COPYRIGHT (C) 1984-2019 MERRILL CONSULTANTS DALLAS TEXAS USA
MXG NEWSLETTER THIRTY-SIX
MXG NEWSLETTER NUMBER THIRTY-SIX February 10, 2000
Technical Newsletter for Users of MXG : Merrill's Expanded Guide to CPE
TABLE OF CONTENTS Page
0. Summary of Y2K issues that were fixed in MXG 17.17. 3
I. MXG Software Version 17.17 contains this newsletter, no printing. 4
II. MXG Technical Notes 9
1. Compatibility issue for products with deaccumulated records. 9
2. Recent hardware timing measurements for moving files 9
3. An example of the use of IMACFILE/MACFILE exit: 10
4. Make sure your MVS REGION size is 64M, or you may get errors. 10
5. MXG Definitions with regard to MXG Software Changes 10
III. MVS Technical Notes 11
1. Typos in previous APAR numbers in NEWSLTRS/CHANGESS. 11
2. APAR OW36535 corrects Start/Stop in type 42 subtype 6. 11
3. Problems with ECSA creep are documented in APAR OW34249. 11
4. CA-SPOOL product APAR GS98190 corrects type 6 record errors. 11
5. How many volumes can you have in a multi-volume MVS dataset? 11
6. How does software compression compare with hardware compression? 11
7. How do you hardware compress a SAS dataset on OS/390? 12
8. APAR OW38615 for RACF type 80 corrects PERMIT record. 13
9. APAR OW37742 for Workload Manager "Goal Mode". 13
10. APAR AR38393 for Hardware Compression bypass compressed. 13
11. APAR OW39277 for VSAM record statistics in LISTCAT. 13
12. LOTUS Domino may need a 2GB region. 13
13. APAR OW37709, new BMF cache options, may help LOTUS Domino. 13
14. Shark/Parallel Access Volume hardware have several APARs. 13
15. IFASMFDP with //SYSPRINT to disk can lose data on I/O error. 13
17. VSAM Catalog reminder: they stop functioning 1/1/2000. 13
18. APAR OW32140 for WLM, CICS, reduces IRASASRV sampling rate. 13
19. APAR PQ25641 for Smart Batch reports possible data loss. 14
20. APAR PW30654 for TCP/IP have FTPEND earlier than FTPSTART. 14
21. APAR OW37263 needed for OpenEdition/USS type 30 accounting. 14
22. APAR OW39746 corrects TYPE74CF CPU busy exceeding interval. 14
23. APAR OW41239 corrects type 85 mounted duration for OAM. 14
24. APAR OW41169 corrects SHARK and type 42.6 activity counts. 14
25. APAR OW38842 causes blank OMVSEXPN in OMVS/USS TYPE30_4. 14
26. APAR PQ32322 for TCP/IP API has FFFFFFFFx for byte count. 14
27. APAR OW42559 for UCCOLDT in DCOLCAPD is invalid. 14
28. SHARK/PAV/FICON APARs to 2.8 require MXG 17.08. 14
29. The size of Hipervolumes on EMC boxes is in DCOLVOLS. 14
IV. DB2 Technical Notes. 14
1. Excess DB2 SMF type 101 records due to Query CP parallelism. 14
2. APAR PQ10864 consolidates child task accounting. 15
3. APAR PQ27561 for increase QTXACLUN after 5.1
4. The Sunrise product can corrupt DB2 101/102 timestamps.
V. IMS Technical Notes. 9
VI. SAS Technical Notes. 15
1. SAS 6.09 TS460 supports UCB's above the line, 0C4 fixed. 15
2. SAS Version 7 ABEND 913-1C fixed in SAS Version 8. 15
3. Options CODEPCT/BLKSIZE(TAPE) don't exist in Version 7. 15
4. Unknown Exception 80000602 due to bad SAS PROFILE. 15
5. WARNING removed in TS 465. 15
6. SMS "Immediate Release" can cause USER ABEND 315. 15
7. SAS V7 zap for SAS data library to exceed 32,767. 16
8. SAS V7 errors PROC SORT with Duplicate SORT KEY variable. 16
9. New PDJULI2 format and JULDATE7() function for Y2K help. 16
10. SAS V8 Developer Release VBS I/O error if '1A'x in data. 16
11. SAS V8 Developer Release has successfully run BUILDPDB. 16
12. Note on impact of YEARCUTOFF on two digit literal dates. 17
13. USER ABEND 0318 usually multi-vol, but other causes listed. 17
14. Format libraries built with V7/V8 cannot be used with V6. 17
15. Under Win NT, creating FORMATS fails with existing .SC2 file. 17
16. SAS V8 Production, Win NT reads only first S370VBS concatenate. 17
17. What level of MXG is needed for SAS Version 8 Release TS M0? 17
18. Variable Blocked SOURCLIB can be used, with some cautions. 17
VII. CICS Technical Notes. 18
1. CICS RCT parm TOKENE= replaced in CICS/TS by the CICS RDO. 18
2. MXG 17.04 or later required for CICS/TS 1.3. 19
VIII. Windows NT Technical Notes. 19
IX. Incompatibilities and Installation of MXG 17.17. 19
X. Online Documentation of MXG Software. 19
XI. Changes Log 19
Alphabetical list of important changes 20
Highlights of Changes 17.001 thru 17.395 20-24
COPYRIGHT (C) 2000 MERRILL CONSULTANTS DALLAS TEXAS USA
0. Summary of Y2K Issues after MXG 16.16 fixed in MXG 17.17:
a. Y2K Notes:
MXG Supports Leap Year in Year 2000 because mapping of dates from
SAS date/datetime values by SAS recognizes Feb 29, 2000 as leapday.
ANALDATE utility (MXG 17.07, Change 17.262) will examine all SAS
date/datetime variables in all datasets in a SAS data library
and reports what dates are found.
Outstanding APAR OW42559 for DCOLLECT variable UCCOLDT in DCOLCAPD
dataset, has no PTF; date wrong, records lost, but IBM will fix it.
b. MXG Software Changes for Y2K Ready after MXG 16.16 (in MXG 17.10+):
Change 17.363 (MXG 17.10): Julian 0cyyddd were changed to 0cyyddd in
products that had not been Y2K tested by anyone before. See text for
the list of products that were changed and those still not tested.
Change 17.352 (MXG 17.10): TYPETMS5 variable OUTDATE was still 0cyyddd.
Change 17.344 (MXG 17.10): Support for ZARA Release 1.3 is required
because only that release of ZARA is Y2K compliant.
Change 17.341 (MXG 17.10): VMXGVTOF VTOC reader had dates CREATED,
EXPIRES, and LASTUSE wrong. Text of Change contains correction.
(VMXGVTOC is archaic, having been generally replaced by DCOLLECT.)
Change 17.231 (MXG 17.07): Soft Audit XPMLKDT/XPMXPDT dates were not
only not Y2K Ready, but were flat out wrong in MXG decoding.
Change 17.227 (MXG 17.07): SAP IMS log record 'AE' was not Y2K Ready
until SAP Release 5.0i. Date was 0DDMMYYF, incompatibly changed to
YYYYMMDD format by that release, and now supported by this MXG change.
Change 17.216 (MXG 17.06): MXG Tape Mount Monitor ASMTAPES must have
been assembled with "ES6" to run under OS/390 R1.3 or later. If it
was assembled with ES5 or ESA, the INITTIME date will be 1900 in
the SMF record, but as long as you are using MXG 17.xx or later, it
will protect that 1900 date and your MXG dataset will be Y2K Ready
even if you don't re-assemble the MXGTMNT program from ASMTAPES.
Change 17.136 (MXG 17.04): TCP SMF record for TELNET in TYPETCPT
IBM did not use century bit 0cyydddF for this TELLOGFT datetimestamp;
instead used non-valid yyyydddF format, which creates year 3800 data.
MXG replaced SMFSTAMP8 format with circumvention code to support.
Change 17.091 (MXG 17.02): TELEVIEW 4.3
Release was incompatible, dates had never been validated, now ok.
Change 17.021 (MXG 17.01): HSM and TMS
Date values are Y2K compliant, but default print format was 6 instead
of 7, so the Julian Date of 2000001 prints as 2E6. The default MXG
format was changed; you can add FORMAT xxxxxx 7. ; to your reports.
I. MXG Software Version 17.17 is now available, upon request.
MXG Newsletter THIRTY-FIVE was the last printed MXG Newsletter.
MXG Newsletter THIRTY-SIX and all future MXG Newsletters will not be
distributed in printed form, but will be available in the MXG Source
Library, member NEWSLTRS, and on the homepage, www.MXG.com.
1. Major enhancements added in MXG 17.17:
Support for OS/390 Release 2.9.
Support for Lotus Domino Server Release 5.02.
CICINTRV dataset creation logic was wrong.
Revised utility to print NEWSLTRS/CHANGESS members.
Revisions and updates with new variables in all ADOCs.
Major enhancements added in MXG 17.10:
Year 2000 errors fixed for TYPEZARA and VMXGVTOF.
Y2K Cosmetic conversion of Julian dates from 00cyyddd to yyyyddd.
TELNET LOGF Time field is Duration, not datetime, not Y2K prob!
RMFINTRV/VMXGRMFI enhancements fixed and documented.
