II11093: INFO ITEMS FOR IMS SYSTEMS. PRIOR II10757 NEXT II11241 *CLOSED*FOR FUTURE UPDATES SEE THE "HTTP://WWW.IBM.COM/IMS" SUPPORT PAGE

APAR status

• Closed as canceled.

Error description

• ******************** JANUARY 10, 2002 **************************
  RETAIN INFORMATION ITEMS WILL NO LONGER BE USED TO DOCUMENT
  COMMON USER ERROR INFORMATION.  THE NEW PROCESS WILL RELAY THIS
  INFORMATION VIA THE SUPPORT PAGE WITHIN THE IMS HOME WEBSITE AT
          "HTTP://WWW.IBM.COM/IMS"
  CLICK ON THE "SUPPORT" KEYWORD IN THE LEFT MENU AREA.
  ****************************************************************
    This apar is an extension of II02886, II04204, II05221,
  II05470, II07766, II08109, II08512, II08893, II09120, II09411,
  II09872, II10332, and II10757.  Its purpose is to document
  common problems, diagnostic procedures, and/or solutions in the
  Systems area of IMS.
    When adding new entries, be sure to add an index item to
  briefly describe the problem symptom.  Include the description
  page number too.  When closing your description update, include
  your name and current date.  Common Search Arguments: IMSIISYS
  565515800 R600 R610
  569517609 569517611 R510 RA02 R100 R112 R012 R106 R116 R126
  568501201 568501202 568501301 R410 R102 R154 R164 R114 R122
  566540801 566540802 566540901 R310 R152 R162
  HMK6600 JMK6601 JMK6602 JMK6603 JMK6604 JMK6605
  HMK5100 JMK5102 JMK5106 JMK5164 JMK5114 JMK5112 JMK5116 JMK5126
  HMK4102 HMH4102 HMJ4102 JMH4112 JMJ4112 JMJ4122 JMK4162 JMK4152
  HMK3102 HMH3102 JMH3112 HMJ3102 JMK3112 JMK3162 JMK3152
  -----------------  PROBLEM INDEX  -----------------------------
  1. ABEND0C4 in DFSFXC50 at offset x'153C' or ABENDU0014 PG.  E1
  2. MSGGIM42001W & GIM35933I running IV_B314J            PG.  E2
  3. MSGGIM30206E running IV_B305J - IV_B311J             PG.  E2
  4. ABEND0A9 ABENDS0A9 RC10F1 000010F1.                  PG.  E3
  5. MSGDFS662I FIRST POSITIONAL PARAM INVALID            PG.  E4
  6. IMS abends under LE have altered behaviour           PG.  E5
  7. MSGDFS681I DFS681I not present in BMP JES JOBLOG     PG.  E6
  8. MSGDFS2499W RC01 for batch job in IMS 6.1            PG.  E7
  9. CMDMCS parm need not be specified in DBC proc        PG.  E8
  10. Security parms require a COLD start to take effect  PG.  E8
  11. DBCTL ECSA creep SP231 Key7 RAT control blocks      PG.  E9
  12. DIAG  Incorrect LUMI x'7F01' Trace record descript. PG. E10
  13. Abend878 rc10                                       PG. E11
  14. ABEND0C1 branch from ACF9292A ACF2                  PG. E12
  15. AbendU0261 SNAP dumps issued/how to get svcdump     PG. E13
  16. MsgDXR116E during IVP IV_F201J rc=17.               PG. E14
  17. msggim40501e processing uq04302 for dfsmsgrt        PG. E15
  18. abendu0476 in DFSPR000 because pcb address is zero  PG. E16
  19. Abend0C1 Abend0C4 with corrupted savearea and regs  PG. E17
  20. MSGIEA602I and hitting MAXUSER limit at IMS 6.1     PG. E19
  21. ABEND0C4 in DFSESPR0 branching to DFSESGL0.         PG. E20
  22. ABENDU0758 with QBUFMAX set in IMS 6.1 environment. PG. E21
  23. Structure checkpoint invalidates CQS logtoken       PG. E22
  24. ABENDU3042 RC=00130017                              PG. E23
  25. MASSIVE STORAGE OVERLAY or other catastrophies      PG. E24
  26. ABEND0C4 in ASMTDLI running CICS transactions       PG. E25
  27. MSGDFS2499W RC01 RC1 using QuickFetch               PG. E26
  28. High SRBTIME in IMS address space                   PG. E27
  29. Poor performance due to PARDLI on for all regions   PG. E28
  XX. FUTURE UPDATES WILL BE MADE WITHIN THE IBM SUPPORT
      WEBSITE AT "WWW.IBM.COM/IMS".  THE SUPPORT OPTION IS ON
      THE LEFT MENU.  (January 10, 2002)
  

