* copyright (C) 1984-2019 merrill consultants dallas texas usa
Yüklə 28,67 Mb.
|
|
because it is a per-connection-type-per interval dataset. See Change 31.093. Sep 2014 Note: If you have tailored TYPEDB2/TYPSDB2 code, you will have to modify your code to create the DB2STAT5 dataset. If you uses _NDB2 to null all datasets, then you will need to add MACRO _WDB2ST5 DB2STAT5 % so that that dataset is created. Otherwise, you will get ERROR: No data set open to look up variables printed before "The data set WORK.ZZDB2ST5 may be incomplete/" Thanks to Joe Babcock, JPM Chase, USA. Thanks to John Hornor, JPM Chase, USA. Change 31.080 -RMF III Enhancements, Fixes, and Notes ASMRMFV -Change 31.021 increased the number of ENC table triplets May 28, 2013 processed from 1 up to 6. However, later analysis of data from several installations confirmed that triplets 2-6 all have zero token values for every enclave and no identification information. Therefore, processing is reverting back to only processing the first triplet as before.
-Note: Users who have implemented change 31.021 or 31.062 will see the number of observations in the ZRBENC dataset significantly reduced with this change. -Validity checking for most RMF Monitor III tables has improved. This is intended to detect malformed RMF III tables that could cause abends or bad data in the PDB build later. For example, table headers cannot have a zero or excessive length. However, users should not see a general increase in skipped records. Table skipping should be a rare event. -ASMRMFV was inconsistent in handling tables with entry counts of zero. Most tables with this condition were already ignored with no output to any file. However, for some tables the header only was output. Now all such tables are ignored by default. -Note: RMF Monitor III table headers only contain control information such as header length, total record length, and the number of entries, entry length, and table offset for triplet groups. So there is no loss of useful analysis data by excluding headers only from the RMFBSAM file. It is the entries themselves that contain relevant information for reporting. -Because headers only for tables with zero entries are no longer output, the ZEROENC/NOZEROENC parameters (and their aliases) for ENC table processing are now OBSOLETE. If coded they will be accepted without error, but will be ignored. -ZEROENC/NOZEROENC will no longer be displayed in filters message RMFV006I. -New parameters SKIPTAB/NOSKIPTAB are now supported. NOSKIPTAB is the default and says to ignore any tables with zero entries. This is the behavior of prior ASMRMFV versions. SKIPTAB says to output these as skipped records. SKIPTAB is intended for possible diagnostic use only and the default NOSKIPTAB should be adequate for all users.
-Note: SKIPTAB/NOSKIPTAB applies only to the following RMF Monitor III tables: ASI, CFI, CPD, CSR, DVT, ENC, ENT, OPD, SPG. -When the POLICY option (alias POL) is in effect a second RMFV027I message will now be produced showing the SRM coefficients for CPU, IOC, MSO, and SRB for that Service Policy. -Error message RMFV007S had incorrect values for DDNAME and Reason Code when a BLDVRP failure occurred. -ASMRMFV prologue source documentation has been updated for the new SKIPTAB/NOSKIPTAB parameters, the obsolete ZEROENC/NOZEROENC parameters, and has more information on the checks used to determine skipped records. -REQUIREMENT: In order to receive these improvements 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. There are NO CHANGES to VMACRMFV for this enhancement. Thanks to Rodger Foreman, Trans Union, USA
VMAC42 were on GMT rather than local, and there is no GMT offset Apr 26, 2013 in that subtype, but SMFTIME was used to derive and use the offset to convert those datetimes to local zone. This caused me to examine and find that all of the TYPE42 datasets with STARTIME and ENDTIME also had GMT values, but these datasets were corrected in this change: 42D1 42D2 42D3 42D4 42EX 42L1 42L2 42S1 42S2 42S3 42S4 42SR 42VT 42X1 42X2 42X3 42X4. Either no one uses these datasets, or they recognized the SMFTIME was local and could visually convert! Thanks to Paul Naddeo, Fiserv, USA. Thanks to Edward Petka, Fiserv, USA.
