Ahhh, finally got to do the test on a capable machine, but very
overloaded with some db procecesses, dont take anything from the
absolute timing i know are very bad..,
For all the other, my test seem ok..
Server Software: Microsoft-IIS/5.0
Server Hostname: ciberlogic
Server Port: 80
Document Path: /examples/jsp/snp/snoop.jsp
Document Length: 713 bytes
Concurrency Level: 200
Time taken for tests: 39.937500 seconds
Complete requests: 10000
Failed requests: 0
Write errors: 0
Total transferred: 8909265 bytes
HTML transferred: 7130000 bytes
Requests per second: 250.39 [#/sec] (mean)
Time per request: 798.750 [ms] (mean)
Time per request: 3.994 [ms] (mean, across all concurrent
requests)
Transfer rate: 217.84 [Kbytes/sec] received
Connection Times (ms)
min mean[+/-sd] median max
Connect: 0 1 5.8 0 281
Processing: 78 791 4566.9 250 39718
Waiting: 15 766 4568.4 250 39718
Total: 78 792 4566.9 250 39718
Percentage of the requests served within a certain time (ms)
50% 250
66% 265
75% 265
80% 281
90% 312
95% 453
98% 812
99% 39656
100% 39718 (longest request)
Saludos ,
Ignacio J. Ortega
> -----Mensaje original-----
> De: Mladen Turk [mailto:[EMAIL PROTECTED]]
> Enviado el: 21 de junio de 2002 8:19
> Para: 'Tomcat Developers List'
> Cc: [EMAIL PROTECTED]; GOMEZ Henri
> Asunto: mod_jk testing (WAS RE: mod_jk 1.2.0 workers pool)
> Importancia: Alta
>
>
> First some things about testing:
>
> DO NOT USE the date.jsp (it could return various html lengths, so from
> there are coming those Failed requests ).
> I've just browsed through the source code of ab and it uses the first
> page as its length ethalon, and date.jsp could retur various length if
> the tests are finishing on the single second interval so
> there is 1 byte
> html length differnce.
>
> Use the /examples/jsp/snp/snoop.jsp instead (think it allways returns
> the same content-length).
>
> The ones that count are only the 'Non-2xx responses'
>
>
> > -----Original Message-----
> > From: Ignacio J. Ortega [mailto:[EMAIL PROTECTED]]
> > Sent: 20. lipanj 2002 22:37
> > To: 'Tomcat Developers List'
> > Subject: RE: mod_jk 1.2.0 workers pool
> >
> >
> > > -----Mensaje original-----
> > > De: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED]]
> > Enviado el: 20
> > > de junio de 2002 20:29
> >
> > I need time to prepare a machine that can handle this level
> > of concurrency, Mladen used -c 200, i need to switch to one
> > of the servers, :(, Thanks to microsoft, W2k professional
> > only can handle 10 simultaneus connections :((((..
> >
> It's a IIS thing I think.
>
> > I think Mlanden uses XP, right? which version ? Home,
> > Professional? do you know the exact number of concurrent
> > connections the XP version you use can handle?
> >
>
> Its Professional. And if I'm running the snoop.jsp with -c 200:
>
> 181857 [Thread-4] ERROR server.JkCoyoteHandler - Error in action code
> java.net.SocketException: Software caused connection abort:
> socket write
> error
>
> And,
>
> 181867 [Thread-4] ERROR common.ChannelSocket - Error, closing
> connection
> java.net.SocketException: Software caused connection abort:
> JVM_recv in
> socket input stream read
>
> Running with -c 10 doesn't produce such a error.
>
>
> MT.
>
>
--
To unsubscribe, e-mail: <mailto:[EMAIL PROTECTED]>
For additional commands, e-mail: <mailto:[EMAIL PROTECTED]>
