With the system-managed tape library support, each scratch allocation (through the ACS routines) will assign one or more tape storage groups and each tape storage group can be associated with 1 to 8 tape libraries. As a result of this, there can be many tape subsystems (and devices) that are considered eligible for each scratch request. A limitation exists today on the number of tape subsystems (device pools) that SMS and MVS allocation can support with a scratch allocation. Today this limit is 253 tape subsystems or 4048 devices (253 x 16 devices). With APAR OA21462, SMS will detect and limit the number of eligible tape subsystems (device pools) to the first 253 subsystems and will ignore the tape subsystems beyond 253.

For example, if an installation is using the TS7700 Virtualization Engine and assigns SGTAPE and SGTAPE associated with 4 two-cluster grids, the maximum number of tape subsystems that could be returned to SMS would be 4 X (32 subsystems each) = 128 which is well within the existing 253 subsystem limit. However, if an installation, for example were using 6 three-cluster grids, and SGTAPE was associated with all 6 three-cluster grids, the maximum number of tape subsystems that could be returned to SMS would be 6 X (48 subsystems each) = 288 which would go over the existing subsystem limit of 253. If there are more than 253 tape subsystems considered eligible for the scratch request, SMS will only return the first 253 tape subsystems to MVS allocation. As you start using libraries with more and more tape subsystems (for example, the TS7700 Virtualization Engine), review your ACS routines and the number of tape libraries, and subsystems, being used for your scratch allocations.