* copyright (C) 1984-2019 merrill consultants dallas texas usa
Yüklə 28,67 Mb.
|
|
that is now supported after IBM z/VM Support UPDATED the documentation today at http://www.vm.ibm.com/PUBS/MON630/MRPRCAPM.HTML Thanks to Wolfgang Kueller,s IT Solutions, AUSTRIA. Change 34.098 If the same POLICY-NAME is used in different SYSPLEX, the ANALACTM report did not print the WLM definitions because of the Apr 23, 2016 filter criteria. This revision adds SYSPLEX variable to protect for this unwise choice. Thanks to Scott Barry, SBBWorks, Inc, USA.
CLEARDB2 unexpectedly also created the WORK.DB2STSBP dataset; the READDB2 READDB2 code to suppress it was not added to the MACKEEP, Apr 21, 2016 and a typo in CLEARDB2 had changed _WDB2SBP to _WDB2SBR. Change 34.095 -Gregorian dates were not displayed correctly in ASMRMFV ASMRMFV message log due to incorrect leap year testing in ADOCRMFV FINDGREG code. Dates of the form ddmmmyyyy were listed Apr 20, 2016 as one day later than actual. Day of the week values in Apr 27, 2016 these messages were also one day later than actual. May 11, 2016 Julian dates of the form yyyy.ddd are correct. -NOTE: RMFBSAM output data was NOT affected only the dates in ASMRMFV log messages. ASMRMFV does not modify dates in the RMFBSAM output file. -A S0C4 Abend could occur when an VSAM I/O Read Error happened processing the Sample Set Header (SSH) RMF III table. This Abend was also possible processing the Data Set Header (DSH) and Service Policy (SVP) tables. -Several buffer handling improvements for performance are added and described below. -Three initial buffer size adjustments are changed to reduce FREEMAIN/GETMAIN overhead when an increase is needed during processing. These are applied to the 32752 RMF III VSAM Record Size to avoid buffer expansions.
------------- ------------------- ----------------- Decompression 2 160 Sample 1 2 Service Policy 1 18
expensive because the IBM ERB3RDEC decompression program must decompress the data first to find out if it fits in the buffer. If not, the buffer must be expanded and the decompression repeated on the subsequent call. -The increases in buffer sizes are offset by new logic that only applies the adjustments to the first buffer acquisition. Any subsequent expansions now only request the actual memory needed, but should be rare. In prior ASMRMFV versions the adjustment was applied to every GETMAIN so the buffer areas became unnecessarily large when several expansions of the same buffer were needed. -In testing with 12 various RMF Monitor III data sets buffer memory used was actually slightly less than the current production ASMRMFV level. Your actual results will vary. REGION=300M was used for these jobs. ASMRMFV 33.274 ASMRMFV 34.095 LSR 16000K LSR 16000K INDEX 32K INDEX 32K SAMPLE 2207K SAMPLE 3198K DECOMP 7521K DECOMP 5118K SVP 96K SVP 1151K ====== ====== *ALL* 25855K *ALL* 25499K
tailorable in ASMRMFV by the user as before. These set the initial buffer size multipliers for the Decompression, Sample, and Service Policy buffers respectively. Any change to these values requires an assembly and link of ASMRMFV. -The BUFFERS parameter in ASMRMFV provides a listing after each RMF Monitor III data set processed to show buffer usage and expansions. Ideally expansions should be zero. The default is NOBUFFERS and the display is suppressed. -NOASIX parameter causes ABEND S0C4 or RMFV075W messages; tests for this parameter were incomplete, but it is an emergency parameter to suppress ASI extension data,if ASI data needed to be bypassed. NOASIX causes data loss and was intended to be used only when recommended by support. Next ASMRMFV will correct; this is just a don't use note! Thanks to Randy Hewitt, HPE Enterprise Services, USA. Thanks to Randy Shumate, Reed Elsevier Technology Services, USA. Change 34.094 PDB.JOBS observations with ABEND='JCL' were created for BUILD005 purge records that were for job transmission. Now, Apr 18, 2016 IF JSTRTIME=. AND SYSEXEC LE ' ' AND SYSTRANS GT ' ' Apr 19, 2016 the TYPE26J2 purge observation is output in PDB.NJEPURGE instead of PDB.JOBS. -Variable INTRDR is now kept in PDB.JOBS. Thanks to Ian Porter, NISSAN-NEDC, ENGLAND.
