Hey Sebb, I have found the error. I had a counter which closed the connection after 10th iteration. hence it was failing from 11th thread.
Thanks, Nirmal On Wed, Dec 30, 2009 at 12:25 AM, nirmal kumar <[email protected]> wrote: > Hey thanks for your response... > > After digging through the logs I could find the cause for such > behavior... When ever i try to turn the thread pool to more than 10 > (say 11 thread ) I get a socket broken exception > "2009/12/30 00:00:39 WARN - > jmeter.protocol.tcp.sampler.TCPClientImpl: Write error > java.net.SocketException: Software caused connection abort: socket > write error > at java.net.SocketOutputStream.socketWrite0(Native Method) > at java.net.SocketOutputStream.socketWrite(SocketOutputStream.java:92) > at java.net.SocketOutputStream.write(SocketOutputStream.java:124) > at > org.apache.jmeter.protocol.tcp.sampler.TCPClientImpl.write(TCPClientImpl.java:69) > at > org.apache.jmeter.protocol.tcp.sampler.TCPSampler.sample(TCPSampler.java:305) > at > org.apache.jmeter.threads.JMeterThread.process_sampler(JMeterThread.java:346) > at org.apache.jmeter.threads.JMeterThread.run(JMeterThread.java:243) > at java.lang.Thread.run(Thread.java:619) > > 2009/12/30 00:00:39 WARN - jmeter.protocol.tcp.sampler.TCPClientImpl: > Write error java.net.SocketException: Software caused connection > abort: socket write error > at java.net.SocketOutputStream.socketWrite0(Native Method) > at java.net.SocketOutputStream.socketWrite(SocketOutputStream.java:92) > at java.net.SocketOutputStream.write(SocketOutputStream.java:124) > at > org.apache.jmeter.protocol.tcp.sampler.TCPClientImpl.write(TCPClientImpl.java:69) > at > org.apache.jmeter.protocol.tcp.sampler.TCPSampler.sample(TCPSampler.java:305) > at > org.apache.jmeter.threads.JMeterThread.process_sampler(JMeterThread.java:346) > at org.apache.jmeter.threads.JMeterThread.run(JMeterThread.java:243) > at java.lang.Thread.run(Thread.java:619) > > Hope the above exception explains the behavior for more jmeter more > response. But now this leads a new problem.. After some goggling i > found a reference > > http://old.nabble.com/TCP-Sampler-td3524656.html > > I have tried to set the tcp.eolByte=136 .. But still I get the same error .. > > Thanks, > Nirmal > > > On Tue, Dec 29, 2009 at 9:19 PM, sebb <[email protected]> wrote: >> On 29/12/2009, nirmal kumar <[email protected]> wrote: >>> Hi >>> >>> I have been running Jmeter for perfomace testing of an xmpp server. I >>> used to TCPSampler to send transaction messages. I have observed the >>> following behaviour : >>> >>> a) Thread Pool Increase: >>> To begin with i have tried to send messages using single thread >>> and measure the throughput ot be 10Messages per second. But as I >>> started to increase the number of thread still my throughput remain >>> the same. Can some explain me the cause for such behaviour. >> >> Perhaps the server cannot handle any more throughput. >> >>> I have measure the CPU, I/0 utilization too.. CPU - 0.4% and I/0=> >>> very minimal. >>> >>> hence inorder to increase the msg/sec i tried installing more jmeter >>> and I could see a increase the messages sent. >>> For 1 Jmeter ==> 10Msg/sec >>> For 3 Jmeter ==> 14.26 Msg/Sec >>> For 5 Jemter ==> 23.4 Msg/Sec >> >> How many threads for each JMeter? How long does each run last (approx)? >> >> It should not be necessary to use multiple instances of JMeter in most cases. >> >>> I would like to have a 300msg/sec. I don't want have too many >>> instances of jmeter machines. >>> Can some one help me in achieving it. >> >> It's possible to achieve much higher throughput with a single JMeter >> instance, so there must be something wrong here. >> >> How long do the TCP samples take? Do they complete successfully? >> >> The TCP Sampler is tricky to configure correctly, because there are >> lots of options - are you sure it has been set up correctly? >> >>> Thanks, >>> Nirmal >>> >>> --------------------------------------------------------------------- >>> To unsubscribe, e-mail: [email protected] >>> For additional commands, e-mail: [email protected] >>> >>> >> >> --------------------------------------------------------------------- >> To unsubscribe, e-mail: [email protected] >> For additional commands, e-mail: [email protected] >> >> > > > > -- > Regards > Nirmal > -- Regards Nirmal --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
