Mailing-List: contact users-help@tomcat.apache.org; run by ezmlm
Precedence: bulk
Reply-To: "Tomcat Users List" <users@tomcat.apache.org>
Received-SPF: pass (athena.apache.org: local policy)
From: Peter Crowther <Peter.Crowther@melandra.com>
To: 'Tomcat Users List' <users@tomcat.apache.org>
Date: Tue, 2 Sep 2008 17:49:48 +0100
Subject: RE: Installing Multiple Instances on Windows Server 2003
Thread-Topic: Installing Multiple Instances on Windows Server 2003
Thread-Index: AckM+M7L7ltr2ZG1Rhmbw0K/S+qepQABvfrQ
Message-ID: <6715CF65287F8F408DA109EC03AC6C0D07C9C65E9C@puma.melandra.net>
References: <19267576.post@talk.nabble.com> <19268553.post@talk.nabble.com>
 <6715CF65287F8F408DA109EC03AC6C0D07C9C65E9A@puma.melandra.net>
 <19269403.post@talk.nabble.com>
In-Reply-To: <19269403.post@talk.nabble.com>
Accept-Language: en-US, en-GB
Content-Language: en-US
acceptlanguage: en-US, en-GB
Content-Type: text/plain; charset="us-ascii"
Content-Transfer-Encoding: quoted-printable
MIME-Version: 1.0

> From: Steve G.B. [mailto:stefano.guidobaldi@yahoo.it]
> I've created a VM with 4 VCores, and all of the cores are
> allocated to the VM.

OK, so 50% CPU =3D 2 cores maxed out.  Out of interest, is it 25% with only=
 one Tomcat started?

> I'm using Sun JVM 1.6, and stressing the Guest with
> Loadrunner on another
> machine (if you ask: this machine with loadrunner isn't the
> bottleneck)
>
> No Databases, no I/O requests, no Network saturation.

OK.

Given that you're getting exactly 50% CPU use (it *is* exact, right?) that =
indicates the test harness is very unlikely to be the problem.  It would al=
most certainly bottleneck at some other CPU value.

> that's why I think it's the JVM.
> For my tests I used the standard demo webapp in Loarunner
> (Mercury Tours),
> and a couple of stupid jsp pages. So Apache and Tomcat both.

I'll highlight that to the folks who know the demo app better: does it run =
properly under load?  I'd assume so...

> Fun thing is that when using Tomcat and Apache combined, I
> can get an 80-85% CPU Utilization.

Yes.  If you've got httpd passing everything through it as well, you'll inc=
rease the CPU load - that's expected!

> Problem is that for my tests I need
> something more simple and the same server.

Yes.  Why add complexity when it's not required? :-)

> How do you saturate an 8-core host?

With some reasonably complex code in the JSPs :-).  Out of interest, if you=
r JSPs call something that loops a couple of million times before returning=
, what happens to the CPU use?

                - Peter

---------------------------------------------------------------------
To start a new topic, e-mail: users@tomcat.apache.org
To unsubscribe, e-mail: users-unsubscribe@tomcat.apache.org
For additional commands, e-mail: users-help@tomcat.apache.org