Retaining and moving dump tapes

When defining vital record specifications for retaining dump tapes, consider these conditions:

The dump class definitions because the dump classes and the vital record specifications you define must work together.
The dump class and DASD volume serial numbers when you define the data set names you use in the data set name masks.
The data set name format used by DFSMShsm for dump tapes includes the dump class and the DASD volume serial number. These are the two most likely variables in the data set name on which you will base your retention and movement policies.
The dump classes that are managed the same.
The dump classes that require separate management.
The dump classes that need to be stored off site.
The need to define additional vital record specifications to support the old data set naming conventions for a period of time.

Figure 1 is the minimum you are required to specify in a vital record specification for DFSMSrmm to keep dump tapes and to prevent DFSMSrmm from releasing DFSMShsm volumes.

Figure 1. Keeping tapes used for dump

RMM ADDVRS DSNAME('bprefix.DMP.**') -
    COUNT(1) CYCLES

bprefix: Specifies the DFSMShsm-defined backup and dump data set prefix.
**: Represents zero or more qualifiers of a data set name.
COUNT(1): Specifies to retain a single cycle. In the example, the COUNT(1) is used to retain a single dump cycle. Since DFSMShsm generates a unique name for each dump cycle, COUNT(1) retains all dump cycles.
CYCLES: Specifies that DFSMSrmm retain data sets based on cycles or copies of a data set.

If you want to be more specific and manage cycles of dumps, using differing policies you can extend the data set name mask as shown in Figure 2.

Figure 2. Managing cycles of dumps

RMM ADDVRS DSNAME('bprefix.DMP.class.V%%%%%%.**') -
    COUNT(1) CYCLES

bprefix: Specifies the DFSMShsm-defined backup and dump data set prefix.
class: Specifies the DUMPCLASS you have defined to DFSMShsm.
V%%%%%%: The six character volume serial number.
**: Represents zero or more qualifiers of a data set name.
COUNT(1): Specifies to retain a single data set that matches the filter mask.
CYCLES: Specifies that DFSMSrmm retain data sets based on cycles or copies of a data set.

You can also define vital record specifications to manage different types of dump classes having different cycles. Figure 3 shows the use of a pseudo-GDG data set name to identify dump cycles for movement to storage locations for disaster recovery. A DFSMSrmm pseudo-GDG is a collection of data sets, using the same data set name, that DFSMSrmm manages like a GDG. A pseudo-GDG data set name contains the ¬ as a placeholder for the characters in the pattern that change with each generation.

In the example, ¬ is used to mask the DFSMShsm generated date and time in the data set names so that all generations of a dump can be logically managed together.

Figure 3. Retaining and moving volumes by cycles

RMM ADDVRS DSNAME('bprefix.DMP.class.V%%%%%%.T¬¬¬¬¬¬.D¬¬¬¬¬') -
    COUNT(100) CYCLES LOCATION(REMOTE) STORENUMBER(2) -
    DELAY(1)

bprefix: Specifies the DFSMShsm-defined backup and dump data set prefix.
V%%%%%%: The six character volume serial number.
¬: Is a place holder for a single character in a data set name mask for a pseudo-GDG data set name. When ¬ is used in a data set name mask, DFSMSrmm manages the data sets matching the data set name mask like a generation data group.
COUNT(100): Specifies to retain all dump cycles managed by DFSMShsm. 100 is the limit for the number of dump cycles.
CYCLES: Specifies that DFSMSrmm retain data sets based on cycles or copies of a data set.
LOCATION(REMOTE): Specifies the storage location called REMOTE to which the number of volumes for the data sets specified in STORENUMBER should be moved.
STORENUMBER(2): Two cycles are kept in the storage location REMOTE. All other cycles up to the COUNT(100) value are retained in the HOME location.
DELAY(1): The current cycle is kept in the library for one day before being removed to the storage location.

In Figure 3 all 100 cycles of a dump taken for a specific volume and class combination are kept if produced by DFSMShsm. For each matching vital record group:

The current cycle is kept in the library for one day before being removed to the storage location.
Two cycles are kept in the storage location.
All other cycles are returned to and kept in the library until DFSMShsm releases them.
You can use any combination of LOCATION, STORENUMBER, NEXTVRS, and DELAY to provide the required level of service.
COUNT(100) is coded so DFSMSrmm keeps all cycles that DFSMShsm produces. You can tailor the example by using different values for the class name.

You might need to define additional vital record specifications to support the old data set naming conventions for a period of time.