You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@lenya.apache.org by Thorsten Scherler <th...@apache.org> on 2009/03/17 13:07:35 UTC
Jmeter tests show performance problems
Hi all,
I made some stress testing with the jmeter script we have in
src/modules/development
I disabled the part of the creating node and tested a couple of pages
from the default content of the default pub.
I tested a clean checkout, no modification at all.
The result is quite interesting, since we have a throughput for 5
concurrent user of ONLY 3.1 request/sec. I can remember that some
versions ago I did the test against formally known 1.4 and we had a far
better throughput.
Anybody has made stress testing with lenya lately and see the same
result as I do? If so any idea what we have changed that can caused this
bad performance?
salu2
--
Thorsten Scherler <thorsten.at.apache.org>
Open Source <consulting, training and solutions>
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@lenya.apache.org
For additional commands, e-mail: dev-help@lenya.apache.org
Re: Jmeter tests show performance problems
Posted by Andreas Hartmann <an...@apache.org>.
Thorsten Scherler schrieb:
> On Tue, 2009-03-17 at 13:38 +0100, Andreas Hartmann wrote:
>> Hi Thorsten,
>>
>> Thorsten Scherler schrieb:
>>> I made some stress testing with the jmeter script we have in
>>> src/modules/development
>>>
>>> I disabled the part of the creating node and tested a couple of pages
>>> from the default content of the default pub.
>>>
>>> I tested a clean checkout, no modification at all.
>>>
>>> The result is quite interesting, since we have a throughput for 5
>>> concurrent user of ONLY 3.1 request/sec. I can remember that some
>>> versions ago I did the test against formally known 1.4 and we had a far
>>> better throughput.
>> did you run the tests against the authoring environment? I think 3 req/s
>> is quite realistic for authoring.
>
> Yeah, we repeated the test against live and we got 11 but I need to review
> the test again tomorrow. If it is really 11 then IMO that is quite slow as well.
On my MBP 2.8 GHz I get 9 req/s for the homepage of the default
publication and 17 req/s for the tutorial page with a single thread.
Apparently the news aggregation has a significant impact on the performance.
-- Andreas
--
Andreas Hartmann, CTO
BeCompany GmbH
http://www.becompany.ch
Tel.: +41 (0) 43 818 57 01
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@lenya.apache.org
For additional commands, e-mail: dev-help@lenya.apache.org
Re: Jmeter tests show performance problems
Posted by Thorsten Scherler <th...@apache.org>.
On Tue, 2009-03-17 at 13:38 +0100, Andreas Hartmann wrote:
> Hi Thorsten,
>
> Thorsten Scherler schrieb:
> > I made some stress testing with the jmeter script we have in
> > src/modules/development
> >
> > I disabled the part of the creating node and tested a couple of pages
> > from the default content of the default pub.
> >
> > I tested a clean checkout, no modification at all.
> >
> > The result is quite interesting, since we have a throughput for 5
> > concurrent user of ONLY 3.1 request/sec. I can remember that some
> > versions ago I did the test against formally known 1.4 and we had a far
> > better throughput.
>
> did you run the tests against the authoring environment? I think 3 req/s
> is quite realistic for authoring.
Yeah, we repeated the test against live and we got 11 but I need to review
the test again tomorrow. If it is really 11 then IMO that is quite slow as well.
salu2
--
Thorsten Scherler <thorsten.at.apache.org>
Open Source <consulting, training and solutions>
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@lenya.apache.org
For additional commands, e-mail: dev-help@lenya.apache.org
Re: Jmeter tests show performance problems
Posted by Andreas Hartmann <an...@apache.org>.
Hi Thorsten,
Thorsten Scherler schrieb:
> I made some stress testing with the jmeter script we have in
> src/modules/development
>
> I disabled the part of the creating node and tested a couple of pages
> from the default content of the default pub.
>
> I tested a clean checkout, no modification at all.
>
> The result is quite interesting, since we have a throughput for 5
> concurrent user of ONLY 3.1 request/sec. I can remember that some
> versions ago I did the test against formally known 1.4 and we had a far
> better throughput.
did you run the tests against the authoring environment? I think 3 req/s
is quite realistic for authoring. It should be a lot better for the live
site, though.
-- Andreas
--
Andreas Hartmann, CTO
BeCompany GmbH
http://www.becompany.ch
Tel.: +41 (0) 43 818 57 01
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@lenya.apache.org
For additional commands, e-mail: dev-help@lenya.apache.org
Re: Jmeter tests show performance problems
Posted by ru...@uniklinik-freiburg.de.
Hi Thorsten,
> Hi all,
>
> I made some stress testing with the jmeter script we have in
> src/modules/development
>
> I disabled the part of the creating node and tested a couple of pages
> from the default content of the default pub.
>
> I tested a clean checkout, no modification at all.
>
> The result is quite interesting, since we have a throughput for 5
> concurrent user of ONLY 3.1 request/sec. I can remember that some
> versions ago I did the test against formally known 1.4 and we had a far
> better throughput.
>
> Anybody has made stress testing with lenya lately and see the same
> result as I do? If so any idea what we have changed that can caused this
> bad performance?
I didn't do strees testing, but I realized that requests to css files take
some time. (firebug)
This is because chaparon is used to generate the css files. After I have
disabled charparon, overwriting resources.xmap and resources-shared.xmap in
my publication
(thanks to the fallback protokoll) the performance was much better.
Maybe we can improve perfomance with a better cache handling.
Using chaparon for css generation seems to me a good thing, so I wouldn't
change this.
Rudolf
>
> salu2
> --
> Thorsten Scherler <thorsten.at.apache.org>
> Open Source <consulting, training and solutions>
>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: dev-unsubscribe@lenya.apache.org
> For additional commands, e-mail: dev-help@lenya.apache.org
>
>
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@lenya.apache.org
For additional commands, e-mail: dev-help@lenya.apache.org