What is the criteria used by CA OPS/MVS to start a new OSF server?
The algorithm that is used to add servers depends on the OSFQADDPROCESS parameter.
If OSFQADDPROCESS=AVAIL a new server is added, when the number of commands in the OSF TSO execution queue is higher than OSFQADD and OSFMIN but lower than OSFMAX.
When OSFQADDPROCESS=QADD a new server is added when the queue depth equals OSFQADD+1 + (n * OSFQADD) until OSFMAX is reached. Where ‘n’ equals the number of dynamically added servers.
OSFQADDPROCESS=QADD is the default and recommended setting to optimize performance. The calculation results in a 'slower' rate of adding servers, so that the overhead of a task starting and stopping in the system is decreased. It also takes into account the fact that with a larger number of servers active, the chances are greater that work will end, in one of the servers, and it will be ready for/get work from the OSF queue, before another server can actually start and be fully initialized.
You can find additional information at our website: CA Technologies Documentation