GRAFWRKX you can specify TABULATE=YES without creating any graphs. Apr 16, 2016 Tabulate was cleaned up so that there is only a single table generated rather than a table for each variable. Change 34.092 -Support for IHDRRMFV "Header" Exit member to select which IHDRRMFV RMF III records are to be read by TYPERMFV from RMFBSAM. IMACRMFV Your selection code can be put in member IHDRRMFV in your VMACRMFV "USERID.SOURCLIB(IHDRRMFV)" tailoring library, or you can VMXGINIT use the macro variable MACRMFVH "instream" to select: Apr 13, 2016 //SYSIN DD * Apr 15, 2016 %LET MACRMFVH= %QUOTE(IF ERBDTYPE='ASIG3'; ); Apr 21, 2016 %INCLUDE SOURCLIB(TYPERMFV); would only populate the ZRBASI dataset. -Removed incorrect second INPUT of CPCGRPNM/CPCGRPLM that could cause STOPOVER INPUT STATEMENT EXCEEDED ERROR. -These ZRBASI variables are now input as PIB versus RB: ASI_LVSHR4KB ASI_LVSHR1MGBYTES ASI_FREEMAINEDFRAMES Thanks to Randy Hewitt, Hewlett Packard, USA. Change 34.091 Support for IMS Log 16x Sign On/Sign Off record creates EXIMS16 new IMS16 dataset. FORMATS dddddd dataset description IMAC16 IMS16 IMS16 IMS SIGN ON / SIGN OFF VMACIMS VMXGINIT
Apr 15, 2016 Thanks to Gene Heikkinen, Blue Cross Minnesota, USA. Change 34.090 TYPE115 Macro _WTY115X wasn't listed in _N115 null macro. VMAC115 TYPECIMS Macro _WIMFMQ wasn't listed in _N116 null macro VMACCIMS UTILBLDP, cosmetic, extra blank lines were printed in the UTILBLDP code that clears the old style macros. Apr 13, 2016 A new QA report will detect _Nxxxx omissions. Thanks to Andre G. Moretto, IBM Global Technology Services/Delta, USA Change 34.089 Support for SAMS VANTAGE User LSPOOLPO record INCOMPAT VMACSAMS changes. These new variables in SAMSLSPC dataset: Apr 14, 2016 SAMSBYFR ='FREE*BYTES' SAMSCLFR ='MAXIMUM*FREE*EXTENT IN*TRACKS' SAMSDSCBPCT='PERCENT*USED*DSCBS' SAMSEAV ='EXTENDED*ADDRESS*VOLUME?' SAMSFREEC ='TOTAL*FREE*SPACE IN*CYLINDERS' SAMSHASG='GLOBAL*HASH*VALUE' SAMSHASL='LOCAL*HASH*VALUE' SAMSLPAR='LPAR WHERE VANTAGE RUNS' SAMSMFEB ='MAX FREE*EXTENT*IN BYTES' SAMSSHR ='DASD*VOLUME*SHARE*STATUS' SAMSSUBS='SUBSYSTEM WHERE VANTAGE*RUNS' SAMSSYSP='SYSPLEX WHERE VANTAGE RUNS' SAMSTCYLS ='TRKMGDSPACE*TOTAL*FREE*CYL' SAMSTEXTNT='TRKMGDSPACE*FREE*EXTENTS' SAMSTINDEX='TRKMGDSPACE*FRAGMENTATION*INDEX' SAMSTMCYLS='TRKMGDSPACE*MAX EXT*CYL PORTION' SAMSTMTRKS='TRKMGDSPACE*MAX EXT*ADDL TRKS' SAMSTTRKS ='TRKMGDSPACE*ADDITIONAL*FREE*TRKS' SAMSTVTRKM='TRKMGDSPACE*TOTAL TRACKS' SAMSTVTRKS='TOTAL*TRACKS ON*VOLUME' and variable SAMSRSVD, a reserved field, is not kept. -SAMSPOOL record