* copyright (C) 1984-2019 merrill consultants dallas texas usa
Yüklə 28,67 Mb.
|
|
this was inconsistent with the operation when 'ALL' was default because only NOxxx filters were coded. Those in effect did override 'ALL', but this change restores the earlier behavior so that any NOxxx filter parameters will be honored even when 'ALL' is coded. No further behavior changes are planned for this condition. -Change 32.037 also altered the behavior when both NOxxx table filter and table xxx selection parameters were used for the same table. The xxx table selection parameter always overrode the NOxxx table filter parameter regardless of input order. This change reverses that behavior so the NOxxx filter parameter always overrides the xxx selection parameter for the same table. No further behavior changes are planned for this condition. -NOTE: Please do not select AND filter on the same RMF Monitor III table in the same ASMRMFV run to avoid unintended results. -Documentation in the ASMRMFV source prologue and in the ADOCRMFV member has been updated appropriately. -When validating a large RMF III CFI table and a section offset exceeded X'FFFF', the calculation for the output CFI record length was incorrect. The entire CFI table was then skipped because the result record length was over 32K. Message RMFV035S was issued followed by Return Code 0016. The register used in the length calculation is now cleared prior to inserting each half word section length for validation so that the CFI table is not skipped. -Updated documentation content, and mixed casing as well! -REQUIREMENT: In order to implement this fix the current ASMRMFV utility program from this MXG change must be installed. See MXG SOURCLIB member JCLASM3 for sample JCL for the assembly and link-edit install steps. Thanks to Warren Cravey, Fidelity Institutional, USA.
VMAC7072 bytes to the right, producing very large numeric values, Mar 10, 2014 but the SCALING/ACTUAL ratio, now SMF70CPA_RATIO was the same! The new z/OS 2.1 RATIO is used instead of raw-CPA to calculate the SMF70CPA Service Unit Per Second value: New: SMF70CPA=16000000*SMF70CPA_RATIO Old: SMF70CPA=16000000/SMF70CPA-in-Record So the new ratio is the inverse of the RMCTADJC, almost: Raw CPA=423 1/423=.002364066 ==> SMF70CPA= 37825.059 SMF70CPA_RATIO =.002364590. 37833.440 The raw-CPA is an integer, which limits the granularity of SMF70CPA value. Integer 424 sets SMF70CPA=37914.692 so the exact 37833 value cannot be set with raw-CPA. Now, since RATIO is a fraction, any numeric value can be set for SMF70CPA in future engines. The numeric difference is small, but with this change, for z/OS 2.1, SMF70CPA is now calculated using the SMF70CPA_RATIO value. Thanks to Scott Barry, SBBWorks, Inc, USA. Change 32.053 Compression ratios SMF21CRR/SMF21CRW are now calculated VMAC21 for non-3590 devices, using these existing variables: VMAC21 SMF21CRR=BYTEREAD/SMF21MDR; VMAC21 SMF21CRW=BYTEWRIT/SMF21MDW; VMAC21 For 3590s, note that: Mar 10, 2014 -Bit SMF21NCT is on ONLY for 3590s and inputs the channel and device bytes read/written SMF21BRN/BWN/DBR/SMF21DBR variables, now used for the compression ratio. -Bit SMF21MFV is NOT on for 3590s, so the previous counts of bytes read/written/compressed will always be missing for 3590s in variables SMF21MCR/MDR/CRR/MCW/MDW/DRW. Thanks to Jorge Fong, NYC Information Technology, USA. Change 32.052 Support for SMF 112-35 Tivoli Enterprise Monitor Server EX112TEM record creates new dataset IMAC112 dddddd dataset description VMAC112 112TEM T112TEMS Tivoli Enterprise Monitor Server VMXGINIT There are also subtypes 32 thru 37 that may be supported Mar 10, 2014 in a later update to this change. Thanks to Michael Oujesky, DTCC, USA.
VMXGUOW NO to bypass reading the CICUOW dataset. This will let VMXGINIT you set _UOWCIC to _NULL_ and save a lot of WORK space by Mar 9, 2014 eliminating the creation of a dataset that is ONLY used if you also have MQ data and have specified MQMADD=YES.
VMACRMFV (Report Class) did not contain the five CPUxxxTM values Mar 8, 2014 (TCB/SRB/RCT/HPT/IIP), suggesting these RMF III datasets were seldom used, because the same data exists in RMF I TYPE72GO dataset. They are now created so that I could compare the CPU times captured by RMF I, RMF III and SMF. See MXG Technical Newsletter 64 MVS Technical Note 1.
