Stupid question: did you try something else than JMeter? Here is what i get with a good warmup and ab:
โ( rmannibucau @ ubuntu )โ( 1.7.0 -:- 3.1.0 ) โ( /tmp/apache-tomee-plus-1.6.0-SNAPSHOT/SimpleGET )ยท> ab -g simple.tsv -n 100000 -c 300 http://localhost:8080/SimpleGET-1.0-SNAPSHOT/generic/ This is ApacheBench, Version 2.3 <$Revision: 655654 $> Copyright 1996 Adam Twiss, Zeus Technology Ltd, http://www.zeustech.net/ Licensed to The Apache Software Foundation, http://www.apache.org/ Benchmarking localhost (be patient) Completed 10000 requests Completed 20000 requests Completed 30000 requests Completed 40000 requests Completed 50000 requests Completed 60000 requests Completed 70000 requests Completed 80000 requests Completed 90000 requests Completed 100000 requests Finished 100000 requests Server Software: Apache-Coyote/1.1 Server Hostname: localhost Server Port: 8080 Document Path: /SimpleGET-1.0-SNAPSHOT/generic/ Document Length: 16 bytes Concurrency Level: 300 Time taken for tests: 19.746 seconds Complete requests: 100000 Failed requests: 0 Write errors: 0 Total transferred: 16900000 bytes HTML transferred: 1600000 bytes Requests per second: 5064.30 [#/sec] (mean) Time per request: 59.238 [ms] (mean) Time per request: 0.197 [ms] (mean, across all concurrent requests) Transfer rate: 835.81 [Kbytes/sec] received Connection Times (ms) min mean[+/-sd] median max Connect: 0 35 202.0 0 3007 Processing: 1 22 18.2 19 677 Waiting: 1 22 18.1 19 676 Total: 2 57 205.5 19 3058 Percentage of the requests served within a certain time (ms) 50% 19 66% 22 75% 23 80% 25 90% 34 95% 48 98% 1019 99% 1024 100% 3058 (longest request) To get some diagram from the output (simple.tsv) run: gnuplot simple.p with simple.p: set terminal jpeg size 1500,1000 set size 1, 1 set output "simple.jpg" set title "Benchmark testing" set key left top set grid y set xlabel 'requests' set ylabel "response time (ms)" set datafile separator '\t' plot "simple.tsv" every ::2 using 5 title 'response time' with lines exit then open simple.jpg JMeter doesn't handle very well concurrency *Romain Manni-Bucau* *Twitter: @rmannibucau <https://twitter.com/rmannibucau>* *Blog: **http://rmannibucau.wordpress.com/*<http://rmannibucau.wordpress.com/> *LinkedIn: **http://fr.linkedin.com/in/rmannibucau* *Github: https://github.com/rmannibucau* 2013/5/30 Caroline <[email protected]> > Hi all, > > I am trying to benchmark the dummy at /SimpleGET/generic in the following > project: http://cl.ly/2H1e1e3P2v0u > > To do so, I do not have the ideal conditions, but I believe I should get > far > better results: > - My Server: > 1.7Ghz i5 4GB MacBook Air at Java 7, TomEE+ 1.5.2 > directly connected to my Clients through UTP and working DHCP > > - My Clients: > 0 to 300 concurrent, in steps of 5 every 2 seconds, generated by the > Stepping Thread Group using JMeter (http://jmeter.apache.org) and > JMeterPlugins > ( > http://code.google.com/p/jmeter-plugins/downloads/detail?name=JMeterPlugins-1.0.0.zip > - place .jar in JMeter/lib/ext) using the Test Plan at > http://cl.ly/3C1a282P381h (pay attention to server IP). > 2.8Ghz i7 8GB iMac at Java 7 > > > My test web service should just do return "<go>return</go>";, however I > still have an error rate of 7 to 11% on 200000 samples generally starting > from around 150 concurrent requests. > > > I am at the end of my wits here. I believe I have dumbed down the test > webapplication far enough (my endgoal is to benchmark a more complex web > application). I believe I have simplified the network connection enough by > avoiding the complete network but directly connecting through UTP. I > believe > I have enough power on both my computers to have better results. I > researched the "connection reset by peer" erorr and I am using the > suggested > settings in my HTTP Request call in JMeter and I have simplified the > network > connection to my maximum. Previously, I was routing this traffic through a > DD-WRT WRT54G, same problems. > > > Thank you in advance, > > Caroline > > > > -- > View this message in context: > http://openejb.979440.n4.nabble.com/Load-testing-via-JMeter-connection-reset-by-peer-at-150-concurrent-threads-tp4663339.html > Sent from the OpenEJB User mailing list archive at Nabble.com. >
