NB 6.5.4
Linux
Storage Lifecycle Policies
AdvancedDisk documentation speaks to disk management in terms of
1) space utilization among volumes in a disk pool: will move to a new volume in a disk pool when insufficient space exists on a volume.
2) load balancing amongst media servers.
However, in a configuration with a single media server with many clients writing to a disk pool, all jobs will go to the same disk, yes? and when that disk reaches a certain capacity, all subsequent jobs will then move to the next disk.
It seems though that tends toward disk contention. If your disk pool has two very large volumes, almost all writing will be to the first disk in the pool.
I would like to spread I/O and disk contention out among a series of disks, accessible concurrently. Would this arrangement work?
DSU-1 configured as DiskPool1 with 1 device
DSU-2 configured as DiskPool2 with 1 device
DSU-3 configured as DiskPool3 with 1 device
DSU-4 configured as DiskPool4 with 1 device
Then group these into single Storage Group, using the LeastRecentlySelected (-sm 2) selection method.
Any other suggestions on how to better distribute backup jobs among multiple disks, but still use the AdvancedDisk space estimation logic?