You are viewing a plain text version of this content. The canonical link for it is here.
Posted to users@tomcat.apache.org by Michael Oullion <th...@hotmail.com> on 2010/06/15 11:05:24 UTC
Perfomance analysis
hello,
I'm doing some benchmark on Web service's Call and I have some strange results.
First of all,
My web service is on a Axis2 1.5.1 framework,
I use a Apache Tomcat 6.0.20 / JRE (server mode) 1.6.0_20-b02.
I use JMeter 2.3.4 to do my benchmark.
When I analyze my result, I have some nice (and stable ) things but some times I observe a huge decrease of Hit/second and a huge increase of response time.
For example, normal response time is (avg) 300 ms and it increase to 1.000 ms and more, at the same time the hit/second decrease from 90 to 40.
I watch the CPU activity to see if the bottleneck is here and I observe that I have the same thing.
When the response time increase, the CPU activity fall down from 40% to 10%.
I watch disk time access too and I don't observe a link between this activity and the fall of performance.
I try to attach my graph to this mail but I don't know if it's okay on the mailing list.
I don't understand why some time the performance fall down.
So, if you are some ideas...
_________________________________________________________________
Your E-mail and More On-the-Go. Get Windows Live Hotmail Free.
https://signup.live.com/signup.aspx?id=60969
RE: Perfomance analysis
Posted by Michael Oullion <th...@hotmail.com>.
Oh sorry, I forgot to explain my architecture.
No, I have one dedicated server which host the tomcat and I use JMeter on a second server ( not dedicated ).
> From: andrew.bruno@gmail.com
> To: users@tomcat.apache.org
> Subject: Re: Perfomance analysis
> Date: Tue, 15 Jun 2010 19:53:56 +1000
> CC: users@tomcat.apache.org
>
> Is jmeter on same box as app?
>
> On 15/06/2010, at 7:05 PM, Michael Oullion <th...@hotmail.com> wrote:
>
> > hello,
> >
> > I'm doing some benchmark on Web service's Call and I have some
> > strange results.
> >
> > First of all,
> > My web service is on a Axis2 1.5.1 framework,
> > I use a Apache Tomcat 6.0.20 / JRE (server mode) 1.6.0_20-b02.
> > I use JMeter 2.3.4 to do my benchmark.
> >
> > When I analyze my result, I have some nice (and stable ) things but
> > some times I observe a huge decrease of Hit/second and a huge
> > increase of response time.
> > For example, normal response time is (avg) 300 ms and it increase to
> > 1.000 ms and more, at the same time the hit/second decrease from 90
> > to 40.
> >
> > I watch the CPU activity to see if the bottleneck is here and I
> > observe that I have the same thing.
> > When the response time increase, the CPU activity fall down from 40%
> > to 10%.
> > I watch disk time access too and I don't observe a link between this
> > activity and the fall of performance.
> >
> > I try to attach my graph to this mail but I don't know if it's okay
> > on the mailing list.
> > I don't understand why some time the performance fall down.
> > So, if you are some ideas...
> >
> > Your E-mail and More On-the-Go. Get Windows Live Hotmail Free. Sign
> > up now.
> >
> > ---------------------------------------------------------------------
> > To unsubscribe, e-mail: users-unsubscribe@tomcat.apache.org
> > For additional commands, e-mail: users-help@tomcat.apache.org
_________________________________________________________________
Hotmail: Powerful Free email with security by Microsoft.
https://signup.live.com/signup.aspx?id=60969
Re: Perfomance analysis
Posted by Andrew Bruno <an...@gmail.com>.
Is jmeter on same box as app?
On 15/06/2010, at 7:05 PM, Michael Oullion <th...@hotmail.com> wrote:
> hello,
>
> I'm doing some benchmark on Web service's Call and I have some
> strange results.
>
> First of all,
> My web service is on a Axis2 1.5.1 framework,
> I use a Apache Tomcat 6.0.20 / JRE (server mode) 1.6.0_20-b02.
> I use JMeter 2.3.4 to do my benchmark.
>
> When I analyze my result, I have some nice (and stable ) things but
> some times I observe a huge decrease of Hit/second and a huge
> increase of response time.
> For example, normal response time is (avg) 300 ms and it increase to
> 1.000 ms and more, at the same time the hit/second decrease from 90
> to 40.
>
> I watch the CPU activity to see if the bottleneck is here and I
> observe that I have the same thing.
> When the response time increase, the CPU activity fall down from 40%
> to 10%.
> I watch disk time access too and I don't observe a link between this
> activity and the fall of performance.
>
> I try to attach my graph to this mail but I don't know if it's okay
> on the mailing list.
> I don't understand why some time the performance fall down.
> So, if you are some ideas...
>
> Your E-mail and More On-the-Go. Get Windows Live Hotmail Free. Sign
> up now.
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: users-unsubscribe@tomcat.apache.org
> For additional commands, e-mail: users-help@tomcat.apache.org