OSI/MHS Configuration and Management Manual

Sizing and Tuning Your OSI/MHS Subsystem
OSI/MHS Configuration and Management Manual—424827-003
8-8
Configuring Supporting Software
You can also use X.25 hunt groups—multiple X.25 lines with the same DTE address—
to distribute incoming or outgoing traffic across multiple lines. X25AM does not
support this feature. To use it, you must obtain the lines from your network supplier
and associate each with a different X25AM process.
An MR group can service only one X.25 line. You should configure at least two MR
groups per line so that while one RTS process is servicing the line, another can be
listening for an incoming association.
See Appendix D, OSI Address Configuration in OSI/MHS, for information about
defining the OSI addresses for OSI/MHS.
You should use a large transport protocol data unit size with OSI/MHS, although some
X.25 service providers impose limits on this size. You define TPDU size as an attribute
to OSI/AS.
Transaction Management Facility (TMF)
OSI/MHS requires a significant amount of disk space for TMF audit trails. The audit
space required for a particular message being relayed is several times the size of the
message, with greater overhead required for small messages than for large ones. For
example, the audit space required for a 4K message is about 16 times the size of the
message. The space required for a 40K message is about 4 times the size of the
message. Additional audit space is required if the message has either or both of the
following characteristics:
•
Local recipients (MS users)
•
Multiple destinations (requiring relay to different MTAs)
The store cleaner places great demand on the TMF audit trails while it is running,
especially if there are many messages to be deleted. This situation is most likely to
arise after a peak traffic period. It will also arise if the store cleaner interval is large,
allowing large numbers of messages to be processed in each interval.
TMF audit trails are overwritten periodically. The frequency depends on message
traffic and the size of the audit trail files. Consider using auxiliary audit trails for
OSI/MHS. (See the TMF Planning and Configuration Guide and the TMF Operations
and Recovery Guide for information about auxiliary audit trails.)
Enscribe and SQL Databases
When you size and distribute MHS databases—for example, MR group PDU stores—
you must decide whether to place the data on the same volume (for ease of
management) or to distribute it across volumes for better performance and fault
tolerance. If you have put all the data on the same disk and find that the disk is
consistently busy, consider distributing data (such as PDU stores) or partitioning files.
Certain OSI/MHS files cannot be distributed: the OSI/MHS configuration database
must be kept in one volume. All MS SQL tables and the SQL catalog must be on the
same subvolume.