Support for Beta91 Balancing Manager SMF record.
Support for Software Innovation's LDMS product.
Added a replica of MICS SNTNSS report from NETSPY (and fixed it).
EXPDBOUT Example to add CICS Statistics datasets to your PDB.
Several ANALRMFR enhancements, this was just most recent.
Test version of VMXGSUM in XMXGSUM, uses SAS View to save DASD+.
Major enhancements added in MXG 17.09:
ABEND 2415, some sites due to no RECFM in //NULLPDS in SAS proc.
Starting with MXG 17.07, VMXGINIT now opens //SOURCLIB, and
some SMS sites get ABEND2415. Adding RECFM=U to the //NULLPDS
DD statement corrects the ABEND. See Change 17.317.
Support for IIS Log.
Support for Windows 2000 Build 2195 NTSMF data.
Support for remaining CA-VIEW Metrics validated.
Support for additional Landmark TMVS subtypes.
More IMS Log revisions for negative values, more in testing.
RMFINTRV now invokes VMXGRMFI, supports 115 workload, SYNC59
ASUMTALO Last-complete-interval now corrected.
Major enhancements added in MXG 17.08:
TYPEIMSA IMS Log revisions correct negative RESPNSTM/SERVICTM.
TYPECIMS IMF SQLCALLS are lost due to INCOMPAT change in MVIMF.
TYPE73 FICON PCHANBY/PNCHANBY wrong in initial FICON support.
TYPE74 PCTPNCHA/PCTPNOTH/PCTDVPND/PCTPNDEV revisions.
Support for APAR OW41147 ORGEXPDT=99999 - Read it: Y2K Critical
Support for CA View Metrics SARSMFUX SMF record.
Support for RACF Unload IRRDBU00 Started Task subtype.
Support for TRMS Version 51A08 (COMPATIBLE).
Support for Landmark DB2 Monitor V 3.2 (INCOMPAT).
Support for APAR OW40579/41407 SMF 42 subtype 4.
Support for APAR PQ28258 for SMF 103 record.
Support for unix PerfMeter Freeware Monitor records.
Support for DFSMS/MVS V1R5 - in place, no changes.
Support for CA VIEW Metrics SARSMFUX SMF record.
Support for SQL*NET NIV adds IPADDR/PORTNR to ORACLE.
Enhanced TYPE1032 Web Server eliminates negatives.
Negative QXSELECT because DB2 overflowed counter!
CECSER wrong if CPUs added or deleted during interval.
Utility to show IMACWORK definitions CPU sources.
TYPE90A replaces TYPE90 member for SMF type 90 data.
CICS Statistics EOD record has missing DURATM.
Major enhancements added in MXG 17.07:
Major IMS Enhancements for MXG IMS log processing.
Support for Domino Server R5.0/R5.01 SMF type 108 record.
Support for Top Secret 5.1 (INCOMPATIBLE).
Support for TMON/MVS V2 PTF TD01655 (COMPAT).
Support for MQ Series Version 2.1 (COMPATIBLE).
Support for TeleView 4.3B subtype 3 record.
Support for OS/400 Release 4.4.0 (LRECLs INCOMPAT)
Support for Mobius View Direct 6.1.2 (INFOPAC).
Support for APAR OW39508 7060 Multiprise EIO and DSD.
Support for OS/390 R2.8 (COMPAT, changes were APARs).
ASUM70PR now creates PDB.ASUMCEC BY CECSER vice BY SYSPLEX.
Major enhancements added in MXG 17.06:
Support for OS/390 R2.8 (Compatible, 16.09 or later supports).
Support for Lotus Notes, SMTPDS, SMTPRS objects in NTSMF data.
Y2K Ready: ASMTAPES must have been Assembled with ES6 parameter.
Major enhancements added in MXG 17.05:
Support for IBM's TPF (Transaction Processing Facility) OS.
Support for APAR OW31701 for ESS Parallel Access Volumes.
Support for OW39128 for RACF, adds DSNAME of PDS used for PROGRAM
Support for STK's VTCS 2.2.0 (INCOMPATIBLE) VSM SMF records
IBM's sample IXGRPT1 for SMF type 88 records is replicated.
New PDB.ASUMCEC corrects errors in PDB.ASUM70PR, more useful.
TYPETASK='OMVS' instead of TYPETASK='STC ' for OMVS/USS jobs.
OMVS/USS jobs filled SPIN, never purged, now output to PDB.
Major enhancements added in MXG 17.04 dated Jul 16, 1999:
Correction for preliminary IMS Log Enhancements.
Major enhancements added in MXG 17.04 dated Jul 16, 1999:
Support for CICS TS 1.3 new field inserted in CICSTRAN, INCOMPAT!
If you have CICS TS 1.3, you must install MXG 17.04 (or the one
line Change 17.156) or many variables in CICSTRAN will be bad.
Support for APAR OW37565 identifies CP or ICF CPUs in TYPE70PR.
ICF CPUs are now detected and deleted automatically in ASUM70PR.
Significant IMS Log processing enhancements available for test.
Utility to examine dates in SAS data library for Y2K.
Support for APAR OW37816, new 2105 cache data in TYPE74CA.
Support for Connect Direct R 3.2 'CT' record.
Support for MIM user record enhanced, new dataset.
Support for RMM European or American Date formats.
Support (preliminary) for NTSMF Windows 2000 Beta Three records.
Enhanced RMFINTRV/VMXGRMFI permits over 100 workloads now works.
Major enhancements added in MXG 17.03:
Support for Type 42 subtypes 7/8 NFS Usage/Users.
Support for APAR OW37091 Measured Usage SMF 89 changes.
Support for SoftAudit Version 7.1 (COMPATIBLE).
Support for STK's NearOAM V2.2 (COMPATIBLE).
Support for 32 (up from 16) sortworks for SYNCSORT.
Support for OS/400 V4.3.0, no change, is in MXG 16.16.
BUILDPDB vars TAPEDRVS/TAPE3480/etc corrected for MULTIDD='Y'.
BUILDPDB vars EXCPNODD/IOTMNODD now corrected for MULTIDD='Y'.
Enhancement for PDB.ASUMTAPE obs with STATUS=MISSEDMNT.
Major enhancements added in MXG 17.02:
Support for DB2 type 102 subtype 199, Dataset I/O.
Support for Lexmark MarkVision Job Statistics
Support for CICS TS 1.2 Journal segment GLRHTYPE=2, used by SAP.
Support for SOFTAUDIT 6.1.2 (COMPATIBLE).
Support for TELEVIEW 4.3 (INCOMPATIBLE).
Support for RACAL IT Security's SRM product for HSM.
Support for i-data afp-software's SMF record.
Support for NTSMF 2.3 (COMPAT), 21 new NT objects added.
Enhanced RMFINTRV/VMXGRMFI permits over 100 workloads. See 17.04.
RACF command keywords specified/ignored now decoded in TYPE80A.
Major enhancements added in MXG 17.01:
Support for DB2 type 102 subtype 199, Dataset I/O.
Support for Index Statistics in T102S022.
Support for deaccumulation of TYPE30_6 data.
Support for APAR OW37708/APAR OW38073 new fields.
Support for APAR OW15406 (IODF Creation now YYYY).
Support for "FICON" channels adds fields compatibly.
Support for Candle V400 Omegamon for CICS Epilog.
Support for IXFP/ICEBERG Subtype 8 and fix for subtype 6.
Support for NTSMF new Quota Server object.
Support for TANDEM F40, G04 and G05 INCOMPATIBLE.
Changing Interval with DURSET didn't change ASUM70PR.
Don't use PDB.TYPETMNT. USE PDB.ASUMTAPE for mounts.
ML-19 of ASMTAPES suppresses TMNT005E messages.
TYPEIMSA processing is wrong in 16.16. Use 17.08.
STC SILO SMF record sometimes short.
Unexpected TCP/IP command of STOU protected.
TMS Julian dates printed as 2E6, format now 7.
Revised and documented utility to modify BUILDPDB.
Text in col 72 cause unexpected failure.
All of these enhancements are described in the Change Log, below.
