PK25674: NEW FUNCTION TO SUPPORT STACKED IMAGE COPIES

APAR status

• Closed as new function.

Error description

• New Function to support stacked Image Copies
  

Local fix

Problem summary

• ****************************************************************
  * USERS AFFECTED: Users with IMS Database Recovery Facility    *
  *                 Version 3 Release 1 installed.               *
  ****************************************************************
  * PROBLEM DESCRIPTION: DRF V3R1 provides users with the        *
  *                      capability to generate output image     *
  *                      copies with HPIC.  DRF V3R1 does not    *
  *                      allow the capability to generate these  *
  *                      image copies stacked on tape.           *
  ****************************************************************
  * RECOMMENDATION: INSTALL CORRECTIVE SERVICE FOR APAR/PTF      *
  ****************************************************************
  DRF needs to be redesigned to allow the capability to stack
  output image copies on tape.
  

Problem conclusion

Temporary fix

Comments

• DRF is changed to accept the HPIC STACK(*) and STACK(*,*)
  options on the DRF UTILGBL statement.  DRF will call HPIC to
  produce image copy output stacked on the same group of tape
  volumes.
  
  ----
  The following Doc changes are required:  SC18-9407-00
  IMS Database Recovery Facility for z/OS, V3R1, User's Guide
  Chapter 2. Administering the IMS Database Recovery Facility
  
  Add a section similar to "Managing Inputs" and call it
  "Managing Outputs".
  "Stacked Image Copy Output
      Stacking can place two or more output image copy data sets
      on the same tape volume or set of tape volumes.  The user
      can indicate to DRF that stacking of output image copies is
      to be done by specifing STACK on the UTILGBL statement.
      DRF will then pass this indication on to HPIC during image
      copy generation.  Refer to "Specifying UTILGBL() global
      control statement parameters" in Chapter 3 for syntax.
  
      Stacking is used to increase the efficiency of tape media
      use and to decrease the number of tape volumes needed by
      allocation.  However, use of STACK with DRF will serialize
      the recovery process and may decrease the over all recovery
      performance.  Once STACK is specified, DRF will process each
      DBDS listed in the recovery list to the same stack.
  
      Before you run your STACK recovery job, ensure that you have
      a sufficient number of tape devices available.  For
      stacking, you will typically need to have one tape device
      available to read your input image copy (if it is on tape)
      and one for your stack output (two if you are generating
      dual copies).  You can refer to "Managing Performance,
      Availability of Tape Devices".
  
  ----
  In the section for "Managing Performance" include an update for
  the following sub-section:
  
  Availability of Tape Devices
  Before you run a recovery job, determine if a sufficient number
  of tape devices are available.  Ideally, if you are restoring
  from image copies that reside on separate tapes, you need to
  have enough tape devices available to handle each parallel
  recovery task.
  
  In a simple example, if you have five image copies and you set
  up your recovery environment to start five recovery subordinate
  address spaces to run in parallel, you need to have at least
  five tape devices available for recovery.  If you have less than
  the number of tape devices available than are needed, MVS
  prompts you with an IEF238D message asking you to reply with the
  additional device name when it becomes available or to cancel.
  
  If increasing the number of available tape devices to match your
  parallel recoveries is not possible, here are two other methods
  by which you can manage your recovery:
  
  *  You can decrease the number of recovery subordinate address
     spaces that run in parallel.  For example, if you have five
     image copies and only one available tape device, you might
     consider overriding your recovery SORT parameters with
     SORTPARM(NUM(1)).  This forces IMS Database Recovery Facility
     to serialize your recovery, but because you have only one
     available tape device it is already serialized.  This method
     eliminates the need for you to repeatedly reply with the tape
     device name on the IEF238D message.
  
  *  You can specify TAPECHK(Y).  For more information about the
     TAPECHK option specification, see Chapter 3 Defining and
     tailoring the IMS Database Recovery Facility environment.
     With TAPECHK(Y), IMS Database Recovery Facility considers the
     value specified in READNUM or ICNUM as a means to control the
     allocation of tape devices in each subordinate address space.
     See Controlling Input Parallelism in this chapter for more
     details about READNUM and ICNUM usage.
  
     -   In a simple example, if you have five image copies, five
         subordinate address spaces, and only one tape device, you
         would specify TAPECHK(Y) and ICNUM(1,13).  The "1" in
         ICNUM informs IMS Database Recovery Facility that only
         one tape device is available.
  
