Backup Exec - Any Disk Storage - Round Robin Device Selection
When backup jobs point to a B2D device pool as the backup destination, logic can be added to allow for a round-robin selection of devices from the device pool. Additionally, job history can be referenced to help determine a minimimum free space requirement for a job.
The issue is - The default logic of selecting the first available device from a B2D device pool often causes an overutilization of the first device in the pool and underutilization of subsequent devices. Users are having to cancel backup jobs that select a device from the pool having critical disk usage without enough free space for the job being run. Even though there are several other devices in the pool with very low usage.
For example: Current logic will chose the first available device having 13GB of free space and begin a backup job that will require 60GB or more. If the backup job is a GRT type backup, the job cannot span volumes and will hang. Non-GRT backups will also pause wanting the user to insert free media. If the B2D device pool is comprised only of internal disk (not removal disk storage) media insertion is impossible.
Solution: A round-robin approach would better spread usage across devices in the pool. Coupled with proper use of retention periods this can greatly lower the chances of a critical or full device being selected thereby allowing more time for media reclamantion to keep up with the backups.
Additionally, even though Backup Exec cannot anticipate the final size of the backup job, historical data from the same job can be referenced to determine the previous byte-count. Apply an N% (user definable) increase to the previous byte-count and select a device having at least that amount of free space.