was also INCOMPATIBLY changed but no new variables were created. Thanks to Emmanuelle Tanguy, ARKEA, FRANCE. Change 34.088 Unused Change Number. Change 34.087 -MXG 34.02, IMS 12.1, IMS 07 record was misaligned due to VMACIMS 8 overlooked added bytes, and incorrect input of DLRAZAAP Apr 12, 2016 that was added in IMS 13.1, not 12.1.2 -Zero divide fix if MXGRDTM=0 (fast read on ASCII). -Variable LINTSY2 is formatted $HEX16. Thanks to Paul Volpi, UHC, USA. Change 34.086 Support for TYPE8069 RACFEVNT 8069 R_PKISERV GENCERT EXTY8069 event, including protection for truncated TP2=322 and IMAC80A doc errors for TP2=343 and TP2=351. VMAC80A
VMXGINIT Apr 8, 2016 Thanks to Joseph J. Faska, DTCC, USA Thanks to William M. Vender, DTCC, USA. Change 34.085 Support for these new z/VM VXSYTEMP dataset variables FORMATS from the extended (third) segment: VMACVMXA NCM_TCT_FCOP='FICON*OPERATIONS' Apr 8, 2016 NCM_TCT_DFCOP='DEFERRED*FICON*OPERATIONS' Apr 12, 2016 NCM_SCT_FCOP='SUMM COUNT*FICON*OPERATIONS' Apr 14, 2016 NCM_TCT_FCXTM='FICON*TRANS-MODE*OPERATIONS' Apr 21, 2016 NCM_TCT_DFCXTM='DEFERRED*FICON*TRANS-MOD*OPS' Apr 24, 2016 NCM_SCT_FCXTM='SUMM COUNT*FICON*TRANS-MODE' Apr 26, 2016 -Variables PCTLPABY and PCTCPCBY are now correctly May 4, 2016 calculated AFTER the deaccumulation. May 16, 2016 -Format $MGVXACH is updated for new Channel Types. May 23, 2016 -Dataset VXSTOVDK variable QDIIOCNT is now deaccum'd. -Dataset VXSYTLCK variables CALS/CALX/SYN deaccum'd. -Dataset VXAPLSL0 is now properly deaccumulated; note that only observations with TICKS GT 0 are output. -Dataset VXSYTLCK variable CALSSCNT is DIF()d and SYNATT4S second DIF() corrected to SYNFTG4S. -Dataset VXBYUSR variables that are now deaccumulated: VMUYPLTL0-5,VMUSTLT0-5,VMUVMTL0-5,VEBALERT/HDWAI/SVSCT/ TPIAI/TVSCT/VEBVIRAI,VMDUFOCT/UOFTM/SLCNT -Dataset VXIODVSW is now deaccumulated with sort order corrected. -Some deaccumulated datasets had observations output with DELTATM value negative; those are now correctly deleted (they are the first instance so no deaccum is possible). -Dataset VXMTREND now has DELTATM=. as it is not accumed. -Dataset VXIODDEV variable VIUTIMIN's TIME12.2 reinstated. These variables had missing values and missing labels: RDEVSKSM64 RDEVWXCT RDEVRXCT SCMIDTIM SCMPDTIM PAVIDTIM PAVPDTIM
-May 16: Variable ASMSSCH in dataset VXSTOASP is only two bytes, so its accumulation wraps at a value of 65536; MXG had incorrectly used FFFFFFFFx, causing large values. But variable SCGSSCH is four bytes and would be safer. Thanks to Graham Harris, RBS, ENGLAND.
