OSI/MHS Configuration and Management Manual

Sizing and Tuning Your OSI/MHS Subsystem
OSI/MHS Configuration and Management Manual—424827-003
8-18
TMFCOM
TMFCOM
Use TMFCOM to monitor TMF transactions related to OSI/MHS processes and GPI
applications. If many TMF aborts are associated with GPI applications, it is possible
that there are too many applications for the level of incoming traffic.
Handling Large Messages
Adjust an existing configuration as follows to accommodate larger messages:
•
Decrease the number of object trees created for each GPI application (because
the object trees are larger).
•
Allocate more space for TMF audit trails.
•
Allocate more swap space for gateways and other processes. (See
Interconnection Problems With Gateways on page 9-52.)
•
If messages will be consistently large, adjust the MAX-MSG-XFER-TIME attribute
accordingly. If you expect a range of message sizes, use the REM-MSG-
THRUPUT and REM-MSG-OVERHEAD attributes instead of MAX-MSG-XFER-
TIME. Expect it to take longer for OSI/MHS to relay, submit, and deliver large
messages than to perform the same operations on smaller messages.
•
Distribute processes to limit contention for virtual memory by processes running in
the same processor.
Summary
This section has provided an overview of factors that affect OSI/MHS performance. It
has listed services required by OSI/MHS, along with references to manuals describing
those services. This section has identified types of attributes that might be adjusted to
improve performance and has included a summary of background and procedures
helpful for performance tuning. Special consideration has been given to large
messages.
Section 9, Troubleshooting Your OSI/MHS Subsystem, describes troubleshooting for
the OSI/MHS subsystem.