Using parallel servers

Parallel servers work like workstation fixed resources: the number of servers required by an operation must be available before IBM Workload Scheduler for z/OS will start the operation. Operations on non-virtual workstations use each one server; so, the number of parallel servers available on a workstation at any time is equal to the maximum number of operations that can be started simultaneously.

In case of virtual workstations, the scheduler considers the parallel servers as a criterion to select the submission destination: the next destination to be selected, as part of the round-robin algorithm, is checked for the number of parallel servers. If the check fails, it moves on the next destination in the list. Parallel servers do not apply for jobs on fault-tolerant agents.

You can choose whether IBM Workload Scheduler for z/OS takes parallel servers into consideration when it plans the schedule, when it starts an operation, at both times or at neither. This is the server usage, specified as P, C, B, or N on the workstation definition. See Specifying workstation fixed resources.

IBM Workload Scheduler for z/OS assumes that operations on computer workstations always use a single server. The maximum number of servers that can be specified on a single workstation is 65535.

Specify the number of servers for an operation on the WORK STATION RESOURCES AND SERVERS panel (Figure 1). You can display this panel by selecting option 2 on the OPERATION DETAILS panel (Figure 1).