* copyright (C) 1984-2019 merrill consultants dallas texas usa
Yüklə 28,67 Mb.
|
|
CICDQR 43 TDQUEUE Transient Data Specific STITDQR DQR A10 CICDS 57,5 Dispatcher Domain, CPU each TCB STIDS DS DSG CICDTB 33 Dynamic Transaction Backout Global STIDTB DTB A05 CICFCR 67 File Control Specific STIFCR FCR A17 CICFEPIP 16 FEPI (pool) STIFEPIP FEP A22 CICFEPIC 17 FEPI (connection) STIFEPIC FEC A23 CICFEPIT 18 FEPI (target) STIFEPIT FET A24 CICIRCB 75 IRC Batch Global STIIRCB IRC A19 CICJCR 49 Journal Control Specific STIJCR JCR A13 CICLDG 27,3 Loader Domain for Program Global STILDG LDG LDG CICLDR 25 Loader Domain for Program Specific STILDR LDR LDR CICLSRFR 40 LSRPOOL File stats each File STILSRFR LS3 A09 CICLSRR 37,3 LSRPOOL Pool stats each LSR pool STILSRR LS1 A08 CICM 81 Monitor Domain Global STIM M MNG CICPAUTO 23 Autoinstall Program STIPGG PGG PGG CICSDG 90 System Dump Global STISDG SDG SDG CICSDR 88 System Dump Specific STISDR SDR SDR CICSMD 7,5 Storage Manager Domain Subpool STISMD SMD SMD CICSMDSA 9,2 Storage Manager DSA and EDSA STISMDSA SMS SMS CICSMT 8,6 Storage Manager Task Subpool STISMT SMT SMT CICST 66 Statistics Domain Global STIST ST STG CICTC* 3 Task Control Global STITC TC A01 CICTCLR* 58 TCLASS Transaction Class Specific STITCLR TC1 A15 CICTCR 34 Terminal Control Specific STITCR TCR A06 CICTDG 87 Transaction Dump Global STIDTG TDG TDG CICTDR 85 Transaction Dump Specific STITDR TDR TDR CICTM 63 Table Manager Global STITM TM A16 CICTSQ 48 TSQUEUE Temporary Storage Global STITSQ TSQ A12 CICTSR* 4 Transaction Statistics Specific STITSR TSR A02 CICUSG 61 User Domain Statistics STIUSG USG USG CICVT 21 VTAM Global STIVT VT A03 CICXMC 12 Transaction Manager TCLASS STIXMC XMC XMC CICXMG 10 Transaction Manager Global STIXMG XMG XMG CICXMR 11 Transaction Manager Transaction STIXMR XMR XMR Change 12.165 Support for LEGENT's TSO/MON 6.1 adds Service Class Name VMACTSOM to TSOMCALL and TSOMSYST datasets when MVS/5.1 is run in Sep 28, 1994 Goal Mode, and adds a number of Workload Manager fields to TSOMSYST dataset. There can be up to ten sets of the Service Class fields, but MXG chooses to keep only the first two sets by default (others can easily be added in MACRO _KTSOSYS in IMACTSOM, if really needed!). The real significant change in 6.1 is internal; the transaction distribution counters were expanded from 1 to 2 bytes (so you can have a larger interval and not overflow!). Thanks to LEGENT for lots of lead time, and test data too!!! Thanks to Arne Lund, LEGENT TSO/MON Development, DENMARK. Change 12.164 Omegamon CICS V551 SMF user dataset OMCIVSAM has zero obs VMACOMCI because 2 extra bytes were added between V551 and V550, Sep 28, 1994 but also because MXG logic error. The MXG logic error: After the IF SUBTYPE=103 and RECSUBTY=1 THEN DO; line, change IF VMAXF GT 1 THEN DO _I_=1 TO VMAXF-1; to read IF VUSEDF GT 1 THEN DO _I_=1 TO VUSEDF-1; The extra 2 bytes: Change the IF statement preceding the INPUT ESRES1 statement from ELSE IF SUBTYPE EQ 102 THEN to read ELSE IF SUBTYPE EQ 102 OR SUBTYPE EQ 103 THEN Unrelatedly, variables EGENAPL ERCDNUM and ESTTIME were removed from the KEEP= for datasets from subtypes 102 and 103, as those variables exist only in the subtype 100, (and never should have been in the KEEP= list!). Thanks to Scott Pearson, King County Medical Blue Shield, USA. Thanks to Shellia Shih, King County Medical Blue Shield, USA. Change 12.163 Requesting CICS reports with PDB=PDB,TYPE=ALL,SUMMARY=YES ANALCISH produced 0 observations and no reports because of logic Sep 27, 1994 errors for this combination. Specifying SUMMARY=NO will produce all reports except summary. Many lines changed. Thanks to Tom Albert, John Alden Life, USA.
