Stacking image copy data sets

Use the STACK_NAME keyword to specify the name of a stack group to which the image copy data set belongs.

The same name is specified for each image copy that is to be included in the group. All image copy data sets that are part of the same stack group are written onto the same tape volume (or volumes) as file number 1, 2, 3, and so on.

Figure below specifies that an image copy data set be created for each area in the database. The output data sets are all written to the same tape volume.

REORGANIZE DBD=
dbdname,IAREA=ALL
  IC STACK_NAME=NAME1,UNIT=TAPE,DISP=(NEW,CATLG),
     DSNAME='
dataset-name-mask'

Dynamic allocation is always used for stacked data sets. When an IC subcommand containing a unique value for the STACK_NAME keyword is encountered, values that are specified for the UNIT, VOLCNT, VOLSER, EXPDT and RETPD keywords (if any) on the IC subcommand are saved; these values will apply to all image copy data sets that are written to the stack. The UNIT keyword must be specified and must refer to a tape device type. All other dynamic allocation keywords, such as DSNAME, pertain to the individual output image copy data set that is being defined. DISP=NEW is required for stacked data sets; there are no restrictions on the remaining parameters for the DISP keyword.

As the image copy function is executing, only one output image copy data set within a stack group can be open at a time. You must not specify the same stack group on more than one IC subcommand within any primary function command, such as IMAGECOPY.

In figure below two image copy data sets are requested for each area in the database. You cannot request that both copies be written to the same stack. You can, however, specify a different STACK_NAME keyword on each IC subcommand. You can also mix stacked and non-stacked image copy processing within the same primary command.

IMAGECOPY DBD=dbdname,IAREA=ALL, ...
  IC STACK_NAME=NAME1, ...
  IC STACK_NAME=NAME1, ...

Another consequence of this restriction applies when the product is performing parallel processing of commands in the offline mode. Because only one output image copy data set within a stack group can be open at a time, the serialization of each stack group affects the amount of parallel processing that is performed.

In figure below, three separate parallel tasks are created (one for each REORGANIZE command). All of these tasks, however, are also creating image copy output data sets within the same stack group. Consequently, only one reorganization task can be performed at a time; the other two tasks will be waiting for their turn to access the output image copy stack.

GLOBAL DBD=dbdname,IAREA=ALL, ...
  IC STACK_NAME=NAME1, ...
REORGANIZE DBD=dbdname1, IAREA=ALL, ...
REORGANIZE DBD=dbdname2, IAREA=ALL, ...
REORGANIZE DBD=dbdname3, IAREA=ALL, ...
Related topic

Was this page helpful? Yes No Submitting... Thank you

Comments