PK21790: GET MSG IEF238D SHOULD GET 'WAIT UNTIL THE TAPE IS AVAILABLE'.

APAR status

• Closed as program error.

Error description

• Get message IEF238D DRFS0001 - REPLY DEVICE NAME OR 'CANCEL'
  when executing a parallel job with READNUM(1,13) specified.
  The message IEF238D should not be issued in this case but
  should be 'wait until the tape is available'.
  This occurs when the Image Copy is placed on multiple tape
  cartridges which have a lot of area for recovery.  In some
  cases the Recovery job hangs.
  

Local fix

• As described, these 2 problems occur when there are several data
  sets to restore with 1 tape device available.  If the number of
  parallel subordinate address spaces are reduced to 1 by
  specifying SORTPARM(NUM(1)) in the SYSIN the revovery will work.
  

Problem summary

• ****************************************************************
  * USERS AFFECTED: All DRF tape users with fewer tape devices   *
  *                 and more DRF jobs to be recovered will be    *
  *                 affected.                                    *
  ****************************************************************
  * PROBLEM DESCRIPTION: DRF waits on allocation recovery after  *
  *                      scheduling more image copy restore than *
  *                      allowed by the READNUM value.           *
  *                      Allocation recovery issues messages     *
  *                      IEF244I and IEF238D, but DRF hangs.     *
  *                      The DRF subordinate address spaces      *
  *                      must be cancelled.                      *
  ****************************************************************
  * RECOMMENDATION: INSTALL CORRECTIVE SERVICE FOR APAR/PTF      *
  ****************************************************************
  When the customer has fewer tape devices and more DRF jobs that
  need to be recovered (e.x. READNUM(1,13), SORTPARM(NUM(3))),
  the next DRF subordinate address space did not wait until
  allocation device is available. This caused IEF244I message to
  be issued to request an available allocation device.
  

Problem conclusion

• DRF attempts to schedule as many image copy restores as allowed
  by the READNUM(x,y) values where 'x' is the number of tape
  drives available and 'y' is the number of concurrent restores
  allowed.
  The idea was to allow more image copy restores to run in
  parallel when there is a mixture of tape and dasd. The problem
  occurred when there were more tape restores than tape drives
  available.
  A new keyword TAPECHK(Y|N) has been added to fix this problem.
  With TAPECHK(Y), DRF stops scheduling image copy restores when
  the number of tape devices in use reaches the tape device
  vailable and more DRF jobs READNUM value. TAPECHK(N) is the
  default. TAPECHK(Y) would be used when the number of image
  copies on tape exceeds the number of available tape devices
  to by pass the new code.
  FRXIDYN0 is modified so the DRF subordinate address space
  must wait until receiving the response of tape device status
  from master address space to allocate next DBDS, when TAPECHK
  is specified.
  FRXMSTR0 is modified to check availability of tape device, and
  send the response to the subordinate address space.
  Other modules and macros modified for the changes are:
  FRXEPSS0
  FRXEDRF0
  FRXPSDS0
  FRXPSDR0
  FRXPDSR0
  FRXPDSS0
  FRXRVGB
  FRXRVIC
  FRXRVRT
  FRXRWSP
  FRXVWSP
  FRXAWEX
  FRXAWFN
  FRXCON
  FRXEDRF
  FRXEWSP
  

Temporary fix

Comments

APAR Information

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

  PK18015

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

  UK14479

Modules/Macros

• FRXAWEX  FRXAWFN  FRXCON   FRXEDRF  FRXEDRF0
  FRXEPSS0 FRXEWSP  FRXGRPT0 FRXIDYN0 FRXMSTR0 FRXMSTR1 FRXMTC
  FRXPDSR0 FRXPDSS0 FRXPSDR0 FRXPSDS0 FRXRVGB  FRXRVRT  FRXRWSP
  FRXVWSP
  

Fix information

• Fixed component name

  IMS DB RECOVERY

• Fixed component ID

  5655I4400

Applicable component levels

• R310 PSY UK14479

     UP06/05/16 P F605