II12069: DB2 OEM PRODUCTS, PROBLEM, FIX (CONT. OF II10701) BUGS IN DB2 AND IRLM ( CONTINUE IN II12697 )

APAR status

• Closed as canceled.

Error description

• ================================================================
  10/20/00 LSK
  00E90100 DB2 CRASH or 00E80100 ABND0C4 00000011 DSNTLRES + 0090
  at R510 or  0C4 00000004 R610 DSNTLRES  + 00C0 on BMC LoadPlus
  ReorgPlus or UnloadPlus job that is cancelled or experiences
  Timeouts . rc00e90100 rc00e80100 pic11 pic011 pic4 pic04 90 c0
  abend0c4 timeout . Please check with BMC for circumvention and
  future fix. Utilities Util reorg plus load unload
  LOADPLUS for DB2/5.1.00 Resolution ID: 79068 Failure ID: 226727
  ================================================================
  9/30/99 MHM
  MSGIECTMS3 - Tape was not scratch. (rc64)
  Potential workaround when SYSUT1 is defined for tape may be
  modifying TMS Exit TMSUX2F, but you may want to discuss this
  further with CA.
  Additional symptoms: ABEND0C4-14 (rc14) using UNIT=CART for DB2
                       Utilities.
  ================================================================
  10/29/99 rjl
  Overlay of db2 data in ecsa with data that has a format of a
  B205 STCK instruction or tod clock value. I.E. B2059... B205D...
  I.E. B2059118 B205D110
  The product is called HOTDATE 2000 by Softworks.
  .
  Contact Softworks for a fix.
  ================================================================
  11/16/99 lsk
   IMS 6.1 abendu03302 u03302 u3302 abend0c4 overlay s0c4
   abends0c4 CODE=S0c4             also
   DXR122E  ABEND UNDER IRLM TCB/SRB MODULE DXRRL246 0c4
   OEM vendor, Boole & Babbage ( now BMC ), when using their
   Mainview product because using its " CDE " feature could cause
   overlays. The BMC fix, PTF BPM7389 , addressed overlay problems
  ================================================================
  11/19/99 MHM
  Abend0c4 in CASR230D or CASR230F when running R510 Reorg and
  using ENFDB2.  May want to check with CA for fix LO38078.
  CASR230D failed when it was at 09/09/97 16.40 S920.
  .
  Abend0c4 in CASR230D when running DSNTEP2    (ljy 2000/08/28)
  CASR230D failed at 09/09/97 16.40 S920 level.
  Fix is CA90 T7M8546.
  =======================================
  UM(11/24/99)
  Missing data after DB2 RECOVER TORBA when
  using Full and Incremental Image copies
  created by oem (Platinum Quick Copy -PQC).
  Oem Utility was not using HPGCLRSN. Problem
  is resolved by applying HIPER PUT TAPE '9090AA'.
   UM ( 11/24/99).
  
  ==========================================================
  
  2-10-00 ESA  DSNTIAUL Utility performance.
  Symptoms: Under loaded system a jcl step running DSNTIAUL
  with a fast access path - SELECT * WITH UR
  was observed to run three to five times longer.  This was
  sporadic behavior and not consistent.  DB2 was attributed
  with using the additional CPU.
  .
  Customer identified BMC APPTUNE as causing the high
  CPU usage.  When not active, no problems with DSNTIAUL
  observed.  No known fix from BMC at this time.
  .
  According to the customer who reported this:
  " BMC replicated the problem at their own site....
  The problem is not just isolated to program DSNTIAUL."
  ===========================================
  04/20/00 rjl
  Abend0C5 in Astex's Start Subchannel hook. Dump captured will
  have title= COMPON=IOS,COMPID=SC1C3,ISSUER=IOSVSSCH,IOSSSCHF
  The problem has been seen with i/o to an ESS/2105 device for
  DB2 VSAM (media manager) data.
  After the abend0c5 an abend04e rc00c200c0 and abend04f
  rc00e50702 will be seen.
  .
  CA Astex support found the problem in their ASXIOHK module and
  provided fix JC27008 for Astex 2.7 (problem occured on 2.5).
  The fix for Astex 2.5 is F250071Z.
  ===============================================================
  
  JC 6/16/00
  The following symptoms might occur with the index created by
  OEM Utility. The OEM allocates the index directory page as 3,
  which is not compatible with DB2's directory page on page number
  4. Also see PQ39064 for details.
  ABEND0C4 -11 DSNB1CPL + 084
  ABEND0C4 -10 DSNISM + 322
  ABEND04E RC00C90101 DSNILGCL :5004 ERQUAL5004
  ABEND04E RC00C90101 DSNISGPI :2002 ERQUAL2002
  ABEND04E RC00C90101 DSNIOST2 :2004 ERQUAL2004
  ABEND04E RC00C90101 DSNISM   :5002 ERQUAL5002
  ABEND04E RC00C90101 DSNKIXDB :5003 ERQUAL5003
  ABEND0C4 -10 DSNKFTCH +1CA0
  ABEND0C4 -04 DSNKFTCH +0EFC
  ABEND0C4 -10 DSNKTRAV +4EB4
  ABEND0C4 -11 DSNUKGET +15A2
  ==========================================================
  JC 6/21/00
  DB2 might get the following error with Platinum REORGs.
  Platinum ZAP is COM11006.
  
