* copyright (C) 1984-2019 merrill consultants dallas texas usa
Yüklə 28,67 Mb.
|
|
Jul 5, 2015 should never have been created and no longer is. Oct 4, 2015 -Text added Oct 4, 2015: Variable R744SNAM=STRUCTURE NAME was added to TYPE74DU and TYPE74HO datasets, but not noted in the change text, nor was it added to the _BTY74HO BY List, but in Oct it was observed that the addition of R744SNAM to TYPE74HO caused a CORRECT INCREASE in number of observations in PDB.TYPE74HO (even though the WORK.TYPE74HO had the same number of observations). Because R744SNAM was not kept, the previously NODUP sort was incorrectly removing obs that were not duplicates, and its addition accidentally prevented their removal. Change 33.234 now updates the BY list to explicitly include R744SNAM. -MXG Internal Note on no longer creating a dataset: The macro variables PTY74MO/WTY74MO in VMXGINIT, and all the _xTY74MO old-style macro tokens must still be defined in VMAC74, just in case they were used in user tailoring. Member EXTY74MO also must exist to protect tailored code, and the_STY74MO token was removed from _STY74 to prevent a dataset not found condition. Thanks to Sandy Stromberg, OPTUM, USA.
ANALCAPD had #BYVAL1 rather than the value of CECSER. Jun 25, 2015
VMACCIMS ERROR due to MXG misalignment of the Trace Table that Jun 24, 2015 caused TRNEXTEN=1 when it should have been zero.
ANALWHO check and see if any data was found and now confirms Jun 24, 2015 in an MXGNOTE that observations were NOT found. The REPORT data that is created is now deleted before ending. JOB parameter is added to allow you to specify a specific job name to seek. Thanks to Dave Ireland, USDA, USA.
VMAC115 ERROR.VMAC116.LENQWHS. INVALID PRODUCT SECTION TRIPLET. VMAC116 are corrected. Change 32.172 created new MQ V8.0 Jun 24, 2015 datasets MQCHIN and MQCHININ, but documented that the Jun 30, 2015 code had not been tested with data. Now it has been. -The MQCHIN 115 records ADP, SSL, and DSN segments are significantly longer than documented and don't agree with their contents in CSQDQCTA; this will be pursued with IBM MQ Support. -The MQCHININ 116 dataset is only output if there was activity; that logic is in _ETY116A if you do want to output all segments. -Jun 30: WTASCTSR is now correctly aligned and non-zero. Thanks to Robert Miles Standish, UBS, USA. Thanks to Thomas Orlando, UBS, USA. Thanks to Scott Barry, SBBWorks Inc., USA. Change 33.150 Variable PCTDLZIP='ZIP*DELAY*PERCENT' is created in the VMAC7072 TYPE72GO dataset. Jun 22, 2015 Thanks to Jan Tielemans, KBC, BELGIUM. Change 33.149 Dataset TYPE115 variable QESDBFPT was not divided by VMAC115 4096. Jun 22, 2015 Thanks to Scott Barry, SBBWorks Inc., USA. Change 33.148 -SMFSRCH can't search for text in compressed DB2 or CICS SMFSRCH SMF records unless the EXITCICS INFILE exit is installed Jun 22, 2015 on z/OS. -At present, on ASCII, SMFSRCH can not search compressed records because the internal decompression algorithm is executed AFTER the _INFILE_ has been populated. This is under investigation for a possible solution. -If SMFSRCH found type 102 records with an IFCID LT 100, it incorrectly called UTILBLDP with a token of 102.83, which should have been 102.083. Now fixed. -SMFSRCH will now suppress processing of 102 records if a subtype of 0 is found, since this means the data is compressed, and you are not using the EXITCICS exit. -If a subtype GE 8000 is found the subtype is then converted to BMC.
ANALQBAT SMF30RQT and SMF30SQT variables rather than values that Jun 20, 2015 are calculated as in ANALINIT. This is a member which you need to tailor in your USERID.SOURCLIB. There are changes you must make (look for CHANGE HERE in the member.) It requires are least SAS 9.3 since it uses ODS and creates PDF output.