VMXGPRAL COMPARE procedure executions so you can insert the SAS VMXGPRA1 code in a WHERE clause to select what will be analyzed or VMXGPRNT printed. Mar 5, 2014
VMAC6 that printed INVALID PRINTIME messages, while the source Mar 4, 2014 of the invalid records is investigated.
VMACEJES to match the standard SMF record structure with triplets Mar 4, 2014 (Offset/Length/Number) for the three segments of data. Apr 17, 2014 New variables ESMFCNSR ESMFCNCO ESMF5NSF ESMF5NCF added. MXGEJES Member MXGEJES is a standalone MXG code execution that is Jun 16, 2014 provided to Phoenix International (to encourage their users to consider MXG Software). Thanks to Ed Jaffe, Phoenix Software International, Inc., USA. Change 32.046 Support for XCOM 11.5 AND 11.6 (INCOMPATABLE). Records FORMATS were restructured and MXG code revised to match. Formats VMACXCOM $MGXCMST and $MGXCMFO now decode STATFLGX and FILEOPT. Mar 3, 2014 A new segment contains the SMS DATA/MGMT/STORE classes. Mar 25, 2014 Mar 25: Code corrected, revised for new variables. Thanks to Rosa Maria Martinez Alonso, Bustia, SPAIN. Change 32.045 Support for MQ PCF Files creates new dataset: EXPCFMD
IMACPCF dddddd dataset description TYPEPCF PCFMD PCFMD PCF MD Record TYPSPCF VMACPCF
VMXGINIT Feb 28, 2014 Thanks to Peter Farrell, Commerce Bank, USA.
EXTY87 creates new dataset: IMAC87 dddddd dataset description TYPE87 TY87 TYPE87 GRS Component TYPS87 VMAC87
VMXGINIT Feb 27, 2014 Thanks to Graham Harris, Royal Bank of Scotland, ENGLAND.
ANALID PRINTWHAT= added with default of BOTH. Other possible Jan 24, 2014 values are DETAIL for the detail report or SUMMARY to get only the SUMMARY report. For the summary report the data is summarized to the SMFIDSUB level.
VMXGHSM in line 1797. Feb 26, 2014 Thanks to Lindsay Oxenham, IBM Global Services, AUSTRALIA. Change 32.040 TMON/VTAM vendor maintenance now populates fields so they VMACTMVT can be properly decoded/documented. Variables SXORTTD and Feb 25, 2014 SXORTT are now decoded and formatted as TIME13.3, and the Feb 27, 2014 SXHRT1D/2D/3D/4D/5D are now labeled INTERVAL vs TIME. Thanks to Paul Volpi, UHC, USA.
ANALBVIR IBM-reports in ANALBVIR/JCLBVIR will likely require you EXBVR011 to revise your reporting, but you will no longer have the EXBVR101 thousands of previous variable names, as the restructure EXBVR111 creates the "subtype" dataset (EG: BVIR30) with only the EXBVR301 header variables for that subtype, and repeated segments EXBVR321 are output in a new dataset (EG: BVIR301), with only ONE EXBVR322 set of variable names. A few variables names were changed EXBVR324 and the new reports have been validated against IBM's. EXBVR331 -In an emergency for an old report, member ZMACBVIR is the IMACBVIR prior BVIR support structure that can still be used for JCLBVIR now, but it will NOT be updated nor supported for future VMACBVIR changes to the BVIR data records. REMOVED JAN 2015. VMXGINIT -This was a MAJOR update, contributed by Joe; the other cites are for the Early Adopters who helped in testing. Feb 25, 2014 -These are now the datasets created from BVIR data: Feb 28, 2014 Mar 2, 2014 DDDDDD MXG MXG Mar 8, 2014 DATASET DATASET DATASET Mar 11, 2014 SUFFIX NAME LABEL DESCRIPTION Mar 14, 2014 BVIR01 BVIR01 VNODE VIRTUAL DEVICE PIT Mar 17, 2014 BVR011 BVIR011 HNODE GRID (PIT) CLUSTER Mar 25, 2014 BVIR02 BVIR02 VNODE