Local fix

Problem summary

• 1. Dependent region gets ABEND0C4 in module DFSFXC50 in CSECT
     SYNCLOW at offset x' 153C '.  This is caused by an overlay
     of a pst with the following Omegamon message:
     ' >> OI061: Broken control block chain terminated command  -
     try again <<'. Addit. Symptoms: ABENDU3058 SC07 DFSBCB30
     DFSCWU00. PSTAQCHN overlaid.
     This overlay can also cause an ABENDU0014 in DFSFLLG0
     with RC20 ( RSN20 ) - Invalid Post Code.  The overlay
     data will begin in the PSTECB with hex value 406E6E40 .
     Look for a pst with field PSTAQCHN hex 83888189 or 03888189
     Module DFSCBTS0 returns this value in REG1 .
     The CANDLE fixes for this problem are QI28720 for version 400
     of Omegamon, and QI28510 for version 300.
  ================================= Weed / Maddix 03/18/98 =======
  2) MSGGIM42001W and MSGGIM35933I indicating a problem with
     requisite sysmod UQ09576 will be received when running
     ivp job IVB314J.  This causes a return code of 04 for
     job.  These messages should be ignored.
  =========================== Alan K. Ho 03/04/98 ============
  3) The new Enhanced Holddata (January 1998) will cause problems
     when running the install/ivp SMPE jobs to process IMS fmids
     (jobs IV_B303J THRU IV_B315J depends on the release of IMS).
     These jobs will have TO be modified to add a bypass of
     'HOLDDATA(HIPER)'.  Failure to do so will cause MSGGIM30206E
     when processing these jobs.  IMS 5.1 apar PQ13554 has been
     created to address this problem and will forward fitted into
     IMS 6.1.
  =============================== Alan K. Ho 03/04/98 ============
  4. ABEND0A9 is issued by the VTAM FRR routine to ensure a
  dump is received in the application region (in this case,
  IMS).  ABEND0A9 is issued by the VTAM FRR due to a prior
  abend which caused that FRR to gain control.  The original
  abend is the source of the problem and can be found in
  MVS LOGREC.  If ABEND0A9 occurs in the IMS control region,
  IMS will terminate.
  =============================== Jeff Maddix 03/06/98 ===========
  5. MSGDFS662I on CICS startup:
  DFS662I job.step.proc--FIRST POSITIONAL PARAM IN PARM FIELD IS
  INVALID
  This message will be issued for releases of IMS 5.1 and 6.1 at
  CICS startup if the CICS SIT parms specify Local DL/I.  Local
  DL/I is not supported after release 4.1.  Also, ABENDU0662 will
  be issued.  To correct, remove the Local DL/I parms from SIT.
  =============================== Linda Weed 03/10/98 ============
  6. Under OS/VS COBOL and COBOL II the user is able to create an
     ESTAE which will gain control when an abend occurs within
     IMS. In the LE environment, this is no longer true. APARs
     PN42272, PN61809 & PN42578 were created to resolve abends
     occuring in the LE/IMS environment due to LE's way of
     handling abends.
  ============================== Kevin Bracken 03/12/98 ==========
  7. MSGDFS681I will not be received in the JES JOBLOG for the
  BMP region because the message is issued by the DBCTL region.
  This is done by design and the message should not be expected
  in the BMP JES JOBLOG.  This point is often noticed in the
  conversion process from local DLI to the DBCTL environment.
  ========================= Jeff Maddix / Fred Wiegert 03/17/98 ==
  8. MSGDFS2499W RC(01) Invalid Allocate For Database xxxxxxx for
  Reason 1 is received when running a batch job that does not
  not use dynamic allocation and does not have a corresponding
  dd statement for all database datasets in the PSB.
  This does not affect the running of job.  The job will
  run the same as it did in prior to IMS 6.1, but this message
  will be issued.  The message is issued validly in this case
  as the dd statement is truly not present.  Prior to IMS 6.1
  this message was suppressed in this case for batch jobs.
  The message can be ignored.  Another option is to specify
  NODYNALLOC in the DFSVSMxx member which will cause dynamic
  allocation to be turned of for this batch job (note this
  will cause MSGDFS2480 to be issued). Adding the corresponding
  DD statements will correct this error, thus this message will