EXTY22PB records create the existing TYPE22_B with the I/O Config IMAC22 changes. New TYPE22PB dataset is created from SECID='0B'x VMAC22 that do not have RECIND=9, with Reconfigured PCIE. VMXGINIT function identifiers. Jun 19, 2015 Thanks to Scott Barry, SBBWorks Inc., USA.
VMACBVIR and subsequent Grid-Cluster Containers. Jun 18, 2015 Variable TMDLYCPY is now kept in BVIR33, overlooked. Jul 7, 2015 Variables MAXAHCT AVEAHCT MAXBHCT AVEBHCT are now kept in BVIR20. Thanks to Giuseppe Giacomodonato, EPVTECH, ITALY. Thanks to Mike Creech, Black Knight Financial Services, USA. Change 33.144 Support for SMF Type 119 new Subtypes 94 and 95 create EXT11994 dddddd dataset description EXT11995 T11994 TYP11994 SSH Client Connect Start FORMATS T11995 TYP11995 SSHD Server Connect Start IMAC119 The $MG110AM/CI/MA/CO formats were updated with new VMAC119 values. VMXGINIT Jun 23, 2015 Thanks to Hyrum Smith, Schwab, USA.
VMACNDM 4 bytes, variable PTUNCH34 now $EBCDIC30 vs 34. Jun 15, 2015 -Truncated fields for CT record - MXG Message is correct, vendor contacted. Change 33.142 Variable QWACATWT was missing the divide by 4096 and the VMACDB2 variable QWACPQCT should have been input &PIB.4. Jun 15, 2015 Thanks to Scott Barry, SBBWorks Inc., USA. ====== Changes thru 33.141 were in MXG 33.06 dated Jun 11, 2015========= Change 33.141 Documentation only. The EXITCICS/CICSIFUE INFILE exit EXITCICS WILL NOT work with VSAM SMF input, on SAS log shown as Jun 11, 2015 DSNAME=SYS1.MAN1, VOLUME=Z21SMF,DISP=SHR,UNIT=3390, TYPE=NONINDEXED,SPANNED=YES, RECORDSIZE=(.,32767),AMP=('AMORG'),RECORDS=0 Maybe the DSNAME=SYS1.MAN1 is a clue, but it is the AMP= parameter that identifies the SMF file as a VSAM file. The SAS log also reported: MXGNOTE: SMF EXIT CICS IS IN USE. ERROR: ERROR DETECTED BY INFILE/FILE EXIT: . FATAL: UNRECOVERABLE I/O ERROR DETECTED IN THE EXECUTION OF THE DATA STEP PROGRAM. Thanks to Greg Meyer, ISUZU, USA. Change 33.140 RMF III RCD records INCOMPATIBLY changed for z13 causing VMACRMFV invalid values in SVCLCNM and other variables in RBRCDR. Jun 11, 2015 The change apparently was RMF APAR OA44101 supporting Jun 16, 2015 Simultaneous MultiThreading (SMT) on the IBM z13 server that extended the Resource Collection Data Period Data (RCDPD) array entries in the RMF Monitor III RCD table. -Test IF RCDPLSC GE 0 corrected to IF RCDPLSC GT 0. When RCDPLSC=0, there is no LC Service Class Extension. Have NOT ACTUALLY seen RCDPLSC=0, so this is for safety and not an actual error. Thanks to Michael Kampert, Schwab, USA.