ADAPTER POINT IN TIME BVIR10 BVIR10 HNODE HSM POINT IN TIME DDDDDD MXG MXG DATASET DATASET DATASET SUFFIX NAME LABEL DESCRIPTION BVIR01 BVIR01 VNODE VIRTUAL DEVICE PIT BVR011 BVIR011 HNODE GRID (PIT) CLUSTER BVIR02 BVIR02 VNODE ADAPTER POINT IN TIME BVR101 BVIR101 HNODE HSM PIT PHYSICAL DEVICE BVIR11 BVIR11 HNODE GRID POINT IN TIME BVR111 BVIR111 HNODE GRID (PIT) CLUSTER BVIR20 BVIR20 VNODE VIRTUAL DEVICE HISTORY BVIR21 BVIR21 VNODE ADAPTER HISTORY BVIR30 BVIR30 HNODE HSM HISTORY BVR301 BVIR301 HNODE CACHE CONTAINER/PARTITION BVIR31 BVIR31 HNODE RESERVED BVIR32 BVIR32 HNODE LIBRARY HISTORY BVR321 BVIR321 HNODE LIBRARY DEVICE TYPE BVR322 BVIR322 HNODE LIBRARY POOLING CSP MEDIA BVR323 BVIR323 HNODE LIBRARY POOLING GUP BVR324 BVIR324 HNODE LIBRARY POOLING GP MEDIA BVIR33 BVIR33 HNODE GRID HISTORY BVR331 BVIR331 HNODE GRID CLUSTER Updates made after MXG 32.02: Feb 26: -The variable DEVCLASS was renamed DEVCLSID due to conflict with existing 30 and 74 name DEVCLASS, and the variable DVCLASID was renamed to DEVCLSID as only one name is needed. Format $MGB32DC was merged into and replaced by $MGB10DC. -Some formats for BVIR bit map ranges were corrected. -SCRVOLCT added to _BBVR324 BY list. -ADOCBVIR updated with why NODUP won't 50% remove. -Feb 28:
Dateparts in BVIR21 and BVIR31 were two years wrong - I had moved their calculation to the header, but had left the recalcs in those two subtypes. -Mar 2:
The loop counter to read the GUP Container Segments only read the first four of the thirty-one segments, so obs were not created in BVIR323 and BVIR324. -But dataset BVIR323 was output twice incorrectly! -Lengths of FORMATTED character variables are now set. -Mar 8:
Variable POOL was added back into BVIR323 and to its BY list. I revised Joe's code and used _I_ for loop where the loop variable is not kept, but I messed up this one. -Mar 11:
Corrections to examine all 32 entries in the POOL loop. Two new fields, POOL to the GUP and POOL and MEDIA to BVIR323 and BVIR324 datasets respectively. A lost pad to the container to bring it to 16 bytes. With this change, all prior problems are resolved!!! -Mar 14:
Type 30 code needed to skip some bytes for BVIRVERS=3. -Mar 17:
Type 20 records BVIRVERS=5 with only 192 bytes caused INPUT EXCEEDED. Rest revised to verify data exists for V5 and remainder of skip bytes were removed as there is only one 20 record. -Mar 25: There are lots of "null" segments ('00'X for the ID and/or sum of metrics zero) in BVIR records that not only wasted space but also were the NODUP culprits. Now only segments with activity or valid IDs are output, massively reducing the number of observations. Selection code added to EXBVR321-322-324-331 exit members, except that the test for POOLACTV GT 0 is internal to VMACBVIR because it is also used to terminate the pool loop. -POOLACTV is tested for these datasets, plus BVIR321 when DEVCLSID GT '00'X BVIR322 when MEDATYPE GT '00'X BVIR324 when PYMDID GT '00'X BVIR323 when POOLACTV (SUM of Metrics) GT 0 -BVIR331 output when BVIR331ACTV GT 0 (Sum of Metrics) -Resulting dataset observation counts: BVIR20,30,31,32,33,301,321 8826 obs BVIR322 44130 obs BVIR323,324 26478 obs BVIR331 7614 obs Thanks to Joe Babcock, JPMorganChase, USA. Thanks to Scott Barry, SBBWorks, Inc, USA. Thanks to Dan Case, Mayo, USA. Thanks to Giuseppe Giacomodonato, EPVTECH, ITALY.
