0

Reading through the BEA WebLogic 9.2 Deployment Descriptor Elements documentation, there are a few parameters that we already use in our web applications weblogic.xml to decrease the time it takes to run continuous integration tests against WebLogic 9.2:

  • jsp-descriptor, precompile, false
  • jsp-descriptor, page-check-seconds, -1
  • container-descriptor, servlet-reload-check-secs, -1

This is what our weblogic.xml looks like:

<?xml version="1.0" encoding="UTF-8"?>
<weblogic-web-app xmlns="http://www.bea.com/ns/weblogic/90" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
                  xsi:schemaLocation="http://www.bea.com/ns/weblogic/90 http://www.bea.com/ns/weblogic/90/weblogic-web-app.xsd">
    <jsp-descriptor>
        <page-check-seconds>-1</page-check-seconds>
        <precompile>false</precompile>
    </jsp-descriptor>

    <container-descriptor>
        <servlet-reload-check-secs>-1</servlet-reload-check-secs>
    </container-descriptor>
</weblogic-web-app>

Are there are any other usual suspects that we could consider to squeeze out some more life from our build servers?

dezwart
  • 154
  • 2
  • 11
  • Checking if http://e-docs.bea.com/wls/docs90/perform/ has any useful information. – dezwart Jun 23 '09 at 07:32
  • Worth noting that the default heap size and nursery space of JRockit is rather limited for enterprise web applications. Bumping this up from the standard 128MB to 2GB increased throughput by 15% and reduced the noise from verbosegc. – dezwart Jun 23 '09 at 23:32

1 Answers1

0

I can suggest to use performance analyzer for JMX Weblogic statistics. We use HLC by DBA InfoPower (http://www.dbainfopower.com) to merge JMX statistics with database and OS performance data to get overall picture and deep dive analysis into root causes of performance degradation.