IBM Support

II07347: ABEND878, ABEND80A WHEN DATASETS SMS MANAGED. THIS APAR WILL BE UPDATED AS NEW INFO COMES TO LIGHT.

Subscribe

You can track all active APARs for this component.

APAR status

  • Closed as canceled.

Error description

  • NOTE that all of the information in this Info APAR applies to
    to batch jobs, Started Tasks (STC) and TSO sessions (TSU).
    
         Following conversion to SMS management, jobs / tasks may
    suffer a variety of virtual storage abends such as abend822
    rc14 (IEF085I REGION UNAVAILABLE, ERROR CODE = d'20' = rc20),
    abend878 rc10 or abend80A rc10.  There are several added
    requirements of virtual storage for SMS managed jobs/tasks
    versus the same job/task not SMS managed.  For SMS managed
    jobs:
    
    1. A x'C0' byte SWB will be built for each data set with DISP =
       NEW or DISP=MOD in the job.  This SWB is built during C/I
       (Converter / Interpreter processing) prior to the job's
       beginning execution.  More than one SWB may be built at
       this time.
    
    2. A x'C0' byte SWB will be built for each ds with DISP=OLD or
       DISP=SHR during step allocation after the job has begun
       execution.  More than one SWB may be built at this time.
    
    3. A JFCBX (JFCB eXtension) is built for each ds in a job
       when:                            (JFCX = x'B0' bytes)
          A: The dataset is a candidate for multiple volumes,
          B: The Dataset name is an alias, or
          C: The dataset is to be cataloged.
    
    Since SMS requires data sets it is to manage to be cataloged,
    this may cause additional JFCX's to be built.
    
    Note: All SWA blocks have a x'10' byte "prefix", making (e.g.)
          a x'B0' byte JFCX actually occupy x'C0' bytes of storage,
          and a x'C0' byte SWB occupy x'D0' bytes of storage.
    
    ********
         An increase in virtual storage use is to be expected when
    jobs become SMS managed, based upon the variables described
    above.  Info apar II05506 may be used as a diagnostic technique
    to confirm the cause of virtual storage abends.
    ********
    
    Possible circumventions:
    1) Decrease the REGION parameter of the job,
       to make more storage available for the Scheduler Work Area.
       SWA is part of the PRIVATE AREA but not part of REGION.
    2) Split the job into two or more smaller jobs, or
    3) For JES3, use inish statement CIPARM PARM= to cause SWA
       blocks to be GETMAINED above the 16meg line.
       For JES2, use command $T JOBCLASS(x),SWA=ABOVE where x is
       jobclass, STC, TSU, or wildcard (* or ?) to identify classes
       to be changed. Be sure to also update the JES2 init deck
       JOBCLASS statement.
    4) To control the placement of SWA for 1 particular JOB, one
       would need to write an IEFUJV exit and pass x'04' in word
       4 of the IEFUJV input parmlist.  See your OS/390 or z/OS
       Installation Exits manual for IEFUJV specifics.  You should
       find an IEFUJV example exit in the MVS sample library that
       comes with your MVS release.
    5) Decrease your CSA allocation
    
    As design changes are made, the particulars of this issue may
    also change.  This info apar will be updated from time to time
    with new details or amendments as they become known.
    
    A. APAR OW23626 (incorporated into base code for OS/390 V2R5 =
       HBB6605 = R605) created the possibility of requiring one
       additional SWB for SMS managed datasets.
    B. APAR OW43715 (incorporated into base code for OS/390 V2R10 =
       HBB7703 = R703) created the possibility of requiring one
       additional SWB for SMS managed datasets.
    C. APAR OW44540 created the possibility of requiring one
       additional SWB for SMS managed datasets.
    D. APAR OW50040 created the probability of REDUCING the
       requirement for SWBs for SMS managed datasets by two.
    

Local fix

  • 1) Decrease batch jobs region size to make more storage
       available for the Scheduler Work Area.
    2) Split the job into 2 or more smaller jobs, or
    3) Use the jes2 init parm JOBCLASS(v), keyword SWA(ABOVE)
       or the jes3 inish parm CIPARM to cause swa blocks to be
       built above the 16m line.
    4) Write an IEFUJV exit to pass X'04' in word 4 to MVS C/I
    5) Decrease CSA allocation.
    
    Additional Keywords:
         s822 abends822 MSGIEF085I, s878 abends878, s80A abends80A,
         IEFJDT02 sp237 key1 k1 sp236 swa dd
    

Problem summary

Problem conclusion

Temporary fix

Comments

  • closed for pin
    DB2INFO JES2INFO VSAMINFO
    
    Additional keywords: Abend822 s822 abends822
    abends878 s878 abends80A s80A rc10 ABEND106 S106 rc28
    

APAR Information

  • APAR number

    II07347

  • Reported component name

    V2 LIB INFO ITE

  • Reported component ID

    INFOV2LIB

  • Reported release

    001

  • Status

    CLOSED CAN

  • PE

    NoPE

  • HIPER

    NoHIPER

  • Special Attention

    NoSpecatt / Xsystem

  • Submitted date

    1993-10-26

  • Closed date

    1997-10-16

  • Last modified date

    2008-05-21

  • 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



Document information

More support for: z/OS family

Software version: 001

Operating system(s): MVS, OS/390, z/OS

Reference #: II07347

Modified date: 21 May 2008