VMACTMS5 that was NOT repeatable here with either z/OS or Windows, ANAL30 because this was an NLSCOMPATMODE issue for non-US site. ANALCPUV The $ is a variant character and was the culprit, but it ANALDB2P turns out that the $ is NOT required in the FORMAT NAME, ANALZPCR since the PROC FORMAT sets the CHAR/NUM nature, so all of GRAFLPAR the PROC FORMATS that create 'on the fly' formats were TIMEBILD revised to remove the $ character from variable FMTNAME. TRNDTALO NOTE: ALL NLS ISSUES ARE ELIMINATED IF YOU USE THE NEW VFMT102 // EXEC SAS,CONFIG='MXG.SOURCLIB(CONFIMXG)' VFMT102 documented in Change 27.356 and recommended in INSTALL, Apr 26, 2013 or if CONFIGN9 (instead of CONFIGV9) is used in your site MXGSASV9 JCL, as CONFIGN9 does NOT change the site's NLSCOMPATMODE option value. Thanks to Frank Lund, Evry AS, NORWAY. Thanks to Rich Anderson, SAS Technical Support, USA. Change 31.077 z/VM 6.2 PRCMFC 5.13 (SMF 113-like) record has 384 extra VMACVMXA bytes added that exposed BROKEN CONTROL RECORD ERROR, due Apr 27, 2013 to MXG code (1664 bytes read for the 208 counters, but 1728 subtracted). No new counters exist (yet?) in z/OS SMF 113 so these new bytes are (presumably) for the future, but I will confirm with IBM VM support and update this note if there is another explanation. Thanks to David J. Schumann, BCBS of Minnesota, USA. Change 31.076 Support for USER CICS fields INFRASTR CSGNTAIP FDRTRRCP IMACICUY and FDRICPTI. IMACICUZ IMACICV1
IMACICV2 UTILEXCL
VMAC110 Apr 25, 2013 Thanks to Mayer Rosenthal, WIPRO, USA. Thanks to Jeff Fracas, WIPRO, USA. Change 31.075 ASCII execution only. Variables RTYPE and RRTYPE were VMAC110 incorrectly input as $CHAR instead of $EBCDIC in current Apr 23, 2013 CICS versions' code. Older versions were correct. And there was NO error if you used UTILEXCL to create a tailored IMACEXCL; that code did use $EBCDIC. Thanks to Erling Andersen, SMTDATA, DENMARK. Change 31.074 Dataset ZRBLCP had incorrect values for these variables VMACRMFV LCPUPOLR LCPUCHIN LCPUCHIX due to misaligned INPUT. Apr 23, 2013 Thanks to Sterling James, DST Systems, USA. Change 31.073 Variable TYPETASK added to ASUMSMFI and TRNDSMFI ASUMSMFI datasets. TRNDSMFI Apr 28, 2013 Thanks to Frank Lund, Evry AS, NORWAY.
FORMATS 16 GB/S Link Type is now decoded. Apr 23, 2013 Thanks to "Bustia REPORTING", La CAIXA, SPAIN. Change 31.071 Cosmetic. PAGEBY SYSTEM added so each system starts on a ANALID new page. Apr 23, 2013 Thanks to Jennifer D. Ayers, West Virginia State Government, USA. Change 31.070 PMSTA01 Report. Cosmetic. If DB2STATS had only one obs, ANALDB2R and always for the last observation, page headers for Apr 19, 2013 the report were suppressed.
Apr 28, 2013 Change 31.068 See Change 31.018. This is IDMS EXIT code, contributed by ADOCIDMP Blake Leggett, Pace Applied Technology, that creates the ASMIDMPA SMF records with IDMS resources that are processed with ASMIDMPJ TYPEIDMP/TYPSIDMP MXG programs. ASMIDMPS -See ADOCIDMP for details for installing the exit. ASMIDMPU -The assembly source members for the Account Information JCLIDMPA exit, Journal Exit, Task Statistics Exit, and User Exit JCLIDMPJ for IDMS are now included. JCLIDMPS -Sample JCL members for these exits are provided with JCLIDMPU documentation on what each exit does and how to install Apr 13, 2013 them. -NOTE: Some tailoring of the sample JCL is required for data set names particular to the user installation. -NOTE: These exits are intended for use with IDMS V12.0 and up. Thanks to Blake Leggett, PACE Applied Technology, Inc., USA. CHANGE 31.067 A new way to map DBID/OBID to their database and object ANALDB2R names by first using JCLSQLID to extract those names at JCLSQLID a point-in-time, from SYSIBM.SYSTABLESPACE which has the VFMT102 complete mappings, rather than reading 105/107 IFCIDS Apr 13, 2013 that has been problematic, and then the POINTIME= option added in ANALDB2R POINTINTIME= can be used to read that dataset created by JCLSQLID: -POINTINTIME=YES reads JCLSQLID output from a static //OBID DD DSNAME= you provide. -POINTINTIME=DSNAME allocates DSNAME to FILENAME OBID and reads that file of JCLSQLID output. -POINTINTIME=NO reverts to the prior technique using the IFCID 105/107 records, but neither technique is 100% perfect as mappings can be changed after the snapshot was taken. This method has been more complete in mapping.