  not be issued. APAR PQ16367 will restore previous behaviour.
  ====================== Butterweck 04/24/98 =====================
  9. In DBCTL environment, the CMDMCS parm of the DBC proc is
  optional.  If it is not specified, the default taken will be to
  allow input from an MCS console with none of the command
  security options in effect.
  =================================Weed 03/27/98 =================
  10. Security parameters (SGN=, TRN= and RCF=) cannot be changed
     with a WARM start (/NRE). Either a cold start (/NRE CHECK-
     POINT 0), or an emergency restart with one of these options
     are required (/ERE COLDCOMM | COLDSYS). Review the IMS/ESA
     Operator's Reference (/NRE and /ERE commands) for details.
     From the Operator's Reference manual:
      For an IMS cold start, the base security definition is
      created from the IMS system definition and EXEC
      parameter specifications. For an /NRESTART warm
      restart, the base security definition is created from the
      IMS checkpoint data.
  =====================  BRACKEN  3/31/98 ========================
  11. ECSA SP231 KEY7 storage creep of RAT control blocks (also
      you will see an eyecatcher of TMWK ) for dbctl threads.
      This problem is caused by Boole & Babbage product IMF 3.2,
      ptf BP17153.  They have a zap to fix it (tracking number is
      30986).
  ====================================== Weed 3/31/98 ============
  12. The LUMI X'7F01' trace record is incorrect as per the
  description in the IMS/ESA Diagnosis Guide & Reference. The
  source LU name is actually contained in words 5-6 and thus
  word 6 is not zero as described. Example Below (note word7
  is not shown in this example):
  WORD0    WORD1    WORD2    WORD3    WORD4    WORD5    WORD6
  7F01000B 00000000 00000001 0C613910 0000006D C9D4E2C1 D3E4F6F2
                                               I M S A  L U 6 2
  ================================= HELLIWELL 3/31/98 ==========
  13. Abend878 rc10 in the DLI address space. From the VSMDATA,
  it will be SP252 KEY0 that has taken all PVT & EPVT. They will
  all be x'00002000' bytes in length. The failing getmain is
  likely to be for a subpool other than 252. I used the GM/FM
  trace specifying SP0 KEY0 to find the looping getmain since
  SP0 KEY0 is converted to SP252. The above is caused by BMC
  product XBM ( Extended Buffer Manager ). This product is
  using PC 30B for the getmain request and is being done in
  module XBMICSIO.
  ===================== Wiegert  4/09/98 =========================
  14. ABEND0C1 - wild branch from ACF2 module ACF9292A+x'02D6'.
      IMS 6.1. running ivp job IV2G207J.  This problem is fixed
      by CA's ACF2 fix L029827. abend0c4 abend0c6 open svc 13
  =========================Butterweck/ Weed 8/19/98 ==============
  15. AbendU0261 U0261 in this case the customer was running
  in a DBCTL environment.  SNAP dumps are issued for the abends
  and are filling up storage in the CICS address space.  There
  is no way to suppress these snap dumps, other than with a zap
  in the module that issues the snaps (DFSPSNP0).  The abend
  is occurring because the application is passing a bad
  parameter.  To get an svcdump for this problem, setting a slip
  on abendU0261 will not work.  You need to set an IF slip in
  the module that issues the abend, or zero out an opcode where
  the abend is issued and slip on 0C1.  Determining the module
  where the abend is issued is difficult.  It is issued by
  DFSDPRH0, but we tried slips in there and they didn't match.
  Then tried a slip in DFSPDLI0 at label ERRPCB - this matched.
  The abend is occurring because the PCB address being passed
  to us is not within the DIRCA range.  The DIRCA contains a
  list of PCB addresses - if the PCB being passed is not in
  the list (or at least within the range of these addresses)
  then the abend is issued.  They need to check the application
  to make sure the PCBs match those that are in the PSB.
  ========================= Plaskett 4/22/98 ==================
  16. DXR116E RC=17 during IVP job IV_F201J (V6 ex: IV2F201J).
  If this error is ignored, and step IV_F204J is then run, a
