Ok, before moving on those ideas, we've made a small change to the
following:
VelocityContext velocityContext = null;
// LEP-6865
if (_innerVelocityContext == null) {
velocityContext = new VelocityContext();
_innerVelocityContext = velocityContext;
}
else {
velocityContext = new
VelocityContext(_innerVelocityContext);
}
velocityContext.put("processor", processor);
// Velocity variables
VelocityVariables.insertVariables(velocityContext, request);
Where "_innerVelocityContext" is a static instance.
Does this seem like a logical move?
My next questions is... IF the many "puts" made by
VelocityVariables.insertVariables
are actually utility classes and not runtime values, should these be
added to the "_innerVelocityContext" so that their introspection
meta-data is cached for subsequent use?
OR does this increase the likelihood of encountering concurrency issues?
Thanks,
Ray
On Thu, 2008-07-24 at 20:30 -0400, Raymond Auge wrote:
> Thanks Will, Malcolm, All,
>
> I'll take in these ideas and see how we can work them into a solution.
>
> Ray
>
> On Thu, 2008-07-24 at 16:39 -0700, Will Glass-Husain wrote:
>
> > One big problem is that you are using Velocity.evaluate() which is not
> > recommended for high-volume use.
> >
> > If you store your templates as text files and use a ResourceLoader to
> > retrieve as a Template object, Velocity has very efficient caching.
> > You can store your templates as files (use FileResourceLoader), in the
> > classpath (ClasspathResourceLoader), or in a database
> > (DatasourceResourceLoader).
> >
> > The (unreleased) Velocity 1.6 (in source control trunk) contains
> > further improvements. But you should see order of magnitude
> > improvements from retrieving your templates with a ResourceLoader (and
> > the caching option on) vs. Velocity.evaluate().
> >
> > See the developer's manual for more details.
> >
> > WILL
> >
> >
> > On Thu, Jul 24, 2008 at 3:46 PM, Malcolm Edgar <[EMAIL PROTECTED]> wrote:
> > > Hi Auge,
> > >
> > > On the Velocity development list there is some ongoing work to address
> > > this issue. I would recommend that you try the performance patches and
> > > get in contact with these people.
> > >
> > > https://issues.apache.org/jira/browse/VELOCITY-606
> > >
> > > regards Malcolm Edgar
> > >
> > > On Fri, Jul 25, 2008 at 7:53 AM, Raymond Auge <[EMAIL PROTECTED]> wrote:
> > >> Hello All,
> > >>
> > >> My name is ray Augé. I am an engineer working for Liferay, Inc. on the
> > >> Liferay Portlet project.
> > >>
> > >> We have a critical issue with Velocity performance where under large
> > >> load Velocity becomes a bottleneck for scalability.
> > >>
> > >> Now, I'm fairly certain that this is perhaps because we are using
> > >> Velocity in a way which may not leverage its scalability features.
> > >>
> > >> Allow me to explain the problem and our usage.
> > >>
> > >> Under heavy load we hit a max throughput and thread dumps during this
> > >> time are completely filled with BLOCKED threads as bellow:
> > >>
> > >> [snip]
> > >> "http-80-Processor47" daemon prio=10 tid=0x00002aabbdb90400 nid=0x5a59
> > >> waiting for monitor entry [0x0000000044c72000..0x0000000044c74a80]
> > >> java.lang.Thread.State: BLOCKED (on object monitor)
> > >> at
> > >> org.apache.velocity.util.introspection.IntrospectorBase.getMethod(IntrospectorBase.java:103)
> > >> - waiting to lock <0x00002aaad093d940> (a
> > >> org.apache.velocity.util.introspection.IntrospectorCacheImpl)
> > >> at
> > >> org.apache.velocity.util.introspection.Introspector.getMethod(Introspector.java:101)
> > >> at
> > >> org.apache.velocity.util.introspection.UberspectImpl.getMethod(UberspectImpl.java:165)
> > >> at
> > >> org.apache.velocity.runtime.parser.node.ASTMethod.execute(ASTMethod.java:184)
> > >> at
> > >> org.apache.velocity.runtime.parser.node.ASTReference.execute(ASTReference.java:203)
> > >> at
> > >> org.apache.velocity.runtime.parser.node.ASTReference.render(ASTReference.java:294)
> > >> at
> > >> org.apache.velocity.runtime.parser.node.SimpleNode.render(SimpleNode.java:318)
> > >> at org.apache.velocity.app.Velocity.evaluate(Velocity.java:322)
> > >> at org.apache.velocity.app.Velocity.evaluate(Velocity.java:195)
> > >> at
> > >> com.liferay.portlet.layoutconfiguration.util.RuntimePortletUtil.processTemplate(RuntimePortletUtil.java:238)
> > >> at
> > >> com.liferay.portlet.layoutconfiguration.util.RuntimePortletUtil.processTemplate(RuntimePortletUtil.java:194)
> > >> at
> > >> org.apache.jsp.html.portal.layout.view.portlet_jsp._jspService(portlet_jsp.java:801)
> > >> [/snip]
> > >>
> > >> Now here is what the Velocity usage looks like for that particular
> > >> thread:
> > >>
> > >> TemplateProcessor processor = new TemplateProcessor(
> > >> servletContext, request, response, portletId);
> > >>
> > >> VelocityContext vc = new VelocityContext();
> > >>
> > >> vc.put("processor", processor);
> > >>
> > >> // Wrap a bunch of puts...