ANAL9914 Type 99 Subtype 14 records. Syntax for SYS1 report, Z13: Jun 9, 2015 %ANAL9914(SYSTEM=SYS1,SYSTYPE=Z13) Thanks to Raymond J. Smith, OPTUM, USA. Change 33.138 -z13 in SMT Mode, TYPE70EN had missing or mis-aligned VMAC7072 values when there was a second zIIP engine. Jun 9, 2015 -z13 in SMT Mode, PCTMVSBY was incorrect by a few percent Jun 11, 2015 because LPARBUSY time was incorrectly subtracted. -z13 in SMT Mode zIIP variables in TYPE70xx datasets, could be missing, or DUPLICATE RECORD message could be printed. REQUIRED FOR SMT. The MXG SMT support has come in several iterations as new sequences of data exposed untested logic but this appears to finally resolve the one-to-many merge. If you are testing SMT on z13, please email MXG Support with subject: SMT UPDATES so we can inform you if there are any further required changes for SMT mode. Thanks to Giuseppe Giacomodonato, EPVTECH, ITALY. Thanks to Scott Wiig, US Bank, USA. Thanks to Douglas C. Walter, CitiCorp, USA. Thanks to Brent Turner, Citigroup, USA.
FORMATS new ABENDAID dataset with many details on each intercept. EXABNAID IMACAA
TYPEAA TYPSAA
VMACAA VMXGINIT
Jun 5, 2015 Change 33.136 Support for MainView for IMS 5.2 (a/k/a IMF, CONTROL/IMS) FORMATS for COMPATIBLE changes. VMACCIMS -New variables added to CIMSTRAN: Jun 3, 2015 TRNDLCMX='MAX*DC*CALL*TIME' TRNDLDMX='MAX*DB*CALL*TIME' TRNESSMX='MAX*ESS*CALL*TIME' TRNDLWMX='MAX*WAIT*NOT-IO*TIME' TRNDLDTH='DLI DB*OR*ESS CALLS*GE THRESHOLD' TRNDLCTH='DLI*DC* GE*THRESHOLD' TRNDLDER='DLI*CALL*ERRORS' TRNESSER='ESS*CALL*ERRORS' TRN1STCP='SCHD*FIRST*DLI*CPU' -New values added to variables SMQFLAG, PROGTYPE FLGSPCHR. Change 33.135 Support for Optional CICS User CMODNAME=VMX CMODHEAD=VMX IMACICVP segments. IMACICVQ IMACICVR
IMACICVT IMACICVU
IMACAAAA UTILEXCL
VMAC110 Jun 9, 2015 Change 33.134 -ANALDB2R mis-reported the number of reports produced ANALDB2R when accounting reports were requested but no data met READDB2 the selection criteria. May 26, 2015 -READDB2 applied selection criteria inappropriately to the 105 106 and 107 IFCIDS which are subsystem specific. -ANALID parameter added to ANALDB2R to pass to READDB2 so that you can see what records were in the input data. -Page headings on PMSPR01 were incorrect for the first page of the second and subsequent DB2 subsystems reports. -If the input to PMLOL02 and PMLOK03 was a SAS dataset on tape, the T102S172 dataset was missing from the PROC COPY that moves the data to DASD (zOS only) since there must be concurrent dataset open for these reports Thanks to Judy Xu, WiPro, USA. CHANGE 33.133 ZOSEM User SMF INPUT STATEMENT EXCEEDED because MXG had VMACOSEM expected an 8-byte Resource field length. Since this May 21, 2015 resource looks like it might be a DSNAME, the resource array variables are expanded to 44 bytes. Thanks to John Schoenbeck, IBM Global Services, USA.
VMAC30 misaligned because the triplet was incorrectly input: May 21, 2015 SMF30_US_COMPRREQ SMF30_US_COMPRREQ_PROB Jan 4, 2016 SMF30_US_QUEUETIME SMF30_US_EXECTIME SMF30_US_DEF_UNCOMPRIN SMF30_US_DEF_COMPROUT SMF30_US_INF_COMPRIN SMF30_US_INF_DECOMPROUT These variables now also have "ZEDC" in their label. -Jan 2016: APAR OA48268 corrects invalid values in the first four variables above, especially when SMF data is from logstream processed by IFASMFDL. Thanks to Scott Barry, SBBWorks, Inc., USA.