ANALTMDB this significant user contribution; all of the code was EXTMDDA written by Peter, and it worked correctly "right out of EXTMDDB the box". Not only does his contribution process the EXTMDDE Landmark file (from INFILE TMDBIN) to create 4 datasets: EXTMDDR Dataset Description FORMATS TMDBDA Interval Statistics IMACTMDB TMDBDB Thread Detail TYPETMDB TMDBDE Exception VMACTMDB TMDBDR Thread Resource Sep 25, 1994 but also Peter provided 17 report programs that are in member ANALTMDB! Thanks to Peter Proppe, Bremer Lagerhaus Gesellschaft, GERMANY.
ACHAP99
Sep 25, 1994 This chapter cites and thanks all contributors to MXG Software, who are hereby officially awarded the title of "Chapter 99 CodeSharks"! Named by Judith S. "99" Merrill, Vice President and Partner, in honor of the diligent sleuthing of the MXG code performed by each CodeShark, she also established our policy that credit should be given to every MXG user who made any contribution to MXG Software, whether they provided an entire program, or found a programming error, or offered a suggestion, or even just found a comma out of place in a comment! The MXG Change Log (member CHANGESS) follows each change text with the "Thanks to" note that acknowledges the user that precipitated that Change. Chapter 99 extracts those notes, lists the major contributors first, and then lists (alphabetic by last name) every single user that has ever precipitated a change! We sincerely thank you and salute you all! Change 12.160 Support for STK's ICEBERG SMF record subtype 5 adds new FORMATS dataset ICEBRGDE for Deleted Data Space Release. The VMACICE statistics for the first five extents released are kept Sep 24, 1994 individually, as is the total I/O time for all extents. Thanks to Diane Eppestine, Southwestern Bell, USA. Thanks to Mike Schweitzer, StorageTek, USA. Change 12.159 MXG 12.01 thru MXG 12.03A. Variables QB3Taaaa/QB4Taaaa VMACDB2 in datasets DB2STAT1/DB2STATS are still missing because Sep 22, 1994 typos crept in with Change 12.126. The statement reading IF 2 LE QBSTPID EQ 49 THEN DO; should read ELSE IF 2 LE QBSTPID LE 49 THEN DO; and the statement reading ELSE IF 80 LE QBSTPID EQ 89 THEN DO; should read ELSE IF 80 LE QBSTPID LE 89 THEN DO; Thanks to Siegfried Trantes, Gothaer Versicherungsbank, GERMANY.
VMAC1415 you install this PTF and do not have MXG 12.01 or later, Sep 22, 1994 the variables at the end of the type 14/15 record will be Sep 27, 1994 bad! The variables corrupted are DEVNR, VOLSER, UCBTYPx, DEVCLASS, DEVTYPE, NEXTENT, EXCPCNT, TRKSALOC,and all of the HIPERBATCH statistics. Change 12.036 added support for the new field (inserted in the middle of the record), but I failed to note the incompatibility and failed to provide a temporary fix (you can install this fix before the PTF is installed for protection, although if you are going to use TYPE1415 you really should request the pre-release as the actual change is more extensive). After TEMPLOC=265+OFFSMF; Insert IF SMF14SDC GE 28 THEN TEMPLOC=TEMPLOC+SMF14SDC-24; -When IBM added the Hiperbatch fields, they did not set a bit to indicate the 20 bytes were present, so MXG could only test if there were 20 bytes, and if present, then INPUT the hiperbatch fields. Unfortunately, IBM APAR UW05634 added 30 bytes of SMS data to the end of the record, and a record with SMS but no Hiperbatch satisfied the MXG test, so false values were read in the Hiperbatch variables (SMFIOREQ,SMFCHITS,SMFNMWTS,SMFPHIOS,SMFCIOS). A new APAR, OW08246, now sets RECIND2='....1...'B when hiperbatch is present, but you may not have that APAR on, so this change tests both length and/or RECIND2 to read in Hiperbatch data. A separate bit, RECIND2='.....1..'B is set by UW06888 when SMS data exists, and that bit is used to decide to input the new SMS fields of MGMTCLAS, DATACLAS and STORCLAS. Until you install MXG 12.04 or later, you can correct the invalid Hiperbatch values that are caused by UW06888 by replacing the line now reading: IF LENGTH-COL+1 GE 20 AND NUCB=1 THEN with these two lines LENLEFT=LENGTH-COL+1; IF LENLEFT=20 OR LENLEFT=50 OR RECIND2='....1...'B THEN The actual change was much more extensive, as UW05634 or UW06888 or maybe some other as-yet-unavailable APAR also adds statistics for Compressed Format data; these new variables now exist in TYPE1415 as missing and will be populated when filled in by IBM: SMF14CDL SMF14CDS SMF14CIS SMF14TKL SMF14TKN SMF14UDL SMF14UDS SMF14XF1 SMF14XF2 Thanks to Paul Dzielak, Smith, Barney, Harris, Upham & Co., USA. Thanks to John Maher, Home Savings of America, USA. Thanks to Lawrence Jermyn, Fidelity Systems, USA. Thanks to Jeff Berger, IBM, USA. Thanks to Don Fink, IBM, USA.