VMAC7072 -Up to 256 engines are supported, but no new variables are Apr 12, 2013 kept in TYPE70. Individual engine data for all engines are output in TYPE70EN dataset.
ANALDBTR ERROR: NO BY STATEMENT USED OR NO BY VARIABLES EXISTED. Apr 11, 2013 The substitution style macros all specify &PDBMXG as the input DD to be used in the sorts (that is the MXG default value) but ANALDBTR and ANALDB2R could be pointing at other DDNAMEs including WORK. This change sets the value of &PDBMXG to the PDB= parameter that was specified when ANALDBTR is called, after storing the original value so it can be restored at the end of ANALDBTR.
ANALACTM after SAS 9.3, but it does work in SAS 9.4, and it will Apr 11, 2013 likely continue to execute in future releases, as it is Jun 15, 2013 not actually being removed, but is no longer documented and is no longer supported if errors occur. With SAS 9.4 this note is printed on the SAS log, but it still works: NOTE: DSGI will no longer be supported after SAS 9.3. Also, option CHART=NONE will suppress the DSGI code. And, the Graphic portion of the report does require that you have SAS/GRAPH - that code automatically skipped if we see that you do NOT have SAS/GRAPH, but the formatted hierarchical report of your WLM policy is still created. Change 31.063 Cosmetic. Page spacing for Header 2 was different than ANALRMFR Header 1; the extra // in Header 2 was removed. Apr 10, 2013 Thanks to Jennifer D. Ayers, West Virginia State Government, USA. Change 31.062 -RMF III Enhancements and Notes ASMRMFV -The RMF Monitor III Processor Data Control Block (CPU) VMACRMFV and CPC Data Control Block (CPCDB) tables are now handled Apr 11, 2013 as two distinct tables just as they are documented in the RMF Programmer's Guide. Before they were logically handled as one table (CPU). -The CPCDB table will appear as a new table entry CPC in ASMRMFV RMFV105I messages. Each LPAR entry is counted as an entry in the output. -Note: The CPU table generally contains physical processor data while the CPC table contains LPAR and logical processor data. CPU table data is output to the ZRBCPU dataset while CPC table data is output to the ZRBLCP dataset by VMACRMFV. The number and content of variables in those datasets is unchanged from prior releases with the exception of 2 minor variables CPUCPOFF and CPUTOTLN in the ZRBCPU file. -The CPU and new CPC table selection parameters in ASMRMFV are co-dependent because the tables are so closely related. When one is selected, so is the other. For example, the CPU parameter when coded also selects The CPC table. Conversely, when one is deselected, so is the other. For example, NOCPU when coded will also result in NOCPC. Thus you do not need to alter any existing CPU/NOCPU parameter settings you may currently have to include or exclude the CPC. The ALL parameter (default) also selects both tables. -When the CPC table is selected in ASMRMFV the LPAR and Logical Processor entries are now blocked up to use as much as possible of a 32K logical record. The much shorter CPU table is unblocked. -ASMRMFV/VMACRMFV will now support up to 680 Logical Processors per LPAR (based on current section data lengths) before the entire CPC table is skipped. -VMACRMFV is updated to support the CPCDB table and blocking of its LPAR/Logical Processor sections. -ASMRMFV will now show the Average LRECL output for each RMF Monitor III table in detail and summary RMFV105I messages as a potential aid to sizing the RMFBSAM file. -Redundant code in ASMRMFV to skip various tables when invalid lengths or errors are encountered has been consolidated into a single internal subroutine. -ASMRMFV prologue source documentation has been updated on CPC/NOCPC parameters, the co-dependency with CPU/NOCPU parameters, and other items related to this change. -REQUIREMENT: In order to receive the performance benefits from CPC entry blocking 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. -Tutorial: This item has been mentioned before, but with the CPC table blocking the use of the EXZRBLCP exit is affected. Your tailoring logic in EXdddddd dataset exits to control output of an MXG dataset needs this structure to always be safe:
OUTPUT _Wdddddd; END;
because when "something" is true, they stop the read of this current record, skipping any un-read segments from being tested for "something". This consideration applies in particular to the following RMF III tables that have blocked input data to a PDB build and to their respective output exits:
------------- ----------- ASI EXZRBASI CPC EXZRBLCP CPD EXZRBCPD CSR EXZRBCSR DVT EXZRBDVT ENC EXZRBENC ENT EXZRBENT OPD EXZRBOPD RED EXZRBRED SHD EXZRBSHD SPG EXZRBSPG UWD EXZRBUWD Change 31.061 New MXGACC03 report uses ANALCNCR to create average and ANALDB2R maximum concurrent threads/DBATs for each chosen INTERVAL Apr 4, 2013 for each Subsystem and AUTHID, tabulated by QWHCATYP, the Attachment Type (TSO,CICS,DDR...), with a tabular report. You can tailor using existing ANALDB2R report parameters to select data, or all DB2ACCT data will be reported. BUT IT IS NOT POSSIBLE TO COUNT THREADS/DBATS WITH ROLLUP DB2PARTY='R' ACCUMAC DB2 Accounting Records. If five one-second transactions ran sequentially from minute 1 to minute 6, their rollup record would report five minutes with QLACRLNU=5 concurrently active when there was one. We could calculate an average of 1 in this case, but if the first minute was a peak of 50 and there were then four minutes of 1, we would calculate an average of 11. Thanks to Paul Walters, Navy Federal Credit Union, USA.