SMFSRCH record type and you had specified that nnn type value in UTILBLDP your USERADD=XXXX/nnn argument. The USERADD= erroneously May 22, 2015 had nnn which generated a call to INCLUDE VMACnnn. May 25, 2015 -SMFSRCH now invokes %ANALID report to tabulate the SMF records, subtypes, versions, etc., that were selected; any undefined user records selected are identified. -SMFSRCH LOOKFOR= supports multiple strings with ANDOR choice, and new SELECTION= argument lets you write your own IF statement for the selection. -UTILBLDP needed revision to delete the ID VIEW that was introduced with the addition of %ANALID. If ID is found in a non-BUILDPDB context the _SID sort is suppressed and the ID VIEW is deleted. -ASCII only, SMFSRCH. LENGTH from INFILE statement is now used in place of LENGTH=LENGTH(_INFILE_) for the written length, after an SMF 79 record with a blank last byte was truncated, and detected by an MXG ERROR message, causing the SAS documentation TO BE READ, that stated that the LENGTH() function returns the length of the string, but EXCLUDING any trailing blanks. Thanks to Mike Obrien, Bank of America, USA. CHANGE 33.130 Missing values in SMF70xxx variables in PDB.ASUM1131 and ASUM113 (now archaic PDB.ASUM113 from no longer supported subtype May 20, 2015 2 Type 113s) if the SMF SYSTEM variable in Type 70 does not match SMF70STN in the PR/SM LCPUADDR segments. This logic has been used since always to read from TYPE70PR IF SYSTEM=SMF70STN; SYSTEM=SYSNAME; (IBM SMF70SNM - From IEASYSXX) to use the SYSNAME for the merge with TYPE113. In this new data with SYSTEM NOT matching SMF70STN, it was observed that SMF70STN matched SYSNAME, but now, the original SYSTEM is unchanged: IF SYSTEM=SMF70STN OR SYSNAME=SMF70STN; IF SYSTEM=SMF70STN THEN SYSTEM=SYSNAME; /* ELSE LEAVE THE SMF 70 ORIGINAL SYSTEM FOR MATCH */ -Do not use Subtype 2 SMF 113 with accumulated data; they may not always match up with 70PR because 113 data is lost. Always use Subtype 1 INTERVAL HIS data. IBM has stated they will NOT update Subtype 2 records. Thanks to Virginie Peigney, CA-SILCA, FRANCE.
VMACTMS5 EE4/EE4M/EE4X and EF4/EF4M/EF4X. May 20, 2015 Thanks to Scott Barry, SBBWorks, Inc., USA. ====== Changes thru 33.128 were in MXG 33.05 dated May 19, 2015========= CHANGE 33.128 z13 with one or more LPARs enabled for SMT-mode will have VMAC7072 zero observations in the TYPE70 dataset from all NON-SMT May 19, 2015 mode systems, UNLESS, ACCIDENTALLY, the SMT-mode systems are all at the end of your input SMF file. The SMTSEGNR variable that identifies this is an SMT-mode record was incorrectly retained. Thanks to Scott Wiig, US Bank, USA.
VMAC74 Change 33.124 replaced with logic to match all segments. May 18, 2015 This change is required for valid TYPE749 data. May 22, 2015 May 22: STD and AVG variables created for Queue/Execute durations, which are now formatted TIME16.6. Thanks to David Cogar, Wells Fargo, USA. Thanks to Carl D. Ellis, Wells Fargo, USA.
VMACDB2H (32) header is present. The product section for APPTUNE May 19, 2015 records is not at the end of the SMF 102 but instead is at the beginning of the record. Change 32.148 revised the header input logic, to support unexpected out-of-order headers, but the LENLEFT calculations used record LENGTH. Now for APPTUNE 102s (QWHSIID GE 8000x), the ENDPROD is used to determine the LENLEFT for all header segments. Thanks to Jan Tielemans, KBC, BELGIUM.