MOBWRKI2 preventing ERROR: FILE WORK.SUMSTSBP.DATA DOES NOT EXIST. Apr 7, 2016 Thanks to Jan Tielemans, KBC, BELGIUM. ======= Changes thru 34.083 were in MXG 34.02 dated Apr 5, 2016======== Change 34.083 IMS56FA ARRVTIME value is wrong for transactions that VMACIMS arrived from a system whose GMT offset is not the same as Apr 5, 2016 this system. Now, the GMT delta between the two systems is added to ARRVTIME when there is a difference so that all datetime values are local to this IMS system. Note that you must use TYPSIMST ("S" for SORT) program to invoke the _SIMS56G for the Chain correction. Use the JCL example in the TYPSIMST member's comments. -The INPQUETM is a zero value if the ENDTIME of the prior transaction in a chain is LATER than the STRTTIME of this transaction. Why this happens is not understood yet. -The _IMSVERS macro is NOT USED for the IMS56FA processing because the actual version is in the 56FA record and is used by MXG to control version differences, so multiple IMS version's 56FA records can be processed together. The MXG NOTE/WARN messages about _IMSVERS are removed. Macro _IMSVERS is used ONLY for 07/08/0A/31/35/36/40/59 IMS log records, but only the 07/08 records need 10.1 or 11.1 or 12.1 to be specified and processed separately. Thanks to Michael J. Lamdin, Verizon, USA. Thanks to David A. Bernhardt, Verizon, USA. Thanks to Matthew E Bogart, Verizon, USA. Thanks to Mark Albert, Verizon, USA. Thanks to Stephen P. Nathan, IBM, USA. Change 34.082 DB2 Trace IFCID 196 variables QWn196HY (0-8), QW0196WY, VMAC102 are formatted $HEX4 and INPUT $CHAR2 and QW0196W9 is Apr 5, 2016 formatted $HEX16. Apr 16, 2016 -Some _S102348-_S102355 dataset sort macros had repeated PROC SORTs, but there were no errors, just wasted time.
VMACSTC wrong with a slightly larger (seconds to a minute) and a Apr 4, 2016 non-integer value for a few observations, depending on the SMFTIME delta to the earlier STC timestamp; the GMT algorithm used in VMACSTC was unique and now uses the normal calculation. -Some VMXGTIME calls were revised, and dataset STCVSM28 has new variable REPDURTM=DURATION*OF*REPLICATION added. Thanks to Rudolf Sauer, T-Systems, GERMANY.
VMACVMXA to new fields added to SYTCUP segment that now create: Apr 5, 2016 LCXPMTST='PARTITION*MULTITHREADING*STATUS' LCXHGPNM='LPAR*GROUP*NAME' -Unrelated, CECSER was not populated in VXMTRSYS because that 1.4 record is written before the 1.5 record that has the CECSER is written. If you use TYPSVMXA to invoke sorting of all datasets, this update will populate CECSER into the VXMTRSYS dataset. Or you could use %INCLUDE SOURCLIB(TYPEVMXA); _SMTRSYS _SMTRPRT
RUN; Thanks to Graham Harris, RBS, ENGLAND. Change 34.079 RMF III ZRBASI dataset new variables are created/decoded: FORMATS ASICX='ADDRESS*SPACE*TYPE' VMACRMFV VALUE $MGRMFCX Apr 3, 2016 'S '='S:STARTED TASK' 'T '='T:TSO' 'B '='B:BATCH' 'A '='A:ASCH' 'O '='O:OMVS' '? '='?:UNKNOWN' 'E '='E:ENCLAVE' 'SO'='S:STARTED TASK WITH OMVS PROCESS' 'TO'='T:TSO WITH OMVS PROCESS' 'BO'='B:BATCH WITH OMVS PROCESS' 'AO'='A:ASCH WITH OMVS PROCESS' 'OO'='O:OMVS WITH OMVS PROCESS' '?O'='?:UNKNOWN WITH OMVS PROCESS' 'EO'='E:ENCLAVE WITH OMVS PROCESS' ; ASICR='WLM*CRITICAL*STATUS' VALUE $MGRMFCR 'C '='C:CPU CRITICAL' 'S '='S:STORAGE CRITICAL' 'SC'='SC:BOTH CRITICAL' ;
VMAC7072 the correct calculation in Change 34.010 was overridden VMXGRMFI by the un-removed original calculation (TCBPART/SRBPART). Apr 2, 2016 -MXG 33.33. RMF72 had MSU72, MSUINTRV, MSUPERHR variables incorrectly containing Software MSU, while MSU4HRAV did have the correct Hardware MSU. MXG 34.01 changed those three variables to correctly contain Hardware MSU, and created three new variables with Software MSU, but the MSU4HRAV was changed to Hardware MSU. This 34.02 change corrects MSU4HRAV also to contain Hardware MSU. The MSU4HRAV in RMFINTRV is only the Software MSU captured in the TYPE72GO Service Class records; the IBM Four Hour Average Software MSU actually used for software costs is in the TYPE70LAC variable in PDB.TYPE70. -RMFINTRV MSU variables for Software MSU all have an "S": Software: MSUSOFT MSUINTRVS MSUPERHRS MSU4HRAV Hardware: MSU72 MSUINTRV MSUPERHR n/a -MSUSOFT will be missing if ONLY type 72 records are read; 70s always precede 72s and CECSUSEC is retained from that type 70 to calculate MSUSOFT. Thanks to Randy Shumate, Reed Elsevier, USA.
