For each pool, you need to decide whether this pool is to be used only for specifically identified purposes or whether it can be also selected based on system scratch pools. The easiest way to prevent a pool from being used as a system–based scratch pool is to use only rack pools for your exit driven scratch pooling. If you do not do this, or you do not use one of the other methods described in Planning to manage scratch pools with EDG_EXIT100, you must use the EDG_EXIT100 installation exit to always set a scratch pool to prevent an incorrect scratch pool from being accepted by DFSMSrmm.

DFSMSrmm identifies pools by using a pool prefix, which is a one-to-five character name followed by an asterisk. See Defining pools: VLPOOL for more information.

For example, if you want to define a pool with a prefix of ABC* for use by a specific application, define the pool to DFSMSrmm in parmlib using one of the statements shown in Figure 1.

VLPOOL PREFIX(ABC*) TYPE(R) RACF(Y) EXPDTCHECK(N) -
                 DESCRIPTION('Application XyZ')
VLPOOL PREFIX(ABC*) TYPE(S) RACF(Y) EXPDTCHECK(N) -
     SYSID(NONE) DESCRIPTION('Application XyZ')

The parameter SYSID(NONE) coded in the second statement in the example, prevents the pool from being used by the DFSMSrmm system scratch pooling as long as no DFSMSrmm system is defined with SYSID(NONE).

For each pool, decide if you will run with scratch volumes from this pool loaded in a cartridge loader. For those pools that are not loaded, consider using the EDG_EXIT100 installation exit to disable the cartridge loader whenever those pools are requested.