  ABEND0C4 DSNGEDLC (UQ40803) + 20AC
  
  ============================================================
  07/12/00 HT
  Non clustered indexes fail on V5 with error:
  00C90101 Type 303 ERQUAL 5002 when PTF UQ36705 (9911)
  is on V5 and a REORG was done on V4 using CDB Software
  Company's REORG.
  Options:
  1.  Recover the indexes with CDB S/W fix #29 on their product
  2.  Use REPAIR utility to zap the bad bit in the header
      (do not know which bit)
  3.  Back off PTF UQ36705 (at least until option 1 or 2 is done)
  .
  =============================================================
  09/20/00  ESA
  Rebuild index fails with dups on a table with a unique index
  .
  DB2 R610 Put 0004
  REORG using OEM ( CDB ) went down with a corrupt index.
  IBM Utility REBUILD INDEX fails.  CDB has fixes.
  Use all IBM Utilities corrects problem.
  .
  .
  Key words: CDB OEM REORG REBUILD INDEX 3rd party vendor
  ==============================================================
  07/26/00  LSK
  abend0c4 DXRRL247 +01C2 after appling uq41353. RHT table
  overlaid by  MAINVIEW for DBCTL problem. Fix BPI7842
  Overlaid the IRLM area came from an Mainview module IMEDDTB9 .
  IMF PTF RR320 ADIST HIPER 00/06/29 29JUN00
  ===============================================================
  7/26/00 CSH
  MSGDSNU017I with ABEND04E DSNUGBAC 00E40119
  See pmr 00994,033,000
  Customer had made incremental copies using BMC utility.
  Customer called BMC and found out BMC's incremental copy copies
  all changed pages since last full copy. It updates all previous
  incremental copies with a small 'i' in SYSCOPY. They have
  another feature that allows you to use the 'i' incremental
  copies for a point-in-time recovery.
  If you use BMC COPYPLUS you cannot use IBM'S MERGECOPY utility.
  ===============================================================
  8/07/99 Baez
  TS was modified then a FASTLOAD received abend0c9 in SS37X02.
  Recovery failed with rc00e40404, DSNUCBRP with reg2 00C20064.
  May want to contact Syncsort (CA) for similar problems.
  ================================================================
  08/08/00  LSK
  abend0c4 in dxrrl050 + 31c at pq14436 overlaid of the IRLM EOT
  block.   CA - Platinum fixes are: smp/e install UCMB115 and
  UGEB401 on tape number 0201AA. Non-SMP/E customers on 0214BA.
    Other symptoms include : abend0c4 DXRRL24H + 01c6 rhwa segment
    overlaid. offset 1c6
  ================================================================
  8/22/00   JAK
  SQLCODE553  on SET CURRENT SQLID = 'xxx'
  SQLCODE557  on GRANT SELECT ON DATABASE xxx TO XXX
  The problem was with the ACF2 6.1 to 6.2 upgrade, the DSN3@ATH
  was not relinked using the new 6.2 modules.
  There is a Computer Associates contact issue number 9226999
  that describes the problem and resolution in detail.
  
