About batch creation logic in Storage Lifecycle Manager

The Storage Lifecycle Manager service (nbstserv) is in charge of creating duplication jobs for storage lifecycle policies. Part of duplication job creation includes grouping the backup (or source) jobs into batches.

One objective of the batching logic is to prevent media contention for tape operations, including virtual tape libraries (VTL).

Batching logic applies to both disk and tape. (Though the method to prevent media contention for disk is to use disk pools and then to limit I/O streams to disk pools.)

The batching logic requires that for each evaluation cycle, nbstserv consider all completed source jobs when determining which duplication job to run next. By default, nbstserv performs the evaluation once every 5 minutes.

nbstserv avoids overloading the Resource Broker (nbrb) queue with jobs. Too many jobs in the queue make the role of the Resource Broker harder and slows down system performance.

By default, nbstserv now creates groups based on the Group images across SLPs parameter in the SLP Parameters host properties. By default, multiple storage lifecycle policies with the same priority can be batched together.

This batching logic change affects how duplication jobs appear in the Activity Monitor. Storage lifecycle policies that have been combined into one job appear under a single policy name: SLP_MultipleLifecycles. If a storage lifecycle policy has not been combined with another, the name appears in the Activity Monitor under the name of the SLP: SLP_name.

Users may see some duplication jobs that, although in the running state, do not duplicate data because they have no resources to read or write. These jobs continue to run until they receive resources to complete the job.

To turn off grouping by duplication job priority, set the Group images across SLPs parameter to No in the SLP Parameters host properties.