EXDB2ACB supposed to have zero observations by default, because it Sep 21, 1994 can be very large (and since it is summarized in DB2ACCT, I don't think most sites need the additional detail.). This change added the comment block (that was supposed to have been there all along) in member EXDB2ACB around the OUTPUT _LDB2ACB; statement so there now will be zero observations created unless you change EXDB2ACB. Also, misleading comments in member IMACDB2 were changed, and the text of Change 12.033 was revised to be correct! Change 12.156 Landmark MONITASK variables MONICISP/MONICASP were blank TYPEMON8 when the transaction in fact had CA/CI splits, because TYPETMON the names were reused for the MONIDBD dataset. To fix: Sep 21, 1994 find the tests for TAFLAG6 that set the variables (four lines) and respell MONICISP as MONACISP and MONICASP as MONACASP, and then immediately before the %%INCLUDE SOURCLIB(EXMONTSK); statement insert two lines: MONICISP=MONACISP; MONICASP=MONACASP; to pick up the temporarily stored values. Thanks to Brent Holm, Blue Cross and Blue Shield of Oregon, USA.
VMACOMCI Now it has been made in the source library. Sep 21, 1994 Thanks to Freddie Arie, Lone Star Gas, TEXAS. Change 12.154 TYPE37 short LAND segment caused INPUT STATEMENT EXCEEDED VMAC37 error. Between the lines inputting BRFSMADR and BRFSMNAM Sep 2, 1994 insert: @; IF LENLAND GE 64 THEN INPUT Thanks to Phillip Wong, Orient Overseas Container Line, HONG KONG. Change 12.153 Step TESTOTHR may abend with 213-04 for //DISK DDname if JCLTEST6 your site sends temporary allocations to VIO; that DDname Aug 30, 1994 is used by VMXGVTOC and must point to real DASD, so the DD statement now has VOL=SER=XXXXXX and comments telling you to provide one of your volume serial numbers. Thanks to Willie Antman, FDIC, USA. Change 12.152 TYPE72 and TYPE72GO datasets were enhanced with four new VMAC7072 variables: Aug 30, 1994 MPL72 - RESIDENT*MULTI*PROGRAMMING*LEVEL CSTORE72 - CSTORE*BYTES ESTORE72 - ESTORE*BYTES TSTORE72 - TOTAL*CSTORE+ESTORE*BYTES Thanks to Willie Antman, FDIC, USA.
IMACTMON completely incompatible; reading the new records with the TYPETMON old member TYPEMON8 causes INVALID DATA FOR ENDDATE, D12, Aug 29, 1994 D34, D56, HH, MM, SS, TH and MXG ERROR2.LANDMARK. The record architecture was changed and only 3 original data sets are now created (MONITASK, with 70 fewer variables, MONISYST with 119 fewer variables, and MONIDBDS with 2 new variables), plus one new dataset (MONIDSA), so I chose to change the member name from TYPEMON8 to TYPETMON (and the corresponding IMAC from IMACMONI to IMACTMON) so that I could reduce the size of the datasets created. If you have installed the TMON de-compression exit from member EXITMON6, you must update IMACTMON to tell MXG to use that exit as described in comments in IMACTMON. These datasets are no longer created by TYPETMON: MONIHIST - History record, actually went away in 1.0 MONIMRO - MRO segments no longer exist in TA record MONIUSER - User segments no longer exist in TA record There is one new dataset, from the "TD" record: MONIDSA - Interval DSA Statistics and only these new fields were added to existing datasets MONITASK - TAENGMTO TADBCTHC MONISYST - TIDBCTC TIDBTHC MONIDBDS - *** unknown, 8 bytes undocumented. In fact, the major changes were the removal of many, many variables: all of the non-ESA constructs are now reserved and I chose to delete those old variables in this support (because dragging them along made the MONITASK/MONISYST datasets much larger for no good purpose). However, I am sure that some of your report programs may fail with VARIABLE NOT FOUND errors because of these deletions, so you will need a little extra review to remove references to these deleted variables. Member DOCVER12 identifies all of the variables that have been deleted, but note that variables PRIINCHR/PRIOUCHR (Input/Output character counts) do not exist in MONITASK, and duration variables JCCPUTM/JCDSPTM/KCCPUTM/KCDSPTM/TCCPUTM/TCDSPTM IRCPUTM/IRDSPTM/OSWTCPTM/SYSTCBTM/VSDSPTM no longer exist in MONISYST.