Availability dates for the IBM products and MXG version required:
Availability MXG Version
Product Name Date Required
MVS/ESA 4.1 Oct 26, 1990 8.8
MVS/ESA 4.2 Mar 29, 1991 9.9
MVS/ESA 4.2.2 Aug 15, 1991 9.9
MVS/ESA 4.3 Mar 23, 1993 10.10
MVS/ESA 5.1.0 - compatibility Jun 24, 1994 12.02
MVS/ESA 5.1.0 - Goal Mode May 3, 1995 13.01
MVS/ESA 5.2.0 Jun 15, 1995 13.05
MVS/ESA 5.2.2 Oct 19, 1995 13.09
OS/390 1.1.0 Feb 22, 1996 14.01
OS/390 1.2.0 Sep 30, 1996 14.05
OS/390 1.3.0 Compatibility Mode Mar 28, 1997 14.14
OS/390 1.3.0 WLM Goal Mode Mar 28, 1997 15.02
OS/390 2.4.0 Sep 28, 1997 15.06
OS/390 2.5.0 Feb 24, 1998 15.06
OS/390 2.6.0 Sep 24, 1998 16.04
OS/390 2.7.0 Mar 26, 1999 16.09
OS/390 2.8.0 Aug 24, 1999 16.09
OS/390 2.8.0 FICON/PAV/SHARK Aug 24, 1999 17.08
OS/390 2.9.0 Mar 31, 2000 17.17
CICS/ESA 3.2 Jun 28, 1991 9.9
CICS/ESA 3.3 Mar 28, 1992 10.01
CICS/ESA 4.1 Oct 27, 1994 13.09
CICS/ESA 5.1 aka CICS/TS V1R1 Sep 10, 1996 14.07
CICS-Transaction Server V1R1 Sep 10, 1996 14.07
CICS-TS V1R1 with APAR UN98309 Sep 15, 1997 15.06
CICS-TS V1R2 CICS/TS 1.2 Oct 27, 1997 15.06
CICS-TS V1R3 CICS/TS 1.3 Mar 15, 1999 17.04
CRR 1.6 Jun 24, 1994 12.02
CRR 1.7 Apr 25, 1996 14.02
DB2 2.3.0 Oct 28, 1991 10.01
DB2 3.1.0 Dec 17, 1993 13.02A
DB2 4.1.0 Tolerate Nov 7, 1995 13.07
DB2 4.1.0 Full support Sep 11, 1996 14.07
DB2 5.1.0 Tolerate Jun 27, 1997 14.14
DB2 5.1.0 Full support Jun 27, 1997 15.02
DB2 6.1.0 Mar 15, 1999 16.09
DFSMS/MVS 1.1 Mar 13, 1993 11.11
DFSMS/MVS 1.2 Jun 24, 1994 12.02
DFSMS/MVS 1.3 Dec 29, 1995 13.09
DFSMS/MVS 1.4 Sep 28, 1997 15.04
DFSMS/MVS 1.4 HSM Sep 23, 1998 16.04
DFSMS/MVS 1.5 ??? ??, 1999 16.04
MQM 1.1.2, 1.1.3, 1.1.4 Apr 25, 1996 14.02
MQ Series 1.2.0 May 26, 1998 16.02
MQ Series 2.1.0 Oct 2, 1999 17.07
NETVIEW 3.1 type 37 ??? ??, 1996 14.03
NPM 2.0 Dec 17, 1993 12.03
NPM 2.2 Aug 29, 1994 12.05
NPM 2.3 ??? ??, 1996 15.08
NPM 2.4 Nov 18, 1998 17.01
RMDS 2.1, 2.2 Dec 12, 1995 12.12
TCP/IP 3.1 Jun 12, 1995 12.12
TCP/IP 3.4 Sep 22, 1998 16.04
DOS/VSE POWER V6.3.0 Dec 19, 1998 16.08
VM/ESA 2.0 Dec 23, 1992 10.04
VM/ESA 2.1 Jun 27, 1993 12.02
VM/ESA 2.2 Nov 22, 1994 12.06
VM/ESA 2.3 ??? ??, ???? 16.08
IMS 4.1 Aug 6, 1994 12.02
IMS 5.1 Jun 9, 1996 14.05
IMS 6.1 ??? ?, 199? 16.04
AS400 3.7.0 Nov 1, 1996 15.01
AS400 4.1.0 Dec 30, 1996 15.08
AS400 4.2.0 Apr 27, 1998 16.02
AS400 4.4.0 Sep 27, 1999 17.07
Availability dates for non-IBM products and MXG version required:
Product Name Required
Windows NT 4.0 and NT 3.51 14.14
Windows NT 4.0 Service Pack 2 15.03
Windows NT 4.0 Service Pack 5 16.04
NTSMF Version 1 Beta 14.11
NTSMF Version 2.0 15.05
NTSMF Version 2.1 15.06
NTSMF Version 2.2 16.04
The Monitor for DB2 Version 2 13.06
The Monitor for DB2 Version 3.0 16.02
The Monitor for DB2 Version 3.1 16.02
The Monitor for CICS/ESA 1.2 - 12.12
The Monitor for CICS/ESA 1.3 - 15.01
The Monitor for CICS/ESA 2.0 - 15.06
The Monitor for MVS/ESA 1.3 - 12.05
The Monitor for MVS/ESA 1.5 - 12.05
The Monitor for MVS/ESA 2.0 - 15.09
Omegamon for CICS V200 User SMF 12.05
Omegamon for CICS V300 User SMF 13.06
Omegamon for CICS V400 User SMF 16.02
Omegamon for CICS V400 type 110 segments 16.02
Omegamon for IMS V110 (ITRF) 12.12
Omegamon for IMS V300 (ITRF) 14.04
Omegamon for MVS V300 13.05
Omegamon for MVS V400 13.06
Omegamon for DB2 Version 2.1/2.2 13.05
Omegamon for VTAM V160 12.04A
Omegamon for VTAM V400 15.15
Omegamon for SMS V100/V110 12.03
ACF2 6.2 16.04
ASTEX 2.1 14.04
NETSPY 4.7 14.03
NETSPY 5.0 14.03
NETSPY 5.2 16.05
Boole & Babbage
IMF 3.1 (for IMS 5.1) 12.12
IMF 3.2 (for IMS 6.1 only) 15.09
IMF 3.2 (for IMS 5.1 and 6.1) 16.04
LMS 3.1 12.12A
MXG IMS-Log Not-Officially-Supported
IMS 6.1 - ASMIMSL6/TYPEIMSA 17.08
IMS 5.1 - ASMIMSL5/TYPEIMSA 17.08
APAF 4.1, 4.3 16.08
II. MXG Technical Notes
1. Compatibility issue. TYPExxxx members may need //PDB or %LET.
Some TYPExxxx members data must be deaccumulated to be valid:
SMF records: TYPE103 TYPE28 TYPEDB2 TYPEHMF TYPEHSM
TYPENTCP TYPEROSC TYPETPX
Non-SMF record: TYPETMDB
MXG adds %INCLUDE SOURCLIB(DIFxxxx) to those TYPExxxx members to do
the deaccumulation; the DIF's used to output to the //WORK DDNAME
but DIF() members now sort their output to the MXG default DDNAME
of //PDB. So if you use %INCLUDE SOURCLIB(TYPExxxx) in your jobs
you may need to add a DDNAME of //PDB plus a PROC COPY, or you may
want to use the new %LET Pdddddd=WORK syntax to change that "PDB"
back to the "WORK" that your old job expects.
Note: this does not apply if you use BUILDPDB to create the data).
- You can add a //PDB DD pointing to a temporary dataset and then
add a PROC COPY after the %INCLUDE of the TYPExxxx member:
//PDB DD DSN=Temp,UNIT=SYSDA,SPACE=(CYL,(xxx,yyy))
//SYSIN DD *
PROC COPY IN=PDB OUT=WORK;
... Your SAS code ....
to create in the //PDB and then copy them back into //WORK for
your old program to find.
- Or, you can change the Pdddddd macro for the dataset;
//SYSIN DD *
... Your SAS code ....
The "Pdddddd" name for each dataset is documented in the IMACxxxx
member. The default is usually "PDB", except for these datasets:
Product Default Destination For Dataset
CICS 110 - PCICTRN=CICSTRAN CICSTRAN
IMS log - PIMSTRN=IMSTRAN IMSTRAN
PIMSxxx=WORK All IMS Temporary datasets
HP MW - HPxxxx =HPPDB All HPxxxxxx
2. Recent hardware timing measurements for moving files:
Path connection: MB/sec MB/min MB/Hr GB/Hr
Internet @56Kbit Dial In .016 1 50 .048
Internet T3 .100 6 360 .351
MVS to 10Mbit Lan .166 10 1000 .976
PC-to-PC 100Mbit Lan 3.3 200 12150 11.8
Pc-to-PC Disk-to-Disk Fast IDE 6.0 360 21600 21.1
A 56Kbit connection can move a maximum of 19 MegaBytes
per hour at 100% utilization and no overhead, so how
can we move 50 MB per hour? Compression in the modems.
The SMF data compresses to between 8 to 1 and 10 to 1,
so the 50 MB MVS file is only 5 MB of compressed data
which can be delivered across a 56Kbit dial-up line.
3. An example of the use of IMACFILE/MACFILE exit:
Your site can choose to write the CADI record as SMF type 6 or as a
user SMF record number. Member TYPE6 (included by BUILDPDB) reads
the type 6, but if your site chose to write, for example, type 239,
you will need to use the IMACFILE exit (which is taken after the
SMF header has been read, but before the decoding of the record) to
change ID=239 to ID=6. You would insert in member IMACFILE:
IF ID=239 THEN ID=6;
and MXG will process your type 239 record with the type 6 logic.
Alternatively, you could use the %LET MACFILE macro variable syntax:
%LET MACFILE= %QUOTE( IF ID=239 THEN ID=6; ) ;
to change the record ID in your //SYSIN stream. See DOCMXG.
4. Make sure your MVS REGION size is 64M, or you may get many strange
SAS errors, especially when executing a %MACRO. For example, using
%BLDNTPDB with a region of only 6MB produced syntax errors "CURRENT
WORD HAS BECOME MORE THAN 200 CHARACTERS LONG" that was not a true
syntax error when SAS had enough memory. Put the REGION=64M on your
JOB card, so it will apply to all steps. The REGION= on the EXEC has
effect only when there is no REGION specified on the JOB card.
