This query provides an excellent overview of the hourly usage patterns for your TS77xx grid clusters. With virtual tape drives, fast ready mount (scratch tape) requests are satisfied almost instantly. When a workload requests a specific virtual tape volume, the objective is to have the volume resident in the cache where the time to mount is fast. Recall mounts, on the other hand, result in significant delays for the requesting workload. The virtual tape must be accessed in the physical tape repository and copied into cache before the mount request can be satisfied. If significant numbers of recall mounts occur, particularly in recognizable patterns, review your archiving policies.
Another key metric to watch is the maximum number of concurrently mounted virtual tape drives. If this metric reaches the number of defined virtual tape drives, generally 128 or 256, workloads requesting mounts are delayed. If this delay occurs during particular hours, shifting some workloads to a lower use time period may solve the problem.
|
Copyright © 2014 CA.
All rights reserved.
|
|