PMDF System Manager's Guide

On OpenVMS, the functions of PMDF's processing jobs can optionally be performed by detached processes (managed by the PMDF Process Symbiont), rather than running in plain OpenVMS batch queues. This improves performance by reducing the amount of process creation overhead associated with plain batch queues. See Chapter 9 for details on how to set up a detached processing environment for PMDF.

Regardless of whether plain OpenVMS batch queues or PMDF Process Symbiont queues are used for PMDF processing, the design of running PMDF jobs in queues allows control of PMDF activity by stopping and starting the queue. The use of queue oriented processing jobs also makes it possible to use PMDF in complex cluster environments, where PMDF can need access to communications hardware or software that resides on a system other than the one where the message processing activity originated. Generic queues can be used to spread the load caused by mail delivery activities across multiple machines in a cluster.

All PMDF processing jobs run by default in the MAIL$BATCH batch queue. MAIL$BATCH can either be a batch queue in its own right or a logical name that translates to the name of a batch queue. Alternate queues can be selected for jobs on a channel by channel basis by using the queue channel keyword. If the Process Symbiont is used, then the MAIL$BATCH queue will be a server queue, feeding multiple Process Symbiont execution queues.

In an OpenVMS cluster environment, note that the MAIL$BATCH queue can be a generic queue feeding specific queues on one or more systems thereby spreading the processing load around. If you do spread the load around, be sure that each system has the necessary resources to handle the types of traffic submitted to the MAIL$BATCH queue. Such restrictions are generally dealt with by creating separate queues for channels requiring software or other resources only available on a few systems and then selecting those queues with the queue channel keyword.