VMAC7072 -CPUWAITM now subtracts LPARBUSY from CPUWAITM for zIIPs. May 16, 2015 earlier code had the subtraction for CPs, with no impact. -SMF70CAN/SMF70CTN variables are again kept in TYPE70PR. Thanks to Jim Horne, Lowe's, USA. Thanks to Giuseppe Giacomodonato, EPVTECH, ITALY. ====== Changes thru 33.124 were in MXG 33.05 dated May 12, 2015========= CHANGE 33.124 MXG support for 74-9 zEDC was revised again, when it was VMAC74 discovered that only the first card was output, and that May 13, 2015 the number of PCIE pairings of the PO/DO segments is not May 18, 2015 the same as the number of Hardware Accelerator FO/10 segments. Original change text here removed, See 33.127. -Variable R749LAID was removed; there is no adapter id. -These new variables were added from the DO segments R749DBYR R749DBYT R749DFMT R749DPKR R749DPKT R749DWUP R749DWUM -May 18: See Change 33.127, which is REQUIRED. Thanks to David Cogar, Wells Fargo, USA. Thanks to Scott Barry, SBBWorks, Inc., USA. ====== Changes thru 33.123 were in MXG 33.05 dated May 12, 2015========= CHANGE 33.123 Support for Tandem ZMS Style of records; the previous MXG VMACTANZ support was for what are now "Legacy Style" records. May 12, 2015 Thanks to Mark Goforth, First National of Nebraska, USA. CHANGE 33.122 Reserved Change. CHANGE 33.121 z13 SMT-mode TYPE70/PR data IS WRONG in MXG 33.03-33.04. VMAC7072 In ONLY RMF 70 records written in SMT-mode, the datasets May 12, 2015 TYPE70, TYPE70PR, RMFINTRV, ASUM70PR, ASUM70LP, ASUMCEC, and ASUM70PR can have incorrect values. MXG 33.03 support for z13 SMT-mode records continued to use LCPUADDR from the PHYSICAL to merge with CPUID in the CPUD segment and the results matched initial RMF reports, but now I find that the LCPUADDR won't always match the CPUID value! And, one mismatch destroys the merge, trashing both the TYPE70 and TYPE70PR datasets. The code had to be redesigned to sort and merge on the physical sequence to pair up the two segments to create the per-LCPUADDR TYPE70PR dataset and populate TYPE70.
PHYSICAL segment is retrieved from the Diag 204 interface and is generated by the hardware/firmware when the core is brought into the configuration. It does not match the logical core id in SMF70VPA which is provided by z/OS in the PR/SM logical processor data section(s) of the home partition"
there was no corresponding CPUID=7 in the CPUD segment. I was also confused when a box with 5 CP cores and 5 zIIP cores in PHYSICAL had SIX unique CORE_IDs, but that was because the home partition had 4 logical CP cores and 2 logical zIIP cores, which results in the 8 CPU data sections, with 6 unique CORE_IDs.
IHDR120 in %INCLUDEd after the header and Product segments have VMAC120 been INPUT, and the "instream" macro variable MAC120H is VMXGINIT defined so that you can select/tailor after those header May 6, 2015 variables have been input.
IMACICMX length 384 was missing a final END; statement, causing a May 6, 2015 syntax error if UTILEXCL told you to tailor IMACICMX.
SAGANAL title for all reports produced by PROCs that support the May 7, 2015 CONTENTS= option to label the result. (Procs FREQ, MEANS, and CONTENTS do not support that feature.) -Report 31 added with top ten MSU hours for CICS DFHSIP. -Number of variables kept reduced to 68 from 1533; syntax CPU: and CPI: for TYPE30 kept TYPE70PR CPU: variables.
