• The initial disposition of the data set (as specified in the DISP operand of the DD statement) is used to position the checkpoint data set each time it is opened. For a discussion of a specific circumstance in which the only checkpoint entry that exists on a data set will be the last checkpoint taken (whether that entry is valid or not), see How Checkpoint Entries Are Written.
• The final and conditional dispositions of the data set have their normal meanings. However, if termination is occurring and an automatic restart at a checkpoint is to occur, the system automatically keeps all data sets that are in use by the job, including the checkpoint data set.
• If you code UNIT or VOL=SER on the checkpoint DD statement, it has no effect unless changed by your installation's ACS routines. If the checkpoint data sets are to be restricted to specific volumes, a storage group should be defined for this purpose.
• Filling the current DASD extent or tape volume is called an end-of-volume (EOV) condition. If EOV is encountered while writing a checkpoint on a direct access or tape volume, a second attempt is made to create the checkpoint entry. If EOV occurs before the second attempt at writing the entry is completed, message IHJ0001 (invalid checkpoint) is issued. Control is returned to you with a X'08' return code in register 15 and a reason code of 027 in register 0.

The following example shows DD statements for the checkpoint data set:

   //ddname   DD   DSN=dsname,UNIT=TAPE,DISP=(MOD,KEEP)
   //ddname   DD   DSN=dsname,UNIT=SYSDA,
   //              DISP=(NEW,DELETE,KEEP),SPACE=(CYL,(15,17)),
   //              VOL=SER=CKPTDS