VMAC102 VARIABLE EXCEEDS 200 BYTES because the LABEL for variable Aug 29, 1994 QW0206FL ended with a */ instead of a single quote. Thanks to Marlin W. Thoma, J.C. Penny, USA. Change 12.149 AS/400 dataset QAPMLTG does not exist, and all references VMACQAPM to it have been deleted. Aug 26, 1994
VMAC50 because the %%INCLUDE SOURCLIB(EXTY50); statement was Aug 26, 1994 misplaced. It and its adjacent RETURN; statement need to be moved after the END; statement in line 012600 (i.e., after the penultimate END; and before the ultimate END;). Thanks to Siegfried Trantes, Gothaer Versicherungsbank, GERMANY. Change 12.147 Omegamon II for SMS Change 12.123 (MXG 12.03/12.03A) had VMACOMSM a typo which caused execution errors. The "&PIB.2 " after Aug 25, 1994 OMFS2INT should be "&PIB.2.". In addition, that Change did not support continued records, but this change added substantial logic to process continuations. A complete logical record contains a VOLUME segment, some number of DATASET segments, each followed by some other number of JOBS segments for jobs using that DATASET on that VOLUME. If the total length of these segments exceeds 32768, the SMF record is truncated at the end of either a DATASET or JOBS segment, and subsequent continued records start with the next JOBS/DATASET segment. Supporting these records required retaining the data from the VOLUME and last DATASET segment so the segments in continued records have the volume and dataset information. There is an exposure to data loss with Candle's architecture - if the first record of a continuation is the last record in today's SMF file, then when those continuations are processed tomorrow, there won't be a volume or dataset record and those data will be missing. I have suggested that Candle change their design, so that when they have to continue segments, that they put the previous VOLUME and DATASET segments in each record so that each record is a self- contained description of activity by jobs against dataset on volumes. Thanks to Bob Lembo, American Stores Company, USA. Change 12.146 Reading the NDM VSAM log (instead of the dumped BSAM log) TYPENDML caused zero observations, because there are ten extra Aug 25, 1994 bytes at the beginning of the VSAM record that are not in the VSAM record. To process the VSAM NDM log, change the statement OFFSMF=-14; to read OFFSMF=-4; Comments were added to TYPENDML documenting this feature. Thanks to Jane Chin, Chevron Information Technology, USA.
EX028CM8 (64x-6Fx), nine new segments (NCV,NGV,NLS,NLW,NRV,NVC, EX028ER3 NVS,NVT, and NVV), and MXG creates nine new datasets for EX028INF these NetWare statistics (which look very useful): EX028ING NPMNWCMD - NETWARE START/STOP/ALTER EX028INH NPMNCINT - NETWARE COMMUNICATIONS INTERVAL EX028INI NPMNRINT - NETWARE ROUTER INTERVAL EX028INJ NPMNLINT - NETWARE LAN BOARD INTERVAL EX028SU1 NPMNLSUM - NETWARE LAN BOARD SUMMARY EX028SU2 NPMNGINT - NETWARE SERVER GLOBAL INTERVAL FORMATS NPMNVINT - NETWARE SERVER VOLUME INTERVAL VMAC28 NPMNVSUM - NETWARE SERVER VOLUME SUMMARY Aug 23, 1994 NPMNVMNE - NETWARE MONITOR EXCEPTION/RESOLUTION New variables were added to NPMINPMT (4), NPMVSVGB (4) and NPMVSVDV (22). All changes were compatibly made, although the new subtypes will cause an MXG message to be printed on the log. This support has only been syntax checked; no data has been available for testing as yet.