IMACICWT
UTILEXCL VMAC110
Apr 1, 2016 Thanks to Niels Ole Kjeldsen, KMD, DENMARK. Change 34.076 Support for dataset TYPE80TK new variables TOKMREVOKED, VMAC80A TOKMREVREAS and TOKMFIREACCS/ Mar 31, 2016 Thanks to Roger X Baker, GLIC, USA Thanks to Frank Bauer, GLIC, USA.
MOBWRK05 if you included data with the CLOCK CHANGE hours. Your Mar 31, 2016 MARCH REPORT MUST RUN BETWEEN APR 2-9 WITH this update. When the clock is changed for winter/summer/daylight time two rows were created in MWRT_LOOKUP89 that caused "REPEATS OF BY VALUES" messages. This change keeps only the first row and eliminates the message. Thanks to Al Sherkow, I/S Management Strategies, Ltd. Thanks to Rudi Claes, KBC, BELGIUM. Thanks to Graham Harris, RBS, ENGLAND. Change 34.074 The bit tests for the new TYPE0201 and TYPE0202 datasets VMAC0203 for decoding variables SMF2IHASHMETH and SMF2ISIGTYPE Mar 31, 2016 were missing the final "B", causing blank values. Thanks to Robert Sample, TOMY, USA. Change 34.073 Dataset TYPE749 (PCIE) is enhanced with new calculated VMAC74 variables used in RMF reports. Mar 30, 2016 Thanks to Michael Friske, FMR, USA. Change 34.072A Support for SMF 102 IFCID 58 Added Segment creates new EX102A58 DDDDDD DATASET DESCRIPTION IMAC102 102A58 T102SA58 Added END SQL STATEMENT EXEC VMAC102 See Change 33.102 (in MXG 34.03) which corrected VMAC102 VMXGINIT and validated the new data. Mar 30, 2016 Change 34.072 RMF 73 Subtype 3 ERROR R723DNST NOT EQUAL TO R723RTYP is VMAC7072 for the seldom-used TYPE72DL, in only one record, which Mar 30, 2016 had one WRS pair with RTYP='CB' and RDNN=6 with six DSNTs Mar 31, 2016 that matched, but the RTYP='DB2' pair also had RDnn=6 but there were no segments with DSNT='DB2'. Other RTYP='DB2' records have RDNN=0. While I believe the record is wrong, and should have RDNN=0, IBM noted that all of the sample counts in R723RW01-R723RW15 are zero and there could be no name table entries, so MXG now circumvents by only reading the name table when there are samples recorded. (Variables R723RN01-R723RN15 are blank if no table.) -But this investigation exposed a logic error in MXG reading the name table; the OFFDSN was NOT incremented, so only the first segment was being input, repeatedly, which is now corrected. Thanks to Lorena Ortenzi, UniCredit Group, ITALY Thanks to Paolo Uguccioni, UniCredit Group, ITALY
VMAC42
Mar 29, 2016 Thanks to Michael Friske, FMR, USA. Change 34.070 I/O Connect time S42CONNTM=AVGCONMS*IOCOUNT/1000 is now VMAC42 calculated and kept in TYPE42DS, TYPE42SR and TYPE42VT Mar 27, 2016 datasets, formatted TIME13.3.
ASUM113 -The SORTED BY list had SM113STM SM113CPU transposed. VMAC113 -Dataset TYPE1131 _KTY1131 "variable keep" macro now Mar 24, 2016 works for both TYPE1131 and ASUM1131. Thanks to Scott Barry, SBBWorks, Inc, USA.