IMACCADI and PDB.PRINT, CADIDPLX CADIJDEI CADIJDLI CADIATYP, were VMAC6 all wrong and equal to CADIXFRM, variable CADIACCT now Apr 4, 2013 has nulls converted to blanks, and new variable CADIJBID contains the JCTJOBID from the CADI segment in SMF 6. Thanks to Randy Schlueter, First Data Corporation, USA. CHANGE 31.059 -The System Parameter Report produced many pages, six ANALDB2R pages per subsystem per interval generated. The report Apr 3, 2013 did not anticipate an IFCID=106 record written every 15 minutes. Now, only the first observation for each subsystem, or the first record after a restart will be printed by PMSPR01=YES. -Cosmetic. Page Number printing in DB2PM-like reports did not align when there were five digits of page number. Thanks to Jennifer D. Ayers, West Virginia State Government, USA.
VMACNMON but now it is read into variables BBBP0049-BBBP0056. Mar 29, 2013 -Records SEA, SEAPACKET, DONATE are supported. Thanks to Lennon L. Merchant, Coca-Cola, USA. Change 31.057 -When variable STARTIME was in SORTBY= argument, some of ASUM4HRS the xxxx_4HRS_AVG variables were zero or missing for the ANAL4HRS first four hours of each subsequent day. Comments that VGETFMT tried to communicate "don't use STARTIME in SUMBY=" Mar 27, 2013 SORTBY- VARIABLES BY WHICH THE DATASET IS SORTED NOT INCLUDING THE STARTIME (ASSUMED TO BE THE LAST VARIABLE IN THE SORT) were revised to be clearer. -The prior requirement for the input data to be pre-sorted is removed; ASUM4HRS will do the sort in its processing. -ANAL4HRS was an accident from an early ASUM4HRS iteration that shouldn't have been kept; now it is only comments to use ASUM4HRS. -Argument NONOTES was not UPCASED and was not recognized when the argument was typed in low case, now corrected. Thanks to Michael Marcus, FUJITSU, USA. Change 31.056 Variable THREADS was created only in ASUMDB2A and used in ASUMDB2A ANALDB2R, by setting THREADS=1 for each DB2ACCT obs, but VMACDB2 for RollUp DB2PARTY='R' records, THREADS=QLACRLNU; is now Mar 26, 2013 used to count the number of threads in each rollup. And variable THREADS is now similarly created in DB2ACCT. Thanks to Robb Hermes, Sentry Insurance, USA. Change 31.055 Cosmetic. Variable CIFROM is formatted $MGNDMNT and VMACNDM variable NDMNODET='FLAG1*FIELD' is formatted $HEX2. and Mar 29, 2013 re-labeled correctly. -Variable NDMRTYPE now kept in NDMGO as multiple subtypes are output into that dataset. Thanks to Michael Oujesky, Bank of America, USA. Change 31.054 Corrections for '1031' record to populate JOBNAME and VMACPRPR FIELDNAME. Mar 24, 2013 Corrections for '1061' record, to use the original INPUT. Apr 11, 2013 CUST_ACC changed to character variable as it must be. Thanks to Geert Debatselier, KBC, BELGIUM
ANALDB2R
Mar 24, 2013 Change 31.052 IIS Weblog decoding URIQUERY than ended with &=1 with no VMACWWW URIQNAME before the equal sign caused INVALID SUBSTR note Mar 22, 2013 that is now eliminated. Thanks to Dennis Longnecker, State of Washington Courts, USA. Change 31.051 -Support for user field CHARGE creates variable USCHARGE IMACICUX in CICSTRAN when comment block in IMACICUX is removed. UTILEXCL -Change 30.283 (30.30/30.01) "***UNKNOWN FIELD" protection VMAC110 was still insufficient, and the IMACEXCL created failed VMACSMF with a 180 syntax error due to missing @;. Circumvention Mar 22, 2013 logic updated yet again, hopefully, but sending the full UTILEXCL log is all that is needed for MXG to add UNKNOWN fields optionally to your CICSTRAN dataset. Thanks to Mayer Rosenthal, WIPRO, USA. Thanks to Denise Willers, WIPRO, USA. Thanks to Jeff Fracas, WIPRO, USA.