SMFSRCH specified. Logic added for USERADDS=NOUSERID was wrong. UTILBLDP Logic corrected, but USERADDS=XXXX/nnnn (even if _IDXXXX May 6, 2015 is already defined in IMACKEEP) will circumvent. -It was also possible to skip all user record selection due to an ENDOFFILE test that is removed. -The TYPE30_D dataset (each DD in SMF 30) is not populated by MXG default, and although the only text is the DDNAME, since SMFSRCH is to find and print all SMF records found, new TY30UD argument is defined in UTILBLDP (default is blank), and TY30UD=YES default is set in SMFSRCH, so that dataset TYPE30_D will be populated if type 30 records are selected, or you can use TY30UD=NO in your SMFSRCH call to prevent any observations to be created in TYPE30_D. -ECHO=YES specified in SMFSRCH to always print the MXG created by UTILBLDP on the log. -In testing, an exposure was observed: multiple uses of the file INSTREAM caused text to be overridden, which did not cause an error, but some sites have made INSTREAM to be a permanent dataset, and this exposure could lead to errors for those sites. The //INSTREAM DD was intended to be temporary and used internally as the destination for writing SAS code that is then executed with %INCLUDE INSTREAM. These members no longer use INSTREAM; instead, each writes/reads from FILENAME TMXGxxxx CATALOG 'WORK.MXGTEMP.TMXGxxxx.LOG'; in the WORK file so no DD is required, and no conflict! A few members (notably UTILBLDP) do externalize INSTREAM, and they were not changed. ANALCOMP VMXGPRAL VMXGPRNT QAWPS8 QA94648 SAGANAL PRINTLOG ONLYJOBS ASUMMSUS ANALRRTM ANALPGM ANALCOMP Thanks to MP Welch, Bank of America, USA. CHANGE 33.116 Support for ZCOST AutoSoftCapping Version V3.0.00 adds EXZCO02S subtypes and data. These are the datasets MXG creates: EXZCO02W DDDDDD Dataset Description EXZCO04C ZCO01 ZCOS01 ZCOST CPC EXZCO04G ZCO02 ZCOS02 ZCOST LPAR EXZCO04L ZCO02S ZCOS02S ZCOST SERVICE CLASS PERIOD EXZCO04P ZCO02W ZCOS02W ZCOST WLM INPORTANCE EXZCO04R ZCO03 ZCOS03 ZCOST SYSPLEX/CPC VMACZCOS ZCO04G ZCOS04GP ZCOST GENERAL PARMS VMXGINIT ZCO04C ZCOS04CP ZCOST CPC PARMS May 5, 2015 ZCO04L ZCOS04LP ZCOST LPAR PARMS May 28, 2015 ZCO04P ZCOS04PG ZCOST GENERAL PERIOD PARMS ZCO04R ZCOS04PL ZCOST LPAR PERIOD PARMS This support was completed in MXG 33.06. CHANGE 33.115 Mainview for MQ BBMQBUFF dataset EFFICIENCY variables are VMACBBMQ now correctly input as RB4 instead of PIB. Variables are Apr 29, 2015 SBPIAWPE SPBRAWPE SBPSAWPE.
VMACCIMS it was incorrectly in the CIMSDBDS dataset KEEP=list. Apr 29, 2015 "zAAP" in Label changed to 'ZIP' (Feb 2006): CPUZIPTM='CPU TIME*ON ZIP*ENGINES' TRXZAOCP='CPU TIME*ZIP ELIGIBLE*RAN ON CP' Thanks to Betty Wong, Bank of America, USA. CHANGE 33.113 Some T102S106 variables (DB2 zPARM flags and values) were FORMATS misaligned, not INPUT, or had wrong/no decoding format. VMAC102 Align: QWP2DMIN,QWP2DSEC Apr 30, 2015 Format: Use $MGD361U for QWP4SECA1_TYPE, QWP4SECA2_TYPE Format: New $MGD106P for QWPBDL, QWPBSDL Input: QWP4S1IL for DB2 V10, was only V11 and wrong. Thanks to Lai Fai, Bank of America, USA.
UTILEXCL NOTE THAT 5.3 now sets STGPROT=YES as the IBM DEFAULT, to VMAC110 INTENTIONALLY cause an ABEND for storage overlays caused Yüklə 28,67 Mb. Dostları ilə paylaş:
|