Why 64M? The MXG Default MEMSIZE=64M is specified in CONFIG member
and restricts SAS, so using the same REGION=64M makes sense, and I
don't regard region size as a consumable resource. MXG programs run
in less than 48M, but if you tailor BUILDPDB for additional records,
buffering for the additional output datasets requires more virtual
storage, so MEMSIZE was raised to 64M a few versions ago for safety.
One site that unwisely read both SMF and Landmark CICS data from two
input files in a single data step needed 101MB of virtual space!
5. MXG Definitions with regard to MXG Software Changes:
COMPATIBLE A change in a data record which did not alter either
the location or the format of all of the previously-
kept MXG variables is COMPATIBLE, and you can continue
to run the old version of MXG software, which will read
the new records without error, but none of any new data
fields or any new record subtypes will be created/kept
until you install the MXG Version with this change.
A change that alters any previously-kept variable is
INCOMPATIBLE, and requires the new version to be used.
TOLERATE In other words, the old MXG Version TOLERATES the new
data records, if they are COMPATIBLY changed.
EXPLOIT Once you use the new MXG Version to read the changed
records, all of the new fields, subtypes, etc, that are
described in this change will be created in the MXG
datasets, so the new MXG Version EXPLOITS the new data,
and you have full support of the new data records.
INCOMPAT A change in a data record that causes the current MXG
version to fail, visibly or invisibly, with or without
error conditions or messages, and the output datasets
may contain wrong values and incomplete observations,
and/or observations may have been lost.
You MUST install the new MXG Version with this change
to process data records that have been INCOMPATIBLY
changed by their vendor.
III. MVS Technical Notes.
1. Some MXG typos in APAR names/numbers in NEWSLTRS or CHANGESS:
PW13647 is PQ13647, PW18797 is PQ18797, PW61950 is PN61950, and
PW56441 is PN56441. All are now corrected.
2. APAR OW36535 corrects Start/Stop timestamps in type 42 subtype 6
(TYPE42DS) records. There is now a PTF available. This problem
had been previously reported in OW10694 and OW25265.
3. Problems with ECSA creep are documented in APAR OW34249, and there
is another APAR, OW29277 (included in JES 2.7/HJE6607) that may be
involved in dangling ECSA storage.
4. The CA-SPOOL product APAR GS98190 (dated 1996) prevents that
product from writing invalid type 6 SMF records with TYPETASK='FI'
and invalid data in most other TYPE6 fields.
5. How many volumes can you have in multi-volume MVS datasets?
A recent note by IBMer Arliss White explains:
PDS, PDSE and HFS data sets are limited to one volume.
Extended format data sets with multiple stripes are limited to
A data set on a VIO simulated device is limited to 65,535 tracks
on one volume.
All other DASD data sets are limited to 59 volumes.
Tape data sets are limited to 255 volumes.
For SMS-managed and non-SMS-managed data sets, you can specify
up to 59 volume serial numbers. If the combined number of
volumes for a cluster and its associated alternate indexes
Exceed 59, unpredictable results can occur.
SAS Version 6 has a limit of five volumes for a data library on
DASD, but SAS Version 7/8 eliminates that constraint.
6. How does software compression compare with hardware compression for
MXG's DB2ACCT dataset? On OS/390 R2.5, a 55 MB SMF file with only
DB2 records was read to create the twelve DB2xxxxx datasets. The
DB2ACCT output was 55MB, and occupied 1110 tracks uncompressed.
Compressing with SAS Software reduced DB2ACCT to 600 tracks.
Hardware compression using extended-format striped tape-format
dataset on ESCON reduced DB2ACCT to only 480 tracks. On older ECL
machines, the CPU cost of hardware compression is significantly
higher than software, but the 70 MIPS-per-engine CMOS processor
shows no difference between the CPU cost of software or hardware
compression. Compressing costs 5 minutes of CPU time per Gigabyte
written, added to a base cost of 11 minutes of CPU time per
Gigabyte written uncompressed:
---MVS CPUTM to Create All DB2 Datasets---
--------but compress only DB2ACCT---------
System Machine Standard Striped SAS Software Hardware
SU_SEC Create Uncompr Compress Compress
ECL 2429 32.3 36.6 46.2 55.2
CMOS SAS V6 2812 24.6 28.9 36.2 35.7
CMOS SAS V7 2812 28.7 40.8 40.6 41.1
As a tape-format striped dataset under V6, the dataset size
jumped up to 1800 tracks, because V6 forced a space wasting
DASD block size of 32760, but V7 required only 960 tracks: it
knows about the extended-format datasets and uses half track
to save space. It is 960 rather than 1100 because there's no
directory in a tape-format SAS dataset.
The same 55MB SMF input file was read under Windows 98 with both V6
and SAS V7 on a 166MHz Pentium and a 500MHz Pentium III with fast
IDE drives, and under Windows NT 4.0 Server on a dual 2x200MHZ
Pentium with SCSI drives. On the 500MHz machine, creating all DB2
datasets compressed on the PC took the same time, about 40 seconds,
as it took that 70 MIPS mainframe to create eleven and compress
only the DB2ACCT! And on all of the PC machines, there is no CPU
cost to compress data; it takes slightly less time to compress than
to create non-compressed! The savings of CPU time to move less
data is greater than the cost of CPU time to compress, on PCs:
---Elapsed time of Data Step---
System Standard SAS Software
166MHZ PENT II V6: 113.8 115.0
2x200MHz Dual Pent: 54.7 48.7
500MHZ PENT III V6: 42.4 40.2
500MHZ PENT III V7: 43.0 41.6
All twelve DB2 datasets were 84MB uncompressed, and their total was
58MB when compressed. The uncompressed DB2ACCT data set (the 1100
track dataset that SAS compressed to 600 tracks) was 55MB on the PC
and compressed to 32MB, (about 620 tracks), so MVS and PC software
compression effectiveness are very similar, although MVS hardware
compression is more effective at no additional cost on MVS.
The Dual Pentium time of 48 seconds on the 2x200 compared with 41
on the 1x500 shows that SAS can exploit multiple engines, although
some of the reduction may also be due to the faster SCSI drives.
Here, under NT Server, which separately measures the CPU time, the
processor time increased from 35 seconds to 42 seconds when
compression was enabled, but the run time was still much less due
to the reduced volume of data moved when it is compressed.
In summary: This paragraph revised to quantify the CPU cost on MVS:
On dedicated PCs, it is significantly faster (less CPU time too) to
create compressed datasets than to create uncompressed datasets.
On "MVS", the CPU cost of compression with either hardware or with
SAS Software is about the same on (newer) CMOS hardware, but that
increase is about 40%, which some might regard as expensive, but
may be only a modest cost when it prevents a job to fail due to
insufficient disk space.
Note added Apr 2004: One site's BUILDPDB increased from 266 to
477 CPU seconds (79% increase, or 3.5 extra CPU minutes per day),
but BUILDPDB does lots of compress/decompress/sort/compress/etc.
which would increase the amount of CPU time.
7. How do you hardware compress a SAS dataset on OS/390 mainframes?
You can not use OS/390 hardware compression for a SAS data library,
but you can hardware compress individual SAS datasets, by creating
each as a SAS-tape-format, OS/390-extended-format, striped dataset,
i.e. as a z/OS, extended-format, compressed, striped dataset:
- To create the dataset in "Tape" format, either write to a DD name
that starts with "TAPE", or use a LIBNAME statement to set the
write engine to "TAPE":
LIBNAME DB2ACCT TAPE;
- The dataset that is pointed to by the //DB2ACCT DD statement in
JCL, must create that dataset in a DATACLAS that has specified
EXTENDED ADDRESABILITY = YES (or MAY)
COMPACTION = YES
- You must also put the dataset in a STORCLAS that specifies
SUSTAINED DATA RATE = 4 (to get one stripe, 8 for two...)
And note that this one SAS dataset, now a sequential dataset can be
written to multiple DASD volumes (although it cannot be used for
Direct Reads - SAS datasets in tape format are read sequential).
8. APAR OW38615 for RACF type 80 records corrects PERMIT DATASET
record, which can have an invalid resource name.
9. APAR OW37742 for Workload Manager "Goal Mode" revises the logic of
the "small consumer" algorithm to limit the duration when lower
importance work in a system from preempting higher importance work.
10. APAR AR38393 for Hardware Compression will now bypass decompression
and compression when copying a compressed VSAM file to a device
with same physical attributes. It was supposed to bypass when it
could, but the code did not work as advertised.
11. APAR OW39277 for VSAM record statistics in LISTCAT wrong when using
"dataset name sharing" with multiple opens to a VSAM dataset under
some conditions discussed in the APAR text. The mismatch between
the LISTCAT and actual number of records in the file could also be
the result of an abnormal close, and the APAR recommends looking at
the type 64 records to determine if this APAR would fix the error.
12. LOTUS Domino may need a 2GB region, and may not perform well if the
site's IEFUSI exit changes the region size for the FORKed/SPAWNed
address spaces. See APAR OW38477 for discussion of why MAXASSIZE
should be used instead of the IEFUSI exit.
13. APAR OW37709 discusses new BMF cache options, for PDSE and HFS,
that can be used to manage the size of the BMF data space cache.