> > >>
> > >> VelocityVariables.insertVariables(vc, request);
> > >>
> > >> // many more puts here...
> > >> vc.put(...);
> > >> vc.put(..);
> > >>
> > >> StringWriter sw = new StringWriter();
> > >>
> > >> try {
> > >> Velocity.evaluate(
> > >> vc, sw,
> > >> RuntimePortletUtil.class.getName(), content);
> > >> }
> > >> catch (Exception e) {
> > >> _log.error(e, e);
> > >>
> > >> throw e;
> > >> }
> > >>
> > >> // eventually do
> > >> return sw.toString();
> > >>
> > >> etc..
> > >>
> > >> Is there an obvious miss-use of Velocity than anyone can identify here?
> > >>
> > >> Wondering if we are supposed to re-use a common inner context? IF so,
> > >> would it be a singleton? a pool of re-usable contexts? something else?
> > >>
> > >> Many thanks for any input,
> > >>
> > >>
> > >> ----------------------------------
> > >> Raymond Augé
> > >> Software Engineer
> > >> Liferay, Inc.
> > >> Enterprise. Open Source. For Life.
> > >> ----------------------------------
> > >>
> > >> Liferay Meetup 2008 – Los Angeles
> > >>
> > >> August 1, 2008
> > >>
> > >> Meet and brainstorm with the creators of Liferay Portal, our partners
> > >> and other members of our community!
> > >>
> > >> The day will consist of a series of technical sessions presented by our
> > >> integration and services partners. There is time set aside for Q&A and
> > >> corporate brainstorming to give the community a chance to give feedback
> > >> and make suggestions!
> > >>
> > >> View Event Details
> > >>
> > >> Register Now
> > >>
> > >
> > > ---------------------------------------------------------------------
> > > To unsubscribe, e-mail: [EMAIL PROTECTED]
> > > For additional commands, e-mail: [EMAIL PROTECTED]
> > >
> > >
> >
> >
> >
>
> ----------------------------------
> Raymond Augé
> Software Engineer
> Liferay, Inc.
> Enterprise. Open Source. For Life.
> ----------------------------------
>
> Liferay Meetup 2008 – Los Angeles
>
> August 1, 2008
>
> Meet and brainstorm with the creators of Liferay Portal, our partners
> and other members of our community!
>
> The day will consist of a series of technical sessions presented by
> our integration and services partners. There is time set aside for Q&A
> and corporate brainstorming to give the community a chance to give
> feedback and make suggestions!
>
> View Event Details
>
> Register Now
----------------------------------
Raymond Augé
Software Engineer
Liferay, Inc.
Enterprise. Open Source. For Life.
----------------------------------
Liferay Meetup 2008 – Los Angeles
August 1, 2008
Meet and brainstorm with the creators of Liferay Portal, our partners
and other members of our community!
The day will consist of a series of technical sessions presented by our
integration and services partners. There is time set aside for Q&A and
corporate brainstorming to give the community a chance to give feedback
and make suggestions!
View Event Details
Register Now
