Nortel reported this issue on their application.
################################

Hi Srini,

We saw high cpu again, but in this time the same application is running
on Solaris 9 and Java 1.5.0_09. I sent you the data that captured output
mdb I ran before using truss and after using truss. Truss and pstack and
output of prstat are also included.

The application is till running, so if you want to check something else,
send me instructions.
Same as before truss brought application from high cpu back to normal.

Thanks
 Jiayi
#####################################

The mdb, truss and pstack info is attached.
Older log info is available - The Radiance case number is 65855138

- Original filing
Hardware Platform: Netra 240
Component Affected: Java Software
OS and Kernel Version: [Please copy and paste output from uname -a]
# uname -a
SunOS rtp4cmt 5.10 Generic_118833-33 sun4u sparc SUNW,Netra-240
Describe the problem: [What issues are you experiencing?]
High cpu usage >80% by Java application. Our investigation found this is caused by the thread of Sun Corba, com/sun/corba/se/impl/orbutil/threadpool/WorkQueueImpl.requestWork().
 
[Did this function previously work properly?]
No, it is observed on Solaris 9 with the same JDK (1.5.0_09) once.
 
[What have you already tried to do?]
We collected output of prstat every 10 seconds and did pstack a few times.
 
[Copy and paste error messages from /var/adm or console log outputs.]
There is error message. When this happened whole cpu is occupied a few of instances of this tread, other threads didn't have chance to run. We tried to collect thread dump, but thread dump was not done until cpu usage was back to normal (<60%).
We don't know how and by what this is triggered.
This system is SRS Net Connect enabled: No
I will be sending an Explorer file: No
List steps to reproduce the problem(if applicable): We don't know how and by what this is triggered.
When was the problem first noticed: [Please give date and time.]
The first time the issue was reported is on May 15, 2007

The problem is: getting worse
Any changes recently?: None
What software is having the problem?: [Please provide product name and version.] name and version.]The JRE being used is 1.5.0_09
Desired start date for this service request: normal (according to service coverage)
Hi Srini,

Here is output of prstat that has MEM, but I couldn't find option for
SWAP.
    PID USERNAME SIZE RSS STATE PRI NICE TIME CPU PROCESS/NLWP

 10331 root 478M 352M sleep 59 0 6:29:23 0.9% java/185

Below is command line used for launching this application.

$JAVA_HOME/bin/java -cp $CLASS_PATH -DisActiveUnit="ACTIVE" -Dbefailover
ClassName="com.nortel.iems.server.common.IEMSBEFailOverManager"
-DIEMS_SSLINES_E
NABLED=$SSLINES_ENABLED -Dwebcontainer.port=$WEBCONTAINER_PORT
-Dhttp.port=$WEBS
ERVER_PORT -Dfqdn.name="$FQDN_NAME" -Dssl.port=$SSL_PORT
-Djavax.net.ssl.trustSt
ore=conf/server.truststore -Dcatalina.home=$TOMCAT_HOME
-Dmysql.home=$MYSQL_HOME
 -Dwebserver.port=$SSL_PORT
-Dresource_check=",$WEBCONTAINER_PORT,$TOMCAT_SHUTDO
WNPORT" -Dwebserver.rootdir=$WEBSERVER_HOME
-Dnms.server.host=$VIRTUAL_IP -Djava
.security.auth.login.config=$NMS_HOME/conf/jaas.conf
-DCRYPTO_CLASS=com.nortel.i
ems.common.IEMSCryptoGraphAPIImpl
-Djava.library.path=/opt/nortel/sspfs/lib/ -se
rver -Xms200m -Xmx200m -XX:+UseParNewGC -XX:+UseConcMarkSweepGC
com.adventnet.nm
s.startnms.NmsMainBE NMS_BE_PORT 2000 ROOT_DIR $NMS_HOME NMS_BE_FAILOVER
true $*
 >> $NMS_HOME/logs/startServerLog.txt 2>&1

This application allocates min and max 200 MB for heapsize

Thanks
 Jiayi
Nortel would like to know if something similar has been noticed.
I have asked them to try it on the latest JDK.
We have closed the case presently, will reopen if Nortel reopnes the issue.
Nortel asked to reopen this bug.