENQUEUE. TYPETIMS 27.298 Expensive algorithm to uncompress ASG/Landmark data. TYPENTSM 27.312 MXG 27.09 only. NTSMF Processing failed, RECFM. DAILYDSx 27.287 MXG 27.09 only. TMC= should have been TAPEDATA=. TYPE74 27.287 MXG 27.09 only. DEBUGGING PUT statement removed. CONFIGxx 27.294 MXG 27.08-27.09 only. Option MAUTOLOCDISPLAY removed. TYPEPROS 27.295 PRO/SMS SMF record misalignment corrected. ASUM70PR 27.294 IFL obs in ASUM70LP/ASUMCELP missing STARTIME fixed. ASUM70PR 27.292 First LPAR had missing values in LPSHAR/TOTSHARE. TYPE42 27.291 Variable S42DSRDD incorrect by factor of 10**6. ASMTAPEE 27.300 ASMTAPE ML-45 enhancements to TMNTnnnn messages. Major enhancements added in MXG 27.09, dated Oct 14, 2009 TYPE110 27.260 TRANSPARENT UNCOMPRESS OF CICS/TS 3.2 RECORDS!!! This means you do NOT have to assemble and install the EXITCICS "infile" exit. HOWEVER: VERY EXPENSIVE COMPARED TO EXITCICS. SEE THE TEXT OF THE UPDATED CHANGE.
TYPE110 27.274 SORT Order for CICFCR changed, A17DSIXP/DTRDS fixed. CICINTRV 27.274 CICS DISPATCH INTERVAL LARGER THAN INTERVAL fixed. READDB2 27.272 READDB2 support for WANTONLY now works. TYPE110 27.270 CICS STID=115 should not exist. ASUM70GO 27.268 Example summarizes PDB.TYPE72GO, changing interval. TYPEVMXA 27.264 z/VM MONWRITE example processes only USER domain. TYPE74 27.263 R744FNAM not in TYPE74DU, ITRM sites need update. ASUMMIPS 27.262 MSU/MIPS for zIIPs and zAAPs added to CPs summary. TYPE120 27.255 WAS SMF1209FI/EV (CPU/Elapsed Times) corrected. RMFINTRV 27.252 Format missing for many duration variables, fixed. Yüklə 28,67 Mb. Dostları ilə paylaş:
|