OSI/MHS Configuration and Management Manual

Troubleshooting Your OSI/MHS Subsystem
OSI/MHS Configuration and Management Manual—424827-003
9-31
Recovery Checklist
errors. If the PDU store becomes overburdened with unroutable PDUs, contact your
HP representative about diagnostic software that can delete the unwanted PDUs.
Recovery Checklist
Consider the following items when an unroutable event occurs:
•
Nondelivery reason code and diagnostic
•
Originator names from the message or report
•
Recipient names from the message or report
•
APPL object configuration and status
•
Route selection criteria defined by the ROUTE objects
Message Tracking
For purposes of troubleshooting and performance tuning, it is valuable to be able to
track the progress of a message (or report or probe) through the MTA. OSI/MHS
provides two ways of doing this:
•
Event messages record the major processing steps in the transit of a message:
association establishment and receipt of incoming messages; routing, including
distribution-list expansion and CUG checking; relay and delivery of messages to
recipients or gateways; storage of messages in a message store, and submission
and retrieval of messages through the message store. Event messages also
report other information useful for billing and general resource management; for
instance, they record when an APPL object is added or deleted and when various
subsystem processes start and stop running.
•
Trace records provide detailed information about protocol exchanges between
OSI/MHS and external entities, and make visible the results of message encoding
and decoding by the MR process.
Both mechanisms allow you to select the information that interests you. In the case of
a trace, you use SCF or a management application to start and stop tracing and to
specify the types of information to collect. Later, when you use PTrace to display the
trace data, you specify the types of displays you wish to see.
In the case of event messages, you use SCF or a management application to specify
the types of accounting and message-tracking events to be reported; you can change
the selection dynamically to suit different circumstances or tasks. (Like trace
collection, event-reporting has an impact on performance, so it makes sense to limit
reporting to the most appropriate events.) You can use filters to limit further—on the
basis of event-message content—the set of messages to process, display, or print.
You can also modify the event-text templates (provided in the source file SMHSTMPL
on the OSI/MHS installation subvolume) to determine which tokens in any event will be
displayed.