VMACSMF new-in-30.30 VMACSMF code to get ID and SUBSYSTEM during Mar 19, 2013 header processing was complicated by a subtype 0 record that was actually a subtype 1, so the correction code in VMAC110 was used in the VMACSMF processing logic. Thanks to Rudolf Sauer, T-Systems, GERMANY. Change 31.049 Support for WAS XD (WebSphere Extended Deployment) under EXXDFGLG z/linux under z/VM reads 6 virtualization log files - see EXXDNSLG IBM 2007 "Best Practices" Red Bok SG24-7343 - to create: EXXDSPLG
EXXDSSLG DDDDDD DATASET Description EXXDTILG
EXXDTSLG XDNSLG WASXDNS Node Statistics Historic Cache IMACXDFG XDSPLG WASXDSP Server Power Consumption Stats Cache IMACXDNS XDSSLG WASXDSS Server Stats Cache IMACXDSP XDTILG WASXDTI TC Module Instance Stats Cache IMACXDSS XDTSLG WASXDTS TC Module Stats Cache IMACXDTI
IMACXDTS and this dataset which IBM recommends for Chargeback TYPEXDFG
TYPEXDNS XDFGLG WASXDFG Fine Grained Power Consumption Stats TYPEXDSP
TYPEXDSS These MXG variables are created in "Fine Grained" data: TYPEXDTI
TYPEXDTS APPNAME APPLICATION*NAME TYPSXDFG BEGINTIME INTERVAL*BEGIN*DATETIME TYPSXDNS CELL CELL*NAME TYPSXDSP CELLPOWER TOTAL*CELL*POWER*PER SECOND TYPSXDSS CELLWORKPOTENTIAL MAXIMUM*CELLPOWER*PER*INTERVAL TYPSXDTI CLUSTER CLUSTER*NAME TYPSXDTS ENDTIME INTERVAL*END*DATETIME VMACXDFG GWID GATEWAY*ID VMACXDNS INTRVLTM INTERVAL*DURATION VMACXDSP MODULENAME J2EE*MODULE*NAME VMACXDSS NODE NODE*NAME VMACXDTI NODEGROUP NODE*GROUP*NAME VMACXDTS NODEPOWER TOTAL*NODE*POWER*PER SECOND VMXGINIT NODEWORKPOTENTIAL MAXIMUM*NODEPOWER*PER*INTERVAL Mar 31, 2013 NUMSERVICED REQUESTS*SERVICED*OF THIS*TYPE ODR ON*DEMAND*ROUTER*NAME POWERCONSUMED POWER*CONSUMPTION*RATE*PER SECOND SERVER SERVER*NAME SERVICEPOLICY SERVICE*POLICY*NAME TCMODNAME TRAN CLASS*APPLICATION*MODULE*NAME TCNAME TRANSACTION*CLASS*NAME TIMESTAMP WAS XD FG*DATETIME*STAMP WORKCOMPLETED WORK*COMPLETED WORKFACTOR ESTIMATED*WORK*FACTOR ZDATE ZEE DATE*ZEE OBS*WAS CREATED Note: This new code will not execute under SAS 9.1.3; the 'M' operand to the SCAN function required when there are duplicated delimiters, did not exist back then. Thanks to Sally Jordan, Insurance Services Office, Inc., USA.
VMAC110 because I inserted a "cosmetic" comment documenting the Yüklə 28,67 Mb. Dostları ilə paylaş:
|