VMAC7072 should have been R723CSAC. Aug 23, 1994 Thanks to Tom Parker, Hogan Systems, USA. Change 12.143 Change 12.060 was supposed to have changed SACCT3 to read IMACPDB SACCT9, but that part of the change was not made; now it Aug 22, 1994 has been! I still don't know how this slipped thru! Thanks to Tom Parker, Hogan Systems, USA. Change 12.142 Tape Mount Pending duration (the duration from when the ADOC30 tape mount was issued until the volume is mounted and the Aug 22, 1994 VOLSER has been verified and the UCB Ready Bit is on) is included in the TYPE30_4 duration variable DSPDLYTM. I have erroneously reported (in documentation member ADOC30 and in the text of Change 10.031) that tape mount delay was included in ACTDLYTM, but it has always been a part of DSPDLYTM. The text in ADOC30 and CHANGESS has been corrected. Thanks to Roh Rhue, PENNCORP Financial, USA. Thanks to Bob Drischel, PENNCORP Financial, USA.
VMAC74 set TYPE74, and is needed for ANALRMFR reports. Aug 19, 1994
ADOCBLSR a technique for significantly reducing VSAM run times, is ANALBLSR now contained in member ADOCBLSR. Also, his BLSR analysis Aug 17, 1994 program, ANALBLSR, has been revised to tell you if it did not find any datasets that were candidates for LSR, or if candidates were found that did not meet the minimum criteria (EXCPs against the file, specified in IOHIGH, or percentage of total I/O to a single VSAM component, specified in PCTIO; both IOHIGH and PCTIO are defined in ANALBLSR and can be changed by the user). Thanks to Chuck Hopf, Merrill Consultants, USA.
VMAC7072 are not properly calculated nor completely understood. Aug 17, 1994 (These are sampled values for using/delay from the new Workload Manager when in Goal Mode, and while they may be useful, they are not critical, resource measurements.) First, I was using R723CSAC instead of R723MTVN as the number of samples in the denominator, but even with that correction, there are observations with PCTUSCUS (using CPU) of 249%! Second, some samples are summed across all address spaces, requiring additional normalization. As time ran out for building MXG 12.04, I will revisit the calculations and revise this text in a future release. For now, regard them as highly suspicious. Thanks to Don Deese, (CPExpert), Computer Management Sciences, USA. Change 12.138 Support for APAF 2.2 incompatibly added new data fields, VMACAPAF and now supports up to 12 LPs. Some new variables were Aug 17, 1994 added to several data sets. Kudos to Amdahl for giving me not only documentation but actual test SMF data! Thanks to Gladys Blaylock, Amdahl, USA.
VMAC74 list for datasets TYPE74CF and TYPE74ST and SYSPLEX was Aug 17, 1994 added to KEEP= list for TYPE74 (to support ANALRMFR).
"To Bruce, from Karen, y'know, the girl you dumped last weekend? Well, I just won the New York State Lottery! F... You!"
IMACTLMS with the previous 5.3 release) creates the same TYPETLMS TYPETLMS dataset, with a few new variables. You must change macro VMACTLMS _TLMSVER in member IMACTLMS from 5.4 to 5.3 if you want Aug 10, 1994 to process records from the old 5.3 release, as MXG's default now expects release 5.4 format records. Thanks to Mike Hill, Abbey Life, ENGLAND.
VMAC102 the iteration across each Index used to obtain RIDS was Aug 10, 1994 using the Length instead of the Number of sections! Change the BY QWT02R2L; to read BY QWT02R2N; in the line DO OFFSET=QWT02R2O TO QWT02R2O+(QWT02R2N*QWT02R2L) BY.... inside the MACRO _C102125. Thanks to Dave Witkowski, Amdahl, USA.
ANALRMFR You get "LIBRARY SMF IS NOT IN VALID FORMAT FOR ACCESS Aug 10, 1994 METHOD SASEB" error (because MXG is trying to read a SAS dataset from the SMF DDname!). To correct, insert %IF &PDB EQ SMF %THEN %LET PDB=WORK; before the line with the comment /* BEGIN CPU REPORTS */. Change 12.133 Some DDF Summary variables (QLSTxxxx in DB2STAT0) were DIFFDB2 not de-accumulated in DIFFDB2, causing incorrect counts Aug 10, 1994 (most notably, BYTES sent/received on the statistics summary report ANALDB2R). Change 10.210 removed some of these deaccumulations in error, and some were new fields added in DB2 3.1. Now, all QLSTxxxx variables, except for QLSTLOCN (a character variable) are now DIF()'ed in DIFFDB2.
Thanks to Don Mosley, Farmland Industries, USA. Change 12.132 BETA93 test records were received, and proved that the FORMATS initial support in MXG 12.03 was incomplete. Subtypes 0 Yüklə 28,67 Mb. Dostları ilə paylaş:
|