That APAR notes these options are intended for heavy users of BMF
caching, "such as LOTUS Domino servers experiencing BMF data space
storage shortage when running high activity mail databases for 5000
or more users". TYPE42 fields added by OW37708 to capture active
and max BMF buffers were supported by Change 17.059 (MXG 17.01).
14. For the new Shark/Parallel Access Volume hardware, a set of APARs
have been recently released, and the list below shows the most
recent first, and its previous (and prereq'd) APAR next:
Support is in MXG 17.05.
15. If you run the IFASMFDP "SMFDUMP" program with //SYSPRINT pointing
to a disk dataset (rather than SYSOUT), you run the risk of losing
SMF data if you ever take an I/O error to SYSPRINT, as discussed in
this new APAR without a fix, OW40176.
16. Deleted. Was incorrect; error was in ASUM70PR.
17. A reminder: VSAM Catalogs will stop functioning on 1/1/2000. Any
existing VSAM Catalogs must be converted to ICF Catalogs by then!
18. APAR OW32140 for WLM and CICS Reduces IRASASRV sampling rate if
there are no Server Address Spaces, i.e., where CICS transactions
are not being classified by WLM.
19. APAR PQ25641 for SmartBatch Release 1 or Release 2 BatchPipePlex
Cross System Piping with EOFREQUIRED=Y, reports that blocks of data
can be lost, and there is no error nor ABEND message generated.
It is possible to detect the loss of data records by comparing
the input block count in the reading job's ASFP391I message in
its job log with the output block count in the writing job's
ASFP392I message in its job log.
The type 91 SMF record contains input and output block counts
for SmartBatch, and we are going to see if we can write an MXG
utility to detect if data was lost. 17Sep1999.
20. APAR PQ30654 for TCP/IP reports that FTP SMF records can have their
FTPEND earlier than FTPSTART for transmissions that span midnight.
There is no correction in this APAR (i.e., IBM should have added
two date fields, one for start and one for end, so there would be
no ambiguity (and the data would be valid if an FTP transmission
crossed more than one midnight!), but they didn't!).
21. APAR OW37263 is required to populate the ACCOUNTn Accounting
Fields in type 30 records for OpenEdition tasks that were spawned.
22. APAR OW39746 corrected problems with TYPE74CF Coupling Facility
CPU busy time exceeding duration of the interval. The problem
is related to reconfiguration and goes away when all of the systems
in the parallel sysplex get IPL'd again.
23. APAR OW41239 corrects the value of mounted duration in OAM SMF
type 85 subtype 87 records.
24. APAR OW41169 corrects several SHARK problems, including capturing
SMF type 42 subtype 6 activity counts for PAV alias exposures.
25. APAR OW38842 causes the OMVS/USS field SMF30EXN, TYPE30_4 variable
OMVSEXNP, the OMVS Executed Program Name, to be blank.
APAR OW41696 says that APAR OW41764 corrects the error.
26. APAR PQ32322 reports TCP/IP API SMF records have FFFFFFFFx for byte
count, (-1 if input with IB4., but MXG stores 4,294,967,295, using
PIB4. informat, since negative value is invalid!). The APAR also
says the records with these values are an extra TERM record that
did not have an associated INIT record. The APAR also reports that
the fix, when available, will also eliminate duplicate records!
27. APAR OW42559 reports variable UCCOLDT in DCOLLECT dataset DCOLCAPD,
and DCOLCAPD records being lost. The UCCOLDT value is 1900dddF
instead of either 0100dddF which was documented or 2000dddF which
is not expected (but either 0100dddF or 2000dddF raw values will
be correctly handled by MXG when IBM issues a PTF to fix the error.
28. MXG required for OS/390 R2.8 was originally listed as MXG 16.16,
but there are a number of subsequent APARS for SHARK/PAV/FICON that
changed RMF records (incompatibly for FICON channel measurements);
those APARs now require MXG 17.08 or later.
29. To know the size of Hipervolumes on EMC boxes, DCOLLECT dataset
DCOLVOLS variable DCVVLCAP shows the actual capacity of each of
your DASD volumes in megabytes.
IV. DB2 Technical Notes.
1. Excess DB2 SMF type 101 records due to Query CP parallelism can be
prevented (now) only by RLF Parallel Mode Disablement, with RLFFUNC
and RLFBIND. Previous APARS provided disablement for dynamic plans
and now DB2 APAR PQ06968 extends disablement to static plans also.
2. But the real solution to excess DB2 type 101 records is available
with the PTF UQ24763 for APAR PQ10864. This APAR implements a V5
dcr which rolls up all child task accounting data into a single
record that is written with the parent task at deallocation time,
and that record should be handled by MXG without difficulty, but
I have not yet seen test data records. Only ANALDB2P might be
impacted, as it used the child records in analysis of parallelism.
3. APAR PQ27561 corrects a large increase in QTXACLUN (claim failures)
after migration to DB2 5.1. The APAR corrects the counts and has a
discussion of what was wrongly counted.
4. The Sunrise product can corrupt the timestamps in DB2 101 and 102
SMF records. For Sunrise 4.11 their fix is sd11024, for Sunrise
4.10 their fix was sd10231.
V. IMS Technical Notes.
There are no IMS notes in this Newsletter.
VI. SAS Technical Notes.
1. SAS 6.09 at TS460 contains maintenance to support UCB's above the
line, and using it instead of TS450 has corrected 0C4 ABENDS. The
SAS Log stopped at the "Welcome to MXG" message.
2. SAS Version 7 under MVS and OS/390 fails with ABEND 913-1C if you
read Shared Multi-Volume DASD SAS Data Libraries. Daily jobs that
ran at the same time and that read the same SAS library under V6
now ABEND under V7 when they try to open the same SAS data library.
But only if the "PDB" or SAS Data Library is Multi-Volume and is
on DASD. Single-volume DASD data libraries have no error, and
multi-volume SAS tape data libraries cannot be shared!
The error is fixed in Version 8 and SAS is developing a ZAP for V7
to fix the problem: see SAS Usage Note V7-SYS.SASIO-0704.
You can circumvent the error by changing the JCL to use DISP=OLD on
each DD statement, because that forces MVS to serialize the jobs so
that only one job runs at a time. But a 10-minute elapsed parallel
job stream could become a long elapsed time when serialized.
3. SAS Version 7 will cause SYSMSG notes that OPTIONS CODEPCT and
BLKSIZE(TAPE)=FULL are not supported in this version, but those
changes do not affect the execution of MXG Software, nor do they
set any condition codes. If I can get rid of them easily without
you doing anything, I will, but for now they are harmless.
4. SAS Error Unknown Exception (80000602), a severe error occurred in
task SQL for module SABXSHL executing in SABXSHL ... was eliminated
by the user deleting his SAS PROFILE catalog from SASUSER!
5. SAS 609 maintenance TS460 introduced a message that can be ignored:
WARNING: This is an experimental version which will be replaced
by XX AMS in Version 7.
SAS Usage Note V6-SERVER-F759 explains the note, which was removed
in SAS TS465, but it causes no problem and can be ignored.
6. Using SMS "Immediate Release" for a SAS Data Library can cause SAS
to USER ABEND 315, so don't do that. See SAS Usage Note 2705.
Incorrect multi-volume allocation can also cause 315 ABENDs; see
other mentions of 315 in member NEWSLTRS.
7. SAS V7 ZAP http:/www.sas.com/service/techsup/unotes/V7/0281.html is
a required zap that also allows SAS data library allocations on MVS
to exceed 32,767 tracks. However, a SAS data library allocation is
still limited to 65,536 tracks (4369 cylinders @15trk/cyl) on one
volume, in Version 6, 7 and 8. SAS V7 and V8 also allow more than
5 volumes in a multi-volume SAS data library.
8. SAS V7 errors a PROC SORT with DUPLICATE SORT KEY VARIABLE if the
same name appears twice in the BY statement (due to some problems
in some host sort programs). MXG had two unintentional instances
where SMFTIME was repeated that were removed in 17.04 so MXG would
execute under V7. But now SAS V8 does not error, and instead
prints only a warning message, and then removes the repeated
variable internally before calling the sort. And for both V7 and
V8, the restriction is only for a BY statement with a PROC SORT;
use with other PROCs is not restricted.
9. New SAS format/informats for IBM date formats added in V7 and V8:
PDJULIw. Reads/Writes Packed Decimal Julian Dates ccyydddF
New SAS function
JULDATE7() Returns 7 digit Julian date from a SAS date value,
in yyyyddd format, which could then be written to an
external file with hex format yyydddC by using:
juldate=JULDATE7(sasdate); put juldate PD4. ;
Unfortunately, MXG cannot exploit these new features, because I
can't depend on you having the new SAS version installed, so MXG
already protects these formats as described in member YEAR2000.
10. SAS V8 Developer's Release for Windows fails with I/O error when it
reads a VBS file that contains '1A'x (which Windows erroneously saw
as a CTRL-Z, which is a Windows End-of-File marker). The problem
has been fixed for the V8 Production Release for Windows and unix.
There is a SAS patch for Windows and Windows NT available at:
but SAS does not create external usage notes for test releases.
There is no unix patch because no unix site has reported the error.
