Voice XML Server Out of Memory Error Message
Voice XML Server Out of Memory Error Message
Voice XML Server Out of Memory Error Message
Contents
Introduction
Problem
Solution
Increase Virtual Memory
Contact Developer Services
Frequently Asked Questions
Introduction
This document describes the problem where the Voice XML (VXML) server runs out of memory and reports:
Problem
In some cases, the VXML Server/Tomcat crashes and displays an OutOfMemoryError message.
These snippets are seen in the VXML server application error logs.
at com.audium.server.controller.AudiumServerConfiguration.
<init>(AudiumServerConfiguration.java:212)
Solution
When you run Customer Voice Portal (CVP) Version 9.0, make sure that ES11 is installed in order to resolve
many security vulnerabilities and issues that can lead to the OutOfMemoryError message. Refer to the
Multiple Vulnerabilities in Cisco Unified Customer Voice Portal Software field notice (FN) for more
information.
A similar ES to the one referenced in the FN has been released for CVP Version 8.5. Click the hyperlink in
order to visit the page where you can download CVP8.5.1_ES24.exe.
Increase Virtual Memory
In many cases, an increase of the virtual memory helps. In order to increase the virtual memory in CVP
Versions 8.5 and 9.0, complete these steps:
Windows 2003
1. Navigate to:
Start > Run > Regedit
HKLM\SOFTWARE\Apache Software Foundation\Procrun2.0\VXMLServer\
Parameters\Java\Options
2. Add the "−XX:MaxPermSize=256M" registry key in order to increase the Virtual Memory
PermSpace.
Windows 2008
1. Navigate to:
Start > Run > Regedit
HKEY_LOCAL_MACHINE\SOFTWARE\Wow6432Node\Apache Software
Foundation\Procrun2.0\VXMLServer\
Parameters\Java\Options
2. Add the "−XX:MaxPermSize=256M" registry key in order to increase the Virtual Memory
PermSpace.
There are several tools available for this, such as VisualVM, JProfiler, and Yourkit. Any of the profilers that
assess VXML applications provide details into which data structures accumulate and do not release memory,
which eventually leads to an OutOfMemory condition.
When garbage collection runs depends on the system load. A general rule is that a high−load system or a
system with a small amount of memory encounters garbage collection often whereas a low−volume system or
one with a large amount of memory runs garbage collection less frequently.
How does garbage collection decide what memory to reclaim from the applications?
This is assessed by the Java Virtual Machine (JVM). Any object becomes a candidate for garbage collection
once all references to it are discarded.