javax.sip.TransactionUnavailableException: Transaction not available -- could not acquire stack lock

[GoogleCodeExporter]

I am doing a some simple load testing using sipp 50calls/sec, but the server 
throws a lot javax.sip.TransactionUnavailableException: Transaction not 
available -- could not acquire stack lock error 

The load test script:
sipp $HOST -sf $1 -inf userIds.csv -m 90000 -r 100 -trace_err -l 1000 
-skip_rlim 

MSS-JBOSS: 1.7.0 on JBoss5
SIP Stack config:

gov.nist.javax.sip.LOG4J_LOGGER_NAME=gov.nist
gov.nist.javax.sip.TRACE_LEVEL=0
gov.nist.javax.sip.DEBUG_LOG=log/mss-jsip-debuglog.txt
gov.nist.javax.sip.SERVER_LOG=log/mss-jsip-messages.xml
gov.nist.javax.sip.SERVER_LOG_OVERWRITE=true
javax.sip.STACK_NAME=Mobicents-SIP-Servlets
javax.sip.AUTOMATIC_DIALOG_SUPPORT=off
gov.nist.javax.sip.DELIVER_UNSOLICITED_NOTIFY=true
gov.nist.javax.sip.THREAD_POOL_SIZE=256
gov.nist.javax.sip.REENTRANT_LISTENER=true
# prevent DOS attacks
gov.nist.javax.sip.MAX_LISTENER_RESPONSE_TIME=120
gov.nist.javax.sip.MAX_MESSAGE_SIZE=10000
# setting up the buffer size to reduce retransmissions and avoid loosing 
messages
gov.nist.javax.sip.RECEIVE_UDP_BUFFER_SIZE=327680
gov.nist.javax.sip.SEND_UDP_BUFFER_SIZE=327680
gov.nist.javax.sip.AGGRESSIVE_CLEANUP=true
gov.nist.javax.sip.MAX_FORK_TIME_SECONDS=0
#no cache specified since in non ha mode we don't replicate
org.mobicents.ha.javax.sip.cache.MobicentsSipCache.standalone=true
org.mobicents.ha.javax.sip.CACHE_CLASS_NAME=org.mobicents.ha.javax.sip.cache.NoC
ache
# set to true so that the stack do more validation on dialog,
# but won't work on chained applications so false by default
gov.nist.javax.sip.AUTOMATIC_DIALOG_ERROR_HANDLING=false

JVM tune up:
-Xms2048m -Xmx2048m -XX:MaxPermSize=512m  -XX:+UseLargePages 
-XX:+UseConcMarkSweepGC -XX:+CMSIncrementalMode -XX:-CMSIncrementalPacing 
-XX:CMSIncrementalDutyCycle=100 -XX:CMSIncrementalDutyCycleMin=100 
-Dorg.jboss.resolver.warning=true -Dsun.rmi.dgc.client.gcInterval=3600000 
-Dsun.rmi.dgc.server.gcInterval=3600000"

What else should I look? If I upgrade to 3.0 on JBoss 7, does it help?

Thanks lot.
Weiping

Original issue reported on code.google.com by wguo1...@gmail.com on 6 Aug 2014 at 1:50

[GoogleCodeExporter]

Sorry, the correct SIPP script is:
sipp $HOST -sf $1 -inf userIds.csv -m 90000 -r 50 -trace_err -l 1000 -skip_rlim

Original comment by wguo1...@gmail.com on 6 Aug 2014 at 1:58

[GoogleCodeExporter]

Please upgrade. we don't support older versions in the community, only in the 
corresponding commercial product TelScale from telestax.com

Monitor the memory usage and see if there is stop the world GC happening. You 
can also use a popular profiling tool such as jprofiler or yourkit and analyze 
if the issue comes from your application or the server.

Original comment by jean.deruelle on 18 Aug 2014 at 7:19

[GoogleCodeExporter]

Original comment by jean.deruelle on 17 Dec 2014 at 9:30

• Changed state: WontFix