7:12 p.m.
Jay,
This is great work and very interesting to follow.
I would like to make a suggestion regarding the tests. Putting the
wiki under test is extremely valuable because it is a complex
application and therefore reflective of our users' applications.
However, it is tricky for others to follow along, because as you
mentioned, it is not trivial to setup. To supplement those tests,
perhaps you can generate a seam-gen application with a test schema
(perhaps stealing from one of the example application) so that we can
compare numbers with yours. What makes a seam-gen application valuable
to test is that it uses the Seam Application Framework, which is a
well known performance bottleneck in Seam because of its excessive use
of getInstance() and joinTransaction().
It shouldn't take too long to generate the seam-gen application and
reverse engineer the schema. The tedious part is generating a whole
bunch of fake data, but emacs or vi should make quick work of that.
-Dan
On Wed, Oct 1, 2008 at 6:02 PM, Jay Balunas <tech4j(a)gmail.com> wrote:
The example I was using was the seam wiki example, but that is not a
trivial
thing to set up. I have not tried the same profiling on the other seam
examples so I am not 100% sure that they will show the same behavior,
although I would think that they would.
As for the profiler report - what are you looking for exactly? Basically
under load (25 users) I noticed threads blocking on the
ViewHandlerWrapper.fillChain method. The details are below.
Regards,
Jay
On Wed, Oct 1, 2008 at 5:56 PM, Alexandr Smirnov <asmirnov(a)exadel.com>
wrote:
>
> Jay, can you send me profiler report or a reference application ?
>
> Jay Balunas wrote:
> > I was able to update and build with RichFaces 3.3.0-SNAPSHOT and saw
> > your changes.
> >
> > The good news is that I no longer see any blocking threads in the
> > "org.ajax4jsf.application.
> > AjaxViewHandler" - none at all which is great.
> >
> > The bad news is that the performance has not increased in any
> > meaningful or repeatable way.
> >
> > During my profiling I noticed a definite increase in the blocked
> > threads performing JNDI lookups. My theory is that the
> > AjaxViewHandler was acting as a gate and effectively slowing down
> > concurrent requests into Seam proper. These threads now fly straight
> > through and into Seam processing. They the end up blocked with the
> > other threads performing the lookups.
> >
> > I will continue to look at the lookup issues, and let you all know
> > what I find.
> >
> > -Jay
> >
> >
> > On Tue, Sep 30, 2008 at 2:05 PM, Jay Balunas <jbalunas(a)redhat.com
> > <mailto:jbalunas@redhat.com>> wrote:
> >
> > Excellent - I will build from sources and test the changes.
> >
> > I'll post the findings up on seam-dev.
> >
> > -Jay
> > ----- "Nick Belaevski" <nbelaevski(a)exadel.com
> > <mailto:nbelaevski@exadel.com>> wrote:
> >
> > > All,
> > >
> > > I've created JIRA issue for this optimization:
> > > https://jira.jboss.org/jira/browse/RF-4554 and have just fixed it.
> > >
> > > As from now ViewHandlerWrapper#fillChain method is not
> > synchronized
> > > anymore.
> > >
> > > Best regards,
> > > Nick Belaevski
> > >
> > > > -----Original Message-----
> > > > From: richfaces-exadel-bounces(a)redhat.com
> > <mailto:richfaces-exadel-bounces@redhat.com>
> > [mailto:richfaces-exadel- <mailto:richfaces-exadel->
> > > > bounces(a)redhat.com <mailto:bounces@redhat.com>] On Behalf
Of
> > Max Rydahl Andersen
> > > > Sent: Tuesday, September 30, 2008 10:32 AM
> > > > To: Jay Balunas; richfaces-exadel(a)redhat.com
> > <mailto:richfaces-exadel@redhat.com>
> > > > Subject: Re: [Richfaces-exadel] Fwd: [seam-dev] Some profiling
> > of
> > > > blockingthreads
> > > >
> > > > Sergey/Nik ? Any news on this one ?
> > > >
> > > > /max
> > > >
> > > > > Hi Guys,
> > > > >
> > > > > I have been doing some performance testing on seam
> > applications
> > > and have
> > > > > been able to post some of the results. One of the areas
that
> > I
> > > have
> > > > > investigated so far was blocking/waiting threads.
> > > > >
> > > > > In my example I am running 25 concurrent users and hitting
the
> > > seam wiki
> > > > > example. I am seeing blocked threads waiting on
> > > > > "org.ajax4jsf.application.AjaxViewHandler".