MXG 17.06 was successfully QA'd under SAS V8 Developer's Release,
with the above fix, and only these minor notes:
- Trying to create the MXG Formats into a directory with existing
FORMATS.SC2 file from an earlier version of SAS causes syntax
errors around the "OTHER" operand. Erase the old FORMATS.SC2.
- V8 won't tolerate a LIBNAME to a nonexistent directory.
Continued testing is planned, but it looks very good so far!
11. SAS V8 Developer's Release for OS/390 has successfully executed the
standard MXG BUILDPDB program and the MXG QA stream, with a few new
warnings and some return code fours under investigation:
BUILDPDB with 1.8GB SMF file (no CICS, no DB2, mostly 30s & RMF):
Release EXCPs CPU Minutes Elapsed Minutes
V6 178K 26.23 49.48
V8 155K 25.61 44.26
-13% -2% -11%
However, SAS Institute has replicated an error that surfaced as an
out-of-memory problem during compile of an MXG BUILDPDB program
that used %LET MACKEEP= and %QUOTE( ) to pass a list of _CDExxxx
old-style macros that expanded to over 60,000 lines of SAS inside
the %QUOTE( ) function. It would compile (accidentally?) if the
list was reordered and made many-per-line instead of one-per-line!
But this parsing error exists only inside a %QUOTE function with
thousands of bytes of text, so it doesn't seem pervasive and it
should be fixed by SAS Production Version 8 availability.
12. This NOT an MXG Y2K issue, because MXG doesn't use any date literal
values that are YY, and certainly not any before 1960, but it shows
what happens when YEARCUTOFF is changed from 1900 to 1960 if you
have any YY date literals in your own SAS programs, perhaps one to
calculate your age in days, using: AGE=TODAY()-'19APR41'D;
If YEARCUTOFF=1900 you get the correct AGE=21353, but if you use
the YEARCUTOFF=1960, you get an incorrect AGE=-15172 value! This
problem really only arises when the YY start date is earlier than
the YEARCUTOFF value of 1960, so your kid's ages will be correct!
MXG sets YEARCUTOFF=1960 so that any YY literals still in your MXG
report programs will be protected for computer dates (i.e., after
1960), but you really should use YYYY: AGE=TODAY()-'19APR1941'D;
so date durations will be calculated independent of YEARCUTOFF.
13. USER ABEND 0318 is usually Multi-Volume related, although there are
a list of other "opportunities" in member NEWSLTRS under "0318",
and now another cause: specifying DSORG=DA on //CICSTRAN instead
or DSORG=PS caused USER ABEND 0318 (the JCL was holdover from back
when SAS Version 5 data libraries were DA,RECFM=U libraries).
14. Format Libraries built with SAS V7/V8 cannot be used with SAS V6.
The physical format of SAS data libraries (where format libraries
are stored as type=catalog) was changed between V6 and V7.
15. Under Windows NT, attempting to create the FORMATS catalog in a
directory that contains a FORMATS.SC2 file that is Read Only will
create SAS ERROR: USER DOES NOT HAVE APPROPRIATE AUTHORIZATION
LEVEL FOR FILE LIBRARY.FORMATS.CATALOG. Erase the existing .SC2
file and recreate by running %INCLUDE SOURCLIB(FORMATS);
16. SAS Version 8.0 under Windows/Windows NT reads only the first file
with concatenated files if the RECFM=S370VBS. This error was fixed
in SAS Version 8.0 TS M1, and this note was revised May 15, 2000.
17. What level of MXG is needed for the SAS Version 8 Release (TS M0)?
MXG 16.16 runs with SAS V8 (TS M0), with a note that SAS options
CODEPCT and BLKSIZE(TAPE) no longer exist.
MXG 17.01 provides new CONFIGV8 without those two options to
remove the note on MVS, and new AUTOEXEC for the same
purpose under ASCII SAS. Change 17.073.
MXG 17.07 exploits 32000-byte character variable length for the SQL
text variable in TYPE102 (but only when under V8 and only
if COMPRESS=YES was also specified). Change 17.253.
MXG 17.08 exploits the new INHERIT option of PROC MEANS in VMXGSUM
to skip a data step now unneeded with V8. Change 17.265.
MXG 18.04 changes V8 default to SEQENGINE=V6SEQ. Change 18.104.
18. Variable Blocked files can be used on MVS for MXG's USERID.SOURCLIB
and MXG.SOURCLIB, with some constraints. First, all libraries in
the //SOURCLIB concatenation must be all VB; SAS does not support
mixing RECFM=VB and RECFM=FB files for //SOURCLIB (S013-64 ABEND),
although all RECFM=VB or all RECFM=FB does work without error.
But second, SAS does not support RECFM=VB in its //CONFIG file at
all (S013-20 ABEND), even if all of the //CONFIG datasets are VB.
Tracking 5203445 is open to request future design change in SAS.
VII. CICS Technical Notes.
*** THIS ORIGINAL NOTE HAS BEEN REVISED AND UPDATED - SEE ADOCDB2 ****
*** DO NOT USE TXID ***
1. The CICS RCT parameter TOKENE=YES does not exist now in CICS/TS, as
the CICS RDO replaced the RCT. The new parameter is ACCOUNTREC:
ACCOUNTREC(TXID) - you DO NOT WANT, but may be your default!
ACCOUNTREC(UOW) - the same as TOKENE=YES, use for CICS TS 1.1.
ACCOUNTREC(TASK) - MXG/IBM Recommended, less detail than (UOW).
Specify ACCOUNTREC(UOW) in the RDO to do what TOKENE=YES was doing,
that is, to cause DB2 to create a DB2ACCT record for each CICS event
and to send the LU 6.2 token to populate QWHCTOKN in the DB2 record,
The DB2ACCT and CICSTRAN multiple observations can be matched and
combined in a "Business Unit of Work" observation in PDB.ASUMUOW.
Note that the "UOW" in ACCOUNTREC(UOW), should really be "UOWID",
in my opinion, because the "UOW" option actually creates DB2ACCT
records for each unique value of the 8-byte UOWID, including the
sequence number at its end. To CICS, a "unit of work" is only one
recoverable unit, or one unique value of that 8-byte UOWID token.
But in MXG, especially in MRO, a "UOW" has always been a business
unit of work, the collection of multiple CICS task records from
TORs, AORs, FORs, plus any DB2ACCT records, that have the same
first 6-bytes of that 8-byte UOWID field, MXG variable UOWTIME,
the arrival time of the business work unit. We may get the LUNAME
and TRANNAME of the work unit from the TOR task record, but get
PROGRAM and resources from the AOR record to create PDB.ASUMUOW.
So while reading about CICS internals, think of a "UOW" as a
specific value of the 8-byte UOWID, but everywhere else in MXG,
know that a "UOW" is a business unit of work, combining many CICS
and DB2 UOWIDs into one unit of work.
The "UOWID" level DB2ACCT records created with ACCOUNTREC(UOW) is
the maximum detail available, and provides tracking of the sequence
of times when elements of a Unit of Work passed from CICS to DB2 and
back in the CICSTRAN and DB2ACCT datasets, but the UOWID level of
detail is not needed in the DB2ACCT dataset for PDB.ASUMUOW, so you
will likely use ACCOUNTREC(UOW) only if you need the detail level
or are still on CICS TS 1.1.
In CICS TS 1.2, IBM introduced their (recommended) ACCOUNTREC(TASK)
option to reduce data volume on the DB2 side, as it causes DB2 to
create the DB2ACCT observation only at the end of each CICS task,
but it still populates the QWHCTOKN (because multiple DB2ACCT obs
can still be created by thread release at intermediate syncpoints).
Unless you really need the additional level of detail per UOWID,
specify ACCOUNTREC(TASK) and PDB.ASUMUOW will be cheaper to create!
Do not use ACCOUNTREC(TXID), equivalent to the RCT's TXIDSO=YES.
Unfortunately, the MIGRATE function to create your RDO from your old
RCT may set TXID as the default, so you need to check your RDO and
change it to either TASK or UOW, because TXID causes QWHCTOKN to be
blank, preventing DB2 match up with CICS. Also, only one DB2ACCT
record is created for multiple CICS transactions, with accumulated
resources, when TXID is specified.
What IBM calls a "multi unit of work transaction" is a just a series
of CICS tasks with the same UOWTIME but different values of UOWID.
Checkpoints, SYNC, Commits, etc, alter that sequence number.
The actual MXG definition of a "Unit of Work" uses both the NETSNAME
(the originator of the CICS transaction) and the UOWTIME to create
observations in the PDB.ASUMUOW dataset.
2. MXG 17.04 or later is required for CICS/TS 1.3 (or at least the
Change 17.156 must be applied to MXG 16.16).
VIII. Windows NT Technical Notes.
There are no Windows NT Technical Notes in this Newsletter.
IX. Incompatibilities and Installation of MXG 17.17.
1. Incompatibilities introduced in MXG 17.17 (since MXG 16.16):
a- No changes in MXG architecture were made between 16.16 and 17.17
that introduced incompatibilities.
2. Installation and re-installation procedures are described in detail
in member INSTALL (which also lists common Error/Warning messages a
new user might encounter), and sample JCL is in member JCLINSTL.
X. Online Documentation of MXG Software.
MXG Documentation is now described in member DOCUMENT.