VMAC73 supported. VMAC73 now detects split records, but IBM Mar 24, 2016 plans a 2ndQ APAR that redesigns the 73 split so that Mar 31, 2016 self-contained records are written (EIX=HEN), and that May 4, 2016 design is already supported in the existing 73 code. These are the variables populated from the extended seg: SMF73ECP SMF73EOC SMF73EOD SMF73EOS SMF73ETC SMF73ETD SMF73ETS CHFRATE CHFACTV CHFDFER CHFXRATE CHFXACTV CHFXDFER So, if you don't have the APAR and want those variables added, you would use n the _STY73EX macro. %INCLUDE SOURCLIB(VMAC73,VMACSMF,IMACKEEP); DATA _VAR73; _SMF ; _CDE73; _STY73EX; The TEMP73EX dataset is created with the split segments from the second (split) record (when EIX GT HEN) and the new _STY73EX replaces the _STY73 dataset sort macro to sort and merge TEMP73EX into TYPE73. This split record had 158 valid channels but only 105 fit in the first 32K record (because IBM writes 256 channel segments in every every record, including offline, so the other 53 Extended Channel segments were in the second record. -May 4: APAR OA50254 eliminates the split 73 records and each record will be self-contained with matching channel path data sections and extended channel path sections. Thanks to Joachim Sarkoschitz, DATEV, GERMANY
VMAC85 tested R85PVRM for specific versions but not for '2020'. Mar 24, 2016 However, that ancient test is no longer needed for the subtypes 78,79, and 88 since all records now have the missing early fields that needed that test, so new z/OS versions' won't need VMAC85 to be updated for R85PVRM. Thanks to Joachim Sarkoschitz, DATEV, DENMARK.
VMACVMXA memory variables, VMDUFACTC, VMDUFIBRC and VMDCTPNS. Mar 21, 2016 -New: You can specify %LET MXGABND=8709; so that the BROKEN CONTROL RECORD ERROR will now also cause the job to ABEND with a USER 8709 abend code, so the error can't be overlooked. (This error usually occurs when a back level of MXG tries to read data from a new zVM version.) Thanks to Graham Harris, RBS, ENGLAND. Change 34.065 CICS/TS 5.3 MNSEGCL=5 INPUT STATEMENT EXCEEDED error due VMAC110 to MXG read of 128 bytes but the segment is only 120. Mar 21, 2016 You have to have enabled TSQUEUE Resource Class data to populate dataset CICSRDQU to encounter this error. Thanks to Bob Duchesneau, Northwestern Mutual, USA.
VMACBBMQ in every E6 record is 4 bytes shorter than ENTL, but the Mar 21, 2016 four bytes are unused, so this heuristic detects the last Mar 28, 2016 record condition and the last segment is now output in the BBMQQUES dataset (which will have more observations.) -Circumvention for incorrect ENTL for E4 which caused the BBMQLMGR to be trashed. ENTL=1336 in header but only 1275 exist.
-BY lists updated for NODUP removal for BBMQBUFF, BBMQCHAN BBMQLMGR and BBMQPAGE, although no duplicates have ever been created, just to be consistent. -Mar 28: Datetimes were incorrectly set to GMT in 34.01. now corrected to local. Thanks to Jim Swinarski, Credit-Suisse, USA. Change 34.063 ERROR START GREATER THAN END creating DBID/OBID format is ANALDB2R corrected with this rewrite of VMFMT102 and the dropping VFMT102 of the system and timestamps from the keys to the format. Mar 22, 2016 The FORMATs are now the same whether you use POINTINTIME Mar 31, 2016 or the T102S105/107 records but POINTINTIME will always be more accurate since it is a snapshot of what DB2 sees while the 105/107 records will only reveal databases that have been opened. Thanks to Jutta Gleixner-Schmid, Allianz, GERMANY ==== Changes thru 34.062 were in FINAL MXG 34.01 dated Mar 21, 2016==== Change 34.062 Cosmetic. With MXGREADSMF=LOGGER or =BOTH, log messages VMACSMF "LAST RECORD" and the SMF Summary box of times/bytes read Yüklə 28,67 Mb. Dostları ilə paylaş:
|