The PMDF Process Symbiont will send OPCOM messages to terminals enabled for CENTRAL and NETWORK class messages if it detects error or warning conditions. The possible error messages are described below in Section 9.5.

Fatal conditions will result in the Process Symbiont terminating and all execution queues under its control immediately changing to the STOPPED state. In such a situation the OPCOM messages preceding the termination of the symbiont will generally indicate the problem.

Any attempt to start the process symbiont prior to running the PMDF startup command procedure will also cause the symbiont to abort with a fatal error. PMDF_STARTUP must run prior to starting any queues under the control of the process symbiont. See the section on "Post-installation Tasks" in the OpenVMS edition of the PMDF Installation Guide for additional restrictions on startup execution order.

Should a detached server process encounter an error condition, it can fail to successfully complete processing of a job. Examination of the server log,

PMDF_LOG:task_server_queue-name.log

or the task specific log, generally one of

PMDF_LOG:post.log PMDF_LOG:return.log PMDF_LOG:channel-name_master.log

will usually point out the error.

Possible failures to look for are process context or parameters, such as the process name, being altered by a login.com, causing the task server to encounter an unexpected condition.

It is also possible for detached server processes to encounter an error condition and terminate abnormally without ever creating a log file. Such terminations will be noted by the Process Symbiont and the process error status sent back to the Job Controller. Setting the execution queue for retention on error with

$ SET QUEUE/RETAIN=ERROR queue-name

will allow the termination status of the job to be seen by examining the queue after a failure. Examining the OpenVMS accounting records for detached process terminations can also be useful.