Additional MSCF server tasks are generated to cope with an increase in the number of jobs sent to the processor from the partner processors, and also to avoid inter-processor deadlock situations. If the number of jobs decreases, the superfluous additional server tasks are cleared down. The number of server tasks on the master processor fluctuates in particular in a shared pubset network. An incorrect configuration of the shared pubset network (e.g. low-performance master processor and high-performance slave processor) can overload the master processor in extreme cases. A dramatic increase in the number of MSCF server tasks can result in a memory saturation or can even bring the system to a standstill. The existence of more than a certain number of server tasks on the master processor (see the “MSCF configuration parameter SERVER-TASK-LIMIT” (Global control parameters )) will mean that jobs originating from the partner processors (slave processors) will therefore be put on hold until there are sufficient free MSCF server tasks on the local processor.
The line "NUMBER OF SERVERS: CURRENT=x FREE=y" in the output of the command SHOW-MSCF-CONFIGURATION INF=*LOCAL indicates the number of actual and free MSCF server tasks. If the value for FREE is "0" and the value for ACTUAL is the same as that set for SERVER-TASK-LIMIT or even greater, this indicates a high load from the partner processors. If this status persists, suitable measures should be taken to resolve it (e.g. increasing the server task limit, master change on the processor with the highest catalog load).