  msgDFS039I RC 16-80 is returned. These occur when the IRLM
  parm SCOPE=GLOBAL is defaulted to. This is only valid when a
  sysplex environment exists. Solution is to code SCOPE=LOCAL.
  ========================= Crandall 4/28/98 ==================
  17. msggim40501e when processing uq04320 for dfsmsgrt.  The
  special instructions for processing do not work.  The problem
  is that dfsmsgrt has a distlib of tmsource but the ptf
  shows a distlib of genlibb.  This should have been corrected
  of IMS built internally.  The smp/e zone shows
  un75598 was sup'd by the fmid jmk5106.
  The following uclin will need to be run to correct this
  problem:
    SET BDY(target zone name) .
    UCLIN .
     REP SRC(DFSMSGRT) DISTLIB(GENLIBB) .
    ENDUCL .
  Repeat the same for the dlib zone.
  ========================= Butterweck 5/4/98 ===============
  18) ABENDU0476 in DFSPR000 because pcb address is zero.
  We have seen this when compiling cobol programs because the
  entry point address is 0 ( x'00000000' ) and " entry dlitcbl
  using pcb1 pcb2 etc..." is specified in the procdeure division.
  In this situation the entry point needs to be that of
  this entry statement.
  One way to resolve this is to change the cobol program removing
  this "entry dlitcbl" statement and modify the "procedure
  division" statement to be "procedure division using pcb1
  pcb2 etc..."
  ======================= Butterweck 5/6/98 ================
  19. Abend0C1 or abend0CX with corrupted registers or
  garbage in the regs.  In this case most of the regs had
  zeroes, reg13 had 0000FFFF.  This abend was followed by
  abendU0113.  The savearea flow for the region that got the
  abend0C1 is: DFSTMS00, DFSFXC30, DFSSYI40, DFSFXC40, DFSICLR2,
  ICLRENQ0, ICLRQMGR, DFSPREQ0, DFSQMGR0, DFSQEQ00, DFSQC050,
  DFSQBFM0, DFSQC030, UNK.  This is due to the problem in apar
  PN77456.  The customer had a back-level module DFSKMPX0.
  Other customers that experience similar problems may have
  local compression edit routines that need to be modified also.
  This change was introduced at IMS 4.1, so customers upgrading
  from IMS 3.1 to 5.1 may see the problem.
  ====================== Plaskett 5/6/98 ======================
  20. IEA602I ADDRESS SPACE CREATE FAILED after upgrading to
  IMS 6.1.  The MAXUSER limit in MVS has been reached.  This
  is due to the fact that module DFSRRSI0, which is a new module
  at 6.1, obtains a system LX via the LXRES macro.  This is done
  to set up the correct environment for the RRS/MVS recovery
  platform.  IMS is required to get the system LX.  By doing so,
  this means that when IMS is brought down, the address spaces
  associated with the CTL region will be marked as non-reusable
  until the next IPL of MVS.  This can cause the MAXUSER limit
  to be reached if IMS is brought down and restarted often.
  The customer needs to increase the RSVNONR value in the
  IEASYSxx parmlib member.  The MVS Initialization and Tuning
  Reference has details about this parameter.  It is also
  discussed in info apar II05264.
  ===================== Plaskett 5/6/98 ======================
  21.ABEND0C4 in DFSESPR0 attempting to branch to DFSESGL0 at
  offset x'cee'. The branch is occuring under label espmstop
  and is occuring because the storage at r15 has been
  freed.  Storage in the dump has been corrupted and many
  of the areas used for diagnosis are corrupted as well.
  The address in r15 previously pointed to module IMEESGL9.
  This occurred on a pre-ga version of IMF.  The customer
  used a ga version and the problem was resolved.
  ==================== Butterweck 5/13/98 ==================
  22. Abend U0758 in IMS V6.1 shared message queues
  environment with QBUFMAX set. If QBUFMAX is set and too low,
  then it may result in all available drrns being exhausted and
  a U0758 abend. It's recommended that QBUF is specified only,
  as then extra space in the QPOOL will be dynamically added as
  and when required. The use of QBUFMAX sets up a high watermark
  on the amount of available QPOOL space and thus if not set
  correctly exposes customers to possible U0758 abends.