XI. Changes Log
You MUST read each Change description to determine if a Change will
impact your site. All changes have been made in this MXG Library.
Member CHANGES always identifies the actual version and release of
MXG Software that is contained in that library.
The CHANGES selection on our homepage at http://www.MXG.com
is always the most current information on MXG Software status,
and is frequently updated.
Important changes are also posted to the MXG-L ListServer, which is
also described by a selection on the homepage. Please subscribe.
The actual code implementation of some changes in MXG SOURCLIB may be
different than described in the change text (which might have printed
only the critical part of the correction that need be made by users).
Scan each source member named in any impacting change for any comments
at the beginning of the member for additional documentation, since the
documentation of new datasets, variables, validation status, and notes,
are often found in comments in the source members.
Alphabetical list of important changes after MXG 16.16 now in MXG 17.17:
Member Change Description
All 17.060 MXG Enhancement %LET &Wdddddd=ddname for //WORK copy.
many 17.171 PRINTWAY JCTJOBID='PSnnnnnn' support in TYPETASK.
many 17.214 Support for OS/390 R2.8 (COMPAT, changes were APARs.
ADOCall 17.379 Revisions and updates with new variables in all ADOCs
ANAL42 17.223 Report failed with more than one SYSTEM.
ANAL88 17.185 IBM's sample IXGRPT1 for SMF type 88 replicated.
ANAL91 17.243 Batch Pipes Analysis for APAR PQ25641.
ANALCNCR 17.070 Concurrency analysis now tolerates missing values.
ANALDATE 17.168 Utility to examine dates in SAS data library for Y2K.
ANALDB2R 17.300 Field-width enhancement to DB2PM-like reports
ANALDSET 17.240 MXG 17.03-17.06. DATA SET NOT FOUND corrected.
ANALDSET 17.343 Revised to use MACKEEP instead of IEBUPDTE, NEXTENT.
ANALDSOP 17.198 ANALDSET enhancements (42s, 30 interval, 62s, 6156).
ANALNSPY 17.358 Added and corrected a replica of MICS SNTNSS report.
ANALRMFR 17.160 All BY variables must be in first 4092 bytes.
ANALRMFR 17.369 Several enhancements, this was just most recent.
ASMDALO 17.061 The MXG DASD Allocation Monitor may 0C4.
ASMIMSL5 17.064 0C4 if &DFSMS left at 0 (for DFP) in IMS 5.1.
ASMIMSXx 17.172 Significant IMS Log processing enhancements for test.
ASMIMSxx 17.228 Major IMS Enhancements for MXG IMS log processing.
ASMIMSxx 17.315 Final IMS Revisions for all know negative values.
ASMTAPES 17.046 ML-19 of ASMTAPES suppresses TMNT005E messages.
ASUM70PR 17.045 Changing Interval with DURSET didn't change ASUM70PR.
ASUM70PR 17.163 ICF CPUs are now detected and deleted automatically
ASUM70PR 17.203 Dedicated CPU error fixed, use new PDB.ASUMCEC.
ASUM70PR 17.232 PDB.ASUMCEC now created BY CECSER vice BY SYSPLEX.
ASUM78CF 17.178 New ASUM78CF member summarizes PDB.TYPE78CF data.
ASUMDB2A 17.170 DB2 5.1/6.1 new variables added to DB2ACCT summary.
ASUMTALO 17.242 Lost output when multiple days are input is fixed.
ASUMTALO 17.320 Last-complete-interval now corrected.
ASUMTAPE 17.010 PDB.ASUMTAPE replacement for PDB.TYPETMNT.
ASUMTAPE 17.041 Don't use PDB.TYPETMNT. USE PDB.ASUMTAPE for mounts.
ASUMTAPE 17.106 PDB.ASUMTAPE with STATUS=MISSEDMNT handling revised.
ASUMUOW 17.324 WTIRIOTM no longer sum of all IR waits in ASUMUOW.
AUTOEXEC 17.392 Options S=72,S2=72 removed from AUTOEXEC and CONFIG.
BUILDPDB 17.025 Adding/Dropping variables from PDB.JOBS/STEPS/PRINT.
BUILDPDB 17.110 Vars EXCPNODD/IOTMNODD now corrected for MULTIDD='Y'.
BUILDPDB 17.111 Vars TAPEDRVS/TAPE3480/etc corrected for MULTIDD='Y'.
BUILDPDB 17.113 PDB.PRINT new variables SMF6PRMD and SMF6USID added.
BUILDPDB 17.176 OMVS/USS jobs fill SPIN, never purge, now forced out.
CICINTRV 17.391 CICINTRV dataset creation logic was wrong.
CONFIGV7 17.073 Revised CONFIG for SAS V7 eliminates warning msgs.
DOCMXG 17.051 Typos in MACKEEP= examples in documentation corrected
EX80ASEG 17.247 New TYPE80A exit for Top Secret unique segments.
EXPDBOUT 17.357 Example to add CICS Statistics datasets to your PDB.
FORMATS 17.222 Support for APAR OW39508 7060 Multiprise EIO and DSD.
IMACACCT 17.327 Order of code revised so &MACACCT now works.
IMACEXCL 17.229 Omegamon Exclude logic needed SMFPSRVR test.
IMACEXCL 17.356 CICS TS 1.3 Excluded Field example did not work.
MXGSAS 17.317 ABEND 2415 due to no RECFM=U in //NULLPDS in SAS proc
PRINTNL 17.381 Revised utility to print NEWSLTRS/CHANGESS members.
RMFINTRV 17.238 Non-contiguous shift definitions now supported
RMFINTRV 17.322 RMFINTRV now invokes VMXGRMFI, supports 115 workload
RMFINTRV 17.360 RMFINTRV/VMXGRMFI enhancements fixed and documented.
SPUNJOBS 17.311 DATASET CONDCODE NOT FOUND with old IMACPDB.
TYPE102 17.006 Support for DB2 type 102 subtype 199, Dataset I/O.
TYPE102 17.020 Typos. _C012297 should have been _C102297.
TYPE102 17.044 Typos. _C102206 should have been _C102106.
TYPE102 17.056 Support for Index Statistics in T102S022.
TYPE103 17.270 Support for APAR PQ28258 for SMF 103 record.
TYPE103 17.304 Enhanced TYPE1032 Web Server eliminates negatives.
TYPE103 17.314 Deaccum of TYPE1032 for duplicate IPADDRESS.
TYPE108 17.384 Support for Lotus Domino Server Release 5.02.
TYPE110 17.096 Support for GLRHTYPE=2 CICS TS 1.2 Journal segment
TYPE110 17.156 Support for CICS TS 1.3 new field (INCOMPATIBLE)!
TYPE110 17.279 CICS Statistics EOD record has missing DURATM.
TYPE115 17.248 Support for MQ Series Version 2.1 (COMPATIBLE).
TYPE21 17.013 TYPE21/PDB.TAPES variable OPEN is always blank.
TYPE28 17.018 NPM 2.4. Datasets NPMINSES/NPMEVSAL trashed.
TYPE28 17.049 Zero obs in dataset NPMSEEND corrected.
TYPE30 17.009 Support for deaccumulation of TYPE30_6 data.
TYPE30 17.176 TYPETASK='OMVS' instead of TYPETASK='STC ' for USS.
TYPE30 17.385 New foreign enclave times and TYPE30MR in OS/390 R29.
TYPE42 17.042 Type 42 subtype 16 (SMF42Gxx) was out of alignment.
TYPE42 17.059 Support for APAR OW37708/APAR OW38073 new fields.
TYPE42 17.124 Support for Type 42 subtypes 7/8 NFS Usage/Users.
TYPE42 17.278 Support for APAR OW40579/41407 SMF 42 subtype 4.
TYPE42 17.355 Type 42 subtypes 7/8 NFS caused INVALID NF-CL TRIPLET
TYPE50 17.007 Variables BSIZE and MXTRSIZE corrected in TYPE50.
TYPE64 17.032 Extended Format datasets, HIGHRBA now calculated.
TYPE7072 17.162 Support for APAR OW37565 identifies CP or ICF CPUs.
TYPE7072 17.299 CECSER wrong if CPUs added or deleted during interval
TYPE73 17.026 Support for APAR OW15406 (IODF Creation now YYYY).
TYPE73 17.027 Support for "FICON" channels adds fields compatibly.
TYPE73 17.286 PCHANBY/PNCHANBY wrong in initial FICON support.
TYPE74 17.161 Support for APAR OW37816, new 2105 cache TYPE74CA.
TYPE74 17.180 Support for APAR OW31701 ESS Parallel Access Volumes
TYPE74 17.182 TYPE74OM (OMVS/USS) had several variables wrong.
TYPE74 17.269 PCTPNCHA/PCTPNOTH/PCTDVPND/PCTPNDEV revisions.
TYPE74 17.378 Broken Type 74.4 RMF caused INPUT STATEMENT EXCEEDED.
TYPE74CF 17.211 Doc. XSYSn variable blank is most observations.
TYPE79 17.023 CPU time for Pre-emptible SRBs added in TYPE79s.
TYPE80A 17.012 RACF type 80 with optional RACFTYPE=7 had STOPOVER.
TYPE80A 17.094 RACF keywords specified/ignored are now decoded.
