You can define scratch pools based on the system where the volumes are used. For example, define a pool prefix A* to cover all volumes in a pool reserved for system MVSA. Define a pool prefix VM* for a pool reserved for volumes used on a VM system.

When you define scratch pools by system, you can only use scratch volumes from those pools on the systems specified for them. If you try to mount a scratch volume on a different system than the one with which it is associated, DFSMSrmm rejects the volume. If a system has no pool defined for it, the system accepts all volumes except those you have defined to a specific system. Make your scratch volumes available for use on all systems if you do not want these limitations.

Recommendation: Use one scratch pool for the entire library. You can use a VLPOOL command similar to the default, shown in Figure 1. When defining scratch pools, do not use no label tapes in the pool.

You can define multiple pools for each system by associating the pools with a system using the DFSMSrmm EDGRMMxx parmlib VLPOOL command with the SYSID operand as described in Defining pools: VLPOOL. DFSMSrmm accepts a volume from any of the pools defined to that system. DFSMSrmm updates mount messages and drive displays to indicate the pool from which the scratch volume should be pulled. The pools are searched in alphabetical order and the first suitable pool encountered is added to mount and fetch messages. You can assign the same name to multiple pools to make it easier to satisfy mount requests.

When you define multiple scratch pools, DFSMSrmm does not use the media name when selecting the pool to use. If you are not using system based scratch pools, the operator can mount any volume. Operators use local conventions and operator procedures to select a scratch volume. For example, the operator should select a reel tape for a mount on a 3420 drive. For a 3490 drive, the operator should select a cartridge tape. In most cases, scratch tapes are already pulled and are available for immediate mounting.

Within the pool for use on a VM system, you can define each volume for use only on VM, or for use on z/OS and VM. DFSMSrmm uses this information to reject non-z/OS volumes on z/OS. You can use VM volumes on VM systems.

You can design pools at the user, group, or application level by using DFSMSrmm calls to SMS ACS to assign storage group names, or you can use the EDG_EXIT100 installation exit to control scratch pool selection. See Using SMS tape storage groups for DFSMSrmm scratch pooling and Using the DFSMSrmm EDG_EXIT100 installation exit for more information about using pool names and implementing exit-selected scratch pools.