VMACIMS was no GMTOFFTM in a prior record; the LINK IMSSTCK in Feb 24, 2014 56FA logic was prior to the input of the first 12-byte Mar 17, 2014 character timestamp, from which DEC and then GMTOFFTM can Mar 23, 2014 be decoded. The LINK for 56FA was located to after that first DEC value is known; once GMTOFFTM has been created, its value will be retained. Mar 17: IMSSTCK links for log records 01x,03x,06x were mis-located prior to the creation of GMTOFFTM; only if those records were prior to other GMT-Offset-containing records, the GMTOFFTM was missing and IMSSTCK was then not corrected to local time zone. Mar 23: END; relocated from IMSSTCK: to before LCODE=55. Thanks to Rosa Maria Martinez Alonso, Bustia, SPAIN. Change 32.037 -RMF III Enhancements, Fixes, and Notes ADOCRMFV -A second RMFV008I message is added for non-VSAM data sets ASMRMFV such as the RMFBSAM and SYSIN DD files that will show DCB CLRMFV attributes DSORG, RECFM, LRECL, BLKSIZE, BUFNO, and DOCLRMFV BUFSP. JCLRMFV -A second RMFV008I message is added for RMF Monitor III JCLCRMFV VSAM data sets that will show ACB attributes DSORG, JCLDRMFV LRECL, CISIZE, BUFNO, BUFSP, and Record Count. Feb 23, 2014 -NOTE: The number of VSAM records read for any one Feb 24, 2014 particular RMF Monitor III data set will usually exceed Feb 26, 2014 the number of records in the file. This is because a physical record may contain data from more than one MINTIME interval and data is processed one MINTIME interval at a time. -NOTE: The BUFNO and BUFSP values for all RMF Monitor III data sets will be the same because they share a common VSAM LSR buffer pool. -Data set and DCB information will now also be shown for the non-VSAM SYSPRINT DD file. -BUFNO=20 was always assigned for the SYSIN DD file. However, when the SYSIN DD file is blocked this wastes memory because usually this file has few records. When a blocked SYSIN file is detected a value of BUFNO=2 will be assigned. -When NOxxx table filtering options were mixed with xxx table selection options, ASMRMFV incorrectly turned on selection for all tables when the first xxx table option was encountered. -The NONE option to suppress all output is now obsolete. This option was inflexible; it did not allow RMF Monitor III table selection and/or filtering. Results were also incomplete as no output statistics were shown. So it was not very useful nor was its usage easily understood.
with no action taken and no error messages issued. -The alternative to the defunct NONE option to preview what might be output by ASMRMFV with user specified selections and/or filters in effect is to simply code in JCL:
or //RMFBSAM DD DSN=NULLFILE Either tells the access method to suppress the RMFBSAM output. In either case RMFBSAM will be empty and cannot be used for a PDB build. -The CLRMFV Clist used in the TSO Clist Method for running ASMRMFV has been upgraded to accept NULLFILE as a value for the OUTHLQ Clist parameter. When OUTHLQ('NULLFILE') is coded the RMFBSAM file will be allocated as DSN=NULLFILE thus also providing an alternative to the obsolete ASMRMFV NONE option. OUTHLQ('DUMMY') can also be used. In this case the SPACE Clist parameter is forced to zero and the OUTMLQ and OUTLLQ Clist parameter settings are irrelevant. -With RMFBSAM output suppression the result is useful to determine how many bytes of output will be generated with the selection and/or filter options used. Also decisions can be made as to which tables should be selected for a future PDB build based on data volume. Then the actual RMFBSAM file can be realistically sized later for allocation. Be sure to review ASMRMFV messages RMFV105I to study the output data statistics including bytes output. -JCLRMFV, JCLDRMFV, and JCLCRMFV example members are each updated to include an example with RMFBSAM output suppression. -The DOCLRMFV documentation for CLRMFV Clist usage has been updated with current information. -Documentation in the ASMRMFV source prologue and in the ADOCRMFV member has been updated to note that the NONE parameter is obsolete and to explain the alternative use of DSN=NULLFILE for RMFBSAM. -Removed residual extraneous characters in the first RMFV002I message when no PARM= field existed in JCL. -NOTE: The ASMRMFV SIZE parameter is still supported and also produces no RMFBSAM output regardless of JCL for that file. When SIZE is coded all selection and/or filtering parameters are ignored. SIZE shows a snapshot of RMF Monitor III VSAM file allocations as well as space and index usage. -NOTE: If the ALL table select ASMRMFV option is coded it overrides any other table select or table filter options that may be present regardless of input order. This is a new behavior with this change. Prior to this change NOxxx table filter options would override the coded 'ALL' value. This is no longer the case.
options to avoid unintended results. -NOTE: If an xxx table select and an NOxxx table filter are coded for the same RMF III table the table select option for this table now always overrides the filter option regardless of input order. This is a new behavior with this change. Prior to this change the last option encountered for the same table would be effective. This is no longer the case.