TYPE80A 17.158 Top Secret causes many SEGMENT SKIPPED messages.
TYPE80A 17.199 Support for OW39128, PDS DSNAME for PROGRAM access.
TYPE80A 17.218 Support for Top Secret Release 5.1 (INCOMPAT)
TYPE89 17.116 Support for APAR OW37091 Measured Usage SMF 89 change
TYPE90A 17.287 Replacement for TYPE90 member for SMF type 90 data.
TYPE91 17.298 Batch Pipes IC/OC counts propagated into 12/13/15.
TYPE94 17.213 Support for type 94 import/export statistics.
TYPE94 17.245 ERROR.VMAC94.AUDITLEN INVALID error corrected.
TYPE97 17.385 New in OS/390 Release 2.9
TYPEBETA 17.368 Support for Beta91 Balancing Manager SMF record.
TYPECIMS 17.303 IMF, MVIMF, CIMS: SQLCALLS not counted, INCOMPAT.
TYPEDB2 17.090 DB2STATS dataset some QXxxxxxx variables were wrong.
TYPEDB2 17.338 BPHITRAT, Buffer Pool Hit Ratio, revised.
TYPEDB2 17.382 DB2 TCB times QWACSPCP/QWACSPTT included in DB2TCBTM.
TYPEDCOL 17.244 DCOLLECT variables DCACSIZ/DCACACIC were missing.
TYPEDCOL 17.281 Support for DFSMS/MVS V1R5 - in place, no changes.
TYPEDCOL 17.307 Support for APAR OW41147 ORGEXPDT=99999 Y2K Critical
TYPEDCOL 17.347 Year 2000. IBM APAR OW42559, UCCOLDT in DCOLCAPD.
TYPEEDGR 17.016 Dataset EDGRDEXT has zero observations.
TYPEEPIL 17.003 Support for Candle V400 Omegamon for CICS Epilog.
TYPEEREP 17.339 INPUT STATEMENT EXCEEDED for EREP records.
TYPEHSM 17.019 HSM _DIFFHSM macro relocated into VMACHSM.
TYPEHSM 17.021 HSM Julian dates printed as 2E6, format now 7.
TYPEICE 17.048 Support for IXFP/ICEBERG Subtype 8 and fix for st 6.
TYPEICE 17.346 INVALID DATA FOR IOSSTIME, Iceberg IXFP subtype 8.
TYPEIDAP 17.100 Support for i-data afp-software SMF record.
TYPEIISL 17.321 Support for IIS Log.
TYPEIMSA 17.011 TYPEIMSA processing is wrong in 16.16. Use 17.08.
TYPEIMSA 17.290 More IMS Log revisions correct negative RESPNSTM.
TYPEIPAC 17.234 Support for Mobius View Direct 6.1.2 (INFOPAC).
TYPEITRF 17.336 Variables IMSVERS,IMSRELEASE,SMBCLASS numeric now.
TYPELDMS 17.371 Support for Software Innovation's LDMS product.
TYPEMIM 17.152 Support for MIM user record enhanced, new dataset.
TYPEMRKV 17.099 Support for Lexmark MarkVision Job Statistics
TYPENDM 17.155 Support for Connect Direct R 3.2 'CT' record.
TYPENOAM 17.122 Support for STK's NearOAM V2.2 (COMPATIBLE).
TYPENSPY 17.154 Zero obs in NSPYTIC3 (again, due to Change 16.147).
TYPENSPY 17.367 NETSPY NSPYAPPL dataset had missing response times.
TYPENTSM 17.055 Support for NTSMF new Quota Server object.
TYPENTSM 17.101 Support NTSMF Version 2.3 (COMPAT), 21 new objects.
TYPENTSM 17.165 Protection for Win 2000 Beta 3. IIS, Web changes.
TYPENTSM 17.209 Support for Lotus Notes, SMPTDS/SMTPRS objects.
TYPENTSM 17.335 Support for Windows 2000 Build 2195 NTSMF data.
TYPEORAC 17.308 Support for SQL*NET NIV adds IPADDR/PORTNR to ORACLE.
TYPEPMTR 17.297 Support for unix PerfMeter Freeware Monitor records.
TYPEQAPM 17.107 Support for OS/400 V4.3.0, no change, is in 16.16.
TYPEQAPM 17.235 Support for OS/400 Release 4.4.0 (LRECLs INCOMPAT)
TYPERACF 17.305 Support for RACF Unload IRRDBU00 Started Task subtype
TYPESARR 17.306 Support for CA View Metrics SARSMFUX SMF record.
TYPESARR 17.309 Support for remaining CA-VIEW Metrics validated.
TYPESFTA 17.092 Support for SOFTAUDIT 6.1.2 (COMPATIBLE).
TYPESFTA 17.123 Support for SoftAudit Version 7.1 (COMPATIBLE).
TYPESRMH 17.085 Support for RACAL IT Security's SRM product for HSM.
TYPESTC 17.040 STC SILO SMF record sometimes short.
TYPESTC 17.195 Support for STK's VTCS 2.2.0 INCOMPATIBLE VSM SMF.
TYPESTC 17.230 Variables STC07FPS/TPS now match STK utility report.
TYPESTC 17.313 Variables STC11CI/CE/TOL were blank.
TYPESYNC 17.145 SYNCSORT variables COREREQ/COREUSED now 8-byte store.
TYPESYNC 17.199 Support for 32 (up from 16) sortworks for SYNCSORT.
TYPESYNC 17.350 Flags CONTIGn,CACHFWn fixed, DYNALOn,UNOPENn added.
TYPETAND 17.037 Support for TANDEM F40, G04 and G05 INCOMPATIBLE.
TYPETAND 17.115 TANDEM G05 and later TANDDISK corrected.
TYPETCP 17.034 Unexpected TCP/IP command of STOU protected.
TYPETCP 17.349 TELNET LOGF Time field is Duration, not datetime!
TYPETELE 17.091 Support for TELEVIEW 4.3 (INCOMPATIBLE).
TYPETELE 17.246 Support for TeleView 4.3B subtype 3 record.
TYPETMDB 17.280 Support for Landmark DB2 Monitor V 3.2 (INCOMPAT).
TYPETMNT 17.216 ASMTAPES needed 'ES6' at ASM for Y2K, this protects.
TYPETMO2 17.169 Landmark TARSPTM contains sum of all conversations.
TYPETMS5 17.021 TMS Julian dates printed as 2E6, format now 7.
TYPETMS5 17.151 Undocumented DENX='DE'x TMS records now supported.
TYPETMS5 17.352 Year 2000. Variable OUTDATE was still 0cyyddd.
TYPETMV2 17.259 Support for TMON/MVS V2 PTF TD01655 (COMPAT).
TYPETMV2 17.333 Support for additional Landmark TMVS subtypes.
TYPETPF 17.200 Support for IBM's TPF Operating System records.
TYPETPX 17.036 TPX Start Up record subtype 1 not properly decoded.
TYPETRMS 17.284 Support for TRMS Version 51A08 (COMPATIBLE).
TYPEUNIC 17.002 TYPEUNIC support for CA UniCenter is for Open VMS.
TYPEZARA 17.344 Year 2000. Support for ZARA Release 1.3 (INCOMPAT).
UTANDSTR 17.241 Tandem Utility to read "Unstructured" with new LRECLs
UTILBLDP 17.054 Revised and documented utility to modify BUILDPDB.
UTILRMFI 17.271 UTILRMFI to show IMACWORK definitions CPU sources.
VMAC102 17.253 DB2 trace SQL text variables use &SASCHRLN length.
VMAC110 17.220 SMF type 110 subtype 0 JCRLL=0 error.
VMAC28 17.018 NPM 2.4, NPMINSES/NPMEVSAL datasets trashed.
VMAC80A 17.316 WARNING: BIT MASK TOO LONG corrected.
VMACDB2 17.300 Negative QXSELECT because DB2 overflowed counter!
VMACIMSA 17.011 MXG ASMIMSLG/L5/L6, STRTTIME is missing due to typo
VMACIMSA 17.228 SAP IMS 'AE'x log record was NOT Y2K Ready.
VMACTMDB 17.319 Landmark DB2 Version 3.0 INPUT STATEMENT EXCEEDED.
VMXGINIT 17.250 MXGVERS specification removed from VMXGINIT.
VMXGINIT 17.251 USER= option protected, &SASCHRLN= created
VMXGRMFI 17.142 Enhanced RMFINTRV permits over 100 workloads.
VMXGRMFI 17.388 Added DROPPGN=,DROPSRV= for easier RMFINTRV tailoring
VMXGSUM 17.193 Dashed-variable-list now correctly supported.
VMXGSUM 17.255 New stats, INHERIT exploitation under SAS V8.
VMXGSUM 17.265 VMXGSUM revisions now implemented in MXG 17.08.
VMXGVTOF 17.341 Year 2000. CREATED,EXPIRES,LASTUSE wrong.
WEEKBLDT 17.029 Text in col 72 causes unexpected failure.
XMXGSUM 17.370 Test version of VMXGSUM using SAS View to save DASD+.
YEAR2000 17.363 Year 2000. Julian 0cyyddd values convert to yyyyddd.
Inverse chronological list of all Changes:
Changes 17.398 thru 17.001 are contained in member CHANGESS.