Volume mounting and dismounting schedule timers

 

This topic provides information on how a system manages work requests to the directly attached optical library devices. It is a high-level view and does not include all program logic.

You can use the Change Device Description (CHGDEVMLB) command to change the queuing and scheduling logic used by a system for directly attached optical media libraries.

Two timer values are associated with optical media libraries that affect the scheduling of volume mounting and preemptive dismounts. You can change both timer values by using the CHGDEVMLB command. The first timer value (UNLOADWAIT) is the unload wait time. This value determines how long the system waits for a new request for a mounted volume before removing it. The second timer value (MAXDEVTIME) is the maximum device wait time. This value determines how long a volume with active requests remains in a drive while other queued requests are waiting to use the drive.

By using these two timer values, you can adjust the volume mount scheduling that is used by the optical media library to match your application’s use of optical volumes.

You can change these timer values at any time; however, the new timer values will not become effective until the next time the device is varied on.

System job priority and limit timers are used to schedule volume mounting. The maximum number of volumes that can be mounted is equal to the number of drives in the optical media library. Keep the following points in mind as you schedule volume mounting:

• A volume can remain mounted in an optical drive for the maximum device wait time if work requests with the same or lower job priority for a different volume have been received. An exception to this is when you are initializing a rewritable volume or using DUPOPT; the volume remains in the drive until the operation is completed.

• Work requests for mounted volumes are serviced before requests of the same or lower job priority for volumes not mounted.

• For a multiple job environment, volumes are mounted based on the job priority for the work request. A work request from a job with a higher system job priority causes the required volume to be mounted to handle that request. The volume remains mounted for the maximum device wait time if work requests continue, dismounts after unload wait time inactivity, or is overridden by a work request from a job with higher system priority. If you are initializing a rewritable volume or using DUPOPT, the volume remains mounted until the operation is completed.

• If the work on a drive is interrupted because of a higher priority request, the maximum device wait time timer for the currently mounted volume is canceled. All future requests for that volume are queued for normal processing by priority.

• If the volume needed for a work request is not mounted within the system timeout (typically 30 minutes), the job fails due to a timeout.

 

Change job priority on active jobs

Due to the work management method used by i5/OS^® optical support, changing run priorities of an active optical job at the user level can result in loss of time allocation and, in some cases, causes jobs to time out.

 

Parent topic:

Performance considerations for directly attached optical media libraries