II02787: JES2INFO: UNABLE TO START JES2. PROCLIB DD STATEMENT INVALID MVS/CONVERTER INTERPRETER 5752SC1B9

APAR status

• Closed as canceled.

Error description

• $$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$
  $                                                              $
  $          J E S 2   I N F O R M A T I O N   A P A R           $
  $                                                              $
  $$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$
  
     This APAR describes error symptoms when the JES2 cataloged
  procedure contains DD statements that cannot be processed by
  the Converter/Interpreter because the dataset is not in the
  Master Catalog and there is no volser on the DD statement,
  or because the unit containing that volser is not online.
  (See JES2 Initiatlization and Tuning Guide under the topic
  "Updating the JES2 Cataloged Procedure".)
  
  Symptom One:
  MsgIEE132I or msgIEC143I is issued during IPL start of JES2.
  
  Symptom Two:
  If MVS has been IPL'ed with a JES2 that has the DD statements
  properly defined, and that JES2 is stopped and the JES2 with
  the improper DD statements is started, an abend0B0 will occur
  and msgIEF197I will be issued.  MsgIEF361I can also be issued.
  
  Symptom Three:
  If the PROCLIB DD statement is coded with UNIT=
  and VOL=SER=  parameters, and the proclib dataset is
  *deleted* from the volume, then the next time JES2 is started
  (whether IPL or not) msgIEC143I rc04 describing abend213 will
  be issued from the Converter subtask.  The customer reporting
  this problem received msgIEC143I rc000, which is not a valid
  Return Code.  That was probably due to a vendor product with
  hooks in Open/Close, but does not affect the problem.  In this
  instance, there was no failure in Allocation because the UNIT
  and volser were specified and the volume was online.  It is not
  until the dataset is opened that it is discovered that the
  dataset does not exist.
  .
  Additional keywords for searchability only:
  JES2INFO 5752SC1BH
  .
  

Local fix

•      See also OY67842.
  

Problem summary

Problem conclusion

Temporary fix

Comments

•    There are several different user errors which can make it
  impossible to start JES2, resulting in one or more of the error
  messages given above.  Typical errors are:
  
    A) Deleting the JES2 proc from Proclib and then IPLing.
       JES2 proc then cannot be found.
  
    B) Deleting a user Proclib named in a DD statement in the
       JES2 Proc.  (This may occur through moving a volume
       offline or archiving the dataset).
  
    C) Improperly cataloging or failing to catalog a user proclib,
       where that dataset is not identified by a volser on the
       corresponding DD statement.
  
     The only way to recover from this situation is to correct the
  JES2 procedure.  The purpose of the following suggestions is to
  document some approaches which have been used successfully in
  the past.
  
  1) Is there another CPU with a functioning JES2 system which
     can be used to change the Proclib on the failing system?
  
  2) Is there an MVS Starter System which can be modified to
     access the Proclib on the failing system?
     NOTE:  Any standalone Starter System or Dump/Restore tape
            requires that the IPL volume or tape be on the same
            processor as the IPL console in a multiprocessing
            environment.
  
  3)  Is there any other functioning system at another location
  which can be used to change the failing proclib (perhaps
  requiring that the proclib be dumped to tape and transported to
  the other location)?
  
  4)  If none of these options is available, then perhaps a
  Secondary JES2 can be started on the failing system.  This
  still presents some problems as TSO users may not log on under
  a secondary subsystem, and initiators cannot be started, even
  under a secondary JES2, unless the primary has been started and
  has posted MVS to release the ENQueue on the started task
  Internal Reader. Therefore, even if a secondary JES is available
  MVS control blocks will have to be modified to identify the
  Secondary JES name as the Primary.  (This requires the ability
  to display and alter storage.)  The following procedure has
  been made to work under MVS/SP R1.3.3 but has not been verified
  for other releases (the release of JES2 does not matter for this
  procedure):  The CVT+X'128' is CVTJESCT, which points to the
  'JESCT'.  JESCT+x'1C'=JESPJESN which is the name of the Primary
  Subsystem.  (These fields are generated by MVS at IPL time from
  the IEEVIPL Parmlib member.  The subsystem name in the JESCT
  should be ZAPped to equal the name of the secondary subsystem
  which will be started as the temporary primary.  That JES may
  now be started and will have the ability to start initiators and
  TSO logons.  The next IPL will reset the JESCT to its original
  value.
  
  5)  Item 5 deleted.  Information is no longer valid
  
  6) If changes must be made to the JES2 proc itself and none of
  the above recommendations are capable, then the following trap
  can be done to catch the JCL for the JES2 proc before it is
  executed:
  
   A) IPL the system specifying a non-existant SMFPRMxx
      SYS1.PARMLIB member. i.e. Reply 0,CLPA,SMF=ZZ
  
   B) When SMF is initializing, you will be prompted for
      additional SMF OPTIONS.  Reply with
      SUBSYS(STC,EXITS(IEFUJV)). This will ensure that IEFUJV is
      invoked for started tasks.
  
   C) Once system is up (as far as it will get), set this SLIP:
      SLIP SET,IF,L=(IEFUJV,0),A=WAIT,ML=50,PL=99,
         DATA=(1R?+8?+0,LT,10),END
      This SLIP will cause the system to go into a WAIT each
      time IEFUJV is called to validate a JCL statement.
  
   D) Start JES2 from the console. When the IEFUJV gets control
      to process a JCL statement, the system will be put into a
      wait and the console will display the registers at entry to
      IEFUJV. Register 1 points to a parameter list. The second
      word of that parameter list is a pointer to the JCL
      statement image. You can use the hardware (system) console
      to display the storage where register 1 points and to alter
      the contents of the JCL image. You can modify the JCL card
      in any fashion except you cannot change the type of card.
      For instance, you cannot change a DD card to an EXEC card.
      In the case of a bad DD card, you can DUMMY it out or modify
      it so that it will not cause allocation errors.
  
   E) Once the changes have been made to the card, restart the
      system.  The system will process the statement you just
      examined/modified and then go into another wait for the
      next statement.  Repeat steps D and E until all the JCL has
      been examined and fixed. Note: the changes you make are not
      permanent.
  
   F) Once all the cards in the JES2 proc have been processed,
      JES2 will come up.  Disable the SLIP set in step C.
  
   NOTE: The SLIP can be modified to include a 'DA=(1R?+4?+2,EQ,
  aaaaaaaa)' Where 'aaaaaaaa' is the hex representation for the
  first 4 characters following the '//' on the bad JCL card (if
  you know what card is 'bad').
  

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