  ************************************
  Added by UM 09/14/00
  PROBLEM SUMMARY:Poor Sorting Performance in DB2V6/R610
  
  Poor sorting performance using Syncsort(TM) with DB2 V6
  REORG SORTKEYS, REORG SHRLEVEL CHANGE, LOAD SORTKEYS, or
  REBUILD INDEX SORTKEYS.  Possible out of storage abends
  such as ABEND80A or ABEND878 when using Syncsort with
  one of the these utilities.
  In DB2 V6, support for building indexes in parallel was
  added to the REORG, LOAD, and REBUILD INDEX utilities.
  When this feature is activated by the SORTKEYS keyword,
  multiple subtasks are attached to perform index key sorts.
  When the sort product invoked is Syncsort, the individual
  sorts contend for virtual storage (both above and below
  the line), with each sort attempting to allocate as much
  storage as it can.  This can result in sub-optimal
  perfomance, as when one sort obtains most of the virtual
  storage available, leaving only small amounts for the
  other sorts, and can occasionally result in 80A or 878
  abends indicating an out of storage condition.
  The problem can be detected by examining the messages
  issued by Syncsort.  There will be one set of messages
  for each sort, and one of the messages will indicate how
  much storage was obtained by that sort. If one message
  indicates (for example) that 19M was obtained, but the
  same message for the next sort indicates that only 1M was
  obtained, then performance of the second sort will suffer.
  PROBLEM RESOLUTION:
  To rectify this situation, we (and Syncsort technical
  support) recommend that you choose one of the following
  3 options:
  1) Modify the IEFUSI exit to allow the utility job to use
     more virtual storage above the line.  Allow for as much
     as Syncsort is configured by default to obtain, times
     the number of sort tasks the utility will start(generally
     this is the number of indexes to be built, but see the
     DB2 V6 Utility Guide and Reference).  This will allow
     the sorts to get as much storage as they need and so
     will allow them to sort as efficiently as possible.
  2) Modify the utility job JCL to specify a REGION size
     (calculated as in (1)) sufficient to allow all the
     sorts to obtain as much storage as they need.
  3) Provide a sort parms data set with DD name $ORTPARM
     containing a Syncsort control statement with keywords
     limiting the amount of storage each sort should obtain.
     Please contact Syncsort technical support to determine
     which keywords should be specified and which values
     will provide the best performance for your environment.
  .
  Note: SYNCSORT assisted several customers in modifying their
  parameters.  DSM ( dynamic storage management ) should be off.
  Below is one customer's successful sort options:
  .
  MINCORE=300K,VSCORE=400K,VSCORET=6M,CORE=MAX,DSM=OFF
  .
  ================================================================
  10/26/2000 Kai-Li Kow
  ABEND0C4 Pic 4 module unknown during connect to DB2 for os/390
  PSW in dump showed the following:
  | ............IWMN |
  | TFY EXECUTION TI |
  | ME INVALID  IWMI |
  | 2NUC#LNTFPLIST   |
  | CALLERPB  XDNTF0 |
  | 01IWMX2NTFWCDAPB |
  |   SYMPWARB...... |
  | IWMW2CLS03/05/99 |
  | UW57049 IEANUC01 |
  | ***WORKLOAD--MAN |
  | AGER***.IWMWMCLS |
  .
  R14 points to DSNLIRTR for the calling module with Key7 issuing
  WLMCLSFY and is trying to store in key 0 storage.
  .
  0C4 on entry (or near the entry) to IWMW2CLS where the caller of
  IWMW2CLS is DSNLIRTR (TCP/IP access) or DSNLCRTD (SNA access).
  Compuware Xpediter (Beta version) product intercepts WLM calls a
  and incorrectly puts us in Key 0, or causes this WCDA_DYN to be
  obtained in key 0 storage.
  .
  The customer should disable this Xpediter intercept processing
  and recycle DB2 to correct this abend0c4 prob.
  Compuware support indicates that this problem was fixed by
  their internal PTM 121684 which will be included in their
  GA version of the product.
  ================================================================
  12/7/00 SGC - SRB Spin Loop - LDSDMSGT LDSDINCR
  An SRB loop can occur when attempting an -ARCHIVE LOG command
  after adding active log data sets via Computer Associates (CA)
  24x7 product.
  The loop will appear in DSNJW107 with registers 3 and 4 both
  equal to zero.
  CA fix: LO46209 Product: 24X7 Release:5.2 Solution #:1 Z52036
  ================================================================
  Kai-Li Kow 05/07/2001
  High CPU in Dist address space caused by looping in BMC APPTUNE
  module found in db2 console dump as follows: ( ASQXISQL )
  | ...h......|.ASQS |
  | IDB2.".Q...¢..`. |
  | .....SB...|-..}. |
  | .{.{...0{h.ASQXI |
  | SQL 002000 3.0.0 |
  | 0 20000327-10.34 |
  |  BQ09250-BQ10317 |
  | -BQ10646-BQ14831 |
  BMC supplied the patches to the customer and recommended upgrade
  to BMC V3 to correct the looping problem.
  ================================================================
  Kai-Li Kow 5/8/01 ( Candle KXDJESSC )
  Lots of entries in logrec showing
  abend0c4 pic10 KXDJESSC in Candle's Omegview job
  either preceding or following storage abends with abend878 or
  abend80a or abend04e rc00e20028 on dsnsvstk or other abend0c4s
  in db2 clean up or recovery modules causing db2 crash.
  .
  The problem with Candle was fixed by reassembling the JES offset
  table.
  ================================================================
  09/05/2001 Kai-Li Kow / Linus Yeh
  HIGH CPU seen in DDF xxxdist asid (on and off)
  Customer has seen TMON reporting 90+% cpu spent in Nucleus
  IGVGPVT IGVEGPVT IGVFSDQE
  This is the PC that called IGVVSTOR (PC 00311).
  The PC caller is SQL CAPTURE for DB2 which an optional part of
  TMON for DB2 by LANDMARK and the fix is TE01093.
  ===================================================
  

Local fix

Problem summary

Problem conclusion

Temporary fix

Comments

• Closed for Internet viewing.. DB2INFO  VTAMINFO  VSAMINFO
  

APAR Information

• APAR is sysrouted FROM one or more of the following:

• APAR is sysrouted TO one or more of the following:

Fix information

Applicable component levels