III table in the same ASMRMFV run to avoid unintended results. -REQUIREMENT: In order to implement this fix the current ASMRMFV utility program from this MXG change must be installed. See MXG SOURCLIB member JCLASM3 for sample JCL for the assembly and link-edit install steps. Thanks to H. Sterling James, DST Systems Inc. CHANGE 32.036 MXG developer utility used when creating a new dataset to UTILNODU determine the sufficient list of variables needed in the Feb 21, 2014 new _Bdddddd By List macro that will remove duplicate observations with the PROC SORT NODUP option (invoked in BUILDPDB, or the TYPSxxxx "sorting" types, or by the use of the _Sdddddd dataset sort macro.) The initial "logical" _Bdddddd is created based on what seems to be useful order (e.g. READTIME JOB JESNR for a job related dataset). Then, the input data file is read twice, the dataset is created and sorted by _Bdddddd, and then the count of observations deleted as duplicate in the PROC SORT log messages must be exactly one-half of the observations read by the sort. If that is not true, an ad hoc approach was used to find the additional vars needed. This utility sorts the dataset twice, once with the original _Bdddddd and a second time with BY _ALL_, and the unmatched observations, those NOT deleted, are printed for easy identification of needed variables. Unlikely an MXG user will ever need this, but someday someone will be happy to find it if that is their task. CHANGE 32.035 Websphere ID=120 Subtype=9 datasets TYP1209C, TYP1209E VMAC120 TYP1209S and TYP1209U had incorrect BY variable that Feb 18, 2014 did not remove duplicates and caused DELTA120TM to be always missing. Variables SM1209CR SM1209CS and SM1209CM are added to those datasets, and the new sort order are: MACRO _BT1209C SYSTEM SMFTIME SM1209CR SM1209CS SM1209CY SM1209EM SM1209CM % MACRO _BT1209E SYSTEM SM1209CR SM1209CS SM1209CH SM1209CM % MACRO _BT1209S SYSTEM SMFTIME SM1209CR SM1209CS SM1209CY SM1209EQ SM1209CM % MACRO _BT1209U SYSTEM SMFTIME SM1209CR SM1209CS SM1209EY SM1209FA SM1209CM % Thanks to Wayne Bell, UNIGROUP, USA.
JCLUOWP because the CICSTRAN DD was being opened for read by JCLUOWV VMXGUOWC before the CICSTRAN had been written. Relocated Feb 17, 2014 the %INCLUDE of VMXGUOW to correct. Change 32.033 Support for XPTR 5.2 creates these 23 new datasets: EXXPTR23 dddddd dataset description EXXPTR24 XPTR23 XPTR23 X/TND USER DISCONNECT EXXPTR25 XPTR24 XPTR24 X/NET COMMAND PROCESSED EXXPTR26 XPTR25 XPTR25 API TRANSACTION EXXPTR33 XPTR26 XPTR26 API SUMMARY EXXPTR34 XPTR33 XPTR33 MIGRATION DETAIL EXXPTR52 XPTR34 XPTR34 STORAGE POLICY TARGETS EXXPT104 XPTR52 XPTR52 USER REQUEST TRACKING EXXPT106 XPT104 XPTR104 JOB ACCUMULATION EXXPT107 XPT106 XPTR106 JOB ACCUMULATION WITH RENAME EXXPT109 XPT107 XPTR107 ARCHIVE CREATED EXXPT112 XPT109 XPTR109 REPORT ACCUMULATION EXXPT113 XPT112 XPTR112 REPORT BROWSE EXXPT118 XPT113 XPTR113 USER COMMAND EXXPT120 XPT118 XPTR118 X/PTR DETAIL OUTPUT RECORD EXXPT124 XPT120 XPTR120 NEW REPORT EXXPT130 XPT124 XPTR124 XNET COMMAND PROCESSED EXXPT131 XPT130 XPTR130 ARCHIVAL RETRIEVAL COMPLETED EXXPT132 XPT131 XPTR131 ARCHICE DETAIL EXXPT133 XPT132 XPTR132 TEST FILE SPACE RELEASE EXXPT140 XPT133 XPTR133 MIGRATION DETAIL EXXPT152 XPT140 XPTR140 VSAM BUFFER STATISTICS EXXPT153 XPT152 XPTR152 USER REQUEST TRACKING IMACXPTR XPT153 XPTR153 REPORT VERSION DELETE VMACXPTR -Some minor issues: there are duplicate records VMXGINIT written for subtype 41 and 50, and the date of the Feb 21, 2014 SUBMTIME in subtype 130 is 2005. -BY lists for earlier datasets have been updated and all datasets with data have been "NODUP" validated. Thanks to Phil Grasser, Norfolk Southern, USA.
VMACDB2 incorrectly deaccumulated when it's a static ZPARM value. Yüklə 28,67 Mb. Dostları ilə paylaş:
|