> > > > > I have investigated it a bit (see below), and have included
> > some
> > > of the
> > > > > discussions on the seam-dev list.
> > > > >
> > > > > I would like to get an opinion of the synchronized fillChain
> > > method, and
> > > > > see
> > > > > if there is another way to handle this.
> > > > >
> > > > > Thoughts?
> > > > >
> > > > > Thanks,
> > > > > Jay
> > > > >
> > > > > ---------- Forwarded message ----------
> > > > > From: Emmanuel Bernard <emmanuel(a)hibernate.org
> > <mailto:emmanuel@hibernate.org>>
> > > > > Date: Thu, Sep 25, 2008 at 10:01 AM
> > > > > Subject: Re: [seam-dev] Some profiling of blocking threads
> > > > > To: Jay Balunas <tech4j(a)gmail.com
<mailto:tech4j@gmail.com>>
> > > > > Cc: Pete Muir <pmuir(a)redhat.com
<mailto:pmuir@redhat.com>>,
> > "seam-dev(a)lists.jboss.org <mailto:seam-dev@lists.jboss.org>"
<
> > > > > seam-dev(a)lists.jboss.org
<mailto:seam-dev@lists.jboss.org>>
> > > > >
> > > > >
> > > > >
> > > > > On Sep 25, 2008, at 09:17, Jay Balunas wrote:
> > > > >
> > > > > org.ajax4jsf.application.AjaxViewHandler
> > > > >> --------------------------------------------
> > > > >> - This was not a large % of the blocked threads but
caught
> > my
> > > eye as a
> > > > >> potential trouble spot because it seemed directly related
to
> > > ajax
> > > > calls.
> > > > >> - This was easier to track down and the root cause is in
the
> > > > >> org.ajax4jsf.application.ViewHandlerWrapper class.
> > > > >> - see
> > > > >>
> > > >
> > >
> >
> > http://anonsvn.jboss.org/repos/richfaces/trunk/framework/api/src/main/java
> > > > /org/ajax4jsf/application/ViewHandlerWrapper.java
> > > > >> - The problem is all the calls to "fillChain"
which is a
> > > synchronized
> > > > >> method in the class.
> > > > >> - We need to find out how often the
"_initialized" field is
> > > false - is
> > > > >> this once a request, once for each component, etc...
> > > > >> - it appears to be called many times - I will look
further
> > into
> > > this.
> > > > >>
> > > > >>
> > > > > I will raise this the richfaces dev guys, and see what can
be
> > > done.
> > > > >
> > > > >
> > > > > Using the double check lock pattern with a volatile should
> > make
> > > things
> > > > > better. It has to be done right though :)
> > > > >
> > > > >
> > > > >
> > > >
> > > >
> > > >
> > > > --
> > > > /max
> > > >
> > > > _______________________________________________
> > > > Richfaces-exadel mailing list
> > > > Richfaces-exadel(a)redhat.com
<mailto:Richfaces-exadel@redhat.com>
> > > > https://www.redhat.com/mailman/listinfo/richfaces-exadel
> > >
> > > _______________________________________________
> > > Richfaces-exadel mailing list
> > > Richfaces-exadel(a)redhat.com <mailto:Richfaces-exadel@redhat.com>
> > > https://www.redhat.com/mailman/listinfo/richfaces-exadel
> >
> >
> >
> >
> > --
> > blog: http://in.relation.to/Bloggers/Jay
> > ------------------------------------------------------------------------
> >
> > _______________________________________________
> > Richfaces-exadel mailing list
> > Richfaces-exadel(a)redhat.com
> > https://www.redhat.com/mailman/listinfo/richfaces-exadel
>
--
blog: http://in.relation.to/Bloggers/Jay
_______________________________________________
Richfaces-exadel mailing list
Richfaces-exadel(a)redhat.com
https://www.redhat.com/mailman/listinfo/richfaces-exadel
--
Dan Allen
Software consultant | Author of Seam in Action
http://mojavelinux.com
http://mojavelinux.com/seaminaction
NOTE: While I make a strong effort to keep up with my email on a daily
basis, personal or other work matters can sometimes keep me away
from my email. If you contact me, but don't hear back for more than a week,
it is very likely that I am excessively backlogged or the message was
caught in the spam filters. Please don't hesitate to resend a message if
you feel that it did not reach my attention.