Guest

Cisco ACE GSS 4400 Series Global Site Selector Appliances

Field Notice: Memory Leak in Global Site Selector (GSS) Software Version 1.2.1 Will Cause System to Fail Over Time


April 15, 2005


Products Affected

Product

Comment

GSS - GSS-4480-K9

Running GSS Software Version 1.2.1

GSS - GSS-4490-K9

Running GSS Software Version 1.2.1

GSS - GSS-4491-K9

Running GSS Software Version 1.2.1

Problem Description

In GSS software version 1.2.1, there is a memory leak in the Java Runtime Environment (JRE) that causes the system to lose memory over time. Eventually, one or more subsystems will fail and attempt to restart.

Background

There is a memory leak in the Java 1.4.2 runtime environment that causes the GSS Software 1.2.1 with this version of Java to lose memory over time.

Problem Symptoms

The GSS loses memory over time which eventually causes one or more subsystems to fail and attempt to restart.

The memory leak may be noticed when a user or system action causes a temporary spike in memory usage, such as configuring an object via the GUI.

The following conditions may determine whether or not the GSS is experiencing the problem:

CLI gss status shows one or more subsystems restarted and GSS reported a Java exception memory error in the log file, CLI operation failed, GUI operation failed, GSS failed to operate properly after a gss stop and gss start or other symptoms are possible.

The failure time is approximately 120 days since the software was last restarted. If the sustaining version v1.2(1.1.2) has not been loaded onto the system, then the failure time is approximately 40 days.

Note: GSS software versions 1.1(x.x.x) are not affected.

Workaround/Solution

Cisco recommends the customer upgrade to GSS Software Version 1.2.2 (registered customers only) where the Java runtime environment has been upgraded to 1.5.0, which is free of the memory leak.

A workaround prior to scheduling the upgrade would be to monitor memory usage with CLI command show processes. If process nodemgr reports a high percentage of memory usage exceeding 50 percent, then the system is at risk and should be reloaded with CLI reload command.

Example:

In this sample, the nodemgr process is using 1.9 percent of system memory and is not presently at risk of failure.

gss.example.com#show processes 

NAME PID MEM CPUTIME START 
---- --- --- ------- ----- 
system 18708 0.4 00:00:00 18:30 
database 18884 0.5 00:00:01 18:30 
tomcat 18887 1.6 00:00:03 18:30 
apache 19171 0.2 00:00:00 18:30 
crm 18894 2.8 00:01:40 18:30 
crdirector 18949 2.1 00:00:52 18:30 
dnsserver 18919 2.7 00:00:01 18:30 
keepalive 18951 0.1 00:00:02 18:30 
boomerang 18933 0.1 00:00:00 18:30 
proximity 18977 0.1 00:00:00 18:30 
sticky 18979 0.1 00:00:00 18:30 
nodemgr 18784 1.9 00:00:48 18:30

DDTS

To follow the bug ID link below and see detailed bug information, you must be a registered user and you must be logged in.

DDTS

Description

CSCeg72641 (registered customers only)

Tracks partial fix made and delivered in the sustaining image (GSS 1.2(1.1.2)) for this JRE memory leak issue. It slows the effect of the leak to about 120 days from 40 days.

CSCeh08531 (registered customers only)

Comprehensive fix for this JRE memory leak issue. Available in GSS 1.2(2)

For More Information

If you require further assistance, or if you have any further questions regarding this field notice, please contact the Cisco Systems Technical Assistance Center (TAC) by one of the following methods:

Receive Email Notification For New Field Notices

Product Alert Tool - Set up a profile to receive email updates about reliability, safety, network security, and end-of-sale issues for the Cisco products you specify.