BEA WebLogic Server Tuning Guide

28
Instead, it only copies existing jars to the .wlnotdelete directory (and this can be limited by the procedure described
above). Other files that are created are ejb cache files. It was seen that deployment from an extracted directory was
almost 50 % faster.
6.2 JMS Considerations
JMS lets you store persistent messages in either a disk-based file (called a JMS file store) or in a JDBC-accessible
database (called a JMS JDBC store). During performance testing, the following was noted:
• Using a JMS File Store resulted in less path length but allowed for less concurrency.
• Using a JMS JDBC Store introduced more overhead but allowed for more concurrency. Thus, using the JMS
JDBC Store resulted in more throughput.
• Using a JMS JDBC Store caused a considerable increase in TMF transactions.
Unfortunately, it is almost impossible to address the performance of JMS on WebLogic Server in a generic fashion.
Message size, acknowledge mode, persistence mode, and the type of consumer are just a few of the things that can
impact performance. With so many variables, it is not possible to extrapolate the performance of another JMS-
based application, based on our performance testing. The only way to understand JMS performance is by testing
your own application (or a proof of concept). More information on this is available in the WLI Tuning Guide.
6.3 Configuring TMF and DP2 for Use with WLS
The Transaction Management Facility (TMF) provides transaction protection and database consistency in demanding
online transaction processing and decision-support environments. To furnish this service, TMF manages database
transactions, keeps track of database activity through audit trails, and provides recovery methods. Configuring TMF
for optimum performance can be a complex task. The optimum configuration depends on the type of transactions in
your WLS application and the rate of audit information that is generated. The TMF Planning and Configuration
Guide identifies events that suggest the need to reconfigure your audit trails or change transaction limits. You can
monitor system activity using the TM View graphical user interface or TMFCOM STATUS and INFO commands to
determine if you need to make changes.
As an integral part of the operating system, much of the TMF functionality is provided by system components such
as the file system and the DP2 disk process. Performance is improved when the primary DP2 disk processes (those
for different disks) are spread across processors.
6.4 Tuning the SQL/MX Database and the OSS File System
System managers should use their preferred performance analysis tool to monitor the SQL/MX database.
System managers also need to carefully watch any files (such as TLOG files) in the OSS File System as those files
have a direct performance impact.
6.5 Tuning TLOG Files
TLOG files need to be carefully monitored for disk busy access. TLOG files should be placed on separate physical
disk drives in order to spread the disk busy load.
6.6 Tuning the Data Cache
Tuning the data cache is very application-specific. The goal is to use system memory as wisely as possible, and to
perform as much of the disk I/O in cache (instead of going to the physical disk). Follow the standard cache tuning
guidelines found in the NonStop Server manuals.