     -   That single tape device is allocated to only one address
         space and it puts the other recoveries into a wait state
         until the device becomes available.  However, if you
         erroneously specified ICNUM(2,13) when only one tape
         device is available, IMS Database Recovery Facility
         attempts to allocate two tape devices but MVS issues
         message IEF238D on the second allocation attempt.
  
  Also, be aware that when you restore from image copies that
  reside on tape and also expect IMS Database Recovery Facility to
  generate HPIC image copy output on tape, you require twice as
  many tape devices.  In this scenario, IMS Database Recovery
  Facility requires one tape device to read the input image copy
  from, and one tape device to generate the output image copy to.
  
  ----
  Chapter 3. Defining and tailoring the IMS Database Recovery
    Facility environment
  (Specifying UTILGBL() global control statement parameters)
  
  In the first paragraph, change the second sentence to say:
   "Most of these options are valid on the IB(), IC(), and ..."
  
  Change the UTILGBL syntax diagram to include STACK as an option.
  Put it between SPACE and STORCLAS in the syntax diagram:
   " ...
        |_SPACE(_ _ _CYL,_ ____________ _),__|
        |        | |_TRK,_|            |     |
        |        | |_blk,_|            |     |
        |        |_primary,____________|     |
        |        |_secondary,__________|     |
        |        |_RLSE(_ _CONTIG_ _),_|     |
        |        |       |_MXIG___|    |     |
        |        |       |_ALX____|    |     |
        |        |_ROUND_______________|     |
        |_STACK__(_ __*__ _)__,______________|
        |          |_*,*_|                   |
        |_STORCLAS__(__stor-class-name__)__,_|
     ... "
  
  After the syntax diagram add:
  " STACK(*|*,*)
        STACK places two or more output image copy data sets on
        the same tape volume or set of tape volumes.
        If you specify STACK(*), HPIC dynamically allocates the
        output image copy data set and requests a single copy
        of the stacked image copy output.  STACK(*,*) is used
        when dual copies are needed.  For more information on
        stacking with HPIC please reference the IBM IMS High
        Performance Image Copy for z/OS, User's Guide.
  
        STACK is only supported on the UTILGBL statement.  It is
        not a recognized option on the ADD command with the IC()
        paramenter.  Invoking STACK will cause DRF to process each
        DBDS in the recovery list serially in one recovery
        subordinate address space.
  
        The VOLSER option also needs to be specified on the
        UTILGBL statement to indicate what tape volumes will be
        used.  ICOUT cannot be specified with STACK.    "
  
  ---
  Chapter 5. Starting the IMS Database Recovery Facility
  In the section describing the "IC" keyword, remove "STACK" out
  of the list "not supported by IMS Database Recovery Facility".
  Indicate at the end that "STACK is supported only on the
  UTILGBL statement".
  
  ----
  Chapter 7. Messages and Codes of IMS Database Recovery Facility
  Update the following message:
  "FRD6124A INVALID DATA ENCOUNTERED. KEYWORD: keyword RSN=rsn
   ...
   VOLSER IS REQUIRED ON UTILGBL
      A VOLSER list is required to be specified on the UTILGBL
      statement if STACK is also specified.
   VOLSER2 WITH A UNIQUE VOLUME IS REQUIRED ON UTILGBL
      A VOLSER2 list is required to be specified on the UTILGBL
      statement if STACK is also specified for dual copies of
      stacked image copy outputs.  The volser list from VOLSER2
      is used for the secondary copy. "
  

APAR Information

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

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

  UK15873

Modules/Macros

• FRXBDMG0 FRXEDRF  FRXEDRF0 FRXEPSS0 FRXEWSP
  FRXGRDS0 FRXGRPT0 FRXGUOW0 FRXIRTH0 FRXLACR0 FRXLBLD0 FRXLEPL
  FRXLGET0 FRXLSRT1 FRXLSTR0 FRXLSTS0 FRXPDSS0 FRXPSDR0 FRXRWSP
  FRXSDR00 FRXVADD0 FRXVASD0 FRXVSTA0 FRXVVER0 FRXVWSP
  

Fix information

• Fixed component name

  IMS DB RECOVERY

• Fixed component ID

  5655I4400

Applicable component levels

• R310 PSY UK15873

     UP06/07/05 P F607