Improve BRMS Drive Allocation Logic for Multi‑Drive Tape Libraries

When two BRMS control groups start at (or near) the same time and share the same tape library, both backup jobs may attempt to use the same drive. Even if multiple drives are available in the library, BRMS can assign only one drive to both jobs, causing one job to lose the resource or fail with CPF415E. IBM has explained this is working as designed, and that jobs can dynamically lose or gain a tape drive depending on the order in which resources are released.

This behavior causes unnecessary job failures and prevents BRMS from taking full advantage of multi‑drive libraries. It also prevents BRMS Network mode from selecting remote volumes, because the error occurs before BRMS reaches the media‑selection step.

IBM’s current recommendation is to either avoid overlapping backups or manually allocate and deallocate drives using VRYCFG commands inside control group exit points. However, this solution is manual, error‑prone, and not documented as an official best practice.

Requested Enhancement:
BRMS should automatically select any available tape drive when multiple drives exist in the library, without requiring manual VRYCFG allocation or strict job sequencing. Specifically:

1. When two or more BRMS jobs start close together, BRMS should intelligently distribute jobs to free drives instead of requesting the same drive simultaneously.

2. BRMS should avoid situations where one active backup loses its drive to another job, except when explicitly configured by policy.

3. BRMS should provide consistent, predictable drive assignment behavior when Device Allocation is set to *SHRD and the library has multiple drives.

4. Official documentation should describe how BRMS decides which drive to use and how to configure BRMS so multiple control groups can run concurrently without conflict.

Business Impact:

• Backup windows are longer because jobs must be serialized to avoid contention.

• Multi‑drive tape libraries cannot be fully utilized without custom scripting.

• Jobs may fail unpredictably, requiring manual intervention.

• BRMS Network Mode does not get a chance to select remote volumes when resource allocation fails early.

This enhancement would allow BRMS to behave more intelligently in multi‑drive environments and reduce unnecessary operational complexity.