  The QPOOL is used to build messages before they get enqueued
  on the shared message queues.
  =================== Matt Helliwell 05/11/98 ==================
  23. If a CQS is down and a structure checkpoint taken, then
  that CQSs log records will be deleted and on a subsequent
  restart message CQS0032A will be issued. Since the logtoken
  has been invalidated through the delete, then the only valid
  responses are COLD or CANCEL. To avoid this situation, a
  structure checkpoint should be taken with all the CQSs up.
  =================== Matt Helliwell 05/11/98 ==================
  24. ABENDU3042 RC=00130017.  This can occur when one of the
  datasets in the dfsesl ( //dfsesl ) concatenation are not apf
  authorized.
  Additional Keywords: rc(00130017) 00130017 U3042 3042
  ============= Butterweck 05/28/98 ========================
  25. Dump analysis showed that a DBCTL purge SSI function was
  being performed.  Module DFSASK00 was processing this function
  from PROCDBCT ( ssob function code was x'0034' 52 SSOBPSGN ,
  ssob subfunction CODE was x'000C' 12 SSPSSUBF).  Program old psw
  at low core address x'2c' was picked up after label DBCROUT, and
  used as a branch address. The old psw routine happened to
  point to a MVCL instruction, and started overlaying ( overlay )
  portions of storage.  If reg12 is not changed by the "branch to"
  routine, it will point to just after eyecatcher AWDBCTL ROUTING.
   This problem can result if apar PN82329 is invalidly processed.
  PN82329 contains HOLDDATA to guide the user through proper
  installation. The AI information from PN82329 / UN90698 follows:
  AI - 08/22/96 THIS PTF CONTAINS A HOLD ACTION THAT STATES
  THE SERVICE MUST BE ACCEPTED AND A MINIMUM OF AN ON-LINE
  SYSGEN TO BE PERFORMED.  THIS SERVICE CAN BE SUCCESSFULLY
  PROCESSED BY EITHER OF THE FOLLOWING METHODS:
  1.2... See apar PN82329 for the special instructions.
  ===================================Weed/Maddix 05/29/98=========
  26. ABEND0C4 in ASMTDLI running CICS transactions under
  label licom executing the instruction st 1,liparm .
  This can occur when the IMS version of the language interface
  module is linked with the CICS application.  CICS provides
  a version of the language interface module that must be
  used when running in the CICS environment.
  This is because pxparms ( DFSPRPX0 ) is not used in the
  same manner.
  Additional Keywords:  DFSLI000 CBLTDLI PLITDLI
  ================================ Butterweck 06/02/98 ======
  27. MSGDFS2499W RC01 using Quick Fetch.  IF slip in DFSMDA00
  after MDA0312 (IMODULE LOAD) showed that QFSFETCH was getting
  control in the type2 SVC expansion, doing getmains.  On exit
  from the type2 SVC, DFSMDA00 was using the wrong DFSMDA member
  (wrong name).  Problem went away when Quickfetch was disabled.
  ************************************* Ostdiek 6/3/98 *********
  28. Slow response time in IMS.  IMF and STROBE report high CPU
  use in CTL rgn especially SRB time.  Systrace shows SRB activity
  for IMS DFSIXMPT, BMC module SOPMVTAX and VTAM module ISTAICPT
  (BMC and VTAM most frequent).  Problem not resolved by cycling
  IMS - and IPL must be done.  It was relieved when BMC UltraOpt
  was stopped.  Ultraopt parm FREEBWA was altered to disallow
  getmains / freemains and CPU usage is back to normal.
  SRBTIME PERF PERFM PERFORMANCE
  ***************************************** Ostdiek 6/4/98 ******
  29. IMS performance slow due to frequent iswitches.  Dumps
  showed that PARDLI flag PSTSWTCH was on for all MPPs causing
  a bottleneck in CTL TCB.  Sample SA flow: DFSDLA30 - DFSISWIT
  PARDLI is normally set only for BMPs.  Found customer was using
  a copy of PXPARMS PARMBLK DFSPRPX0 with RCPNDLI set on for all
  regions.  GENLIBB copy of DFSPRPX0 had RCPNDLI off but reslib
  copy had it on.  Presumably this was set by a usermod.  Perf
  degradation is worse due to Candle's AFOper AFOperator which
  was involved getting more frequent MVS locks like CML.
  **************************************** Ostdiek 6/4/98 ******
  

Problem conclusion

Temporary fix

Comments

• IMS Information APAR
  

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